Melt viscoelastic properties of peroxide cured polypropylene-ethylene octene copolymer thermoplastic vulcanizates.
Thermoplastic vulcanizates (TPVs) are prepared by the dynamic vulcanization process, where crosslinking of an elastomer takes place during its melt mixing with a thermoplastic polymer under high shear. The resulting morphology consists of micron sized finely dispersed crosslinked rubbery particles in a continuous thermoplastic matrix. TPVs have rubber like properties but can be processed as thermoplastic using conventional thermoplastic processing equipments (1-3). An extensive study of the TPVs based on various rubber-plastic blends were examined by Coran and Patel (4), (5). In the present study, polypropylene (PP) and ethylene octene copolymer (EOC) based TPVs were prepared by coagent assisted peroxide crosslinking system. With the addition of peroxide two competing reactions take place simultaneously: crosslinking in EOC and chain scission of PP. The latter causes lowering of molecular weight and narrow down the molecular weight distribution, hence melt becomes more Newtonian and viscoelasticity decreases. On the contrary, crosslinking and chain branching occur in EOC phase, which considerably change the viscoelastic properties. The mechanical and rheological properties of the blends depend not only on the constituent polymers but also on the morphology of the blend. TPVs consist of densely packed crosslinked rubber particles in thermoplastic matrix and their rheological properties can be compared to that of highly filled polymers (6). Furthermore, rheological study provides valuable information about the flow properties of the compound in the melt processing equipments.
Goettler et al. (7) was the first to study about the technical importance of the TPVs in terms of rheological characteristics using a capillary rheometer. Han et al. (8) described the comparative rheological study of PP, PP/ethylene propylene diene rubber (EPDM) uncrosslinked and dynamically crosslinked blends using various rheological instruments to measure steady shear flow, uniaxial extension, and oscillation flow properties of the compounds. Steeman et al. (9) reported that TPVs have a yield stress for flow and the value increases with increase in elastomer component in TPVs. An extensive study of rheological response of PP/EPDM blends and the TPVs with phenolic resin as curing agent was explored by Jain et al. (10). Rheological properties of various rubber-plastic uncrosslinked and dynamically crosslinked blends with special reference to the effect of blend ratio, shear rate, melt elasticity, and degree of crosslinking have been interpreted and described by several authors (11-13). Katbab et al. (14), (15) discussed the formation of morphology and correlated with the mechanical and rheological properties for the nonreinforced and reinforced (carbon black and silica fillers) TPVs. Chatterjee and Naskar (16) analyzed the rheological response of nanosilica filled PP based TPVs. Rheological behavior of dynamically and statically vulcanized blends of PP and styrene butadiene styrene triblock copolymer (SBS) with reference to the effect of various curing agents was studied by Ichazo et al. (17). The linear viscoelastic properties of PP/styrene ethylene butadiene styrene block copolymer (SEBS) blends and PP/EPDM TPVs are described in terms of composition ratio and also discussed about the oil diffusion and distribution in the blend components by Sengers et al. (18). Recently, Leblanc (19) studied the nonlinear viscoelastic property of the various commercially available TPVs with varying hardness by a Fourier transform rheometer (FT-rheo) to get insight into the morphology developed and the subtle role played by the extractible content such as oil and plasticizers. Li and Kontopoulou (20) studied the evolution of morphology and rheological response of PP/EOC TPVs with peroxide as curing agent and correlated with the gel point of the EOC phase. Generally, coagents are multifunctional vinyl monomers, which are highly reactive toward free radical either by addition reaction and/or by hydrogen abstraction. Chain scission also could be retarded by stabilizing the PP macroradicals by addition reaction across the double bond in the coagent. Our recent work had revealed the effect of different peroxides (21) and different coagents (22) on fixed blend ratio of PP/EOC TPVs. Parent et al. (23) studied the variation of molecular weight and branching distribution in PP obtained by the treatment of coagent assisted peroxide. Furthermore, the obtained product architecture was analyzed by the melt rheological characterization such as oscillatory shear and extensional deformation. Generally, the flow properties of the TPVs are influenced by various factors such as polymer characteristics (molecular weight, molecular weight distribution, degree of branching), crosslinked dispersed particles (size, degree of dispersion, crosslink density), interfacial interaction and the processing conditions.
In our previous paper (24), mechanical properties and morphology of the TPVs prepared by three different mixing protocols were investigated in detail. The present article aims to understand the melt viscoelastic property and modulus recovery in molten TPVs prepared by three different methods. The morphology of various TPVs was correlated with the observed dynamic rheological response. Initially, the discussion is focused on the fixed concentration (optimized concentration: 2 phr) of curative in TPVs in three different mixing methods. Furthermore, to get a better insight in the reinforcement mechanism, the dependence of dynamic functions was studied by varying the concentrations of curative.
The general purpose polyolefin elastomer Exact 5371 [specific gravity, 0.870 g/[cm.sup.3] at 23[degrees]C; comonomer (octene) content 13%; melt flow index (MFI), 5.0 g/10 min @ 190[degrees]C/2.16 kg], was obtained from Exxon Mobil Chemical Company. Polypropylene (PP; specific gravity, 0.9 g/[cm.sup.3] at 23[degrees]C; melt flow index, 3.0 g/10 min @ 230[degrees]C/2.16 kg) was obtained from IPCL, India. Dicumyl peroxide (DCP; Perkadox-BC-40B-PD) having an active peroxide content of 40% was obtained from Akzo Nobel Chemical Company, The Netherlands. Its half life time ([t.sub.[1/2]]) at 138[degrees]C is 1 hr and specific gravity is 1.53 g/[cm.sup.3] at 23[degrees]C. The coagent, triallyl cyanurate SR 507A; (specific gravity, 1.12 g/[cm.sup.3] at 23[degrees]C) was obtained from Sartomer Company.
Preparation of TPVs
All PP-EOC TPVs were prepared by melt blending of the components in the Haake Rheomix 600 s at a temperature of 180[degrees]C with a rotor speed of 70 rpm. The composition of PP/EOC TPVs employed in this study are shown in Table 1. TPVs were prepared by three different mixing methods, which were systematically represented in Fig. 1. In all the mixing methods, time to perform the dynamic vulcanization process was maintained for 4 min.
[FIGURE 1 OMITTED]
TABLE 1. Composition of TPVs. Component Phr Ethylene octene copolymer (EOC) 100 Polypropylene (PP) 50 Dicumyl peroxide (DCP) 0-5 (a) (varied) Triallyl cyanurate (TAC) 2.0 (a) Concentration of DCP was optimized at 2.0 phr (21), which corresponds to 3 milliequivalents concentration.
Prepared TPVs were designated as follow: the first letter denotes the method by which the TPVs are prepared, followed by the second letter and a number indicating the type of peroxide and concentration of corresponding peroxide, respectively. For example, PD2 corresponds to the composition of 2 phr concentration of dicumyl peroxide prepared by phase mixing method. After mixing, the blends were removed from the Haake chamber at hot condition and sheeted out in a two roll mill at room temperature (25[degrees]C) after a single pass. Sheets were then cut and pressed in a compression molding machine (Moore Press, Birmingham, UK) at 190[degrees]C for 4 min at 5 MPa pressure. Aluminum foils were placed between the mold plates. The molded sheets were then cooled down to room temperature under the same pressure.
To estimate the degree of crosslinking in the EOC phase in the TPVs prepared, the samples were subjected to a series of solvent extraction. All the measurements were done according to ASTM D 2765. About 2 g of TPV sample were first extracted for 24 h in hot xylene at 80[degrees]C to remove the uncrosslinked EOC phase and low molecular weight PP fragments. It seemed to be a qualitative estimation because of the restricted extraction of EOC phase in the presence of PP phase. The samples was completely dried after extraction and weighted. Then the initial loss can be calculated from the
[X.sub.c](%) = ([[W.sub.1]/[W.sub.0]]) x 100 (1)
[X.sub.c] = Percentage of residue after xylene extraction at 80[degrees]C
[W.sub.0] = Initial weight of the specimen and
[W.sub.1] = Weight after xylene extraction.
The dried sample was further extracted for 72 h in boiling xylene at 140[degrees]C to completely remove the PP and soluble EOC phase. The gel content can be calculated as
Gel(%) = 100 - (([[[W.sub.1] - [W.sub.h]) - [W.sub.PP]]/[W.sub.EOC]] x 100) (2)
[W.sub.1] = weight of the sample after xylene extraction at 80[degrees]C
[W.sub.h] = weight of the sample after xylene extraction at 140[degrees]C
[W.sub.PP] and [W.sub.EOC] = weights of PP and EOC in the PP/EOC TPVs, respectively.
Equilibrium solvent swelling measurements were carried out on the PP/EOC TPVs to determine the crosslink density of the EOC in presence of PP. The overall crosslink density was calculated using the modified Flory-Rehner equation (21), (25), (26). From the degree of swelling, the overall crosslink density was calculated relative to the (EOC + PP) phases and expressed as ([upsilon] + PP). The latter was done in order to avoid the need to correct for a part of the PP, being extracted as amorphous PP. A circular piece of 2 mm thickness was made to swell in cyclohexane for about 48 h to achieve equilibrium swelling condition. Initial weight, swollen weight, and de-swollen or dried weight were measured and substituted in the Flory-Rehner equation to get the values of crosslink density.
Phase morphology of the cryofractured samples of various blends was investigated by a JEOL JSM 5800 Digital Scanning Electron Microscope (SEM). All the blends were cryofractured in liquid nitrogen to avoid any possibility of phase deformation during cracking process. For uncrosslinked PP/EOC blends, low melting temperature EOC component was preferentially extracted by treating in hot xylene at 70[degrees]C for 30 min. For dynamically cross-linked PP/EOC TPVs, uncrosslinked PP was preferentially extracted by treating in hot xylene at 120[degrees]C for 60 min. All the treated samples were dried under vacuum oven at 70[degrees]C for 5 h to remove the traces of solvents. The dried samples were then gold sputtered and examined under SEM.
Melt rheological behavior of the TPVs were analyzed in a Rubber Process Analyser (RPA 2000, Alpha Technologies) equipped with biconical dies. All the specimens were directly loaded between the dies maintained at 180[degrees]C, and the test were carried out in both frequency sweep and strain sweep mode.
Frequency Sweep. The frequency was logarithmically increased from 0.33 to 33 Hz at a constant strain amplitude of 5%. A strain of 5% was selected to ensure that the rheological behavior was located in the linear visco-elastic region.
Strain Sweep. The strain amplitude sweep was performed from 1 to 1200% at a constant frequency of 0.5 Hz.
Modulus Recovery. To determine the recovery kinetics, each sample underwent the following test sequences: (a) frequency sweep, (b) strain sweep, (c) relaxation time of 2 min, (d) frequency sweep, (e) strain sweep, (f) relaxation time of 2 min, (g) frequency sweep, and (h) strain sweep. All the tests were carried at a temperature of 180[degrees]C. Average of three measurements was reported for each sample. Similar sequences were pursued by Osman and Atallah (27) in characterizing the viscoelastic property of calcium carbonate filled polyethylene.
RESULTS AND DISCUSSION
Gel Content and Overall Crosslink Density
Variation in the gel content values and the overall crosslink density values of TPVs with same concentration of peroxide prepared by three different methods arc shown in the Table 2. Gel content was measured in two stages for the samples prepared. In the first stage, PP was extracted at 80[degrees]C by hot xylene. It is believed that this can extract uncrosslinked EOC phase along with very low molecular weight PP fractions. During low temperature extraction (80[degrees]C) process, PP remains almost insoluble in xylene (may be very low molecular weight species were extracted out). However, at high temperature (140[degrees]C) soluble PP fractions are removed leaving only crosslinked EOC gels. Maximum value of gel content is observed for phase-mixed TPV and the least for preblended TPV at both stages of extraction. The differences can be roughly compared to the extent of crosslinking in the EOC phase. It can be concluded that, higher gel content in the PD2 indicates that crosslinking reaction of the EOC proceeds predominantly with less influence on the PP phase. Overall crosslink density values are in complete agreement with the gel content measurements. In equilibrium solvent swelling measurements, a significant contribution of both the crosslinking in EOC phase and the degradation in PP phase are coupled to give the values. This further confirms the limited extent of degradation in the PP phase with the higher crosslinking degree in EOC phase for PD2 TPV.
TABLE 2. Gel content and crosslink density of TPVs. Compound name [X.sub.c](%) (a) Gel (%) (b) CD2 78.3 [+ or -] 0.6 39.3 [+ or -] 1.1 PD2 97.2 [+ or -] 0.8 60.8 [+ or -] 1.2 SD2 90.5 [+ or -] 1.0 55.5 [+ or -] 1.0 Compound name Overall crosslink density Crosslink density ([upsilon] + PP) x ([upsilon]) x [10.sup.-4] [10.sup.4] (mol/ml) (c) (mol/ml) (d) CD2 2.27 [+ or -] 0.25 1.55 [+ or -] 0.2 PD2 5.24 [+ or -] 0.3 1.55 [+ or -] 0.2 SD2 3.57 [+ or -] 0.2 1.55 [+ or -] 0.2 (a) Percentage of residue after Xylene extract of TPVs at 80[degrees]C. (b) Gel content after boiling xylene extract of TPVs at 140[degrees]C (on samples previously extracted at 80[degrees]C temperature). (c) Equilibrium solvent swelling measurements in PP/EOC TPVs in cyclohexane at room temperature for 72 hr. (d) Crosslink density of only EOC with the same amount of peroxide as that in TPVs.
A schematic representation of morphology evolution of 50/100 PP/EOC uncrosslinked and dynamically vulcanized blends with same amount of peroxide prepared by three different methods is shown in Fig. 2 and corresponding SEM photomicrographs were shown in Fig. 3. It is interesting to note that in uncrosslinked blend, EOC forms the matrix phase with PP as dispersed one. During dynamic vulcanization, phase inversion occurs, that is, PP forms the matrix and crosslinked EOC form the dispersed phase. With the addition of peroxide, EOC gets cross-linked and PP undergoes degradation. This leads to increase in viscosity ratio, which plays an important role in driving the morphology of TPVs. The less viscous PP encapsulates the more viscous crosslinked EOC phase to minimize the mixing energy. In this particular blend system, proportion of EOC is twice than that of PP (50/100 PP/EOC), which results to form densely packed three-dimensional network structure. These crosslinked EOC particles are dispersed in the form of aggregates and/or agglomerates in the PP matrix. Furthermore, these EOC aggregates or agglomerates are embedded in the PP macromolecules via joint shell mechanism or segmental inter-diffusion mechanism (14), (15), (28), (29), (30). Addition of curative (peroxide and coagent) in PP/EOC blends results to give several reactions: crosslinking in EOC, degradation in PP, generation of in situ graft copolymer, limited extend of crosslinking in PP phase and formation of branched PP chains. However different methods have their own influence on the aforementioned reactions which had been discussed in detail in one of our earlier communications (24). Among the three different methods employed at equal concentration of curative dosage, the efficiency of crosslinking in the EOC phase apparently has an important effect on the evolution of morphology, i.e., the particle size. The finest particle size (better morphology) was observed for PD2 whilst CD2 exhibits the largest rubber particles (see Fig. 3). This may be attributed due to the high degree of crosslinking in EOC with minor effect on PP phase in the PD2. In contrast, smaller particle size is associated with higher surface area having higher probability of adsorption of PP chains on their surface. The PP chains are adsorbed on the surface of cross-linked EOC domains via physical and/or chemical bonding. Within measurement accuracy, no clear distinction between the particle size of PD2 and SD2 is seen. However. PD2 seems to exhibit somewhat lower particle size than SD2 at this particular concentration of peroxide. The morphological evolutions are in agreement with the gel content and overall crosslink density values.
[FIGURE 2 OMITTED]
[FIGURE 3 OMITTED]
Strain Dependence of Viscoelastic Behavior. Strain sweep test was carried out to characterize the strain dependence of viscoelastic properties of the samples and to determine the linear viscoelastic region (LVE). Figure 4a-c display the variation of storage modulus (G'). loss modulus (G") and loss factor (tan [delta] = G"/G') respectively, as a function of strain amplitude for 50/100 PP/EOC uncrosslinked and dynamically vulcanized blends. Irrespective of the different methods attempted, dynamically vulcanized blends show increased dynamic modulus and dynamic viscosity (also called as dynamic functions) than uncured blends. In general, both the modulus (storage and loss modulus) are increased by the process of dynamic vulcanization. However, increase in storage modulus was larger than loss modulus throughout the strain range. Since shear stress develops almost linearly with the strain amplitude, complex modulus-stress plot look quite similar to those given in Fig. 4a. Storage modulus of the sample (uncured and dynamically cured) exhibit a linear region at low strain and nonlinear region at high strain amplitude. Uncrosslinked blend shows a longer linear region up to around 70% strain followed by a nonlinear region. But a progressive nonlinear behavior, that is, decrease of dynamic functions with the increase of strain amplitude can been clearly seen for TPVs. Such behavior resembles to the so-called Payne effect of rubber highly filled with active fillers like carbon black and silica (31), (32). Principally, rheology of TPV can be analogically compared with that of the rubber filled with active fillers (6). According to Payne, nonlinearity is due to the disintegration of secondary agglomerates filler network structure. However, in case of TPVs, the secondary structure corresponds to the crosslinked EOC domains dispersed in the PP matrix in the form of aggregates and/or agglomerates. Indeed nonlinearity of TPVs can then be imagined to be associated with both the mechanism: disintegration of agglomerated EOC domains and debonding of crosslinked EOC domains from the PP matrix phase. In uncrosslinked blends, the nonlinearity at higher strain level is due to the disintegration or breakdown of trapped molecular entanglement network and/or debonding of macromolecular chains of PP anchored to the continuous EOC phase.
[FIGURE 4 OMITTED]
To understand the behavior of TPVs in the strain sweep test, different rheological parameters were observed and tabulated in Table 3. Among the three different methods practiced, PD2 has a high modulus at low strain with a very short linear viscoelastic range and CD2 shows prolonged flat region, indicating that higher strain is required for the onset of nonlinearity than the others. Storage modulus at low strain amplitude (2% strain) had the following order; phase > split < preblend and the reverse is true for modulus at high strain amplitude (1200% strain). It is expected that better interfacial interaction due to cross curing tendency of the PP-EOC phase leads to give higher modulus value at high strain amplitude for preblending method. It seems that main reason for this behavior lies in the morphology developed and the particle size of the dispersed phase. In case of filled polymer system, there is a direct consequence of the particle size and filler networking (magnitude of aggregates and agglomerates). This is due to the fact that the smaller the particle, the smaller is the inter-aggregate distance and hence higher is the probability towards the formation of filler network.
TABLE 3. Rheological parameters and values observed from the strain sweep experiments. First run Comp [G'.sub.0.sup.(a)] [G'.sub.[infinity].sup.(b)] PP-EOC 2.0 0.13 CD2 65.8 1.44 PD2 90.5 0.86 SD2 73.0 1.09 First run Comp [[gamma].sub.c.sup.(c)] tan [delta] = [1.sup.(d)] [DELTA] [G'.sup.(e)] PP-EOC 51.0 - 1.87 CD2 17.2 350.8 64.3 PD2 14.2 150.6 89.6 SD2 15.0 220.1 71.9 Second run Comp [G'.sub.0] [G'.sub.[infinity]] [[gamma].sub.c] PP-EOC 2.5 0.13 40.2 CD2 55.3 1.21 11.5 PD2 79.5 0.81 8.1 SD2 72.7 1.01 5.5 Second run Comp tan [delta] = 1 [DELTA]G' PP-EOC - 2.37 CD2 88.9 54.1 PD2 80.5 78.7 SD2 73.7 71.7 Third run Comp [G'.sub.0] [G'.sub.[infinity]] [[gamma].sub.c] PP-EOC 2.8 0.13 15.0 CD2 52.8 1.11 10.5 PD2 80.1 0.79 7.2 SD2 73.7 1.0 4.5 Comp tan [delta] = 1 [DELTA]G' PP-EOC - 2.67 CD2 75.0 51.7 PD2 80.4 79.3 SD2 62.3 72.7 Average of three measurements were reported. (a) Low strain storage modulus (modulus at 2% strain; kPa). (b) High strain storage modulus (modulus at 1200% strain; kPa). (c) Critical strain amplitude [strain at 90% [G'.sub.0]; 0.9 [G'.sub.0] (%)]. (d) Strain amplitude at G' = G" (crossover point; %). (c) [DELTA]G' = [G'.sub.[infinity]] - [G'.sub.0] (change in storage modulus value with strain amplitude sweep).
It was observed that deviation from the linear viscoelastic behavior of TPVs starts at lower strain than uncrosslinked blends. Transition point in the region where the linear to nonlinear viscoelastic behavior starts is defined as critical strain amplitude ([[gamma].sub.c]) and for uniformity it may be taken as the strain value at the storage modulus equal to 90% of the initial storage modulus (0.9 [G'.sub.0]). It is clear from the Fig. 4a and Table 3, phase mixed TPV shows lower critical strain amplitude followed by split addition and preblending TPV. It can be stated, generally that presence of secondary clusters increase the modulus and enhance the nonlinearity.
Strain dependency of loss modulus (viscous modulus) of uncrosslinked and dynamically crosslinked blends is presented in Fig. 4b. In contrast to the G' behavior of TPVs, G" shows a maximum at the critical strain amplitude region whereas uncrosslinked blends is indeed linear at low strain region followed by decrease in loss modulus value with increasing deformation. Dependence of loss modulus on strain amplitude depends on the rates of network breakdown and reformation as well as sliding of macrochains at the domain surface. The magnitude of loss maxima are substantially emphasized for the phase mixed TPVs. It may be related to the energy dissipation produced by the breakdown of crosslinked EOC aggregates (32).
By definition, loss factor tan [delta] is determined by both loss and storage moduli (tan [delta] = G"/G'). Uncrosslinked blends exhibit higher tan [delta] value in the entire strain amplitude range, which is due to the dominance of viscous nature. However, TPVs exhibit more elastic behavior at low strain region, that is, G' > G", with increasing deformation TPVs become progressively more viscous and two curves intersect at tan [delta] = 1 (G' = G"; Fig. 4c). The effect of smaller particle size and high degree of cross-linking in the EOC phase are manifested by a substantial shift of the intersection point to the low strain region. G-plot (G" vs. G') of the strain sweep data is shown in Fig. 4d. Generally, G' is related to the extent of filler network formation and G" to the breakdown and reformation of these structures. In uncrosslinked blends, there is a linear relationship between G' and G" in the strain region studied. In case of TPVs, a semicircular arc is observed which may be a direct manifestation of morphology developed. Similar semicircular arc is observed for the carbon black filled SBR rubber and smaller particles exhibited more pronounced arc (33). Phase mixed TPVs show more pronounced and bigger arc followed by preblended and split added TPVs. Variation in the TPVs prepared by different methods is associated with the different state and magnitude of crosslinked EOC network (aggregates and/or agglomerates) formation and reformation.
Frequency Dependence of Viscoelastic Properties. Dynamic frequency sweep tests were performed at fixed low strain amplitude (selected from linear viscoelastic region) to further study the network formation and microstructural changes. Figure 5a shows the logarithmic plot of dynamic functions (complex modulus G*, complex viscosity [eta]*) as a function of frequency ([omega]) for the uncrosslinked (TPO) and dynamically crosslinked blends (TPVs). As shown, [eta]* decreases with increase in [omega] for all the blends (TPO and all TPVs), which clearly demonstrates the pseudoplastic behavior and significant non-Newtonian properties (shear thinning behavior). Irrespective of different methods, TPVs exhibit higher viscosity values in the entire range of frequency than the uncrosslinked blends and a more prominent increase of [eta]* in the low frequency region and the curve tends to converge at the higher frequency.
[FIGURE 5 OMITTED]
Zero shear viscosity ([[eta]*.sub.0]) values, obtained experimentally at the [omega] = 0.1 Hz are shown in Table 4. The higher [[eta]*.sub.0] value is observed for phase mixed TPVs. Moreover, viscosity curve at low frequency can be fitted in the Power law equation.
TABLE 4. Rheological parameters and values observed from the frequency sweep experiments. First run Comp [G'.sub.0] (a) [[eta]'.sub.0] (b) n (c) PP-EOC 3006 0.944 -0.3132 CD2 139,000 43.58 -0.8210 PD2 235,000 73.92 -0.8941 SD2 169,000 53.11 -0.8412 Second run Comp [G'.sub.o] [[eta]'.sub.0] n PP-EOC 3490 1.097 -0.3362 CD2 126,000 39.56 -0.7931 PD2 190,000 59.75 -0.8594 SD2 158,000 49.62 -0.8051 Third run Comp [G'.sub.0] [[eta]'.sub.0] n PP-EOC 4630 1.40 -0.3735 CD2 116,000 36.54 -0.7829 PD2 183,000 57.61 -0.8556 SD2 153,000 48.02 -0.8051 Average of three measurements is reported. (a) Complex modulus at low frequency (0.05Hz; kPa). (b) Zero shear viscosity (complex viscosity at low frequency: Pa s). (c) Shear thinning component (slope of the straight line obtained by plotting log ([eta] *) and log ([omega]) between the frequency 0.1 and 1.0 Hz).
[eta]* = k[[omega].sup.n] (3)
where [eta]* is a complex viscosity, k is sample exponential factor, [omega] is the oscillation frequency and n is the shear thinning exponent. The value of n can be determined from the slope of straight line obtained in the low frequency region in the log ([eta]*) vs. log ([omega]) plot. The value of n increased drastically for the dynamically vulcanized blends, that is, more pseudoplastic behavior is obtained by dynamic vulcanization (Table 4). Higher moduli in the low shear rate region and more shear thinning behavior with increasing shear rate can afford to give better processing characteristics. This can be attributed due to the state and mode of three-dimensional network structure formed by the dispersed crosslinked EOC particles and the overlapping of the matrix phase, that is, PP macromolecules absorbed by physical and/or chemical crosslinking on the surface of domains. Thereby interface becomes less mobile, which leads to increase in the shear viscosity at low frequency region. As the frequency increases, the network structure, mainly aggregates/agglomerates tends to collapse and deform to exhibit higher shear thinning behavior (34). This implies that the crosslinked EOC domains could behave as the filler in the PP matrix. It was previously reported that, TPVs with high amount of rubber have been shown to behave like highly filled molten thermoplastic (7), (8), (14).
Frequency sweep at small amplitude does not significantly deform the microstructure of the complex fluid; it allows understanding the reinforcing effect. Figures 5a-c show the dependence of G', G", and tan [delta] as a function of angular frequency. Both the moduli increase with increase in [omega] for all the blends. In TPOs, G" is larger than G' in the low frequency region, demonstrating viscous nature of the compound. However, with increase in frequency the two moduli cross each other called crossover frequency ([[omega].sub.c]), that is, G' = G", denotes the transition from viscous to rubbery response. In uncrosslinked blend [[omega].sub.c] observed at higher frequency (30 Hz). But in case of TPVs, G' becomes large than G" over the whole range of experimental frequency, hence no [[omega].sub.c] is observed. Therefore the TPVs can be described as "visco-elastic solid." In other words, dynamically vulcanized samples show high elasticity in the experimental frequency range. In TPOs, loss factor tan [delta] is high and decays very fast with increase in [omega]. But TPVs show very low values in the entire range of frequency applied and the curve becomes almost flat, reflecting the presence of three-dimensional network structures and their influence on the moduli. In a nutshell, in contrast to TPOs, TPVs show a gradual increase in both the modulus and decrease in their [omega] dependency. In general, dynamic functions at low frequency is strongly depend on the physical network formation (hydrodynamic and secondary structure) by the inclusion, whereas at high frequency is dominated by matrix (along with hydrodynamic effect).
Modulus Recovery. Experiments on recovery of dynamic functions after the application of large strain amplitude perturbation were performed for all the samples investigated. Figure 6 shows that the comparative subsequent strain sweep results performed immediately after a relaxation time of 2 min and a frequency sweep (for clarity only phase mixed TPV is shown in Fig. 6). The typical qualitative feature of Payne effect, that is, decrease of G' and the appearance of G" max in the subsequent strain sweep test is observed. However, linear range decreased ([[gamma].sub.c] at lower strain amplitude) and both the moduli (G' and G") decrease with the formation of G" max at lower strain amplitude in the subsequent strain sweep test. The variation of modulus value at low and high strain amplitude in all the three strain sweeps are shown in Table 3. These evolutions are associated with the nonreversal deformation of the network formed by the disintegration of crosslinked EOC aggregates and/or agglomerates and also de-wetting of the EOC domains from the PP matrix as well as rupture of chain entanglements and chains connecting aggregates. This corresponds to the so called phenomenon Mullins effect or stress softening effect (35). It is important to note that all the dynamic stress softening is achieved in the second sweep (2nd run) and only minor effects take place in the subsequent sweep (3rd run). Furthermore, it is observed from the figures, that the linear complex modulus (modulus at 2% strain) of TPVs is reformed to a large extent (more than 90% of initial or linear modulus is recovered) in the subsequent strain sweep. In the literature, it had been shown that, when the calcium carbonate filled Polyethylene (highly loaded particulate filled polyethylene) is subjected to subsequent strain sweep (27), the onset of nonlinearity shifted to the higher deformation (strain %) and a drastic decrease was observed in the initial storage modulus. However, in case of TPVs the reverse is observed, that is, initial storage modulus is regained to large extent and the onset of linearity is shifted to the lower deformation.
[FIGURE 6 OMITTED]
Dependence of dynamic functions as a function of frequency for both the uncrosslinked and dynamically cross-linked blends, before and after strain sweep followed by relaxation time is plotted in Fig. 7a,b. A marginal reduction in the dynamic functions is observed only in the low frequency region and the curves converge at the higher frequency. It is worth noting that no Mullins effect or strain history effects occur between the 2nd and 3rd runs. The dynamic functions are almost similar in both 2nd and 3rd runs. These results strongly suggest that the EOC domains formed energetically elastic network structure. In other words, the network is not significantly affected by the strain sweep or all the strain effects are recovered during the 2-min relaxation time between the subsequent sweeps.
[FIGURE 7 OMITTED]
Reinforcement Mechanism. The factors which affect the rheological characteristics of the TPVs are size of the dispersed domain, state of dispersion, intrinsic characteristics of the blend components, volume fraction, interfacial interaction, and processing conditions (2), (7). Melt rheological characteristics of the TPVs can be analogically compared with the filled polymer composites (8). In general, active fillers (such as carbon black, silica) have the tendency to form agglomerates, especially at higher loading, leading to the formation of cluster or chain-like structure or network structure. Depending on the activity of the filler surface, the network could be of direct contact mode or joint shell mode (32). In case of fillers, well wetted by the polymer molecules (carbon black in hydrocarbon rubber) the filler network might be formed by a joint shell mechanism (32), (33). A schematic representation of joint shell mechanism is shown in Fig. 8. Because of the better interfacial interaction between the polymer and filler, the polymers molecules are adsorbed onto the surface of the filler either chemically or physically (33). Thereby the segmental mobility of the polymer molecules are restricted and the modulus might increase in the vicinity of the filler particle. Smaller particles have higher tendency to form filler network and have more immobilized rubber shell in comparison with the larger filler particle. Similarly, in TPVs, the difference between the surface energy of the PP matrix and the crosslinked EOC particles leads to the adsorption of the PP molecule onto the surface of the crosslinked EOC particles through segmental interfusion mechanism (28), (29). For better understanding and to avoid complication, discussion is narrowed down to only preblended and phase mixed TPVs.
[FIGURE 8 OMITTED]
To explore the reinforcement characteristics of the TPVs prepared, the concentration of the curative (peroxide) was varied in both preblending and phase mixing method and the corresponding gel content and overall crosslink density values are shown in Fig. 9. The modulus evolution and viscosity curve for the phase mixed TPVs are presented in Fig. 10a,b. We observe a decrease in dynamic functions (G* and [eta] *) at high concentration of peroxide (4.6 phr). As previously mentioned, phase mixing method involves the addition of EOC curative master batch to the molten PP to form TPVs. From the method of preparation, it can be visualized that high concentration of peroxide have more number of reactive species in the EOC phase which readily undergo crosslinking reaction before the optimum dispersion of EOC domains during mixing with molten PP. Furthermore, tightly crosslinked domains restrict the disintegration of crosslinked EOC particles which results to form coarsest crosslinked EOC particles. In case of preblending method, particle size continuously decreases with increase in concentration of peroxide combined with the significant lowering of molecular weight in the PP phase ([beta]-chain scission). A speculative model of morphology development in dynamically vulcanized blends prepared by preblending and phase mixing method as a function of concentration of peroxide is shown in Fig. 11. The flow behavior of the polymer blends can be well illustrated by the Power law model (Eq. 3). The Power law constants such as flow behavior index (n) and consistency index (k) have been calculated from the linear fit of the log-log plot of complex viscosity and the frequency (Fig. 10b) for the preblended and phase mixed TPVs as a function of peroxide concentration is shown in Table 5. All the samples show pseudo-plastic behavior, that is, n value is less than one for all the samples. However, n value is found to increase with increase in concentration of peroxide. It is well known that higher n value indicates a less pseudoplastic behavior. When compared at equal concentration of peroxide, preblended TPVs show more pseudoplastic nature than phase mixed TPVs. For example, n value of CD2 and PD2 was found to be 0.82 and 0.89, respectively. The consistency index is defined as the viscosity value at unit shear rate. It is observed that k value increases with an increase in peroxide concentration for the preblended TPVs. But, in case of phase mixed TPVs, k value increases with increase in concentration of peroxide up to 3.6 phr, beyond which k value tends to decrease. This indicates that PD4.6 sample shows lower resistance to flow.
[FIGURE 9 OMITTED]
[FIGURE 10 OMITTED]
[FIGURE 11 OMITTED]
TABLE 5. Flow behavior index (n) and consistency index (A) value for the preblended and phase mixed TPVs as a function of peroxide concentration. Compound ID n k (kPa [s.sup.n]) CD0.6 0.8012 7.24 CD2.0 0.8210 12.02 CD3.6 0.8562 14.45 CD4.5 0.8819 18.62 PD0.6 0.8662 13.1 PD2.0 0.894 15.4 PD3.6 0.905 19.5 PD4.5 0.913 7.76
It is believed that crosslinked EOC domains are dispersed in the form of aggregates and agglomerates. Furthermore, these aggregates and/or agglomerates are connected via joint shell mechanism to form three-dimensional network structure. It is inseparable in the given experimental window, to determine the network formation through direct bonding mode or indirect mode (through polymer chains bridging on the different domains). It is well known that, smaller the particle size, higher is the surface area, shorter the interaggregate distance and hence formation of more developed network via segmental inter-diffusion mechanism (36), (37). Larger crosslinked EOC particles account to the decrease in the G' and [eta] * in both strain sweep and frequency sweep tests. This invokes the possibility of less adsorbed PP molecular chains and also aid in the roll over of polymer chains over the coarser crosslinked EOC domains or aggregates when deformed in the shear mode. The increase in particle size also reduces the network formation by increasing the distance between the particles. Although the overall crosslink density and the gel content of the TPVs increases with increase in concentration of peroxide in phase mixed TPVs, dynamic functions decrease at higher concentration of peroxide. However, in case of preblending method, within the curative concentration limit, initial storage modulus value increases with increase in curative concentration. Therefore, it can be expected that network formation overrides the effect of degradation in PP matrix effect on the rheological measurements. These clearly emphasis that the primary factors which determine the melt viscoelastic characteristics of the TPVs are the particle size and their network formation via adsorption tendency.
Other important factor for the reinforcement mechanism is the chain bridging effect, that is, influence of matrix molecular weight on the property improvement (38). When the matrix chain length is sufficiently long, PP chains are immobilized by chain adsorption at one or several points along their chain lengths which affect their stress response on the dynamic experiments. In other words, the generated liable bonds alter the mobility and dynamic stiffness. High molecular weight species or long chain molecules efficiently transfer the stress from the crosslinked domains to the matrix. This contributes to the reduction in the interlayer slip between matrix and dispersed domains which gives rise to increase in dynamic function in frequency sweep and retraction of dynamic functions after subsequent strain perturbation (38). A schematic representation illustrating difference in binding the particles and the particle size of the TPVs prepared by preblending TPVs (conventional method) and phase mixed TPVs at equal concentration of curative (2 phr: optimized concentration) is shown in Fig. 12. It is very clear form Table 3 and Fig. 4a that PD2 shows higher modulus value at low strain and more nonlinear behavior. It is very clear from the morphological analysis that PD2 exhibits very small particle size whilst CD2 shows the largest. It can be stated that, Payne effect depends on the state and degree of network formation. The network formed via joint shell mechanism would be much less rigid which may disintegrate at a relatively lower level of applied strain amplitude (32), (33). Smaller crosslinked EOC particles have higher tendency to adsorb the PP chains on the surface and hence form more trapped PP chains by connecting the particles. The adsorbed chains on the surface of the domains via physical and/or chemical adsorption generally broaden the relaxation spectrum towards longer time which has been reflected in the higher dynamic modulus and viscosity in the frequency and strain sweep experiments (32), (36), (37).
[FIGURE 12 OMITTED]
Rheological and morphological characteristics of peroxide cured PP/EOC TPVs have been analyzed with special reference to the effect of blending sequence or mixing protocol. Dynamic rheological behavior of the samples was studied and correlated with the final morphology. Irrespective of blending sequence used, dynamically cured blends show higher dynamic modulus and viscosity than uncured blend. At this particular composition ratio, TPVs respond as a viscoelastic solid as a result of formation of three-dimensional network structure. Among the methods employed, phase mixed TPVs (PD2) shows higher dynamic modulus and more nonlinear behavior. Interpretation of results of increase in dynamic functions and the nonlinearity still remains a subject of debate. The concept of modulus recovery and network formation yields a good interpretation. The network structure formed by the dispersed phase mainly depends on the particle size and their interaction. Recovery test carried out by subsequent strain sweep and frequency sweep proved the presence of the memory effect (Mullins effect) and the existence of the energetically elastic network domains. Both polymer matrix characteristics and the dispersed phase characteristics are equally important in contributing to the overall dynamic properties of the TPVs. There exists a direct consequence between the particle size and the degree of crosslinking in the dispersed phase. The higher the degree of crosslinking in the dispersed phase, the smaller is the particle size. Lower the particle size (higher the surface area) better is the adsorption on the matrix phase which leads to form better and more developed network structure. When the matrix chain length is sufficiently short, lower is the efficiency to transfer the applied stress or strain between the aggregates and/or agglomerates and also to the matrix phase.
(1.) A.M. Gessler and W.H. Haslett (to Esso Research and Engineering Company), U.S. Patent 3,037,954 (1962).
(2.) G. Holden, N.R. Legge, and H.E. Schroeders, Thermoplastic Elastomer--A Comprehensive Review, Hanser, Munich, New York (1987).
(3.) K. Kocsis, Polymer Blends and Alloys, Marcel Dekker, New York (1999).
(4.) A.Y. Coran and R. Patel, Rubber Chem. Technol., 53, 141 (1980).
(5.) A.Y. Coran and R. Patel, Rubber Chem. Technol., 55, 116 (1982).
(6.) K. Naskar, Rubber Chem. Technol., 80, 504 (2007).
(7.) L.A. Goettler, J.R. Richwine, and F.J. Wille, Rubber Chem. Technol., 55, 1448 (1982).
(8.) P.K. Han and J.L. White. Rubber Chem. Technol., 68, 728 (1995).
(9.) P. Steemean and W. Zoetelief, Paper presented at ANTEC, Orland, USA (2000).
(10.) A.K. Jain, N.K. Gupta, and A.K. Nagpal, J. Appl. Polym. Sci., 77, 1488 (2000).
(11.) C.R. Kumer, S.V. Nair, K.E. George, Z. Oomen, and S. Thomas, Polym. Eng. Sci., 43, 1555 (2003).
(12.) P. Ezzati, I. Ghasemi, M. Karrabi, and H. Azizi, Iran. Polym. J., 17, 265 (2008).
(13.) B. Kuriakose and S.K. De, Polym. Eng. Sci., 25, 630 (1985).
(14.) A.A. Katbab, H. Nazockdast, and S. Bazgir, J. Appl. Polym. Sci., 75, 1127 (2000).
(15.) A.A. Katbab, S. Bazgir, and H. Nazockdast, J. Appl. Polym. Sci., 92, 2000 (2004).
(16.) K. Chatterjee and K. Naskar, Polym. Eng. Sci., 48, 1077 (2008).
(17.) M. Ichazo, M. Hernandez, J. Gonzalez, C. Albano, and N. Dominguez, Polym. Bull., 51, 419 (2004).
(18.) W.G.F. Sengers, P. Sengupta, J.W.M. Noordermeer, S.J. Picken, and A.D. Gotsis. Polymer, 45, 858 (2004).
(19.) J.L. Leblanc, J. Appl. Polym. Sci., 101, 4193 (2006).
(20.) Z. Li and M. Kontopoulou, Polym. Eng. Sci., 49. 34 (2009).
(21.) R.R. Babu, N.K. Singha, and K. Naskar, J. Appl. Polym. Sci., 113, 1836 (2009).
(22.) R.R. Babu, N.K. Singha, and K. Naskar, J. Appl. Polym. Sci. 113, 3207 (2009).
(23.) J.S. Parent, A. Bodsworth, S.S. Sengupta, M. Kontopoulou, B.I. Chaudhary, D. Poche, and S. Cousteaux, Polymer, 50, 85 (2009).
(24.) R.R. Babu, N.K. Singha, and K. Naskar, Personal Communication.
(25.) P.J. Flory and J. Rehner, J. Chem. Phys., 11, 512 (1943).
(26.) K. Naskar and J.W.M. Noordermeer, Rubber Chem. Technol., 77, 955 (2004).
(27.) M.A. Osman and A. Atallah, Polymer, 47, 2357 (2006).
(28.) F. Goharpey, A.A. Katbab, and H. Nazockdast, Rubber Chem. Technol., 76, 239 (2003).
(29.) F. Goharpey, A.A. Katbab, and H. Nazockdast, Polym. Eng. Sci., 45, 84 (2005).
(30.) G. Naderi, P.G. Lafleur, and C. Dubois, Polym. Eng. Sci., 47, 207 (2007).
(31.) G. Heinrich and M. Kluppel, Adv. Polym. Sci., 160, 1 (2002).
(32.) M.J. Wang. Rubber Chem. Technol., 71, 520 (1998).
(33.) J.B. Donnet, R.C. Bansal, and M.J. Wang, Carbon Black: Science and Technology, Marcel Dekker, New York (1993).
(34.) L. Cui, Z. Zhou, Y. Zhang, Yinxi Zhang, X. Zhang, and W. Zhou, J. Appl. Polym. Sci., 106, 811 (2007).
(35.) J. Rainier, C. Gauthier, L. Chazeau, L. Stelandre, and L. Guy, J. Polym. Sci. Part A Polym. Phys., 45, 286 (2007).
(36.) X. Xu, C. Gao, and Q. Zheng, Polym. Eng. Sci., 48, 656 (2008).
(37.) J. Frohlich, W. Niedermier, and H.D. Luginsland, Compos. A, 36, 449 (2005).
(38.) M.I. Aranguren, E. Mora, J.V. DeGroot, and C.W. Macosko, J. Rheol., 36, 1165 (1992).
R. Rajesh Babu, Nikhil K. Singha, Kinsuk Naskar
Rubber Technology Centre, Indian Institute of Technology, Kharagpur 721302, India
Correspondence to: Kinsuk Naskar; e-mail: firstname.lastname@example.org
Contract grant sponsor: Council of Scientific and Industrial Research (CSIR), New Delhi, India.
Published online in Wiley InterScience (www.interscience.wiley.com).
[C] 2009 Society of Plastics Engineers
|Printer friendly Cite/link Email Feedback|
|Author:||Babu, R. Rajesh; Singha, Nikhil K.; Naskar, Kinsuk|
|Publication:||Polymer Engineering and Science|
|Date:||Mar 1, 2010|
|Previous Article:||Nanocomposites of silver nanoparticle and dinonylnaphthalene disulfonic acid-doped thermoreversible polyaniline gel.|
|Next Article:||Thermal and mechanical properties of polyurethane rigid foam/modified nanosilica composite.|