Introduction

Polymer blending has been used as a successful method of developing new polymeric materials with the improved or even synergy of many specific properties for different applications. It is well recognized that controlling the morphology is the key role in producing polymer blends with prescribed rheological, physical and mechanical properties [1]. On the other hand, most polymer blends are incompatible and exhibit nonuniform morphology due to their high molecular weight and unfavorable interaction. This is the reason why during the last three decades numerous research works have been devoted to select strategies for compatibilization of polymer blend constituents in order to control and stabilize the morphology of polymer blend systems [2].

Polypropylene (PP) as a mostly used candidate for polymer blending is highly crystalline and commercially reliable. PP due to its low production cost, excellent processability, good mechanical properties, thermal, moisture and corrosion resistance and recyclability has been used for many applications such as packaging, automotive parts and in countless other areas of modern life. One major drawback of PP is its poor impact strength and brittle behavior at cryogenic temperatures [3]. The addition of a dispersed rubbery phase to PP matrix has been commonly used when impact resistance at low temperatures is important [4,5,6,7,8,9]. Ethylene propylene diene (EPDM) and ethylene propylene rubber (EPR) are the most common choices but recently metallocene-catalyzed ethylene/α-olefin copolymers (ECs) have become the preferred choices and offer benefits compared to other elastomeric polyolefins. Among ECs, octene-based ethylene α-olefin or ethylene octene copolymer (EOC) has received more attention [6, 10,11,12,13,14] and is continuously replacing EPDM in almost all fields [15,16,17]. Da Silva et al. [18], Wang et al. [19] and Svoboda et al. [20] found that the addition of EOC into PP greatly improved the impact resistance or notch toughness of PP matrix when compared with the conventionally used EPDM elastomer. Recently, Razavi et al. [21] have prepared EOC toughened PP by continuous electron-induced reactive processing (EIReP) and showed that EIReP improves tensile and impact properties of EOC toughened PP. Hu et al. [22] showed that β-modification and annealing would work synergistically on the low-temperature toughness of PP/EOC blends to toughen the blend and reduce the EOC content required for effective toughening over the temperature range tested. There are different reports in the literature regarding the miscibility of PP/EOC blends [4, 16, 23]. McNally et al. [16] observed that the PP/EOC blends are partially miscible up to 10 wt% EOC, while Kontopoulou et al. [4] showed a complete immiscibility for whole blend ratio of PP/EOC.

Among the previous studies on the PP/EOC blends, however, a suitable comprehension of the microstructure–rheology relationship and toughening mechanism are missing. To fill this void, we used a combination of rheology technique and morphological observation to determine the microstructure–rheology relationship in detail. To achieve an in-depth knowledge of this issue, we tried to unearth an interconnectivity mechanism to probe rheological results. To the best of our knowledge, this study is the first in which the intensified long-chain branching caused by in situ grafting was proposed as the main mechanism of droplet–droplet interconnectivity for justification of enhanced elastic response at low frequency in the rheological measurements. Interconnectivity mechanism was reported in a few investigations [24,25,26]. Rheology was the main technique we used to decipher the microstructure of this blend.

Experimental section

Materials

A commercial-grade polypropylene (HP525J) with Mw = 4 × 105 g mol−1, Mn = 5 × 104 g mol−1 and MFR = 3 g (10 min)−1 supplied by Jam Petrochemical Company, Iran, was used as a polymer matrix. According to the supplier, the molecular weight distribution of PP is bimodal. An ethylene octene copolymer (LC370) with Mw = 6.2 × 105 g mol−1, Mn = 3.2 × 105 g mol−1 and MFI = 3 g (10 min)−1 with a total octene content of 32 wt% purchased from LG Chem company, Korea, was used as an elastomeric impact modifier. The original EOC contained a small amount of long-chain branches.

Sample preparation

PP/EOC blend samples varying in blend ratio 20/80, 50/50, 65/35 and 80/20 were considered. All the samples were prepared by melt compounding using a 60-cm3 internal mixer (Brabender Plasticorder W50) at 190 °C and rotor speed of 90 rpm for 20 min. PP/EOC blend samples with different compositions are given in Table 1.

Table 1 Compositions of PP/EOC blends used in the present study
Full size table

Characterization

The melt rheological measurements were taken on samples using a rheometric mechanical spectrometer (Paar Physica UDS 200). The melt linear and nonlinear viscoelastic measurements were taken under purging of nitrogen by utilizing a parallel plate fixture with a diameter of 25 mm and a constant 1 mm gap under controlled shear deformation oscillatory mode. The frequency sweep tests were performed at frequencies varying from 0.04 to 1000 rad s−1 at 190 °C. The nonlinear rheological behavior of the samples was investigated by performing the start-up of steady shear flow test in which the samples were imposed to a constant shear rate and the transient stress was monitored with time. Each test was repeated 3 times to ensure reproducibility of the results.

Dynamic mechanical thermal analysis (DMTA) was carried out using dynamic thermal analyzer (Triton instrument TTDMA). The bar shape samples of about 27 mm × 10 mm × 1 mm were cut from compression molded sheets. The loss tangent (tanδ) was measured at the frequency of 1 Hz as a function of the temperature in the temperature range from − 80 to 100 °C. The constant heating rate of 5 °C min−1 was maintained.

Differential scanning calorimetry (DSC) measurements were taken using a PerkinElmer instrument at a heating and cooling rate of 5 °C min−1 under a dry nitrogen atmosphere. To draw a meaningful comparison, each thermograph was normalized by the weight of the sample. Prior to the DSC recording, the heat history of the samples was eliminated by heating them at a rate of 5 °C min−1 and holding them for 1 min at 190 °C, and subsequently cooled by 5 °C min−1 to 0 °C in DSC.

Fourier transform infrared (FTIR) spectra of the sample were collected on a Bruker (Vector 33).

The morphology of the blends was studied by scanning electron microscopy (SEM) on cryo-fractured surfaces of the samples using a MIRA\\TSCAN (IROST) electron microscope operating at 7.00 kV. In order to obtain better contrast between the two phases, the EOC phase was etched out by selective n-heptane solvent dissolution technique. The surface of the samples was gold-sputtered to avoid charging.

Results and discussion

Rheology

Figure 1 shows the results of storage modulus (G′) and loss modulus (G″) as a function of frequency (ω) and the corresponding weighted relaxation spectrum (H(λ)) for neat PP and EOC at 190 °C. The H(λ) of the samples was evaluated from dynamic moduli (G′, G″, ω) using the Physical Universal Software US-200. As can be observed, the PP shows a terminal behavior with slop almost equal 2 for G′ versus ω at a low frequency range (Rosian type behavior), whereas EOC exhibits a pronounced positive deviation with a slop of 1 at the same frequency range. The greater elastic response of EOC can be attributed to long relaxation mode associated with its long-chain branches. This is evidenced by two distinct relaxation time peaks shown by EOC in Fig. 1b. The decrease in the slope of G′ versus ω at low frequency range and longer relaxation modes caused by chain-branched polymer has been reported by other researchers [4, 27,28,29]. According to Tabatabaei et al. [29], the larger relaxation time peak observed for long-chain-branched polymer is due to changes in relaxation mechanisms from simple reptation to arm retraction. The long-chain branching of EOC is also revealed by results of transient stress in the start-up of steady shear flow experiment (see Figure S1).

Figure 1
figure 1

a Storage and loss modulus and b weighted relaxation spectrum of the PP and EOC samples at 190 °C

Full size image

Figure 2 shows the results of G′ versus ω and Cole–Cole plot obtained for the PP, EOC and their blends with different blend ratio (80/20, 50/50 and 20/80) measured at 190 °C. In order to better understand the rheological behavior of the blends, the results of variation of storage modulus with EOC concentration measured at low angular frequency (0.04 rad s−1) were compared with those values calculated using the logarithmic form of mixture law [29] (see in the inset of Fig. 2a). As it was expected, the G’ shows positive deviation from mixture law as a melt elastic response of G′ associated with elastic response of droplet deformation and/or interface that is greater for the blends in which EOC phase is in the form of droplet dispersed in the PP matrix compared to the blends in which PP is dispersed phase, e.g., the maximum positive deviation is exhibited by PP80/EOC20 blend. From these results and knowing the fact that EOC is more elastic, the larger elastic response of PP in the PP80/EOC20 can be attributed to additional elastic response. (The higher the melt elasticity of droplet, the lower the deformation-induced elasticity.) This positive deviation of G’ from mixture law is expected to tend to minimum values for the blends with co-continuous morphology in which the droplet deformation-induced elastic contribution diminishes. A similar trend of the melt linear viscoelastic changes with the blend ratio can be found from the Cole–Cole results shown in Fig. 2b. Figure 2c–e presents SEM micrograph of PP80/EOC20, PP65EOC35 and PP50/EOC50 blends. As shown in these figures, radius of EOC droplets size increases with enhancement of EOC contents and in the PP50/EOC50 blend approaches the coarsen morphology.

Figure 2
figure 2

a Storage modulus versus frequency for PP, EOC and their blends at 190 °C. The inset shows the storage modulus as a function of the blend compositions at a low frequency compared to the mixture law results. b Cole–Cole plot for PP, EOC and their blends at 190 °C. SEM micrograph of c PP80/EOC20, d PP65EOC35 and e PP50/EOC50 samples (dark areas indicate EOC phase that etched with normal heptane)

Full size image

In order to provide more insight into morphology and interfacial interaction of the blends and according to the degree of immiscibility between the PP and EOC phases (see Figure S2 and Table S1), the Palierne model was fitted to the linear viscoelastic results of the PP80/EOC20 and PP20/EOC80 blends. The Palierne model has extensively been used to describe the linear viscoelastic behavior of immiscible polymer blends with matrix-dispersed morphology. This model enables one to obtain valuable information about elastic response associated with droplet deformation and/or interfacial interaction. Also, this model has proved to be highly effective in describing the interphase formed in immiscible blends [30, 31]. Palierne model expresses the complex modulus \( \left( {G^{*} \left( \omega \right)} \right) \) of an immiscible blend as follows [30]:

$$ G^{*} \left( \omega \right) = G_{\text{m}}^{*} \left( \omega \right) \frac{{1 + 3\varphi_{\text{d}} H^{*} \left( \omega \right)}}{{ 1 - 2\varphi_{\text{d}} H^{*} \left( \omega \right)}} $$
(1)

where

$$ H^{*} \left( \omega \right) = \frac{{4\left( {\frac{\alpha }{R}} \right)\left[ {2G_{\text{m}}^{*} \left( \omega \right) + 5G_{\text{d}}^{*} \left( \omega \right)} \right] + \left[ {G_{\text{d}}^{*} \left( \omega \right) - G_{\text{m}}^{*} \left( \omega \right)} \right]\left[ {16G_{\text{m}}^{*} \left( \omega \right) + 19G_{\text{d}}^{*} \left( \omega \right)} \right]}}{{40\left( {\frac{\alpha }{R}} \right)\left[ {G_{\text{m}}^{*} \left( \omega \right) + G_{\text{d}}^{*} } \right] + \left[ {2G_{\text{d}}^{*} \left( \omega \right) + 3G_{\text{m}}^{*} \left( \omega \right)} \right]\left[ {16G_{\text{m}}^{*} \left( \omega \right) + 19G_{\text{d}}^{*} \left( \omega \right)} \right]}} $$
(2)

where the subscripts “m” and “d” refer to the matrix and dispersed phase, respectively, α is the interfacial tension, ϕd is the volume fraction of the dispersed phase, and R is the volume average radius of dispersed droplets.

As it can be observed in Fig. 3, the storage modulus of both blend samples at low frequency is higher than those predicted by the Palierne model with greater extent for the PP80/EOC20 blend. On the other hand, it has been shown that the droplet-induced elastic response is greater for the blends comprising Newtonian droplets in the viscoelastic matrix compared to blends containing low deformable viscoelastic droplets dispersed in Newtonian matrix [32]. Accordingly, the PP80/EOC20 blend with highly viscoelastic EOC droplet is expected to show lower droplet-induced elastic contribution compared to PP droplets in PP20/EOC80. However, we can surmise that there may be an additional mechanism for enhancing the elastic modulus of these blends with much stronger effect in the PP80/EOC20 blend. It is worth mentioning that such a positive deviation of storage modulus from Palierne model has also been reported by other researchers for polymer blend systems. For example, Choi et al. [33] observed a positive deviation from Palierne model for the PMMA/core–shell PBA blends with strong interfacial adhesion and related this to the contribution of the additional stress due to the interactions at the particle/matrix interface. Mostofi et al. [34] also observed a distinct positive deviation from Palierne model for PP/PET/SEBS ternary blend sample and attributed this to a core–shell-type morphology wherein PET core was encapsulated by highly elastic SEBS shell, and formed aggregates in PP matrix which were responsible for the additional elastic response. A positive deviation in low frequency has also been reported by Bousmina et al. [26]. They illustrated that whereas the Palierne model accounts quantitatively for the data obtained on PMMA/PS blends and PMMA/rubber blends at low rubber content, it does not even qualitatively predict the secondary plateau for concentrated PMMA/rubber blends. They illustrated that the deviation from the viscoelastic emulsion model could be analyzed in terms of particle connectivity and aggregation. As we reveal in the following, none of the aforementioned mechanism could be applicable to our samples.

Figure 3
figure 3

Comparison between the experimental data (storage modulus) and the Palierne model predictions for the a PP20/EOC80 and b PP80/EOC20 samples

Full size image

Figure 4 compares the results of storage modulus versus frequency of processed PP and EOC and their unprocessed ones at 190 °C. The decrease in the melt elasticity of PP after melt processing is due to the well-recognized thermo-mechanically induced chain scission process [35]. In contrary to PP, the melt processing of EOC led to an appreciable increase in its melt elasticity. A similar behavior was observed by Razavi et al. [21]. This observation, unlike Razavi, could not be caused by cross-linking since processed EOC was completely dissolved in n-heptane at 50 °C. It should be noted that as it is obtained for time sweep test in Figure S3, after melt processing, G′ both PP and EOC remain thermally stable up to 60 min, which is significantly higher than the time period required for rheological measurements.

Figure 4
figure 4

a Storage modulus versus frequency and b weighted relaxation spectrum for processed PP and EOC and their unprocessed ones at 190 °C

Full size image

Figure 4b presents the weighted relaxation spectrum of PP and EOC samples before and after processing at 190 °C. As it can be seen, two broad peaks of unprocessed PP shift to the two sharp peaks at shorter relaxation times (λ) after being melt processed as a result of the thermo-mechanically induced chain scission process. As it is depicted in Fig. 4b, the unprocessed EOC shows two distinct peaks so that after melt processing, the first peak (shorter λ) remains unchanged, while the peak due to the branched molecules (longer λ) becomes bimodal and shift to the long relation time range. By considering these results and knowing that EOC phase does not undergo cross-linking, shifting the second peak to the longer λ can be attributed to an intensified long-chain branch contribution. A similar result was reported for EOC sample by Svoboda and Poongavalappil [36,37,38]. They found that either increasing temperature or octene content in the EOC led to a decrease in cross-linking efficiency but resulted in a rise in chain branching. This was explained in terms of degradation through β-scission due to the presence of “labile hydrogen atoms.” Therefore, this suggests that in the blend samples containing both PP and EOC, chain scission and grafting process between thermo-mechanically induced PP and EOC macro-radicals are two competitive reactions [39,40,41]. From the above described rheological results, we could conclude that during the melt blending of PP with EOC, some macro-radicals generated from thermo-mechanical degradation of PP are grafted onto more stable macro-radicals of branched EOC, mainly across the interface. This will evidently result in chain-branched EOC with an intensified degree of chain branching with even longer grafted chains.

Moreover, the negative deviation in degree of crystallinity of PP and EOC components of blend samples shown in Table S1 can be induced by enhancement of steric hindrance and disordering in each component due to the creation of new intensified long-chain-branched EOC. Further insight into understanding the mechanism of radical grafting between PP and EOC macro-radicals was obtained by FTIR analysis (see Figure S4).

However, it is worth mentioning that in the case of PP80/EOC20 sample the positive deviation of storage modulus at low frequency range and/or Palierne model shown in Figs. 2 and 3 could hardly be only due to the grafting-induced interfacial enhancement. Figure 5 presents the weighted relaxation spectrum of PP, EOC, PP80/EOC20 and PP20/EOC80 samples measured at 190 °C. It can be seen that the relaxation pattern of PP20/EOC80 blend sample is similar to that of the EOC matrix and only the bimodal peak tends to a single peak. Because of the presence of PP macro-radicals in the PP20/EOC80, the formation extent of radical grafting-induced long-chain branching is expected to be greater than that in EOC sample which leads to the reducing difference between distinct distributions of EOC chains with long relaxation time. Conversely, for the PP80/EOC20 blend, can hardly find its similarity to that of the PP matrix. This pattern not only shows a broad peak due to the formation of radical grafting-induced long-chain branching but also presents a peak at very high relaxation time which may be attributed to the EOC droplet interconnectivity generated by bridging of grafted long-chain-branched molecules and inter-molecular entanglements formed between long-chain-branched and PP molecules. Therefore, the EOC droplet interconnectivity may be responsible for the additional elastic response in the PP80/EOC20 blend. This behavior has been also observed by Aoki et al. [42] for the dynamic viscoelastic behavior of molten ABS polymers in which dispersed rubber particles are narrow in size distribution. They found that the longtime relaxation spectra can be associated with the force acting between neighboring particles because in the longtime region the matrix AS copolymer is Newtonian; hence, its elastic properties must be due to entanglement between grafted AS copolymers. The interconnectivity mechanism in polymer blends was also reported by other researchers [24,25,26]. Results of DMTA experiment in Figure S5 and Table S2 can support our explanation regarding the interconnectivity mechanism of the PP80/EOC20 blend. The possibility of interconnectivity in PP80/EOC20 blend sample is schematically described in Scheme 1, and particles interconnections are illustrated by closed black curves. It should be noted that droplets interconnectivity is not appreciable for the PP50/EOC50 blend due to an almost coarsen type morphology. Note, however, that the suggestion of interconnectivity mechanism is still a mere speculation, and therefore, in order to provide more support for this proposition, this is expected to have super-toughening which can conceivably be confirmed using impact test in future research.

Figure 5
figure 5

Weighted relaxation spectrum of the PP, EOC, PP80/EOC20, and PP20/EOC80 at 190 °C

Full size image
Scheme 1
scheme 1

Schematic model for interconnectivity between the adjacent dispersed particles in PP80/EOC20. Yellow and blue areas are attributed to the PP phase and EOC phase, respectively. Long-chain branches of EOC are shown by the red line, and interconnectivity between EOC particles is illustrated by closed black curves; the bottom figure shows this mechanism clearly

Full size image

Conclusions

To sum up, in this research microstructure development and melt viscoelastic properties of PP/EOC blends with matrix-dispersed-type morphology varying in composition were investigated. It was demonstrated that the thermo-mechanically induced PP macro-radicals generated during the melt mixing can react with more stable EOC macro-radicals resulting in the formation of much longer-chain-branched EOC molecules mainly across the interface. These molecules could, in turn, generate a long range of molecular entanglements which were capable of creating EOC droplets interconnectivity.