Chemical Engineering Science 64 (2009) 4719 -- 4731
`
`Contents lists available at ScienceDirect
`
`Chemical Engineering Science
`
`journal homepage: w w w . e l s e v i e r . c o m / l o c a t e / c e s
`
`Rheological and thermal properties of blends of a long-chain branched polypropylene
`and different linear polypropylenes
`Seyed H. Tabatabaei, Pierre J. Carreau∗, Abdellah Ajji
`
`CREPEC, Chemical Engineering Department, Ecole Polytechnique, C.P. 6079, Succ. Centre ville, Montreal, QC, H3C 3A7 Canada
`
`A R T I C L E
`
`I N F O
`
`A B S T R A C T
`
`Article history:
`Received 28 August 2008
`Received in revised form 2 April 2009
`Accepted 3 April 2009
`Available online 14 April 2009
`
`Keywords:
`Polymers
`Polypropylenes
`Linear and branched polymer blends
`Rheological properties
`Thermal properties
`Miscibility
`Immiscibility
`
`Blends of a long-chain branched polypropylene (LCB-PP) and four linear polypropylenes (L-PP) having
`different molecular weights were prepared using a twin screw extruder. The linear viscoelastic properties
`suggested the immiscibility of the high molecular weight L-PP based blends, and the miscibility of the low
`molecular weight L-PP based blends. In addition, the Palierne emulsion model showed good predictions
`of the linear viscoelastic properties for both miscible and immiscible PP blends. However, as expected,
`the low-frequency results showed a clear effect of the interfacial tension on the elastic modulus of the
`blends for the high molecular weight L-PP based blends. A successful application of time–temperature
`superposition (TTS) was found for the blends and neat components. Uniaxial elongational properties were
`obtained using a SER unit mounted on an ARES rheometer. A significant strain hardening was observed
`for the neat LCB-PP as well as for all the blends. The influence of adding LCB-PP on the crystallinity,
`crystallization temperature, melting point, and rate of crystallization were studied using differential
`scanning calorimetry (DSC). It was found that the melting point and degree of crystallinity of the blends
`first increased by adding up to 20 wt% of the branched component but decreased by further addition.
`Adding a small amount of LCB-PP caused significant increase of the crystallization temperature while no
`dramatic changes were observed for blends containing 10 wt% LCB-PP and more. Furthermore, the crys-
`talline morphology during and after crystallization of the various samples was monitored using polarized
`optical microscopy (POM). Compared to the neat linear polymers, finer and numerous spherulites were
`observed for the blends and LCB-PP. Dynamic mechanical (DMA) data of the blends and pure components
`were also analyzed and positive deviations from the Fox equation for the glass transition temperature, Tg,
`were observed for the blends.
`
`Crown Copyright 䉷 2009 Published by Elsevier Ltd. All rights reserved.
`
`1. Introduction
`
`Due to the higher melting point, lower density, higher chemical
`resistance, and better mechanical properties of polypropylene (PP)
`in comparison to polyethylene (PE), it is widely employed for many
`industrial applications. However, the linear structure of L-PP limits
`its applications for processes where good extensional properties and
`melt strength are required such as thermoforming, film blowing,
`foaming and fiber spinning. On the other hand, it is well known that
`branched polymers have enhanced extensional properties and their
`blending with linear counterparts can improve their elongational
`behavior, particularly for PE (Ajji et al., 2003; Lohse et al., 2002; Steffl,
`2004). With the recent development of branched PP, it is expected
`that the elongational properties of L-PP can be effectively enhanced
`when blended with a long-chain branched polypropylene (LCB-PP).
`
`∗ Corresponding author. Tel.: +1 514 340 4711x4924; fax: +1 514 340 2994.
`E-mail address: pcarreau@polymtl.ca (P.J. Carreau).
`
`Long-chain branches are commonly introduced to linear PP via
`electron beam irradiation and post reactor chemical modification
`(Auhl et al., 2004; Tian et al., 2006a). Their effects on the processabil-
`ity have been reported in the literature (Gotsis et al., 2004; Stange
`and M¨unstedt, 2006). Gotsis et al. (2004) showed that branching up
`to an optimum level improved the processability in foaming and
`thermoforming processes, while further branching did not have a
`dramatic effect. Stange and M¨unstedt (2006) found that the strain
`hardening of branched polypropylenes caused not only a high melt
`strength, but also showed a significant homogeneity of deformation
`in the elongational experiments, which allowed forming foams with
`higher expansion ratios than L-PP.
`In several cases (Ajji et al., 2003; Stange et al., 2005; McCallum
`et al., 2007; Fang et al., 2008), the rheological behavior of blends of
`linear and branched polypropylene and polyethylene has been inves-
`tigated. It has been reported that branched polymers exhibit higher
`shear thinning, elasticity, and strain hardening compared to linear
`ones. Stange et al. (2005) showed that adding a small amount of
`LCB-PP significantly influences the rheological properties, especially
`
`0009-2509/$ - see front matter Crown Copyright 䉷 2009 Published by Elsevier Ltd. All rights reserved.
`doi:10.1016/j.ces.2009.04.009
`
`Page 1 of 13
`
`BOREALIS EXHIBIT 1044
`
`

`
`4720
`
`S.H. Tabatabaei et al. / Chemical Engineering Science 64 (2009) 4719 -- 4731
`
`the elongational behavior of linear PP blends. In addition, they found
`that the strain hardening of PP blends decreased as the strain rate
`increased while for the neat LCB-PP, enhancement of strain hard-
`ening was observed. McCallum and coworkers (2007) realized that
`the blending of branched and linear PP not only promoted the melt
`strength, but the mechanical properties increased as well. Ajji et al.
`(2003) showed that adding a small amount of low density polyethy-
`lene (LDPE) increased the strain hardening of linear low density
`polyethylene (LLDPE) resins. They also found that 10–20 wt% of LDPE
`is sufficient for improving the extensional property of LLDPE. Fang
`et al. (2008) concluded that an increase in the length of short
`branches and, possibly, the presence of a few long branches in
`metallocene LLDPEs and comparable molecular weights with LDPE
`could improve the miscibility of LLDPE/LDPE blends.
`To our knowledge no work has been published regarding the ef-
`fect of molecular weight of linear polypropylene on the rheological
`behavior of blends of linear and long-chain branched polypropy-
`lene. Using different rheological characterization methods, it will be
`shown that molecular weight has a crucial role on the miscibility of
`L-PP and LCB-PP. In addition, the influence of adding LCB-PP on the
`thermal properties, crystallization, and solid state behavior of the
`blends will be explored.
`
`2. Experimental
`
`2.1. Materials
`
`Four commercial linear polypropylenes (PP40, PP28, PP08, and
`PP04) and a commercial branched polypropylene (LCB-PP) were se-
`lected. The PP28 and PP08 were supplied by ExxonMobil Company
`and had a melt flow rate (MFR) of 2.8 g/10 min (under ASTM condi-
`tions of 230 ◦C and 2.16 kg) and 0.8 g/10 min, respectively, while the
`PP40, PP04, and LCB-PP were obtained from Basell Company and had
`a MFR of 4 g/10 min, 0.4 g/10 min, and 2.5 g/10 min, respectively. The
`main characteristics of the resins are shown in Table 1. The molecular
`weights of the L-PPs were evaluated from the relation between the
`zero-shear viscosity and the molecular weight (Fujiyama and Inata,
`2002). The molecular weight distribution (MWD) was measured us-
`ing a GPC (Viscotek model 350) with 1,2,4-Trichlorobenzene (TCB) as
`a solvent at a column temperature of 140 ◦C. The melting point, Tm,
`and the crystallization temperature, Tc, of the resins were obtained
`using differential scanning calorimetry. Blends containing 20, 40, and
`60 wt% LCB-PP were prepared using a twin screw extruder (Leistritz
`Model ZSE 18HP co-rotating twin screw extruder) followed by wa-
`ter cooling and pelletizing. The temperature profile along the barrel
`(from hopper to die) was set at 160/180/190/200/200/200/200 ◦C.
`The extrusion was carried out at 80 rpm. During blending, 3000 ppm
`of a stabilizer, Irganox B225, was added to avoid thermal degrada-
`tion of the polymers. To make sure that all samples have the same
`thermal and mechanical history, unblended components were ex-
`truded under the same conditions.
`
`2.2. Rheological measurements
`
`Dynamic melt rheological measurements were carried out us-
`ing a Rheometric Scientific SR5000 stress controlled rheometer with
`parallel plate geometry (diameter of 25 mm and a gap of 1.5 mm).
`All measurements were carried out at 190 ◦C under nitrogen atmo-
`sphere to avoid thermal degradation. Molded discs of 2 mm thick and
`25 mm in diameter were prepared using a hydraulic press at 190 ◦C.
`Time sweep tests were first performed at a frequency of 0.628 rad/s
`for 2 h. Material functions such as complex viscosity, elastic modulus,
`and weighted relaxation spectrum in the linear viscoelastic regime
`were determined in the frequency range from 0.01 to 500 rad/s. In
`order to obtain more accurate data, the frequency sweep test was
`carried out in four sequences while the amount of applied stress in
`each sequence was determined by a stress sweep test.
`To measure the uniaxial elongational viscosity, an ARES rheome-
`ter equipped with a SER universal testing platform from Xpansion
`Instruments was used. The model used was SER-HV-A01, which is a
`dual windup extensional rheometer. It is capable of generating elon-
`gational rates up to 20 s−1. Measurements were performed at 190 ◦C
`under nitrogen atmosphere.
`
`2.3. Thermal analysis
`
`Thermal properties of the various specimens were determined
`using a TA instrument differential scanning calorimeter (DSC) Q
`1000. The samples were heated from 50 to 220 ◦C at a heating rate
`of 10 ◦C/min to eliminate initial thermal history, and then cooled
`to 50 ◦C at the same rate. The melting point and the degree of
`crystallinity were determined from the second heating ramp, also
`performed at a rate of 10 ◦C/min. Crystallinity values are reported,
`using a heat of fusion of 209 J/g for fully crystalline polypropylene
`(PP) (Arroyo and Lopez-Manchado, 1997).
`
`2.4. Polarized optical microscopy (POM)
`
`Crystallization monitoring was performed using polarized opti-
`cal microscopy (OPTIHOT-2) to follow spherulites growth. For the
`non isothermal crystallization tests, films with a thickness of 50 ♯m
`were prepared using a twin screw extruder equipped with a slit
`die. The films were first heated on a programmable hot stage (Met-
`tler FP82HT) from room temperature to 200 ◦C at a heating rate of
`10 ◦C/min and were kept at that temperature for 1 min to eliminate
`initial thermomechanical history, and then cooled to room temper-
`ature at the same rate.
`
`2.5. Dynamic mechanical analysis (DMA)
`
`The solid state behavior of different samples was characterized
`using a TA instrument dynamic mechanical analyzer (DMA) 2980 in-
`side an environmental test chamber (ETC). The temperature ranged
`from −40 to 100 ◦C at a rate of 2 ◦C/min and frequency of 1 Hz was
`
`Table 1
`Main characteristics of the neat polymers.
`
`Resin code
`
`Company
`
`MFR (g/10 min)
`
`Nomencl.
`
`♪
`
`o
`
`a (kPa s)
`
`♪
`
`o
`
`b (kPa s)
`
`Mw (kg/mol)
`
`Mw/Mn
`
`Pro-fax 6523
`Basell
`4.0
`PP40
`9.8
`PP4612
`ExxonMobil
`2.8
`PP28
`14.4
`PP5341
`ExxonMobil
`0.8
`PP08
`43.6
`Pro-fax 6823
`Basell
`0.4
`PP04
`58.3
`Pro-fax 814
`Basell
`2.5
`LCB-PP
`18.6
`aZero-shear viscosity values obtained from the Carreau–Yasuda model, T = 190 ◦C.
`bZero-shear viscosity values obtained from the area under the weighted relaxation spectrum curves, T = 190 ◦C.
`
`9.8
`16.5
`49.5
`67.5
`19.5
`
`501
`543
`772
`812
`N/A
`
`2.8
`3.9
`2.7
`4.3
`2.3
`
`◦
`
`C)
`
`Tm (
`
`159.8
`161.0
`160.0
`159.6
`158.4
`
`◦
`
`C)
`
`Tc (
`
`119.2
`114.4
`117.3
`116.9
`128.4
`
`Page 2 of 13
`
`

`
`S.H. Tabatabaei et al. / Chemical Engineering Science 64 (2009) 4719 -- 4731
`
`4721
`
`applied to the rectangular shape samples. To generate low temper-
`atures and to control temperature during heating, liquid nitrogen
`was used. The glass transition temperature was determined from the
`maximum of the G(cid:4)(cid:4) curves.
`
`3. Results and discussion
`
`3.1. Rheological characterization of the neat PPs
`
`The complex shear viscosities normalized by the zero shear
`viscosities, obtained using the Carreau–Yasuda model (see Table 1),
`are plotted as a function of frequency for the neat PPs in Fig. 1.
`The LCB-PP exhibits a pronounced shear-thinning behavior due to
`the presence of long-chain branches. As the molecular weight of
`the L-PPs increases the behavior becomes more shear-thinning and
`the transition from the Newtonian plateau to the power-law region
`occurs at lower frequencies.
`The plot of the loss angle, ♥(tan ♥ = G(cid:4)(cid:4)/G(cid:4)), as a function of fre-
`quency is shown in Fig. 2. A monotonic decrease in the loss angle
`is observed for the L-PPs while the LCB-PP shows an inflection in
`the curve with a tendency towards a plateau at high frequencies.
`Wood-Adams et al. (2000) related the magnitude and breath of the
`plateau to the weight fraction of branched chains. The larger elastic-
`ity of the LCB-PP compared to L-PPs at low frequencies is attributed
`to more entanglements due to the presence of long-chain branches.
`However, as the frequency increases, the number of entanglements
`decreases due to the more shear-thinning character of the branched
`PP (see Fig. 1).
`To compare the relaxation behavior of the L-PPs and LCB-PP,
`the weighted relaxation spectra evaluated from dynamic moduli
`(G(cid:4), G(cid:4)(cid:4), ♽) using the NLREG (non linear regularization) software
`(Honerkamp and Weese, 1993) are plotted in Fig. 3 (the vertical
`dash lines represent the range of frequencies covered during the
`experiments). The area under the spectrum curve represents the
`zero-shear viscosity of the melt and it is reported in Table 1. A
`good agreement between these values and those obtained using the
`Carreau–Yasuda model is observed, suggesting that the relaxation
`spectra are accurate. It is obvious that the LCB-PP shows a longer
`relaxation time than the L-PPs, indicating that long-chain branches
`affect more the relaxation time than the larger molecules present in
`PP08 and PP04. The larger relaxation time for LCB-PP is attributed
`to changes in stress relaxation mechanisms. The simple reptation,
`as expected for linear polymers, no longer suffices to relieve stress
`
`Fig. 1. Normalized complex viscosity as a function of frequency for neat PPs;
`T = 190 ◦C.
`
`Fig. 2. Loss angle versus frequency for neat PPs; T = 190 ◦C.
`
`Fig. 3. Weighted relaxation spectra for neat PPs; T = 190 ◦C (the vertical dash lines
`represent the range of frequencies covered during the experiments).
`
`when there are enough branches present and slower events such as
`arm retraction must occur.
`
`3.2. Rheological characterization of the blends
`
`The complex shear viscosities as a function of frequency for the
`PP40/LCB-PP and PP04/LCB-PP blends (two extreme cases) are shown
`in Figs. 4(a) and (b), respectively. It is clear that adding the LCB-PP
`causes a pronounced shear-thinning behavior due to the presence
`of long-chain branches (the power-index calculated for frequencies
`ranging from 0.1 to 10 rad/s is reported in the legend). In addition,
`it is obvious that the viscosities of the PP40 based blends are in-
`termediate between those of the neat components while those of
`the PP04 based blends are closer to that of the PP04. The behav-
`ior is typical of linear polymer melts and the complex viscosity of
`the blends follows the log additivity rule for low molecular weight
`based blends. This is shown in Figs. 5(a) and (b), where the com-
`plex viscosities at different frequencies are plotted as a function of
`the LCB-PP content. The logarithmic additivity rule is expressed as
`
`Page 3 of 13
`
`

`
`4722
`
`S.H. Tabatabaei et al. / Chemical Engineering Science 64 (2009) 4719 -- 4731
`
`Fig. 4. Complex viscosity as a function of frequency for: (a) PP40/LCB-PP blends and
`(b) PP04/LCB-PP blends; T = 190 ◦C.
`
`(Utracki and Schlund, 1987):
`(♽) =
`♢ log(♪∗
`(♽))1 + (1 −
`♢) log(♪∗
`log ♪∗
`(♽))2
`(1)
`where
`♢ is the branched PP content in weight percent and ♪∗ is
`the complex shear viscosity. The PP40 based blends obey the log
`additivity rule, as depicted in Fig. 5(a), suggesting miscibility of both
`PP components. However, significant deviations from this empirical
`relation are observed in Fig. 5(b) for the PP04 based blends in the
`entire composition range, suggesting that these two PP parts are
`immiscible. Deviations from the log mixing rule for blends of linear
`and branched PPs and PEs have been reported in the literature (Fang
`et al., 2008; Ho et al., 2002; Liu et al., 2002). It is believed that not only
`the amount of LCB-PP influences the miscibility of the blends, but also
`the molecular weights of both components as well as the branching
`structure (e.g. star-like or tree-like) of the LCB-PP are important
`factors affecting the miscibility.
`The zero-shear viscosity obtained from the Carreau–Yasuda
`model is shown as a function of the LCB-PP content in Fig. 6.
`Good agreement with the log additivity rule for the blends hav-
`ing components with close melt flow indexes (i.e. PP40/LCB-PP
`and PP28/LCB-PP blends) is found, suggesting miscibility of the PP
`components. However, large deviations from the empirical rule are
`observed for the PP04/LCB-PP and PP08/LCB-PP blends, suggest-
`ing that the two components are immiscible. It should be noted
`that at low frequencies (Newtonian region), a large increase in the
`
`Fig. 5. Complex shear viscosity at different angular frequencies as a function of
`branched PP content for: (a) PP40/LCB-PP blends and (b) PP04/LCB-PP blends;
`T = 190 ◦C (the dash lines show the additivity rule).
`
`number of entanglements due to the LCB-PP addition is expected
`for all the samples. Therefore, it is unlikely that the deviations from
`the log mixing rule could be explained for the PP04/LCB-PP and
`PP08/LCB-PP blends only.
`To see the effect of molecular weight of L-PP on the relaxation
`behavior of the blends, the weighted relaxation spectra of the
`PP40/LCB-PP and PP04/LCB-PP blends are illustrated in Figs. 7(a)
`and (b), respectively. The addition of LCB-PP changes the relaxation
`mechanism from simple reptation to arm retraction, which retards
`the movement of chains along their backbone; hence, the maxima
`in the curves shift to longer times and the spectrum shape becomes
`broader. Note that for the PP40/LCB-PP blends, the positions of the
`peaks are intermediate to those of the neat components, indicating
`again miscibility, while for the PP04/LCB-PP blends non uniform
`changes in the peaks are observed.
`The behavior can also be analyzed using Cole–Cole plots of ♪(cid:4)(cid:4)
`versus ♪(cid:4), as illustrated in Fig. 8. The semicircular shape for the
`PP40/LCB-PP blends (Fig. 8(a)) is another evidence of miscibility
`(Kwang et al., 2000; Utracki, 1991), while some synergistic effects for
`the PP04/LCB-PP blends (Fig. 8(b)) are observed. It should be noted
`
`Page 4 of 13
`
`

`
`S.H. Tabatabaei et al. / Chemical Engineering Science 64 (2009) 4719 -- 4731
`
`4723
`
`Fig. 6. Zero-shear viscosity as a function of LCB-PP content; T = 190 ◦C (the dash
`lines show the additivity rule).
`
`∗m
`
`∗B
`
`(2)
`
`(3)
`
`m(♽)+19G∗m(♽)+2G∗d(♽)+G∗m(♽))+(3G∗ d(♽))(16G∗ d(♽))
`
`v
`
`Fig. 7. Weighted relaxation spectrum for:
`(a) PP40/LCB-PP blends and (b)
`PP04/LCB-PP blends; T = 190 ◦C (the vertical dash lines represent the range of fre-
`quencies covered during the experiments).
`
`achieved for ♡/Rv = 0 while for (80/20) PP04/LCB-PP blend, the best
`fit was obtained for ♡/Rv = 100. A value of the interfacial tension
`equal to 0 for the (80/20) PP40/LCB-PP blend is indicative of miscibil-
`ity. However, the non zero value of ♡/Rv for the (80/20) PP04/LCB-PP
`blend is indicative of the presence of a dispersed phase. Assuming
`an interfacial tension of 0.1 mN/m for the PP pair (PP04/LCB-PP),
`the corresponding droplet radius is estimated to be around 1 ♯m.
`Hence, the hypothesis of immiscibility for these blend components is
`reasonable.
`The role of adding the long-chain branched component on the
`temperature sensitivity of complex viscosity of the blends is exam-
`ined via the time–temperature superposition (TTS) principle. The
`results for the unblended polymers as well as for two blends are
`depicted in Figs. 11(a) and (b), respectively (to facilitate the compar-
`ison between data, the curves have been shifted by a multiplication
`factor as indicated). The shift factor, aT, was obtained from the tem-
`perature dependency of the zero-shear viscosity and was larger for
`the branched polymer compared to the linear one. From Fig. 11, it is
`clear that TTS holds for all the samples. van Gurp and Palmen (1998)
`proposed a refined analysis to check the validity of the TTS princi-
`ple. The TTS principle is respected when the plot of the loss angle,
`
`that the curves of the high Mw based blends are closer to that of the
`neat L-PP component, which confirms the results demonstrated in
`Fig. 4.
`Figs. 9(a) and (b) illustrate the storage modulus of the PP40/LCB-
`PP and PP04/LCB-PP blends, respectively. For the PP40 based blends,
`at low frequencies, the storage modulus of the LCB-PP is larger while
`the effect becomes inversed at high frequencies. For the blends con-
`taining a high molecular weight component (Fig. 9(b)) some syner-
`gistic effects at low frequencies are seen for all compositions, which
`is possibly due to the immiscibility of these blends. The increase of
`elasticity at low frequencies is common in immiscible blends and
`has been interpreted in the context of emulsion models (Chun et al.,
`2000; Palierne, 1990; Utracki, 1991).
`Palierne (1990) developed a model to predict the linear viscoelas-
`tic properties of immiscible emulsion-type blends. For a narrow dis-
`tribution of droplet diameters (Bousmina and Muller, 1993) and
`constant interfacial tension, the complex modulus of the blend is
`expressed by:
`1 + 3
`H∗(♽)
`(♽) = G
`(♽)
`1 − 2
`H∗(♽)
`G
`where H∗ is defined as:
`(cid:3)
`(cid:2)
`m(♽))(16G∗m(♽)+19G∗m(♽)+5G∗d(♽))+(G∗d(♽)−G∗ d(♽))
`
`
`
`
`
`(2G∗
`(cid:2)
`(cid:3)
`
`
`
`
`
`(G∗
`
`♡ R
`
`v
`
`♡ R
`
`H∗(♽) =
`
`10
`
`where
` is the volume fraction of the droplets of volume average
`m and G∗
`radius Rv, ♡ is the interfacial tension, and G∗
`d are the complex
`moduli of the matrix and droplets, respectively.
`Due to the low optical contrast between the PP components it
`was impossible to obtain the dimensions of drops in the PP blends
`and hence Rv. Following Fang et al. (2005) and Hussein and Williams
`(2001, 2004), ♡/Rv was used as a single parameter to find the best
`fits of the experimental data for the blends containing 20 wt% LCB-
`PP with low and high molecular weight L-PP components.
`Figs. 10(a) and (b) demonstrate the influence of ♡/Rv (0, 100, and
`500) on the storage and loss moduli predicted by the Palierne model
`for the (80/20) PP40/LCB-PP and (80/20) PP04/LCB-PP, respectively.
`It is obvious that the G(cid:4)(cid:4)values predicted by the model are not sensi-
`tive to the value of ♡/Rv and are in good agreement with the exper-
`imental data. In contrast, the G(cid:4) values are dramatically influenced
`by the value of ♡/Rv at low frequencies. For the (80/20) PP40/LCB-
`PP blend the best fit of the model with the experimental data was
`
`Page 5 of 13
`
`

`
`4724
`
`S.H. Tabatabaei et al. / Chemical Engineering Science 64 (2009) 4719 -- 4731
`
`Fig. 9. Storage modulus as a function of frequency for: (a) PP40/LCB-PP blends and
`(b) PP04/LCB-PP blends; T = 190 ◦C.
`
`Fig. 8. Cole–Cole plots for: (a) PP40/LCB-PP blends and (b) PP04/LCB-PP blends;
`T = 190 ◦C.
`
`♥(tan ♥=G(cid:4)(cid:4)/G(cid:4)), as a function of complex modulus, G∗=(G(cid:4)2+G(cid:4)(cid:4)2)1/2,
`superimpose in a single curve for all temperatures. The analysis of
`van Gurp and Palmen (1998) was examined for (60/40) PP40/LCB-PP
`and (60/40) PP04/LCB-PP blends as well as their neat components
`(the results are not shown). A single curve for each sample for var-
`ious temperatures was observed, indicating that the TTS principle
`holds for all the samples. Macaubas and Demarquette (2002) inves-
`tigated the applicability of TTS for a blend containing 10 wt% PP in
`PS and a blend containing 10 wt% PP in HDPE using the van Gurp-
`Palmen analysis. The immiscible 10 wt% PP/HDPE blend was ob-
`served to respect the principle, but not the immiscible 10 wt% PP/PS
`blend. This was explained by large differences of the flow activation
`energy, horizontal shift factor, and of the interfacial tension for PP
`and PS compared to those of PP and HDPE. In our case, although
`different rheological characterization methods suggested immisci-
`bility of the PP04/LCB-PP blend systems, no significant differences
`for the flow activation energy and the shift factor were found for the
`blend components. In addition, a small value of interfacial tension
`between the components is expected. Hence, these observations
`can explain the validity of TTS for the (60/40) PP04/LCB-PP blend.
`In contrast to oscillatory shear data, uniaxial extension is very
`sensitive to molecular and microstructural parameters (M¨unstedt
`et al., 1998; Wagner et al., 2000). The transient elongational viscos-
`ity ♪+
`e (t) of the resins and blends at different strain rates and 190 ◦C
`
`is illustrated in Fig. 12 (as in Fig. 11, the curves have been shifted by
`a multiplication). As expected, the linear polypropylenes respect the
`linear viscoelastic behavior over a large strain range where the tran-
`sient elongational viscosity is equal to three times that the transient
`viscosity in simple shear determined using the relaxation spectrum
`according to the following equation:
`N(cid:4)
`i=1
`
`e (t) = 3
`♪+
`
`♮iHi(1 − e
`−t/♮i )
`
`(4)
`
`To obtain the transient elongational viscosity from the above equa-
`tion 100 modes (N) were used.
`A significant strain hardening is observed for the LCB-PP as well
`as for all the blends. A sharper increase in the extensional viscosity
`curves and deviations from the linear viscoelasticity at shorter times
`were found as the weight fraction of the LCB-PP increases from 0.20
`to 1. Stange et al. (2005) found that adding a small amount of the
`same long-chain branched PP to a linear PP (different than those used
`in this work) caused strain-hardening and the effect was attributed
`to long-chain branching. Similar results for blends of LDPE and LLDPE
`were reported by Ajji et al. (2003) and Wagner et al. (2004). Note that
`for the PP04/LCB-PP blends at all strain rates the shape of the curves
`
`Page 6 of 13
`
`

`
`S.H. Tabatabaei et al. / Chemical Engineering Science 64 (2009) 4719 -- 4731
`
`4725
`
`Fig. 10. Sensitivity of the Palierne model predictions of G(cid:4) and G(cid:4)(cid:4) to different values of ♡/Rv for: (a) (80/20) PP40/LCB-PP blend and (b) (80/20) PP04/LCB-PP blend; T = 190 ◦C.
`
`are similar to that of the PP04, in accordance with the shear data
`(see Figs. 4 and 8), which is probably due to immiscibility of these
`blends. As for the shear properties, it is clear that the elongational
`properties are dominated by the high molecular weight component.
`The behavior of the PP28/LCB-PP blends was similar to that of
`the PP40/LCB-PP blends and that of the PP08/LCB-PP was close to
`the PP04/LCB-PP. For the sake of simplicity and brevity, the results
`for the PP28/LCB and PP08/LCB blends are not shown.
`
`3.3. Thermal characterization
`In the melting curve of the neat L-PPs, a small peak around 148 ◦C
`was observed and attributed to the presence of a small amount of
`beta (♢) or hexagonal crystalline form of PP (Jang et al., 2001). The
`magnitude of the peak was reduced when up to 5 wt% of the branched
`PP was added and disappeared by further addition of the LCB-PP. As
`the peak was not seen for the neat LCB-PP, it could be concluded
`that the presence of long-chain branches prevented the formation of
`
`beta crystals. Tian et al. (2006b) studied the crystalline structures of
`linear and long chain branched PPs using wide-angle X-ray diffrac-
`tion (WAXD). They found that linear PPs could crystallize in the ♡
`and ♢ forms, while branched PPs crystallized only in the ♡ crystalline
`form, in agreement with our DSC results.
`Melting and crystallization temperatures obtained from the peak
`positions as well as the degree of crystallinity of the various materials
`and blends are presented in Table 2 The melting point of the blends
`first increases by adding 20 wt% LCB-PP to the linear PPs while further
`addition reduces it. The reason for this behavior is unclear at present
`and different trends for the melting point of blends of linear and
`branched PPs have been reported in the literature (McCallum et al.,
`2007). Generally, a single melting peak was observed for all the
`blends. However, as the melting points of the components are very
`close, this cannot be considered as an indication of miscibility.
`Adding LCB-PP to the linear PPs causes a significant increase in
`the crystallization temperature, Tc, for all LCB contents, with the
`maximum change occurring at 20 wt% LCB-PP (see Table 2). Simi-
`lar observations for blends of other linear and branched PPs were
`
`Page 7 of 13
`
`

`
`4726
`
`S.H. Tabatabaei et al. / Chemical Engineering Science 64 (2009) 4719 -- 4731
`
`reported by McCallum et al. (2007). To determine at which con-
`centration the transition for the crystallization temperature occurs,
`additional blends containing 2, 5, and 10 wt% LCB-PP were prepared
`and analyzed and the results are shown in Fig. 13. It is clear that Tc
`of PP28 changes from 114.4 to 123.3 ◦C after adding only 2 wt% LCB-
`PP while further addition of LCB-PP causes a slight increase of the
`crystallization temperature. Dramatic increases in the crystallization
`temperature have also been reported in literature (Zhang et al., 2007)
`
`when a nucleating agent was added to a neat polymer. It was in-
`terpreted by a lower free energy for the heterogeneous nucleus for-
`mation compared to that of the homogeneous nucleus formation. In
`our case, one possibility for the dramatic rise of Tc can be explained
`by a large increase of the heterogeneous nuclei sites due to addition
`of the long-chain branches. However, the residual catalysts used to
`
`Fig. 11. Time-temperature superposition for:
`(60/40) PP40/LCB-PP
`(a) PP40,
`blend, LCB-PP and (b) (60/40) PP04/LCB-PP blend (the reference temperature
`is 225 ◦C).
`
`Fig. 12. Transient elongational viscosities as a function of time at different Hencky
`strain rates, ˙♦, for: (a) PP40/LCB-PP blends and (b) PP04/LCB-PP blends; T = 190 ◦C
`(the solid lines represent the linear behavior calculated from the relaxation time
`spectrum).
`
`Table 2
`Melting point (Tm), crystallization temperature (Tc), and degree of crystallinity (Xc) of the neat PPs as well as the blends ( ± indicates the standard deviation of the crystallinity).
`◦
`◦
`◦
`◦
`
`Sample
`
`Tm (
`
`C)
`
`Tc (
`
`C)
`
`PP28
`(20/80) PP28/LCB-PP
`(40/60) PP28/LCB-PP
`(60/40) PP28/LCB-PP
`LCB-PP
`PP40
`(20/80) PP40/LCB-PP
`(40/60) PP40/LCB-PP
`(60/40) PP40/LCB-PP
`LCB-PP
`
`161.0
`162.1
`161.6
`160.6
`158.6
`159.8
`161.3
`160.9
`160.0
`158.6
`
`114.4
`126.5
`127.5
`127.7
`128.4
`119.2
`126.9
`127.5
`127.8
`128.4
`
`Xc
`40.0 ± 0.9
`42.2 ± 1.4
`41.3 ± 0.7
`35.6 ± 1.1
`35.0 ± 0.9
`38.8 ± 1.9
`40.8 ± 1.1
`40.2 ± 0.7
`37.3 ± 1.2
`35.0 ± 0.9
`
`Sample
`
`Tm (
`
`C)
`
`Tc (
`
`C)
`
`PP04
`(20/80) PP04/LCB-PP
`(40/60) PP04/LCB-PP
`(60/40) PP04/LCB-PP
`LCB-PP
`PP08
`(20/80) PP08/LCB-PP
`(40/60) PP08/LCB-PP
`(60/40) PP08/LCB-PP
`LCB-PP
`
`159.6
`161.9
`161.7
`160.5
`158.6
`160.0
`161.5
`160.5
`160.2
`158.6
`
`116.9
`125.5
`126.4
`127.1
`128.4
`117.3
`126.0
`127.3
`127.6
`128.4
`
`Xc
`36.8 ± 1.1
`41.6 ± 0.7
`37.5 ± 0.5
`36.9 ± 1.2
`35.0 ± 1.0
`37.8 ± 1.5
`40.2 ± 0.8
`39.6 ± 0.8
`39.1 ± 0.3
`35.0 ± 1.0
`
`Page 8 of 13
`
`

`
`S.H. Tabatabaei et al. / Chemical Engineering Science 64 (2009) 4719 -- 4731
`
`4727
`
`Fig. 13. DSC cooling thermographs for PP28, LCB-PP, and their blends at 10 ◦C/min.
`
`Fig. 15. Plots of log[− ln(1 − X(t))] as function of log(t) for PP28, LCB-PP and the
`40 wt% PP28/LCB-PP blend.
`
`and cooling rate, ♚, is expressed as (Tian et al., 2007):
`t = To − T
`
`♚
`
`(5)
`
`where To is the onset crystallization temperature. All the curves show
`an S shape, similar for all the components and blends, as seen in
`Fig. 14. However, a lower transition appeared at the later crystalliza-
`tion stage for samples containing the LCB-PP. In general, the rate of
`crystallization is first controlled by nucleation and then by growth of
`crystals and packing (Tian et al, 2007). As mentioned above, adding
`a branched polymer to a linear one increases the number of nuclei
`sites resulting in a larger crystallization rate for the blends in the
`early stages of the crystallization. However, due to more impinge-
`ments of the crystalline parts, crystalline growth is stopped at later
`stages. Tian et al. (2007) studied the crystallization kinetics of a linear
`PP and PPs containing different levels of long chain