LIBRARY
Michigan State
University

 

 

 

PLACE ll RETURN BOX to romovo this chockom from your "cord.
TO AVOID FINES return on or bdon duo duo.

DATE DUE DATE DUE DATE DUE
Mm l E 283

         
  

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

U1
El

 

 

 

 

 

 

fl

MSU Is An Afflrmutivo Adm/Emmi Oppoctunily Intuition
m

 

m1

STUDIES ON CELLULOSE ACETATE/POLY(ETHYLENE

TEREPHTHALATE) BLENDS

By

Ajay Gupta

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

MASTER OF SCIENCE

DEPARTMENT OF CHEMICAL ENGINEERING

1994

ABSTRACT

STUDIES ON CELLULOSE ACETATE/POLY(ETHYLENE TEREPHTHALATE)
BLENDS

By
Ajay Gupta

Blends of Cellulose Acetate (CA) and poly(ethylene terphthalate) (PET) have been
prepared and compatibilized by reactive extrusion. The effect of composition, extent of
transesterification, method of extrusion and addition of plasticizer on the morphology was
studied. In the first part ot the work, viscosity of unplasticized CA has been estimated
to be of the order of 1011 Pa.s. Isothermal crystallization studies were done on CA blends
with PET bottle recycle to determine optimum processing conditions. Blends of CA and
glycol modified PET were compatibilized via traneseterification reaction and the extent
of reaction was characterized by Soxhlet extraction and electron microscopy. Optimum
particle size for impact resistance was concluded to be ~ 1 micron. Fibrillar morphology
was found desirable for high tensile properties. Maximium tensile strength of 12,000 psi,

tensile modulus of 0.55 Mpsi, and Izod impact strength of 1.27 ft.lb/ inch were obtained.

to my parents

iii

ACKNOWLEDGEMENTS

I wish to thank my advisor, Dr. Ramani Narayan, for creating this project for me, for
his excellent guidance and support, and for giving me the freedom to learn. I also wish
to thank my colleagues, friends and staff at Composite Materials and Structures Center,
Chemical Engineering Department, Department of Packaging, Electron Optics Center and
Michigan Biotechnology Institute for helping me in their own ways and for making my

stay at MSU rewarding and enjoyable.

iv

TABLE OF CONTENTS

Chapter Page
List of Tables ix
List of Figures x
Abbreviations xv
Introduction 1
Terminology 3
Structure of the thesis 4
Background
Polymer Blends 8
Blend Morphology 10
Compatibilization 12
Reactive Extrusion 14
Viscoelastic properties 17
Materials 20
Poly(ethylene terephthalate) 20
Cellulose Acetate 23
Catalysts 26

Chapter Page
Plasticizer 29
4. Blend Preparation 30
Grinding and sieving 32
Drying 33
Extrusion 34
5. Materials Characterization
Injection Molding 40
Compression Molding 41
Tensile and Izod Impact tests 42
Rhemetrics Mechanical Spectrometer 43
Thermal Analysis- DMA, DSC, TGA 45
Electron Microscopy- SEM, TEM 49
Extraction 52
6. Viscosity of concentrated polymer solutions 53
Background 53
The Mixture Rule 60
Reduced Variable Treatment 64
The use of Equation 6.6 68
The Kelly and Bueche Equation 70

Viscosity of Cellulose Acetate 74

vi

Chapter

7. CA/PET Blends
Crystallization - Background
PET Blends and Alloys
Crystallization of PET with CA
CA/EKX blends
Theoritical prediction of modulus

8. Reactive Compatibilization
Transesterification
Characterization of the reaction
Processing of CA with PET-bottle recycle
Compatibilized CA/EKX blends
Discussion

9. Morphology of CA/EKX Blends
Background
Experimental
Results

10. Conclusions and Recommendations
Extrusion
Transesterification

CA/EKX blends

vii

Page

76
76
78
79
84
103
108
108
110
113
115
128
132
132
134

136

143
144

145

Chapter

Page
Plasticized CA/EKX blends 146
Achievements and Prospects 147

Bibliography 149

viii

Table

3.1

3.2

3.3

3.4

4.1

5.1

6.1

6.2

7.1

7.2

8.1

8.2

8.3

9.1

LIST OF TABLES

Data for PET bottle regrind.

Data for EKX-105 grade PET.

Data for CA.

Data for DEP.

Temperatures used for extrusion.

Typical temperatures used for injection molding.
Pure PMMA viscosities extrapolated by mixture rule.

Pure PMMA viscosities extrapolated by the Kelly & Bueche
equation.

Crystallization in CA/PET blends.

Morphology of CA/EKX blends.

Extraction results.

Temperatures used for extrusion of CA/PET.
Temperatures used for injection molding CA/PET.

Calculated viscosity ratios for plasticized CA/EKX blends.

ix

Page

22
23
25
29
39
41

61

73
82
100
112
114
114

140

Figure

2.1
3.1
3.2
3.3
3.4
4.1
4.2

4.3

4.4
5.1
5.2
5.3
5.4
5.5

6.1

LIST OF FIGURES

Droplet deformation and break-up in shear flow field.
Poly(ethylene terephthalate).

Dimer of CA (DS = 2).

Flow designation of CA/plasticizer formulations.
DMAP molecule.

Experimental protocol.

Caliberation of feeder for EKX pellets.

Schematic of the Baker-Perkins Twin Screw Extruder, CMSC,
MSU

Configuration of the co-rotating twin screws
ASTM D638 type I injection molded specimen.
Environment chamber of the RMS-800.

Image formation in the SEM.

Sample preparation for TEM.

Apparatus used for extraction.

PMMA/DEP data for 120°C; ’- -’ mixture rule, ’--’ best fit not
using data at d>=1.

Page

1 1
20
24
25
27
31
32

35

37
43
45
49
5 1

52

62

Figure

6.2

6.3

6.4
6.5
6.6

6.7

6.8

6.9

6.10

7.1
7.2
7.3
7.4
7.5
7.6
7.7
7.8
7.9

7.10

PMMA/DEP data for 140°C; ’- -’ mixture rule, ’--’ best fit not
using data at d>=1.

PMMA/DEP data for 100°C; ’- -’ mixture rule, ’--’ best fit not
using data at 6:1.

Viscosity of CA estimated by the mixture rule.
Pure CA viscosity estimated by reduced variable method.
Prediction of pure PMMA viscosity with Equation 6.6.

Prediction of PMMA viscosity using Kelly and Bueche equation
at 120°C.

Prediction of PMMA viscosity using Kelly and Bueche equation at
140°C.

Kelly and Bueche Equation fit for CA/DEP data at 10 rad/s.

Viscosity curve of pure CA estimated by the Kelly and Bueche
equation.

Crystallization exotherms of 100% PET
Crystallization exotherms of 90% PET/ 10% CA.
Crystallization exotherms of 70% PET/30% CA.
Enthalpy of recrystallization of CA/PET blends.
Strength at break of CA/EKX blends.

Tensile modulus of CA/EKX blends.

Elongation at break of CA/EKX blends.

Izod impact strengths of CA/EKX blends.

Glass transition temperatures of CA/EKX blends.

Melting points of CA/EKX blends.

xi

Page

62

63
63
67

69

72

72

74

75
8O
80
81
81
85
85
86
86
88

89

Figure

7.11

7.16

7.17

7.18

7.19

7.20

7.21

7.22
7.23
7.24
8.1

8.2

Enthalpy of melting of CA/EKX blends.

20% CA/80% EKX blend; Microtruded; no catalyst.
Sectioning parallel to direction of flow.

40% CA/60% EKX blend; Microtruded; no catalyst.
Sectioning parallel to direction of flow.

50% CA/50% EKX blend; Microtruded; no catalyst.
Sectioning parallel to direction of flow.

50% CA/50% EKX blend; twin screw extruded, no catalyst.

Sectioning perpendicular to direction of flow.

60% CA/40% EKX blend; Microtruded; no catalyst.
Sectioning parallel to direction of flow.

70% CA/30% EKX blend; Microtruded, no catalyst.
Sectioning parallel to direction of flow.

70% CA/30% EKX blend; twin screw extruded, no catalyst.

Sectioning parallel to direction of flow.

80% CA/20% EKX blend; twin screw extruded, no catalyst.

Sectioning parallel to direction of flow.

80% CA/20% EKX blend; twin screw extruded, no catalyst.

Sectioning perpendicular to direction of flow.

90% CA/10% EKX blend; twin screw extruded; no catalyst.

Sectioning parallel to direction of flow

Complex viscosities of CA/EKX blends.

Prediction of complex shear modulus for CA/EKX blends.
Prediction of Young’s modulus for CA/EKX blends.

CA backbone grafted with PET.

Strength at break for 80/20 blends.

xii

Page

90

92

92

94

94

95

95

97

97

98

98

101
106
106
109

115

Figure
8.3
8.4
8.5
8.6
8.7
8.8

8.9

8.10

8.11

8.12

8.13

8.14

8.15

8.16

8.17

8.18

Tensile modulus for 80/20 blends.

Elongation at break for 80/20 blends.

Izod impact strength of 80/20 blends.

Tgs of 80/20 CA/EKX blends.

Melting points of 80/20 blends.

Enthalpy of melting of 80/20 CA/EKX blends.

80/20 CA/EKX blend + no catalyst;
Sectioning parallel to direction of flow.

80/20 CA/EKX blend + no catalyst;
Sectioning perpendicular to direction of flow.

80/20 CA/EKX blend + 0.05% FASCAT 4201;
Sectioning parallel to direction of flow.

80/20 CA/EKX blend + 0.05% FASCAT 4201;
Sectioning perpendicular to direction of flow.

80/20 CA/EKX blend + 0.2% FASCAT 4202;
Sectioning parallel to direction of flow.

80/20 CA/EKX blend + 0.2% FASCAT 4202;
Sectioning perpendicular to direction of flow.

80/20 CA/EKX blend + 0.05% DMAP;
Sectioning perpendicular to direction of flow.

80/20 CA/EKX blend + 0.05% Stannous Octoate
Sectioning perpendicular to direction of flow.

80/20 CA/EKX blend + 0.05% Stannous Octoate;
Sectioning parallel to direction of flow.

80/20 CA/EKX blend + 0.9% Titanium butoxide;
Sectioning parallel to direction of flow.

xiii

Page
116
116
117
118
119

120

121

121

122

122

124

124

125

125

126

127

Figure

8.19

8.20
9.1
9.2
9.3
9.4
9.5

9.6

9.7

9.8

40/60 CA/EKX blend + 0.1% DMAP;
Microtruded three times. Transverse section.

Variation of 80/20 CA/EKX blend properties.
Estrnation of T1; for CA/DEP blends.
Viscosities of EKX/DEP blends.

Estimation of viscosity for EKX/DEP blend.

Estimation of Tgs for the EKX/DEP blends.

T85 of the two phases in the EKX/CA+DEP blends.

30/70 CA(H)/EKX blend, 1000X;
Sectioning perpendicular to direction of flow.

50/50 CA(MS)/EKX blend, 1000X;
Sectioning perpendicular to direction of flow.

50/50 CA(H2)/EKX blend;
Sectioning perpendicular to direction of flow.

xiv

Page

128
129
136
137
137
138

139

141

142

142

CA
DEP
DMA
DMAP
DS

DSC

PET
rpm
SEM
St. 0c.
TEM
TGA

TiBuO4

ABBREVIATIONS

Cellulose Acetate.

Diethyl pthalate.

Dynamic mechanical analysis.
4-dimethylaminopyridine.

Degree of substitution.
Differential scanning calorimetry.
"Transpet EKX-105" grade glycol modified PET.
Fourier transform infra-red.
Poly(ethylene terephthalate).
Revolutions per minute.

Scanning electron microscope.
Stannous 2-ethyl hexanoate.
Transmission electron microscope.
Thermo-gravimetric analysis.

Titanium Butoxide.

XV

Chapter 1

Intro d’uction

 

The large increase in the consumption of plastics in the past few decades has resulted in
a dramatic solid waste problem. The plastic waste in the municipal solid waste (MSW)
is projected to be 9.8% by the year 2000 from 7.2% in 1987 which was about 50% of
the total plastics sold in the US (Michigan DNR, 1987) Out of this plastic waste stream,
it is estimated that only about 1% is recycled [Narayan, 1990]. This has brought about
severe shortage of landfill sites and pollution. In 1988, 180 million tons of MSW was
produced out of which 70% was landfilled. Two thirds of the usable landfill sites in the
US have closed in the last 11 years and one third of the remainder will close by 1995
(Recycling Source Book, 1993). Since recycling is still quite expensive, there is an
increasing interest in the use of renewable resources to make plastic materials. The main
reasons for this interest are [Narayan, 1991]:
0 Environmental Concerns.
0 Alternate feedstocks to non-renewable imported petroleum feedstocks.
O The need to utilize the nation’s abundant agricultural feedstocks.
0 Legislation and public opinion.

Natural polymers such as starch, polyhydroxy butyrate valerate and cellulose and

their derivatives such as poly(lactic acid), cellulose acetate, cellulose propionate etc are

2

being increasingly used as a replacement (partially or in full) for synthetic polymers
derived from petroleum. Moreover, these polymers have the advantage that they are
chemically or biologically degradable and thus can be recycled back to earth instead of
being reused in low grade applications.

These natural polymers are relatively more difficult to process and may be more
expensive than synthetic polymers. Therefore, it is of interest to find new methods of
modifying these materials to make them easier to process and improve the properties.

At the same time, new applications are being sought for the reclaimed polymers
for which the demand is outstripped by the supply. Very often, it is not possible to clean
and purify the collected plastic waste products and they have be utilized in lower grade
uses rather than being used in the original application.

Blending of two or more polymers to make a material with desirable properties
is becoming increasingly popular. Alloys and blends represent inexpensive routes to
satisfy both end-user requirements and suppliers desires for product differentiation.
Tailoring of properties, the use of existing equipment and the low cost of developing of
new products have made polymer blends demand outstrip that for engineering base
polymers.

This thesis deals with the study of a renewable polymer, Cellulose Acetate and
it’s blends with a synthetic polymer, poly(ethylene terephthalate) to create a new material

with superior properties that could possibly be commercially viable as an engineering

grade composite.

Terminology

Many of the terms used in the field of polymer blends have subtle differences in

the implied meaning of the terms. The following is the description of some of the terms

used in this text:

Polymer blend: A mixture of two or more polymers. This is different from a
block or graft copolymer in the sense that there are no chemical bonds between
the polymers.

Polymer alloy: A blend with properties superior than the component polymers.
Miscible blend: A blend which forms a single thermodynamically compatible
phase.

Immiscible blend: A polymer blend in which the component polymers exist as
distinct phases.

Compatible polymers: The term "compatibility" is used for polymers when a
desired or beneficial result occurs when they are mixed. It does not imply
miscibility or immiscibility.

Compatibilized blend: An immiscible polymer blend which has been
compatibilized by some method such as surface modification, grafting, addition
of a compatibilizing agent etc.

Polymer composite: A combination of a polymer with a reinforcing agent or a

filler.

4
Structure of the thesis: Problems and Solutions

Cellulose Acetate (CA) is a derivative of the natural polymer cellulose in which
some or all the hydroxyl groups are replaced with acetate groups to reduce hydrogen
bonding and make the polymer easier to process. However even with this chemical
modification, it is necessary to use a plasticizer to aid in processing. The plasticizers
used impart qualities like clarity and impact resistance but result in a decrease in strength
and modulus. Also, the plasticizers that are being currently used are low molecular
weight plasticizers, which have a tendency to leach out with time and leave behind a
brittle polymer. It was sought to avoid the use of a monomeric plasticizer by blending
CA with a synthetic polymer which could act as a processing aid. The polymer selected
to blend with CA was poly(ethylene terephthalate) (PET) which is widely used in
consumer product such as bottles, fibers, films etc. and thus large quantities of post-
consumer wastes are being collected and is available in the market at prices much lower
than that of virgin PET. Other reasons for selecting PET were:

0 Due to it’s high crystallinity, it could probably act as a reinforcing phase for the
amorphous CA matrix.

0 Low melt viscosity as compared to unplasticized CA.

0 The presence of ester groups in the backbone presented an opportunity to
compatibilize the blend easily via a trans-esterification reaction.

0 Hydrogen bonding between CA and PET could help in compatibilizing the blend.

A description of all the materials used for this work is given in Chapter 3.

5
Most mixtures of polymers are not miscible [Paul, 1978 (a)]. Indeed,

incompatibility is it often desirable as provides the opportunity to achieve a synergistic
combination of properties in which the properties of the blend are superior to the
properties of the component homopolymers. To achieve this effect it is necessary to
control the morphology of the blend. It is also necessary to compatibilize the blend so
as to reduce interfacial tension between the phases. Both these issues were investigated
in this thesis.

The most important factors in controlling the morphology of a blend are viscosity
and volume ratios, processing conditions, compatibilization etc. Since CA has always
been processed with a monomeric plasticizer, available literature on the viscosity and
properties of CA do not contain data on unplasticized CA. To study the blend of CA and
PET, it was necessary to know the viscosity of unplasticized CA. This proved to be a
challenging problem as the viscosity of unplasticized CA could not be measured by
conventional methods. Chapter 6 is devoted to the estimation of viscosity of unplasticized
CA.

The next step was to blend CA and PET without any effort at compatibilization.
Chapter 4 describes the method of preparing the blends. Extrusion was the process of
choice because the operations of melting, mixing, compatibilization and processing can
be integrated in one step. Chapter 5 discusses in detail the methods and instruments used
for the characterization of the blends. Chapter 7 contains data, results and discussion for
the uncompatibilized CA/PET blends. An effort has also been made to compare the

values for modulus and strength obtained to various theoretical models in order to

6

evaluate the scope for further improvement through better control of morphology and/or
compatibilization.

Chapter 7 also contains a section on crystallization kinetics of CA/PET blends.
This was found necessary as processing of CA/PET blends was difficult due to the
crystallization of PET in the extruder. Isothermal crystallization kinetics were studied in
order to better understand and solve the problem.

By blending CA with PET, large improvements in strength and modulus were
obtained. These compared well with theoretical predictions for a compatibilized blend.
However, low impact strength and morphological studies suggested that better adhesion
could result in better properties.

The presence of ester groups in both CA and PET offered the possibility of
modifying the polymers via a transesterifiaction reaction. It was sought to catalyze this
reaction during the extrusion process to provide an "in situ" compatibilizer. Reactive
extrusion in a single or twin-screw extruder is an economical method to produce and
process polymer blends. Various catalysts were investigated and the extent of the reaction
was characterized by solvent extraction. The trans-esterification reaction and the results
for the compatibilized blends are presented and discussed in Chapter 8.

To study the effect of viscosity and volume ratios on the morphology of the blend,
a novel method was investigated. Instead of using polymers of different molecular
weights to vary the viscosity of the polymers in the blends as has been traditionally done,
the viscosity of CA and PET was varied by using as plasticizer which was compatible

with both the polymers. A redistribution of the plasticizer between the two phases

7

occurred and the glass transition temperature was used as the measure of plasticizer in
the phase. This viscosity of the phase could thus be calculated. The results of the study
are described and discussed in Chapter 9.

Finally, Chapter 10 contains the conclusions from the thesis and some
recommendation for further work.

The background and literature survey for each topic has been presented as it has
come up in the thesis. However, Chapter 2 contains some background on the subjects
of polymer blends, their thermodynamics, morphology and rheology, reactive extrusion,
compatibilization and viscoelasticity so as to give an overview of the field of polymer

blends and the goals that have been sought to accomplish by this work.

Chapter 2

Background

 

2.1 Polymer Blends

Polymer blends are finding increased acceptance by the industry due to many
advantages over component homopolymers. Alloys and blends represent fast and
inexpensive routes to satisfy both end-user requirements and suppliers’ desires for
product differentiation. Tailoring of properties, the use of existing equipment and the low
cost of developing of new products have made polymer blends demand outstrip that for

engineering base polymers. Some other advantages that may be obtained by blending are

[Kulshreshtha, 1993].

O Extending the performance of expensive resins.

0 Reuse of plastic scrap.

0 Increase in toughness, flame retardance, modulus, crack resistance etc.
0 Improved dimensional stability, processibility and ozone resistance.

Polymer blends may be miscible, partially miscible or immiscible. The miscible
and partially miscible blends are characterized by a single glass transition temperature,
exhibit homogeneity under magnification in an electron or phase contrast microscope and

have physical properties intermediate between those of blend components. Films of

9

miscible blends tend to be transparent. However films of immiscible polymer blends can
also be transparent if the components have the same refractive index, form a two layered
film or the phase size of one of the phases is smaller than the wavelength of light [Paul,
1978 (a)].

Blends of immiscible polymers have a high interfacial tension and poor adhesion
between the two phases. This interfacial tension contributes to higher viscosities,
difficulty in imparting desired degree of dispersion to random mixtures and to their
subsequent lack of stability to phase separation during use. However, sometimes
immiscibility is desirable to achieve a synergistic combination of the two polymers to
form a polymer alloy.

The miscibility or immiscibility of two polymers is dictated by the

thermodyanmics of the mixture. The free energy of mixing AGmix is given by

A0,", = mm — msm, (2.1)

This quantity, consisting of enthalpic (AH) and entropic (AS) parts, is a function of
composition and temperature T. For miscibility, AGmix must be negative. Because of the
small number of moles of each polymer in the blend, the entropy of mixing is very small
and so for most of the polymer blends, the free energy of mixing is positive. Specific
interactions, like hydrogen bonding, dipole-dipole interactions or intramolecular repulsion
effects can also lead to some degree of miscibility. This may explain the partial
miscibility among hologenated polymers and those containing oxygen (e.g. ester groups)

[Paul, 1978 (a)].

Polymer blends are currently formulated by many methods including solvent

10

casting, extrusion in a single or twin screw extruder etc. Reactive extrusion in a single
or twin-screw extruder is an economical method to produce and process polymer blends

as the operations of melting, mixing, compatibilization and processing can be integrated

in one step.

2.2 Blend Morphology

Blends of immiscible polymers can be organized into a variety of morphologies.
The properties and therefore the subsequent use of an immiscible blend are strongly
dependent on this arrangement of the two phases. If one phase is dispersed in another,
the matrix component dominates the properties. A parallel arrangement as in fiber
composites allow both the phases to contribute to the mechanical properties. Both phases
can be continuous to form an " interpenetrating network". These structures are structurally
isotropic and adhesion between the two phases is not as critical as in the other
morphologies. The advantage of having a two phase structure over a homogenous
structure is that the two phases can combine synergistically to give mechanical properties
which are greater than those of the homopolymers. Unique properties may be imparted
to the blend by the control of rheological factors during the processing of the blend.
(Paul, 1978 (a)]. The effect of processing parameters, rheology, volume fractions etc.
on the morphology of the blend and the effect of morphology on the properties of the
blend have been investigated by a number of authors [Kuleznev, 1978; Utracki, 1991;

Favis, 1991; Willis, 1991; Tsebrenko, 1984]. Tanaka [Tanaka, 1992] has investigated

11

the effect of trans-esterification reaction 'on the morphology of binary blends of
polyesters.
Under the appropriate flow fields and Viscoelastic conditions, droplets of the

dispersed phase deform and break up as shown in Figure 2.1 [Han, 1981; Torza, 1972].

 

(a) (b)

J

(C) . (d)

l
I

; i

i i i

 

 

 

Taylor [Taylor, 1934] was the first to study droplet breakup in Newtonian fluids.
Rumscheidt & Mason [Rumscheidt, 1961] studied the deformation of newtonian droplets
over a wide range of viscosity ratios in uniform shear flow and hyperbolic flow. Two
phase flow in nonuniform shear flow (poiseuille flow) has also been investigated
[Goldsmith, 1962; Gauthier, 1971; Chaffey 1965; Hetsroni, 1972]. However, polymers
exhibit Viscoelastic behavior specially at high shear rates. According to Van Oene, the
component with the larger normal stress function will form droplets in the other
component, and difference in viscosity, shear rate (or shear stress), extrusion temperature
and residence time spent in mixing influence only the homogeneity of the dispersion and

not the mode of dispersion [Van Gene. 1972]. Tavgac [Tavgac, 1972] studied the effect

12

of fluid elasticity on the extent of the deformation of the droplets in uniform shear and
extensional flow fields. Lee noted that the deformation of Viscoelastic droplets in
hyperbolic flow appears to be the same as that of Newtonian droplets having a viscosity

equal to the zero-shear viscosity of the Viscoelastic droplet [Lee, 1972].

2.3 Compatibilization

Ideally, two or more polymers may be blended together to form a wide variety
blends with desired morphologies and properties. But most polymers are
thermodynamically immiscible and do not give a homogeneous product upon blending.
However in many cases it is desirable to have a two phase morphology, for example, the
mbber microphase in polystyrene acts as inhibitors to crack propagation and thus increase
the impact strength of polystyrene. In such cases property enhancement is achieved by
the synergy of the two phases to give "alloys" rather than simple blends with properties
in between those of the homopolymers. However, the high interfacial tension and poor
adhesion between the phases would be undesirable. High interfacial tension results in
high viscosities, phase separation and poor dispersion while processing of the blend. Poor
adhesion may lead to very weak and brittle mechanical behavior and may make the
generation of highly structures morphologies (in coextrusion, fibers etc.) impossible.
Thus it is highly desirable to add a "compatibilizer" (interfacial agents) which can

alleviate to some degree the problems mentioned above [Paul, 1978 (a)].

13

Compatibilization is accomplished generally by two routes:

1. Addition of a third ingredient as a co-compatibilizing agent. Block or graft
copolymers are often chosen as compatibilizing agents and may be added as a
third component or be produced "in situ" through reactive blending.

2. By modification of the matrix, or dispersed phase, or both to form a co-reactive
alloy. The co-reactive technique is, for instance, used for toughening of nylon
through reaction of amine end groups of nylon with anhydride grafted modifier.
Some of the other modified polymers which have been used in such co-reactive
alloys are isocynate terminated polycaprolactone, oxazoline functionalized

polystyrene, acrylic acid functionalized polyethylene etc [Curry, 1990].

The addition of a compatibilizer can be expected to (1) reduce the interfacial
tension between the phases, (2) permit finer dispersion during mixing, (3) decrease the
tendency of phase separation and (4) improve interfacial adhesion. These factors usually
result in an increase in mechanical properties such as impact strength, ductility,
toughness, dimensional stability and thus these properties can also be used as an indirect
measure of compatibilization. Direct methods to monitor the extent of interfacial activity

or reaction would be FTIR spectroscopy, NMR, separation of phases by extraction,

electron microcopy etc.

14

2.4 Reactive Extrusion

Reactive extrusion is the use of chemical reactions during polymer extrusion to
form desired products. Free radical initiators, cross-linking agents and other reactive
additives can be injected into the extruder to cause these reactions [Cheung, 1990]

Reactive extrusion has been receiving considerable attention as a commercial process

because of the following advantages [Hyun, 1988]:

1. The extruder takes advantage of the high viscosity to transfer through drag flow.
2. The extruder allows for good control of the reaction temperature.
3. The improved surface/volume ratio leads to higher reaction rates throughout the

entire extruder channel.

4. Reactive extrusion can eliminate several solidification-remelting stages, resulting
in significant economic advantages as well as minimizing polymer degradation
during processing. Even ultra-high molecular weight polymers can be processed
by an extruder.

The use of twin screw extruders for reactively compatibilizing polymer blends is arguably

the most economical means of polymer blend processing as it integrates the operations

of melting, mixing, compatibilization and processing in one step.
The extruder being used for reactive extrusion should have the following

characteristics [Tucker, 1987]:

1. Good temperature control for the barrel, preferably fluid heated barrels for
uniform heat distribution. The barrel temperature will influence specific energy

consumption to some extent, and shear stress at the wall of the extruder bore.

15

Good distributive mixing at low shear rates. Thermal homogeneity in the melt
must be maintained by continual distributive mixing, because the low thermal
conductivity of polymers in general permits large temperature gradients to exist
in the absence of good mixing. Distributive mixing also brings the compatibilizer
in contact with the phase boundary.

Screws designed to give good dispersive mixing. This refers to the reduction in
domain size of the dispersed phase (for immiscible polymers). This type of
mixing is directly related to the shear stress applied to the polymer and material
properties of the blend components.

Convenient and accurate feeding of the raw materials, compatibilizing agents, low
molecular weight additives etc. It may be necessary to keep the polymer(s) dry
before feeding or one or more of them may need to be fed at some distance
downstream from the first feed port.

Venting ports may be necessary to remove byproducts of the reaction or for

maintaining an inert atmosphere inside the barrel.

The reactions that are usually carried out in extruders are [Tucker, 1987]:

1.

Controlled degradation: This is done by exposing the polymer to high
temperatures, shear and usually a free radical generator such as oxygen or an
organic peroxide. Narrow molecular weight distribution and high concentration
of reactive sites for grafting can be achieved by this method. This type of reaction

is generally carried out on polyolefins.

16

In-situ polymerization: This includes both addition and condensation reactions.
Reactions requiring high residence times may not be possible by this method.
Examples of commercially viable in-situ extrusion processes are production of
polyurethane elastomers, the condensation reaction of polyether irnides, and the
anionic polymerization of nylon 6 from caprolactum.

Graft reactions: Graft reactions can be used to make two-phase dispersion
products in which grafting is done to promote better mechanical properties, for
instance, a rubber phase dispersed in a brittle thermoplastic and modified polymer
products with the grafting of monomers or additives onto the polymer backbone.
Polyolefins are often modified with maleic anhydride, acrylic acid, methacrylic
acid etc. to improve mechanical properties and to promote adhesion to non-
polymer surfaces such as glass, ink and fillers. Halogenation of butyl rubber is
done to enhance chemical resistance and improve mechanical properties.
Functionalization: Functionalization of end groups in polymers usually provides
improved chemical or thermal stability, mechanical and optical properties and
may alter the morphology of the material. For example, ethylene-acrylic acid
ionomer is modified with an amine to improve mechanical and optical properties

in thick extruded sheets used as glazing laminates.

17
Shear Rate in Extruder

In the extruder, the materials are subjected to complex shear and elongational
deformations and complex temperature profiles along the various cross-sections of the
extruder barrel.. Therefore, it is difficult to characterize the type and magnitude of the
strain rate in the extruder by a single number [Wu, 1987]. However, an effective strain
rate may be determined and used as a crude but convenient parameter. The shear rate in
a co—rotating twin screw extruder has been analyzed by Burkhardt and coworkers
[Burkhardt, 1978]. The shear rates usually encountered in twin screw extruders are of
the order of a few hundred s". For this work, an arbitrary value of 100 rad/s has been
chosen for all calculations as this was the highest rate at which viscosity could be
measured conveniently. Since a constant screw speed was used throughout this work, any

error in using this shear rate would only cause a systematic error.

2.5 Viscoelastic properties

In Viscoelastic studies of polymeric materials the method of sinusoidal excitation
and response is very useful. In this case the applied force and the resulting deformation
both vary sinusoidally with time, the rate being specified by the frequency in
cycles/ second. For linear Viscoelastic behavior, the strain is out of phase with the stress.

This results from the time necessary for molecular rearrangements and is associated with

relaxing phenomena.

18

The stress a and the strain 5 can be expressed as follows :

o =oosin(tot+6) (2.2)

e =eosin(<ot) (2-3)

where w is the angular frequency and 6 is the phase angle. Then

0 = oosin(wt)cosb +oocos(<ot)sin6 (2-4)
The stresses can be considered to consist of two components, one in phase with the strain
(00 cos 6) and the other 90° out of phase ( 0, sin 5 ). When these are divided by the

strain, we can separate the modulus into an in—phase (real) and out-of-phase (imaginary)

component. These relationships are

0 =EOE’sin(tot) +eOE”cos(tot) (2.5)
o o .

E ’ =—°cos6 and E ” =——°s1n6 (2-6)
e0 60

where E’ is the real part of the modulus and E” is the imaginary part. The complex

representation of the modulus can be expressed as :

E‘=E’+iE” (2'7)
For the other defamation types, for example, the ratio of the peak stress to peak strain

for shear, the relationships are similar

:G tnga+Glfl (2.8)

where G* is the shear complex modulus, G’ is the real part, also called the storage

19

modulus and G" the imaginary part, also called the loss modulus. The storage modulus
is related to the storage of energy as potential energy and its release in the periodic
deformation. The loss modulus is associated with the dissipation of energy as heat when

the material is deformed. The phase angle 5 is given by

tan =— (2.9)

The loss tangent tan 6 is called internal friction or damping, and is the ratio of energy
dissipated per cycle to the maximum potential energy stored during the cycle.

The Cox-Merl rule allows a direct correlation to be made between the steady
shear viscosity and the complex viscosity. The two viscosities are equal when the shear

rate (in reciprocal seconds) equals the frequency (in radians/second).

Chapter 3

Materials

 

3.1 Poly(ethylene terephthalate)

PET or poly(oxyethylene oxy terephthaloyl) is a thermoplastic polyester produced by
melt condensation polymerization from terephthalic acid or dimethyl terephthalate and

ethylene glycol [Figure 3.1].

||
— C—O— CH2- CH2“

 

 

Figure 3.1: Poly(Ethylene Terephthalate).

The ester (-COO-) link influences the properties of the polymer in the following ways

[Brydson, 1989, pp 653-654]:

1 It is, chemically, a point of weakness being susceptible to hydrolysis,
ammonolysis and ester interchange, the first two reactions leading to chain

scission. In some cases the reactivity is influenced by the nature of the adjacent

groupings.

20

21

2 As a polar group it can adversely affect high frequency electrical insulation
properties.
3 The polar ester group may act as a proton acceptor allowing interactions with

other grouping either of an inter- or an intramolecular nature.
4 The ester link appears to enhance chain flexibility of an otherwise polymethylenic
chain. At the same time it generally increases interchain interactions and in terms

of the effects on melting points and rigidity the effects appear largely self-

cancelling.

Engineering grade PET resins have excellent melt flow characteristics and close mold
tolerances. PET is hygroscopic and to prevent excessive hydrolysis while processing, the
moisture content of the resin entering the machine should not be more than 0.02%. To
achieve maximum crystallinity, high surface gloss and minimize warpage, the mold
temperature should be high (210°F-250°F) [Modern Plastics Encyclopedia, 1990]. PET
copolyesters are available with slower rates of crystallization, lower Tm and T8, which
permit a broader range of processing conditions. The use of isopthalic acid as a partial
replacement for terephthalic acid also retards crystallinity and this has been used
commercially with 1,4-cyclohexylene glycol instead of ethylene glycol [Brydson, 1989,
pp 680]. Use of higher homologous component instead of ethylene glycol also reduced
the melting points of the polymer e.g. PBT has a melting point of 224°C. PET has wide

applications in automotive parts, electrical/electronic components, small appliances,

bottles, furniture etc.

22
3.1.1 PET bottle regrind
The used blow molded PET bottles are sorted, ground and cleaned for a variety
of end uses such as fiberfill for pillows and sleeping bags, carpet fiber, injection molded
parts, and fihn and sheet. Eastman Kodak and Wellman are the major suppliers of PET

recyclate in the US. Some data for PET bottle regrind is given in Table 3.1.

Table 3.1: Data for PET bottle regrind.

 

Molecular weight Mw ~ 40,000

Density (unoriented crystalline polymer) ~ 1.45 g/cm3

Glass transition temperature T8 = 78°C

Melting Temperature T"n ~ 249°C

Enthalpy of Melting Hm = 51.43 J/g

Maximum rate of cryStallization - at 125°C

Solubility parameter [Brandrup, 1989] 21.7 - 22.7(MPa)“2

Cost (bulk) recycled flake $0.36-$0.40/1b
pelletized $0.44—$0.50/lb
plant scrap $0.40-$0.45/1b

 

 

 

3.1.2 Transpet "EKX"-105

EKX grade of PET is a copolymer of PET and neo-pentyl glycol (2,2 Dimethyl -1,3
propanediol), was supplied by Hoechst Celanese in pellet form. The inclusion of neo-
pentyl glycol in the polymer backbone reduces the rate and extent of crystallinity
permitting the use of a broader range of processing conditions. The molded parts are

clear and can be used in cosmetic applications such as bottles, vials, toothbrushes and

23

display boards. EKX can be sterilized by gamma radiation and has medical applications

such as tubing, bottles etc. EKX is also currently being thermoforrned into trays. Table

3.2 lists some data for EKX.

Table 3.2: Data for EKX-105 grade PET.

 

Molecular weight Mw ~ 50,000
Density 1.30 gin/cm3

Glass transition temperature T3 = 785°C
Melting temperature Tm = 220°C
Enthalpy of melting Hm = 17.8 J/g
Maximum rate of crystallization - at 159°C

Solubility parameter [Brandrup, 1989] 21.7 - 22.7(MPa)“2
Cost (bulk) $1.05/lb

 

 

 

3.2 Cellulose Acetate

Cellulose Acetate is derived by chemical modifying Cellulose (a polysaccharide
consisting of B-D-glucopyronosyl units). Due to strong hydrogen bonding of hydroxy
groups, cellulose is not thermoplastic. Replacing hydroxyl groups with acetate groups
reduces the hydrogen bonding, increases the interchain separation and makes the polymer
less polar [Brydson, 1989]. The number of acetate groups per anhydroglucose unit is
called the degree of substitution (DS). The CA unit has got a 4Cl chair configuration with
maximum number of groups in the equatorial plane (Figure 3.2). The C(3) carbon is less

reactive, perhaps due to Hydrogen bonding with 0(5) pyranose ring atom.

24

 

Figure 3.2: Dimer of CA (DS = 2).

Steric hindrance by a substituent already attached to the C(2) position may also play a
role towards the subdued reactivity [Bikales, 1971]. CA has the advantage that it is
derived from a renewable resource (cellulose) obtained from wood pulp or cotton linters.

The processibility of pure CA is poor and plasticizers such as dioctyl pthalate or
diethyl pthalate are generally used. Plasticized CA has got excellent clarity, high gloss,
colorability, machinability, toughness and hardness. For molding and extrusion, CA with
19%-37% plasticizer is used (flow grades "H5" to "S"). Figure 3.1 shows the flow
designation of commercially available CA formulations.

CA is widely used in tool handles, playing cards, optical applications-opthalmic
frames and face shields, cosmetic and personal care items, tubing, store fixtures and
displays, sporting goods, writing instruments, sheeting and furniture profiles, extruded

tape, audio tape, and electrical items.

25

4O

 

 

 

 

 

Percent Plasticizer
N
LII

 

 

 

 

 

 

 

 

 

 

 

10
S MS M MH H H2 H3 H4 H5

Flow designation
Figure 3.3: Flow designation of CA/plasticizer formulations.

CA (unplasticized) grade JLF-68 was supplied by Hoechst Celanese in flake form.
Plasticized CA samples (made from JLF-68) were supplied by Rotuba Extruders in pellet

form. Table 3.3 gives some data for CA.

Table 3.3: Data for CA.

 

Molecular weight Mw = 55,100
Mn = 11,800

Density 1.30 g/cm3

Degree of substitution 2.45

Glass transition temperature T8 = 191.5°C

Melting temperature Tm = 230°C

Enthalpy of melting Hm = 12.1 J/g

Maximum rate of crystallization - at

Solubility parameter [Brandrup, 1989] 22.7-26.0 (MPa)“2

Cost (bulk) CA flakes $1.35/lb
plasticized CA $1.3-$1.5/lb

 

 

 

26

3.3 Catalysts

Catalysts were used to catalyze the trans-esterification reaction between the ester
groups of PET and CA. Tin and Titanium compounds are increasingly used as trans-
esterification catalysts. The Tin-Carbon bond is quite stable to heat. Also, the Tin-Carbon
bond does not add to the carbonyl group, nor does it react with alcohols. Triorganotins
and tetraorganotins are toxic, specially the lower triorganotin compounds. Diorganotin
compounds are substantially less toxic than the analogous triorganotins. Mono-organotins
compounds present no special toxological problems. In general, they show decreasing
toxicity with increasing alkyl chain length, but of lower order of toxicity than

diorganotins [Kirk-Othmer, 1983 (pg 52-57)].

3.3.1 Dibutyltin dilaurate

[CH3(CH2)1,,C02]ZSn[(CH2)_-,CH3]2 (CAS # 77-58-7) FW = 631.83

Dibutyltin dilaurate (FASCAT 4202, Atochem) is an off-yellow liquid. Diorganotins are
often used as trans-esterification catalyst at 200°C - 230°C. Dibutyltin dilaurate is used
for polycondensation of dimethyl terephthalate to PET. It has been used as a catalyst for
reaction injection melding of poly(urethane-urea)s [Ryan, 1991]. It is also used as a
stabilizer for PVC to prevent degradation by heat and sunlight. [Kirk-Othmer, 1983, pg

64]. It has a tin content of 18.8 % by weight (elemental analysis).

27

3.3.2 4-Dimethylaminopyridine

[C7H10N2] (CAS # 1122-58-3) FW = 122.17
DMAP (Aldrich Chemical Company) is a widely used and versatile catalyst. It is a
colorless, crystalline solid with a melting point of 114°C. It is a highly efficient catalyst

for a variety of acylation reaction and is many times used on an industrial scale.

\N/

Figure 3.4: DMAP molecule.

3.3.3 Dibutyltin oxide

[Sn[(CH2)3CH3]ZO] (CAS # 818-08-6) FW = 248.92
Dibutyltin oxide (FASCAT 4201, Atochem) is an infusible and insoluble solid. This

catalyst is used at 220°C - 240°C. It has a tin content of 47.7 % by weight (elemental

analysis).

28

3.3.4 Stannous 2-ethylhexanoate

[C16H3004Sn] (CAS # 301-10-0) FW = 405.1
Stannous 2-ethylhexanoate (SIGMA Chemical Company), sometimes referred to as
stannous octanoate is a clear, very light yellow and somewhat viscous liquid. The
primary use of stannous octanoate is as a catalyst with certain amines for the manufacture
of polyester urethane foams. Other industrial applications include its use as a catalyst in
silicones, including room temperature vulcanizing, silicone rubbers and silicone-oil
emulsions, in epoxy formulations and in various urethane coatings and sealants [Kirk-

Othmer, 983, pg 68]. it has a tin content of 29.3 % by weight (elemental analysis).

3.3.5 Titanium(IV) Butoxide

[Ti[O(CH2)3CH3]4] (CAS # 5593-70-4) FW = 340.36

Titanium Butoxide (Aldrich Chemical company) is a colorless combustible liquid.
Titanium Butoxide has been used to catalyze trans-esterification reaction of Polycarbonate
and PET at 275°C [Pilati, 1984] where some discoloration and gas evolution were
observed due to side reactions. Titaniun Butoxide has a Titanium content of 14.7% by

weight (elemental analysis).

29

3.4 Plasticizer

Plasticizers are additives that soften and flexibilize rigid polymer chains. Processing
temperature is lowered by a reduction in glass transition temperature and due to internal
lubrication. To study the morphology of CA/EKX blends as a function of volume fraction
and viscosity, it is found convenient to vary the viscosity of both CA and EKX by adding
a plasticizer which is distributed between the two phases during extrusion. The samples
of CA supplied by Rotuba extruders were plasticized by diethyl pthalate (DEP). DEP was
also found compatible with EKX (only one T2 was observed for EKX/DEP blends).

Table 3.4 shows some data for DEP.

Table 3.4: Data for DEP.

 

Melting Point -3°C

Boiling Point 298°C

Glass transition temperature ~ - 65°C
Density 1.118 gm/cm3
Solubility index [Brandrup, 1989] 20.5 (MPa)“2

Cost (bulk) $0.76/ lb

 

 

 

Chapter 4

Blend Preparation

 

Preparation of blends in the twin screw extruder was the first step in the experimental
protocol. The raw materials - Cellulose Acetate and PET had to be of the right particle
size before they could be fed to the extruder. Where necessary, they had to be mixed
with catalysts, plasticizer etc. before being fed to the extruder. In many experiments, the
plasticizer was fed to the extruder directly with a syringe along with the CA/PET feed.
The extrudate was collected and further processed by compression molding or injection
molding for further characterization. Figure 4.1 shows the main steps involved in the

experimental procedure followed.

30

Drying

v v /

E Twin Screw extrusion N ggggfig]
Extraction

V
DSC @

1
[Compression Molding:

\ h

.

E Injection Molding j\.
l l

Tensile \ m
testing W

Figure 4.1: Experimental Procedure

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

32

4.1 Grinding and Sieving

Before processing and drying of the raw materials and blends, it was found
necessary to grind/ sieve the blend components. The extruder that was used had only one
feeder and having the blend components of about the same particle size helped in proper
feeding and mixing of the blend in the proper ratio. The larger pellets were transported
faster than the fine particles and so the feed to the extruder was not homogenous.
Besides, not all the fine particles were fed because of the design of the feeder. Drying

was also more efficient after grinding. Figure 4.2 shows the feed rates obtained with the

feeder for "EKX" pellets.

A

 

 

\

 

0
111111111

 

5"
UT

 

 

/

\
\

 

Feed rate (Kg/hr)

O
p—t
L

 

 

.9
u:

 

T #1 I I

 

 

 

 

 

 

 

a 11111111
-4

1 I f

7 8 9101112

Screw RPM (% of maximium RPM)
Figure 4.2: Calibration of feeder for EKX pellets.

Cellulose Acetate was supplied by Hoechst Celanese in flake form though there were

about 20% by weight of fine particles which the feeder of the extruder could not feed

33

efficiently. So these were removed with a 1 mm sieve. These fine particles were utilized
to make blends of Cellulose Acetate and plasticizer, for extrusion in the microtruder and
for characterization of CA with DSC etc.

"EKX" was supplied by Hoechst Celanese in pellet form. These had to be ground
before mixing with Cellulose Acetate flakes or drying. A granulator for plastic materials
available in the department of Packaging, MSU was used for this purpose. The
granulator had a 2 mm classification sieve which was cleaned before grinding each batch
of material. This granulator was able to grind even large pieces (about 5 cm diameter and

10 cm long) of the extrudate obtained when blends were made without using a die.

4.2 Drying

The Cellulose Acetate, PET (or EKX) and the blends obtained by extrusion were
always dried before extrusion or injection molding to minimize degradation of the
material. Improper drying of the feed could also result in poor injection molded samples.
The materials were dried in a convective oven at 70°C for at least 8 hours before
processing. Plasticized samples of Cellulose Acetate were always dried at 60°C to
minimize loss of the plasticizer. The materials to be dried were kept in aluminum trays
such that the thickness of the layer was not more than 5 cm. To minimize reabsorption
of water during processing, about 300 grams of material was removed from the oven at
one time and put into the feeder of the extruder or injection molder. Thus whenever

blends needed to be extruded, the blend components were mixed in the proper ratio

before drying.

34
4.3 Extrusion

Extrusion was one of the most important steps of the experiments. This is where
the blend materials were mixed and the "in-situ" compatibilization was done through
trans-esterification reaction.

The extruder used was a Baker-Perkins co-rotating intermeshing twin screw
extruder. Figure 4.3 shows a schematic diagram of the extruder). The diameter of each
of the screw was 3 cm, the length was 42 cm. There were two feed ports on the barrel,
two barrel valves and a venting port. The materials were fed through the first port and
the other feed port and venting port were kept closed. Each screw had two sets of six
mixing paddles and a Camel back discharge screw at the end. The die that was used had
two holes of 3 mm diameter. This die could be easily removed and was not used
whenever there was too much dieswell or the extra residence time if the material in the
die resulted in unacceptable degradation. The temperature was measured at three points
on the barrel, at one point in the die and at four points inside the barrel (melt
temperature), defining the conditions in zone 1, zone 2, zone 3 and the die. The barrel
was cooled by water whose flow could be controlled manually by four valves. There was

no cooling water provided to the die.

pm: .820 .535 320m 53... 233.53 2: co 22523. a... BEE

Aembfioma X moeoN. VA 1 1N oeoNli VAI'I_ oeeN . v

 

>\// 7

g

 

 

. r ,, _: _ :._:\\\\.__
L 31.11! i...

 

z ~
~
\ 2,.
Houseman? 258.5 28 “85582 05 .v
03:08:55. :02 05 .5 tom 85:23 .m
85:80:55. sham .0 atom 30.; .N

8338.805. :02 .m 82.3 65m A

36

The extruder shafts were composed, of slip-on screws, kneading paddles and
orifice plug segments. In all the experiments the configuration of these elements was kept
as depicted in Figure 4.4. The transversely neighboring paddles are always kept at 90
degrees to each other, while axially kneading paddles can take a number of orientations
depending on the amount of forwarding action desired in each mixing zone. Barrel valves
are used in conjunction with orifice plugs to control the amount of cross-sectional area
available for axial flow. The orifice plugs are discs of diameter close to that of the
barrel, and the barrel valve is a triangular-shaped vane positioned directly over the
orifice plugs [Nichols, 1991]. The barrel valves were kept open 12% to provide some

backmixing.

9588 SE $358-8 2: mo meta—Smacou ”v.9 25E

detach—o ESE

A

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. ... 17.
H fl .1 l
1 II .1 IJII
rIJ I
H U
L F 1 1 I
f L l E l
m m m N g

maocom Joan—ego .v 88?:— mg .N
mwgm 00? .m mBOuow 80m A

38

The extruder was purged (cleaned) with Polystyrene before and after each run.
the temperature of the barrel and the die was kept at 200°C (and screw RPM = 200) for
purging before the run and then the temperature was increased to the required
temperature. This was done because the extruder could be cleaned better when the PS
was relatively more viscous and could push out the less viscous polymer already in the
extruder. When PS coming out of the extruder was clear, it was assumed that no other
material (except PS) was in the extruder. This generally took about 300-500 grams of PS.
Now the screw was stopped, the temperature was raised if required and the feeder was
cleaned by a vacuum cleaner. The dried material was fed in batches of about 300 grams
straight out of the dryer to prevent reabsorption of atmospheric water vapor. In the
beginning, the extrudate was a mixture of the material and PS and the first 400 gram was
discarded. Since injection molding required about 1000 gram of material, about 1500 g
of material were fed to the extruder. Collection of the material was stopped when the
load (torque) of the extruder which was kept constant during extrusion (about 85% for
most blends) started decreasing. Although there was still some material left in the
extruder, this was subjected to an extended residence time in the hot extruder and could
thus have been degraded. After the run was complete, the feeder was cleaned and PS
added to purge the extruder again. If another blend of similar ratios was to be extruded
after the previous run, the extruder was not purged with P8. The material was added
after cleaning the feeder and again the first 400 gram of extrudate was discarded.

Materials which had enough melt strength and less die swell were extruded with

the die. The strands could be pulled and continuously pelletized by a pelletizer. Since

39

the pelletizer was not able to pinch or cut soft material, the strands had to be cooled by

using a fan (a water tank was available but was not used to minimize water absorption

by the material). Materials which had to much die swell and/or were highly viscous were

extruded without the die to avoid unnecessarily high residence times. These extrudates

could not be pelletized and so had to be ground in a grinder before further processing.

Some extruded strand was always saved to prepare samples for Electron Microscopy.

The melt temperatures in zone 1 and zone 3 ran lower than the set points whereas

the temperature in zone 2 ran higher than the set point. The discharge temperature ran

quite close (i2°C) to the set point. Typical extrusion temperatures are shown in Table

4.1. The temperature in the first zone is kept lower as the materials are to be conveyed

ahead without melting in this zone. A RPM of 200 was used for all materials.

Table 4.1: Temperatures used for extrusion.

 

 

 

 

 

 

 

 

 

Zone Set temperature (°C) =Melt temperature (°C) I.
feed zone 210 170
1 210 230-235
2 250 230-238
Die 220—240 220-242
g: ====.=.-_—==

 

 

Chapter 5

Materials Characterization

 

After preparing the blends in the extruder, the blends were further processed (injection
molded, or compression molded) if necessary and characterized by thermal analysis
equipment (TGA, DMA, and DSC), mechanical tests (Tensile and Impact), the
viscoelastic properties were measured with a dynamic mechanical spectrometer and the
morphology was studied by electron microscopy (SEM and TEM). The trans-

esterification reaction was characterized by DSC, SEM and Soxhlet extraction.

5 .1 Injection Molding

A New Britain (model 75) injection molder was used for molding the materials
into ASTM D638 type I specimens (Figure 5.1). Before injection molding, the materials
were ground to a 2 mm sieve size and dried (the same conditions as those for extrusion).
Before each run the injection molder was purged with Polystyrene till a few clear
dogbones were obtained. If two runs with only slightly different samples were injection
molded, the injection molder was not purged with PS but rather the first 400 g (about 20
dogbones) after adding the new material were discarded. It was found necessary to spray

a release agent on the mold after 10 dogbones that were molded. The cooling time was

40

41

kept high enough to allow solidification and, some shrinkage of the material but short
enough to keep the sprue soft enough to be easily separated of the injected part when the
old was opened. The injection time was typically 4-7 seconds. The time of injection
should allow the mold to be completely filled but if the pressure is maintained for longer
times, the material cannot shrink and the injected part cannot be pushed out from the
mold. When the injected part was not ejected from the mold, it had to be removed by
hand and sometimes the sprue stuck in the runner had to be forced out with a brass rod.
This took about a minute and so the material in the barrel which by then was subjected
to a long exposure to high temperature, were discarded by running the screw for a few

seconds. Table 5.1 shows some typical temperatures used for injection molding.

Table 5.1: Typical temperatures used for injection molding.

 

 

 

 

Zone Temperature (°C) Jl

1 1 215 1
2 240
nozzle 240

 

 

 

 

5.2 Compression Molding

When less number of samples were required as for DMA and RMS, compression

molding was used to conserve material. However, compression molding has got the

42

disadvantages that the blend is not homogenous and the dispersed phase will not get
aligned as in injection molding. So compression molding was only used to mold pellets
of plasticized CA and extruded blends for DMA and RMS. About 2 inches of the
material extruded without the die used in 4 inch X 6 inch X 2 mm mold gave good void-

free sheets suitable for characterization work.

5.3 Tensile and Izod Impact Tests

The samples for these tests were conditioned for 40 hours at room temperature
and 50% humidity prior to testing.

Tensile tests for strength, elongation and Young’s modulus were conducted on a
UTS machine SEM-20 (ASTM D638M). A laser extensometer was used to measure the
strain. The speed of testing was kept constant at 0.2% for all the tests. At least 5 good
readings were obtained for each sample.

The injection molded specimens (type I) were also used for Izod impact tests.
With a diamond saw, a section with length 63.5 mm was cut from the injection molded
dogbone as shown in Figure 5.1 confirming with ASTM D256 (these cut sections could
also be used for DMA). A T MI notch cutter and impact tester were then used to notch
and test the specimens. At least 10 good readings were obtained for each sample. A thin
specimen (under 6.36 mm thick) generally absorbs more energy due to crushing, bending
and twisting than do wider specimens. Therefore, specimens 6.36 mm or over are

recommended (Note 10, ASTM D256 - 90b).

43

b

Figure 5.1: ASTM D638 type I injection molded specimen (all dimensions in mm).
The cuts and notch for Izod impact test and DMA specimens are also shown.

5.4 Rheometrics Mechanical Spectrometer (RMS)

A Rheometrics Mechanical Spectrometer Model RMS-800 was used to
characterize the rheological properties of the pure polymers, their blends and plasticized
polymers. The main parts of the rheometer are
a. Pneumatics: Pressurized air is used for the air bearings of the motor and of the

transducer. Gas is used for the environmental chamber where the sample is heated. The

44

computer functions as the heart of the system, controlling the test sequence, making
measurements and reducing the resulting data. The test sequence parameters are
programmed via the computer keyboard and from the panel of the instrument.

b. Environment Chamber: This is a computer controlled, forced gas convection oven
in which the sample and the test fixtures are placed. The sample and test fixtures can be
heated or cooled before the start of the test. The environmental chamber uses a platinum
resistance thermometer as the control sensor in the control loop. Figure 5.2 shows a
schematic of the environment chamber with the test fixtures and the sample.

c. Transducer Linear Drive: When the temperature is equilibrated in the environmental
chamber, the test sample is loaded between the test fixtures. This is done by moving the
transducer assembly until the test fixtures are separated sufficiently to allow sample
insertion. The transducer assembly is connected to a motor driven platform that can be
raised or lowered. The output from the transducer is coupled to a linear force scaling and

stepper cutout card. this scaled axial force is input to the computer through an analogue

to digital converter.

Modes of Operation:

The RMS-800 can be operated in various modes like the rate sweep, temperature
sweep, frequency-temperature sweep, strain sweep, time sweep etc. In this work the
strain sweep, temperature sweep and the frequency-temperature sweep modes were used.
All measurements were taken under dynamic strain so that complex viscosities, real and

imaginary modulii etc could be calculated.

45

 

 

 

 

 

 

 

 

1. Oven, left half 2. Oven, right half
3. Locking latch 4. Sample viwing port
5. Handle 6. Thermocouple -to- computer
7. Upper test fixture coupling connector panel
8. Lower test fixture 9. Sample
10. Heat baffle
: i 1
. . l
'::'j
10 1-
—— 7 I 3 ~-
1‘“ j l *‘i‘ r— 2
9 :1 4 __L a.
:3 J --,.~/ 5
6 8 l
—: fit?— 10 —. I -—

 

 

 

Figure 5.2: Environment chamber of the RMS-800.

Sample Preparation: Samples in the form of discs 2 mm thick and 25 mm diameter
were required for the RMS. Pellets were compression molded to give a sheet 2 mm
thick. The temperature was kept such that a sheet free of voids was obtained. Circular
samples were then cut from this sheet using a die. Sometimes it was found necessary to

grind the pellets to obtain void free samples.

5 .5 Differential Scanning Calorimetry (DSC)

DSC is a technique of non-equilibrium calorimetry in which the heat flow into or

a the sample and a reference is measured as a function of time or temperature. The

46

sample and the reference are kept at the same temperature by changing a current passing
through the heaters under the two pans. This is an accurate and easy method to use and
is very useful to measure temperatures and heats of transition, specific heat, thermal
emissivity, rate and degree of crystallinity, purity, rate of decomposition, the rate of
reactions etc. Since the heating is controlled with a computer, it is possible to follow
complex heating algorithms. The experiment can be followed in real time. This is useful
when large temperature changes have to done manually at a particular point in the
heating program (like quenching or suddenly cooling to study rate of crystallinity).

A DuPont 910 DSC module was used with a DuPont 9900 Computer/Thermal

Analyzer for all the experiments. The sample size was always kept 10 - 20 mg.

5.6 Dynamic Mechanical Analysis (DMA)

DMA involves the determination of the dynamic mechanical properties of
polymers and their assemblies. As a result of this analysis, the relationships between the
dynamic properties and structural parameters (crystallinity, molecular orientation,
molecular weight, crosslinking, copolymerization, and plasticization, etc.) and
environmental or external variables (temperature, pressure, time, frequency, type of
deformation, surrounding atmosphere, humidity, etc.) can be explained [Murayama, (pg.
2)].

The commonly used frequency range in DMA of polymeric materials is from 10‘2

Hz to 10° Hz.. However when low frequencies are used , the results of DMA can be

47

compared with those of other thermal analysis techniques. A frequency of 1 Hz has been

used for the all the experiments in this thesis.

Glass Transition (TE) and Melting Point (Tm):

The properties of polymers change with temperature. In particular, the properties
like thermal expansion coefficients, modulii undergo an abrupt change in the region of
glass transition. The temperature of this abrupt change is defined as the glass transition
temperature. In dynamic mechanical studies, the dynamic modulus (storage modulus)
decreases rapidly, and the loss modulus and tan 6 exhibit maxima. The glass temperature
can also be determined by DSC, TMA, NMR studies, change in dielectric constant,
refractive index etc. Since so many different methods exist for the determination of T8,
different values are often obtained. Even the same method will often yield different
results depending on the conditions used. The Tg obtained by DMA is dependent on the
rate of heating, frequency, crystallinity, phase size, etc. Throughout this work the
maxima of loss modulus peak was taken as the T, as this was found to give unambiguous
results as compared to the maxima of tan 6 or the inflexion point of storage modulus.
The tan 6 peak at a frequency of one cycle per second is at a temperature of about 5°C
to 15°C above the glass temperature as measured by dilatometry or DSC. The maximum
in the loss modulus at low frequencies is very close to the Tg [Nielson, 1974]. The rate
of heating and thermal history was kept the same whenever comparisons of the results

was required.

48

Sample preparation: The samples for DMA were either cut from the injection molded
dogbones (as for Impact test specimens) or from compression molded sheets (the same

sheets could be used for both DMA and RMS).

5.7 Thermo-Gravimetric Analysis (TGA)

In TGA the weight of a sample is measured as a function of time or temperature.
This is a relatively simple technique giving useful information about the moisture content,
plasticizer content, degradation temperature etc. The results obtained may depend on the
sample size, atmospheric conditions, type of pan and heating rate. The same heating
profile was used whenever comparison of data was required. The sample was heated
under nitrogen, covered or uncovered. Care was taken not to touch the sample pans or
pan covers by hand.

To find plasticizer content (Chapter 6), in the case when the rate of weight loss
did not fall to zero, the final weight was calculated by a second order deferential
equation. To find the degradation temperature of polymers, a high resolution "HiRes"
mode was used. Here the heating rate is automatically reduced whenever there is a large
change in sample weight. This gave sharper peaks at the degradation temperature.

DuPont TGA modules 951 and HiRes 2950 were used with a DuPont 9900

Computer/ Thermal analyzer for all the experiments. The sample weight was always ~ 20

mg.

49

5.8 Scanning Electron Microscopy (SEM)

Image formation: The SEM is employed to observe the surface morphology of a sample.
The normal SEM image is formed when secondary electrons from the atoms of the
sample are given out as a result of inelastic scattering by the electron beam. These
secondary electrons are then detected by an Everhart—Thornley detector. The production
of secondary electrons is very sensitive to the changes in topography of the sample. The
projecting areas of the samples giving out a large number of these electrons and thus
appear brighter. The electrons cannot escape from low lying areas like crevices and thus
these areas appear dark in the final image. A resolution of 4 to 6 nm is possible with this

technique.

._- OBJECTIVE LENS

 

: :<—F1NAL APERTURE

BEAM OF ELECTRONS —» 1:3:
J DETECTOR

 

‘—SF¥ SAMPLE

7 BE
BEAM-SAMPLE
, INTERACTION
VOLUME

Figure 5.3: Image formation in the SEM.

50

Sample Preparation: The samples for SEM. are typically 2 to 4 mm size.The samples
are fractured after freezing under liquid nitrogen. This " freezes in" the actual
morphology of the sample and minimizes the artifacts introduced by fracturing. Surfaces
obtained by room temperature fracturing were also used for the mica and talc samples.
These were then mounted on Aluminum stubs and coated with gold in a sputter coater.
When a backseattered electron image (BEI) is desired, the sample has to be coated with
Carbon. A coating of about 20 nm thickness and the use of graphite paint was found to
be sufficient to prevent charging of the sample with a 15 kV accelerating voltage. Any

staining, critical point drying or etching of the samples was done before the coating step.

5.10 Transmission Electron Microscopy (TEM)

Image Formation: Transmission Electron Microscopy (TEM) is used to give information
about the composite morphology. The image is formed by the electrons transmitted
through a very thin section of the sample. The amount of secondary electrons produced
depends on both density and thickness of the sample. This difference in scattering of the
electrons between different portions of the sample gives the contrast necessary to form
the TEM image. The areas which are thinner and/or have less atomic number are
"electron transparent", i.e. , they allow more electrons to pass through and thus appear
as bright areas in the image. The "electron dense" areas scatter more electrons as well
as absorb electrons and appear as darker areas in the image. A resolution of up to 0.2

nm is possible with the TEM.

51

Sample Preparation: Sample preparation for'TEM is a difficult step and that is why this
technique is used only where some information is required that is not available by SEM
or when a higher resolution is required. A cuboid is cut from the sample of dimensions
approximately 4x6x10 mm and one of its edges is trimmed to a trapezoid using a razor
blade. (Fig 5.4) Thin sections of 60-150 nm are then be cut by a microtome using either
a glass or a diamond knife. The size and thickness of the sections have to be adjusted to
give the best image in the TEM. The sections are collected on copper grids, coated with

carbon or stained as required (Osmium tetraoxide staining was required for CA/EKX

samples) .

 

Figure 5.4: Sample preparation for TEM.

52
5.10 Extraction

The Extraction apparatus consisted of a Soxhlet extracton tube, an Allihn
condenser, a 500 ml round bottom flask and cellulose extraction thirnbles (Figure 5.5).
With the Soxhlet apparatus, one phase can be extracted from the blend by using a
suitable solvent. To extract CA, Acetone was used since PET is insoluble in Acetone.
The sample is kept in a cellulose thirnble and about 300 ml of Acetone is allowed to boil
in the flask. The condensed acetone vapors drip into the thirnble dissolving CA . When
the Acetone in the Soxhlet tube reaches a certain height, it is automatically siphoned back
into the flask. Thus, a closed system is set up and the blend can be continuously refluxed
with fresh solvent. The blend was extracted for 24 hours. The thirnble was weighed

before and after the extraction to calculate the weight of the residue.

AllihnCondensor

 

 

Soxhlet Extraction Tube

 

Thimble

 

 

 

 

Figure 5.5: Apparatus used for extraction.

Chapter 6

The viscosity of concentrated polymer
solutions

 

The modeling of concentrated polymer solution viscosity is interesting both from
the theoretical and practical points of view. On a fundamental level, attempts are made
to model polymer viscosities with molecular parameters like branching, friction factor,
free volume etc. Many empirical and semi-empirical relations have also been found
useful. In chapter 7, the morphology of CA/PET systems with and without plasticizer is
studied as a function of viscosity and volume ratios. Thus it is necessary to know the
volume and viscosity of the two phases. The usefulness of several strategies and
correlations was investigated by using data available in literature. It was also found
necessary to understand the underlying theory and assumptions before selecting the

methods to use. The equations by Kelly and Bueche were found to give satisfactory

results.

6.1 Background

The rheological and thermodynamic properties of dilute polymer solutions are
fairy well characterized both experimentally and theoretically. The purpose of these

theories and equations was to provide a rational extrapolation of viscosity data to infinite

53

54

dilution. However, the viscosity behavior of, polymer solutions at high concentrations,
especially in the region approaching pure polymer, has not yet been investigated
extensively. The effect of high concentrations on rheological properties has not yet been
fully understood. The problem of the viscosity of concentrated solutions of high
molecular weight polymers is a very complicated one from both the theoretical and
experimental point of view because in fairly concentrated solutions one encounters
viscoelastic behavior. The effect of crystallinity has not been addressed at all by any of
the authors.

Various attempts have been made to reach a better understanding of concentrated
polymer solutions and establish analytically the relation between viscosity and
concentration of the solution. The approaches commonly used are assumptions of an
effective viscosity [Peterlin, 1954], application of Network theory [Lodge, 1958],
reduced variables [Simha & Zakin, 1960, 1962; Ultracki & Sirnha, 1963], derivations
of theoretical equations based on free volume concepts [Fujita & Kushirnoto, 1961; Kelly
& Bueche, 1961], and calculation of concentration dependent intermolecular potentials
from fluctuation theory [Fixman & Peterson, 1964]. The equations from these efforts
have not been able to cover the entire range of concentrations successfully and have
parameters which are often difficult to determine. An empirical equation for the entire
concentration range in terms of Huggins Constant has been developed [Lyons &
Tobolsky, 1970] but it is applicable only for low molecular weight polymers. The model
due to Rouse also predicts the viscosity at lower molecular weights [Graessley W.W.,

1984]. An excellent review on the subject has been published [Berry & Fox, 1968].

55

Another difficulty in the development of theories for the viscosity of concentrated
polymer solutions has been the in obtaining accurate data at high concentration specially
near the glass transition temperature. Published data covering a wide range of
concentrations is available for only a limited number of systems. The data that have been
used by the above authors have been obtained by capillary viscometer, falling cylinder
viscometer and tensile creep apparatus. Viscosities at higher shear rates which would be
of practical importance in processing equipment has not been investigated.

In the development of these theories, the viscosity of the pure polymer has often
been used as a known parameter. Thus these theories and equations should prove useful
in the interpolation of data at higher concentrations. However, in the case of Cellulose
Acetate it is not possible to obtain the viscosity of the unplasticized polymer or even
fairly concentrated solutions at higher shear rates and temperatures as it decomposes
below it’s processing temperature. Thus, the extrapolation of the available data to low
levels of diluent concentration is therefore of interest in the case of Cellulose Acetate and
other similar polymers. An attempt is made to fit the various equations to the available

data [Bueche, 1955] and analyze which of them is suitable for this purpose.

6.2.1 Molecular theory for viscosity

The viscosity of polymers and their solutions is found to depend on various
parameters like molecular weight or chain length, chain dimensions, temperature, chain

branching, polydispersity, diluent, shear rate etc. To correlate the experimentally

56

observed values of viscosity with these variables the zero shear viscosity no is considered
to be a product of two factors, a structural sensitive factor F, dependent primarily on
number of atoms (or groups) Z in the chain backbone and a temperature or density

dependent friction factor per chain atom t. Thus the following empirical relation is

obtained

n = F(Z)C(p) (6.1)

A relation similar in form to the Vogel equation has been used to give the temperature

dependence of r as

1
«(T-T0)

 

Int: = lnCo + (6.2)

The inherent friction factor to is presumed to be constant, independent of molecular
weight and temperature although this may not be entirely true. If the molecular weight
dependence of a and T0 are known, then we can calculate the viscosities at constant

friction factor, f, from the isothermal viscosities according to

1

”W ‘1‘"‘1‘ 73-75

(6.3)

with the values of a and To appropriate to each Z to correct for the dependence of r on
Z. The correction l/a(T-To) is typically independent of Z when Z > ca. 400.

It is found that the isothermal viscosity can be approximated by the relation

*1, = KTZ" (6.4)

where

57

a=3.4, 'Zch;

2.52a23.4, ZsZc

KT depends on the material and Zc represents a critical value of Z, above which a =
3.4. Equation 6.4 is found to be consistent with equation 6.1. [Berry & Fox, 1968]

[Fox & Allen, 1962] noted that a parameter X given by

X = 20.61), M) <1>2/v2 (6-5)
is a more significant parameter than either M or Z. Here v2 is the polymer specific
volume, {52] is the unperturbed radius of gyration and (62 is the volume fraction of the

polymer. When the viscosity at constant friction factor is plotted against X it is noted that

“c °< X, for X sXC (6.6)

“c °< X“, for X 2X6

Moreover Xc is found to be nearly the same for all the polymers. This behavior has been
observed to be followed by many systems. [Bueche, 1953, 1955; Fox and Allen, 1964;
Johnson, 1952; Kelly & Bueche, 1961; Burke & Weiss, 1975] Thus, the effect of diluent

can be accounted for by the inclusion of <62 in the definition of X.

58

6.2.2 The concept of free volume

The concept behind the free volume theory for viscosity is that the rate controlling
step in viscous flow at low shear stresses is the formation of a void into which a chain
segment can jump and hence produce flow. Doolittle [Doolittle, 1951] showed that the
viscosity of liquid paraffirrs can be accurately defined as a simple function of relative free
space except for values in the neighborhood of freezing points of each compound.
Unfortunately there is no unique manner of defining this quantity. Doolittle considered
this free space to arise due to the total thermal expansion of the liquid without change
in phase. He defined a ‘fractional free volume’ as the fractional increase in the volume

resulting from expansion

f = vf Iva = (v — v0) /v0 (6.7)
where vf=volume of free space per gram of liquid at any temperature; vo=volume of 1
gram of liquid extrapolated to absolute zero and v = volume of 1 gram of liquid at any

temperature. The viscosity data could then be correlated by the expression

In T] = A + B [f (6.8)
where A and B are constants. Ferry modifies Doolittle’s equation as follows:
1n n = In A + B(v — vf) /vf (6.9)

It is assumed that above the glass transition temperature, the fractional free volume

increases linearly with temperature

59

f zfg + “[T _ Ts) (6.10)
where f, is the fractional free volume at an arbitrary reference Tg and orf is the thermal

coefficient of expansion of the fraction free volume. ozf has been approximated by the

relations [Peticolas, 1963; Porter, 1966].

af = a, - 0:0 for T 2. Tg (6.11)

and

af= ctg - 0:0 for T<Tg

Here orl ,a8 and etc are the expansion coefficients for the liquid, glass and occupied

volume respectively.

6.2.3 Dependence of fractional free volume on diluent

The fractional free volume usually increases by the addition of the diluent. This
dependence of f on (b is not simple as T1; and a, are known to vary with ¢. However
some investigators [Wood, 1954] have suggested that simple additivity in the free volume

might adequately give the dependence fie). Thus for a two component system

fM : ¢1f1 + ¢2f2 (6.12)
Here subscript M, 1 and 2 refer to the mixture, diluent and polymer respectively. f1 and

f2 follow similar relations as Equation 6.10

60
6.3 The Mixture Rule

Generally, the simplest method of estimating the viscosity of a solution at a

temperature well above Tg is to use a mixture rule such as

logn 2 (bplognp + chlognL (6-13)
The viscosities of pure polymer and liquid are 17,, and 17L, respectively, while <13], and ch
are the corresponding volume fractions.
The data of Fujita & Maekawa [Fujita & Maekawa, 1962] for diethyl phthalate
viscosities for the temperatures 20°C to 110°C was used. The Arrhenius equation

(Equation 6.14) was used to find the viscosity at other temperatures.

11 = Ke (IE/RT) (6. 14)

Here K is a constant characteristic of the material, E is the activation energy for the flow

process, R is the gas constant and T is the temperature in degrees kelvin. Thus

32 = e “(73.71) (6.15)

Figures 6.1-6.3 show the PMMA/DEP data from Bueche at three temperatures. The
dotted lines in each case refer to the prediction by Equation 6.13. However, treating the
viscosity of pure polymer as an unknown and the fitting the equation to the remaining
data we can determine how well the viscosity of pure polymer is predicted by the mixture

rule. Table 6.1 shows the results obtained.

.—
:1.

 

1

61

Table 6.1: Pure PMMA viscosities estimated by mixture rule.

 

 

 

 

 

PURE PMMA PURE PMMA

TEMPERATURE VISCOSITY VISCOSITY
(°C) (actual) (estimated)
140 5.25 X 107 4.31 X 103

 

 

The prediction of pure PMMA viscosity is very poor at lower temperatures but gets
better at high temperatures. Figure 6.4 shows the predicted viscosities of pure CA as a

function of frequency at 240°C.

62

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12
1;;10
5:5 8
V 6
§ 2 // )/V
.5. 0 /
8?
—— -2 _
-4 f
0 0.2 0.4 0.6 0.8 1
Volume fraction PMMA

Figure 6.1: PMMA/DEP data for 120°C; ’— -’ mixture rule, ’--’ best fit
not using data at 6:1.

 

 

 

 

 

 

 

 

 

 

 

 

 

8
”'3? 6 /
92/ 4 //
12> / I
.._. /
.3. 2 /
.‘2 0
> /
%°-2 //
._ 1/

-4 ,

O 0.2 0.4 0.6 0.8 1

Volume fraction PMMA

Figure 6.2: PMMA/DEP data for 140°C; ’- -’ mixture rule, ’--’ best fit
not using data at 6:1.

63

 

 

 

 

 

 

 

 

 

 

 

 

 

20 1
A 15 /,1/
m /
a“: 10 1/ /
3.2: //
g 5 , //
> / /
8° //
—-— /

0 /’
/
-5 1
0 0.2 0.4 0.6 0.8 1
Volume fraction PMMA

Figure 6.3: PMMA/DEP data for 100°C; ’- -’ mixture rule, ’--’ best fit
not using data at d>=1.

1 .00E+7

 

1 .00E+6

 

A

1 .00E+5

 

1 .00E+4

 

 

1 .00E+3

 

 

 

1 .00E+2

VISCOSITY (Pas

 

 

1.00E+1

 

 

1.00E+0 3. . H...” . Hm... . .mm.
1.005-1 1.00E+0 1.00E+1 1.00E+2
FREQUENCY (rad/s)

Figure 6.4: Viscosity of CA estimated by the mixture rule. Lower four curves are
for CA/DEP blends.

 

 

 

64
6.4 Reduced Variable Treatment .[Kraus & Gruver, 1965]

This method allows the superposition of the viscosity-shear rate curve for any
polymer containing a diluent on the viscosity curve of a pure polymer. The Newtonian

viscosity according to the semi-empirical equation by [Fox & Allen, 1964] is

m. = (N, /6X,2'4)1'1$§l IM)pZ]3"‘f, (6-16)
with
X. = 92.05.31 M)

where

77»: = Newtonian Viscosity

N = Avogadro’s Number

i502} = Mean square radius of gyration of the molecule

M = Molecular weight

p = density

fo = segmental friction factor

Z = number of main chain atoms per molecule

Zc = critical value of Z for entanglement

In a system containing diluent it appears generally true that [Bueche, 1956; Fox

& Allen, 1964]

z,’ c = z, p (6.17)

where Z,’ is the value of Zc for the polymer-diluent system and c is the concentration of

65

the polymer. Thus the Newtonian viscosity of a plasticized polymer is

n, = (N,/6)(Z,p>°-4(1s21 IM)(Z.)3"f (648)

Dividing Equation 6.18 by Equation 6.16 we obtain

mum/11,10) = <b""(flf.)(is21 Isis (6.19)
where d) is the volume fraction of the polymer and the subscript "0" refers to pure

polymer. To account for non-Newtonian behavior, the shear rate dependence of viscosity

can be stated as [Bueche, 1959]

11 In” = FMNS') (6-20)
where 77 is the viscosity at shear rate 3. The function F is independent of the molecular

weight. Bueche [Bueche, 1959] suggests that this relation may be extended to solutions

by introduction of the concentration as shown in the following equation

n In, = F011; M) (6.21)

Recent studies [Markowitz & Brown, 1963] have shown that better concentration

superposition might be obtained with <62 instead of d).

n In” = Fm: m?) (612)
The use of this method in its original form requires prior knowledge of the viscosity of
unplasticized polymer. However, here the unplasticized polymer viscosity has been
treated as an unknown and the best possible superposition of data is obtained on the

calculated viscosity of the pure polymer.

66

The steps involved in this process are

1. Assume a Newtonian viscosity for the unplasticized polymer, nN(1).

2. Calculate the shift factor ag=nN(¢)/nN(1).

3. Shift the data to give a curve of n/ad, vs. a¢s/¢2.

4. When all the data is shifted, a curve is fitted to the data using an equation similar to

the Cross Equation

C2 _ C1
Tl/ 61¢ =C1 + C (6023)
1 + C3(Tl-§/¢2) ‘

 

The deviation of the data from this curve is calculate and different values of MO) are
chosen till this value is minimized.

Figure 6.5 shows the shifted data and the curve for the unplasticized Cellulose
Acetate. The absicca shows a,,s/<i>2 and the ordinate is n/ag. Unfortunately, due to the
large value of 1),..(1) obtained, the shift factor is small and the shear rates for which the

curve is obtained are much reduced which are not useful for processing equipment.

*/a
T‘ t

67

 

 

 

 

 

 

 

 

 

 

8

1x10 ._ W
xxxLX-xgt
1x107 ; Wax-<19
3 ”ease.

1x106;
1x105;
1x104 m

1.05-5 1.0E—4 1.0E-3 1,013-2 1.0E-1 1.0E+0

2
«rs/d»

 

Figure 6.5: Pure CA viscosity estimated by reduced variable method.
"0" H2, "+" H, "x" MH and "." MS grade CA/DEP blend.

68
6.5 The use of equation 6.6

As mentioned earlier, the behavior predicted by Equation 6.6 has been observed to be
followed by a number of systems. To use this correlation, the isothermal viscosities, 771 first
have to be converted to viscosities at constant friction factor, 17,. The diluent affects both F(X)
and (“(p). The dependence of F(X) on the volume fraction of polymer is taken care of by the

inclusion of d: in the definition of X. Thus only the effect of as on t has to be investigated. A

relation similar to Equation 6.2 can be used for plasticized polymers

1

lnCM : lnCo + ————M_
«WT—To )

(6.24)
With the help of the W—L-F equation this can be recast as [Williams, Landel and Ferry, 1955]

KM

logC“ = logCo +
1 + (T - TgM)/ A”

 

(6.25)

where

and

1/ KM = 2.303aM(T8M — To“)
Data on 17(T) at constant diluent volume fraction may be analyzed to obtain parameters
or“ and ToM as functions of composition since the temperature dependence of F(X) is small.
Knowing these the parameters KM and AM can be determined as functions of d) if TEM is known

as a function of 6. However, it is observed that KM and A“ are only slightly dependent on d>

69

for ¢>0.5. Treating these as constants we can rewrite Equation 6.25 as [Berry, 1966]

C1
(c2 + T - T8”)

 

Inc“ = Inc, + (6.26)

where Cl and C2 are constants with respect to (1). These constants can be used to compute "r

according to the relation [Berry, 1966]

108711445) = 108n7(¢2) - C (6°27)
with
C ~1—C‘ 1 [ C1 1
C2+T-Tg °2 C2+T—Tg W

The parameters Cl and C2 can be optimized by using the data for concentrated mixtures such that

77; is proportional to (623". Figure 6.6 shows the viscosity of PMMA at 120°C as predicted with

this scheme.

A12

p—c
O

+

log (viscosity + constant

 

ON-fi-GOO

O O C

O
O
l I

-0.35
-0.25
-O.15
-0.05

log PMMA volume fraction

Figure 6.6: Prediction of pure PMMA viscosity with Equation 6.6
" +" actual data from Bueche, "0" data fitted to Equation 6.6, "x" data
fitted to Equation 6.6 without the data at (b = 1.

70

As can be seen from Figure 6.6, the data can be fitted to Equation 6.6 after the viscosities are
corrected for constant friction factor, but this method cannot is not able to predict the large

increase in viscosity at very high polymer concentration. It should be noted that at d), = 0, the

constant = 0, so "r = m.

6.6 The Kelly and Bueche equation

The basis of the Kelly and Bueche theory is the free volume concept i.e. the effect of
diluent on a polymeric system is pictured as a relative increase in free volume contributed by
the diluent, thereby loosening the local liquid structure [Kelly & Bueche, 1961]. Cohen and
Turnbull [Cohen & Turnbull,1959] derived an expression for the dependence of viscosity on a
critical free volume v’ , required for the displacement of a flow unit and arrived at an equation
similar to Equation 6.9. This theory has been combined with the W-L-F equation by Kelly and
Bueche to arrive at a very useful equation for predicting the viscosity of concentrated polymer

solutions. The W-L—F equation takes the free volume of the polymer to be [Williams, Landel

& Ferry, 1955]

f = 0.025 + 4.8 X 10‘4(T — T8) (6.28)
In this relationship 0.025 represents the free volume at glass temperature, and the additional term

expresses the increase in free volume due to the expansion of the polymer as the temperature is

raised from Tg to T.

Kelly and Bueche make the assumption that the free volume contributed by the diluent

71

may be added to that of the polymer. Thus the free volume of the system for <62 volume fraction

of the polymer is

f = 6210.025 + 4.8 X10‘4(T — Tg)] +(1 — 6910.025 + «5(7 - T9] (6.29)
where a, and Tg’ are the thermal expansion coefficient and glass temperature of the solvent

respectively. Using this value off, the following equation is obtained.

n = Kp4exp(¢[o.025 +4.8 x 10'4(T — Tg)] + (1 - ¢)[0025 + a.(T - T1911"1 (6.30)

where K is a combined constant including the molecular weight. The corresponding equation for
M < Mc is of the same form except for a first power concentration dependence.

This relation requires a knowledge of the glass temperature of polymer and solvent and
a, for the solvent. Since in practice, a, and Tg’ may be difficult to determine by direct
measurements, several reasonable estimates might be tried until the best fit of the data is
obtained. The coefficient as is usually of the order 103. The value of glass temperature for Di
Ethyl Phthalate is taken to be -65°C for fitting the data [Kelly & Bueche, 1961]. It should be
noted that at higher polymer concentrations, the factor (l-qb) becomes small, so precise values

of as and Tg’ are not essential. Figures 6.7 and 6.8 show the PMMA/DEP data at 120°C and

140°C by Bueche fitted to Equation 6.30.

72

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

12
10 I
”a 8 /
v f/
a» 4 /
'53 / /
O 2 /-/"
a /
/
:90 -2 +4 /
.. * /
-4 , /
-6
0 0.2 0.4 0.6 0.8 1
Volume fraction PMMA
Figure 6.7: Prediction of PMMA viscosity using Kelly and Bueche equation at
120°C, " + " PMMA/DEP data from Bueche, " . . . . " best fit, "_" best fit without

using data at 6:1.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

8
4.
6 /
A /
g 4
EL
.3‘ 2
§ 0 /
e0 _2 Z
.2 j /
-4 I
-6
0 0.2 0.4 0.6 0.8 1
Volume fraction PMMA
Figure 6.8: Prediction of PMMA viscosity using Kelly and Bueche equation at
140°C, " + " PMMA/DEP data from Bueche, " . . . . " best fit, "_"best fit without

using data at 42:1.

73

The Kelly & Bueche equation is seen to give a very good fit the PMMA/DEP data and

is able to predict the large increase in viscosity at high polymer concentrations. Table 2 shows

the actual and estimated values of viscosities.

Table 6.1: Pure PMMA viscosities extrapolated by the Kelly and Bueche equation.

 

 

 

PURE PMMA PURE PMMA
TEMPERATURE VISCOSITY VISCOSITY ERROR
(°C) (Actual) (estimated) ( %)
120 5.01 X 10” 4.97 X 1011 0.8
140 5.25 X 107 5.248 X 107 0.03

 

 

 

 

 

 

Fujita & Kushirnoto [Fujita & Kushirnoto, 1961] have tried to eliminate the parameter K in

Equation by using a viscosity 11" (corresponding to volume fraction of ) and using viscosity

ratios (shift factors). The final equation obtained is:

T, *2
_ 1 =f(T,<l>1*)+ [f( in )l 1

. (6.31)
lnac 13(1) (6,-6.6

 

 

where
ac= n(T.¢1)(l-¢1*)/n(T,¢1*)(1-¢1)
B(T) is a parameter relating the fractional free volume f to (111
The values of B(T) and f(T, $1) can be evaluated by plotting data in the form of -1/lnac against

1/(qbl - d1," ). However there is considerable uncertainty in determining these from the intercept

74

and slope of the straight line obtained because in general the value of fractional free volume are

small and in determining 8(T) these errors are squared.

6.7 Viscosity of Cellulose Acetate

The viscosity and volume fraction CA data from the four CA/DEP blends was fitted to

Equation 6.30 and the parameters K and as were optimised. Figure 6.9 shows the fit obtained

for the data at 10 rad/s.

 

12.0 1

 

10.0 j /

8.0 j /

 

 

 

 

 

8 . /
e: 6.0, /
a? 403 /
m . ‘ /
§ 1 M
5 2.0, //
. /’
-2.oj //

 

/

-4,0‘.,., .4.-. 24. -.,. ”a.
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
volume fraction of CA
Figure 6.9: Kelly and Bueche Equation fit for CA/DEP data at 10 rad/s.

 

 

 

 

 

 

 

 

 

 

 

 

75

The value of Kand as used were 0.45 and 7.91 X 10‘4 respectively. The same value of as was
used for data at other frequencies, varying the value of K.
Figure 6.10 shows the estimated viscosity of unplasticized CA and the viscosity curves

of the four CA/DEP blends.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1.0E-1 1.ors+o 1.0E+1 1.0E+2
1.0E+12
1.0E+11 . - '
1.0E+10‘

--~ pure CA
1.0E+9 H2
1.0E+4 . H

3 .,_s_s__ __ Mn

0‘9“” “VOTE-ii.“ .
‘ ~m ' Man -_
:E 7‘3“- i: :EMW’ Ys‘ V»\ L MS
1.0E+3, ’ a 3,, “Jr v.3 r».
I ‘ ”canasgbww:
sweats}: . ‘83.“
ijfi,é\}§s
""35:
1.0E+2 - .---... . ....... . .
1.0E-1 1.0E+0 1.0E+1 1.0E+2

Figure 6.10: Viscosity curve of pure CA estimated by the Kelly and Bueche
equation.

Chapter 7

CA/ PET Blends

 

Blends of Cellulose Acetate with two kinds of poly(ethylene terephthalate) were made
without adding any catalyst for compatibilization. Blends were made by extrusion in the
microtruder as well as the twin screw extruder. Rate of crystallinity studies were done on CA
blends with highly crystalline PET as the high rate of crystallization was found to make
extrusion difficult. Blends of CA with an amorphous grade of PET (EKX-105) were successfully
extruded and characterized by thermal analysis and mechanical testing. The morphology of
blends extruded both in the microtruder and in the twin screw extruder was studied by Scanning

Electron Microscopy.

7 .1 Crystallization - Background

A majority of engineering plastics like PET, PAT, PC, Nylon 66, PPS etc are
semicrystalline. The crystallinity dictates the mechanical and thermal properties and the
processing conditions of the polymers and their blends with other polymers and fillers. A review
of the published literature on the crystallization of polymers in blends and alloys clearly indicates
that the crystallization behavior and morphology of the component polymers are significantly

modified by the presence of the other component [Nadkarni & Jog , 1989 (a)].

76

77

In a polymer blend, both the components may be crystalline or one may be crystalline
and the other amorphous. In addition, the blends may be miscible or immiscible. In the case of
immiscible blends, if there is a large difference in melting points of the polymers, the component
with the higher melting point will crystallize in the molten phase of the other component which
will affect the crystallization process through the effect of its viscosity on the mobility of the
crystallizing polymer chains. On the other hand the lower melting polymer will crystallize in the
presence of the solidified particles of the other component which may act as nucleating agents.

In the case of miscible blends, the presence of a single Tg widens or narrows down the
temperature range of crystallization of the constituent polymers. If the Tg shifts nearer to the
melting point of a polymer, both the degree and rate are seen to be reduced. The molecules of
the crystallizing polymer may be hindered by those of the second component and in some rare
cases co-crystallinity is observed. In such co-crystalline blends the components are mixed in the
crystalline as well a the amorphous regions, for instance, a blend of low density polyethylene
with a copolymer of ethylene, polypropylene and 1-4, hexadiene [Starkweather, 1982].

Due to the interplay of these factors, the morphology of a blend is more sensitive to the
processing conditions and this can be utilized effectively to obtain a broader range of property
combinations from a single blend composition.

Crystallinity behavior has been studied with DSC, DMA, dilatometry, density
measurements, Wide Angle X-Ray Spectroscopy, X-Ray diffractometry, Infrared Spectroscopy
and depolarized light-intensity measurements. As far as the effect of blending on the crystal
structure is concerned, the published literature is limited, yet generally the crystal structure of

the constituent polymer does not change appreciably by blending.

78
7 .2 PET Blends and Alloys

PET is semi—crystalline polymer with a melting point of about 250°C. The crystallization

of PET with other polymers has been widely studied due to its commercial importance.

Blends of PET with Crystalline Polymers

Wilfong et al [Wilfong, 1986] blends of PET and polyolefins (LLDPE, HDPE, PP etc).
In these the crystallization of PET takes place in the melt of the other polymer and the rate of
crystallization is reduced. A miscible blend of PET/PAT is reported by Escala and Stein [Escala,
1979] The Tg of PET is lowered resulting in an increase in crystallization rate of PET and a
corresponding decrease in crystallinity for PAT. The rate of crystallization of PET is increased
by the addition of polyphenyl sulphide (PPS) [Shingankuli] indicating enhanced nucleation due

the presence of solidified PPS.

Blends of PET with Amorphous Polymers

Partially or completely miscible blends that have been studied are PET/ PC [Nassar,1979],
PET/polyarylate [Robeson, 1985] and PET/polyester carbonate blends [Ahroni, 1983] All of
them exhibited a slower crystallization rate and a lower degree of crystallinity of PET in the
blend due to an increase in the Tg of PET. An immiscible blend of PET/PMMA has been studied
by Nadkarni and Jog [Nadkarni, 1987]. PET in the blend exhibited a lower degree of
crystallinity, although the rate of crystallization was increased.

The effect of inorganic nucleating agents on crystallization of PET has also been

investigated [Pryzgocki, 1975].

79

7 .3 Crystallization of PET with Cellulose Acetate

The poly(ethylene terephthalate) used for these experiments was ’bottle regrind’. The
isothermal crystallization kinetics of pure PET and two blends of PET with Cellulose Acetate
was studied with the help of DSC. Blends of 90% PET/10% CA and 70% PET/30% CA (by
weight) were prepared in the twin screw extruder. A 20 mg sample of the extrudate was then
subjected to the heating program in a closed DSC pan. The temperature was raised above the
melting point of PET and held there isothermal for 10 minutes to ensure that all crystallites
melted. Then the temperature was brought down manually as fast as possible to the temperature
required and held isothermal. After the crystallization was complete the sample was melted again
and the enthalpy of melting was recorded. This was done at three temperatures (200°C, 210°C

and 220°C) for the three blends (0%, 10% and 30% CA content).

80

 

0.50“

0.25‘

DSC Heat Flow lH/gl

 

 

 

-O.25

o . El I {o I {5 20
Time (min)

Figure 7.1: Crystallization exotherms of 100% PET.

 

 

'0735
25 Overlay 11.00 TA Inst.2200

 

 

 

 

 

 

1.00
202.25%:

0354”.”
a
\ .1
a N” - 211.97°c
I
o
F.
II.
J
I
5 0.254

0.001

.0.” r f r '1 f ' v T v V

0 s 10 15 20 as 30

 

 

T13. (can) Overlay V1.00 TA Ira-c.2200

Figure 7.2: Crystallization exotherms of 90% PET/10% CA.

use Heat Flew (Ila)

1.00

81

 

0.754

 

 

 

 

 

r
5

1':
Tile into)

I
20

Overlay V1.00 TA Inec.2200

:5

Figure 7.3: Crystallization exotherms of 70% PET/30% CA.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

go“; ( 1 1 1
:50 i - - 200C _-
O 2
E405 \i> ' 21°C -
"5 ; ’ \i\
‘27; - A 220C
30 ‘ ~—
9 \ ——
$320“
0 .
a 3 \
7810:
.5: .
E :
m 0 .
-10 0 10 20 30 40 50 60 70 80 90 100

CA % (by weight)
Figure 7 .4: Enthalpy of recrystallization of CA/PET blends.

82

The rate of crystallization as seen is to be .quite fast and therefore crystallization at
temperatures less than 200°C could not be studied. It is seen that the rate of crystallization
decreases with increasing percentage of CA even though 10% CA seems to aid in crystallization
perhaps by providing sites of nucleation. Figure 7.4 shows the enthalpy of melting plotted as a
function of CA percentage. A reduction in the extent of crystallization of PET is seen. Again
for 10%CA at 220°C the reduction is not much. The following table summarizes the results. The
rate of crystallization is characterized by a "characteristic time" (T’) defined as the time taken

for the crystallization to complete after the maxirna in the crystallization exotherm is observed.

Table 7.1: Crystallization in CA/PET blends.

 

 

 

 

 

 

 

 

 

 

CA/PET Blend Temp(°C) T’(s) Hm(J/g) Tm(°C)
200 128 50.98 250.33
0 % CA 210 184 52.68 249.01
220 1218 51.43 248.58
200 92 42.48 251.57
10 % CA 210 111 40.41 247.44
220 151 37.69 250.78
200 90 32.85 252.62
30 % CA 210 156 35.74 246.58
220 1252 31.67 244.70

 

 

 

 

 

 

 

83

The fast rate of crysrallization and high melting point of PET made extrusion difficult due
to the following reasons
1. The Cellulose Acetate degrades during processing at temperatures above 260°C.

2. If the temperature is kept low (~ 260°C), the PET starts crystallizing in the cold spots of
the extruder clogging up the barrel and die.

Because of these reasons an amorphous grade of PET (EKX-105) supplied by Hoechst-Celanese

was investigated. Copolymerization with a second component (neo-pentyl glycol) reduces both

the crystallinity and the melting point to ~220°C. This was found suitable to extrude with

Cellulose Acetate.

To extrude CA with PET bottle recycle, cold spots and high residence times in the die
and barrel of the extruder should be avoided. It was later realized that blends with high CA
content could be extruded as the CA matrix helped in the flow of PET even if it had crystallized.
A 80% CA/ 20% PET blend was extruded without much degradation. the properties have been

included in Chapter 8 along with other compatibilized blends of the same composition.

84
7.4 CA/EKX blends
Blends of Cellulose Acetate and "TRANSPET EKX-105" grade PET were
extruded in the twin screw extruder in composition ranging from 0-95 % CA content by
weight. Blends with higher CA content had to be extruded without a die as explained in
Section 4.3. The temperatures used are given in Table 4.1. Only 95% CA/5% EKX
blend showed some discoloration and blends with higher CA content showed

unacceptable degradation.

7.4.1 Mechanical Properties

Figures 7.5 to 7.8 show the strength at break, tensile modulus, elongation at
break and Izod impact strength respectively of the CA/EKX blends. Due to technical
difficulties, enough material could not be extruded for the mechanical testing of 80%

EKX/20% CA blend though we can expect that it’s properties would follow the general

trends.

85

1 4000

 

1
12000 A

 

 

. 10000 l

8000'

 

Strength at break (p51)

0 r 1

o ‘ 1o 30 4o 50 60 70 80 90 95
CA % (by weight)
Figure 7 .5 : Strength at break of CA/EKX blends.

 

 

 

 

 

 

 

 

Tensile Modulus (106 psi)

 

 

 

 

 

 

 

 

 

l (

0‘10‘30 40 50 60 70 80 90 95
CA % (by weight)

Figure 7 .6: Tensile modulus of CA/EKX blends.

 

86

 

 

 

 

 

 

 

Elongation at break (%)

 

 

 

 

 

 

 

 

 

0 10 30 40 50 60 70 80 90 95
CA % (by weight)

Figure 7.7: Elongation at break of CA/EKX blends.

Izod Impact Strength (FtLb/lnch)

 

0 10 30 40 50 60 70 80 90 95 Dexel

CAIPET blends (% CA)
Figure 7.8: Izod impact strengths of CA/EKX blends.

 

87

It is seen that the modulus increases, consistently with increasing CA content
whereas there is a dip in strength and elongation at break around 50% CA content. This
is due to the morphology of the blend. The phase size of the dispersed phase becomes
quite large and there is no well defined continuous phase at intermediate concentrations.
At higher concentrations the CA phase becomes the matrix resulting in an increase in
properties. At 70% and 80 % CA content, the morphology is fibrillar and it seems that
fiber-pull and fiber elongation are important energy absorbing mechanisms and result in
an increase in impact strength (also see Section 8.7). The morphology of the blends
will.be discussed in more detail later. These results show that the excellent inherent
properties of CA can be recovered by avoiding the use of monomeric plasticizers like
DEP and DOP. This increase in mechanical properties is at the cost of clarity and impact
strength (2 ft lb/ in of notch for CA with 33% DEP).

There is a large decrease in Izod impact strength due to blending. This is most
probably due to the large interfacial stress between the two phases. Compatibilization of
the two phases generally increases the impact strength of a composite due to better stress
transfer at the interface. There is usually an optimum for the dispersed phase size which
gives he highest impact strength. Compatibilization should also reduce the phase size of
the dispersed phase which could also result in an increase in strength and elongation.
Another possible advantage of reduction in dispersed phase size (below the wavelength
of light) could be increased transparency of the blend, specially thin films. The effect of

reaction on the properties and morphology of the blends is discussed in Chapter 8.

88
7.4.2 Thermal Analysis

Thermal tests were performed on the extruded samples to check for any changes
in glass transition temperatures, melting point, crystallinity etc. A reduction in melting
points, or a change in glass transition temperatures can be taken as indicative of
miscibility of the two phases. Co-crystallinity is also desirable in a composite as it results

in a high strength and modulus. Figure 7.9 show the glass transition temperatures of the

two phases as measured by DSC and DMA.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

250 .
l
200 / #_
1
$150
.‘93 —: Tg(PET)-DSC
E
(3 —u— Tg(PET)-DMA
19100
57b“""“" , 507“” Tg(CA)-DMA
l‘ ' 1 ' ' “ —.— Tg(CA)-DSC
so 1
o .

 

 

 

 

 

 

0 10 20 30 40 50 60 70 80 90 100
CA % (by weight)

Figure 7 .9: Glass transition temperatures of CA/EKX blends.

The data does not suggest any signs of miscibility. The DSC consistently shows
Tg of EKX as 5-10°C lower than that measured by the DMA. The Tg of CA does not

show any such regular trend.

89
Figure 7.10 shows the melting points of the CA/EKX blends from DSC

endotherms. Since the melting points of both CA and EKX phases are quite close,
separate melting endotherrns are not observed but rather a composite peak is observed.

Thus it is not possible to tell the extent of crystallization of each phase.

250

 

N
A
O

 

N
(A)
O

 

 

N
N
O

 

N
.3
O

 

Melting Point (Centigrade)

 

 

 

 

 

 

 

200‘
0 10 20 30 40 50 60 70 80 90 100
CA °/o (by weight)
Figure 7.10: Melting points of CA/EKX blends.

 

 

 

 

 

The melting point observed is quite constant. Thus it can be said that blending is
not causing a change in the structure of crystallites of either phase otherwise some
change in melting point would be observed. Figure 7.11 shows the enthalpies of melting
of the CA/EKX blends. It is seen that the crystallization of the CA. phase is suppressed
(otherwise the enthalpy data would have been along the upper dashed line instead of the

bold line). For the same reason, it can be said that the crystallinity of EKX is not

90

affected by the presence of the CA phase. Another explanation can be that the reduction
in crystallinity of EKX is offset by the crystallinity of CA phase so as to give enthalpy

data along the full line (which assumes only EKX crystallinity from 0 to 100%).

 

 

 

 

 

 

 

 

 

18s \n
j ”3\
16 - T\ .\
a 1 ‘H \s ‘
514: NJ \
0112‘ \ \ \-
11310: \ :1
2 : L3 \\
“5 8 , \m
(>3: 6: L\
a . \
g 4: :R‘t
LU :
2: \fi
0 -

 

 

 

 

 

 

 

 

 

 

 

 

0 10 20 30 40 50 60 70 80 90 100

CA °/o (by weight)
Figure 7 .11: Enthalpy of melting of CA/EKX blends.

7.4.3 Morphology of CA/EKX blends

The morphology of the blends was studied by looking at the extruded material
with a Scanning Electron Microscope. The samples were fractured in directions parallel
(longitudinal) and perpendicular (transverse) to the direction of flow after freezing under
liquid Nitrogen. Only the areas representative of the blend were photographed. The
dimensions mentioned are approximate and read directly off the photographs. The

average diameter was calculated by averaging the diameters of 10-40 particles.

91

In general, the twin screw extruder gave more homogenous blends, had better
dispersion and distribution of the minor phase than the Microtruder. In blends with less
CA content, the CA was dispersed in the form of spherical particles. The particle size
increased with increasing CA content and at around 40% CA the particles start to
coalesce and get elongated. In the Microtruded 50/50 blend there is no well defined
matrix. In blends with more than 50% CA, the EKX phase gets elongated into a fibrilla
morphology consisting of long thin fibrils with diameter 0.5-10 microns and lengths more
than 100 microns. The Microtruded blends had thinner fibrils than those extruder in the
twin screw extruder but the was more variation in the size of the fibrils. At 90% CA,
the EKX phase was in the shape of droplets with some particles elongated into fibrils.

Figure 7.12 shows a longitudinal section of Microtruded 10% CA/90% EKX
blend. This morphology is typical of low CA content blends where the CA phase is in
the shape of spherical particles. The dispersion is not very homogenous in this blend and
there is no evidence of interfacial adhesion between the two phases (there is an annular
space between the CA particles and the EKX matrix which arises due to the difference
in extent of contraction of the materials upon cooling).

Figure 7.13 shows a section of a Microtruded 40% CA/60% EKX blend. In this
blend (as well as in Microtruded 30% CA/70% EKX blend), the CA phase is present in
the form of elongated relatively thick, rod like structures of varying cross-sectional
shapes and sizes. The EKX is clearly the matrix and again the blend is non-homogenous.
The corresponding blends extruded in the twin screw extruder are homogenous and the

CA phase is present in the form of spherical particles 1 to 10 microns in diameter.

 

A #4:... '829 l. 1" I ' I If
Figure 7.12: 10% CA/90% EKX blend; Microtruded; no catalyst.
Sectioning parallel to direction of flow.

 

. ' r .
, >3, 7 I
181w 1 808 m? ~-10.au on
Figure 7.13: 40% CA/60% EKX blend; Microtruded; no catalyst.
Sectioning parallel to direction of flow.

93
Figure 7.14 shows a longitudinal section of Microtruded 50% CA/50% EKX

blend. There is no well defined matrix and a mechanically interlocking morphology of
two phases can be seen (the fractured surface of the CA phase has a smoother texture
than the EKX phase). Figure 7.15 shows a transverse section of the corresponding twin
screw extruded blend. The EKX is again the matrix and the CA particles are spherical
with diameters ranging from 2 to 15 microns.

Figure 7 . 16 shows a longitudinal section of a Microtruded 60% CA/40% EKX
blend. The EKX phase is now elongated with phase size ranging from 0.25 to 10 microns
in diameter and upto 100 microns long. These fibrils are aligned in the direction of flow.
The corresponding twin screw extruded blend is more homogenous and the EKX fibrils
are shorter with a diameter of about 10 microns.

Figure 7.17 shows a longitudinal section of a Microtruded 70% CA/30% EKX
blend. the blend is now quite homogenous. The EKX phase can clearly be seen as long
and thin fibrils (only the channels from where the fibrils were pulled out can be seen ,
though one fibril is seen on the right hand side of the micrograph) with diameter of about

0.4 microns.

 

Figure 7.14: 50% CA/50% EKX blend: Microtruded; no catalyst
Sectioning parallel to direction of flow.

(1‘

i‘ ,3

‘ 4 ‘a " x 1

‘ -‘ a: :fli‘. I. W 1!. ‘_‘ii
Figure 7.15:!50% CA/50% EKX blend; twin screw extruded no
catalyst. Sectioning perpendicular to direction of flow.

 

 

Figure 7.16: 60% CA/40% EKX blend: Microtruded: no catalyst.

Sectioning parallel to direction of flow.
Mn

 

'4L‘J’.\~ 1. 1.1""); » ’
Figure 7.17: 70% CA/30% EKX blend: Microtruded, no catalyst.
Sectioning parallel to direction of flow.

96

Figure 7.18 shows a low magnification micrograph of a twin screw extruded 70%
CA/30% EKX blend. The EKX fibrils are seen to be several hundred microns in length
and have an average diameter of about 2 microns. the fibers are excellently aligned
along the direction of flow. This may be due to the uniaxial strain due to the pelletizer.
Instead of the screw pushing the blend out through the die, in this case, the blend was
pulled out of the extruder by the pelletizer and this would have caused the EKX
dispersed phase to be pulled into fibers and highly aligned in the direction of flow.

Figure 7.19 shows the longitudinal section of a twin screw extruded 80%
CA/20% EKX blend in which the EKX fibers have not been pulled out and can be
clearly seen the micrograph. Again, the fibers are quite thin (diameter ~ 1 micron), as can
be seen in the transverse section of the same blend in Figure 7.20.

In the 90% CA/10% EKX blend most of the EKX is dispersed as droplets though
some droplets have been elongated. The blend extruded in the twin screw extruder
(Figure 7.21) is homogenous whereas the microtruded is blend is non-homogenous (there

is a large difference in the sizes of the dispersed droplets and the elongated EKX phase).

o
0"

¢ "A. s
’0» o,
f,
,1 f .
r. W,-

Wu
’1

I

"1

I. r L" ..cé.\“ ‘r.:'~

1. .».... . . -:.ou.e 61-1093 -

Figure 7.18: 70% CA/30% EKX blend: twin screw extruded, no
catalyst. Sectioning parallel to direction of flow.

 

A‘

7; _'_~:.an*‘_‘:3‘

CA/‘20% EKX-blend; twin screw extruded. no
catalyst. Sectioning parallel to direction of flow.

 

 

Figure 7. 20: 80% CA/20% EKX blend; twin screw extruded, no
catalyst. Sectioning perpendicular to direction of flow.

 

Figure 7. 21. 90% CA/10% EKX blend. twin screw extruded. no
catalyst. Sectioning parallel to direction of flow

99

In all these micrographs the CA is seen as a distinct phase than EKX with no sign
of adhesion between the two phases. They also illustrate the phenomenon of ’in situ’
fiber formation in polymer blends. By removing the die and pulling the strand by the
pelletizer, the fibrillation is seen to be enhanced. According to Elemendorf and Van der
Vegt [Elemendorf, 1991] the continuous elongation of droplets does not cause enough
instability so as to cause break up of the fiber. According to Vinogradov [Vinogradov,
1974], the fibrillar structure in an elongation flow may break up if the residence time in
the capillary is longer than the break up time for fiber—like domains. Removing the die,
(which has got a converging flow followed by a short capillary (see Figure 4.2), thus
also prevents the fibers from breaking up.

Table 7.2 summarizes the morphology of CA/EKX blends extruded in the

Microtruder and the twin screw extruder.

100

Table 7.2: Morphology of CA/EKX blends.“

 

CA %
(by weight)

Microtruded

Twin screw extruded

 

10

CA spherical, not homogenous;
particle diameter ~ 1.5 microns.

CA spherical, homogenous;

particle diameter ~1 micron.

 

20

CA spherical, not homogenous;
particle diameter ~1.5 microns.

CA spherical, homogenous;

particle diameter ~ 2 microns

 

30

CA elongated, not homogenous;
d ~7 microns, l > 30 microns.

CA spherical, homogenous;

particle diameter ~ 3 microns.

 

40

CA elongated, homogenous;
(1 ~ 2-10 microns

CA particles spherical, some

fused together; d ~ 6 microns.

 

50

No well defined matrix.

CA dispersed phase, large
particles; d ~ 10 microns.

 

60

EKX elongated,homogenous;
d ~2 microns, l > 50 microns.

EKX elongated, relatively
thicker; (1 ~ 10 microns.

 

7O

EKX fibrillar, long, thin;

(1 ~ 0.4 microns, l > 50 microns.

EKX fibrils highly aligned,
long; (1 ~ 2 microns,

 

80

EKX fibrillar, long, thin;
(1 ~ 0.4 microns.

EKX fibrillar, thin, some
breakup; d ~ 1 micron.

 

9O

 

 

Some EKX particles elongated,
very non-homogenous.

 

CA spherical, little elongation,

homogenous (1 ~ 1 micron.

#=

 

 

101
7.4.4 Viscosity

The zero shear viscosity, no, of liquid mixtures has been described by a number

of simple relations called the additivity rules one of which is [Utracki, 1983]

Inno=2Xi1nni

where X,- and n,- are the weight fraction and viscosity of component 1'. These additivity
rules are followed by ideal mixtures, devoid of any large thermodynamic interactions.
Polymer blends usually do not follow these equations and generally show positive,
negative or both positive and negative deviations from the log additivity rule [Utracki,
1982]. Figure 7.22 shows the complex viscosities of CA/EKX blends as a function of

composition for four different frequencies (in rad/s).

LOE+12
LOE+11
LOE+10
LOE+9
LOE+8
LOE+7
LOE+6
LOE+5
LOE+4
LOE+3
LOE+2
LOE+1
LOE+0

complex viscosity (Pas)

 

o —
+ -o-
LL] L...
c; c:
o _

30E+l
40E+l
50E+l
70F+l
&OE+I
90E+l

20E+l

‘I
uJ
‘3
NO
weight fraction of CA

Figure 7.22: Complex viscosities of CA/EKX blends.

102

The CA/EKX blends are seen to havea negative deviation from the log additivity
rule (shown by the dashed line). The predictions by log additivity rule were calculated
by using the estimated viscosity of CA (see section 6.6) of 2.09 X 1011 Pa.s and 352 Pa.s
for EKX, both at 0.1 rad/s. This negative-deviation type behavior is typical of immiscible
blends and other blends like Polyamide-6/PET, LDPE/PS etc. also are negative -
deviation blends [Utracki, 1983]. One of the proposed explanations for this behavior is
that due to lack of adhesion between the domains of the two polymers, there is slippage.
For such systems with slippage, it should be expected that viscosity would be dependent
on the morphology and type of strain. It is interesting to note that the drop in viscosity
corresponds to the drop in properties such as Izod impact strength, strength at break and

elongation at break.

103

7 .5 Theoretical prediction of modulus

Several theories exist to predict the elastic properties of multiphase polymers as

a function of composition and temperature. For a heterogeneous blends of two polymers,

a broad range of responses to mechanical deformations is possible, depending o blend

morphology, on the degree of molecular mixing or interpenetration, compatibility at the

interfaces and on the size of the phase separated regions- as well as on the properties

of blend constituents [Dickie, 1978].

There are three principle groups of models that have been applied to the problem

of predicting modulus-composition dependence:

1.

Mechanical Coupling Models: These are empirical mechanical models used to
predict modulus. They are not however, morphologically or mechanically realistic
models of blend structure and response. The models of Takayanagi, Kraus and
Rollman and Fujino fall in this category. They employ various arrangements of
the familiar spring and dashpot models to describe the dynamic mechanical
response of a heterogenous system.

Self-Consistent Models: In these models, the mechanical response of an idealized
representative composite structure (comprising of a dispersed phase particle
embedded in a matrix) is compared with that of a homogenous body having the
properties of the composite. the Kerner model [Kerner, 1956], modified Kemer
models and van der Poel’s derivation [C, van der Poel, 1958] fall in this
category.

Bounds on the modulus: In these the strategy is to calculate upper and lower

104

bounds on the composite moduli. The models of Paul [Paul, 1960], Hashin
[Hashim 1962] and Hashin and Shtrikman [Hashim 1963] fall in this category.
Often the bounds are too widely separated to provide an useful prediction of the
response but coincide for a few special cases [Hill, 1963].
In the following figures, the prediction of three models have been compared with the
experimental data. Similar analysis for poly(urethane-urea) blends has been done by Ryan
et al [Ryan, 1991] and for SAN/PC blend by Kodama [Kodama, 1993]. The Kerner

equation

G = G1¢llvl + ¢2G2IG1:
(bl/VI + ¢2GllGl

C

(7.1)

 

requires that the materials be isotropic, that the dispersed phase is spherical and there is
good adhesion at the interface. Here v = 1(1-1’1) and G’ = (7-5121)G1 + (8-10v1)G2,
where Gc is the composite shear modulus, d) is the volume fraction and v is the Poisson’s
ratio. Subscripts 1 and 2 refer to the matrix and filler respectively.

The Davies model [Davies, 1971]

Gel/5 = 4,1011” + (1 — (1)0021” (7.2)
has been derived for two incompressible interpenetrating continuous phases with perfect
bonding at the interface.

For materials without a well defined morphology, Budiansky [Budiansky, 1965]
has modified Kerner’s equation. The model assumes that both components are continuous

but the material is isotropic and macroscopically homogenous. The composite modulus

105

is obtained from a solution to the quadratic equation

MG. - G,)(Gc + «02) + <l>,(G2 — chGc + «01) = o (7.3)

where a = 2(5‘4Vc)/(7-5Vc) and v, is the Poisson’s ratio of the composite and in this case
is taken to be the weighted average of the Poisson’s ratio of the two components.

For all calculations, the Poisson’s ratio have been taken as 0.44 for both EKX and
CA. Pure CA complex shear modulus is taken as 1.305 MPa and Young’s modulus is
taken as 5.665 X 105 psi. The Davies equation requires the materials to be
incompressible, i.e. Poisson’s ratio = 0.5 but the error introduced is negligible. Figure
7.22 is a semi-log plot of complex shear modulus (measured by DMA) versus volume
fraction of CA and Figure 7.23 is a semi-log plot of Young’s modulus (measured by
Tensile tests) versus volume fraction of CA. In these figures the boxes are experimental
data; the upper full curve is the prediction of Kerner equation for a blend with CA as the
matrix; the lower chain curve is the prediction of Budiansky equation; the upper chain
curve is the prediction of the Davies equation and the lower full curve is the prediction

of Kerner equation with EKX as the matrix.

106

10000.0

 

 

1000.0 252:1
/ - H / ,-
z / _

100.0 _ /

 

10.0 ./

 

Complex shear modulus (MPa)

 

 

 

 

 

 

 

1.0
-0.1 0.1 0.3 0.5 0.7 0.9 1.1

CA volume fraction
Figure 7 .23: Prediction of complex shear modulus for CA/EKX blends.

 

1.0

 

 

 

 

 

 

 

 

 

 

Young's modulus (X 10A-6 psi)

 

 

 

 

 

 

 

0.1
-0.1 0.1 0.3 0.5 0.7 0.9 1.1

CA volume fraction
Figure 7 .24: Prediction of Young’s modulus for CA/EKX blends.

 

107

It can be seen that the observed modulus is consistent with the prediction of the
models. Indeed, the experimental data obtained seems to be on the higher side of the
predicted range. This could be interpreted as an indication of good interfacial adhesion
since the models used are derived with this assumption [Allen et a1, 1974]. However
here, modulus values for 100% CA have been estimated. Actual values could be much

higher which would predict higher values for the blends. The decrease in impact strength

could also be indicative of poor interfacial adhesion.

Chapter 8

Reactive Compatibilization

 

The Scanning Electron Micrographs and low impact strengths of uncompatibilized blends
of CA and PET suggested poor interfacial adhesion. Advantage is taken of the ester
groups in the polymer backbone of PET and the ester groups of CA to form a
compatibilizing copolymer via trans-esterification reaction. It was attempted to
characterize the extent of the reaction and the compatibilizing effect with extraction,

mechanical tests, thermal analysis and Scanning Electron Microscopy.

8.1 Trans-esterification

Condensation polymers normally have potentially reactive groups inherent in the
backbone and at the chain ends. For instance polyamides have carboxylic acid and amine
endgroups and amide groups in the backbone. PET has ester groups in the backbone and
hydroxyl and carboxylic acid end groups. The ester groups of Cellulose Acetate are also
potentially reactive. Therefore it was attempted to form a copolymer of CA and PET via
trans-esterification which could act as a compatibilizer for the two phases. Interchange
reactions involving chain end groups, ester groups and hydroxyl groups of CA are also
possible. Trans-esterification would result in a copolymer consisting of a CA backbone

with PET side chains. the breaking of the ester bond in the PET backbone would result

108

109

in shorter chain lengths (Figure 8.1). On the other hand a reaction between the end

groups of PET and the hydroxyl and/or the ester groups on the glucopyranosyl units

would give longer PET side chains.

 

Figure 8.1: CA backbone grafted with PET.

Three types of interchange reactions are theoretically possible:
RICO-OR2 + R3OH e R,CO-OR3 + RZOH
RICO-OR2 + R3COOH ... R3CO-OR2 + R,COOH
RlCO-OR2 + R3CO-OR, ... R1CO-OR, + R3CO-OR2
Since the end-group concentration is very small compared to the ester groups,

trans-esterification would be the more predominant reaction.

110

In recent years considerable interest has arisen in the study of blending of
polyesters and in the exchange reaction which may occur during the melt mixing process.
Montaudo et a1 [Montoudo, 1992] have investigated interchange reaction between
poly(thylene adipate) and PET. DrOscher [Drbscher, 1981] investigated the ester
interchange between PET and an oligoester with di(oxyethylene) units. Tranesterification
between poly(butylene terephthalate) and Polyarylate led to reduced crystallinity and
melting points [Miley & Runt, 1992; Espinoza, 1993] and for PET/Polyarylate systems
increased miscibility led to a slight densification [Robeson, 1985]. The
PET/Polycarbonate system has been investigated by many authors [Pilati, 1985; Dewaux,
1982; Nassar, 1979]. The difference between these systems and the CAIPET system is
that in CA the ester groups are not in the backbone and so the CA chain remains intact.
Unlike the other systems in which the final result of a very high degree of interchange
reaction would be a random copolymer, in the CA/PET system, small PET segments

would be grafted at many places along the CA chain.

8.2 Characterization of the reaction

For miscible blends factors determining Ta and Melting Point change, viz.,
interchain versus intrachain interaction, crystallites formed etc. So thermal analysis
provides an easy method to monitor the progress of the reaction. IR and NMR
spectroscopies are preferred among the few direct and quantitative analysis methods,

despite their lack of sensitivity at low reaction levels [Espinoza, 1993].

111

Cellulosic graft copolymer are somewhat easier to separate from their
homopolymers than other systems due to greater solubility differences. Extraction
procedures and solution followed by fractional precipitation have both been used. A
density gradient ultracentrifugation technique has also been used [Krassig, 1965]. Another
advantage of cellulosic grafts is that the cellulose (or it’s derivative) backbone can be
destroyed by acid hydrolysis and the side chain isolated for studying the molecular weight
distribution and other properties.

It was decided to use extraction with a Soxhlet tube to characterize the reaction
since a small extent in reaction would translate as a much larger change in the amount
that could be extracted. Acetone was used a the solvent to extract the CA phase. The
transesterification catalyst was added to a 80/20 (by weight) feed of CA/EKX blend. The
extrudate obtained was dissolved in phenol (common solvent) and reprecipitated in
Methanol. The precipitate was filtered out and washed with methanol and then with
water. The precipitate was then dried in a vacuum oven. By this method it was hoped to
get a porous and "loose" structure was obtained in all the CA phase was easily accessible
to the solvent. Unfortunately it was difficult to get consistent results by this method and
it was decided to perform extraction of powdered (or rather " shredded" since the blend
was fibrous) samples.

About 10 g of the dried shredded sample thus obtained was added to a (dried)
extraction thirnble and the CA phase was extracted with Acetone in a Soxhlet apparatus
for 24 hours at the end of which the thirnble was again dried and weighed giving the

weight of the residue. The difference between the original weight of the precipitate and

112

the residue gave the weight of the extract. The samples and the thimble were dried in a
vacuum oven at 60°C for at least 5 hours. Table 8.1 shows the results obtained.

Extraction was also carried out on a physical mixture of 80% CA and 20% EKX to

compare with other blends.

Table 8.1: Extraction results.

 

 

 

 

 

 

 

 

S. no. Catalyst weight % of total —
(by weight) extracted.
1 Physical mixture 27.4
2 No catalyst 13.1
3 0.05% FASCAT 4201 11.4
4 0.2% FASCAT 4202 9.0
5 0.05% DMAP 9.1
6 0.05% Sn. Oc. 17.3
7 0.9% Ti.BuO4 7.5

 

 

 

 

 

These results indicate that less material can be extracted with more
compatibilization. It is assumed that in all the blends, the extractable phase is equally
accessible to Acetone, though this may not be entirely true in the case of the physical

blend (which was not extruded like the rest of the blends). It is also interesting to see that

113

even without adding catalyst, there is some evidence of compatibilization. This may be
due to the esterification catalyst(s) already present in the EKX which are known to

promote transesterification or simply due to hydrogen bonding.

8.3 Processing of CA with PET-bottle recycle

With the isothermal crystallization studies on CAIPET blends, it was realized that,
crystallization of PET, it was necessary that long residence times and cool spots in the
extruder and die should be avoided. However, to minimize degradation of CA, the
temperatures have be kept as low as possible. Low die temperatures also help in the
processing of the extrudate as PET has a very low melt strength. Another phenomenon
that was making processing difficult was the strain induced crystallization of PET. PET
molecules are known to orient during elongational flows and crystallize even at
temperatures very near it’s melting point [Van der vegt, 1967].

In view of the above limitations on the processing conditions, it was decided to
extrude a 80% CA/20% PET blend without the die to prevent elongational flow. The
high percentage of CA would also help in the transportation of any PET which had
crystallized.

Table 8.2 shows the temperatures used for extrusion of CAIPET blend. The RPM
was 200 as before and the feed rate was 10%. The conditions for injection molding are

shown in Table 9.3

114

Table 8.2: Temperatures used for extrusion of CAIPET.

 

 

 

 

 

 

 

 

Zone Set Temperature (°C) Melt temperature (°C) ll
feed zone 210 154
1 280 283
2 275 274
Die 225 239

 

Very slight discoloration was observed. The strand could be directly pulled out

from the screw end by the pelletizer roll and pelletized.

Table 8.3: Temperatures used for injection molding CAIPET.

 

 

 

 

 

 

Zone Temperature (°C)
1 250 l
2 268

nozzle 235

 

The injection time was 1.8 s, the holding time 5 s and the cooling time was 18 s for a

water cooled die. The mechanical properties of 80% CA/ 20% PET blend are included

in the next section for comparison with other blends of similar composition.

 

 

115
8.4 Mechanical Properties
Figures 8.2-8.5 show the strength at break, tensile modulus, elongation at break
and Izod impact strength respectively for CA/EKX blends with and without catalysts and

also for a CA/PET blend all containing 80% by weight of CA.

 

 

 

6300

Strength at break (p51)

 

 

 

 

 

 

10 catalyst
TiBuO4

SC Oct
PET'BR

FASCAT 4201

Figure 8.2: Strength at break for 80/20 blends.

116

 

 

 

 

MEN—ME

$520

60%

53:.

No? F<Um<u

SQ» H<Um<u

@258 c:

Figure 8.3: Tensile modulus for 80/20 blends.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

a _

1
E xmonn um :onmcoE

mmuhmm

.uUO.um

v03mHH

Nomv h<um<m

Homv h<um<m

un>H141u 0:

Figure 8.4: Elongation at break for 80/20 blends.

117

 

 

 

 

 

 

 

 

 

 

lziod impact strength (ltlb/inch)

 

 

 

 

 

 

 

 

TIBuO4
St.Oct.
DMAP

a:
‘9
[—
LL]
0..

no catalyst
SCAT 4201
ASCAT 4202

FA
F

Figure 8.5: Izod impact strength of 80/20 blends.

8.5 Thermal Analysis

Figure 8.6 shows the glass transition temperatures of the 80/20 blends as
measured by DSC and DMA. It is seen that there is not much change in Tg of CA phase,
but there is a decrease in the Tg of EKX phase. This result would be expected as a result
of tranesterification as the chains of EKX are being broken and grafted onto the CA
backbone. If reduction in the Tg of EKX phase is taken as a measure of the extent of
reaction, it is seen that the FASCAT catalysts and DMAP are more effective than the
others. This also agrees with the extraction results where a decrease in weight percentage

extracted was observed with the FASCAT catalysts and DMAP. Addition of Titanium

118

butoxide (TiBuO4), however, causes little change in the Tg of EKX phase, while resulting

in a large decrease in weight percentage extracted.

200
A 180

 

—‘-- Tg(EKX)(DSC)

O\
O

140
120
100

-—+-— Tg(CA)(DSC)

—x— Tg(EKX)DMA

OO
O

—o— Tg(CA)DMA

O\
O

 

 

 

Temperature (Centigrade

Nb
CO

 

O

Q)
1..
3
D.

l iBuO4
St.Oct.
DMAP

no catalyst

FASCAT 4201 '
FASCAT 4202

Figure 8.6: T85 of 80/20 CA/EKX blends.

Figure 8.7 shows the melting points for the 80/20 compatibilized and
uncompatibilized blends. There is hardly any change in the melting point, and obviously,
this parameter cannot be used as an indication of the extent of the reaction. It seems that
the grafted CA chains are not hindering or modifying the crystalline structure of EKX

(it was concluded in Section 7.3 that CA phase was not crystallizing in the presence of

EKX; see Figure 7.11).

119

I“)
[\J
{A

 

 

 

I
I

7’30 i/

 

IN)
fl
()1

 

 

 

210

 

temperature (ccnti grade)

 

 

 

 

 

 

 

205
°00 1
e; -— N <1- ..z a.
E 3‘ 9' 5 9 z
:3 — E— - m a
o < < I“
: U U
U) U)
< <
L!- LL.

Figure 8.7: Melting points of 80/20 blends.

Figure 8.8 shows the enthalpy of melting of the 80/20 blends. It is seen that with
except for blends containing TiBuO4 or DMAP, the extent of crystallization actually
increase as compared to that of pure EKX. This may be due to morphological reasons
and due to increased orientation of the EKX molecules during extrusion. This effect of

morphology will be discussed later.

120

 

O\

Enthalpy of melting (J/g)
w A (It

 

 

 

 

 

 

N

 

fl

 

. _ ._——A«,_—.____74»>.———

 

 

 

 

 

 

 

 

0 l 1
1’ o o O o <
A
'5? ‘3" 9' 55‘ 3 2
8 E— l— E: CD 3
0 <2 <1
1: U U
m m
< <
LL. LL.

Figure 8.8: Enthalpy of melting of 80/20 CA/EKX blends.

8.6 Morphology

Figures 8.9 and Figure 8.10 show the twin screw extruded 80/20 CA/EKX blends
without any catalyst to compare with the blends with catalysts. These are the sameas
Figure 7.19 and Figure 7.20 respectively. The EKX phase is elongated into fibers of
diameter ~ 1 micron and quite homogeneously distributed.

Figures 8.11 and 8.12 show the morphology of 80/20 CA/EKX blend with 0.05 %
FASCAT 4201 catalyst. The dispersion and phase size of the EKX phase is the same as

that without any catalyst. The morphology is still fibrillar.

 

331.)
Figure 8. 9. 80/20 CA/EKX blend + no catalyst
Sectioning parallel to direction of flow.

 

a. k“;
Figure 8. 101,80/20 CA/EKX blend + no catalyst
Sectioning perpendicular to direction of flow.

 

   

       

£35.. A ‘1-"in \.'-_ ~ .‘ ‘Mk In.
Figure 8.11: 80/20 CA EKX blend + 0.05% FASCAT 4201;
Sectioning parallel to direction of flow.

_ A...".:' .vv. ' 1' ‘ '1 ‘ 1
Figure 8.12: 80/20 CA/EKX blend + 0.05% FASCAT 4201:
Sectioning perpendicular to direction of flow.

123
Figure 8.13 and Figure 8.14 show the SEM micrographs of 80/20

CA/EKX blends in which 0.2% by weight of FASCAT 4202 catalyst was added. It is
seen in the longitudinal section that the particles are elongated but there are many small
dispersed dr0plets visible. The transverse section shows that the EKX phase is now very
well dispersed with a phase size of less than a micron (Compare Figure 8.14 with Figure
8.12 which are at the same magnification.

Figure 8.15 shows the transverse section of 80/20 CA/EKX blend with
0.05% of DMAP as catalyst. The dispersion is better as compared to the blend with no
catalyst or that with FASCAT 4201. There was difficulty in pelletizing this blend and
also there was color of the extrudate was dark brown. This is possibly due to the
degradation of the catalyst itself and not due to the degradation of CA, as it is observed
that the mechanical properties are not effected to a large extent. Also, the discoloration

was reduced when the blend was dissolved in phenol to reprecipitate for extraction.

                
       

r - h ’- ~ '
. AJ 4213 «A r defiant“ A - :1
Figure 8.13: 80/20 CA/EK blend + 0.2% FASCAT 4202;
Sectioning parallel to direction of flow.

       

    

H ..'.".‘ Ila-‘3 ‘. J '- 2. ".‘o'r-t‘i‘ 3" 1' 5
Figure 8.14: 80/20 CA/EKX blend + 0.2% FASCAT 4202;
Sectioning perpendicular to direction of flow.

           

. _ ‘0 V . - ’7 u t

b y ‘ . I . ‘ P.‘ '- ’ ‘ a ' _ m h

Figure 8.15: 80/20 CA/EKX blend + 0.05% DMAP
Sectioning perpendicular to direction of flow.

 
     

4?
. *:

  

" . ' .2 I '_ ‘ . . . r , .

3. > t ., 5%. .J " ‘ Q.
.; " .r'.-;' ;'- ' ' ~. ..

Figure 8.16: 80/20 CA/EKX blend + 0.05%

Sectioning perpendicular to direction of flow.

   

1' V '
a _ l I l | 0 i
Stannous Octoate

 

126
Figure 8.16 shows the transverse section of 80/20 CA/EKX blend with 0.05% of

Stannous Octoate. The dispersion is the same as that for the blend with no catalyst. the
particles have been elongated into fibers as can be seen in the longitudinal section (Figure
8.17). The thin white strands are an artifact due to improper fracture causing the pulling
out of EKX fibers from the CA matrix.

Figure 8.18 shows the longitudinal section of 80/20 blend with 0.9% by weight
of Titanium butoxide. The morphology has been completely changed. There are a large
number of very fine particles of less than a micron size but also some elongated large

particles with diameters of ~ 10 microns.

 

‘ as
Figure 8.17: 80/20 CA/EKX blend + O. 05% Stannous Octoate?
Sectioning parallel to direction of flow.

 

1:113; -' '
Figureh8. 18: 80/20 CA/EKX blend + 0 9% Titanium butoxide;
Sectioning parallel to direction of flow.

Figure 8.19 shows the transverse section of a 60% EKX/40% CA blend with
0.1% DMAP extruded three times in the microtruder. The catalyst was added in the first
run. This micrograph clearly shows good adhesion between the two phases. The
dispersed phase (CA) looks like it is coated with EKX and the fracture has not taken

place along the interface.

128

 

211

Figure 8.19: 40/60 CA/EKX blend + 0.1% DMAP;
Microtruded three times. Transverse section.

8.7 Discussion

The extraction results coupled with thermal analysis and change in morphology
and mechanical properties strongly suggests that a compatibilization reaction is occuring
between the two phases. The FASCAT catalysts. DMAP and Titanium butoxide are seen
to be effective as compatibilizing agents. It is interesting to study the relationship
between the extent of compatibilization. morphology and properties of the blends (Figure

8.19).

129

 

 

--a-- strength

—+— elongation

 

/ \ modulus

fif/Nq + izod

/\ —*— % extracted

 

1‘

 

 

 

 

V/ 717‘
:3” 1
h.

.[cl

 

 

 

 

 

 

 

 

 

 

 

_4p__

T1BuO4
St.Oct.
DMAP

no catalyst

FASCA'1‘4201
FASCAT 4202

Figure 8.19: Variation of 80/20 CA/EKX blend properties.

Except for Stannous octaoate, all the rest of the catalysts effect the morphology.
The phase size is seen to be reduced. This is very characterisitic of a compatibilizing
reaction and similar results have been reported earlier [Tanaka, 1992; Willis, 1991;
Gleinser, 1994, Xanthos, 1990]. The morphology is dictated by two driving forces -
1) unfavorable thermodynamics which drives the morphology towards phase separation
and 2) the compatibilizing reaction which tends to reduce the interfacial tension and thus
the tendency to phase separate. When the phase separation proceeds much faster than the
chemical reaction, a domain structure is first formed. then the chemical reaction proceeds
mainly on the domain surface and helps in compatibilizing the blend. On the other hand,

when the chemical reaction is much faster than the phase separation, the reaction

130

proceeds homogeneously throughout the melt and a large concentration fluctuation is
never established. As a distinct and much finely dispersed domain structure is seen, it
is the former process which is occuring in CAIPET blends.

It is seen that reduction in the diameter of fibrils has an adverse effect on the
pr0perties of the blend. This suggests that fibre pull out and possible yielding of the EKX
phase before failure are important energy absorbing mechanisms. When there is good
adhesion between the two phases and the fiber diameter is > 1 micron, only shear
yielding can occur. The fiber is not pulled out and the crack proceeds to go through the
fiber. This results in properties which are a little less than that which would be expected
if both the energy absorbing mechanisms were occuring. This is the mode of fracture in
the blend compatibilized with Stannous octoate. When there is good adhesion and the
fiber is thinner and possibly crystallized due to orientation, both of the above two cannot
take place and there is a large decrease in properties. This seems to be the mode of
fracture of the blends compatibilized by FASCAT 4201, FASCAT 4202 and DMAP. The
morphology of the blend compatibilized with Titanium butoxide is very different. Due
to good adhesion (as indicated by extraction) and lack of long fibers, there is a large
decrease in impact strength. But there are some thick fibers of about 10 micron diameter
present. These are not contributing towards the energy absorption by fiber pull out but
are absorbing energy during tensile elongation due to yielding. This results in an increase

in elongation before failure and compensates for the energy not absorbed due to fiber pull

OUI .

131

It can be concluded that fibers and particles less than 1 micron are not effective
as crack inhibitors and cannot effectively absorb energy by yielding. Long fibers are
effective only when they can be pulled out of the CA matrix. This requires sharp
interface and poor adhesion. This is a very surprising result since generally much
improvement in mechanical properties is observed with better adhesion. This also
suggests that CA fails by crazing and particles smaller than a critical size ( ~ 1 micron) are
not effective in either terminating or initiating crazes. For PVC the critical size is less
than 0.1 micron and for PS, the critical size is about 1 micron. This suggests that the

critical particle size for every individual polymer is different.

Chapter 9

Morphology of CA/ PET Blends

 

The morphology of CA/PET blends were studied as a function of volume ratio and
viscosity ratio. A novel method was employed to vary the viscosity ratio. Standard curve
of viscosity anf glass transition temperature were drawn as a function of plasticer content
for the pure polymers. The equation of Kelly & Bueche was used for this purpose (see
Section 6.6). The glass transition temperatures of the two phases in the blends were then

used to calculate the viscosity and thus the viscosity ratio of the phases.

9.1 Background

Physical properties of immiscible polymer blends have been shown to be closely
related to the size and shape of the deformable minor phase. Most investigations
generally discuss phase morphology variations in terms of shear stresses, composition
ratios, shear rate and effect of compatibilization. Very few papers are found in which the
effect of viscosity ratio has been studied exclusively.

Starita [Starita,1972] studied the blend of five polystyrenes (of different molecular
weights) and five polyethylenes (of different molecular weights) over a range of
compostion ratios in an extruder. The type of microstructure in the blended systems were

determined by solvent leaching, electron microscopy and DTA studies. He concluded that

132

133

when the ratio of viscosity of minor dispersed phase to the viscosity of matrix (p) is >
1, the minor component is coarsely dispersed. When this ratio is < 1 or ~ 1, the minor
phase is finely dispersed. However in this paper, the effect of compatibilization on the
phase mo1phology has not been studied.

Favis & Therrien [Favis, 1991] have done a similar study using Polycarbonate
and Polypropylene of various molecular weights. It was found that proximity to the die
wall also effected the morphology in the strands. The influence of viscosity ratio becomes
more pronounced at higher composition. It was also found that over a wide range, the
effect of screw speed and volumetric flow rate on the phase size was not significant
(these parameters have more effect on highly elastc dispersed phases such as elastomers).

Min & White [Min, 1984] have studied PE (three molecular weights) and PS at
two different temperatures and one of the PE with Nylon-6 and polycarbonate. The
polarity of the phases was found to effect the morphology of the blends. It is concluded
that when 0.7 < p < 1.7, the droplets are elongated by the shear stress and form
fibrils, increases in shear stress decreases the domain size of the dispersed phase and
when p > 2.2, undeformed droplets are observed.

Willis et a1 [Willis, 1991] have studied the effect of compatibilization on
PE/Polyamide and PE/PA blends. They blended three grades of each polymer but have
not reported the effect of change in viscosity of the components. Their main conclusion
is that the effect of interfacial modification on morphology predominates over that of
composition and viscosity ratio. A narrowing of the region of dual phase continuity was

observed due to the addition of the compatibilizer.

134

Tsebrenko et a1 [Tsebrenko, 1980] have investigated the formation of fibres in
polyoxymethylene and a copolymer of ethylene and vinyl acetate. Here p varied from
0.35 to 27.7. Fibres are formed at viscosity ratios close to unity. Fibre formation is also
governed by extrusion shear stress.

No paper can be found in which both the viscosity ratio and composition of the
blend is varied and subsequently the effect of compatibilizer on the phase morphology
is studied. The easiest way to vary the viscosity of the component(s) in the blend is to
vary the molecular weight of the component. The effect of varying shear stress (RPM)
or temperature in the extruder is difficult to characterize and moreover do not seem to

influence the phase morphology much [Favis, 1991].

9.2 Experimental

To study the morphology of CA/EKX blends as a function of volume ratio and
viscosity ratio, samples of Cellulose Acetate or EKX grade PET of different molecular
weight could not be obtained. It was decided to follow a novel appraoch to vary the
viscosity ratio. Diethyl pthalate was found to be miscible with both CA and EKX
Moreover, diehyl pthalate is commercially the most popular plasticizer for CA. It was
envisioned that when plasticized samples of CA were blended with EKX, the plasticizer
would get redistributed in the two phases changing the viscosity of the two phases. The
plasticizer would also change the glass transition temperature Tg of the phase . By

measuring the T55 of both the phases, it would be possible to calculate the volume

135

fraction of plasticizer present in the phase this would make calclation of the viscosity of
the phase possible. This approach makes the assumption that both the phases are
completely immiscible and there is no reaction between the two components.

The equation of Kelly and Bueche (section 6.6) was found suitable to model the
variation in viscosity as a function of volume fraction of polymer. The variation of T3
with volume fraction polymer also follows from the same equation [Kelly, 1961]. Since,

at Tg, the system is assumed to have reached a critical free volume fraction A = 0.025,
.025 = c[.025 + 4.8 X 10'4 (T — Tg)] + (1 - c)[.025 + 015(T — 19] (9.1)
and since T = T8 of the system:

[4.8 X 10'4 + «5(1 — c)ng
[4.8 X10‘4c + «5(1 - c)]

(9.2)

 

Tg (system) =

Four blends of CA/DEP were obtained from Rotuba extruders (grades MS, MH,
H and H2). Samples for viscosity measurement on the RMS were compression molded
and the weight fraction of DEP was found out by TGA. Samples for DMA were also cut
from the same compression molded slab and these were used to measure the Tgs of the
blends. Blends of EKX with different amounts of DEP were extruded in a twin screw
extruder and similar measurements were made on these also. The viscosity and TE models
were obtained by fitting the data to equations 6.30 and 9.2 respectively.

Blends of the plasticized CA were blended with EKX in different weight ratios.

The morpholgy was studied with SEM and the Tgs were measured by DMA.

136
9.3 Results

Figure 9.1 shows the variation of Tgof CA as a function of volume fraction of

CA.

200.0

 

 

150.0 ’ . i

100.0

 

 

50.0 i / |
0.0 * .

I
-50.0 . . i 7 l '
‘/ 1 l l l .
l j '
-100.0. % I ! . s .
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
volume fraction of CA

Figure 9.1: Estmation of T8 for CA/DEP blends.

 

 

1 g (centlgrade)

 

 

 

 

The value of as used was 7.91 X 10“. Viscosity fit for CA is shown in Figures 6.9 and
6.10. Figure 9.2 shows the measured complex viscosities for six EKX/DEP blends (the
polymer volume fractions are indicated on the figure). Figure 9.3 shows the Kelly and
Bueche equation fitted to the viscosity data at 100 rad/s for EKX/DEP blends. The values
of K and as used were 0.0374 and 0.001 respectively. Figure 9.4 shows the estimation
of Tgs of the EKX/DEP blends as a function of EKX volume fraction. It seems that at
more than 15% DEP, there is some phase seperation and so only the values for

concentrated solutions were used to optimize the parameters.

137

 

1x106;

J

 

1x105;

y—d
><
p—d
O
.h

 

\\
:_\

 

 

complex viscosity (Pa.s)

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

3 22 \
1X10 \ 827 \'\
2 wmk‘
1x10 \ V
.678 \
1x101 ‘5
1x100 '1 ... 1
-l 0 l 2
1x10 1x10 1x10 1x10
frequency (rad/s)
Figure 9.2: Viscosities of EKX/DEP blends.
4.0 ! 1 I l
1 l l l /“
3-0 l l : x /
2.0 l l l ‘ l l /+/l
”.7; 10 ‘ ] I L + / I l
a | l
E: 00 . l l , .
35;, 0 l l l :// l
2 2.0 I l- //'r i
{a ' l/ l l
3 -3.0 M l g
-4.0 h E ‘ l
-50 }
-6.0 l . i , l 1
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0

volume fraction EKX

Figure 9.3: Estimation of viscosity for EKX/DEP blend.

138

100.0

 

80.0

. ' - l '

. l 1 .

1 . I /

'. ‘ l I
60.0 I / +
40.0 J ‘ . & fl,
20.0 l ‘ 1 /l/
0.0 I . r ;

l 1 1
-20.0 1 7 . /' 1 :r
l

-400 ‘fi

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

temperature (centigrade)

 

 

 

 

 

 

-60.0 1r

 

 

. i i
~80.0‘ 1 l a .

0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
volume fraction EKX

Figure 9.4: Estimation of Tgs for the EKX/DEP blends.

 

 

 

 

 

 

 

With the help of these equations it was thus possible to calculate the viscosity of
each phase if the glass transition temperature of the phase was known. Subsequently eight
blends of CA/EKX/DEP were extruded (ratios by weight)

CA(MS)/ EKX blends -30/70, 40/60 and 50/50.
CA(H)/EKX blends - 30/70 and 40/60.
CA(H2)/EKX blends- 30/70, 40/60 and 50/50.

Figure 9.5 shows the T83 of the CA/DEP phase and the EKX/DEP phase for these
blends as measured by DMA. These values of Tg were now used to calculate the volume
fraction DEP and the viscosity of the two phases and thus the volume and viscosity ratio.

The estimated viscosity values for 100 rad/s were used for this purpose.

th
vi:
Al:
pha
be t

mor.

139

 

180 .

160 6\

 

 

p—a
A
O

H
N
O

 

 

p—a
O
O

 

1'", 1

EM pnaae \
~. ~\ )

\

\

\
“N

0 0.2 0.4 0.6 0.8 1
weight fraction plasticized CA in blend

Figure 9.5: Tgs of the two phases in the EKX/CA+DEP blends;
"*" EKX/CA(HZ) blends; "o"EKX/CA(H) blends; "x" EKX/CA(MS) blends.

 

/

 

O\
O

temperature (centigrade)

A
O

 

 

N
O

 

 

 

 

 

 

 

Table 9.1 shows the values calculated for the EKX/plasticized CA blends using
the estimated viscosities at 100 rad/s. The addition of DEP to the blend changes the
viscosity ratio by an order of two but unfortunately viscosity ratios ~ 1 are not obtained.
Also, in the blends in which DEP volume fraction is more than 15 %, there may be some
phase seperation in which case, that Tg readings and the estimated viscosities would nOt
be dependable. Due to these reasons and unavailablity of more plasticized CA samples,

more blends were not made.

~\-\-.-—.u~:rv n‘na Q~A-wv~t

 

 

 

 

 

 

 

 

 

 

 

m+mvod _+mmo.~ v+mwwd 5.3 03: m. $2 a
v+m§.~ _+m£.m m+mwm.w 3.0m Ywi v. m2 w
v+m36 ~+mmvs c+m:.m once cdfl m. m2 h
v+m_mo.c _+mvm.v c+m>md cwdm ~62 v. = o
m+mvmd m+mfll n+me.m mvém ng m. m m
v+mmmg ~+m:o.~ m+mvv.m me >63 m. N: v
film—9N m+moc.w c+mnmd flame méfl v. N: m
m+m:.~ N+m~ng n+m$.m $0.0m Owcfi m. «E N
o+mm_.m m+mow.m a+mtkw new 0.9: m. .o v. .m. <0 83 ~
833 own—E 82¢ 32:. n53

0:8 mmEXMm meEU mun—DOE awn—EU E 5:23 new: <0 .02

9885 «o .3885 go €885 mo d. ho uh Emma? mo oufiw 9.2m

 

 

 

 

 

 

 

 

.853 Xv—m\<0 Edema—a .2 8:8 SE88; nan—:Bmu Q5 «35.

 

141

Figure 9.6 shows the transverse section of blend 5 and Figure 9.7 shows the
transverse section of blend 9. In these two blends and blend numbers 1, 2, 3, 6, 7 and
8, the CA phase is dispersed in the EKX matrix in the form of spherical particles. The
change in viscosity ratio has apparently had not much effect on the distribution or the
domain size of CA phase. However, in blend 4, the two phases are co-continuous (Figure

9.8) as opposed to dispersed CA morphology for blend 1(Figure 7.15).

 

Figure 9.6: 30/70 CA(H)/EKX blend; 1000X
Sectioning perpendicular to direction of flow.

 

4 A . .
Figure 9. 7: 50/50 CA(MS)/EKX blend.1000X
Sectioning perpendicular to direction of flow.

 

ISKU 111003.. 121.00 1.31.9 .
Figure 9. 8: 50/50 CA(H2)/EKX blend: '
Sectioning perpendicular to direction of flow.

Chapter 10

Conclusions and Recommendations

 

10.1 Extrusion

Cellulose Acetate has a high shear viscosity, of the order of 1011 Pa.s as estimated
in Chapter 6. The extent of degradation of CA was observed to be highly dependent on
the amount of shear. There was considerably less degradation even at high temperatures
in the Micsrotruder. Therefore, high shear rates and long residence times should be
avoided while extruding CA. The self-wiping twin screws configuration elimainates
stagnant zones where degradation could occur. The mixing elements impart high shear
to the melt and so only as many as that required for good dispersion should be used.
Twin screw extruders can be designed for very little axial mixing by using either very
low helix angle screws or straight segments which impose no axial forwarding. The
straight segments have the advantages of greater free path for vapor evolution, and of
increased surface generation without proportionate reduction in residence times [Todd,
1975]. To enhance mixing and avoid product overheating, the application of successive
kneading zones is feasible in short axial length due to excellent pressure generating
capability of the screw elements [Eise, 1983].

PET bottle recycle was found to crystallize faster than even virgin PET, perhaps

due to impurities present in it (Section 7.3). To extrude blends with high PET content,

143

144

it is necessary to avoid any cold spots in the die and barrel where PET could start to
crystallize. PET could also crystallize due to the orientation of molecules and this may
be fascilated by elongationaal flow in the barrel or die. Blends of CAIPET-bottle recycle
were successfully extruded without the use of the die which used to clog due to
crystallization of PET. The use of higher temperatures in the die resulted in low melt
strength. Extrusion of PET bottle recycle was easier with a higher content of CA as the
CA matrix helped to transport the dispersed PET phase.

Another method that could prove feasable is the use of an extruder with two feed
ports. The PET could be fed from the first port along with some CA and melted at
higher temperatures. The rest of the CA could be added downstream where the

temperature would be lower. This setup could possibly reduce the thermal degradation

of CA.

10.2 Trans-esterification

The use of trans-estrification reaction to compatibilize CAIPET or CA/EKX
blends has been shon to be a simple and effective method with the use of reactive
extrusion. The reaction is intrinsically difficult to characterize due to the lack of
formation of any new types of functional groups which could be easily detected by
techniques such as FTIR spectroscopy. Recourse had to be taked to indirect methods to
characterize the reaction. Soxhlet extraction of the CA phase with Acetone was found

more sensitive to the reaction, since the reaction involves species of high molecular

145

weights and so a relatively small extent of reaction translates as a much greater change
in the weight of CA that could be extracted. The amount of PET chains in the extracted
phase was still an unknown and this could be found out by FTIR spectroscopy (if the
amount of PET is greater than the senstivity of the instrument) or CHN analysis. Another
way to do this would be to destroy the CA backbone by acid hydrolysis and titrating for
the amount of glucose units formed.

Scanning Electron Microscopy also provided strong evidence of
transesterification. Due to compatibilization, the particle size of the dispersed PET phase
is seen to be reduced. An increase in adhesion between the two phases is seen in Figure

8.19 where the sample was extruded three times with a catalyst.

10.3 CA/EKX blends

CA blends could be easily extruded with a glycol modified grade of PET called
Transpet EKX-105 (supplied by Hoechst Celanese). EKX had a low degree of
crystallinity and a melting point nearer to that of CA, so higher temperatures were not
necessary during processing.

Thermal analysis of CA/EKX blends, both compatibilized and uncompatibilized,
showed no evidence of miscibility. The mechanical properties of the blends were found
to be highly dependent on morphology of the blends. Dispersed particles of EKX in CA,
helped in increasing the impact strength of the blend. By analyzing the morphology-

property dependence of compatibilized blends, it was concluded that the critical size of

146

EKX particles necerssary to impart impact toughening was ~ 1 micron (section 8.7). A
dispersed phase elongated into fibrils helped to increase the tensile properties of the
blends. Blends with intermediate compositions gave a large dispersed phase particle size
or a co-continous morphology. Both of these had an inverse effect on the mechanical
properties (Figures 7.4-7.8). The tensile properties of uncompatibilized CA/EKX blend
were found to agree with those predicted by models. This could be due to the presence
of transesterification catalysts present in EKX or due to hyrogen bonding between CA
and EKX. Further compatibilization by adding more catalysts had an adverse effect on
mechanical properties. It was concluded that pull-out of EKX fibrils was an important
energy absorbing mechanism and is thus a desirable morphology for CA/EKX blends.

Addition of plasticier should increase the impact strength and make processing easier.

10.4 Plasticized CA/EKX blends

Diethyl pthalate was found to be compatible with both CA and EKX (Chapter 9).
DEP was redistributed in the two phases lowering the Tg and the viscosity of the two
phases. However, the change in the viscosity ratio was not much and thus this teqnique
could not be used to study the dependence of morphology on viscosity and volume ratios.
Addition of 25 % by weight of DEP to the EKX reduces the Tg to room temperature and
this phase dispersed in CA could increase the impact strength.

A plasticizer more compatible with CA and relatively incompatible with PET (or

EKX) would decrease the viscosity of CA only. This would change the viscosity ratio

147

and could make the technique used in Chapter 9 more useful in studying the morphology
of CA/ PET blends. On the other hand, a plasticizer for EKX, which is incompatible with
CA would reduce the Tg of EKX phase only. This would provide impact toughening to
the blend without sacrificing properties such as modulus, strength and service

temperature.

10.5 Achievements and Prospects

The excellent properties of CA/PET blends places them in the category of
"engineering grade resins" like nylons, ABS, PMMA etc. Significant improvement in
properties such as tensile modulus and strength has been achieved by elimination of the
use of plasticizers such as DEP. However, due to blending with another polymer, the
transparency and impact strength asssociated with plasticized CA have been lost. The
impact strength of CA/PET blends could be improved by addition of plasticizers or a
toughening rubbery phase (like rubber modified PS, Polybutyl acrylate modified
PMMA). Such an impact modified grade of CA/PET blend could have hugh commercial
possibilities.

It has been demonstrated that recycled bottle grade PET can be utilized to prepare
an engineering grade material when blended with CA. Since CA itself is made from
Cellulose, a renewable natural resource and is possible biodegradable, this blend can also

be termed as "environment friendly".

148

Currently CA is an underutilized polymer with limited uses when plasticized with
DEP. It is also facing increasing competition from styrenic polymers which are cheaper.
It is thus an oppurtune time to find new uses for CA. Also, recycled PET is much
cheaper than either glycol modified PET or virgin PET. These factors and the ease of
blending and compatibilizing CA/PET blends would make them commercially attractive
alternative to both suppliers of raw materials and compounders.

CA/ PET blends could be potential alternatives to nylons in structural applications,
in household articled such as haircombs, in extrusion operations where the low melt
viscosity of nylons is a disadvantage and in applications where the lower cost of CAIPET
blends would compensate for any loss in propereties.

Glass filled Polyphenyl sulphide articles such as molded bulb sockets, motor
housings, switch components are also potential uses for CA/PET blends due to the high
cost of PPS.

Due to better processibility and heat distortion temperature, another possible
application could be in tiles and laminates where PMMA and epoxy are currently used.

CAIPET could compete with currently available cellulosics (CA, CAB and CAP,
all use monomeric plasticizers) in applications where transparency is not of importance
and greater dimensional stability and heat resistance is required. Potential uses include

tool handles, safety goggles, steering wheels, tabulator keys, and telephone housings.

 

BIBLIOGRAPHY

Bibliography

Ahroni, S. M., J. Macromol. Sci. (B), 22, 813 (1983).

Allen, G. et al, Polymer, 15, 28 (1974).

Berry, G. C. and Fox T. G., Adv Polymer Sci.,5, 261 (1968).
Berry, G. C. V. C. Long, and L. M. Hobbs, Polymer, 5, 31 (1964).

Bikales, N. M. and L. Segal, eds., Cellulose and Celulose derivatives, Wiley
Interscience, NY, 1971, Part IV, page 151.

Brandrup, J ., and E. H. Immergut, eds, Polymer Handbook, Third ed., Wiley-
Interscience, New York (1989).

Brydson, J. A., Plastic Materials -5th edition, Butterworths, London (1989).
Bueche, F., J. Appl. Phys, 24, 423 (1953).

Bueche, F ., J. Appl. Phys, 26, 738 (1955).

Bueche, F. J., J Chem. Phys, 25, 599 (1956).

Bueche, F., J. Appl. Phys, 30, 1114 (1959).

Burke, J . J. and Weiss V., ed., Characterisation of Materials in Research,
Syracuse Univ. Press, Syracuse (1975). '

Burkhardte, K. H. Hermann, and S. Jakopin, SPEANTEC Tech. Papers, 36, 498
(1978).

Chaffey, C. E. et a1, Rheol. Acta, 4, 56 (1965).

149

150

Cheung, P. D. Suwanda and S. T. Balke, Polymer Engg. and Sci., 30 (17), 103
(1990).

Chia, L. and Ricketts S. , Basic Techniques and Experiments in Infrared and
FT-IR Spectroscopy, Perkin Elmer publication 0993-8420, March 1988.

Cohen, M. H., and D. Tumbull, J. Chem. Phys, 31, 1164 (1959).

C. van der Poel, Rheol Acta, 1, 198 (1958).

Davies, W. E. A., J. AppLPhys, 4, 1176 (1971).

Dickie, R. A. in Polymer Blends Vol. 1, Academic press, 1978, Chapter 8.
Doolitle, A. K., J. Appl. Phys, 22, 1471 (1951).

Eise, K., J. Curry, and J. F. Nangeroni, Polym. Eng. Sci., 23, 642 (1983).

Elemndorf, J. J. and A. K. van der Vegt, in Progress in Polymer Processing,
V012, Utracki, L. A., ed., (1991).

Escala, A. and R. S. Stein, Adv. Chem. Ser., 176 (1979).

Favis, B. D. and D. Therrien, Polymer, 32, 8, 1474 (1991).

Fixman, M. and J. M. Peterson, J. Am. Chem. Soc., 86, 3524 (1964).
Fox, T. G. and Allen V. R., Polymer, 3, 111 (1962).

Fox, T. G. and Allen V. R., J. Chem Phy.,4l, 344 (1964).

Fujita, H. and Kushimoto A., J. of Chemical Physics, 34 (2), 393 (1961).
Gauthier F. et al., Tran. Soc. Rheol., 15, 297 (1971).

Gleinser, W., et al, Polymer, 35, 128 (1994).

Goldsmith, H. L. and S. G.Mason, J. Colloid Sci., 17, 448 (1962).

Graessley, W.W., Viscoelasticity and flow in Polymer Melts and Concentrated
Solutions, American Chemical Society, 1984.

Han, C. D., Multiphase flow in Polymer Processing, Academic Press, New
York (1981).

151
Hanrahan, B. D., S. R. Angeli, and J. Runt, Polym. Bull, 15, 455 (1986).

Hashin, Z. J. Appl. Mech., 29, 143 (1962).
Hashin, Z., and S. Shtrikman, J. Mech. Phys. Solids, 11, 127 (1963).

Hetsroni G et al., Progess in Heat and Mass Transfer, G Hestroni, S. Sideman,
and J. P. Hartnett, eds., Vol. 6, p 591, Pergamon, London (1972).

Hill, R., J. Mech. Phys. Solids, 11, 357, (1963).
Hirschfeld, T. B., Anal. Chem, 48, 721 (1976).
Hyun, M. E., and S. C. Kim, Polymer Engg. and Sci., 28 (11), 743 (1988).

Johnson, M.F., W. W. Evans, 1. Jordan, and J. D. Ferry, J. Coll. Sci., 7, 498
(1952).

Kelly, F. N. and F. Bueche, J Palmer Sci.,50, 549 (1961).
Kerner, E. H. , Proc. Phys. Soc. Lond. (B), 69, 808 (1956).

Kirk-Othmer (eds), Encyclopeia of Chemical tehnology, Third Ed. , Wiley
Interscience, New York, 23 (1983).

Kodama, M., Polym. Engg.Sci., 33 (17), 1141 (1993).
Koenig, J. L. et al, Appl. Spectrosc, .31, 292 (1977).
Koenig, J. L., Adv. Polym. Sci., 54, 101 (1983).

Kraus G. and J. T. Gruver, J Palmer Sci.,A2, 797 (1964).
Kuleznev V. N. et al., European Pol. J., 14, 455-461 (1978).

Lee W. K., Ph. D. Thesis (Chem. Eng), Univ. of Houston, Houston, Texas,
(1972).

Lodge, A. S., Rheologica Acta, 2, 158 (1958).
Lyons, P. F., and A. V. Tobolsky, Polym. Eng. Sci., 10, 1 (1970).

Markowitz, H., and D. R. Brown, Trans. Soc. Rheol., 7, 125 (1963).

152

Michigan Dept. of Nat. Res, Statewide Market Study for Recyclable Plastics,
Feb. (1987).

Min K. et a1, Pol. Engg. Sci., 24 (17), 1327 (1984).
Modern Plastics Encylopedia ’90, McGraw Hill (1990).

Nadkarni, V. M. , and J. P. Jog, Handbook of Polymer Science and Technology
Vol.4, ed. N. P. Cheremisinoff, Marcel Dekker, Inc., New York, (1989).

Nadkarni, V. M., and J. P. Jog, Polym. Eng. Sci., 27, 451 (1987).

Narayan, Ramani, Rationale and Design of Environmentally Degradable Plastics:
Recycling back to nature, 83rd Annual Meeting, Air and Waste Management
Asso., Pittsburgh, Pennsylvania, June 24-29 (1990).

Narayan Ramani, in ASC Symposium Series 476, 1 (1991).

Nassar, T. R., D. R. Paul and J. W. Barlow, J. Appl. Polym. Sci.,23, 85 (1979).

Nichols, Kevin, PhD. thesis, Department of Chemical Engineering, Michigan
State University (1991).

Nielson, L. E. , Mechanical Properties of Polymers and Composites Vol. 1,
Marcel Dekker Inc., New York (1974).

Paul, B., Trans. AIME, 218, 36 (1960).

Paul, D. R., and Newman S. , eds. , Polymer Blends Vol.1, Academic Press, New
York (1978).

Paul, D. R., and Newman S. , eds. , Polymer Blends Vol.2, Academic Press, New
York (197 8).

Peterlin, A., J. Polymer Sci., 12, 45 (1954).

Peticolas, W. L., Rubber Chem. Technol., 36, 1422 (1963).

Porter, R. S., and J. F. Johnson, Chem Rev., 66, 1 (1966)

Pryzgocki, W and A. Wlochowicz, J Appl. Polym. Sci., 19, 2683 (1975).

Recycling Saurcebook, Thomas J Cickonski and Karen Hill, eds. , Gale Research
Inc., Detroit (1993).

153
Ryan, A. J., J. L. Stanford and R. H. Still, Polymer, 32 (8), 1426 (1991).

Rumscheidt, F. D., and S. G. Mason, J. Colloid Sci., 16, 238 (1961).

Sakellarides, S. L., and A. J. McHugh, Polym. Engg. Sci., 27 (22), 1662
(1987).

Sheridan, L. A., Chem. Eng. Progress, 71 (2) (1975).
Shingankuli, V. L., J. P. Jog, and V. M. Nadkarni (submitted).
Simha, R. and J. L. Zakin, J. Chem. Phys, 33, 1791 (1960).
Simha, R. and J. L. Zakin, J. Colloid Sci., 17, 270 (1962).
Starita, J. M., Trans. Soc. Rheol., 16 (2), 339 (1972).

Starkweather, H. W. Jr, Polymer Compatibility and Incompatibility, ed. Karel
Solc, Harwood Academic Publ., Chur (1982).

Tanaka, H., Macromolecules, 25, 4453 (1992).

Tavgac, T., Ph. D. Thesis (Chem. Eng), Univ. of Houston, Houston, Texas
'(1‘12193'71312’, G. 1., Proc. R. Soc. Lon, Ser. A, 146, 501 (1934).

Todd, D. B., Chem. Eng. Progress, 71 (2), 81 (1975).

Torza, S. et al., J. Colloid Interface Sci., 38, 395 (1972).

Tsebrenko, M. V., M. Rezanova, and G. V. Vinogradov., Polym. Engg. Sci.,
20 (15), 1023 (1980).

Tucker, C. S., in Plastics Engg., pg 27-30, May (1987).

Utracki, L. A., and M. R. Kamal, Polym. Eng. Sci., 22, 96 (1982).
Utracki, L. A., Polym. Eng. Sci., 23, 602 (1983).

Utracki, L. A., J. Rheol., 35, 1615 (1991).

Utracki, L. and Simha R., J. Polymer Sci., A1, 1089 (1963).

Van der Vegt, A. K., and P. P. A. Smit, S. C. I. Monograph, 26, 313 (1967).

154
Van Oene, H., J. Colloid Interface Sci., 40, 448 (1972).

Vinogradov, G. V. et a1, Int. J. Polym. Metals, 3, 99, 1974.
Wilfong, D. L. et al, J. Mater. Sci., 21, 2014 (1986).

Williams, M. L., R. F. Landel, and J. D. Ferry, J. Am. Chem. Soc., 77, 3701
(1955).

Willis, J. M. et al, J. Mat. Sci., 26, 4742 (1991).

Wood, L. A., in Synthetic Rubber, G. S. Whitby, ed., John Wiley and Sons,
New York (1954).

Wu, 8., Polymer Blends Vol. 1, Academic press, p. 243 (1978).
Wu, 8., Polym Eng. Sci., 27, 335 (1987).

Xanthos M., M. W. Young, and J. A. Biesenberger, Polym. Eng. Sci., 30, 355
(1990).

"I111111111111th