tllllllIIWIIIHWIIIIHllllllﬂllllHIWWIHIINWI

   

lCHiGAN STATE UNIVERSITY LIBRARIES

lull/IllINN/lIll/llIl/r’ll/Il/lll/llIll/Ill!!!” l

3 1293 02080 6117

 

 

 

 

LIBRARY

Michigan State
University

 

 

 

This is to certify that the

thesis entitled

MORPHOLOGY VARIATIONS IN EXTRUSION 0F POLYPROPYLENE/BF
RUBBER BLENDS

presented by

Jonathan Michael Lopez

has been accepted towards fulﬁllment
of the requirements for

M. S . degree in Chemical Engineering

 

 

[WM/«1:3

-W—‘ﬂ

U @tprofessor

 

 

:a/Ie/ij

O~7639 MS U is an Afﬁrmative Action/Equal Opportunity Institution

 

PLACE lN RETURN BOX to remove this checkout from your record.
TO AVOID FINE return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

QCIHW

 

 

 

 

 

 

 

 

 

 

 

 

 

11100 chlFiC/DﬁﬁnpﬁS—p.“

 

MORPHOLOGY VARIATIONS IN EXTRUSION OF POLYPROPYLENE / EP
RUBBER BLENDS

By

Jonathan Michael Lopez

A THESIS

Submitted to
Michigan State University
in partial fulﬁllment of the requirements
for the degree of

MASTER OF SCIENCE

Department of Chemical Engineering

1999

ABSTRACT

MORPHOLOGY VARIATIONS IN EXTRUSION OF POLYPROPYLENE I EP
RUBBER BLENDS

By

Jonathan Michael Lopez

This study investigates the morphology of two polymer blends containing
polypropylene and one of two different grades of ethylene propylene (EP) rubber. Image
analysis of SEM micrographs of the extrudates taken from different processing
conditions reveals migration and varying degrees of stretching of the EP rubber drops
from the wall to the center region of the die. The least volume fraction of the rubber
phase was found to occur in the wall region, while the maximum degree of stretching was
found to occur at the shear-free, center region of the capillary. At the center, the
stretched rubber drops assumed a filament-like structure. Stretching that occurred in the
wall region was found to be in the vorticity direction.

Greater apparent slip was found in the larger die diameters, where more
migration was expected to occur. The amount of migration that occurs was explained in
the terms of two competing fluxes, the flux due to viscosity variations and the flux due to
collision. As viscosity increase with aspect ratio of the drops, the flux due to viscosity is
expected to be signiﬁcantly greater than the flux due to collisions in a small diameter die,
where more stretching is expected to occur. This leads to a higher viscosity barrier in the

core region for migration, so that less migration occurs in the small diameter dies.

In Loving Memory of Marina Gallardo Lopez

iii

ACKNOWLEDGMENTS

This research was supported by Montell Polyolefins and by the Michigan
Materials and Processing Institute. I am pleased to acknowledge the assistance of Dr. Chi
Chang Shu and Dr. M.D. Wolkowicz at Montell Polyoleﬁns in identifying materials and

preparing the blends for this study.

I would also like to express my sincere thanks to Dr. K. J ayaraman for not only
helping me to deﬁne the problem in the early stages, but also for providing invaluable
guidance during the course of this work. His energy and keen insight on the subject were

inspirational and motivated me to perform quality research.

Lastly, I would like to thank my mother, father, brother, and my fellow colleagues
for all of their support and patience. Special thanks go out to Darrell, Paul, and Sang-
Yoon, who each made working at the ofﬁce an enjoyable experience in their own special

way. Thanks also go out to Guangda, Frank, Jason, Ben and Jen for their moral support

and friendship.

iv

TABLE OF CONTENTS

LIST OF TABLES .......................................................................................................... vii

LIST OF FIGURES ....................................................................................................... viii

LIST OF SYMBOLS ......................................................................................................... x
CHAPTER 1

INTRODUCTION ............................................................................................................. 1

1.1 Entrance Pressure Effects .............................................................................. 2

1.2 Mooney Method for Determining Slip Velocity ........................................... 3

1.3 Apparent Slip ................................................................................................... S

1.4 Objectives ......................................................................................................... 6
CHAPTER 2

PREVIOUS WORKS ........................................................................................................ 7

2.1 Flow Lines on Injection Molded Parts .......................................................... 7

2.2 Classical Slip .................................................................................................... 8

2.3 Apparent Wall Slip in Capillary Flow .......................................................... 9
CHAPTER 3

MORPHOLOGY OF POLYPROPYLENE / EP RUBBER BLENDS ...................... 12

3.1 Introduction ................................................................................................... 12

3.2 Materials and Methods ................................................................................. 12

3.2.] Experimental .......................................................................................... 12

3.2.2 SEM analysis of morphology .................................................................. 13

3.3 Results and Discussion .................................................................................. 16

3.3.1 Cross-stream migration ......................................................................... 16

3.3.2 Stretching diﬁ’erential ............................................................................. 22

3.4 Conclusions .................................................................................................... 30
CHAPTER 4

RHEOLOGY OF POLYPROPYLENE / EP RUBBER BLENDS ............................. 32

4.1 Introduction ................................................................................................... 32

4.2 Materials and Methods ................................................................................. 32

4.3 Phenomena in Capillary Flow Rheometry .................................................. 32

4.3.1 Start-up pressure transients ................................................................... 32

4.3.2 Nonlinear Bagley plots ........................................................................... 33

4.3.3 Viscous heating eﬁects ........................................................................... 33

4.4 Capillary Testing of Polypropylene 6323 .................................................... 34

4.4.1 Bagley corrections .................................................................................. 34

4.4.2 Mooney analysis ..................................................................................... 34

4.5 Capillary Testing of 80/20 PP/EP Blends .................................................... 38

4.5.1 Bagley corrections .................................................................................. 38

4.5.2 Mooney analysis ..................................................................................... 40

4.5.3 Recoverable strain in blends .................................................................. 44

4.6 Conclusions ................................................................................................................ 46

CHAPTER 5 CONCLUSIONS AND RECOMMENDATIONS ................................ 47

5.1 Conclusions .................................................................................................... 47

5.2 Recommendations for Further Work .......................................................... 48
APPENDICES

APPENDIX A. IMAGE ANALYSIS ............................................................................. 50

A.l Determination of Area Fraction .................................................................. 50

A.2 Determination of Average Aspect Ratio, Z ................................................ 50

APPENDIX B. DETERMINATION OF VOLUME FRACTION .............................. 52

BIBLIOGRAPHY ........................................................................................................... 55

vi

LIST OF TABLES

Table 3.1 Diameters (in mm) of dies used for extruder and capillary rheometer ............. 13

Table 3.2 Experimental conditions of CO-043 blend extrudates obtained for SEM
analysis .............................................................................................................................. 15

Table 3.3 Range of distance (pm) of regions from center in Figure 3.3 .......................... 15

Table 3.4 Area fractions of CO-O43 blend samples for regions indicated in Figure 3.3,

W = 20 from micrographs perpendicular to ﬂow ........................................................... 17
Table 3.5 Average aspect ratios and areas of rubber drops in regions 0-3

for sample B ...................................................................................................................... 22
Table 3.6 Capillary number in regions 0-3 for sample B ................................................. 24

Table 3.7 Average aspect ratios and areas of rubber drops in wall region
for samples A-F ................................................................................................................. 28

vii

LIST OF FIGURES

Figure 1.1 Schematic of a contraction in capillary flow. ................................................... 3
Figure 1.2 Mooney plot for a homogeneous polymer. ....................................................... 4
Figure 1.3 Behavior of a system undergoing lubricated flow. ........................................... 6

Figure 3.1 Micrographs of extrudate cores for (a) CO-043 blend, 7 =1,661 s'1 and (b)

app

MDE 251 blend, ya”, = 1,331 s" at 200°C, D = 2 mm. .................................................... 14
Figure 3.2 Drop size distributions in core for 80/20 blends at 200°C .............................. 14
Figure 3.3 Cross-section of extrudate strand. ................................................................... 15

Figure 3.4 Perpendicular to flow micrographs of regions 1-3 for CO-O43 blend extrudate,

D= 1.5mm, yapp=2000s;l .............................................................................................. 18
Figure 3.5 Flow plane micrographs of true center and regions 1-3 for CO-043 blend
extrudate, D = 1.5 mm, 9"... = 2000 s" .............................................................................. 19

Figure 3.6 Volume fractions from ﬂow plane micrographs for CO-043 blend extrudate,
D = 1.5 mm, 7 = 2000 s", t,.,=70 ms. ........................................................................... 20

“PP

Figure 3.7 Flow plane area distributions for regions 1—3 of CO-O43 blend extrudate,
D=1.5 mm, 7 =2000s". ............................................................................................. 22

“PP

Figure 3.8 Plot of average aspect ratio for CO-043 blend, D = 1.5 mm,
yap, = 2000 s" ................................................................................................................... 24

Figure 3.9 Plot of viscosity ratio versus shear stress ........................................................ 26

Figure 3.10 Viscosity ratio along radial position for CO-043 extrudate, D=1.5 mm,

yap, = 2000:;1 ................................................................................................................... 26
Figure 3.11 Plot of 2G’ versus shear stress for components ............................................ 27
Figure 3.12 Micrographs of wall regions perpendicular to ﬂow in extrudates A-D ........ 29

Figure 3.13 Micrographs of wall regions perpendicular to ﬂow in extrudates E and F... 30

Figure 4.1 Bagley plot for 6323 PP for 1.0—mm diameter data at 200°C ......................... 35

viii

Figure 4.2 Bagley corrections for 6323 PP at 200°C ....................................................... 35

Figure 4.3 Mooney plot for 6323 PP at 200°C, L/D=10 .................................................. 36
Figure 4.4 Mooney plot for 6323 PP at 200°C, UD=2O .................................................. 36
Figure 4.5 Mooney plot for 6323 PP at 200°C, UD=30. ................................................. 37
Figure 4.6 Slip velocities for 6323 PP at 200°C .............................................................. 37
Figure 4.7 Bagley corrections for 80/20 6323 PP/MDE 251 at 200°C ............................ 39
Figure 4.8 Bagley corrections for 80/20 6323 PP/CO-043 at 200°C ............................... 39

Figure 4.9 Mean residence time versus wall shear stress for various die geometries ...... 41
Figure 4.10 Mooney plot for 80/20 6323 PP/MDE 251 at 200°C, IJD=20 ..................... 42
Figure 4.11 Mooney plot for 80/20 6323 PP/CO-043 at 200°C, UD=20 ........................ 42

Figure 4.12 Schematic of dispersed phase concentration variation over radial
position .............................................................................................................................. 44

Figure 4.13 Recoverable strain for 80/20 PP/EP blends at 200°C, UD=30, D=1 mm 45

ix

LIST OF SYMBOLS

Ca : Capillary number.

D : Diameter of die, mm.

EP : Ethylene propylene.

EPDM : Ethylene propylene diene terpolymer.
HDPE : High density polyethylene.

K : Coefﬁcient in power law model for viscosity, Pa - s.
I? : Drop aspect ratio.

L/D : Len gth-to-diameter ratio.

MFR : Melt ﬂow rate index.

M,n : Number-average molecular weight.

Mw : Weight-average molecular weight.

n : Exponent in power law model for viscosity.
NMR : Nuclear magnetic resonance.

AP : Pressure drop, MPa.

APen : Pressure drop due to entrance effects, MPa.
PDI : Polydispersity index.

PMMA : Poly(methyl methacrylate).

PP : Polypropylene.

PS : Polystyrene.

R : Radius of die, mm.

s : Reciprocal of power law model exponent, n.
SBR : Styrene butadiene rubber.

SEM : Scanning electron microscope.

SIM : Shear-induced migration

TEM : Transmission electron microscope.

TPO : Thermoplastic oleﬁn.

ire, : Residence time, ms.

us : Slip velocity, m/s.

Greek Symbols

5 : Thickness of lubricant layer.

4) : Volume fraction.

Yam, : Apparent shear rate, s".

n : Shear viscosity, Pa - s.

'r}c : Viscosity of core in lubricated ﬂow model.
11, : Viscosity of lubricant layer in lubricated ﬂow model
I... : Wall shear stress, MPa.

CHAPTER 1

INTRODUCTION

The occurrence of surface irregularities on injection molded parts has been the
subject of much research in the ﬁeld of polymer processing, particularly in the
automotive industry where the demand for aesthetic quality can rank just as high as the
demand for structural performance. Much debate has ensued as to what causes these
surface defects, with wall slip, the loss of material adhesion between the polymer melt
and the mold wall, being one of the leading theories. In order to better understand the
mechanism by which these defects form, the flow behavior of these materials was
investigated.

Polymers are generally known for their viscoelastic flow behavior, and they are
used in structural applications which require some degree of stiffness. Rubbers, on the
other hand, are somewhat more ﬂexible and are generally tougher. Blending rubber with
polymer thus yields a material which exhibits properties of both rubber and polymer.
Polypropylene blended with ethylene propylene (EP) rubber falls within the class of
materials known as thermoplastic oleﬁns (TPOs). TPOs are of practical importance,
since they are used extensively in the automotive industry for making bumpers and other
car body parts.

A variety of rubbers, such as ethylene propylene diene terpolymer (EPDM) and
styrene butadiene rubber (SBR), have been found to exhibit wall slip in capillary ﬂow
(Geiger, 1989; Moumiac et al., 1992; White et al., 1991; Vinogradov, 1967). Wall slip

has also been reported to occur in the flow of homogeneous polymers such as

polyethylene (Hatzikiriakos and Dealy, 1991, 1992). Slip effects may not be as
pronounced in polypropylene systems because another phenomenon, viscous heating,
may also be present to a considerable degree (Rosenbaum and Hatzikiriakos, 1997). In
capillary ﬂow, for example, the effect of viscous heating increases with increasing die
diameter, while the occurrence of slip has just the opposite dependence on diameter.
Because slip has been shown to occur in both pure rubber and homopolymer
systems, it is reasonable to assume that blends of polypropylene and EP rubber are prone

to similar behavior in capillary flow.

1.1 Entrance Pressure Effects

The dependence of wall slip velocity on wall shear stress for a given melt can be
determined from experiments using a capillary rheometer or with a lab-scale extruder.
Pressure and ﬂow rate data are used to calculate wall shear stress and apparent shear rate,
respectively. Entrance effects of the polymer melt into the die are usually signiﬁcant, and
must be determined in order to calculate the true wall shear stress. In capillary ﬂow, the
magnitude of the entrance pressure effects increases with the contraction ratio, which is

the ratio of the barrel diameter to the diameter of the die. A schematic of this contraction

is shown in Figure 1.1.

-—) (—- Barrel Diameter

 

 

Die Diameter

Figure 1.1 Schematic of a contraction in capillary flow.

The entrance pressure loss, AP“, is obtained by constructing a Bagley plot of
pressure drop against length divided by diameter (UD) with data obtained from dies of at

least two different UD ratios. The pressure drop, AP, at a given apparent shear rate, 70p], ,

will usually increase in direct proportion to increasing UD ratios, such that the apparent
shear rate curves will be linear. Extrapolation of these lines to an L/D of zero will yield

AP“. The true wall shear stress, 1W, is then calculated by

_(AP-AP..) (1-1)

T ,
” 4L/D

1.2 Mooney Method for Determining Slip Velocity

When slip ﬂow occurs, the methods and corrections usually employed for
determining rheological properties in capillary flow experiments must be modiﬁed
considerably (Macosko, 1995; Hatzikiriakos et al., 1992). The following relation may be
derived for the apparent shear rate in isothermal ﬂow of a power law ﬂuid with wall slip

in a capillary of radius R.

7a..=ﬂi+———4 :— (1'2)
” R 3+s K

Here, us is the wall slip velocity, 1,, is the wall shear stress, K is the coefficient of the
power law viscosity model and s is the reciprocal of the power law exponent 11. Above a
critical wall shear stress, the flow curves (of wall shear stress vs. apparent shear rate)
from dies of different diameters begin to diverge, exhibiting a diameter dependence.
Equation (1-2) forms the basis of Mooney plots, which are commonly used to
determine slip velocity in a variety of homogeneous polymer systems. Mooney plots are
plots of the apparent shear rate against l/D at different ﬁxed values of wall shear stress,
for dies of different lengths and radii but the same L/D ratio (Figure 1.2). Extrapolation
of these lines to 1/D = 0 yields the apparent shear rate that would be obtained at a given
shear stress without wall slip. The slip velocity is determined by dividing the slope of the
constant wall shear stress lines by 8. This procedure works well for several grades of
polyethylene but not for polypropylene—presumably because viscous heating has a
greater effect on the polypropylene melt viscosity (Kazatchkov et al., 1995; Rosenbaum
and Hatzikiriakos, 1997). An inherent failing of the Mooney technique is that it neglects
the effects of temperature variations and of high pressure on viscosity, which can lead to
inaccurate slip-velocity calculations. Another complication reported with rubber
compounds is an additional dependence of slip velocity on die geometry (Moumiac et al.,

1992; Malkin et. a1, 1993).

  

 

 

data taken at constant
wall shear stress

7}? z
\ /

C E’

8 .

c6 / z

a , ,x ,
a) [2" z z

5, I,’ ,2? ’,o’

E X I ’ I I I I 8“:

o

a / I I I z '0 1

8: ,9’

<2 / ’

0 ND

Figure 1.2 Mooney plot for a homogeneous polymer.

1.3 Apparent Slip

In multicomponent systems, the mechanism by which slip occurs can be
somewhat more complicated, so much so that the Mooney procedure for inferring slip is
no longer reliable. In these systems, a lubricant layer that can undergo classical slip,
separate from the bulk ﬂow, may form. This lubricant layer can form under one of three
conditions: (1) migration of a lubricating or low viscosity component to the wall; (2)

migration of a low surface energy component to the wall; or (3) inward migration of

drops in ﬂowing emulsions. As illustrated in Figure 1.3, this lubricant layer has some
thickness 5 and an associated viscosity 7]]. An example of the ﬁrst case is the tube ﬂow

of water and oil systems. The second case was illustrated by Chan and Feng (1997) who

<——oo

Lubricated layers
/ viscosity n1

 

Core with viscosity 11c

 

 

Tll<nc

Figure 1.3 Behavior of a system undergoing lubricated ﬂow.

performed a study with an HDPE blend mixed with Dynamar, a ﬂuoroelastomer. They
found that the Dynamar migrated and adhered to the wall, while the bulk melt detached
from this wall layer. The third case is suspected to occur in polymer blends, with the

rubber droplet phase migrating from the wall to the core region.

1.4 Objectives

To investigate the phenomenon of slip that may occur within these polymer

blends, the main objectives of this study are:

1. To examine the morphology of the ﬂowing blends under various extrusion conditions.

2. To relate trends in ﬂow behavior to trends in morphology.

CHAPTER 2

PREVIOUS WORKS

2.1 Flow Lines on Injection Molded Parts

Flow lines, the surface defects or irregularities on injection molded thermoplastic
parts, are most prevalent when the material is a polymer blend, a rubber modiﬁed
thermoplastic or a ﬁlled polymer (Han et al., 1996; Hobbs, 1996). These ﬂow lines
usually appear as alternating bands with different textures and degree of gloss, with the
rough bands having a lower gloss and a duller appearance than the smooth bands
(Hamada, 1996). Because the repeat patterns of the bands are different on the top and
bottom surfaces of the part, these ﬂow lines are suspected to stem from asymmetric
instabilities that may develop in the melt ﬂow (Chang, 1994; Hobbs, 1996). Hobbs
(1996) conducted a study on the effect of gloss reducing additive particles on this
phenomenon and found that although the width of the bands was greater than the particle
size, the onset of these defects occurred at lower wall stresses, presumably due to the
presence of the larger particles in the additive. These surface defects are thus assumed to
be associated with slip ﬂow just as, in extrusion operations, the occurrence of surface
defects in the extrudate, such as “sharkskin” and “melt fracture,” has been related to slip

ﬂow (Denn, 1990, 1992; El Kissi and Piau, 1990; Malkin et al., 1993).

2.2 Classical Slip

In solving the relevant equations for ﬂuid ﬂow, a common boundary condition
that is assumed is the no-slip condition. For a majority of ﬂuids, this assumption is
usually true; however, in the processing of polymer melts, where the shear stress at the
wall can become sufﬁciently high, this condition may not hold (Benbow, 1961). The
high shear stress can induce slip ﬂow, which refers to the loss of adhesion between the
ﬂowing melt and the wall. The onset of slip occurs at some critical wall shear stress, w,
for a given polymeric material.

The techniques for quantifying slip were first developed by Mooney in 1931 (see
Introduction). The underlying assumption in the Mooney technique is that the wall shear
stress, slip velocity, temperature, and pressure gradient are constant along the length of
the die and are independent of diameter.

Numerous authors have conducted slip studies of various homogeneous polymers
to determine the mechanism by which slip occurs in these systems. Shidara and Denn
(1993) found that polystyrene and polyethylene have very different slip characteristics.
Polyethylene was found to exhibit wall slip, while polystyrene did not exhibit slip to an
appreciable degree. Polydispersity also seems to have a major effect on the slip
characteristics of the polymer, as polyethylene with a narrow molecular weight
distribution is more likely to slip than broad molecular weight distribution polyethylene
(Schowalter, 1988). Denn (1990) concluded that the more stretched the polymer chain,

the more likely it is to slip.

2.3 Apparent Wall Slip in Capillary Flow

Apparent wall slip in the capillary ﬂow of various polymer blends has been
investigated extensively by numerous authors. The degree of slip evidenced in blends
appears to depend on three interrelated factors, which may also account for the greater
occurrence of ﬂow lines with these systems. These are signiﬁcant changes in the
morphology of the melt ﬂow, the associated blend rheology, and the different degrees of
adherence of the components or phases to the mold walls.

Rofe and de Vargas et al. (1996) studied apparent wall slip effects in dilute
xanthan solutions using nuclear magnetic resonance imaging (NMR) techniques and
concluded that there were two main factors which correlated with the apparent slip in
these systems. First, they found that the apparent slip velocity was related primarily to
the wall shear stress as validated by their NMR and Mooney analysis experiments.
Second, results from gel permeation chromatography (GPC) strongly suggested that the
degree of slip depended on the molecular weight of xanthan in solution. This was
corroborated experimentally by recirculating the xanthan solution through a syringe
pump. The pump destroyed a large portion of the high molecular weight xanthan and slip
was no longer found to occur.

In suspensions, the mechanism by which apparent slip occurs can be somewhat
more complicated. The particles in the suspension may migrate and establish
concentration gradients, causing significant changes in rheology. Majors et al. (1989)
made the ﬁrst observation of shear-induced particle migration in circular conduits by
using non-invasive NMR imaging techniques. The particles were found to migrate to

regions of low-shear rate in the ﬂow ﬁeld, which caused an overall blunting of the

velocity profile. This blunting of the velocity profile has been corroborated by other
authors (Kamis et al., 1966; Arp and Mason, 1977; Sinton and Chow, 1991; Koh et al.,
1994).

Leighton and Acrivos (1987) proposed a mechanism for migration of suspensions
of neutrally buoyant, mono-modal spherical particles in Newtonian ﬂuids subjected to
inhomogeneous shear ﬂows. The key features of their mechanism are based on
considering the effects of spatially varying viscosity and particle interaction frequency
within the ﬂow field. Utilizing the NMR measurements of the changing concentration
and velocity proﬁles and the ﬂux expressions proposed by Leighton and Acrivos,
workers from MIT and Los Alamos derived a diffusion equation for the particulate phase
(Phillips et al., 1992). The main assumption behind their shear-induced migration (SIM)
model is that the suspension is a Newtonian ﬂuid in which the viscosity is a function of
the volume fraction of solids. An inherent limitation of this model is that it neglects the
torques and pressure drops required to sustain ﬂow. Graham et al. (1998) made the
necessary modifications to the SIM model to account for the effects of non-local ﬁeld
variables, the curvature of the mean particle path and the concentration dependence of the
particle interaction coefﬁcient. The modified model was shown to provide very good
agreement with experimental data by accurately predicting the migration of particles
from the wall to the core. The model was even able to reproduce the plateau region in the
concentration proﬁle near the center as reported by Hampton et al. (1997) in his
migration studies of pressure driven ﬂow in circular conduits.

For emulsions containing non-rigid, deformable particles or drops, the

morphology can be complicated further by the phenomena of drop break—up, coalescence,

10

and stretching. Bousmina and Muller (1998) studied the rheology and morphology of
crosslinked polymethylmethacrylate rubber blends and found that the degree of stretch of
the rubber drops varied going from the center to the wall region. Flow plane micrographs
of the extrudate at various positions from the center of the capillary were obtained using a
transmission electron microscope (TEM). The particles were found to become more
stretched and aligned in the direction of the ﬂow when going from the center to the outer
radius. The alignment in the skin region of the extrudate was to found to result in a
decrease in both viscosity and post-extrusion swell of the blends. Bousmina and Muller

did not, however, report any radial concentration variations (i.e., no particle migration).

11

CHAPTER 3

MORPHOLOGY OF POLYPROPYLENE / EP RUBBER BLENDS

3.1 Introduction

The two-phase of morphology of a system containing viscoelastic drOps dispersed
in a viscoelastic matrix is suspected to have signiﬁcant implications for the ﬂow
behavior. Break-up, migration, and stretching of the drOps are likely to occur, which can
drastically affect the melt rheology of these blends.

In order to gain a better understanding of how the ﬂow behavior can be affected
by these phenomena, the morphology of two blends containing 80/20 weight ratios of
polypropylene and one of two different grades of EP rubber were examined with the aid

of scanning electron microscopy (SEM).

3.2 Materials and Methods
3.2.] Experimental

Blends of polypropylene (MFR=11) and ethylene propylene (EP) rubbers—
EPSYN-MDE 251 (Mw=285,000 and Mw/Mn=1.98) and Dutral CO-043 (Mw=195,000
and Mw/Mn=5.24), were obtained from Montell Polyolefins, Inc. The PP was batch
mixed, by Montell, with the EP in a Banbury mixer in 80/20 weight ratios at 200°C. The
ﬂow experiments were run on a 3A in single-screw extruder attached to a Brabender
plasticorder and also on a Kayeness 6052 DM Dynisco Capillary Rheometer. The dies
used are listed in Table 3.1. The pressure transient was monitored to ensure that steady-

state values of the pressure and ﬂow rate were obtained.

12

Table 3.1 Diameters (in mm) of dies used for extruder and capillary rheometer.

 

 

L/D
10 15 20 30
Extruder --- 1.0, 1.5, 2.0 1.0, 1.5, 2.0 1.0
Capillary Rheometer 0.5, 1.0, 1.5 --- 0.5, 1.0, 1.5 0.5, 1.0, 1.5

 

3.2.2 SEM analysis of morphology

Preliminary SEM micrographs taken of the extrudate cores, revealed that the
blend containing the CO-O43 rubber had a ﬁner dispersion of rubber drops than the blend
containing MDE 251 (Figure 3.1). The drop size distribution was found to be much
narrower in the CO-043 blend than in the MDE 251 blend (Figure 3.2). To facilitate the
development of a predictive model, the CO-043 blend with the narrower distribution was
chosen for further investigation.

To examine the changes in morphology in the melt ﬂow, extrudate strands from
the 0.5-, 1.0-, and 1.5-mm diameter dies of UD=20 at shear rates of 500 and 2000 s’1
were sent to Montell for scanning electron microscope (SEM) analyses. The six
extrudate strands were labeled A-F as indicated in Table 3.2. Cross-sections of the strand
were microtomed and the EP rubber was dissolved with methylcyclohexane, leaving
visible voids in the PP matrix. Locations where micrographs were taken of the core-,
mid-, and wall regions are indicated in Figure 3.3 and Table 3.3. Drop density and area
fraction of the rubber phase in each of the micrographs were determined using Jandel

Scientiﬁc’s Sigma Scan Pro 3.0 image analysis software.

13

 

Figure 3.1 Micrographs of extrudate cores for (a) CO-043 blend, 70", = 1,661 s'

1

and (b) MDE 251 blend, 7..., = 1,331 s" at 200°C, D = 2 mm.

 

yw=1.661 s"

‘ —80/20 PP/O43
‘ - - - - 80/20 PP/251

1 .5
Diameter (p, m)

Figure 3.2 Drop size distributions in core for 80/20 PP/EP blends at 200°C.

Table 3.2 Experimental conditions of CO-043 blend extrudates obtained for SEM
analysis.

 

Diameter, D* Apparent Shear Rate, 1",”,

 

 

 

Sample ID (mm) (8'1)
A 0.5 2043. 1
B 0.5 486.45
C 1 .0 1994.45
D l .0 498.61
E 1 .5 1999.86
F 1 .5 500.87
*Dios have LID=20.

 

Figure 3.3 Cross-section of extrudate strand.

Table 3.3 Range of distance (mm) of regions from center in Figure 3.3

 

 

Extrudate Diameter (mm) 1 2 3
0.5 <100 170-190 230-250
1 <200 300-350 470-500
1 .5 <300 450-500 720-750

 

15

3.3 Results and Discussion
3. 3.1 Cross-stream migration

Micrographs of regions 1-3 for the 1.5-mm diameter sample at an apparent shear
rate of 2000 s‘1 are shown in Figure 3.4 for the CO-043 blend. Flow plane micrographs
of sample B were also taken and are shown in Figure 3.5. (It should be noted that a true
center micrograph was obtained in the ﬂow plane region, while region 1 in the
perpendicular to ﬂow micrograph was confirmed to be ~250 pm from the true center.)

Because of the stretching that occurs in the ﬂow plane, the area fractions in the
perpendicular to ﬂow view are biased. (For a detailed discussion on how the area fraction
was calculated, see Appendix A). As seen in Table 3.4, the radial variations of the area
fraction differ from one extrudate to another. For extrudates A, B, and D the greatest
concentration of rubber appears to be in region 2, the region between the core and the
wall; for extrudates C and F, there appears to be a gradient in the rubber concentration
with the maximum being at the core; and for extrudate E, the concentration gradient is in
the opposite direction, with the maximum concentration of rubber at the wall. The area
fraction at the center of the extrudates (region 1) appears to decrease with die diameter,
suggesting that stretching in the ﬂow plane seems to increase with decreasing die
diameter.

The radial variation of the area fraction of the rubber in each of the samples
provides strong evidence of migration. As the stretching appears to be primarily in the
direction of ﬂow, the volume fractions for extrudate B were determined based on the ﬂow
plane area fractions measured in each region. (See Appendix B for a discussion of how

the volume fraction was determined.) As seen in Figure 3.6, the volume fraction of the

16

Table 3.4 Area fractions of CO-043 blend samples for regions indicated in Figure 3.3,
IJD = 20, from micrographs perpendicular to ﬂow.

 

 

 

D 7W Area Fraction of Drops in Region
Sample ID (mm) (8") 1 2 3

A 0.5 2043.1 0.0916 0.1 10 0.0830
B 0.5 486.45 0.091 1 0.144 0.0863
C 1.0 1994.45 0.1 18 0.1 12 0.0869
D 1.0 498.61 0.135 . 0.155 0.136
E 1.5 1999.86 0.132 0.140 0.151
F 1.5 500.87 0.177 0.130 0.0835

 

17

   

Region 1 Region 2

 

Region 3

Figure 3.4 Perpendicular to ﬂow micrographs of regions 1—3 for CO—043 blend
extrudate, D = 1.5 mm, 7“”: 2000 s".

   

Center Region 1

   

Region 2 Region 3

Figure 3.5 Flow plane micrographs of true center and regions 1—3 for CO-043 blend
extrudate, D = 1.5 mm, 7“”, = 2000 s".

0.24 I-

0.22 I-

0.20 r

0.18 -

0.16 -

Volume Fraction, d>

0.14 -

 

o. 1 2 L l I l A J l I
0 1 00 200 300 400 500 600 700 800

Distance from center, r (urn)

Figure 3.6 Volume fractions from ﬂow plane micrographs for CO-043 blend
extrudate, D = 1.5 mm, 70”, = 2000 s", tres=70 ms.
rubber seems to reach a maximum in region 2, which seems to concur with the trends in
the perpendicular to ﬂow view for all of the other samples, except C and F.

One would expect that the maximum concentration of rubber be at the core and
not in the region between the center and the wall, so the following explanation is offered.
llt has been shown that the dispersion of rigid, spherical particles in suspensions can have
a major effect on the overall ﬂow characteristics of the system, causing such effects as
the blunting of the velocity proﬁle (Sinton and Chow, 1991; Koh et al., 1994) and the
reduction of extrudate swell (Newman and Trementozzi, 1965). It has also been found
that the length at which the ﬂow becomes fully developed depends on the bulk volume
fraction of the dispersed phase and the ratio of the average drop size to the radius of the
die (a/R). Hampton et al. (1997) concluded from their study on the demixing of neutrally

buoyant suspensions of spheres that as both the bulk volume fraction and a/R decreases,

20

the entrance length required for fully developed ﬂow increases. For suspensions of
volume fraction of 0.2 and a/R = 0.0256, they found that an entrance length on the order
of 103 diameters was required for fully developed ﬂow.

In the CO-043 blend, where the bulk volume fraction is about 0.22 and the
average particle size is roughly 1 pm, yielding a maximum value of a/R on the order of
103, the ﬂow can not be assumed to be fully developed. The situation is further
complicated by the fact that the dispersed phase consists of deformable, viscoelastic
drops and not rigid spherical particles. Caution thus must be exercised when interpreting
the concentration profiles (Figure 3.6) obtained from the capillary rheometry of these
blends, as these profiles are not representative of what occurs at steady-state.

Cross-stream migration of the rubber drops is further corroborated by the
differential in the size of the drops in each of the regions. As seen in Figure 3.7, the area
distribution in the parallel to ﬂow view for the wall layer (RegiOn 3) of sample B is
shifted towards smaller-sized drops in comparison to the area distributions in regions 1
and 2. The volume-averaged drop size in region 3 of the ﬂow plane was found to be
smaller than the sizes determined in the other regions (Table 3.5). This suggests the
inward migration of the larger size drops from the wall to the core, since the particles are

moving away from the region of greatest shear stress (i.e., the wall).

21

 

0.3
"t

0.25 -,
E‘ 0.2 _ : Region 1
q, — —Reglon 2
i - - - Region 3
h .
u.

 

 

 

Area (um’)

Figure 3.7 Flow plane area distributions for regions 1-3 of CO-043 blend
extrudate, D = 1.5 mm, 7"... = 2000 s".

Table 3.5 Average aspect ratios and areas of rubber drops in regions 03 for sample E.

 

 

 

 

Region in Region in
Perpendicular to Flow View Flow Plane View
_Quantity 0 1 2 3 0 1 2 3
E v - 1.51 1.56 2.33 8.41 6.27 3.38 1.62
E n — 1.42 1.48 1.88 4.59 3.05 2.29 1.50
g - 1.06 1.06 1.24 1.83 2.06 1.47 1.08
Av (umz) - 0.435 0.149 0.452 1.90 0.630 0.917 0.326
AII (pmz) - 0.212 0.099 0.237 0.632 0.242 0.282 0.163
Apm - 2.05 1.50 1.91 3.00 2.60 3.25 2.00

 

 

 

 

3.3.2 Stretching differential
A number of studies have been conducted on the deformation of viscoelastic
drops in a viscoelastic matrix (Mighri et. al, 1998; Levitt et. al, 1999; Hobbie and Migler,

1999). Bousmina and Muller (1996) found that in the ﬂow plane view of PMMA/rubber

22

blend extrudates, the rubber phase become more stretched and oriented near the core the
wall. This is in stark contrast with what is observed in the ﬂow plane views for sample B.
As seen in Figure 3.5, the drops become more oriented and stretched in the core, where
they assume a ﬁlament-like structure (Figure 3.5). In the wall region, the stretching
yields ellipsoidal structures that are oriented in the vorticity direction, which is
perpendicular to the ﬂow (Figure 3.4). To quantify the degree of stretch occurring in
each of the regions, we deﬁne aspect ratio, which is the ratio of the length of drop to the

diameter of the drop, or

Length of Drop (3-1)
Diameter of Drop

 

Aspect Ratio 2 i =

A plot of the aspect ratio of the drops in the different regions (Figure 3.8) for both
the perpendicular and ﬂow-plane views shows that the maximum stretching is found in
the outer skin region for the perpendicular to ﬂow; however, in the ﬂow plane, the
stretching is the greatest at the center.

It is useful at this point to deﬁne the capillary number, which is a dimensionless
ratio between the viscous stresses exerted on the drop by the external ﬂow ﬁeld and the

interfacial stresses that tend to restore the drop to a spherical shape. This is given as

Ca Tw (r/R) (3-2)

l"/ a

23

9i) -

BI)

71)

BI)

5i)

     
 

-o— Flow Plane

Aspect Ratio

 

 

 

4'0 b + Perpendicular to Flow

31) -

2L) -

1.0 - '

0.0 . . . . 1 . . .
0 100 200 300 400 500 600 700 800

Distance from center, r (um)

Figure 3.8 Plot of average aspect ratio for CO-043 blend, D = 1.5 mm, 1' = 2000 5".

“PP

where Ca is the capillary number, I“, is the wall shear stress, r is the radial position, R is

die radius, a is the average drop size, and P is the interfacial tension. For Ca<1, the
drops are prone to retain their shape, while for Ca>>1 the drops tend to deform or stretch.
Data obtained from the supplier of the rubber shows that F ranges from 0.5-1.0 mN/m.

As seen in Table 3.6, the capillary number is considerably greater than one in all regions.

Table 3.6 Capillary number in regions 0-3 for sample B.

 

Center 1 2 3
Capillary Number 0 42 95 88

Table 3.5 summarizes the weighted- and number-averages of the aspect ratio and

area of each region and also gives the dispersity indexes (PDI) of the distributions, where

24

PDI is deﬁned generically here as the ratio of a weighted-average quantity to a number-
average quantity. The drops in region 2 of the ﬂow plane were not only considerably
larger than the drops in regions 1 and 3, but they also had the greatest variation in size
with a PDI of 3.25. Consequently, for the corresponding perpendicular to ﬂow view, the
drops in region 2 were found to be the smallest when compared to the drop sizes in
regions 1 and 3, which suggests that the majority of the stretching in this region occurred
in the ﬂow plane.

The stretching at the center, shear-free zone of the extrudate warrants further
explanation. A plot of the viscosity ratio of the EP rubber phase to the polypropylene
matrix versus shear stress shows that the rubber phase has a consistently greater viscosity
than the polypropylene for a wide range of shear stresses, with the MDE 251 having a
higher viscosity ratio than the CO-043 (Figure 3.9). Plotting the viscosity ratio versus
radial position for extrudate E (Figure 3.10) reveals that the viscosity ranges from 1.8 at
the center to 30 near the wall. A number of studies conducted on drop break-up in purely
Newtonian systems in simple shear ﬂow (Grace, 1982; Torza et al., 1972; Karam and
Bellinger, 1968) have shown that drop break-up occurs in these systems at viscosity
ratios of less than 4. At the center of extrudate E, however, there is no evidence of
substantial drop break-up, but stretching is evident.

At low shear stresses, the normal stress difference can be approximated by two
times the storage modulus G’((0). A plot of 2G’ versus shear stress (Figure 3.11) shows
that both rubbers have a consistently lower normal stress than the 6323 PP. Despite the

low viscosity ratio of the CO-043 rubber at the center of the extrudate, the lower normal

25

Viscoslty R8110, “EJ116323 pp

1000

 

 

 

 

' I
D
’ I
I
1 I
00 IE. I ’
I I o
. I I O
b I l . .
I I . . EP = 00'043
10% . I I ' . IEP=MDE251
: o
, O
I . Q .
n O
l- O 9 . .
1 I I I III-II I I A II4_III I L I‘LL-II
1000 10000 100000 1000000
Shear Stress, 1: (Pa)
Figure 3.9 Plot of viscosity ratio versus shear stress.
35 l.
O
30 -
25 l’
.2 o
‘5
C 20 '
e .
8 15 -
.2
> O
10 '
O
. O
5 r 0
at” ° 0
0 I I L I I I I I I
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9

rIR

Figure 3.10 Viscosity ratio along radial position for CO-043 extrudate,
D=l.5 mm, 7'... = 2000s'l .

26

 

 

1.E+06 g
1.E+05 I ‘ ‘
E ' A
- 06323 PP . o .
1.E+04 7 . I
E '382’22. . ° -
a » ‘ . . *
o
r 1.E+O3 E . ‘ ‘
g ’ 0 A I
. o ‘ I
1.E+02 I ‘ : I
E I '
1.E+01 I
1E+00 A . . .A.I.n A I ......1 1 . . ..‘.‘.
1.E+02 1.E+03 1.E+04 1.E+05
Shear Stress (Pa)

Figure 3.11 Plot of 2G’ versus shear stress for components.

stress of the rubber as compared to that of the PP matrix, makes the rubber drops prone to
stretching in the direction of the ﬂow.

The average aspect ratios and areas for the rubber drops in the wall regions for
samples A-F are listed in Table 3.7. The most obvious trend is that the average area or
size of the drops tend to increase with increasing diameter at a given shear rate. The PDI
of the area distributions appears to be both a function of diameter and shear rate. For a
shear rate of 2000 s", the PDIs of the area distributions seems to remain constant
regardless of diameter; however, at the lower shear rate, the 0.5-mm diameter was found
to have a narrower distribution of drop sizes than what was observed at the larger
diameters. The drop aspect ratio, as expected, was shown to increase with shear rate.

Another feature that was prominent in the wall region of each the extrudates was

the elongation of the rubber drops in the vorticity direction, which is perpendicular to the

27

Table 3.7 Average aspect ratios and areas of rubber drops in
wall region for samples A-F.

 

 

 

 

 

 

Yapp = 5008.1 Yam, : 20008.1
Quantity 0.5 mm 1.0 mm 1.5 mm 0.5 mm 1.0 mm 1.5 mm
Diameter Diameter
7,, (MPa) 0.098 0.099 0.097 0.165 0.157 0.152
E v 1.84 2.01 2.18 2.50 2.22 2.33
i n 1.62 1.75 1.86 2.03 1.83 1.88
g 1.14 1.15 1.17 1.23 1.22 1.24
Av (umz) 0.141 0.376 0.570 0.149 0.157 0.452
An (m2) 0.085 0.180 0.275 0.076 0.082 0.237
Apm 1.66 2.09 2.07 1.96 1.91 1.91

 

 

ﬂow (Figures 3.12-3.13) This phenomenon was reported by Hobbie and Migler (1999) in
their study of shear- induced deformation of dilute emulsions of viscoelastic polymer
melts and found to occur at some critical shear stress where the normal stress of the
dispersed phase and of the matrix matched. Using 2G’ as an estimate of the normal
stress, extrapolation of the curves for PP and the two rubbers was found to overpredict
the wall shear stress at which vorticity elongation occurs in these blends. This suggests

that 2G’ is a poor estimate of the normal stress at high shear stresses.

28

 

150 m 15.0 um

Sample A, 7“", =2043 s", D=0.5 mm Sample B. 7,,” =486 S". D=0.5 mm

 

 

Sample C, 7W =1994 s", D=1.0 mm Sample D, 7“” =499 s", D=1.0 mm

Figure 3.12 Micrographs of wall regions perpendicular to ﬂow in
extrudates A-D.

29

15.0 um

      

Sample 13, 7am, =2000 s", D21 .5 mm Sample F, I'm, =5008'1. D=1-5 mm

Figure 3.13 Micrographs of wall regions perpendicular to ﬂow in
extrudates E and F.

3.4 Conclusions

This study has shown that in emulsions of EP rubber dispersed in a PP matrix, the
rubber drops undergo both migration and stretching along different directions as one
moves from the center of the capillary to the outer radius. From the ﬂow plane
micrographs of the 1.5-mm diameter extrudate taken at an apparent shear rate of 2000 s",
the maximum concentration of rubber drops was found to be in the region between the
center and the wall; however, caution must be exercised when interpreting these results as
it was shown that the ﬂow was not fully developed. Stretching was found to yield
ﬁlament-like structures in the ﬂow plane view at the center or shear-free zone of the
extrudate, while the maximum amount of stretching in the perpendicular to ﬂow view

seemed to occur in the wall region, where the drops assume an ellipsoidal-like structure.

30

The sizes of the drops in the wall region were found to increase with die diameter and
with shear rate and elongation of the drops took place in the vorticity direction.

A reasonable estimate of the volume fractions could not be obtained for the other
extrudates, since stretching may also be occurring in the ﬂow plane of these samples,
resulting in a distorted view of the true drop size in the perpendicular to ﬂow view.
Further work would entail obtaining a more detailed picture of the 3-D structure of the

rubber drops to gain a better estimate of volume fraction.

31

CHAPTER 4

RHEOLOGY OF POLYPROPYLENE / EP RUBBER BLENDS

4.1 Introduction

In this chapter, the rheology of the blends is discussed in light of the information
gained from the morphology study. Speciﬁcally, a connection is made between the
phenomena of drop migration and deformation to apparent slip, viscosity and extrudate

swell. Neat polypropylene is discussed ﬁrst as a reference case.

4.2 Materials and Methods

Neat polypropylene (MFR=11) and blends of polypropylene and ethylene
propylene (EP) rubbers—EPSYN-MDE 251 (Mw=285,000 and Mw/Mn=1.98) and Dutral
CO-043 (Mw=195,000 and Mw/Mn=S.24), were obtained from Montell Polyolefins, Inc.
The PP was batch mixed with the EP in a Banbury mixer in 80/20 weight ratios at 200°C.
The ﬂow experiments were run on a Kayeness 6052 DM Dynisco capillary rheometer.

The dies used are listed in Table 3.1.

4.3 Phenomena in Capillary Flow Rheometry
4.3.1 Start-up pressure transients

Hatzikiriakos and Dealy (1994) performed a study on the start-up pressure
transients in the capillary rheometry of various polyethylenes and found that the time

required to achieve a steady pressure reading at constant piston speed, also known as the

32

“rise time,” increases with UD and the amount of polymer in the reservoir and decreases
with diameter and piston speed.
In all of the ﬂow tests performed in this study, the pressure transient was

monitored to ensure that data obtained was at steady state.

4.3.2. Nonlinear Bagley plots

The entrance pressure effects were determined using the Bagley method as
described in Section 1.1. Curvature may occur in the Bagley plot if the viscosity of the
system has a signiﬁcant pressure dependency (Laun, 1983; Laun and Schuch, 1990), or if
the slip velocity is strongly affected by pressure (Hatzikriakos and Dealy, 1992). In
either case, care must be taken in extrapolating the Bagley plot curves to obtain the

proper ends corrections.

4.3.3 Viscous heating eﬁ‘ects

Another phenomenon than can occur in the capillary ﬂow of polymer melts is
viscous heating effects. When a viscous material is deformed in a ﬂow ﬁeld, some of the
work of deformation is converted into thermal energy, which can cause a substantial rise
in temperature within the ﬂow. The high viscosities and low thermal conductivities of
polymers make them especially prone to viscous dissipation. Hatzikiriakos and Dealy
(1997) concluded from their numerical analyses of capillary ﬂow undergoing slip and
viscous heating effects, that viscous heating increases with apparent shear rate and die

diameter. Since the Mooney technique neglects the effects of temperature and pressure,

33

data obtained from systems that undergo a significant amount of viscous dissipation
appear as curved lines on the Mooney plot. Slip effects can therefore be masked by
viscous heating in eliminating the diameter-dependency of the ﬂow curves (Katzatchkov

and Hatzikiriakos, 1997).

4.4 Capillary Testing of Polypropylene 6323
4.4.1 Bagley corrections

The Bagley plots for all L/D’s were slightly curved especially at high shear rates,
indicating that the slip velocity has some dependence on pressure (Hatzikiriakos and

Dealy, 1992). The ends correction AP," was thus estimated by a second-order

polynomial extrapolation of the data to the y—axis (Figure 4.1 shows the Bagley plot for
the 1.0-mm diameter data). A plot of the Bagley corrections (Figure 4.2) shows that the

AP," are a diameter dependent, which is due to the different centraction ratios for each

die diameter.

4.4.2 Mooney analysis

Plots of the corrected data on Mooney coordinates are shown in Figures 4.3-4.5.
Estimates of the slip velocity as a function of wall shear stress were determined from
the 1.5- and 1.0-mm diameter data (Figure 4.6). Slip velocity was found to decrease with
increasing UD, a trend also reported to occur with high density polyethylenes (HDPEs)
by Hatzikiriakos and Dealy (1992). Since viscous heating has also been determined to
occur in the capillary ﬂow of polypropylene, some of the reduction in the “apparent” slip

velocity may be due to the increasing effect of viscous heating with increasing [JD

34

Pressure Drop, AP (MPa)

Bagley Correction (MPa)

25-

 

 

5 10 15 20 25
LID

Figure 4.1 Bagley plot for 6323 PP for 1.0-mm diameter dies at 200°C.

 

30

78091 S"
0 4998.29
I 3222.74
A 2079.58
X 1349.90
X 863.45
0 559.42
+ 364.84

- 231 .06
— 145.94
0 97.29

 

 

 

 

35 40

 

10 L'
’ I
’ I
. .I
OI
e
1 L o . I ‘ 00.5mm
. , A
i I l1.0mm
. o I A‘A A1.5mm
. o .
. ‘A
. ' A
o
‘l
0 ll— I I I I I A I I I I I I I I LI
0.01 0.1 1

Wall Shear Stress (MPa)

Figure 4.2 Bagley corrections for 6323 PP at 200°C.

35

Apparent Shear Rate, 7.” (1Is)

Apparent Shear Rate, 7.9, (1Is)

4500 -

 

 

 

 

 

 

 

 

 

 

 

 

4000 - I... MPa
3500 _ - o 0.060
- I 0.070
3000 , A 0.080
_ ' x 0.090
2500 , + X 0.100
o 0.110
2000 . 2 + . + 0.120
- -0.125
1500 . + ° x -0.130
O 0.133
1000 P ; x X
x x ‘
500 - A ‘ :
! 9
O J I I l I
0 0.5 1 1.5 2 2.5
110 (mm")
Figure 4.3 Mooney plot for 6323 PP at 200°C, L/D =10.
3500 A.
3000 » . ‘w' MP3
0 0.060
2500 _ no.070
" _ A0080
x 0.090
2000 - O " + 3110.100
- + O 0.110
1500 _ + . +0120
0 -0.125
x -0.130
1000 - 0 x o 0.134
x x x
x
500 - ‘ A :
2 2 .
o l l l l I
0 0.5 1 1.5 2 2.5
110 (mm“)

Figure 4.4 Mooney plot for 6323 PP at 200°C, UD =20.

36

g §

§

Apparent Shear Rate, 7.” (tie)
" 5
§ 8

§

1000

Slip Veloclty (mm/s)

 

 

 

 

 

100

10

 

 

1‘”, MPa
0 ° o 0.062
30.070
, _ " A 0.080
- x 0.090
- - + x 0.100
- + 0 0.110
+ + 0.120
. o -0.125
c -0.130
x x X o 0.1339J
X
x X ‘
A
A I I
9 o ’
0.5 1 1.5 2 2.5
110(mm")
Figure 4.5 Mooney plot for 6323 PP at 200°C, UD =30.
O
O
O I
o u0=10 0 ° I '
- u0=20 I I ‘
: AL/D=30 ° ‘ A ‘
I ' A
E a .
_ I
‘ A
5 I
.7 IA
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16

Wall Shear Stress (MPa)

Figure 4.6 Slip velocities for 6323 PP at 200°C.

37

 

4.5 Capillary Testing of 80/20 PP/EP Blends
4. 5.1 Bagley corrections
The Bagley plots for both blends at all three diameters were curved, so that a

second-order polynomial ﬁt was used to obtain AP“. A plot of AP versus wall shear

(ll

stress shows that AP”, for the MDE 251 blend has a similar dependence on contraction

ratio as does the neat PP (Figure 4.7), with the small dies (large contraction ratio) having

a greater Bagley correction than the large dies (small contraction ratio).

The plot of AP versus wall shear stress for the CO-043 blend, however, does not

show any consistent trends with contraction ratio. The effects of morphology are
believed to have a significant effect on the entrance pressure for this particular system.
Han et al. (1995) concluded in his study of polystyrene (PS)/poly(methyl methacrylate)
(PMMA) blends, that because of the significant deformation of the dispersed PMMA
droplets as they entered the entrance region of the die, the use of Bagley corrections in
predicting wall shear stresses in blends is invalid. The data presented from this point on

are not corrected for entrance pressure effects.

38

Bagley Correctlon (MPa)

 

 

10:
b
b ..
r o I
O.-
b
.OI‘
. I
...A 00.5mm
1 i . . ‘ ILOmm
' I At5mm
D “
P I,
h A
p 5“
p
01 # A A LLAAA‘ 9 A A A AAAAI
001 DJ 1

Bagley Correction (MPa)

Wall Shear Stress (MPa)

Figure 4.7 Bagley corrections for 80/20 6323 PP/MDE 251 at 200°C.

 

 

10 :
b
b
L I I
O
i I 9
O
. Ir ‘
. A
A “ 00.5mm
1 ‘ A I . I1.0mm
- O
, A1.5mm
p ‘ ‘ ‘
. I
O
. I
‘ o
' o
O
b
O A A A A A A A L. A A A A A A A A
0.01 0.1

Wall Shear Stress (MPa)

Figure 4.8 Bagley corrections for 80/20 6323 PP/CO—043 at 200°C.

39

4.5.2 Mooney analysis

Although the Mooney plot can not be used in this case to infer slip, re-plotting the
uncorrected data on Mooney coordinates can provide some insight as to how the
morphology is changing with respect to die diameter and apparent shear rate. The main
changes in morphology that are expected to occur are migration and stretching of the
rubber drops. Before proceeding, it is useful to deﬁne mean residence time, which is

represented as

 

5 31/0 (4-1)
a

where L is the length of the die, D is the diameter of the die, L? is the mean velocity, and

7“,”, is the apparent shear rate. Plotting the mean residence time versus wall shear stress

(Figure 4.9) shows that the residence times are lowest for the smallest [JD and that for a
given IJD, the smallest diameter has the longest resident time. As residence time varies
inversely with shear rate, less migration is expected to occur at high shear rates than at
low shear rates. In terms of stretching, high shear rates lead to more elongation,
especially in the smaller diameter dies.

Figures 4.10 and 4.11 show the Mooney plots for UD=20 of the uncorrected
MDE 251 and CO-O43 blend data. The trends between the blends are similar as both
plots show the same type and degree of curvature, even though the MDE 251 was found
to have a wider drop size distribution than the CO-043 (Figures 3.2 and 3.3). The
progression of increasing apparent viscosity (decreasing apparent shear rate) with
decreasing die diameter suggests that apparent slip in both blends increases with

diameter. As shown in Chapter 3, the wall region in all of the CO-043 extrudates showed

40

10000

1000

+1.5 mm, 15 mm
+1.0 mm, 10 mm
+0.5 mm, 5mm
100 +1.5 mm, 30mm
+1.0 mm, 20 mm
+0.5 mm, 10 mm
+1.5 mm, 45 mm
10 —-—1.0 mm, 30mm
+0.5 mm, 15 mm

Mean Residence Time (ms)

1 A A A AAAAAI A A A AAAAAI

0.010 0.100 1.000
Wall Shear Stress (MPa)

Figure 4.9 Mean residence time versus wall shear stress for various die geometries.

41

Apparent Shear Rate, 7.9..(118)

Apparent Shear Rate, 7.9,,(113)

2500

2000

1 500

1 000

500

 

0.5

+

OIDXX O

'0IDXX.+

—A

1ID ("1m")

1

1.5

1... MPa

 

 

00.076
I0.090
A0.100
x 0.110
X0120
00.130
+0.140
-0.150
-0.160
00.170

 

 

2.5

Figure 4.10 Mooney plot for 80/20 6323 PP/MDE 251 at 200°C, UD=20.

1400

1200

1000

800

600

400

200

 

OIDX X

'OIDX X

+

DOIDXX

1... MPa

 

o 0.050
I 0.060
A 0.070
x 0.080
x 0.090
o 0.100
+ 0.110
- 0.120
- 0.130

 

 

 

 

0.5

1ID (mm")

1.5

N

Figure 4.11 Mooney plot for 80/20 6323 PP/CO-043 at 200°C, UD=20.

42

2.5

a depletion of rubber drops. Migration of the rubber drops is believed to occur more
readily in the large-diameter dies than in the small—diameter dies, especially at high shear
rates, since more stretching is expected to occur in the small diameter die.

From the work of Phillips et al. (1992), who developed a constitutive equation for
concentrated suspensions that undergo shear-induced migration, it has been shown that
there are two main mechanisms for particle flux in these systems. These are the particle
ﬂuxes due to the collision frequency and the viscosity gradient, which are denoted as N,

and M, respectively. The expressions for the ﬂuxes are given as

a . 4-2
NC : —Kca2¢a_(¢7) ( )
r

2 - 2
Nn =—Kn a W dn (4-3)
7] dr

 

where a is the particle radius, 0 is the volume fraction, 11 is the suspension viscosity, and
7 is the shear rate. The constants K, and K n are the model parameters determined from
experiment. The flux due to collisions NC is in the direction of decreasing shear rate, so
that drop migration is driven towards the center, shear-free zone of the die, while the flux
due to the viscosity gradient N" is in the opposite direction, so that the drops are driven to
regions of lower viscosity (i.e. the wall).

The suspension viscosity, 1], depends not only on the drop concentration, but is
also a function of drop aspect ratio. Stretching at the core is expected to occur to a

greater degree in the small diameter die, since the large contraction ratio yields more

extensional flow. The flux due to the viscosity gradient will be significantly greater than

43

the ﬂux due to collisions in small diameter dies as compared to large diameter dies. This
translates to an overall lesser migration of the drops from the wall to the core in small
diameter dies, so that the maximum rubber concentration lies closer to the wall in a small
diameter die than in a large diameter die (Figure 4.11). Less migration means less
depletion of rubber from the wall layer and thus a higher apparent viscosity (i.e., less

apparent slip).

small diameter die

     

large diameter die
<I>(r)

 

 

Figure 4.12 Schematic of dispersed phase concentration variation over
radial position.

4. 5.3 Recoverable strain in blends

Extrudate swell is the response of a polymer melt to recover strain and return to
its original shape on emergence from the capillary. Data on the extrudate swell of 6323

PP and the 80/20 PP/EP blends was converted to recoverable strain as follows
(I 2
Recoverable strain = [B] —1 (4'5)

where the d is the diameter of the extrudate as it exits a die of diameter, D.

As seen in the plot of recoverable strain versus wall shear stress (Figure 4.12), the blends

have a characteristically lower extrudate swell than the 6323 PP. Newman and

Trementozzi (1965) conducted a study on the inﬂuence of ﬁllers on extrudate swell in

poly(styrene-acrylonitrile) and found that the extrudate swell in these systems decreased

with increased ﬁller content. The presence of ﬁller was also found to increase the

tendency for plug ﬂow by developing a region of low filler concentration at the surface,

causing a significant reduction in the local viscosity from that of the interior, and also by

the “trapping” of the matrix polymer. The more stretched the rubber drops, the more

constrained the matrix, so that less material is able to recover from the imposed strain.

(d/Dlz-1

 
  
  

3 .
2.5 , +6323 PP
-o- MDE 251 Blend
—8— CO—043 Blend
2 .
1.5 -

0.5 _ k/KMB/ﬂ

0 l I I L l l l l

0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20
Wall Shear Stress, 1., (MPa)

 

Figure 4.12 Recoverable strain for 80/20 PP/EP blends at 200°C, UD =30,
D: 1 mm.

45

4.6 Conclusions

The Bagley plots for neat polypropylene and the two blends were found to be
slightly curved, and were thus ﬁtted with a second-order polynomial in determining the
ends correction. For neat polypropylene and the blend containing MDE 251, the Bagley
correction was found to depend on contraction ratio; however, for the CO-043 blend, the
Bagley correction did not follow any consistent trend with diameter.

An estimate of the slip velocity for neat polypropylene showed that slip velocity
decreased with increasing LID, presumably because of the increasing effect of pressure.
For the blends, Mooney plots of the uncorrected data showed a reverse trend in slip
velocity with diameter, as the large diameters exhibited more slip than the smaller die
diameters. Greater apparent slip was found in the larger diameters, where more migration
is expected to occur. The amount of migration that occurs is explained in the terms of
two competing ﬂuxes, the ﬂux due to viscosity variations and‘the ﬂux due to collision.
As viscosity depends on the aspect ratio of the drops, the ﬂux due to viscosity is expected
to be significantly greater than the ﬂux due to collisions in a small diameter die, where
more stretching is expected to occur. This leads to a higher viscosity barrier in the core
region for migration, so that less migration occurs in the small diameter.

The recoverable strain of the blends were found to be considerably lower than that
of neat PP. The reduction in the recoverable strain of the blends is attributed to the
reduction in the local viscosity in the wall layer due to migration of the rubber drops, and

by the constraining of the matrix by the stretched rubber drops.

46

CHAPTER 5

CONCLUSIONS AND RECOMMENDATIONS

5.1 Conclusions

In the extrusion of blends containing polypropylene and EP rubber, the following
trends in morphology were observed:

1. Migration of large rubber drops from the wall to the core leading to a

depletion in the wall region.

2. Maximum stretching of the drops at the center or shear-free zone of the die,

due mainly to the normal stress differences between the rubber and the matrix.
This is in contrast with what Bousmina and Muller (1996) found in their study
of PMMA/rubber blends that the greatest stretching occurred in the wall
region. V

3. Stretching in the wall region along the vorticity direction.

Flow tests conducted on neat polypropylene showed that the Bagley corrections
increased with increasing contraction ratio. The slight curvature in the Bagley plot was
attributed to the pressure dependence of slip velocity, which decreased with increasing
UD. Curvature observed in the Mooney chart was also attributed to the pressure
dependence of slip velocity and to possible viscous heating effects.

In the MDE 251 blend, the Bagley corrections were found to increase with
increasing contraction ratio; however, no discernible trend was found to occur in the
blend containing CO-043. Mooney plots of the uncorrected data for the blends showed

that greater apparent slip occurred in the larger diameter die, suggesting that migration

47

occurs more readily in the larger diameter die. This can be explained in terms of N,- and
N", the ﬂuxes due to collision between the EP rubber drops and the viscosity gradients,
respectively. In the smaller diameter die, more stretching at the center is expected to
occur, since the large contraction ratio yields more extensional ﬂow. Since viscosity

increases with aspect ratio, there is a greater viscosity barrier for migration in the small

diameter die.

5.2 Recommendations for Further Work

The ultimate goal of this research is to come up with a model which relates trends
in rheology to trends in morphology in these blends. To do so, requires a comphrensive
understanding of how the morphology changes under different conditions (i.e., die
diameter, shear rate, etc.). Further work should focus on obtaining a more detailed

picture of the drop deformation and migration under different conditions.

48

APPENDICES

APPENDIX A

IMAGE ANALYSIS

A.1 Determination of Area Fraction

The SEM micrographs were scanned into I PEG image files and analyzed using
Sigma Scan 3.0 from Jandel Scientiﬁc. The areas of the voids left from dissolution of the
EP rubber were determined by one of two methods. In regions 1 & 2 (Figure 3.3 and
Table 3.3), where the voids were relatively circular, the diameters of the voids were

measured and the areas were determined by

71'D2

 

Area of Void i = A, =

For region 3 and the ﬂow plane micrographs for sample E (Figure 3.3 and Table 3.3),
where the voids took an irregular shape, the actual areas were measured using the
software. After a sufficient number, n, of voids were counted (2 800), the total area
occupied by the measured voids, AM, was determined and the area fraction of the CO-O43
rubber phase was calculated by

ii

. ,-= A
Total Area Fraction = '—"’—'—

IO!

A.2 Determination of Average Aspect Ratio, 6
A feature in Sigma Scan 3.0 allows measurement of the major and minor axis
lengths of object areas. The major axis is deﬁned by searching all the border pixels of an

object and choosing the two pixels that are the farthest apart. Once a major axis is

50

defined, the minor axis can be calculated. Minor axes are drawn between two pixels
deﬁning the longest line perpendicular to the major axis. The aspect ratio for a void is

then calculated by

 

I? _ major axis length
minor axis length I.

The average aspect ratio is then determined by

n A
e, = e, '
12:; [Amt ]

 

51

APPENDIX B

DETERMINATION OF VOLUME FRACTION

The volume fraction is related to the area fraction by some factor a, or

’1
2’5"}
j=l

A

It)!

 

X61

To determine a, one must solve the relevant mass balance for ﬂow through a capillary.
For a capillary of radius R, the integral mass balance written in terms of volume

fraction ¢(r) is
2 R —
—R2 Iv(r)¢(r)rdr = 17¢

where v(r) is the velocity proﬁle, 17 is the mean velocity, and (3 is the bulk volume
fraction. The volume fraction can be written in terms of area fraction as
¢(r) = 61¢. (r)

Substitution of this expression into the mass balance, yields
R—a20 Rv(r)¢A (r)rdr = 1747

To non-dimensionalize, we let

Re-writing the mass balance in non-dimensional form and solving for a gives

'v(€) — 4
2a —_—¢.(€)r:d§=¢=>a= .
1 V ZJV(§)¢(

17

 

(C XdC

52

V(C )

For plug-ﬂow, T = 1 so that
v

a : ¢
216(4):“

 

For non-plug ﬂow through a long tube, the velocity profile cannot be assumed to be 1.

The velocity proﬁle uz (r) and volumetric ﬂow rate Q for a power law ﬂuid are given as

"R [RAPj/nlr [r )lH/n]
uz(r)= — l— —
l+n 2KL R

_ er3 [ RAP j“

(Middleman, 1977)

 

 

—1+3n 2KL

The mean velocity is determined from

 

Q nR (RAP j”

17: :
ItR2 l+3n 2KL

Dividing the velocity proﬁle by the mean velocity yields

u—z— 1+3" 1— L l+l/n
L7 l+n R

The value of a was determined for both plug and normal ﬂow. For normal ﬂow, the

 

value of n was set to 0.33, the value for 6323 polypropylene. The values of a for normal
and plug ﬂow were 1.26 and 1.31, respectively. Since plug ﬂow has been shown to occur
in ﬁlled systems (see Chapter 2), the plug ﬂow value for a was used in determination of

volume fraction.

53

BIBLIOGRAPHY

 

Bibliography

1. Arp, RA. and S.G. Mason, J. Colloid Interface Sci., 61, 44-61 (1977).
2. Benbow, J.J., R.N. Charley, and P. Lamb, Nature, 192, p.223 (1961).
3. Bousmina, M. and R. Muller, Rheol. Acta, 35 (4), 369-381 (1996).

4. Chan, C.-M. and J. Feng, J. Rheol., 41 (2), 319 (1997).

5. Chang, M.C.O., SPE Tech Papers, 40, p.36 (1994).

6. Denn, M.M., Ann. Rev. Fluid Mech., 22, 13-34 (1990).

7. Denn, M.M., Theoretical and Applied Rheology, eds. P. Moldenaers and R.
Keunings, p. 45-49 (1992).

8. El Kissi, N. and J.M. Piau, J. Non-Newt. Fluid Mech., 37, 55-94 (1990).
9. Geiger, K., Kautsch und Gummi Kunstst B, 19 (7), 507 (1989).
10. Grace, H.P., Chem. Eng. Commun., 14, 225-277 (1982).

l 1. Graham, A.L., A.A. Mammoli, and MB. Busch, Rheol. Acta, 37 (2), 139-150
(1998).

12. Hamada, H. and H. Tsunasawa, 1. App. Polym. Sci., 60(3), 353-362 (1996).

13. Hampton, R.E., A.A. Mammoli, A.L. Graham, N. Tetlow, and SA. Altobelli, J.
Rheol., 41 (3), 621-640 (1997).

14. Han, C.Y., M.D. Wolkowicz, D. Dong, and SS. Dagli, SPE Tech Papers, 42,
(1996).

15. Han, J.H., C.C.-Feng, D.-J. Li, and CD. Han, Polymer, 36 (12), 2451-2462 (1995).
16. Hatzikiriakos, S.G. and J.M. Dealy, J. Rheol., 35 (4), 497-523 (1991).

17. Hatzikiriakos, S.G. and J.M. Dealy, J. Rheol., 36 (4), 703-741 (1992).

18. Hatzikiriakos, S.G. and J.M. Dealy, Polym. Eng. Sci., 34 (6), 493-499 (1994).
19. Hobbie, E.K. and KB. Migler, Phys. Rev. Letters, 82 (26), 5393-5396 (1999).
20. Hobbs, S.Y., Polym. Eng. Sci., 36 (11), 1489-94 (1996).

21. Karam, HI. and J.C. Bellinger, Ind. Eng. Chem. Fundam, 7, 576-581 (1966).

55

22.

23.

24.
25.
26.
27.
28.

29.

30.

31

32.
33.
34.
35.
36.

37.

38.

39.
40.
41.

42.

Karnis, A., H.L. Goldsmith, and S.G. Mason, J. Colloid Interface Sci., 22, 531-533
(1966).

Kazatchkov, I.B., S.G. Hatzikiriakos, and C. Stewart, Polym. Eng. Sci., 35 (23),
1864-1871 (1995).

Koh, C.J., P. Hookman, and LG. Leal, J. Fluid Mech., 266, 1-32 (1994).

Laun, H.M. and H. Schuch, J. Rheol., 33, 119 (1990).

Laun, H.M., Rheol. Acta, 22, 171 (1983).

Leighton, D.T. and A. Acrivos. J. Fluid Mechanics, 181, 415-439 (1987).
Levitt, L., C.W. Macosko, S.D. Pearson, Polym Eng. Sci., 36 (12), 1647 (1996).

Macosko, C.W., Rheology: Principles, Measurements, and Applications, VCH,
New York, 1994.

Majors, P.D., A.L. Graham, and M. Gottlieb, Proc. 10th Intl. Congr. Rheology,
Sydney, Australia, 2, 137-139 (1988).

. Malkin, A. Ya, A.V. Baranov, and ME. Vikulenkova, Rheol. Acta, 32 (2), 150-155

(1993).

Middleman, S., Fundamentals of Polymer Processing, p. 88 (1977).

Mighri, F., A. Ajji, P.J. Carreau, J. Rheol., 41: (5) 1183-1201 (1997).

Mooney, M., J. Rheol., 2, 210 (1931).

Moumiac, Ph., J.F. Agassant and B. Vergnes, Rheol. Acta, 31 (6), 565-574 (1992).
Newman, S. and Trementozzi, Q.A., J. App. Polym. Sci., 9, 3071-3089 (1965).

Phillips, R.J., R.C. Armstrong, and RA. Brown, J. Non-Newt. Fluid Mech., 64,
157-171(1996).

Rofe, C.J., L. de Vargas, J.P.-Gonzalez, R.K. Lambert, and RT. Callaghan, J.
Rheol., 40 (6), 1115-1128 (1996).

Rosenbaum, EB, and S.G. Hatzikiriakos, AIChE J., 43 (3), 598 (1997).
Schowalter, W.R., J. Non-Newt. Fluid Mech., 29 (1-3), 25-36 (1988).
Shidara, H. and M.M. Denn, J. Non-Newt. Fluid Mech., 48 (1-2), 101-110 (1993).

Sinton, SW. and AW. Chow, J. Rheol., 35, 735-772 (1991).

56

 

43. Torza, S., R.G. Cox, and S.G. Mason, J. Colloid Interface Sci., 38, 395-41 1 (1972).
44. Vinogradov, G.V. and LI. Ivanova, Rheol. Acta, 6 (3), 209-222 (1967).

45. White, J.L., M.H. Han, N. Nakajima, and R. Brzoskowski, J. Rheol., 35 (1), 167-
189 (1991).

57