Rheologica Acta

, Volume 46, Issue 4, pp 507–520

Rheology and morphology of polystyrene/polypropylene blends with in situ compatibilization

Authors

  • Yanli Huo
    • Department of Chemical EngineeringKatholieke Universiteit Leuven
  • Gabriel Groeninckx
    • Department of Chemistry, Division of Molecular and Nanomaterials, Laboratory of Macromolecular Structural ChemistryKatholieke Universiteit Leuven
    • Department of Chemical EngineeringKatholieke Universiteit Leuven
Original Contribution

DOI: 10.1007/s00397-006-0158-3

Cite this article as:
Huo, Y., Groeninckx, G. & Moldenaers, P. Rheol Acta (2007) 46: 507. doi:10.1007/s00397-006-0158-3

Abstract

Rheology and flow-induced morphology were studied in immiscible polypropylene (PP)/polystyrene (PS) blends with a droplet–matrix microstructure. Two reactive precursors, maleic anhydride grafted PP and amino terminated PS, were added during the melt-mixing process to form a graft copolymer. The effects of both the amount of compatibilizer and the shear history on the rheological and morphological behavior were investigated systematically. Small amplitude oscillatory experiments and scanning electron microscopy were used to study the phase morphology. Shear history has an important effect on the morphology of the uncompatibilized blends. The droplet size refines with increasing shear rate. The decrease of this effect with increasing degrees of in situ compatibilization is mapped out. The results are discussed in terms of interfacial tension and the interfacial coverage. It turns out that most of the conclusions that were previously obtained on physically compatibilized blends are also valid for chemically compatibilized ones.

Keywords

In situ compatibilizationMorphologyRheologyPolymer blends

Introduction

Blending of polymers is a common and economical means of creating new materials with improved properties. Most polymers are immiscible and hence produce two-phase blends. The weak interaction between the two components results in poor interfacial adhesion. In addition, the morphology is usually coarse and unstable, which causes bad mechanical properties. The addition of premade block or graft copolymers (physical compatibilization) can produce a finer morphology and increase the interfacial adhesion. The results of this compatibilization are influenced by molecular weight and architecture of the block copolymer (Macosko et al. 1996; Riemann et al. 1997; Van Hemelrijck et al. 2004). In situ formed block or graft copolymers, generated at the interface by a reaction during the melt-mixing process (chemical compatibilization), can act as efficient compatibilizers. In comparison with premade compatibilizers, these copolymers are more likely to be located at the interface between the two components (Milner and Xi 1996). Their compatibilization efficiency will also depend on their molecular weight and structure (Joen et al. 2004).

In in situ compatibilized polymer blends, the compatibilizer is generated using a polymer or precursor with functional groups. It can react with one of the components of the blend to form the copolymer at the interface. This is illustrated by the well-known reactive blends based on polyamide (PA). The reactive polymer contains functional groups such as maleic anhydride (Dedecker and Groenincks 1999) or acrylic acid (Jo and Kim 1992), which can react with the amine end group of the PA. Two reactive precursors can also be used to generate block or graft copolymers without involving any of the blend components (Moan et al. 2000; Harrats et al. 2004). The amount and the reactivity of the functional groups, the molecular weight, and the structure of the reactive chains and the diffusion kinetics of the precursors from their parent phases to the interface all affect the interfacial reaction (Yin et al. 2001, 2002; Harrats et al. 2004). Interfacial reactions were studied using bilayers of end-functional polymers to eliminate the effect of mixing (Jiao et al. 1999; Jones et al. 2003; Yin et al. 2003). The results indicate that the extent of reaction at polymer interfaces is related to the molecular weight and structure of functional polymers, reactant concentration, and thermodynamic interactions.

The rheological properties and the flow-induced morphology of immiscible blends are interconnected. This relationship was investigated intensively for uncompatibilized blends under steady shear flow, both for model systems at ambient temperature and blends of real molten polymers (e.g., Vinckier et al. 1996; Graebling et al. 1993; Gramespacher and Meissner 1992). In physically compatibilized blends, the rheological properties are influenced by the amount, molecular weight, and architecture of the added copolymers. Again data are available on model and real systems (Germain et al. 1994; Riemann et al. 1997; Velankar et al. 2004). Several groups also studied the rheological properties of in situ compatibilized blends. For example, Asthana and Jayaraman (1999) investigated PA 6/polypropylene (PP) blends with different extents of reaction at the interface. They reported that the interfacial tension, as expected, dropped progressively with increasing extent of reaction. Moan et al. (2000) studied the effects of a random terpolymer formed during the reactive blending on morphological and rheological properties of a PA 12/linear low-density polyethylene blend. In both studies an additional relaxation mechanism was observed, which did not exist in the uncompatibilized blends. The literature on the rheology–morphology relation in chemically compatibilized blends is, however, rather limited. This is, among others, related to the inherent difficulties associated with prolonged experiments at elevated temperatures. A systematic investigation of the rheology and the resulting flow-induced morphology in reactively compatibilized blends is the subject of the present study.

Materials and methods

Materials

PP and polystyrene (PS), two of the most widely used plastics with good thermal stability, were selected as the blend components. Two reactive precursors, maleic anhydride grafted PP (PP-g-MA) and amino-terminated PS (PS-NH2), were added during the melt-mixing process to form graft copolymers. The PS (R2/827 from BASF) has a weight average molecular weight Mw of 51,510 g/mol and a Mw/Mn ratio of 1.05. The PP (MFI-37 from Borealis) has a melt flow index of 37 g/10 min. The PP-g-MA (from Eastman Kodak), with 7.8wt% of MA, has a Mw of 9,100 g/mol and a Mw/Mn ratio of 2.3. The PS-NH2 (from Polymer Source) has a Mw of 33,500 g/mol and a Mw/Mn ratio of 1.04. The zero shear viscosities of PP and PS at 205 °C are 536 and 93 Pa s, respectively. With PS as the dispersed phase, the viscosity ratio (p) is around 0.2 at this temperature. A critical capillary number can be associated with the viscosity ratio. It reflects the conditions at which viscous stresses start to dominate the shape restoring interfacial stresses: above this critical capillary number droplet break up occurs. The critical capillary number for our PS in PP blend is approximately 0.5 (Grace 1982). All the rheological measurements were performed at 205 °C. The linear dynamic moduli of PP and PS at this temperature are shown in Fig. 1.
https://static-content.springer.com/image/art%3A10.1007%2Fs00397-006-0158-3/MediaObjects/397_2006_158_Fig1_HTML.gif
Fig. 1

Linear dynamic moduli of PP and PS at 205 °C

The blend components were dry-blended at ambient temperature and fed to a DSM twin-screw midi extruder. The blending conditions were the same for all samples: a temperature of 215 °C, rotation rate 100 rpm, and mixing time of 5 min. The latter was adequate due to the high reactivity of MA and the amine groups (Harrats et al. 2004). During melt blending the mixing chamber was saturated with N2 gas to avoid oxidative degradation. The concentration of compatibilizer x is expressed as the weight percentage of the reactive mixture PP-g-MA + PS-NH2 on the total blend. The composition of the blends with the reactive polymers was 0.8(100 − x) wt% PP/0.2(100−x) wt% PS/x wt% compatibilizer with x varied from 0 to 5. The two reactive precursors were added at the stoichiometric anhydride/amine ratio. After melt blending, the extruded filaments were quenched in cold water and cut into pellets, which were squeezed into disc-shaped plates of 25 mm diameter and 1 mm thickness using a heating press (Collin) at 215 °C. These discs were used for the rheological measurements.

Rheological measurements and Palierne model

Rheological behavior and the resulting morphology were investigated in a systemic way with varying compatibilizer concentrations and shear histories. The rheological measurements were performed on a dynamic stress rheometer (Rheometric Scientific) with 25 mm diameter/0.1 rad cone and plate geometry at 205 °C in a nitrogen atmosphere. First, the sample was presheared at a given shear rate for a fixed duration in time. This generated a morphology, which is determined by the balance between break up and coalescence. More specifically, preshearing at 1.0, 1.5, and 2.0 s−1 was performed for 1,800 strain units. At lower shear rates of 0.5 and 0.1 s−1, the applied strain was 1,200 and 240, respectively. For uncompatibilized blends it was verified that 1,200 strain units were sufficient to reach steady state. It should be noted that for each experiment a new sample from the same batch was used to avoid degradation effects. After the flow was stopped, small amplitude oscillatory measurements were performed. The applied strain in the oscillatory tests was in the linear viscoelastic region and the morphology did not change during the dynamic measurements. The flow-induced morphology could then be probed by analyzing the frequency dependence of the storage moduli, using the Palierne model (Palierne 1990; Jacobs et al. 1999).

A simple version of the Palierne model has successfully been applied to obtain morphological information of uncompatibilized blends by several researchers (Graebling et al. 1993; Friedrich et al. 1995; Vinckier et al. 1996). For a blend with a droplet–matrix morphology, the storage moduli display a shoulder in the low-frequency region due to a relaxation mechanism that reflects the shape relaxation of the droplets. From fitting the Palierne model to the dynamic moduli, the ratio of the volume average radius (Rv) over the interfacial tension (α) can be derived. If the volume average radius is determined by an independent method and the droplet size distribution is not too large, the interfacial tension can be obtained in this manner. Inversely, the flow-induced droplet size can be determined for a blend with a known interfacial tension. For compatibilized blends, the more general version of the Palierne model has to be used (Jacobs et al. 1999). With some assumptions it can be reduced to a simplified form with only two parameters, α/Rv and β′′/Rv, in which the parameter β′′ is the interfacial shear modulus. The latter is related to the presence of the compatibilizer at the interface. The two-parameter version of the Palierne model, as discussed in detail by Van Hemelrijck et al. (2004), will be used in the present work.

Morphology detection

After the rheological experiments the specimen was cooled using cold air and subsequently fractured in liquid nitrogen. A smooth surface was achieved by a microtome (Leica Ultracut Uct) equipped with a glass knife. The dispersed PS phase was extracted with chloroform at room temperature to enhance the contrast. The fractured surfaces of the etched samples were sputter-coated with gold. The blend morphology was observed by means of scanning electron microscopy (SEM, Philips XL30 FEG).

To determine the droplet size, a transparency was laid over the micrograph and the droplets were traced by hand (Sundararaj and Macosko 1995). The manual trace was then scanned into the image analysis software (Scion Image). The diameter of each droplet (Di) was calculated from the corresponding area (Ai). Typically 200–300 particles were analyzed per sample. Corrections to the particle size were performed using Schwartz–Saltykov method (Underwood 1970). The correction was done by first dividing the particle size distribution into 10–15 linear size ranges and by characterizing each size range with the midpoint of the range. The particle size was then multiplied by a matrix of coefficients resulting from a set of equations to get the real particle size distribution in three dimensions. The number average diameter (Dn) and the volume average diameter (Dv) were calculated based on the real particle size distribution by Eqs. 1 and 2, respectively:
$$ D_{{\text{n}}} = \frac{{{\sum\limits_i {{\left( {N_{{\text{v}}} } \right)}_{i} D_{i} } }}} {{{\sum\limits_i {{\left( {N_{{\text{v}}} } \right)}_{i} } }}} $$
(1)
$$ D_{{\text{v}}} = \frac{{{\sum\limits_i {{\left( {N_{{\text{v}}} } \right)}_{i} D^{4}_{i} } }}} {{{\sum\limits_i {{\left( {N_{{\text{v}}} } \right)}_{i} D^{3}_{i} } }}} $$
(2)
with (Nv)i as the number of particles having diameter Di. The polydispersity (d) was characterized by means of the ratio:
$$ d = {D_{{\text{v}}} } \mathord{\left/ {\vphantom {{D_{{\text{v}}} } {D_{{\text{n}}} }}} \right. \kern-\nulldelimiterspace} {D_{{\text{n}}} } $$
(3)

Results and discussion

Rheology and morphology of uncompatibilized blends

Small amplitude oscillatory measurements are useful to probe the droplet–matrix morphology, generated by a flow history, without altering the morphology (Scholz et al. 1989; Vinckier et al. 1996). Figure 2a shows the shear rate dependence of the storage moduli for the uncompatibilized PP/PS blends. The solid lines refer to fitting the curves at shear rates of 0.1 and 2.0 s−1 with the emulsion model of Palierne using a single fitting parameter (Rv/α) (Palierne 1990). As a comparison, the component contribution to the storage modulus according to Dickie’s model (Dickie 1973) is added as well. In this model the contribution from the interface is not included. At high frequencies the responses for the various preshear rates are identical and are well described by Dickie’s model. This indicates that at these frequencies only the components contribute to the spectrum irrespective of the flow-induced morphology. At low frequencies the curves show a pronounced “shoulder,” which systematically shifts to higher frequencies when the preshear rate increases. This is logical because smaller droplets have a shorter shape relaxation time and consequently the shoulder appears at higher frequencies, reflecting the refinement of the droplet–matrix morphology with increasing shear rate.
https://static-content.springer.com/image/art%3A10.1007%2Fs00397-006-0158-3/MediaObjects/397_2006_158_Fig2_HTML.gif
Fig. 2

a Shear history dependence of the storage modulus of the uncompatibilized 80/20 PP/PS blend. The solid lines are the Palierne model fittings to the data at shear rates of 0.1 and 2.0 s−1. The dashed line refers to the component contribution to the storage modulus according to Dickie’s model. b The corresponding relaxation spectra of the uncompatibilized blend after different shear histories

The shape relaxation time of the droplets can be determined by calculating the continuous relaxation spectrum based on the dynamic moduli. A nonlinear regression program (Honerkamp and Weese 1993) was used for that purpose. The relaxation spectra of the blends after different preshear histories, calculated from the moduli of Fig. 2a, are shown in Fig. 2b. For comparison, the relaxation spectrum of the matrix (PP) is added as well. A relaxation peak, reflecting the shape relaxation, can clearly be seen in all the spectra of the blend.

From the fit of the Palierne model (Palierne 1990) to the dynamic moduli, the ratio of the volume averaged droplet radius (Rv) over the interfacial tension (α) can be derived. Rv can, in principle, be determined when the interfacial tension (α) is known from an independent method and vice versa. The reported values for the interfacial tension between PP and PS range from 4 to 7 mN/m, depending on the molecular weight of the components and the temperature (Funke et al. 2001; Palmer and Demarquette 2003). To avoid any ambiguity concerning the interfacial tension, the values of Rv/α as obtained from fitting the Palierne model to the data will be directly used here. They are plotted as a function of the preshear rate in Fig. 3. Alternatively, we could use the droplet relaxation time τs as a function of preshear rate from Fig. 2b. A value of Rv/α can then be calculated with an approximate equation (Graebling et al. 1993), derived from the Palierne model (Palierne 1990):
$$ \frac{{R_{{\text{v}}} }} {\alpha } = \frac{{4\tau _{{\text{s}}} }} {{\eta _{{\text{m}}} }}\frac{{10{\left( {p + 1} \right)} - 2\phi {\left( {5p + 2} \right)}}} {{{\left( {19p + 16} \right)}{\left[ {2p + 3 - 2\phi {\left( {p - 1} \right)}} \right]}}} $$
(4)
in which ηm is the matrix viscosity and φ the volume fraction of the dispersed phase. These values for Rv/α were added to Fig. 3. They are in good agreement with those obtained from fitting the Palierne model directly. The dashed line in Fig. 3 describes the values of Rv/α as expected on the basis of a critical capillary number of 0.5 (see above). It can be seen that the reduction in droplet size is inversely proportional to the preshear rate and that the absolute values of Rv/α coincide with those expected on the basis of the critical capillary number. The deviation at low shear rate (0.1 s−1) is due to the fact that the applied strain is insufficient to achieve a steady-state morphology. Also, a droplet size hysteresis may play a role. This behavior is very similar to that of some uncompatibilized model blends that are liquid at room temperature (Vermant et al. 2004).
https://static-content.springer.com/image/art%3A10.1007%2Fs00397-006-0158-3/MediaObjects/397_2006_158_Fig3_HTML.gif
Fig. 3

Rv/α as a function of preshear rate for the uncompatibilized 80/20 PP/PS blend. The dashed line represents the theoretical prediction based on the critical capillary number (Grace 1982). The micrographs of the blend after shearing at 1.0 and 2.0 s−1 are added

Rheology and morphology of compatibilized blends

Effect of compatibilizer concentration

As for the uncompatibilized blends, the dynamic moduli were used to probe the effect of compatibilizer concentration on the flow-induced morphology. Figure 4a shows the results for a series of compatibilizer concentrations after shearing at 1.0 s−1 for 1,800 strain units. The results of fitting the data with the Palierne model using two parameters are shown as solid lines in Fig. 4a for 0.5 and 5wt% compatibilizer. The dynamic response of pure PP and the uncompatibilized blend subjected to the same shear history are added to this figure as well. The spectrum of the blend with 0.5wt% compatibilizer was shifted vertically by one decade to clearly show the detailed shape of the curve. An alternative way to represent the data is shown in Fig. 4b in which G′/ω2 of the blends and the pure PP are plotted vs the frequency. It can be seen that when the amount of the compatibilizer is less than 2wt%, the curve exhibits two shoulders. The additional shoulder is seen at lower frequencies than that of the shape relaxation and shifts to higher frequencies with increasing compatibilizer concentration. The extra relaxation process, related to the presence of a compatibilizer, was also observed in some physically compatibilized polymer blends (Riemann et al. 1997; Van Hemelrijck et al. 2004) and in reactively compatibilized blends (Asthana and Jayaraman 1999; Moan et al. 2000). When increasing the amount of compatibilizer to 5wt%, the two shoulders merge into a single, very pronounced one.
https://static-content.springer.com/image/art%3A10.1007%2Fs00397-006-0158-3/MediaObjects/397_2006_158_Fig4_HTML.gif
Fig. 4

a Frequency dependence of the storage modulus of pure PP and 80/20 PP/PS blends with varying amount of compatibilizer. The solid lines represent fitting the data for 0.5 and 5wt% compatibilizer. b Frequency dependence of G′/ω2 of pure PP and 80/20 PP/PS blends with varying amount of compatibilizer. All the blends were subjected to a preshearing at 1.0 s−1 for 1,800 strain units

To quantify the characteristic relaxation times, the relaxation spectra were calculated from the dynamic moduli in Fig. 4a. The results are shown in Fig. 5, as well as those of PP and the uncompatibilized PP/PS blend, subjected to the same shear history. Two characteristic relaxation times, τs and τβ, can clearly be seen in the curves of compatibilized blends, except for the blend with 5wt% compatibilizer. In comparison with that of the uncompatibilized blend, the shape relaxation time of the compatibilized blends, τs, slightly shifts to lower values. The extra relaxation time, τβ, decreases substantially with increasing amount of compatibilizer. The same trends were also observed in the physically compatibilized blend systems (Riemann et al. 1997). When increasing the amount of the compatibilizer further to 5wt%, the two relaxation times merge, as was also reported for the physically compatibilized system [polyisoprene/poly(dimethyl siloxane), PI/PDMS] when the block copolymer concentration is more than 1% (Van Hemelrijck et al. 2004).
https://static-content.springer.com/image/art%3A10.1007%2Fs00397-006-0158-3/MediaObjects/397_2006_158_Fig5_HTML.gif
Fig. 5

Relaxation spectra of pure PP and 80/20 PP/PS blends with varying amount of compatibilizer calculated from the dynamic moduli. All the blends were subjected to a preshearing at 1.0 s−1 for 1,800 strain units

The terminal viscosity of the blends is also changed by compatibilization. Figure 6 shows the complex viscosity of blends with different amounts of compatibilizers subjected to the same shear history (shearing at 1.0 s−1 for 1,800 strain units). The data of pure PP are added as well. At low compatibilizer concentrations, the terminal viscosity slightly increases, and this increase becomes more pronounced as the compatibilizer amount reaches 5wt%. The increase of the terminal viscosity was also observed in physically compatibilized PI/PDMS blends (Van Hemelrijck et al. 2004).
https://static-content.springer.com/image/art%3A10.1007%2Fs00397-006-0158-3/MediaObjects/397_2006_158_Fig6_HTML.gif
Fig. 6

Frequency dependence of the complex viscosity of pure PP and 80/20 PP/PS blends with varying amount of compatibilizer. All the blends were subjected to a preshearing at 1.0 s−1 for 1,800 strain units

The morphology of blends with varying amounts of compatibilizer was analyzed by SEM. Figure 7 shows the results for a series of compatibilizer concentrations after shearing at 1.0 s−1 for 1,800 strain units. As a comparison, the morphology of the uncompatibilized PP/PS blend, subjected to the same history, is also shown. When the amount of compatibilizer is 0.5wt%, the droplet size is slightly smaller than without compatibilizer. Increasing the compatibilizer concentration results in a substantial further decrease in droplet size. This will further be analyzed below.
https://static-content.springer.com/image/art%3A10.1007%2Fs00397-006-0158-3/MediaObjects/397_2006_158_Fig7_HTML.jpg
Fig. 7

Scanning electron micrographs of 80/20 PP/PS blends with varying amount of compatibilizer subjected to the same shear history (shearing at 1.0 s−1 for 1,800 strain units): a 0wt%, b 0.5wt%, c 1wt%, and d 2wt%

Effect of shear history

The shear history has a significant effect on the morphology of the uncompatibilized blends. This persists in compatibilized blends as long as the compatibilizer concentration is sufficiently low. Figure 8a,b shows the shear history dependence of G′/ω2 for blends with 1 and 2wt% compatibilizer. With 1wt% compatibilizer, the curves for the various shear rates clearly differ. When the shear rate is higher than 0.5 s−1, the low frequency relaxation starts to emerge. Similar phenomena are observed at high shear rates (more than 1.0 s−1) for the blend with 2wt% compatibilizer (Fig. 8b). For both compatibilizer concentrations, the storage moduli still depend on the previous shear rate. However, the effect of shear history on the dynamic response starts to become less pronounced in the 2wt% compatibilized blend compared to the 1wt% compatibilized blend. The shear rate dependence is nearly completely suppressed in the 5wt% compatibilized blend (shown later in Fig. 10). In blends with low compatibilizer concentrations, lowering the shear rate causes τβ to decrease as τs to increase. This causes the two relaxation times to merge at low shear rates. The physical nature of the phenomenon is not completely clear yet, but it is probably related to the redistribution of compatibilizers at the interface because the interface is not completely immobilized by the compatibilization.
https://static-content.springer.com/image/art%3A10.1007%2Fs00397-006-0158-3/MediaObjects/397_2006_158_Fig8_HTML.gif
Fig. 8

Shear history dependence of G′/ω2 of the 80/20 PP/PS blends with a 1wt% and b 2wt% compatibilizer

Figure 9 shows the corresponding micrographs of the compatibilized blends with 1 and 2wt% compatibilizer after shearing at different shear rates. Subjected to the same shear rate, the blend with 2wt% compatibilizer has a smaller droplet size. In both cases the droplet size decreases with increasing shear rates (quantitative analysis, see further).
https://static-content.springer.com/image/art%3A10.1007%2Fs00397-006-0158-3/MediaObjects/397_2006_158_Fig9_HTML.jpg
Fig. 9

Scanning electron micrographs of 80/20 PP/PS blends with a 1wt% and b 2wt% compatibilizer subjected to different shear histories (0.1s–1, 0.5s–1, 1.5s–1, and 2.0s–1)

Increasing the compatibilizer concentration to 5wt% nearly eliminates the dependence of the storage moduli on shear rates (Fig. 10a). Micrographs of the phase morphology after shearing at 0.5 and 1.0 s−1 are also added in Fig. 10a. The morphology is indeed hardly affected by shearing. The corresponding relaxation spectra of the blend with 5wt% compatibilizer after different shear histories are shown in Fig. 10b. At first glance only a single shoulder is present in the relaxation spectra, which is nearly identical for different shear rates.
https://static-content.springer.com/image/art%3A10.1007%2Fs00397-006-0158-3/MediaObjects/397_2006_158_Fig10_HTML.gif
Fig. 10

a Shear history dependence of the storage modulus of a 80/20 PP/PS blend with 5wt% compatibilizer. The micrographs of the blend after shearing at 0.5 and 1.0 s−1 are added. b Corresponding relaxation spectra

Discussion

The quantitative results of the SEM analysis (Rv and d) and those from the rheological analysis (α/Rv and β′′/Rv) are given in Tables 1 and 2. In Table 1 the effect of compatibilizer concentration on the various parameters is summarized for a preshearing of 1.0 s−1 for 1,800 strain units, whereas in Table 2 the effect of shear rate is summarized for two compatibilizer loadings. When the quality of the fitting of the Palierne model was not very good, the fitting parameters are put between brackets.
Table 1

Effect of compatibilizer concentration on droplet radii Rv, polydispersity d, fitting parameters α/Rv and β′′/Rv, the calculated interfacial tension α and interfacial shear modulus β′′, and the calculated interfacial coverage c0 (shearing at 1.0 s−1 for 1,800 strain units)

wt% (x)

Rv (μm)

d

α/Rv (N/m2)

β′′/Rv (N/m2)

α  (mN/m)

β′′ (mN/m)

c0a

0

4.66

1.21

972

 

4.5

  

0.5

4.25

1.13

999

3.906

4.2

0.0166

0.205

1

3.50

1.12

877

7.547

3.1

0.0264

0.340

2

2.80

1.21

(1,863)

(126.6)

(5.2)

(0.355)

0.549

5

2.21

1.27

1,383

276.7

3.1

0.612

1.117

aIn copolymer chains per square nanometer

Table 2

Effect of shear rate on droplet radii Rv, polydispersity d, fitting parameters α/Rv and β′′/Rv, the calculated interfacial tension α and interfacial shear modulus β′′, and the calculated interfacial coverage c0 for the blends with 1 and 5wt% compatibilizers

Shear rate (s−1)

Rv (μm)

d

α/Rv (N/m2)

β′′/Rv (N/m2)

α  (mN/m)

β′′ (mN/m)

c0a

1wt% (x)

5wt% (x)

2.0

 

2.67

1.19

1,319

6.002

3.5

0.0160

0.259

1.5

 

3.10

1.16

1,151

6.103

3.6

0.0189

0.301

1.0

 

3.50

1.12

877

7.547

3.1

0.0264

0.340

0.5

 

4.81

1.24

(1,162)

(104.9)

(5.6)

(0.504)

0.467

0.1

 

6.77

1.13

(678)

(134.2)

(4.6)

(0.908)

0.657

 

1.0

2.21

1.27

1,383

276.7

3.1

0.612

1.117

 

0.5

2.65

1.19

1,202

278.9

3.2

0.739

1.340

 

0.1

3.54

1.16

899

168.9

3.2

0.598

1.790

aIn copolymer chains per square nanometer

The effect of shear history and compatibilizer concentration on the droplet dimensions, as determined by SEM, is shown in Fig. 11. For the uncompatibilized blends the SEM pictures point toward an inverse proportionality between shear rate and droplet size. This is consistent with the results from the fitting of the Palierne model to the moduli (shown in Fig. 3). Adding compatibilizer clearly reduces the droplet size and the dependency of the size on shear rate.
https://static-content.springer.com/image/art%3A10.1007%2Fs00397-006-0158-3/MediaObjects/397_2006_158_Fig11_HTML.gif
Fig. 11

Effect of the shear history and compatibilizer concentration on the droplet size of 80/20 PP/PS blends

As mentioned above, the dynamic data of the compatibilized blends were analyzed using the Palierne model with an interfacial shear modulus. Two parameters, α/Rv and β′′/Rv, are now obtained with the fitting procedure (see Tables 1 and 2). If the droplet radius Rv is known from an independent measurement such as SEM in our investigation and the droplet size distribution is not too large, α and β′′ can be calculated. These values were added to the tables. The interfacial coverage c0 of the compatibilizer can also be estimated (Van Hemelrijck et al. 2004), knowing the droplet radii and assuming that all the reactive precursors have reacted to form copolymers and no micelles of the copolymers are present in bulk phases:
$$ c_{0} = \frac{{z\rho _{d} R_{{\text{v}}} N_{{\text{A}}} }} {{300\,M_{{\text{w}}} }} $$
(5)
where z is the fraction of copolymer with respect to the amount of the dispersed phase. ρd (kg/m3), NA, and Mw (kg/mol) are the density of the dispersed phase, the Avogadro number (6.02 × 1023), and the molecular weight of the copolymer, respectively. The average number of the anhydride groups on a chain of PP-g-MA is about 3. Assuming that all the anhydride groups react with the amino group on the PS-NH2 chain, the molecular weight of the copolymer equals (\( M_{{\text{w}}} ,\,_{{{\text{PP - g - MA}}}} + 3 \times M_{{\text{w}}} ,_{{{\text{PS - NH}}_{{\text{2}}} }} \)) 109.6 kg/mol. The resulting values of the interfacial coverage are added to Tables 1 and 2.
Van Hemelrijck et al. (2004) showed that the value of β′′ was proportional to the interfacial coverage in a series of physically compatibilized blends. It can be verified here to what extend this relationship also holds for the chemically compatibilized blends under investigation. Figure 12 shows the interfacial shear modulus as a function of interfacial coverage for the series of shear rates and compatibilizer concentrations studied here. From this figure it is clear that the interfacial shear modulus strongly increases with interfacial coverage, as was the case for the physically compatibilized blends. The uncertainty on the numerical values is larger for the chemically compatibilized blends. Hence, the power law of the dependency cannot be determined accurately from the data shown in Fig. 12, although they suggest a more than proportional relation.
https://static-content.springer.com/image/art%3A10.1007%2Fs00397-006-0158-3/MediaObjects/397_2006_158_Fig12_HTML.gif
Fig. 12

Evolution of the interfacial shear modulus as a function of the interfacial coverage for different compatibilizer concentration. The arrows indicate the direction of the shear rate from 2.0 to 1.0 s−1 in the 1wt% blend and from 1.0 to 0.1 s−1 in the 5wt% blend

Conclusions

The rheology and the flow-induced morphology of in situ compatibilized PP/PS blends were investigated using dynamic measurements and SEM. In uncompatibilized blends, the shear history has a pronounced effect on the morphology. The shear effect still exists in the compatibilized blends but it gradually disappears at higher compatibilizer concentration. Two relaxation mechanisms are observed at higher shear rates and lower compatibilizer contents. At higher compatibilizer levels, here 5wt%, the rheology and morphology hardly change with shear rates. The results are in line with those obtained on physically compatibilized systems. The interfacial tension and interfacial shear modulus were estimated by fitting the data with Palierne model. As for physically compatibilized blends, the interfacial shear modulus substantially increases with interfacial coverage.

Acknowledgements

Financial support from the Research Council, Katholieke Universiteit Leuven (grant GOA 03/06) and the FWO-Flanders (project G. 052304) are gratefully acknowledged. The authors wish to thank Prof. Jan Mewis for helpful discussions.

Copyright information

© Springer-Verlag 2007