Bimodal polyethylenes form one pot synthesis : effect of
flow induced crystallization on physical properties
Kukalyekar, N.P.
DOI:
10.6100/IR631464
Published: 01/01/2007
Document Version
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers)
Please check the document version of this publication:
• A submitted manuscript is the author's version of the article upon submission and before peer-review. There can be important differences
between the submitted version and the official published version of record. People interested in the research are advised to contact the
author for the final version of the publication, or visit the DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
Citation for published version (APA):
Kukalyekar, N. P. (2007). Bimodal polyethylenes form one pot synthesis : effect of flow induced crystallization on
physical properties Eindhoven: Technische Universiteit Eindhoven DOI: 10.6100/IR631464
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners
and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.
• Users may download and print one copy of any publication from the public portal for the purpose of private study or research.
• You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal ?
Take down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.
Download date: 01. Aug. 2017
Bimodal Polyethylenes from One Pot Synthesis;
Effect of Flow Induced Crystallization on
Physical Properties
PROEFSCHRIFT
ter verkrijging van de graad van doctor aan de
Technische Universiteit Eindhoven, op gezag van de
Rector Magnificus, prof.dr.ir. C.J. van Duijn, voor een
commissie aangewezen door het College voor
Promoties in het openbaar te verdedigen
op woensdag 19 december 2007 om 16.00 uur
door
Nileshkumar Kukalyekar
geboren te Yellapur, India
Dit proefschrift is goedgekeurd door de promotoren:
prof.dr. S. Rastogi
en
prof.dr. P.J. Lemstra
Copromotor:
dr. J. C. Chadwick
A catalogue record is available from the Eindhoven University of Technology Library
ISBN: 978-90-386-1176-1
Copyright © 2007 by N.Kukalyekar
Printed at the Universiteitsdrukkerij, Eindhoven University of Technology, Eindhoven.
Cover design: Nileshkumar Kukalyekar and Paul (Verspaget & Bruinink)
The research described in this dissertation was financially supported by the Dutch Polymer Institute.
(DPI) project # 321b.
|| Ea`Irama ||
Dedicated to,
My parents and Gayatri.
Table of contents
Summary
1. Introduction
1.1. Structure property relationship
1.2. Vision
1.3. Scope of the thesis
1.4. Synthesis of polyethylene
1.4.1 Polyolefin catalysis
1.4.2 Philips catalysts
1.4.3 Ziegler-Natta catalysts
1.4.4 Metallocene catalysis
1.4.5 Post metallocene catalysis
1.5. Why supported catalysts
1.5.1 Silica supports
1.5.2 Alumina supports
1.4.3 Magnesium chloride
1.6. Polymerization and particle growth
1.7. Synthesis route to overcome processability problem and achieve
ultimate properties
1.8. Oriented structures for ultimate properties
1.8.1 Polymer Crystallization
1.8.2 Methods of attaining orientation in polyethylene
1.8.3 Ultimate properties
1.9 References
1
1
2
3
4
4
5
5
7
7
8
8
8
9
9
10
11
11
13
14
14
2. Immobilization of catalysts on MgCl2/AlRn(OEt)3-n supports, Confirmation
and refutation of single centre nature of immobilized catalysts using
melt rheometry
19
2.1. Introduction
19
2.1.1 Characterization of narrowly dispersed PE by GPC and
reliability of GPC
20
2.2. Materials and methods
22
2.2.1 Materials
22
2.2.2 Catalysts
22
2.2.3 Support preparation and catalysts immobilization
23
2.2.4 Polymerization procedure
23
2.2.5 Sample preparation for rheometry
23
2.2.6 Polymer characterization
24
vi
Table of contents
2.3. Results and discussion
2.3.1 Immobilization and activation of [1-(8-quinolyl)indenylCrCl2
2.3.2 Titanium and vanadium amidinate complexes.
2.3.3 Titanium and vanadium pincer and phebox complexes
2.3.4 Titanium and zirconium based systems
2.4. Conclusion
3. Synthesis and characterization of bimodal polyethylene prepared in a
single reactor using a co-immobilized dual catalyst
3.1. Introduction
3.2. Materials and methods
3.2.1 Materials
3.2.2 Catalysts
3.2.3 Catalysts immobilization and polymerization
3.2.4 Sample preparation for melt rheometry
3.2.5 UV-VIS spectroscopy
3.2.6 Polymer characterization
3.3. Results and discussion
3.3.1 Use of hydrogen for the synthesis of bimodal polyethylene
3.3.2 Bimodality by dual catalyst system.
3.3.2.1 UV-VIS spectroscopy
3.3.2.2 Polymerization
3.3.2.3 Thermal analysis
3.3.2.4 NMR characterization
3.3.2.5 Rheology
3.4. Conclusion
3.5. References
24
25
28
29
32
33
37
37
40
40
40
40
41
41
41
41
41
44
45
47
50
53
55
58
59
4. Effect of long chains on the flow induced crystallization of bimodal polyethylene
61
4.1. Introduction
61
4.2. Materials and methods
64
4.2.1 Materials
64
4.2.2 GPC characterization
64
4.2.3 X-ray characterization
64
4.2.4 Differential scanning calorimetry
65
4.2.5 Experimental protocol for flow induced crystallization
66
4.3. Results and discussion
66
4.3.1 Flow induced crystallization at 137 °C
68
4.3.2 Degree of orientation.
72
4.3.3 Thermal analysis
74
4.3.4 Suspension of only shishes in the PE matrix
75
4.4. Conclusion
78
4.5. References
78
Table of contents
5. Towards ultimate properties of bimodal polyethylene
5.1. Introduction
5.1.1 Deformation of polymers
5.1.2 Deformation, desired morphology and the ultimate properties
5.2. Materials and methods
5.3. Results and discussion
5.3.1 Possibilities of generating oriented structures
5.3.2 Solid state drawing
5.3.3 Solid state drawing and mechanical testing
5.3.4 Melt extrusion of bimodal blends
5.4. Conclusion
5.5. References
vii
81
81
82
83
84
84
85
85
87
91
92
92
6. Investigation of the synergistic effect on catalyst activity of a co-immobilized
dual catalyst system containing vanadium and iron catalysts
95
6.1. Introduction
95
6.2. Materials and methods
96
6.2.1 Materials
96
6.2.2 Catalysts
96
6.2.3 Catalyst immobilization and polymerization
97
6.2.4 Polymer characterization
97
6.3. Results and discussion
97
6.3.1 Synergistic effect
98
6.3.2 UV-VIS analysis
101
6.3.3 Chain branching
103
6.3.4 Ligand exchange reaction
104
6.3.4.1 Synthesis of V complex
106
6.3.4.2 Synthesis of V complex
106
6.3.4.3 Polymerization results
106
6.3.4.4 Polymer morphology
108
6.4. Conclusion
108
6.5. References
109
Technology Assessment
Samenvatting
Acknowledgements
Curriculum vitae
111
115
119
121
viii
Table of contents
Summary
The ultimate properties of a polymer are governed not only by the chemical structure of the
polymer chains but also by the processing conditions applied during fabrication of the final
product, in particular as a result of orientation of long-chain molecules. The intrinsic properties
of a polymer are governed mainly by its molecular weight; the higher the weight-average molar
mass (Mw), the better are the physical properties of a polymer. High molecular weight is essential
for better mechanical properties such as tensile strength, toughness and wear properties.
However, high molecular weight polymers, due to a prohibitively high melt viscosity, become
intractable and difficult to process using conventional processing techniques. Up to a critical
molecular weight (Mc, typically twice the average molecular weight between entanglements) of a
polymer, the viscosity increases linearly with molecular weight (Mw<Mc ; η0 ~ Mw), but above
Mc the viscosity increases rapidly (Mw>Mc ; η0 ~ Mw3.4). Intractability of a high molecular weight
polymer limits the means of its processing and its applications. Therefore the finished products
become too expensive due to the high cost involved in specially designed processing techniques.
Thus, for a successful product design, chemistry, physics and engineering have to be addressed
in parallel with equal stress. In practice, polymer (plastics) processing involves a compromise
between properties (for which high molecular weight is desirable) and (fast) processability,
which normally requires lower molecular weight.
Often, a polymer with low molecular weight is mixed with a high molecular weight polymer.
However, melt-phase mixing (blending) of two polymers is difficult, particularly when one of
the components has high molecular weight, and therefore a very high melt viscosity. Solution
blending is not a realistic option on an industrial scale due to the high costs. In industry, bimodal
polymers are produced by sequential polymerization using two or more reactors in series; low
molecular polymer is produced in the first reactor and high molecular weight polymer in the
second.
In this thesis bimodal polyethylenes are synthesized by immobilizing two different catalysts on a
single support, in order to overcome the problem of miscibility of the high and low molecular
weight components. This route overcomes the usage of cascade reactors and provides the
required homogeneous distribution of high molecular weight polymer (Mw > 1 million g/mol) in
a matrix of low molecular weight polyethylene.
In addition to the synthesis, significant attention has been paid to the morphology and rheology
(flow properties) of the materials produced. Under defined flow conditions, different relaxation
times of the low and high molecular weight chains lead to crystallization of the high molecular
x
Summary
weight material, forming a so-called shish structure. This means that the long chains are
pulled/oriented in the flow direction and can act as nucleation sites for crystallization of the
low(er) molecular weight chains present in the blend, resulting in the known shish-kebab
morphology; a fibrous extended-chain nucleus comprising oriented (high molecular weight)
chains, decorated with folded-chain crystals, whereby the low molecular weight polymer can
crystallize at relatively high temperatures. The desired ultimate properties, such as the Emodulus and tensile strength, are linked to this structure development during flow.
Several catalysts were immobilized on supports of type MgCl2/AlRn(OR’)3-n to synthesize
polyethylenes differing in molecular weight and molecular weight distribution. Melt rheometry
was effectively used to differentiate between so-called single-centre and multi-centre catalysts.
Investigation of the melt rheological properties of polyethylenes prepared with chromium,
titanium, vanadium and zirconium catalysts immobilized on MgCl2 supports confirmed narrow
(Schulz-Flory) molecular weight distribution, indicative of single-centre catalysis, for Cr-, Vand Zr-based systems, but not for Ti. In the case of polymers prepared with MgCl2-immobilized
Ti complexes and having narrow MWD (Mw/Mn 2-3) according to GPC, deviation from a
Schulz-Flory distribution is evident from a decrease in storage modulus with decreasing angular
frequencies, whereas polyethylenes prepared with analogous vanadium complexes exhibit a
constant (plateau) modulus over a wide frequency range. The presence of a plateau modulus is
characteristic for narrow MWD polyethylene, so that melt rheology provides a valuable tool to
prove, or disprove, the presence of a Schulz-Flory distribution in cases where GPC does not
provide a definitive answer.
and
[2,6-{(2-chloro-4,6Co-immobilization
of
1-(8-quinolyl)indenyl
CrCl2
dimethylphenyl)N=C(Me)}2C5H3N]FeCl2 on a MgCl2/AlEtn(OEt)3-n support was used to produce
bimodal polyethylene in a single reactor, termed a one pot process. In this way, intimate miscible
blends of high (Mw = 106 g/mol, Mw/Mn = 2.0) and low (Mw = 1 × 105 g/mol, Mw/Mn = 5.5)
molecular weight PE were synthesized. The relative proportions of the high- and low-molecular
weight components are controlled by varying the loadings of the two precatalysts. Rheological
studies of bimodal blends suggest homogeneous mixing of the high and low molecular weight
components within the time frame of the experiments under low strain.
Homogeneously mixed very high and low molecular weight polyethylenes synthesized in this
way are model systems to study the influence of molecular weight on crystallization under shear
conditions. Due to distinct differences in the relaxation times, the influence of molar mass on
crystallization and the resultant crystalline morphology during shear flow was investigated by
time-resolved X-ray techniques. The observations are that highly stable shish structures are
generated from high molecular weight chains, providing nucleation sites for further
crystallization. Shish structures thus formed are found to be stable above the equilibrium melting
temperature of unconstrained extended chain crystals of polyethylene, in the absence of flow.
Summary
xi
Structure development during flow is linked to the mechanical properties of the bimodal
polyethylene. Several of the bimodal polyethylenes synthesized, differing in the proportions of
the high and low molecular weight components, were used for solid state drawing. During solid
state drawing at 137 °C, orientation of the high molecular weight chains was found to be
influenced by the presence of the low molecular weight matrix. This was confirmed by a strong
dependence of the mechanical properties of bimodal PE on the amount of high molecular weight
component.
Finally, as part of the investigations on polymerization catalysis, several experiments were
carried out involving the co-immobilization of [PhC(NSiMe3)2]VCl2(THF)2 and [2,6-{(2-chloro4,6-dimethylphenyl)-N=C(Me)}2C5H3N]FeCl2 on a MgCl2/AlEtn(OEt)3-n support, which
indicated a synergistic effect on the catalyst activity. This effect was not evident in the case of
co-immobilized chromium and iron precatalysts. Several possible reasons for this synergistic
effect were investigated; however no definitive reason could be found. Nevertheless, two new
vanadium complexes, namely [2,6-{(2-chloro-4,6-dimethylphenyl)-N=C(Me)}2C5H3N]VCl3 and
[2,6-{(2,4,6-trimethylphenyl)-N=C(Me)}2C5H3N]VCl3, were synthesized during the course of
this investigation. These complexes are not active for ethylene polymerization under
homogeneous conditions, but were active after immobilization on a MgCl2/AlEtn(OEt)3-n
support. [2,6-{(2-chloro-4,6-dimethylphenyl)-N=C(Me)}2C5H3N]VCl3 gave high molecular
weight polyethylene with narrow (Schulz-Flory) molecular weight distribution.
xii
Summary
Chapter 1
Introduction
Polyolefins, which include large volume materials such as polyethylene (PE) and polypropylene
(PP) and specialty materials such as elastomers, have become established as commodity plastics
due to their cost/performance balance and economical and environmentally friendly production.
During the last 50 years, polyolefins have undergone a tremendous growth and still have further
growth potential due to their versatility in properties and applications, low cost and easily
available raw materials
1.1 Structure Property Relationship
Nowadays polymers, ranging from the large volume commodity plastics to the small volume
specialty polymers, have replaced metals and other natural products in many applications and
therefore properties such as strength and long life of a polymer become a requisite in such
applications. Ultimate polymer properties are dependent on several aspects, starting from the
chemical and physical structure of polymer chains to the mode and conditions of their
processing. The intrinsic properties of a polymer (for a given microstructure) are governed to an
important extent by its molecular weight; the higher the molecular weight (MW), the better are
the physical properties of a polymer. High MW is essential for better mechanical properties,
fatigue resistance and wear properties. However, high MW polymers, due to their high melt
viscosity, become intractable and difficult to process using conventional processing techniques.
Up to the critical molecular weight (Mc, typically twice the average molecular weight between
entanglements) of a polymer, the viscosity increases linearly with molecular weight, but above
Mc the viscosity of the polymer increases with the 3.4th power of the molecular weight.
Intractability of a high MW polymer limits the means of its processing and its applications.
Therefore the finished products become expensive due to the high cost involved in specially
designed processing techniques.
The problem of intractability of high MW polymer melts can be overcome by techniques
involving solution routes. The discovery of solution spinning of ultra high molecular weight
polyethylene (UHMW-PE, Mw> 106 g/mol) led to the production of high-modulus, high-strength
fibres used in highly demanding applications.1 These strong fibres are generated by spinning a
1
Chapter 1
semi-dilute solution of UHMW-PE in decalin or paraffin oil; subsequent extraction of solvent
from gel filament and hot drawing leads to a strong fibre at the maximum draw ratio. The
strength of industrially produced Dyneema ® fibre exceeds 3 GPa, with a modulus of around 120
GPa.
High-strength, high-modulus fibres are used in several applications such as lightweight ropes,
nets and textiles. However, as-spun fibres can not be used efficiently in other applications where
the strength is required not only in the fibre direction but also in the transverse direction. For
such applications, fibre composites are prepared by reinforcing or gluing the fibers together with
other resins2 or with a low molecular weight material similar in chemical composition to that of
the fibres.3 The fibres can possess high strength and high modulus, approaching the theoretical
limits of the material; however, the overall properties of the composite are lowered due to the
lower modulus and strength of the adhesive material.
In practice, even for highly demanding applications, the strength of fibres is not always required
and oriented tapes or wires are sufficient for many applications. These oriented products can be
prepared using conventional processing techniques such as melt processing and solid state
drawing. However, for convenient processing the polymer composition has to be tailored without
compromising much on the properties. In this thesis, in order to achieve a golden mean between
the ease of processing and ultimate properties, an attempt is made to combine engineering with
polymer chemistry and physics.
1.2 Vision
The goal of this thesis is to combine the advances in catalysis, polymer crystallization and
processing to obtain high-strength, high-modulus tapes or wires. In this thesis, a possibility of
generating high-strength, high-modulus tapes or wires by using low MW polyethylene as an
adhesive for a matrix of highly oriented high MW polyethylene chains is discussed.
A fibre composite is prone to fail at low force due to the low strength and low modulus of matrix
material, as well as poor adhesion between fibres and the matrix.3 However, if the matrix
material is the same as that of the fibres and organized in a manner such as to strengthen the
composite in the direction normal to fibres, the composite can resist deformation in both fibre
and transverse direction for higher applied load. Upon flow in the molten state and subsequent
crystallization followed by hot drawing, oriented structures such as shish and interlocked kebabs
are reported in the literature.4 It is possible to generate this structure by forming fibrous extended
chain nuclei and then growing lamellar crystals on them. The lamellar crystals, due to
interpenetrated structures, assist in preventing fibrillation. A schematic representation of
interlocked shish-kebab structures is depicted in Figure 1.1. The most important requirement to
achieve this morphology is the minimum distance between shishes. If the shishes are spaced far
2
Introduction and scope
apart the lamellar crystals start twisting, leading to a lower modulus in the direction normal to
the fibre direction.5 So, by optimizing the proportion of the oriented fibers and the interlocked
kebab (formed from L-MW-PE), the spacing between the shishes can be optimized. To satisfy
this condition, there is a need for tailoring of the composition of high MW and low MW PE in
the blend, in combination with the appropriate processing conditions. Appropriate composition
of the high and low MW components in PE can widen the molecular weight distribution (MWD)
to attain easy melt processability, and recent advances in catalysis can assist in obtaining the
required miscible blends of high and low molecular weight polyethylene.
Shish
Shish
Interlocked
Kebab
Figure 1.1 Shish and inter-locked kebab structures. (Shading is done for differentiation)
1.3 Scope of the Thesis
The route of heterogeneous catalysis, which is preferred by the polyolefin industry, is chosen for
the synthesis of polyethylene. In Chapter 2, a method of heterogenization of catalysts and the use
of supported catalysts in ethylene polymerization is described. A qualitative method for
confirmation or refutation of single centre characteristics by using melt rheometry is also
described. In Chapter 3 the synthesis of miscible bimodal blends of high and low molecular
weight polyethylene by immobilizing two different catalysts on a MgCl2/AlRn(OEt)3-n support is
reported. In Chapter 4, the bimodal blend is used to investigate the effect of high molecular
weight polyethylene chains on flow-induced crystallization. It is successfully demonstrated that
it is possible to make a suspension of only shishes in the matrix of low molecular weight
polyethylene. In Chapter 5, it is demonstrated that by tailoring the composition of bimodal PE
and controlling the crystallization, high-strength, high-modulus tapes or wires can be prepared.
In Chapter 6, a synergistic effect of catalyst co-immobilization on catalyst activity is studied
using several characterization techniques. Finally, the industrial relevance and technological
impact of the results presented in this thesis is described in Technology Assessment.
In the present chapter, the chemistry of polyethylene synthesis will be outlined, followed by a
brief overview of the tremendous amount of work done in the area of oriented polyethylene for
achieving ultimate properties.
3
Chapter 1
Catalysts Synthesis and
Immobilization
Processing and
Mechanical Properties
High strength
-High Modulus
Polyethylene
Flow Induced Crystallization
of Bimodal PE
Polyethylene
Synthesis and
Characterization
Synthesis of Bimodal
PE
Figure 1.2 Scope of the thesis.
1.4 Synthesis of Polyethylene
Catalysts play an important role in the production of polyolefins. In this thesis the synthesis of
polyethylene is performed using supported catalysts. Because catalysis is an integral and
important part of this thesis, it is important to review the genesis and progress of polyolefin
catalysis. Information about the various classes of polyethylene and processes for their industrial
production is given in the Appendix.
1.4.1 Polyolefin Catalysis
Catalysts play an important role in many chemical reactions. In the chemical industry, 60 % of
the chemicals produced involve catalytic processes.6 Supported catalysts7 are used world wide
for the production of polypropylene and 60 % of polyethylene. Most of the commercial
polyolefin catalysts are heterogeneous Ziegler-Natta or chromium oxide systems. Metallocenes,
being able to produce polyolefins with well defined molecular architecture, are versatile
catalysts; however, they occupy a relatively small market share.
Historically the synthesis of polyethylene dates back more than 100 years, when von Pechmann
polymerized diazomethane,8 although this route was not industrially viable. In 1933 Fawcett and
Gibson at ICI discovered the high pressure free radical ethylene polymerization process, which
was the first breakthrough in the industrial production of PE.9 Subsequently, several researchers
worked relentlessly in search of highly efficient catalysts for ethylene production. During the
1930s Marvel and Friedrich discovered that BuLi polymerized ethylene to produce linear
polyethylene.10 At BASF AG, Max Fischer obtained high molecular weight liquid and solid byproducts with a mixture of TiCl4 and Al powder.11 The decade of the 1950s proved to be the
most important decade in the history of polyolefin catalysis. In 1951 a breakthrough in
polyolefin catalysis occurred when the Phillips catalyst was invented by Hogan and Banks at
4
Introduction and scope
Phillips Petroleum Company. The biggest breakthrough, however, came in 1953 when Karl
Ziegler at the Max Planck Institute of Coal Research in Mülheim discovered heterogeneous
catalysts which were capable of producing HDPE at low temperature and pressure.12
1.4.2 Phillips catalysts
Phillips types of catalysts are based on chromium oxides supported on amorphous silica. The
ability of the Phillips catalyst to polymerize ethylene without the use of any activator makes it
unique among all olefin polymerization catalysts. The Cr/SiO2 catalysts are used in the
commercial production of more than 30 % of all PE produced worldwide. This invention was
clearly revolutionary as these catalyst systems can make more than 50 different types of HDPE
and LLDPE. The structure of the active centers and the ethylene polymerization mechanism of
the Cr/SiO2 catalyst is extensively reviewed by Groppo at al.13
1.4.3 Ziegler-Natta Catalysts
In 1963 Karl Ziegler and Giulio Natta were awarded the Nobel Prize for Chemistry for their
breakthrough discovery of polyolefin catalysts. Ziegler, while working on ethylene
oligomerization with aluminium alkyls, the reaction which he termed the ‘Aufbaureaktion’,
discovered in 1953 that HDPE could be synthesized under mild conditions using combinations of
a transition metal compound and an aluminium alkyl.14 In 1954 Giulio Natta and his co-workers
at Milan Polytechnic polymerized propene to obtain a tacky solid using the Ziegler catalyst.
Extraction with boiling solvents led to the isolation of amorphous and crystalline fractions,
which Natta characterized with X-ray diffraction, observing stereoregularity in the crystalline
fraction. He then coined the terms isotactic, syndiotactic and atactic for the stereo isomers, which
are shown in Figure 1.3.
A
B
C
Figure 1.3 Polypropylene stereoisomers A. Isotactic B. Syndiotactic and C. Atactic.
Ziegler-Natta catalysts are typically heterogeneous catalysts formed by a salt of a transition
metal from groups IV–VIII and a metal alkyl, called the co-catalyst or an activator, of an element
from group I-III. There are many patents concerning combinations of metal alkyls and transition
metals, however only a few are used in practice. Aluminium alkyls such as AlEt3, Al-i-Bu3 and
5
Chapter 1
AlEt2Cl are preferred as co-catalysts, while transition metal compounds containing Ti, V and Cr
are mainly used. In 1964 Cossee and Arlman15 proposed a monometallic mechanism for ZieglerNatta olefin polymerization, wherein the active centre- contains a transition metal-carbon bond.
The Cossee-Arlman mechanism is depicted in Figure 1.4.
Polymerization proceeds by two steps, first coordination of monomer to the active centre and
then migratory insertion of coordinated monomer into the metal-carbon bond. In this mechanism
a vacant coordination site is regenerated, which enables further chain propagation.
R
R
M
+
M
R
R
M
M
Figure 1.4 Cossee-Arlman mechanism for Ziegler-Natta olefin polymerization.
Termination of a growing chain takes place by a chain transfer reaction; chain transfer to
monomer, metal alkyl or chain transfer agent is known to terminate the growing chains. Thermal
cleavage of an active centre involving β-hydrogen elimination to generate a hydride species also
terminates the reaction.16
R
R
CH2
CH
CH2
CH2
H
M
CH2
CH2
M
CH2
CH
+
CH2
R
R
CH 2
CH3
M
CH3
+
H
M
H
H
R
CH2
M
CH
H
CH2
R
M
CH
+
H
Figure 1.5 Termination reactions in olefin polymerization.
The discovery of the high activity ‘δ’ MgCl2-supported catalyst for ethylene polymerization in
6
Introduction and scope
1968, extended to isospecific propylene polymerization in 1975, was a milestone in the process
development of polyolefin production.17 Evolution of highly active and stereospecific Ziegler–
Natta catalysts has eliminated the steps of catalyst residue and atactic polypropylene removal.12
Despite the shortfalls of multi-centre nature of these catalysts and non-homogeneous comonomer incorporation, Ziegler-Natta catalysts are still the dominant catalyst systems in
commercial polyolefin production.
1.4.4 Metallocene catalysts
The evolution of single-centre catalysts started when Breslow18 and Natta19 used Cp2TiCl2 in
combination with alkyl aluminium compounds for ethylene polymerization. Metallocene
catalysts are organometallic coordination compounds containing two cyclopentadienyl rings or
substituted cyclopentadienyl rings bound to a central transition metal atom. Different
substitutions at the cyclopentadienyl ring give rise to several different ligand structures. These
catalysts each have a unique active centre nature and are hence called Single-Centre catalysts.
This single-centre nature leads to uniformity in the polyolefins produced using these catalyst
systems.20
A revolutionary discovery by Kaminsky and Sinn in 1980s, when they reported very high
activities of single-centre catalysts using methylaluminoxane (MAO) as a cocatalyst,21 triggered
a great deal of scientific research in the area of single-centre catalysts and their immobilization.22
Although water poisons Ziegler-Natta catalysts, addition of small amounts of water to
Cp2TiEtCl/AlMe2Cl and Cp2TiCl2/AlMe2Cl systems increased the catalyst activity23,24
tremendously. Sinn and Kaminsky observed that the weakly active Cp2ZrMe2/AlMe3 system
could be activated by addition of water. The formation of MAO by partial hydrolysis of AlMe3
was the cause of the increased activity.
1.4.5 Post Metallocene catalysts
Apart from metallocene catalysts,20e many different types of homogeneous catalysts have been
developed for olefin polymerization. These include half-metallocene complexes such as
dimethylsilyl-bridged amidocyclopentadienyltitanium complexes, termed constrained geometry
catalysts, which can generate ethylene copolymers with long-chain branching via incorporation
of vinyl-terminated chains.25 During the 1990s a considerable amount of research was carried out
on non-metallocene catalysts involving late-transition metals. The discovery of highly active (αdiimine) nickel catalysts,26,27 which were able to produce linear and branched polyethylene,
opened the doors to a treasure of late transition metal catalysts. Greater tolerance to oxygen and
functional groups have made these catalysts potentially attractive for the use of polar
comonomers.28 The α-diimine ligands, synthesized by condensation of a diketone with 2
equivalents of alkyl or aryl amines catalysed by Lewis or Brønsted acid, stabilize organometallic
7
Chapter 1
complexes. Brookhart and Gibson reported a series of iron and cobalt complexes containing a
bis(imino)pyridyl ligand.29,30 These complexes are highly active for ethylene oligomerization and
polymerization. However, in contrast to the majority of homogeneous olefin polymerization
catalysts, these systems typically give polyethylenes with relatively broad polydispersities.
Recently developed bis(phenoxy-imine) group 4 transition metal complexes, known as FI
catalysts reported by Fujita, are also highly active in ethylene polymerization.31,32
1.5 Why Supported Catalysts?
Most of the commercial polyolefin catalysts are heterogeneous Ziegler-Natta or chromium oxide
systems. Generally, homogeneous catalyst systems are not capable of producing polyolefins with
regular spherical morphology, whereas this can be achieved with heterogeneous catalysts.
Industrially viable gas and slurry polymerization processes require so called drop-in catalysts;
this condition is fulfilled by heterogenization of homogeneous catalysts.22 Different types of
supports and the mechanism of polymer particle growth will be discussed in the coming section.
Supported catalysts are prepared by effecting physical and/or chemical immobilization of
catalysts on the surface of solid carriers. Carriers used in Ziegler-Natta catalysts are silica and
magnesium chloride (MgCl2), while the Phillips catalyst involves the use of silica or
silica/alumina. MgCl2 is a crystalline, ionic solid into which the catalyst active centers are
inserted via isomorphic lattice substitution. Silica is an amorphous solid on which chromium
oxide is fixed by calcination at high temperature. Several other support materials have been
explored and have been reported in the literature. The most widely used support materials will be
discussed in the following section.
1.5.1 Silica supports: Silica is the most widely explored support material used for
immobilization of single-centre catalysts. A general review of silica-based catalysts has been
compiled by Pullukat et al.33 As discussed earlier the Phillips catalyst, which is one of the most
widely used catalysts in the industrial production of PE, is immobilized on silica. Silica exists in
many forms but typically amorphous silica, which includes anhydrous amorphous silica and
surface-hydroxylated amorphous silica, is used as a carrier for catalysts. The structure of catalyst
active centres anchored on silica depends upon the surface structure and chemical properties of
the support. Sauer et al.34 have discussed in detail theoretical studies of van der Waals forces at
surface sites.
1.5.2 Alumina Supports: Alumina, which is mainly γ-Al2O3, has also been used as a support
material for the immobilization of polyolefin catalysts. Mixed supports prepared by cocondensation of silica and alumina have also been developed.
8
Introduction and scope
1.5.3 Magnesium Chloride:
The discovery in the late 1960s of “activated” magnesium chloride as a support for TiCl4,
followed by the incorporation into the catalyst system of electron donors giving high
stereospecificity in propylene polymerization, was a landmark event in the polyolefin
industry.35,36 Since then, magnesium chloride has been literally at the core of the polyolefins
synthesized by heterogeneous catalysts. Similarity in the ionic radii of Mg2+ and Ti4+ and large
surface area make MgCl2 the most suitable support material for Ziegler-Natta catalysts. In the
late 1970s MgCl2-supported TiCl4 catalysts replaced TiCl3 catalysts, leading to dramatically
increased catalyst activities. Doi37 proposed that the increased activity resulted from electron
donation from MgCl2 to Ti, stabilizing the monomer coordination to the active centre. However,
Farina calculated a higher electrophilicity of Ti atom when TiCl4 was supported on MgCl2, due
to an ‘electron sponge’ effect of the support.38
MgCl2 has also been used for the heterogenization of single-centre catalysts, although not as
extensively as silica. Marks39 has reported the activation of [(C5Me5)2ThMe2] by MgCl2 , where
the Lewis acidity of MgCl2 leads to the abstraction of methide anion to generate a catalytically
active actinide center. Sivaram et al. have described a method wherein a solid catalyst is obtained
by addition of hexane to a solution of [Cp2TiCl2] and MgCl2 in tetrahydrofuran.40 This catalyst,
in the presence of AliBu3 or AlEt2Cl, was active in ethylene polymerization. Soga and
Kaminaka41 supported zirconocene on MgCl2 and, using AlR3, produced polyethylene with broad
molecular weight distribution.
Magnesium chloride and ethanol adducts have also been used to prepare supports for singlecentre catalyst immobilization. Sacchetti et al.42 at Montell (now Basell) demonstrated that
partial dealcoholization of spherical adducts of composition MgCl2·3EtOH, followed by reaction
with trialkylaluminium and subsequently with a metallocene, is an effective method for catalyst
immobilization and activation. MgCl2-based supports containing MAO have also been used.43
Recently, several groups have used MgCl2/AlRn(OR')3-n type of supports, obtained by reaction of
AlR3 with solid or soluble adducts of MgCl2 and an alcohol, for immobilization and activation of
early- and late-transition metal catalysts.44-46 Supports of type MgCl2/AlEtn(OEt)3-n, prepared by
the reaction of MgCl2·1.1EtOH with AlEt3, have formed the basis for the catalyst immobilization
and polymerization studies described in this thesis.
1.6 Polymerization and particle growth
Polymerization starts when a catalyst particle, typically 10-100 micrometers in diameter, comes
into contact with cocatalyst and monomer, either in the gas or liquid phase. Monomer diffuses
through the catalyst boundary layer to reach the active centres, where the polymerization reaction
takes place. The rate of polymerization depends on the type of catalyst, catalyst surface area and
9
Chapter 1
pore size. There are several models reported for the mechanism of polymer particle growth.47,48
Spherical Polymer
Figure 1.6 Schematic representation of polymer particle growth from an immobilized catalyst
particle.
The spherical supports of 10-100 micrometer diameter generate spherical polymer particles of
around 100-3000 micrometers. The particle size distribution of the polymer depends upon the
particle size distribution of the catalyst. Termed as Reactor Granule Technology by Galli,49,50
replication of the spherical morphology of catalyst particles to the polymer particle is a key
feature in heterogeneous olefin polymerization. Efficient fragmentation and uniform particle
growth are important in the replication process and are dependent on high surface area, a
homogeneous distribution of catalyst active centres throughout the catalyst particle and free
access of the monomer to the innermost areas of the catalyst particle.
1.7 Synthesis route to overcome processability problem and achieve ultimate properties
Broad molecular weight distribution (MWD) is essential for the ease of processing. However,
there is a possibility of a loss of desired mechanical properties by broadening the MWD. In such
a situation, it is desirable to have a bimodal molecular weight distribution, wherein the
composition of a polymer can be altered by tailoring the proportions of high and low molecular
weight components. A schematic representation of molecular weight distribution is shown in
Figure 1.7.
Bimodality in polymers can be achieved by two routes: modifying the polymer architecture
during synthesis in one or several reactor(s), or post-synthesis blending of different polymers.
Post-synthesis blending can be performed by extrusion or solution blending, but effective
blending of polymers having high molecular weights is difficult. The supported catalysts give a
possibility of immobilizing two catalysts capable of generating polymers differing in chemical
structure, branching content and/or molecular weights. Thus the polymer architecture can be
tailored in the reactor. In this thesis in Chapter 3, the synthesis of bimodal PE in a single reactor
using a co-immobilized supported catalyst is described. The bimodal polyethylenes, having a
varied composition of low and high molecular weight components, are model systems for
obtaining the interlocked shish-kebab morphology as described in Figure 1.1.
10
Introduction and scope
Processability
Matrix
Creep Resistance
Mechanical
Strength
Melt
Strength
During
Extrusion
wt %
Reduced
Impact
Strength
Molecular Weight
Figure 1.7 A schematic representation of contribution of various molecular weight fractions to
the property profile of polyethylene with monomodal (---) and bimodal ( ) MWD.
1.8 Oriented structures for ultimate properties
Linear polymers, due to the strong covalent bonding, are potentially strong and stiff in nature.
This intrinsic strength and stiffness can be effectively utilized to improve the mechanical
properties of the flexible chain polymers, such as polyethylene, by chain extension and
alignment. Since the 1970s, a tremendous amount of research has been done in the area of
oriented polyethylene and several reports have been made of high-strength and high-modulus
fibres. Pioneering work in solid state drawing was reported by the group of Ward51, while high
strength fibres from melt processing originated from Keller’s group and were developed by
Odell and Bashir.4,52 Thereafter commercially available gel-spun fibres, which were based on the
solution process reported by Pennings and co-workers,53 were developed by Smith and Lemstra.1
The developmental route to stiff and strong fibres from flexible polymers proceeds via
crystallization and therefore it becomes essential to recollect some aspects of crystallization.
1.8.1 Polymer crystallization
Polymers stay in the random coil state in melt and solution; however, they can crystallize to form
ordered three dimensional structures. The ability of polymer chains to participate in more than
one crystalline unit, thus forming crystalline and amorphous domains, distinguishes polymer
crystals from crystals of small molecules. The presence of both crystalline and amorphous
domains makes the crystallizable polymer semicrystalline.
In polymer crystallization, it is well established that a polymer molecule having chemical and
structural regularity when cooled transforms itself from a random coiled melt state to an ordered
11
Chapter 1
crystalline state.54,55 The crystallinity, when structure and composition is taken into
consideration, depends upon branching, stereoregularity and comonomer content. After the
breakthrough discovery of Ziegler-Natta catalysts,11,16 stereoregular polymers were studied
extensively to gain insight into and to propose new fundamental concepts in polymer
crystallization. Linear polyethylene, consisting of the simplest structural unit, is the most widely
studied polymer in polymer crystallization. Keller, Fischer and Till discovered that linear
polyethylene when crystallized from dilute solution forms thin platelet like structures called
lamellae.56-58 Keller suggested that these crystals were folded chain crystals, which is now
accepted by the scientific community.
Flexible polymers crystallize in the form of chain folded thin lamellae perpendicular to the
surface and connected to the amorphous region by other chains. When a polymer is crystallized
from melt the lamellar crystals form spherical aggregates, termed spherulites, wherein crystal
lamellae grow radially outwards from a common centre. In Figure 1.8, schematics of folded
chain crystals can be seen.
Figure 1.8 Schematic model of folded chain crystals.
Crystallization in polymers can be divided into two types: quiescent crystallization and
crystallization due to flow. In quiescent crystallization, upon cooling from melt the polymer
chains form spherulites, literally spheres, which were first reported in 1945 in polyethylene.59
Since then, they are recognized as characteristic of quiescent crystallization of a polymer melt.
Spherulites are the spherical structures generated from crystal lamellae growing radially
outwards from a common centre. Despite detailed studies, the origin of spherulitic growth,
including nucleation and formation of spherical aggregates of crystals, is not well understood.60
After experiencing flow and subsequent or parallel crystallization, a polymer melt can generate
oriented shish-kebab structures in addition to the spherulites.61,62 These crystalline structures are
depicted in Figure 1.9. Shish-kebabs are crystalline assemblies consisting of a central core of
smooth extended chain crystals called as shish and laterally grown folded chain crystals called
kebabs.61a The formation of oriented structures and the kinetics of crystallization have direct
implications on polymer processing and properties, making flow-induced crystallization studies
essential.
12
Introduction and scope
Kebab
Shish
Figure 1.9 Schematic model of spherulite and shish-kebab structure.
Earlier studies performed using dilute polymer solutions showed that vigorous stirring was
required to obtain full extension of chains to generate shish-kebab like structures.63 Immediate
extension, upon flow, of longer molecules suggested that high molecular weight chains are
preferable to form shish-kebab morphologies,64 thus giving impetus for studies involving chain
elongation in extensional flow rather than shear flow, which is considered to be a weaker flow.
However, recent studies have confirmed that the shear flow can cause orientation in the polymer
melt.65,66 This will be discussed in Chapter 4.
1.8.2 Methods of attaining orientation in polyethylene
In practice, to orient the flexible polymer systems, polymer can either be deformed in the solid
state or crystallized in the oriented form to obtain extended chain morphology. Crystallization of
polymers in the oriented form is achieved by melt spinning and solution spinning, while solidstate deformation is achieved by drawing or extrusion or a combination of both.5
Solid-state deformation is performed above the α-relaxation temperature but below the melting
temperature of PE. The degree of deformation is represented by the draw ratio, which governs
the modulus of drawn fibres, while high molecular weight is required for high strength fibres.
Considering this argument, it appears that high molecular weight polyethylene with high draw
ratio would generate strong and stiff fibres. However, a large number of entanglements does not
allow high draw ratio. If the entanglements are reduced by crystallizing H-MW PE from dilute
solution, then high draw ratio can be obtained. The fibres generated from this process possess
high modulus and high strength. As entanglements play a vital role in draw ratio, efforts were
made to compression mould less entangled (disentangled) UHMW-PE powder to obtain films
with high draw ratio.67
As described above, a reduced number of entanglements can increase the draw ratio. So, a dilute
solution of high molecular weight polyethylene was considered to be an ideal scenario to obtain
stiff and strong fibres. Pennings observed excellent mechanical properties of fibres obtained via a
solution process.53 It was stated that the gel-like layer at the surface permitted crystallization of
oriented structures at higher temperature. Commercially available high-modulus, high-strength
13
Chapter 1
Dyneema® fibres, which were developed by Smith and Lemstra, are produced by solution or gel
spinning of UHMWPE.1 The pioneering work by Smith and Lemstra followed by others shows
the presence of disentangled amorphous regions that tend to facilitate solid state deformation
desired in the production of high modulus, high strength fibres.
In melt spinning, polymer melt is extruded through a spinneret and subjected to drawing while
being crystallized.3,4 This method, though simple, is quite complicated if the ultimate modulus
and strength is desired, as it requires several drawing stages at different temperatures. The
extension and cooling rates govern the mechanism of fibre formation and the fibre properties.
1.8.3 Ultimate properties
Several groups have reported different techniques listed above to obtain strong and stiff fibres
and fibre composites. Each technique has its own advantages and disadvantages, depending
upon the application of the end product and the cost involved. Reported values of modulus and
strength are summarized in Figure 1.10.
In conclusion, as pointed out by Ward,5 these advances were achieved and can only be achieved
by lateral thinking and bringing together ideas from a wide range of expertise. Without
engineering the inventions will not progress very far, while for successful engineering the
chemical and physical structure of the polymer is vital.
Modulus GPa
Drawn
Single
Crystal
Mat
Drawn
Gel
Fibers
400
380
360
340
320
300
280
260
240
220
200
180
160
140
120
100
80
60
40
20
0
Strength GPa
10
9
Theoretical
Values of PE modulus
In Chain Direction
8
Theoretical
values of
Chain Strength
7
6
Drawn
Single
Crystal
Mats
Surface
Grown
Fibers
5
4
3
2
Drawn
Melt Crystallized
PE
Surface grown
and
drawn gel
fibers
1
Drawn Melt
Crystallized
Normal Isotropic
0
Figure 1.10 Scales showing tensile strength and modulus for polyethylene. (Reproduced from
Barham, P. J.; Keller, A. J. Mat. Sci. 1985, 20, 2281).
14
Introduction and scope
1.9 References
1. (a) Smith, P.; Lemstra, P. J. J. Mat. Sci. 1980, 15, 505. (b) Smith, P.; Lemstra, P. J. Polymer
1980, 21, 1341. (c) Smith, P.; Lemstra, P. J. ; Kalb, B.; Pennings, A. J. Polym. Bull. 1979, 1,
733.
2. Parker, D. S.; Yee, A. F. J. Thermoplast. Compos. Mater. 1989, 2, 2.
3. Bashir, Z.; Odell, J. A. J. Mat. Sci. 1993, 28, 1081.
4. (a) Odell, J. A.; Grubb, D. T.; Keller, A. Polymer, 1978, 19, 617. (b) Bashir, Z.; Odell, J, A.;
Keller, A. J. Mat. Sci. 1984, 19, 617. (c) Bashir, Z.; Keller, A. Colloid. Polym. Sci., 1989,
267, 116.
5. Ward, I. M. Structure and Properties of Oriented Polymers, Chapman and Hall 1997, Second
Edition.
6. (a) ‘Recognizing the Best in Innovation: Breakthrough Catalyst’. R&D Magazine,
September 2005, 20. (b) For an overview of catalysis in industrial process see: Catalysis: An
Integrated Approach, Second Edition (van Santen, R. A.; van Leeuwen, P. W. N. M.;
Moulijn, J. A; Averill, B. A. 1999, Elsevier, Amsterdam.
7. Potter, D. J. B.; Tattum, L. Polypropylene Annual Review September 1998, 4.
8. Von Pechmann, H. Ber. Dtsch. Chem. Ges. 1898, 31, 2643.
9. Ballard, G. H.; Seymour, R. B.; Cheng, T. History of Polyolefins, 1986, D. Reidel Publishing
company, Dordrecht.
10. Friedrich, M. E. P.; Marvel, C. S. J. Am. Chem. Soc. 1930, 52, 376.
11. Fischer, M. 1943, DBP874215, 18, 12.
12. Mülhaupt, R.; Macromol. Chem. Phys. 2003, 204, 289.
13. Groppo, E.; Lamberti, C.; Bordiga, S.; Spoto, G.; Zecchina, A. Chem. Rev. 2005, 105, 115.
14. http://nobelprize.org/nobel_prizes/chemistry/laureates/1963/ziegler-lecture.pdf.
15. (a) Cossee, P. J. Catal, 1964, 3, 80. (b) Arlman, E. J. J. Catal, 1964, 3, 89. (c) Arlman, E. J.;
Cossee, P. J. J. Catal. 1964, 3, 99.
16. Huang, J.; Rempel, G. L. Prog. Poly. Sci. 1995, 20, 459.
17. Galli, P.; Vecellio, G. Prog. Polym. Sci. 2001, 26, 1287.
18. Breslow, D. S.; Newburg, N. R. J. Am. Chem. Soc. 1957, 79, 5072.
19. Natta, G.; Pino, P.; Mazzanti, G.; Giannini, U. J. Am. Chem. Soc. 1957, 79, 2975.
20. (a) Kaminsky, W.; Kulper, K.; Brintzinger, H. H., Wild, F. R. W. P.; Angew. Chem. Int. Ed.
Engl. 1985, 97, 507. (b) Jordan, R. F.; Adv. Organomet. Chem. 1991, 32, 325. (c) Kaminsky,
W. Catal. Today. 1994, 20, 257. (d) Mohring, P. C.; Coville, N. J.; J. Orgamomet. Chem.
1994, 479, 1. (e) Brintzinger, H. H.; Fischer, D.; Mülhaupt, R.; Rieger, B.; Waymouth, R. M.
Angew. Chem. Int. Ed. Engl. 1995, 34, 1143. (f) Bochmann, M.; J. Chem. Soc. Dalton Trans
1996, 255. (j) Kaminsky, W.; Arndt, M. Adv. Polym. Sci. 1997, 127, 144.
21. Sinn, H.; Kaminsky, W.; Vollmer, H.; Woldt, R. Angew. Chem. Int. Ed. Engl. 1980, 19, 390.
22. (a) Hlatky, G. G. Chem. Rev. 2000, 100, 1347. (b) Severn, J. R.; Chadwick, J. C.; Duchateau,
15
Chapter 1
R.; Friederichs, N. Chem. Rev. 2005, 105, 4073.
23. Reichert, K. H.; Meyer, K. R.; Makromol. Chem. 1973, 169, 163.
24. Long, W. P; Breslow, D. S. Liebigs Ann. Chem. 1975, 463.
25. McKnight, A. L.; Waymouth, R. Chem. Rev. 1998, 98, 2587.
26. Johnson, L. K.; Killian, C. M.; Brookhart, M. J. Am. Chem. Soc. 1995, 117, 6414.
27. Killian, C. M.; Tempel, D. J.; Johnson, L. K.; Brookhart, M. J. Am. Chem. Soc. 1996, 118,
11664.
28. Ittel, S. D.; Johnson, L. K.; Brookhart, M. Chem. Rev. 2000, 100, 1169.
29. Britovsek, G. J. P.; Gibson, V. C.; Spitzmesser, S. K.; Tellmann, K. P.; White, A. J. P.;
Williams, D. J. J. Chem. Soc. Dalton Trans. 2002, 1159.
30. Small, B. L.; Brookhart, M.; Bennett, A. M. A. J. Am. Chem. Soc. 1998, 120, 4049.
31. Mitani, M.; Mohri, J.; Yoshida, Y.; Saito, J.; Ishii, S.; Tsuru, K.; Matsui, S.; Furuyama, R.;
Nakano, T.; Tanaka, H.; Kojoh, S.; Matsugi, T.; Kashiwa, N.; Fujita, T. J. Am. Chem. Soc.
2002, 124, 3327.
32. Matsui, S.; Mitani, M.; Saito, J.; Tohi, Y.; Makio, H.; Matsukawa, N.; Takagi, Y.; Tsuru, K.;
Nitabaru, M.; Nakano, T.; Tanaka, H.; Kashiwa, N.; Fujita, T. J. Am. Chem. Soc. 2001, 123,
6847.
33. Pullukat, T. J.; Hoff, R. E. Catal. Rev.-Sci. Eng. 1999, 41, 389.
34. Sauer, J.; Zahradnic, R.; Int. J. Quantum. Chem. 1984, 26, 793.
35. Kashiwa, N. J. Polym. Sci.: Part A: Polym. Chem. 2004, 42, 1.
36. Barbè, P. C.; Cecchin, G.; Noristi, L. Adv. Polym. Sci. 1987, 81, 1.
37. Doi, Y.; Soga, K.; Murata, M.; Suzuki, E.; Ono, Y.; Keii, T. Polym. Commun. 1983, 24, 244.
38. Farina, M.; Puppi, C. J Mol Catal, 1993, 82, 3.
39. Marks, T. J. Acc. Chem. Res. 1992, 25, 57.
40. Sensarma, S.; Sivaram, S. Polym. Int. 2002, 51, 417.
41. Soga, K.; Kaminaka, M. Macromol. Chem. Phys. 1994, 195, 1369 .
42. Sacchetti, M.; Pasquali, S.; Govoni, G. US Patent 5698487; Chem. Abstr. 1996, 124, 177225.
43. Guan, Z.; Zheng, Y.; Jiao, S. J. Mol. Catal. A: Chem. 2002, 188, 123.
44. (a) Severn, J. R.; Chadwick, J. C. Macromol. Rapid Commun. 2004, 25, 1024. (b) Severn, J.
R.; Chadwick, J. C. Macromol. Chem. Phys. 2004, 205, 1987. (c) Severn, J. R.; Duchateau,
R.; Chadwick, J. C. Polym. Int. 2005, 54, 837. (d) Severn, J. R.; Kukalyekar, N.; Rastogi, S.;
Chadwick, J. C. Macromol. Rapid Commun. 2005, 26, 150. (e) Severn, J. R.; Chadwick, J.
C.; Van Axel Castelli, V. Macromolecules 2004, 37, 6258.
45. (a) Huang, R.; Liu, D.; Wang, S.; Mao, B. Macromol. Chem. Phys. 2004, 205, 966. (b)
Huang, R.; Liu, D.; Wang, S.; Mao, B. J. Molec. Catal. A: Chem. 2005, 233, 9197. (c) Xu,
R.; Liu, D.; Wang, S.; Mao, B. Macromol. Chem. Phys. 2006, 207, 779.
46. (a) Nakayama, Y.; Bando, H.; Sonobe, Y.; Kaneko, H.; Kashiwa N.; Fujita, T. J. Catal. 2003,
215, 171. (b) Nakayama, Y.; Bando, H.; Sonobe, Y.; Suzuki, Y.; Fujita, T. Chem. Lett.
(Japan) 2003, 32, 766. (c) Nakayama, Y.; Bando, H.; Sonobe, Y.; Fujita, T. Bull. Chem. Soc.
Jpn. 2004, 77, 617. (d) Nakayama, Y.; Bando, H.; Sonobe, Y.; Fujita, T. J. Mol. Catal. A:
16
Introduction and scope
Chem. 2004, 213, 141.
47. McKenna, T.; Soares, J. B. P. Chem. Eng. Sci. 2001, 56, 3931.
48. Zakharov, V. A.; Bukatov, G. D.; Barabanov, A. A. Macromol. Symp. 2004, 213, 19.
49. Galli, P.; Haylock, J. C. Macromol Symp. 1992, 63, 19.
50. Galli, P. J. Macromol. Sci., Pure. Appl. Chem. 1999, A36, 1561.
51. (a) Capaccio, G.; Ward, I. M. Polymer 1974, 15, 233. (b) Capaccio, G.; Ward, I. M. Polymer
1974, 15, 233; Polym. Eng. Sci. 1975, 15, 219. (c) Capaccio, G.; Ward, I. M. Polymer 1975,
16, 239. (d) Capaccio, G.; Crompton, T. A.; Ward, I. M. J. Polym. Sci. Polym. Phys. 1976,
14, 1641. (e) Capaccio, G.; Crompton, T. A.; Ward, I. M. Polymer 1976, 17, 645. (f)
Capaccio, G.; Crompton, T. A.; Ward, I. M. J. Polym. Sci. Polym. Phys. 1980, 18, 301.
52. (a) Keller, A.; Odell, J. A. J. Polym. Sci. Polym. Symp. 1978, 63, 155.
53. (a) Pennings, A. J.; Lageveen, R.; Devries, E. S. Colloid. Polym. Sci. 1977, 255, 532. (b)
Zwijneburg, A.; Pennings, A. J. Colloid. Polym. Sci. 1975, 253, 452.(c) Zwijneburg, A.;
Pennings, A. J. Colloid. Polym. Sci. 1976, 254, 868.
54. Wunderlich, B. 1976, Macromolecular Physics. Academic Press: San Diego; Vols. 1-3
55. Mandelkern, L. 1964, Crystallization of Polymers. McGraw-Hill: New York.
56. Keller, A. Phil. Mag. 1957, 2, 1171.
57. Fischer, E. W. Z. Naturf. 1957, 12a, 753.
58. Till, P. H. Jr. J. Poly. Sci. 1957, 24, 301.
59. Bunn, C. W.; Alcock, T. C. Trans. Faraday Soc. 1945, 41, 317.
60. Abo el Maaty, M. I.; Hosier, I. L.; Bassett, D. C. Macromolecules, 1998, 31, 153.
61. (a) Keller, A; Kolnaar, H.W.H. In: Meijer, editor. Processing of Polymers, Vol 18, Chapter 4.
(b) Keller, A.; Hikosaka, M.; Rastogi, S.; Toda, A.; Barham, P. J.; Goldbeckwood, G. J. Mat.
Sci, 1994, 29, 2579. (c) Mackley, M. R.; Keller, A. Polymer, 1973, 14, 16. (d) Pope, P.;
Keller, A. Colloid. Polym. Sci. 1978, 256, 751.
62. Varga, J. J. Mat. Sci. 1992, 27, 2557.
63. Pennings, J. A.; Kiel, A. M. Kolloid-Z. Z. Polymere 1965, 205, 160.
64. Mackley, M. R.; Keller, A. Phil. Trans. Roy. Soc. 1975, 278, 29.
65. Somani, R. H.; Yang, L.; Zhu, L.; Hsiao, B. Polymer, 2005, 46, 8587.
66. Kumaraswamy, G . J Macromol. Sci. Part C: Polym. Rev., 2005, 45, 375.
67. (a) Bastiaansen, C. W. M.; Meyer, H. E. H.; Lemstra, P. J. Polymer, 1990, 31, 1435. (b)
Bastiaansen, C. W. M. Ph.D. Thesis 1990, Eindhoven University of Technology.
17
Chapter 1
18
Chapter 2
Immobilization of catalysts on MgCl2/AlRn(OR’)3-n
supports. Confirmation and refutation of single centre
nature of immobilized catalysts using melt rheometry
Abstract
Several catalysts have been immobilized on MgCl2/AlRn(OEt)3-n supports to produce
polyethylenes differing in molecular weight and molecular weight distribution. Investigation of
the melt rheological properties of polyethylenes prepared with chromium, titanium, vanadium
and zirconium catalysts immobilized on MgCl2 supports has confirmed narrow (Schulz-Flory)
molecular weight distribution, indicative of single-centre catalysis, for Cr-, V- and Zr-based
systems, but not for Ti. In the case of polymers prepared with MgCl2-immobilized Ti complexes
and having narrow MWD (Mw/Mn 2-3) according to GPC, deviation from a Schulz-Flory
distribution was evident from a decrease in storage modulus with decreasing angular frequencies,
whereas polyethylenes prepared with analogous vanadium complexes exhibited a constant
(plateau) modulus over a wide frequency range. The presence of a plateau modulus is
characteristic for narrow MWD polyethylene, so that melt rheology provides a valuable tool to
prove, or disprove, the presence of a Schulz-Flory distribution in cases where GPC does not
provide a definitive answer.
2.1 Introduction
Most of the commercial methods of polyolefin production are based on heterogeneous catalysts.
As described in Chapter 1, catalysts are immobilized on an inert carrier before subjecting to
polymerization in order to avoid reactor fouling and to achieve high catalyst activity, high bulk
density and spherical morphology of the polymer. Silica and MgCl2 are the most widely used
inert materials for immobilizing catalysts in the production of polyethylenes. Silica is the most
explored support for heterogenization of single centre alpha-olefin polymerization catalysts,
while MgCl2 is used as a support for Ziegler-Natta catalysts in commercial processes for the
production of polyolefins.1
19
Chapter 2
In recent years considerable efforts have been made to develop effective approaches for the
immobilization of single centre catalysts on suitable supports, the use of heterogeneous or
immobilized catalysts being a prerequisite for application of such catalysts in slurry or gas phase
processes for polyolefins. Although silica is the most commonly used support material, there has
been increasing interest in the use of magnesium chloride, in view of the established use of
MgCl2 as a support for Ziegler-Natta catalysts. Particular attention has been paid to the supports
prepared by reaction of an aluminium alkyl with solid or hydrocarbon-soluble adducts of
magnesium chloride and an alcohol.2,3
Controlled particle size and porosity can be obtained by partial dealcoholation of MgCl2·3EtOH,
prepared by cooling an emulsion of a molten MgCl2/ethanol in paraffin oil.2 Supports of type
MgCl2/AlRn(OEt)3-n, obtained by the reaction of AlR3 with MgCl2·nEtOH, have been
successfully used for the immobilization of a range of early- and late-transition metal catalysts
for ethylene polymerization.4,5 The molecular weight distributions of the polyethylenes obtained
have ranged from narrow (Mw/Mn = 1.8 – 2.2), in the case of immobilized chromium and
vanadium catalysts, to broad (Mw/Mn = 4-12) in the case of bis(imino)pyridyl iron catalysts.
In this chapter, the immobilization of several single-centre and multi-centre catalysts on
MgCl2/AlRn(OEt)3-n type of supports and their MAO-free activation for the production of
polyethylene will be discussed, along with the use of melt rheometry to characterize the
molecular weight distributions of the polymers obtained.
2.1.1 Characterization of narrowly dispersed PE by GPC and reliability of GPC
Determination of the polydispersity of a polyolefin is routinely carried out using high
temperature gel permeation chromatography (HT-GPC). A relatively broad molecular weight
distribution is obtained when the catalyst contains a range of different active species, such as is
generally the case with Ziegler-Natta catalysts, whereas the presence of a single type of
catalytically active species should, when chain transfer reactions take place, lead to a SchulzFlory distribution with Mw/Mn = 2.0 and Mz/Mw =1.5. In principle, a homogeneous catalyst
generating uniform active species should give a polymer with a narrow (Schulz-Flory)
distribution and this is very often the case. The problem arises when the molecular weight
distribution obtained with a catalyst presumed to be a single centre deviates from the SchulzFlory distribution. For example, an Mw/Mn value between 2 and 3 still indicates a narrow
molecular weight distribution, but the deviation from the value of 2.0 may be due to presence of
more than two active species, or to inaccuracies in the GPC measurement. It is generally
recognized that molecular weight and polydispersity measurement by HT-GPC is not a simple
task, particularly for high molecular weight polymer, and for this reason Mw/Mn values in the
range 2-3 are in practice often accepted as being indicative of single centre catalysis.
20
Confirmation and refutation of single centre nature
In addition to GPC, molecular weight distribution can be determined by rheometry. Rheometry
has advantages over chromatographic techniques, being cheaper, faster and applicable for
polymers which are insoluble in common solvents. However, deducing the information about
molecular weight distribution from rheological data involves complicated mathematical models.
Several rheologists have proposed mathematical models which can relate molecular weight
distribution to steady shear viscosity. Most of the models are based on the tube and reptation
theory of de Gennes.6 However, incorporation of polydispersity in this model is difficult because
the relaxation time of the individual chain has to be addressed. In the models based on double
reptation mixing rules, the relaxation modulus G(t) is expressed in terms of molecular weight
distribution. This model is successful in predicting molecular weight distribution of polydisperse
polymers (Des Cloizeaux model).7 Mead inverted the double reptation model to generate
molecular weight distribution.8 Guzman et al. proposed a regularization-free method to calculate
molecular weight distribution from dynamic moduli data.9 This is a parameter-free method and
also quantifies the uncertainty in rheological molecular weight distribution determination. Bailly
et al. have done a considerable amount of simulation work on predicting the viscoelastic moduli
of polydisperse entangled melts.10 Although many models are successful in determining polymer
molecular weight distribution, high molecular weight limits the validation of these models by
chromatographic techniques, which are difficult to use for the analysis of high molecular weight
polymers. A relatively simple method, in which oscillatory shear behaviour is measured as a
function of angular frequency, has been described by Zeichner and Patel.10 In this method, the
storage modulus, G', and the loss modulus, G'', are determined. At low frequencies G'' > G',
whereas at high frequencies G' > G''. The intersection of the modulus vs. frequency curves gives
the crossover modulus, Gc, which is a measure of the polydispersity of the polymer. The
polydispersity index (PI) is derived from the reciprocal of Gc. This method is effective for
polyolefins having relatively broad MWD, such as polypropylene produced with Ziegler-Natta
catalysts,12 but is less suitable for the high molecular weight, narrow MWD polyethylenes under
investigation in the present work.
The main discussion point in the present chapter will be the use of melt rheometry to distinguish
between single-centre and multi-centre catalysts. At the end of this chapter we will be able to
confirm or refute the single-centre nature of the immobilized catalysts used in this work.
2.2 Materials and Methods
2.2.1 Materials
All manipulations were performed under an argon atmosphere using glove box (Braun MB-150
G1 or LM-130) and Schlenk techniques. Light petroleum (b.p. 40-60°C) and dichloromethane
were passed over a column containing Al2O3 and stored over 4Å molecular sieves. All the
solvents were freeze-thaw degassed at least twice prior to use.
21
Chapter 2
AlEt3 (1.3 M solution in heptane) was purchased from ACROS. AliBu3 (1 M solution in hexane)
was purchased from Akzo Nobel.
Ethylene (3.5 grade supplied by Air Liquide) was purified by passing over columns of 4Å
Molecular Sieves and BTS copper catalyst.
2.2.2 Catalysts
(η1:η5-Me2NCH2CH2C5Me4)CrCl2 (1) was kindly donated by SABIC Euro Petrochemicals. {2,6[ArN=C(Me)]2C5H3N}VCl3 (2) and {2,6-[ArN=C(Me)]2C5H3N}FeCl2 (3) (Ar=2,6-diisopropyl)
were prepared as reported previously.13 1-(8-quinolyl)indenyl CrCl2 (4) was received from
Basell polyolefins.14 Immobilization of the vanadium(III) and titanium(III) amidinate complexes
(5-8) on MgCl2/AlEtn(OEt)3-n supports and their activation has been reported previously by our
group.4e Pincer and phebox titanium and vanadium complexes (9-12) were prepared by Dr. A.V.
Chuchuryukin at the University of Utrecht.
iPr
N
N
iPr
N
V
Cl
Cl
Cl
iPr
Pri
(1)
iPr
(2)
N
iPr
N
N
CrCl2
Fe
Cl
iPr
N
Cl
Pri
(3)
(4)
Scheme 2.1. Structures of catalysts used for immobilization on MgCl2.
2.2.3 Support Preparation and Catalyst Immobilization
Support preparation was performed by the addition of AlEt3 to a slurry of an adduct MgCl2 ·
1.1EtOH (average particle size d50 82µm) in light petroleum (AlEt3/EtOH = 2) at 0°C, after which
the mixture was kept at room temperature for 2 days with occasional agitation. The resultant
support was washed with light petroleum three times and dried under argon flow and
subsequently under vacuum until free flowing. The Al content of the support was determined by
the H. Kolbe Microanalytisches Laboratorium, Mülheim an der Ruhr, Germany. The ethoxide
content in MgCl2/AlEtn(OEt)3-n support was determined by gas chromatography (GC) analysis of
22
Confirmation and refutation of single centre nature
the ethanol content of a solution obtained by dissolving 50 mg of support in 2.5 mL of BuOH
containing a known quantity of PrOH as an internal standard.
Catalyst immobilization was effected by mixing the support (100 mg) with precatalyst solutions
in solvent (2.0 mL toluene or dichloromethane, containing typically 1.0 µmol of precatalyst) and
keeping at room temperature overnight. The choice of solvent was based on the solubility of the
catalyst in a particular solvent. In some cases, for quick immobilization of catalyst, the vial
containing the catalyst solution and the support was kept in a water bath at 50 °C for 2 hours.
The slurry of the immobilized catalyst was diluted with light petroleum and used directly in
ethylene polymerization.
2.2.4 Polymerization procedure
Polymerization was carried out in a l L Premex autoclave by charging the immobilized catalyst
(100 mg, containing 1.0 µmol precatalyst), slurried in approx. 100 mL light petroleum, to 400
mL light petroleum containing the desired amount of cocatalyst, at 50 °C and an ethylene
pressure of 5 bar.
Fig 2.1 PREMEX polymerization reactor.
After catalyst injection, polymerization was continued at constant pressure for 1 h and with a
stirring rate of around 1000 rpm. After venting the reactor, 20 mL of acidified ethanol were
added and stirring was continued for 30 min. The polymer was recovered by filtration, washed
with water and ethanol and dried in vacuum oven overnight at 70 °C.
2.2.5 Sample Preparation for Rheometry
The polymer was stabilized with 0.2 wt % Irganox 1010 (added as a solution in acetone, after
which the polymer was dried in vacuum oven overnight at 60 °C). Discs of 8 mm diameter and
23
Chapter 2
25 mm diameter having a thickness of 0.5 mm were compression moulded at 180 °C and 200 bar
pressure.
2.2.6 Polymer Characterization
Molecular weight and molecular weight distributions of the resulting polymers were determined
by means of size exclusion chromatography on a PL-GPC210 at 135 °C using 1,2,4trichlorobenzene as solvent. Oscillatory shear measurements were performed on an ARES
rheometer from Rheometric Scientific. Frequency sweep measurements were carried out at 180
°C.
Differential scanning calorimetry (DSC) was carried out with a QA100 differential scanning
calorimeter (TA Instruments). The samples (1-5-2.5 mg) were heated to 160 °C at a rate of 10
°C/min for the determination of the first melting temperature (Tm1), then cooled down at the same
rate to 20°C. A second heating cycle at 10 °C/min was then applied for the determination of the
melting temperature Tm2.
2.3 Results and Discussion
Within the Polymer Chemistry department of Eindhoven University of Technology, extensive
studies have been made of the immobilization of a wide range of early- and late-transition metal
ethylene polymerization catalysts on MgCl2/AlRn(OEt)3-n supports. In earlier reports, among the
catalysts giving narrow molecular weight distribution polyethylene (Mw/Mn 1.8-2.1, determined
by high temperature GPC) were (η1:η5-Me2NCH2CH2C5Me4)CrCl2 (1) and {2,6[ArN=C(Me)]2C5H3N}VCl3, where Ar = 2,6-diisopropylphenyl (2).4d,13 In contrast, the
immobilization of TiCl4 and {2,6-[ArN=C(Me)]2C5H3N}FeCl2 (3) resulted in broader MWD
(Mw/Mn 4.1-5.6). The structures of catalysts 1-3 are shown in Scheme 2.1. In addition to GPC,
melt rheometry was used to characterize the molecular weight distributions of these polymers
and it was demonstrated that the viscoelastic response of linear polyethylene over a wide range
of angular frequency could be used to confirm the differences in MWD indicated by GPC,
thereby differentiating between genuinely single-centre and multi-centre catalysts in these
systems. The earlier findings are summarized in Figure 2.2, in which the polyethylene
viscoelastic response is plotted as storage modulus (G') versus angular frequency (ω).
24
Confirmation and refutation of single centre nature
6
10
5
G' (Pa)
10
10
-3
10
-2
10
-1
10
0
10
1
10
2
Frequency (rad/s)
Figure 2.2. Linear viscoelastic response in oscillatory shear of polyethylene synthesized using
catalysts 1 (■), 2 (○), 3 (▲) and TiCl4 (●).
Frequency sweep experiments were carried out on polyethylene discs over the frequency range
in the viscoelastic regime, which is determined by performing strain tests. All samples show
plateau modulus at higher frequency confirming high molecular weight of the polymers. The
characteristic plateau behaviour over the whole frequency range,15 indicative for samples
comprising long chains with similar viscoelastic response and relaxation times (i.e. narrow
MWD), is clearly apparent for the polymers synthesized using the chromium and vanadium
catalysts 1 and 2, but not for the polymers synthesized with TiCl4 or the iron catalyst 3. For the
latter polymers, the steady decrease in the storage modulus with decreasing frequency is a result
of the broader molecular weight distributions, the broadening in the chain length distribution
resulting in a broad relaxation time spectrum.
2.3.1 Immobilization and activation of [1-(8-quinolyl)indenyl]CrCl2
The confirmation of retention of the single-centre characteristics of the chromium catalyst 1 after
immobilization and activation on a MgCl2-based support triggered an investigation of
immobilization and activation of another chromium complex, [1-(8-quinolyl)indenyl]CrCl2 (4),
the structure of which is also shown in Scheme 2.1. This type of complex, containing a rigid
spacer group between the cyclopentadienyl ring and the nitrogen atom, has been reported to give
improved stability in homogeneous polymerization.16 Immobilization of this catalyst on
MgCl2/AlEtn(OEt)3-n supports prepared by the reaction of an adduct MgCl2·1.1EtOH with excess
AlEt3 was carried out simply by contacting 100 mg of support with 2 mL of dichloromethane
containing 1 µmol 4 at ambient temperature. The results of ethylene polymerizations carried out
with either AlEt3 or AliBu3 as cocatalyst and at temperatures in the range 50-70 °C are listed in
Table 2.1.
25
Chapter 2
Table 2.1. Ethylene polymerization using [1-(8-quinolyl)indenyl]CrCl2 (4) on MgCl2-based
supportsa
Entry
Activity
Mw
Mn
Mw/Mn
Cocatalyst Temp.
(°C) (kg/mol Cr·bar·h) (kg/mol) (kg/mol)
Tm1d
(°C)
Tm2e
(°C)
1b
AlEt3
50
4800
1631
544
3.0
143.6
133.9
2b
AliBu3
50
2000
1300
820
1.6
142.4
134.0
3c
AliBu3
50
740
1324
824
1.6
142.7
133.2
4c
AliBu3
60
900
750
338
2.2
142.5
131.6
5c
AliBu3
70
1600
850
403
2.1
143.2
134.0
a
Polymerization conditions: 500 mL of light petroleum, immobilized catalyst 100 mg, AlEt3 or
AliBu3 1 mmol, ethylene pressure 5 bar, time 1 h.
b
Support composition: MgCl2·0.11AlEt1.72(OEt)1.28.
c
Support composition: MgCl2·0.13AlEt1.99(OEt)1.01.
d
Melting temperature from first heating cycle of nascent polymer, as obtained from the reactor
e
Melting temperature from second heating cycle.
It is apparent that the choice of cocatalyst has a considerable effect on catalyst activity. AlEt3
gave higher activity than AliBu3, but also resulted in a somewhat broader molecular weight
distribution. Homogeneous, MAO-activated polymerizations with 4 and related complexes at
ambient temperature have been reported to give polydispersities between 2 and 4, and molecular
weights in the range 100 – 500 kg/mol.14 The data in Table 2.1 reveal that significantly higher
weight average molecular weights are obtained after immobilization of the catalyst on a
magnesium chloride support, as has been observed for numerous other catalysts.4 The lower
molecular weights obtained in homogeneous polymerization can at least partly be attributed to
chain transfer with AlMe3 present in MAO. The results in Table 2.1 also show that different
batches of support, varying in chemical composition, gave different catalyst activities, and that
high molecular weight polyethylene was obtained with both AlEt3 and AliBu3. The activities
obtained here, and indeed also the activities obtained4d with complex 1 on MgCl2/AlRn(OEt)3-n
supports, are higher than those of AlR3-activated Cr(III) complexes recently claimed17 to be
highly active in comparison with previously reported half-metallocene chromium complexes.
The results of frequency sweep measurements carried on the polymers prepared under similar
conditions but with different cocatalysts (Table 2.1; entries 1 and 2) are shown in Figure 2.3A.
26
Confirmation and refutation of single centre nature
60
50
6
10
50
6
30
20
10
5
10
0
10
-2
10
-1
10
0
10
1
10
Phase Angle (°)
20
40
G' (Pa)
30
10
Phase Angle (°)
G' (Pa)
40
5
0
10
2
10
10
Frequency (rad/s)
-2
10
-1
10
0
10
1
2
10
Frequency (rad/s)
A
B
Fig. 2.3. Linear viscoelastic responses in oscillatory shear of polyethylene: (A) synthesized using
catalyst 4 with AliBu3 (■) or AlEt3 (●) as cocatalyst; (B) synthesized using catalyst 4 with
AliBu3 as cocatalyst at 50 °C (■), 60 °C (●) and 70 °C (▲). Respective open symbols represent
phase angle.
It is evident that the storage modulus of the polymer prepared at 50 °C with AliBu3 as cocatalyst
remained constant over the entire frequency range, confirming the narrow molecular weight
distribution, while the drop in G' at lower frequencies for the AlEt3-cocatalyzed sample indicates
a broader MWD. This is in line with the GPC results, but it should also be taken into account that
the melt rheological properties of polyethylene can be strongly influenced by long-chain
branching, even at low levels of branching.18 Catalysts known to give polyethylenes with
significant amounts of long-chain branching include constrained geometry catalysts19 and certain
zirconocenes.20 Long-chain branching can occur when a vinyl-terminated macromonomer is
incorporated into a growing polyethylene chain. C-H bond activation of the polyethylene
backbone has also been proposed as a mechanism for the formation of long-chain branching.21
The presence of long-chain branching can result in the disappearance of the characteristic plateau
modulus for narrow MWD polyethylene at lower frequencies.20d For a more complete
rheological characterization, we therefore also investigated the frequency dependence of the
phase angle (G''/G'). Generally, low phase angle values suggest good elastic recovery, whereas
higher values are indicative of viscous behaviour. Figure 2.3A shows higher phase angle values
for the AlEt3-cocatalyzed sample than for the polymer prepared using AliBu3, indicating a more
viscous nature due to a more prominent loss modulus. There is no evidence for any decrease in
the phase angle at low frequency, which would be indicative of elastic behaviour arising from
long-chain branching. Long-chain branching would moreover be expected to give rise to
physical network formation and a second, low-frequency plateau for G'; this is also not observed.
The DSC data in Table 2.1 show high melting temperatures (Tm1 = 142-144 °C) for the nascent
polymers and normal melting temperatures (Tm2 = 131-134 °C) for melt crystallized samples.
These observations are typical for linear polyethylene. The possibility of long-chain branching
can not be totally ruled out, but the degree of branching, if at all present, is too low to have an
27
Chapter 2
effect on the crystallization behaviour of these polymers. The GPC data concerning the effect of
polymerization temperature on MWD (Table 2.1; entries 3-5) are less definitive, the value of
Mw/Mn increasing from 1.6 at 50 °C to 2.1-2.2 at higher temperatures. Melt rheometry on these
samples, however, revealed (Figure 2.3B) a slight deviation of the storage modulus from plateau
behaviour in the case of the polymers synthesized at 60 and 70 °C, confirming a broadening in
the molecular weight distribution. The phase angle profiles show an increase in viscous
behaviour at low frequencies, indicating the presence of some low molecular weight chains in
the polymers synthesized at the higher temperatures.
2.3.2 Titanium and vanadium amidinate complexes.
In an earlier publication from our group, the immobilization and activation of vanadium(III)
amidinate catalysts on a MgCl2/AlEtn(OEt)3-n support has been reported.4e Polyethylene with
moderate molecular weight (Mw ~ 700 kg/mol) and narrow molecular weight distribution (Mw/Mn
= 2.0) was obtained. Analogous titanium(III) amidinates gave Mw ~ 600 kg/mol and Mw/Mn =
2.3. It was concluded, on the basis of these narrow molecular weight distributions, that the
single-centre behaviour of these catalysts was retained after immobilization on the support.
However, the increase in the value of Mw/Mn on going from vanadium to titanium, albeit slight,
raises some doubt concerning the single-centre nature of the titanium-based systems.
Si
Si
Si
thf
N
thf
Cl
N
V
N
Cl
N
N
Cl
Ti
N
Si
thf
V
N
Cl
N
Si
(5)
(6)
Scheme 2.2. Catalyst structures
(7)
thf
Cl
thf
Cl
N
Cl
Ti
N
Cl
thf
Si
(8)
In order to resolve this question, the melt rheological properties of these polymers have now
been investigated. The structures of the amidinate catalysts, first prepared by Hessen and
coworkers,22 are shown in Scheme 2.2. Polymerization was carried out at 50 °C, using AliBu3 as
cocatalyst.4e The results of frequency sweep tests carried out on the polymers prepared with
[Me3SiNC(Ph)NCH2CH2NMe2]MCl2(THF), where M = V (5) or Ti (6), are shown in Figure 2.4.
A very slight decrease in storage modulus with decreasing frequency is apparent for the
vanadium-catalyzed polymer, whereas a much steeper decrease in G' is observed for the polymer
prepared with the titanium complex 6. Similar differences between vanadium and titanium are
evident from Figure 2.5, which shows the results obtained with polymers prepared with
[PhC(NSiMe3)2]MCl2(THF)2, where M = V (7) or Ti (8). The phase angle behaviour also
confirms the less viscous nature of the vanadium-catalyzed polymers. These results provided a
28
Confirmation and refutation of single centre nature
first indication of a much stronger deviation from single-centre behaviour in the case of MgCl2immobilized titanium catalysts than for vanadium catalysts with the same ligand structure and
triggered the investigation and comparison of other titanium- and vanadium-based systems.
60
60
50
50
10
20
G' (Pa)
30
40
Phase Angle (°)
G' (Pa)
40
6
30
20
10
10
5
10
0
10
-2
10
-1
10
10
0
10
1
10
Phase Angle (°)
6
10
2
5
0
10
-2
10
-1
0
10
10
1
10
2
Frequency (rad/s)
Frequency (rad/s)
Fig. 2.4
Fig. 2.5
Figure 2.4. Linear viscoelastic response in oscillatory shear of polyethylene synthesized using
catalysts 5 (■) and 6 (●).
Figure 2.5. Linear viscoelastic response in oscillatory shear of polyethylene synthesized using
catalysts 7 (■) and 8 (●).
2.3.3 Titanium and vanadium pincer and phebox complexes.
In recent years, many advances have been made in the synthesis and development of nonmetallocene single-centre catalysts for olefin polymerization.23,24 Much attention has been given
to NNN-pincer type complexes, as a result of both their catalytic properties and ease of
synthesis. Much less work has been carried out on NCN-pincer complexes containing a σ-aryl
donor ligand.25 Recently, a series of pincer complexes, including the vanadium and titanium
complexes 9 and 10, has been synthesized and tested in ethylene polymerization. Full details of
the synthesis and crystal structures of these complexes will be part of a future publication.26 It
has been found that, whereas these and related complexes25b give low activity and poor stability
under homogeneous polymerization conditions, using MAO as activator, very high activities can
be obtained after immobilization on magnesium chloride. The results of HT-GPC
characterization of polymers synthesized with complexes 9 and 10, immobilized on
MgCl2/AlEtn(OEt)3-n supports, are given in Table 2.2. The titanium pincer catalyst 10 gave very
high molecular weight polyethylene, but the polydispersity values in Table 2.2 indicate a
significant deviation from a Schulz-Flory distribution, particularly for polymerization at 70 °C.
29
Chapter 2
N
N
O
O
VCl2
TiCl3
N
N
(9)
N
N
N
V
N
Ti
Cl
Cl
Cl
O
O
O
Cl
Cl
(10)
(11)
(12)
Scheme 2.3. Catalyst structures
Table 2.2. Ethylene polymerization using vanadium and titanium pincer and phebox
catalysts on MgCl2-based supportsa
a
b
Entry
Catalyst
Temp.
(°C)
Mw
(kg/mol)
Mn
(kg/mol)
Mw/Mn
Tm1a
(°C)
Tm2b
(°C)
1
9
50
926
447
2.1
143.8
134.7
2
9
70
922
462
2.0
142.0
134.0
3
10
50
4500
1500
3.0
141.6
133.9
4
10
70
2200
370
5.9
141.0
134.9
5
11
70
690
360
1.9
141.4
133.5
6
12
50
1406
403
3.5
142.9
134.9
Melting temperature from first heating cycle of nascent polymer, as obtained from the reactor
Melting temperature from second heating cycle.
In contrast, the vanadium pincer catalyst 9 gave narrow distribution irrespective of the
polymerization temperature, indicating retention of single-centre behaviour. This was confirmed
by determination of the melt rheological properties of these polymers, the results of which are
shown in Figures 2.6 and 2.7.
30
Confirmation and refutation of single centre nature
60
60
50
50
20
G' (Pa)
30
40
Phase Angle (°)
G' (Pa)
40
6
30
20
10
5
10
0
10
-2
10
-1
10
0
10
1
10
10
Phase Angle (°)
10
6
10
10
2
5
0
10
Frequency (rad/s)
-2
10
-1
10
0
10
1
10
2
Frequency (rad/s)
Fig. 2.6
Fig. 2.7
Figure 2.6. Linear viscoelastic response in oscillatory shear of polyethylene synthesized at 50 °C
using vanadium (■) and titanium (●) pincer complexes 9 and 10 (Table 2.2; entries 1 and 3).
Figure 2.7. Linear viscoelastic response in oscillatory shear of polyethylene synthesized at 70 °C
using vanadium (■) and titanium (●) pincer complexes 9 and 10 (Table 2.2; entries 2 and 4).
The vanadium-catalyzed polymers show almost constant values of G' over the entire frequency
range, although there was a slight deviation for the polymer synthesized at 70 °C, whereas the
polymers prepared with the analogous titanium complex exhibit strong decreases in G' with
decreasing frequency.
A further comparison between vanadium- and titanium-based systems was made with complexes
11 and 12 (Scheme 2.3) containing the 2,6-bis(2'-oxazolinyl)phenyl (phebox) ligand. These
complexes, like the pincer complexes 9 and 10, were synthesized by Chuchuryukin et al.26 via
transmetallation reactions involving arylgold(1) phosphine precursors.27 The GPC data in Table
2.2 (entries 5 and 6) and the melt rheological behavior (Figure 2.8) of polymers prepared with
the immobilized complexes reveal, once more, a fundamental difference between vanadium and
titanium, narrow MWD being obtained with the vanadium complex 11, but not with the related
titanium complex 12.
In Figure 2.8 it is apparent that the plateau modulus of the vanadium-catalyzed polymer is lower
than the characteristic value of 2 MPa. It is well known that this lowering in plateau modulus can
be caused by shrinkage in the sample, which occurs when the stresses in the sample are released
in the molten state. In a measurement on a parallel plate rheometer in dynamic mode, the stress
constant is related to the third power of the sample radius. Shrinkage of the sample can therefore
have a strong effect on the determination of the true plateau modulus.
31
Chapter 2
60
50
6
G' (Pa)
40
30
20
Phase Angle (°)
10
10
10
5
0
10
-2
-1
10
0
10
10
1
2
10
Frequency (rad/s)
Figure 2.8. Linear viscoelastic response in oscillatory shear of polyethylene synthesized using
vanadium (■) and titanium (●) phebox complexes 11 and 12 (Table 2.2; entries 5 and 6).
2.3.4 Titanocene and zirconocene-based systems.
In earlier studies on the immobilization of single-centre catalysts on MgCl2/AlEtn(OEt)3-n
supports, the metallocenes Cp2TiCl2 and Cp2ZrCl2, as well as CpTiCl3 and related halftitanocene complexes, have been investigated.4a,b GPC analysis of the resulting polyethylenes
indicated narrow polydispersities, with Mw/Mn values mostly in the range 2 – 3, and it was
concluded that this was an indication that the single-centre characteristics of these catalysts were
retained after immobilization on the support. However, the deviation from single-centre
behaviour exhibited by the titanium amidinate, pincer and phebox complexes described above,
apparent from the polymer melt rheological properties, prompted us to carry out similar
frequency sweep experiments on titanocene- and zirconcene-based systems to ascertain which of
these may genuinely be classified as single-centre. A polyethylene prepared4b with Cp2TiCl2 had
an Mw of 1,633 kg/mol and Mw/Mn = 2.9, while Cp2ZrCl2 gave4a an Mw of 896 kg/mol and
Mw/Mn = 2.7. Figure 2.9 shows the frequency dependence of G' for polymers prepared with these
complexes.
60
50
6
G' (Pa)
40
30
20
Phase Angle (°)
10
10
10
5
0
10
-2
10
-1
10
0
10
1
10
2
Frequency (rad/s)
Figure 2.9. Linear viscoelastic response in oscillatory shear of polyethylene synthesized using
Cp2ZrCl2 (■) and Cp2TiCl2 (●).
32
Confirmation and refutation of single centre nature
It is evident that the single-centre behaviour of the zirconocene is retained after immobilization
on a MgCl2/AlEtn(OEt)3-n support, whereas this is not so for the titanocene. The melt rheological
properties of the polymers prepared4b with the MgCl2-immobilized half-titanocene complexes are
illustrated in Figure 2.10, which in each case clearly shows significant decreases in G' with
decreasing frequency. In other words, these results deviate from the rheological behaviour
expected for polymers having narrow (Schulz-Flory) MWD. The narrow molecular weight
distributions (Mw/Mn 2.0 – 2.5) determined4b using GPC can therefore not be taken as reliable
proof of single-centre catalysis in these systems.
6
G' (Pa)
10
10
5
10
-3
-2
10
10
-1
10
0
10
1
2
10
Frequency (rad/s)
Figure 2.10. Linear viscoelastic response in oscillatory shear of polyethylene synthesized using
(PhCMe2Cp)TiCl3 (o),(Indenyl)TiCl3 (■),(tBuCp)TiCl3 (▲) and (tBu3P=N)CpTiCl2 ().
The results obtained with the MgCl2-supported titanium catalysts, in contrast to those of
analogous vanadium and zirconium complexes, indicate the formation of more than one type of
active species. A physical explanation for deviation from a pure Schulz-Flory distribution, such
as non-uniform monomer concentration within the growing polymer particle, can be discounted,
as this would not depend on a particular transition metal. One possibility for the formation of
multiple active species could involve a disproportionation of CpTiCl3, and analogous halftitanocenes, on the support surface, forming Cp2TiCl2 and TiCl4. Immobilization of TiCl4 itself
on a MgCl2/AlEtn(OEt)3-n support leads to a Ziegler-Natta type catalyst with high activity and
giving relatively broad molecular weight distribution.4b
2.4 Conclusions
Melt rheometry is a valuable diagnostic tool to determine the presence, or absence, of a narrow,
Schulz-Flory molecular weight distribution in polyethylene. A low sensitivity of the storage
modulus, G', to changes in angular frequency gives a plateau modulus characteristic of narrow
MWD polyethylene, whereas a decrease in G' with decreasing frequency is indicative of a
broadening of the molecular weight distribution. The method is particularly useful in cases where
33
Chapter 2
GPC data indicate narrow polydispersity, but where the ratio Mw/Mn differs from the SchulzFlory value of 2.0, taking into account the experimental difficulties associated with GPC analysis
of high molecular weight polymers and the fact that measured Mw/Mn values in the range 2 – 3
are often taken to be close enough to 2 to be indicative of the presence of a single type of active
species in polymerization.
Application of the method to polyethylenes prepared with a range of early-transition metal
complexes on MgCl2-based supports reveals that the retention of single-centre behaviour of the
catalyst after immobilization on the support is dependent on the transition metal. A constant
storage modulus over a wide range of angular frequencies, consistent with a Schulz-Flory MWD,
was observed for polyethylenes prepared with MgCl2-immobilized chromium, vanadium and
zirconium complexes, confirming that in these systems the single-centre characteristics of the
catalyst can be retained after immobilization and activation on the support. In contrast,
significant decreases in G' with decreasing frequency were observed for the polymers prepared
with MgCl2-immobilized titanium complexes, indicating a broadening in molecular weight
distribution and the presence of more than one active species. The previous assumptions of
single-centre catalysis in the case of the Ti-based systems, based on GPC, can therefore be
discounted. The finding that immobilization on magnesium chloride of vanadium and titanium
complexes having similar ligand structures leads to retention of single-centre behaviour in the
case of vanadium but not with titanium indicates that high-activity vanadium-based systems are
particularly interesting candidates for the synthesis of polyethylene homopolymers and
copolymers with narrow composition distribution.
2.5 References
1. (a) Hlatky, G. G. Chem. Rev. 2000, 100, 1347. (b) Severn, J. R.; Chadwick, J. C.; Duchateau,
R.; Friederichs, N. Chem. Rev. 2005, 105, 4073.
2. Sacchetti, M. ; Pasquali, S. ; Govoni,G. U.S.Patent 5,698,487, 1997. Chem Abstr. 1996, 124,
177225.
3. Nakayama, Y.; Bando, H.; Sonobe, Y.; Kaneko, H.; Kashiwa, N.; Fujita, T. J. Catal. 2003,
215, 171. (b) Nakayama, Y.; Bando, H.; Sonobe, Y.; Suzuki, Y.; Fujita, T. Chem. Lett.
(Japan) 2003, 32, 766. (c) Nakayama, Y.; Bando, H.; Sonobe, Y.; Fujita, T. Bull. Chem. Soc.
Jpn. 2004, 77, 617. (d) Nakayama, Y.; Bando, H.; Sonobe, Y.; Fujita, T. J. Mol. Catal. A:
Chem. 2004, 213, 141.
4. (a) Severn, J. R.; Chadwick, J. C. Macromol. Rapid Commun. 2004, 25, 1024. (b) Severn, J.
R.; Chadwick, J. C. Macromol. Chem. Phys. 2004, 205, 1987. (c) Severn, J. R.; Chadwick, J.
C.; Van Axel Castelli, V. Macromolecules 2004, 37, 6258. (d) Severn, J. R.; Kukalyekar, N.;
Rastogi, S.; Chadwick, J. C. Macromol. Rapid Commun. 2005, 26, 150. (e) Severn, J. R.;
Duchateau, R.; Chadwick, J. C. Polym. Int. 2005, 54, 837.
34
Confirmation and refutation of single centre nature
5. (a) Huang, R.; Liu, D.; Wang, S.; Mao, B. Macromol. Chem. Phys. 2004, 205, 966. (b)
Huang, R.; Liu, D.; Wang, S.; Mao, B. J. Mol. Catal. A: Chem. 2005, 233, 91. (c) Xu, R.;
Liu, D.; Wang, S.; Mao, B. Macromol. Chem. Phys. 2006, 207, 779.
6. Doi, M.; Edwards, S. F. The Theory of Polymer Dynamics, Oxford: Clarendon Press, 1996.
7. Des Cloizeaux , J. Macromolecules 1990, 23, 4678.
8. Mead D. W. J.Rheol. 1994, 38, 1797.
9. Guzman J. D.; Schieber J. D.; Pollard R. Rheol. Acta. 2005, 44, 342.
10. (a) Leugue, A.; Bailly, C.; Keunings, R. J. Non Newtonian Fluid Mech. 2006, 133, 28. (b)
van Ruymbeke, E.; Bailly, C.; Keunings, R.; Vlassopoulos, D. Macromolecules 2006, 39,
6248. (c) Liu, C. Y.; Halasa, A. F.; Keunings, R.; Bailly, C. Macromolecules 2006, 39, 7415.
(d) Liu, C. Y.; Keunings, R.; Bailly, C. Macromolecules 2007, 40, 2946.
11. Zeichner, G. R.; Patel, P. D. Proc. 2nd World Congress Chem. Eng. 1981, 6, 333.
12. Chadwick, J. C.; van der Burgt, F. P. T. J.; Rastogi, S.; Busico, V.; Cipullo, R.; Talarico, G.;
Heere, J. J. R. Macromolecules 2004, 37, 9722.
13. Huang, R.; Kukalyekar, N.; Koning, C. E.; Chadwick J. C. J. Mol. Catal. A: Chem. 2006,
260, 135.
14. Enders, M.; Fernandez, P.; Ludwig,G.; Pritzkow, H. Organometallics 2001, 20, 5005.
15. Vega, J. F.; Rastogi, S.; Peters, G. W. M.; Meijer, H. E. H. J Rheol. 2004, 48, 663.
16. (a) Mihan, S.; Lilge, D.; De Lange, P.; Schweier, G.; Schneider, M.; Rief, U.; Handrich, U.;
Hack, J.; Enders, M.; Ludwig, G.; Rudolph, R. US 6699948. (b) Enders, M. Macromol.
Symp. 2006, 236, 38.
17. Xu, T.; Mu, Y.; Gao, W.; Ni, J.; Ye, L.; Tao, Y. J. Am. Chem. Soc. 2007, 127, 2236.
18. (a) Bersted, B. H. J. Appl. Polym. Sci. 1985, 30, 3751. (b) Shroff, R. N.; Mavridis, H.
Macromolecules 1999, 32, 8454. (c) Wood-Adams, P. A.; Dealy, J. M.; deGroot A. W.;
Redwine, O. D. Macromolecules 2000, 33, 7489.
19. Lai, S.-Y.; Wilson, J. R.; Knight, G. W.; Stevens, J. C.; Chum, P.-W. S. US 5272236.
20. (a) Malmberg, A.; Kokko, E.; Lehmus, P.; Löfgren, B.; Seppälä, J. V. Macromolecules 1998,
31, 8448. (b) Harrison, D.; Coulter, I. M.; Wang, S.; Nistala, S.; Kuntz, B. A.; Pigeon, M.;
Tian, J.; Collins, S. J. Mol. Catal. A: Chem. 1998, 128, 65. (c) Kolodka, E.; Wang, W.-J.;
Charpentier, P. A.; Zhu, S.; Hamielec, A. E. Polymer 2000, 41, 3985. (d) Gabriel, C.; Kokko,
E.; Löfgren, B.; Seppälä, J. V.; Münstedt. H. Polymer 2002, 43, 6383. (e) Piel, C.; Stadler, F.
J.; Kaschta, J.; Rulhoff, S.; Münstedt, H.; Kaminsky, W. Macromol. Chem Phys. 2006, 207,
26.
21. Reinking, M. K.; Orf, G.; McFaddin, D. J. Polym. Sci.: Part A: Polym. Chem. 1998, 36,
2889.
22. Brandsma, M. J. R.; Brussee, E. A. C.; Meetsma, A.; Hessen, B.; Teuben, J. H. Eur. J. Inorg
Chem. 1998, 1867.
23. Gibson, V. C.; Spitzmesser, S. K. Chem. Rev. 2003, 103, 283.
24. Resconi, L.; Chadwick, J. C.; Cavallo, L. In Comprehensive Organometallic Chemistry III,
Crabtree, R. H.; Mingos, D. M. P., Eds.; Elsevier: Oxford, 2006; Vol. 4, p. 1005.
35
Chapter 2
25. (a) Donkervoort, J. G.; Jastrzebski, J. T. B. H.; Deelman, B.-J.; Kooijman, H.; Veldman, N.;
Spek, A. L.; van Koten, G. Organometallics 1997, 16, 4174. (b) Matsunaga, P. T. WO
99/57159.
26. Chuchuryukin, A. V.; Huang, R.; van Faassen, E. E.; van Klink, G. P. M.; Lutz, M.;
Chadwick, J. C.; Spek, A. L.; van Koten, G., manuscript in preparation.
27. Contel, M.; Stol, M.; Casado, M. A.; van Klink, G. P. M.; Ellis, D. D.; Spek, A. L.; van
Koten, G. Organometallics 2002, 21, 4556.
36
Chapter 3
Synthesis
and
characterization
of
bimodal
polyethylene prepared in a single reactor using a coimmobilized dual catalyst
Abstract
Synthesis of polyethylene (PE) having bimodal molecular weight distribution was carried out in
a single reactor, using a supported dual catalyst system. Ethylene polymerization was performed
and
[2,6-{(2-chloro,4,6with
the
complexes
1-(8-quinolyl)indenyl
CrCl2
dimethylphenyl)N=C(Me)}2C5H3N]FeCl2 immobilized on a MgCl2/AlEtn(OEt)3-n support.
Intimate blends of high (Mw = 1 × 106 g/mole) and low (Mw = 1 × 105 g/mole) molecular weight
PE were synthesized, the relative proportions of the high- and low-molecular weight components
being controlled by varying the loadings of the two precatalysts. This route overcomes the
usually practiced cascade reactor process wherein two reactors are connected in series. High
temperature gel permeation chromatography and differential scanning calorimetry analysis
confirmed bimodality in the polymer molecular weight distribution and the degree of control
offered by this synthesis route. Melt viscosities of the reactor blends were found to be lower than
that of the high molecular weight PE, indicating good miscibility of the high- and low-MW
components.
3.1 Introduction
Molecular weight (MW) and molecular weight distribution (MWD) are important criteria
influencing the physical and mechanical properties of polymers. Generally, MW contributes to
the mechanical properties and MWD to the processing of the polymers. High MW is essential
for better mechanical properties, fatigue resistance and wear properties, while on the other hand
low MW is important for easy processability. Control over MW and MWD allows the polymer
37
Chapter 3
architecture to be tailored to meet the competing demands of processability and properties.
Polyethylene with broad and bimodal MWD is of particular interest in this respect. A schematic
representation of contribution of fractions of molecular weight to various polymer properties is
depicted in Figure 3.1.
Processability
Matrix
Creep Resistance
Mechanical
Strength
Melt
Strength
During
Extrusion
wt %
Reduced
Impact
Strength
Molecular Weight
Figure 3.1 A schematic representation of contribution of various molecular weight fractions to
the property profile of polyethylene with monomodal (---) and bimodal ( ) MWD.
Bimodality or multimodality in polymers can be achieved by two routes: designing polymer
architecture during synthesis in the reactor, or post-synthesis blending of different polymers.
Post-synthesis blending can be performed by extrusion or solution blending, but effective
blending of polymers with very different molecular weights is difficult. Up to the critical
molecular weight (Mc, typically twice the average molecular weight between entanglements) of a
polymer, the viscosity increases linearly with molecular weight, but above Mc the viscosity of the
polymer increases with the 3.4th power of the molecular weight. Melt blending is therefore
impossible when one of the components has very high molecular weight. In the solution blending
process, phase separation of the two fractions can result in an imperfect blend and the need to
remove traces of solvent from the product makes this approach less attractive. Industrial
processes for the production of bimodal polyethylene have therefore been based on the use of
two or more polymerization reactors in series. A typical cascade process, shown in Figure 3.2,
involves a first reactor in which ethylene polymerization is carried out in the presence of a
hydrogen concentration sufficient to give relatively low molecular weight, and a second reactor
with a much lower hydrogen concentration, leading to formation of the high molecular weight
fraction.1 This process also allows the production of bimodal copolymers with an “inverse” comonomer distribution, in the sense that an α-olefin co-monomer is incorporated into the high
rather than the low molecular weight polyethylene fraction. This is advantageous for the
formation of tie molecules between the crystalline lamellae, leading to improved stress crack
resistance.1-3
38
One pot synthesis of bimodal PE
Low MW
PE
Bimodal PE
Catalyst
+
Co-catalyst
Hydrogen
Ethylene
Solvent
Figure 3.2. Industrial cascade reactor process for the synthesis of bimodal PE.
An alternative approach for the synthesis of bimodal polyethylene involves the use of hybrid
catalyst systems in a single polymerization reactor. Intimate blends of high and relatively low
molecular weight polyethylene can be obtained via the co-immobilization of two different
catalysts on a support and significant research efforts are being made in this field.4-12 The most
commonly used support material for the immobilization of single-centre catalysts is silica and
extensive efforts have been made to develop immobilized catalysts for use in slurry and gasphase polymerization processes.13-14 In recent years, however, there has been increased interest
in the use of magnesium chloride-based supports, such as those of type MgCl2/AlRn(OR’)3-n,
prepared by reaction of AlR3 with an adduct of MgCl2 and an alcohol.15-16 Our own studies in
this area have mainly involved the use of MgCl2/AlEtn(OEt)3-n supports, prepared by reaction of
AlEt3 with solid MgCl2/EtOH adducts, for the immobilization of both early- and late-transition
metal catalysts.17-21 These and other22-23 studies have shown that the use of magnesium chloride
supports can give high catalyst activity without the use of methylaluminoxane (MAO) or a borate
activator. Furthermore, controlled (spherical) polyethylene particle morphology can be obtained.
This chapter addresses the synthesis of bimodal polyethylene via the co-immobilization of two
different precatalysts on a MgCl2/AlEtn(OEt)3-n support. The catalyst selected for the lower
molecular weight polymer fraction was the bis(imino)pyridyl iron catalyst {2,6[ArN=C(Me)]2C5H3N}FeCl2, where Ar = 2-chloro-4,6-dimethylphenyl. Previous studies with
this catalyst, immobilized on magnesium chloride, gave polyethylene molecular weights (Mw) of
around 100 kg/mol.21 For the high molecular weight fraction, we selected the chromium catalyst
[1-(8-quinolyl)indenyl]CrCl2. Studies by Enders et al.24 have shown that the rigid spacer group
between the cyclopentadienyl (indenyl) ring and the nitrogen atom in this quinoline complex
gives good stability in homogeneous polymerization. The single-centre characteristics of this and
a related chromium catalyst are retained after immobilization on magnesium chloride, giving
39
Chapter 3
high molecular weight polyethylene (Mw > 1 000 000 g/mol) and narrow polydispersity (Mw/Mn
= 2).19,25 An unsuccessful attempt to synthesize bimodal PE in a single reactor using hydrogen as
a chain transfer agent will also be discussed.
3.2 Materials and methods
3.2.1 Materials
All manipulations were performed under an argon atmosphere using glove box (Braun MB-150
G1 or LM-130) and Schlenk techniques. Light petroleum (b.p. 40-60 °C) and dichloromethane
were passed over a column containing Al2O3 and stored over 4Å molecular sieves. All the
solvents were freeze-thaw degassed at least twice prior to use.
AliBu3 (1 M solution in hexane) was purchased from Akzo Nobel. TiCl4 was purchased from
Aldrich.
Ethylene (3.5 grade supplied by Air Liquide) was purified by passing over columns of 4Å
Molecular Sieves and BTS copper catalyst.
3.2.2 Catalysts
Catalyst (1) and catalyst (4) are mentioned in Chapter 2. {2,6-[ArN=C(Me)]2C5H3N}FeCl2 (13)
(Ar=2-chloro-4,6-dimethyl) was prepared according to a literature procedure.26
N
N
CrCl2
N
1
Fe
Cl Cl
4
N
Cl
Cl
13
Scheme 3.1. Precatalysts used for immobilization on MgCl2 support.
3.2.3 Catalyst Immobilization and Polymerization
Support preparation was performed using the procedure as reported in Chapter 2. For singly
immobilized catalyst systems, a solution of 1 µmol precatalyst in 1 mL dichloromethane was
mixed with 100 mg support and the resultant slurry was kept overnight at room temperature. In
the case of the dual catalyst system, catalyst immobilization was effected by mixing the support
(100-120 mg) with solutions of two precatalysts. A higher amount of support (200 mg) was used
when the two precatalysts exceeded 1.5 µmol loadings. Both the precatalysts were dissolved
separately in dichloromethane and required amounts of them were added in quick succession to
40
One pot synthesis of bimodal PE
the support. The slurry of the support and precatalysts was kept overnight at room temperature
until the supernatant solution became colourless, indicating complete immobilization of the
catalysts. Polymerization was performed following the procedure described in Chapter 2.
3.2.4 Sample Preparation for Melt Rheometry
The polymer was stabilized with 0.2 wt % Irganox 1010 (added as a solution in acetone,
followed by drying in vacuo overnight at 60 °C). Discs of 8 mm diameter and 25 mm diameter
having a thickness of 0.5 mm were compression moulded at 180 °C and 200 bar pressure.
3.2.5 UV-VIS Spectroscopy
The HP8453 UV-VIS spectrometer was used for UV-VIS characterization. 50 mg of support was
transferred to a quartz cuvette equipped with an airtight cap. For individual catalyst
immobilization, a dichloromethane solution containing 0.2 µmol of catalyst was then added;
while for co-immobilization a dichloromethane solution containing 0.13 µmol Cat 4 and 0.07
µmol Cat 13 was used. UV-VIS spectra of the supernatant solution were recorded at definite
time intervals to probe the kinetics of immobilization.
3.2.6 Polymer Characterization
Molecular weight and molecular weight distributions of the resulting polymers were determined
by means of size exclusion chromatography on a PL-GPC210 at 135 °C using 1,2,4trichlorobenzene as solvent. Oscillatory shear measurements were performed on an ARES
rheometer from Rheometric Scientific. Frequency sweep measurements were carried out at 160
°C. Polymer morphology was studied on a Philips Environmental Scanning Electron Microscope
(XL-30 ESEM-FEG). DSC measurements were carried out on a Q1000 DSC from TA
Instruments. NMR experiments were performed on a Bruker Avance 500 solid state NMR
spectrometer.Further details of the characterizations carried out are given in the respective
subsections.
3.3 Results and Discussion
3.3.1 Use of Hydrogen for the Synthesis of Bimodal Polyethylene
Since the use of molecular hydrogen as a chain transfer agent was first reported by Vandenberg,
Ettore and Luciano, and Natta27, it has been very widely used to control the molecular weight of
polyolefins produced using Ziegler-Natta catalysts. In propylene polymerization, the addition of
hydrogen generally leads to significant increases in catalyst activity, as a result of reactivation
via chain transfer of “dormant” sites formed by 2,1- rather than the usual 1,2-insertion of the
41
Chapter 3
monomer.28 On the other hand, the presence of hydrogen in ethylene polymerization frequently
leads to decreased activity, depending on the type of catalyst used.
As described in the introductory part of this chapter, the cascade reactor process involves the use
of hydrogen to produce bimodal polyethylene on an industrial scale. At least two reactors are
required, although the formation of bimodal polyethylene in a single polymerization step in the
presence of hydrogen has been reported by Soares et al.29, who used a metallocene immobilized
on silica/MAO. From previous studies it is known that the MgCl2/AlEtn(OEt)3-n immobilized
catalyst systems are able to produce polyethylenes having very high molecular weights and
varied molecular weight distribution17-19, prompting an investigation in the present work of the
hydrogen response of the highly active but multi-centre immobilized TiCl4 and the moderately
active but single-centre (η1:η5-Me2NCH2CH2C5Me4)CrCl2. The polymerization reactor was
pressurized with the required amount of hydrogen before adding ethylene. Ethylene
polymerization was performed in the presence of hydrogen until the polymerization was stopped
by adding acidified ethanol to the reactor. The polymerization results are listed in Table 3.1.
Table 3.1. Effect of hydrogen on catalyst activity, molecular weight and molecular weight
distribution.
Entry
Catalyst
Hydrogen
(bar)
1
TiCl4
0
2
TiCl4
1
3
TiCl4
2
4
1
0
5
1
1
6
1
2
Polymerization temperature: 50°C.
Activity
(kg/mol M.bar.h)
Mw
(kg/mol)
Mn
(kg/mol)
Mw/Mn
10200
2035
649
3.1
4600
872
175
5.0
2800
476
133
3.6
1500
2557
1420
1.8
140
70
Support composition: MgCl2·0.35AlEt2.66(OEt)0.34.
Addition of hydrogen to the TiCl4-catalyzed ethylene polymerization resulted in a dramatic
decrease in the catalyst activity. One bar of hydrogen decreased the catalyst activity to half and 2
bar hydrogen pressure decreased the activity to almost 1/3 of the original activity. The activity of
catalyst 1 was also very sensitive to hydrogen and as the catalyst activity was relatively low even
in the absence of hydrogen, this system was not investigated further.
Taking into account the significant lowering in molecular weight in the presence of hydrogen
obtained with the immobilized TiCl4 catalyst (i.e. good hydrogen response), a two-step
polymerization process, equivalent to the cascade reactor process, was applied. The first step
involved ethylene polymerization in the presence of hydrogen (2 bar pressure) to produce
relatively low molecular weight PE (Mw ~ 400 kg/mol). In the second step all the gaseous
42
One pot synthesis of bimodal PE
contents were vented off and the reactor was flushed with ethylene to remove residual traces of
hydrogen; further polymerisation of ethylene would then be expected to produce high molecular
weight PE. The polymerization time in the first stage (under 2 bar hydrogen pressure) was varied
in order to obtain polymers with different proportions of low- and high-molecular weight
components, keeping the total polymerization time for the combined first and second steps
constant at 60 min. Polymerization details and GPC results are listed in Table 3.2.
The results in Table 3.2 indicate that high catalyst activity is restored after venting off hydrogen
from the reactor. However, it was observed that instead of getting bimodality in the molecular
weight distribution, only a moderate broadening in MWD was obtained. This observation can be
attributed to the fact that hydrogen decreases the catalyst activity dramatically. The relatively
small amount of the low molecular weight PE component is therefore overshadowed by the high
molecular weight component produced after venting off hydrogen from the reactor. As shown in
entry 2, upon venting off hydrogen after 10 minutes, the activity of the catalyst was higher than
the catalyst activity when polymerization was carried out without hydrogen.
Table 3.2. Influence of two-step polymerization on molecular weight distribution.
Entry
Timea
(Min)
Activity
(kg/mol M.bar.h)
Mw
(kg/mol)
1
2
3
4
5
0
10
30
40
60
10200
14200
6800
6000
2775
2035
1644
1916
1173
476
Mn
Mw/Mn
(kg/mol)
649
432
671
245
133
3.13
3.80
2.85
4.68
3.58
Catalyst: TiCl4.
Total polymerization time: 60 min.
a
: Time denoting polymerization carried out in presence of hydrogen.
Support composition: MgCl2·0.13AlEt1.99(OEt)1.01.
In entries 3 and 4, although the overall activities were less than for entry 1, they were much
higher than the activity obtained (entry 5) when the total polymerization was carried out in the
presence of hydrogen. The results indicate that there is no linear relationship regarding the
reversibility of catalyst activity in the presence/absence of hydrogen.
In order to understand the possible mechanism behind the apparent increase in the catalyst
activity after the venting of hydrogen, we studied Scanning Electron Micrographs (SEM) of the
polymers prepared as indicated in entries 1 and 2. The polymer beads were embedded in epoxy
resin and cut transversely to observe the polymer bead morphology in the core and the shell. The
micrographs depicted in Figure 3.3 are the transverse sections taken with a Philips
43
Chapter 3
Environmental Scanning Electron Microscope (XL-30 ESEM-FEG). Figure 3.3(a) shows a very
porous morphology for the polymer particle produced without hydrogen, whereas Figure 3.3(b)
shows a relatively compact morphology for the particle obtained from the two-step
polymerisation. The porous morphology can be attributed to a monomer mass transfer limitation,
restricting the transport of ethylene through the growing particle. Due to the mass transfer
limitation, the core of the polymer particle remains hollow as some of the active centres do not
receive monomer. The more compact morphology obtained in the two-step polymerization could
be due to the presence of a small molecule such as hydrogen increasing the monomer transport
through the growing particle, as has been suggested by Parasu Veera.30 The suppression of mass
transfer limitation at the initial stage of polymerization could therefore be the reason for the
increased catalyst activity.
Although it was found that the catalyst activity was reversible after venting off hydrogen from
the reactor, the above results indicate that this approach is not suitable for the synthesis of
bimodal PE in a single reactor. Further studies were therefore concentrated on the coimmobilization of two catalysts on one support; this is discussed in detail in the following
section.
(a)
(b)
Figure 3.3. Scanning electron micrographs showing (a) porous particle morphology for entry 1
in Table 3.2; (b) compact particle morphology for entry 2.
3.3.2 Bimodality by Dual Catalyst System
Separate and co-immobilization of the quinolyl chromium precatalyst 4 and the
bis(imino)pyridyl iron precatalyst 13 was carried out by mixing the support with solutions of the
precatalysts in dichloromethane. In order to achieve high catalyst activity, the precatalyst
loadings on the support were limited to 10 µmol/g. Previous studies with similar supports have
shown an inverse relationship between loading and catalyst activity.17 Quantitative catalyst
immobilization was evident from the complete discoloration of the precatalyst solution.
44
One pot synthesis of bimodal PE
3.3.2.1 UV-Visible Spectroscopy
Many polyethylene catalysts, such as bis(imino)pyridyl iron complexes and metallocenes, have a
strongly conjugated ligand system and therefore show strong absorption in UV-VIS
spectroscopy. Many researchers have used UV-VIS spectroscopy to study the mechanism of
polyolefin catalysis with such conjugated systems.31 In the present work, UV-VIS was used to
investigate the relative rates of immobilization of different catalysts on MgCl2-based supports.
1.2
1.2
4
4
1.0
Absorbance (AU)
Absorbance (AU)
1.0
0.8
0.6
0.4
13
0.2
0.8
0.6
13
0.4
0.2
400
600
800
1000
400
Wavelength (nm)
600
800
Wavelength (nm)
(a)
(b)
Figure 3.4. UV-VIS spectrum of (a) Cat 4 and Cat 13 in solution after adding to the support in
individual immobilization (b) Co-immobilization of Cat 4 and Cat 13 on the support.
As shown in Figure 3.4(a), Cat 4 shows a strong absorption peak at 315 nm, while Cat 13
absorbs at 715 nm. These two peaks are not well separated, and for Cat 4 there is also a small
absorption peak at 690 nm. However, for qualitative analysis, assignment of the peaks to the
respective catalysts appears to be effective. When the solutions of Cat 4 and Cat 13 were added
to the support for co-immobilization, absorption peaks at 315 nm and 715 nm appeared (Figure
3.4(b)), similar to the peaks obtained with each individual catalyst.
Upon addition of catalyst solution to the support, the colour of the supernatant solution faded
away with time, as the catalyst became immobilized on the support. Therefore, decrease in peak
intensity in the UV-VIS spectrum gives a direct measure of the rate of catalyst immobilization.
Decrease in the intensity at 315 nm and 715 nm was followed with time in order to study the
kinetics of catalyst immobilization for both singly- and co-immobilized catalysts.
In Figure 3.5(a) absorption peaks at 715 nm, corresponding to the singly immobilized Cat 13,
recorded at different time intervals are plotted. As evident from the plot, fast decrease in the peak
intensity suggests easy immobilization of the iron catalyst on the support. The time required for
complete immobilization of the catalyst onto the support was found to be around 7 hours. Cat 4
was slow in immobilization and was not completely immobilized even after 21 hours, as shown
45
Chapter 3
0.5
1.2
A
1.0
Absorbance (AU)
Absorbance (AU)
0.4
B
0.3
0.2
A
B
0.8
0.6
C
0.4
C
D
0.2
0.1
0.0
600
700
800
300
400
Wavelength (nm)
500
Wavelength (nm)
(a)
(b)
1.2
Absorbance (AU)
1.0
0.8
0.6
0.4
0.2
0.0
0
100
200
300
400
Time (Min)
(c)
Figure 3.5. (a) UV-VIS absorption peak with time (A= 0 min, B=70 min, C =420 min) for singly
immobilized catalyst 13.
(b) UV-VIS absorption peak with time (A= 0 min, B=120 min, C =420 min D=1300
min) for singly immobilized catalyst 4.
(c) Immobilization profile with time for co-immobilized Cat 4 (□) and Cat 13 (●).
in Figure 3.5(b). A normalised immobilization profile (UV response of the Fe complex is
multiplied by a constant for a good comparison with the response of the Cr complex) of coimmobilized iron and chromium catalyst with time is depicted in Figure 3.5(c). As was found for
the singly immobilized catalysts, immobilization of Cat 13 was faster than Cat 4. In Fig 3.5(c)
one can observe that the absorbance value for Cat 4 after 420 min is lower than that found in Fig
3.5(b). This observation might suggest that when co-immobilized a synergistic effect is observed
in the rate of immobilization of the two precatalysts, or it can be an experimental error, although
the observation was consistent in repeated experiments. The co-immobilization experiments
were carried out keeping the same metal to support ratio as in the singly immobilized catalyst
46
One pot synthesis of bimodal PE
system. It is important to mention that only one combination of loadings of the two catalysts was
investigated, and the kinetics of immobilization would definitely vary according to the Fe:Cr
ratio when the loadings are changed. However, these studies were carried out to probe the
kinetics of immobilization in a qualitative manner and to check the distribution of two catalysts
on the support. It is probable, on the basis of these results, that the more easily immobilized iron
catalyst occupies the most suitable centres on the support, with the chromium catalyst then
becoming immobilized on the remaining support centres.
These observations, though they do not confirm a statistical distribution of these two catalysts on
the support, however do not rule out the same. Given the fact that low catalyst loadings were
used for these studies and polymerizations, it is difficult to determine the distribution of catalysts
on the support with tools such as Scanning Electron Microscopy and Energy Dispersive X-ray.
3.3.2.2 Polymerization
Ethylene polymerization was carried out at 50 °C in light petroleum slurry, using either AlEt3 or
AliBu3 as cocatalyst. Polymerizations were first carried out with each individual catalyst
immobilized on the magnesium chloride support. The results are given in Table 3.3. It is
apparent, particularly for the chromium catalyst, that the highest activities were obtained using
AlEt3. However, the somewhat broader polydispersities obtained with AlEt3 prompted the
selection of AliBu3 as cocatalyst for polymerizations with the co-immobilized catalysts.
Confirmation of retention of the single-centre nature of the chromium catalyst upon
immobilization is discussed in detail in Chapter 2. Relatively broad molecular weight
distribution is a typical feature of polyethylene prepared with bis(imino)pyridyl iron catalysts;
these systems, in contrast to the chromium catalyst, are not single-centre.21
Table 3.3. Effect of cocatalyst in ethylene polymerization with Cat 4 and Cat 13 immobilized on
a support having the composition : MgCl2·0.11AlEt1.72(OEt)1.28
Entry
Catalyst
Cocatalyst
Activity
(kg/mol of
M.bar.h)
Mw
(kg/mol)
Mn
(kg/mol)
Mw/Mn
1
2
3
4
4
4
13
13
AlEt3
AliBu3
AlEt3
AliBu3
4800
2000
8000
6000
1409
1324
103
97
744
824
13
16
1.9
1.6
7.7
5.9
The results of polymerizations carried out after co-immobilization of different proportions of
catalysts 4 and 13 are given in Table 3.4. The combined activity reported in this table is the
activity per mol of total metal (Fe and/or Cr) in the catalyst. Catalyst loadings were selected
47
Chapter 3
taking into account the different activities of the individual catalysts. In the immobilization,
dichloromethane solutions of each catalyst were added to the support in quick succession, aiming
to achieve a statistical distribution of both catalysts throughout the support particle. The polymer
composition column in Table 3.4 contains the weight fractions of the high molecular weight
(HM) and low molecular weight (LM) polyethylene fractions produced, determined on the basis
of GPC analysis
Table 3.4. Ethylene polymerization with Cat 4 and Cat 13 in tandem.
Exp
Catalyst
Loading
(µmol/100mg)
4
13
PE1
0
1
PE2
0.5
PE3
Combined
Activity
kg.(mol M)-1
(bar)-1(h)-1
PE
Composition
(wt %)
M
w
M
n
(kg/mol)
(kg/mol)
M w/M
HM
LM
6000
0
100
97
16
5.9
0.8
5460
15
85
150
16
8.9
1
0.5
3470
37
63
555
20
27.0
PE4
1
0.3
3385
45
55
742
25
29.5
PE5
2
0.3
2440
68
32
443
28
15.5
PE6
2
0.1
2095
80
20
686
81
8.5
PE7
1
0
2000
100
0
1324
824
1.6
n
The trends in overall molecular weight and molecular weight distribution apparent from the data
in Table 3.4 are indicative of the increasing amount of high molar mass PE with increase in the
loading of precatalyst 4. GPC plots of polyethylenes prepared with the individual catalysts 4 and
13, using AliBu3 as cocatalyst, are shown in Figure 3.6, while the GPC traces obtained for the
bimodal samples (Exp PE2-PE6) are shown in Figure 3.7. The peaks for the low- and highmolecular weight fractions, representing the polymer produced with catalysts 13 and 4,
respectively, are clearly distinguishable. The areas under the GPC peaks were determined by
using 15 degree polynomial fitting in Matlab. The areas under the GPC peaks of the high and
low molecular weight components were found to be in good agreement with the expected
composition based on the catalyst loading.
Several polymerizations, which were performed in duplicates and triplicates, confirmed the
reproducibility of the experimental results. GPC characterization of the samples synthesized in
duplicates provided the reproducible desired composition. A good control over the
polymerization conditions assured the repeatability.
48
One pot synthesis of bimodal PE
1.2
0.8
1.0
0.7
d WF/ dlog M
d WF/ dlog M
0.6
0.5
0.4
0.3
0.8
0.6
0.4
0.2
0.2
0.1
0.0
0.0
2
3
4
5
6
7
3
4
Log M
5
6
7
8
log M
(a)
(b)
Figure 3.6. GPC plots of (a) PE 1, synthesized using Cat 13 and (b) PE 7 synthesized using Cat
4.
d WF/ dlog M
PE 6
PE 5
PE 4
PE 3
PE 2
3
4
5
6
7
log M
Figure 3.7. GPC plots of PE synthesized using co-immobilized Cat 4 and Cat 13 on
MgCl2/AlEtn(OEt)3-n support.
A comparison of expected yield and actual yield, in grams of polyethylene, is represented in
Figure 3.8. A good agreement between the actual yield and the expected yield indicates that the
combined activities of the co-immobilized catalysts correspond to the sum of their individual
activities. In other words, in this system there is no synergetic effect of co-immobilization on
catalyst activity, in contrast to systems in which an iron or chromium catalyst was immobilized
with a nickel diimine catalyst.33
49
Chapter 3
40
35
30
Yield (g)
25
20
15
10
5
0
PE1
PE2
PE3
PE4
PE-5
PE-6
PE7
PE Samples
Figure 3.8. Comparison showing expected yield (□) and actual yield (■).
In the latter case, increased activity of the iron or chromium catalyst after co-immobilization
with a nickel diimine was ascribed to the presence of Ni-catalyzed branched polyethylene.
3.3.2.3 Thermal Analysis
Thermal analysis of the synthesized bimodal polyethylene was carried out by means of
Differential Scanning Calorimetry (DSC). A very high first melting point, 144.6 °C, was
observed for nascent high molecular weight PE (PE7), while nascent low molecular weight PE
(PE1) showed a first melting peak at 136 °C. The melting temperature of melt crystallized PE 7
was 135 °C and that of PE 1 was 134 °C. DSC thermograms are shown in Figure 3.9.
144.6 °C
134 °C
Heat Flow
Heat Flow
136 °C
20
40
60
80
100
120
140
160
135 °C
20
Temp ( °C)
40
60
80
100
120
140
160
Temp ( °C)
(a)
(b)
Figure 3.9. Heating runs of (a) PE1 and (b) PE7 comparing the first (nascent) and second (melt
crystallized) melting point.
The observed differences in the melting temperature on the first and the second heating runs is a
long standing issue. From the results shown in Figure 3.9 it is evident that the difference
increases with the increasing molar mass. Engelen et al.34 used electron microscopy to show that
the nascent crystals are also folded chain crystals. Considering similar crystal thickness in the
50
One pot synthesis of bimodal PE
nascent and melt crystallized samples, the difference in the melting temperature cannot be simply
explained by the Gibbs-Thomson equation. Recently, Lippits et al.35,36 have attributed the cause
of the difference in the melting temperature to the melting kinetics of the crystals. One of the
main causes for the kinetics in melting was attributed to topological differences in the amorphous
phases of the semi-crystalline polymers, obtained on synthesis and after melting. These
topological differences in semi-crystalline polymers arise due to differences in the chain folding
such as loose or tight folds on the crystal surface. Differences in the chain conformations within
the amorphous phase of nascent and melt crystallized samples have been elucidated by dynamic
NMR.37 To recall, in the nascent high molecular weight PE, possible chain conformations are
restricted giving rise to a significantly broader peak width in the 13C spectra compared to the
melt crystallized sample of the same polymer. These differences in the chain conformations are
attributed to tight folds present in the amorphous region of the crystal. Since the melting process
invokes adoption of a constrained chain in the crystal into the random coil state, the melting of a
polymer crystal will not only be influenced by the crystalline but also by the amorphous regions,
especially when chain segments in the amorphous region are constrained. Thus it has been
suggested that the melting should take into account not the crystal thickness but the number of
CH2 units cooperatively connected having the restricted chain conformation.38 This causes
increase in the melting point of the nascent polyethylenes. However, the loose folds in the melt
crystallized samples means that the number of CH2 units required for co-operative motion is
equal to the crystal thickness, resulting in the normal melting temperature, 135 °C, as predicted
by the Gibbs-Thomson equation.
DSC can be effectively used to fractionate the molar mass by altering the heating rates. In order
to confirm the bimodality in the molecular weight distribution of the PE samples synthesized,
two kinds of DSC experiments were carried out with two different temperature profiles: Profile
A with a normal heating rate of 10 °C/min and Profile B with slow heating rates. For the slow
heating cycle profile, PE samples of around 3-5 mg were heated to 120 °C at 10 °C/min and then
a heating rate of 1 °C/min was applied up to 160 °C. A similar protocol was used for the second
heating run. The melting temperatures of all samples, from PE1 to PE7, are listed in Table 3.5
and the DSC thermograms of the nascent PE samples are shown in Figure 3.10. From Figure
3.10 it is apparent that on following profile A, fast heating, a single endothermic transition peak
is observed, though the peak shows an asymmetric profile with an increasing amount of high
molar mass. The asymmetric peak gets well resolved into two peaks, peak 1 and peak 2, once the
heating rate is reduced to 1 °C/min, following the profile B. Independent of the adopted profile
the samples were cooled to room temperature at a rate of 10 °C/min and heated again (second
heating) following the Profile B. On the second heating a single melting peak was observed.
51
Chapter 3
Table 3.5 DSC analysis of PE samples.
PE
PE1
PE2
PE3
PE4
PE5
PE6
PE7
a
Tm Peak
Profile A
Nascent Melt
crystallized
136.0
134.0
137.5
137.4
138.2
137.0
140.0
137.5
142.0
137.2
142.8
135.8
144.6
135
Tm Peak NascentProfile B
Peak 1
Peak 2
(°C)
(°C)
133.6
133.7
a
139.0
134.7
140.0
135.3
140.3
135.0
141.9
135.0
142.4
-
Tm melt
crystallized (°C)
Profile B
133.0
133.7
134.6
134.6
134.3
134.0
134.0
a tail appeared showing a peak at 139 °C.
PE-7
PE-7
PE-6
PE-6
Heat Flow
Heat Flow
PE-5
PE-4
PE-5
PE-4
PE-3
PE-3
PE-2
PE-2
PE-1
PE-1
120
100
120
130
140
150
140
Temp (°C)
Temp (°C)
(a)
(b)
Figure 3.10. DSC thermograms of nascent PE samples, PE1 – PE7, listed in Table 3.4. (a)
Heating rate 10 °C/min and (b) heating rate 1 °C/min.
A comparison of melting temperatures obtained from two different temperature profiles, as
depicted in Table 3.5, shows that the faster heating rate yielded higher melting points than those
obtained from the low heating rate. This observation held good for both the first (nascent) and
second (melt crystallized) melting points of all samples. Generally, for polymers where
reorganization causes crystal thickening, melting temperature increases with decreasing heating
rate. When such a crystal thickening is not possible, melting point decreases with decrease in the
heating rate. Increase in the melting temperature with higher heating rates is attributed to
superheating which is caused by thermal inertia and time-dependent melting processes.39 In this
present scenario, although the reorganization is possible as the crystals are not extended chain
crystals, melting temperature decreases with a slower heating rate. This clearly suggests that the
52
One pot synthesis of bimodal PE
increase in the melting temperature at the heating rate of 10 °C/min is due to the superheating
effect.
PE1 (Mw = 105 g/mol) showed a sharp and symmetric melting peak at 136 °C, when temperature
profile A was applied. In Figure 3.10(a), as one goes up the series, broadening of the melting
peak with a shift towards higher temperature can be seen. In the series up to PE4, where the
composition is dominated by the low molecular weight component, broadening of the peak with
retention of the symmetry is observed. However, when the proportion of high molecular weight
content is increased beyond 50 weight % the peak becomes asymmetric, although no clear
shoulder or second peak is visible. PE7 (Mw = 106 g/mol), showed a sharp melting peak at 144
°C, even higher than the equilibrium melting temperature (141 °C) of PE.
As shown in Figure 3.10(b), when temperature profile B was applied the peak melting
temperature for PE1 shifted to 133.6 °C and an asymmetric peak was observed. Peak broadening
was observed for PE2 and PE3 while PE4, PE5 and PE6 clearly showed two peaks. The positions
of the two peaks showed a good match with the peak melting temperature for the nascent high
and low molar mass polyethylenes at these heating rates. This observation suggests formation of
two different populations of crystals, one from high molar mass and the other from low molar
mass. However, during the second heating cycle both peaks merged to show only one melting
peak which was around 134 °C. The single melting peak on the second heating run even in the
bi-modal samples can be attributed to the expected melting temperature for the melt crystallized
samples, independent of the molar mass. To resolve the extent to which chains of two different
molar masses are intermixed, DSC observations alone are not conclusive. This issue of
intermixing is addressed in the section on rheology of this chapter.
3.3.2.4 NMR characterization to determine the degree of branching
It is known that bis(imino)pyridyl iron catalysts generate PE with unsaturated chain ends.40
When a dual catalyst system, comprising a bis(imino)pyridyl iron catalyst and another catalyst
which can incorporate alpha-olefin, is used then there is a possibility of generation of branched
PE. In ethylene/1-hexene copolymerization experiments carried out with the chromium catalyst 4
immobilized on a MgCl2/AlEtn(OEt)3-n support, moderate co-monomer incorporation was
observed. It is therefore possible that the combination of catalysts 4 and 13 on a single support
could result in the formation of some long-chain branching if vinyl-terminated chains generated
by the iron catalyst are incorporated as comonomer at chromium centres. The formation of shortchain branching is also possible, as bis(imino)pyridyl iron catalysts can also produce vinylterminated oligomers.41
In order to establish whether any branching could be detected in the bimodal polyethylene
prepared with the co-immobilized catalysts 4 and 13, NMR characterization of samples PE2 and
53
Chapter 3
PE7 was done at Max Planck Institute, Mainz. Of the bimodal polymers prepared, PE2 (15 wt%
of UHMW-PE and 85wt % of low molar mass) is the sample most likely to contain branching
due to the high loading of iron relative to chromium. For comparison and as a reference, sample
PE7 (100wt % of UHMW-PE), prepared with the chromium catalyst only, was also analyzed.
Solution-state NMR experiments did not produce good results due to very low signal to noise
ratio. In order to obtain better signal to noise ratio a high concentration of polymer sample is
required and therefore melt-state NMR was used to determine the degree of branching.
The measurement was carried out on a Bruker Avance 500 solid-state NMR spectrometer
operating at a proton and carbon Larmor frequency of 500.13 and 125.75 MHz, respectively. A
commercial Bruker 13C-1H optimized high temperature 7 mm MAS probe-head was used for the
measurement. Nitrogen gas was used for all pneumatics to limit thermal oxidation. The
experimental and quantification methods followed have been reported in the Ph.D. thesis of
Katja Klimke.42,43 The measurement was conducted at 3 KHz spinning speed and at a sample
temperature of 150 °C. Quantification of the degree of branching was achieved by comparing the
peak areas of end group and backbone in the 13C single pulse excitation spectrum, which was
acquired using 10 µs 13C 90° excitation pulses and 180° pulse train heteronuclear dipolar
decoupling.
3
α
2
(a)
(b)
Figure 3.11. NMR spectra of (a) PE7 showing no peak for alpha carbon atom and (b) PE2
showing a peak at 34 ppm for carbon atom alpha to the branching point.
As apparent from the NMR spectrum (b), a peak appearing at 34 ppm, which corresponds to
three carbons in the α-position relative to the branching point, confirms that there is some
branching in the sample. Two other peaks, which can be attributed to penultimate and prepenultimate chain-end carbon atoms, were apparent in the spectrum, but the peak for carbon in
the β-position relative to the branch point (normally appearing at ~27 ppm) was absent. If the
presence of impurities and solvent traces in the sample is excluded, the absence of the β-C peak
can be attributed to its long T1 relaxation time. Quantification of the peak areas showed that the
54
One pot synthesis of bimodal PE
degree of branching in sample PE2 was approximately 3 branches per 104 carbon atoms, while
the level of branching in PE7 was less than 1 branch per 105 carbon atoms.
3.3.2.5 Rheology
The flow properties of polymers in the melt state depend on chemical composition, molecular
friction coefficient, molecular weight and molecular weight distribution. Rheology of
copolymers and polymer blends is often complicated, depending upon the nature and
composition of the components involved. In the case of polymer blends where both the
components are of same molecular structure but vary in the molecular weight and molecular
weight distribution, the rheological properties are governed by molecular weight and molecular
weight distribution. The bimodal PE blends discussed in this chapter have two components
having the same molecular structure. The flow properties of these blends therefore depend on the
relative proportions of the high and low molecular weight components and the molecular weight
distribution. Rheological investigation was carried out to probe into the elastic and viscous
properties of the polymers synthesized and to study miscibility in these intimate bimodal PE
blends.
First we recall some of the general rheological observations on a visco-elastic melt. Rheological
studies on a polymer melt reflect the frequency response of chain dynamics at different length
scales. At low frequencies polymer chains have sufficient time to rearrange and relax, thus
highlighting the viscous response of the visco-elastic melt. With increasing frequency, as the
presence of physical entanglements becomes evident, the elastic response of the viscoelastic
polymer melt is observed as a plateau region. At higher frequencies chain dynamics within the
entanglements are realised. This translation of the chain dynamics to the experimental findings is
depicted in Figure 3.12.
Figure 3.12. Relaxation in flow of linear polymers. (reproduced from a presentation by Prof.
Tom McLeish) CR – constraint release (diffusion), CCR – convective constraint release, CLF –
contour length fluctuations.
55
Chapter 3
Figure 3.12 describes relaxation in polymers with respect to normalised shear rate. At low shear
rates the relaxation takes place by a combined influence of reptation, constraint release and
contour length fluctuations. In an entangled polymer melt, contour length fluctuations are
supposed to lift the chain confinements from the chain ends while constraint release occurs along
the chain. So, at low shear rate or at low frequencies, the longer chains have a more prominent
effect on relaxation behaviour. At low frequencies, chains have sufficient amount of time to relax
and thus the energy is dissipated and the melt appears more viscous. At high shear rates the
relaxation is mainly governed by strong stretching as well as by convective constraint release
(CCR) and contour length fluctuations (CLF). So, at these high frequencies, chains do not have
sufficient time for relaxation and only the stretching of chains within the entanglements is
realised. Thus the polymer melt exhibits elastic nature.
In the present bimodal blends a peculiar relaxation pattern is expected, as the two components of
the blend possess considerably different relaxation times. Small amplitude oscillatory shear
experiments were carried out to determine the viscoelastic response of the bimodal polyethylenes
reported in Table 3.4. Measurements of storage (G’) and loss (G’’) moduli were performed using
parallel plate geometry. The linear viscoelastic region of each bimodal PE was determined by
carrying out a dynamic strain sweep test between 0.1 and 100 % strain at 100 rad/s and 160 °C.
The highest strain in the linear viscoelastic region for each sample, at the frequency of 100 rad/s,
was chosen as the experimental strain for that sample. Dynamic frequency sweep tests were
performed over the angular frequency range between 0.001 rad/s and 100 rad/s at 160 °C for a
fixed strain, where the strain varied from 0.5 % to 7 % for high to low molar mass.
The series of samples prepared by controlled synthesis provides a valuable opportunity to
investigate the variation in viscoelastic response of a polymer melt with increasing content of
high molar mass in the matrix of low molar mass polymer. These studies also provide insight
into the extent of molecular miscibility in the synthesised blend. In Figure 3.13(a) and (b), the
storage moduli (G’) of all bimodal PE samples as a function of frequency are shown. As these
data are analogous to the relaxation modulus of the polymer, the results can give some
information about variation in the viscoelastic response of a polymer melt with change in the
molar mass distribution. In Figure 3.13 (c), tan delta (G”/G’) values of the polymers listed in
Table 3.4 are plotted against frequency.
56
One pot synthesis of bimodal PE
10
5
10
4
10
3
10
2
10
1
6
G' (Pa)
5
10
10
-4
10
-3
-2
10
-1
0
10
10
1
10
10
10
2
-4
-3
-2
10
-1
10
10
0
10
1
10
2
10
Frequency(rad/s)
Frequency(rad/s)
(a)
(b)
5
4
3
tan-delta
G' (Pa)
10
2
1
0
-4
10
-3
10
-2
-1
10
10
10
0
1
10
2
10
Frequency(rad/s)
(c)
Figure 3.13. Viscoelastic responses of (a) PE1 (■), PE2 (●), PE3 (▲), and (b) PE4 (◊), PE5 (♦),
PE6 (○) and PE7 (∆) comparing storage moduli as a function of angular frequency. (c) Flow
recovery characteristics, tan delta versus frequency.
The viscoelastic response of the low molecular weight PE (PE1) is shown in Figure 3.13 (a). As
apparent from the figure, the storage modulus increases with increasing frequency; however, no
plateau region is realised, suggesting fast chain dynamics within the probed frequency region. It
is to be noted that because of highly viscous response of the low molar mass material the phase
angle (tan delta value) is high as denoted in Figure 3.13 (c). On the addition of 15 wt% of
UHMW-PE component in the matrix of low molecular weight PE, the anticipated increase in the
storage modulus is observed at all frequencies. The melt shows a considerable shift from viscous
to elastic response, which is apparent from the phase angle. At lower frequencies, a plateau
seems to appear. The low modulus of this low frequency plateau suggests less number of
57
Chapter 3
entanglements in the chain. This may be explained by the dilution of physical entanglements
between the high molecular weight chains, that arises due to differences in the chain dynamics
of the low and high molar mass polymer. In the time frame of high molar mass, faster chain
dynamics of the low molar mass will cause constrained release of physical entanglements. This
constrained release will result in the dilution of entanglements, leading to lower plateau modulus
at lower frequencies. This is depicted in Figure 3.13 (b). With the increasing amount of high
molecular weight in the polymer blends (from PE3 to PE6), decrease in the constrained release
caused by the low molar mass will result in an increase in plateau modulus at lower frequencies.
The modulus will spread over the wider frequency range, ultimately shifting to a single plateau
arising from the intrinsic molar mass between the entanglements for linear polyethylene (PE7).
The phase angle shows the anticipated continuous decrease with the increasing amount of high
molar mass content – suggesting a change from viscous to elastic response of the melt.
The gradual increase in modulus with increasing molar mass from PE2 to PE7 over the explored
frequency range suggests the homogeneous miscibility of the high and low molar mass
polyethylene components produced via one pot synthesis. From the rheological studies
performed here the influence of branches on viscoelastic response of the polymer melt is not
apparent. From the adopted approach it will not be feasible to make a distinction between the
constrained release promoted by low molar mass component or by branches. Considering the low
amount of branch content in the polymer blend, the influence of branches on the rheological
response will be considerably low compared to the influence of low molar mass present in a
blend.
These rheological investigations suggested that both molecular weight and molecular weight
distribution govern the melt properties of these reactor blends. Despite the differences in the
relaxation time and melt viscosity of the high and low molecular weight components, these
polymers were melt miscible, which was apparent from the gradual trend in the rheological
behaviour of the bimodal polyethylenes.
3.4 Conclusions
Intimate bimodal blends of high and low molecular weight polyethylene were successfully
synthesized using a co-immobilized dual catalyst system comprising a chromium and an iron
complex on a magnesium chloride support. The precatalysts involved in this system, which
retained their identity upon immobilization, showed no instances of synergic or poisoning effects
as far as their activities were concerned. Thermal analysis performed on the nascent blends show
crystallization of the low and high molar mass crystals on synthesis as depicted by distinct
melting peaks, while on melting of the melt crystallized sample a single peak is observed. To
probe the extent of molecular miscibility in the melt state of the blends, rheological studies were
performed. The studies showed a regular trend in the viscoelastic response of the polymer melt
58
One pot synthesis of bimodal PE
with increasing amount of high molar mass content. The studies suggest homogeneous mixing of
the high and low molar mass components within the time frame of the experiments under low
strain. Synthesis of bimodal polyethylene using hydrogen as a chain transfer agent was not
achieved; however, the reversibility in the catalyst activity in the presence/absence of hydrogen
was confirmed.
3.5 References
1. Alt, F. P.; Böhm, L. L.; Enderle, H.-F.; Berthold, J. Macromol. Symp. 2001, 163, 135.
2. Seguela, R. J. Polym. Sci.: Part B: Polym. Phys. 2005, 43, 1729.
3. Hubert, L.; David, L.; Seguela, R.; Vigier, G.; Degoulet, C.; Germain, Y. Polymer 2001,
42, 8425.
4. Han, T. K.; Choi, H. K.; Jeung, D. W.; Ko, Y. S.; Woo, S. I. Macromol. Chem. Phys.
1995, 196, 2637.
5. Ahn, T. O.; Hong, S. C.; Huh, W. S.; Lee, Y.-C.; Lee, D.-H. Polym. Eng. Sci. 1999, 39,
1257.
6. Kim, J. D.; Soares, J. B. P.; Rempel, G. L. J. Polym. Sci. Part A: Polym. Chem. 1999, 37,
331.
7. De Souza, R. F.; Casagrande, O. L; Macromol. Rapid Commun. 2001, 22, 1293.
8. Liu, H.-T.; Davey, C. R.; Shirodkar, P. P. Macromol. Symp. 2003, 195, 309.
9. Vaughan, G. A.; Szul, J. F.; McKee, M. G.; Farley, J. M.; Lue, C.-T., Kao, S. C. WO
03/008468.
10. Mihan, S.; Schmitz, H. WO 2005/103099.
11. Razavi, A. WO 2006/045738.
12. Hong, S. C.; Mihan, S.; Lilge, D.; Delux, L.; Rief, U. Polym. Eng. Sci. 2007, 47, 131.
13. Hlatky, G. G. Chem. Rev. 2000, 100, 1347.
14. Severn, J. R.; Chadwick, J. C.; Duchateau, R.; Friederichs, N. Chem. Rev. 2005, 105,
4073.
15. Sacchetti, M.; Pasquali, S.; Govoni, G. US 5698487.
16. Nakayama, Y.; Bando, H.; Sonobe, Y.; Kaneko, H.; Kashiwa, N.; Fujita, T. J. Catal.
2003, 215, 171.
17. (a) Severn, J. R.; Chadwick, J. C. Macromol. Rapid Commun. 2004, 25, 1024. (b) Severn,
J. R.; Chadwick, J. C. Macromol. Chem. Phys. 2004, 205, 1987.
18. Severn, J. R.; Duchateau, R.; Chadwick, J. C. Polym. Int. 2005, 54, 837.
19. Severn, J. R.; Kukalyekar, N.; Rastogi, S.; Chadwick, J. C. Macromol. Rapid Commun.
2005, 26, 150.
20. Severn, J. R.; Chadwick, J. C.; Van Axel Castelli, V. Macromolecules 2004, 37, 6258.
21. Huang, R.; Kukalyekar, N.; Koning, C. E.; Chadwick, J. C. J. Mol. Catal. A: Chem. 2006,
260, 135.
59
Chapter 3
22. (a) Nakayama, Y.; Bando, H.; Sonobe, Y.; Suzuki, Y.; Fujita, T. Chem. Lett. (Japan)
2003, 32, 766. (b) Nakayama, Y.; Bando, H.; Sonobe, Y.; Fujita, T. Bull. Chem. Soc. Jpn.
2004, 77, 617. (c) Nakayama, Y.; Bando, H.; Sonobe, Y.; Fujita, T. J. Mol. Catal. A:
Chem. 2004, 213, 141.
23. (a) Huang, R.; Liu, D.; Wang, S.; Mao, B. Macromol. Chem. Phys. 2004, 205, 966. (b)
Huang, R.; Liu, D.; Wang, S.; Mao, B. J. Molec. Catal. A: Chem. 2005, 233, 91. (c) Xu,
R.; Liu, D.; Wang, S.; Mao, B. Macromol. Chem. Phys. 2006, 207, 779.
24. (a) Enders, M.; Fernández, P.; Ludwig, G.; Pritzkow, H. Organometallics 2001, 20, 5005.
(b) Mihan, S.; Lilge, D.; De Lange, P; Schweier, G.; Schneider, M.; Rief, U.; Handrich,
U.; Hack, J.; Enders, M.; Ludwig, G.; Rudolph, R. US 6699948. (c) Enders, M.
Macromol. Symp. 2006, 236, 38.
25. Kukalyekar, N.; Huang, R.; Rastogi, S.; Chadwick, J. C. Macromolecules submitted.
26. Souane, R.; Isel, F.; Peruch, F.; Lutz, P. J. C. R. Chimie 2002, 5, 43.
27. Boor, J., Jr. Ziegler-Natta Catalysts and Polymerizations, Academic Press: New York,
1979.
28. (a) Guastalla, G.; Giannini, U. Makromol. Chem., Rapid Commun. 1983, 4, 519. (b)
Chadwick, J. C.; Morini, G.; Albizzati, E.; Balbontin, G.; Mingozzi, I.; Cristofori, A.;
Sudmeijer, O.; van Kessel, G. M. M. Macromol. Chem. Phys. 1996, 197, 2501.
29. Chu, K. J.; Soares, J. B. P.; Penlidis, A. Macromol. Chem. Phys. 2000, 201, 552.
30. (a) Parasu Veera, U.; Weickert, G. Polym. React. Eng. 2003, 11, 33. (b) Parasu Veera, U.;
McKenna, T.; Weickert, G. J. Sci. Ind. Res. 2007, 66, 345.
31. Peruch, F.; Cramail, H.; Deffieux, A. Macromolecules 1999, 32, 7977.
32. Coevoet, D.; Cramail, H.; Deffieux, A. Macromol. Chem. Phys. 1998, 199, 1451.
33. Huang, R.; Koning, C. E.; Chadwick, J. C. Macromolecules 2007, 40, 3021.
34. Engelen, Y. M. T.; Lemstra, P. J. Polym. Comm. 1991, 32, 343.
35. Lippits, D. R.; Rastogi, S.; Hohne, G. W. M. Phys. Rev. Lett. 2006, 96, 218303.
36. Lippits, D. R.; Rastogi, S.; Hohne, G. W. M. ; Mezari, B.; Magusin, P. C. M. M.
Macromolecules 2007, 40, 1004.
37. Rastogi, S. ; Lippits, D. ; Peters, G. W. M. ; Graf, R. ; Yao, Y. ; Spiess, H. W. Nature
Materials, 2005, 4, 635.
38. Hohne, G. W.H. Polymer, 2002, 43, 4689.
39. Thoda, A.; Hikosaka, M.; Yamada, K. Polymer, 2002, 43, 1667.
40. Britovsek, G. J. P.; Gibson, V. C.; Spitzmesser, S. K.; Tellmann, K. P.; White, A. J. P.;
Williams, D. J. J. Chem. Soc. Dalton Trans. 2002, 1159.
41. Huang, R.; Koning, C. E.; Chadwick, J. C. J. Polym. Sci.: Part A: Polym. Chem. 2007,
45, 4054.
42. Klimke, K. Ph.D. thesis, Mainz University, 2006.
43. Pollard, M.; Klimke, K.; Graf, R.; Spiess, H.W.; Wilhelm, M.; Sperber, O.; Kaminsky,
W. Macromolecules, 2004, 37, 813.
60
Chapter 4
Effect of long chains on the flow
crystallization of bimodal polyethylene
induced
Abstract
From Chapter 3 it was evident that co-immobilization of two different catalysts on a MgCl2
based support and subsequent ethylene polymerization resulted in the synthesis of intimate
bimodal blends of polyethylene. The synthesis route provided a blend of polyethylene having
two different molar masses, one component of which is intractable due to high molar mass, with
distinct differences in the relaxation times, mixed at the molecular length scale. These intimate
blends are model systems to study the influence of molar mass on crystallization and the
resultant morphology during shear flow. In this chapter we investigate the influence of flow
induced orientation of high molecular weight polyethylene from the bimodal reactor blend. Due
to distinct differences in the relaxation times, the influence of molar mass on crystallization and
the resultant morphology during shear flow has been investigated by time resolved X-ray
techniques. The observations are that highly stable shish structures are generated from high
molecular weight chains, providing nucleation sites for further crystallization. Shish structures
thus formed are found to be stable above the equilibrium melting temperature of unconstrained
extended chain crystals of polyethylene. It was observed that the shear flow, when applied at 142
°C, could generate a suspension of only shishes in the matrix of low molecular weight
polyethylene.
4.1 Introduction
Flow induced crystallization (FIC) in polymer melts is an important phenomenon, which
influences industrial processing techniques such as extrusion, injection and blow moulding and
consequently the resulting product properties. In polymer crystallization, it is well established
that a polymer molecule having chemical and structural regularity, when cooled, transforms itself
from a random coiled melt state to an ordered crystalline state.1 Since the early observations of
Treloar2 and Gent3, where they found that a crosslinked rubber when stretched induces
61
Chapter 4
crystallization, considerable efforts have been made to understand the crystallization caused due
to flow. Now it is proven and accepted beyond doubt that external flow causes orientation and
extension of polymer chains in the melt and thus affects the crystallization kinetics, structure and
morphology.4,5 The polymer, after experiencing external flow, shows two types of structures:
spherulites and shish-kebab.1,6-9 Spherulites are the spherical structures generated from crystal
lamellae growing radially outwards from a common centre. Despite detailed studies, the origin of
spherulitic growth including nucleation and formation of spherical aggregates of crystals is not
well understood.10-13 Shish-kebabs are crystalline assemblies consisting of a central core of
smooth extended chain crystals called shish and laterally grown folded chain crystals called
kebabs.4,5 The formation of oriented structures and kinetics of crystallization have direct
implications on polymer processing and properties, making the FIC studies essential.
Kebabs
Shish
(a)
(b)
Figure 4.1. Schematic representations of (a) spherulite and (b) shish-kebab structure.
Historically, shish-kebab structures were observed in a dilute solution and recognised as formed
due to orientation and flow; however, their origin was pointed out by Pennings14 in his
pioneering work on crystal formation in stirred solutions. Elongational flow in dilute solution
was chosen in order to minimize the interactions between polymer molecules and to maximise
the behaviour of individual chains during flow. Thereafter a lot of attention was paid to structure
development during elongational flow and shear flow. The shear flow was considered to be too
weak to extend polymer chains, thus was less investigated. However, now it is proven that even
very weak shear flow has an effect on polymer crystallization kinetics and morphology.15,16
Keller and Kolnaar4 in their review stated that shear flow, although not sufficient for full
extension of long polymer chains, can extend some parts of the chain, which can develop the
shish-kebab structures.
It was thought that long-chain polymer fractions were responsible for the formation of primary
nuclei in stressed melts, as the stretched long chains, due to their high relaxation time, remained
anisotropic for longer time and the shorter chains relaxed quickly when the flow was stopped.
Keller and co-workers proposed the coil-stretch transition concept17-21 to support the argument
behind the formation of primary nuclei. An abrupt change in birefringence, upon increasing the
strain rate in dilute polymer solution, indicated fully extended chains at a critical strain rate.
62
Flow induced crystallization in bimodal PE
These experimental evidences for coil stretch transition during crystallization in dilute solution
were instrumental for the hypothesis of shish, formed from extended chains, and kebabs, formed
from random coils. The existence of shish structures in polymer melts, resulting in shish-kebab
morphology, is very similar to what was observed in dilute solutions. These studies suggested the
existence of coil-stretch transitions in the entangled polymer melts. However, there is no
experimental evidence to prove this transition in concentrated solutions and entangled melts.
Extensive simulations by Muthukumar et al.22 show that even very high flow rates could not
extend all the chains and the coiled chains have a large effect on coil-stretch transition.
Molecular simulations showed that even in the monodisperse simulated polyethylenes both
stretched and coiled conformations existed at a given flow rate. They showed that two chains
having identical length could be stretched differently depending upon their initial conformation.
This observation suggests that polydispersity is not the requisite for shish-kebab formation as
thought earlier.
In recent years many research groups have extensively studied shear induced oriented structures
by X-ray techniques. Combined in-situ SAXS and WAXD experiments showed that the early
stage of shish-kebab formation is initiated by primary nucleation and the shish-kebab structures
influence the subsequent polymer crystallization, which involves un-stretched random coils.
Somani et al.23 reported that polymer chains in the shish are not fully stretched and were
entangled with neighbouring chains.
Short term shearing experiments are intriguing and can give considerable information about the
early stages of shish-kebab formation. Different groups applied different types of shear cells to
provide orientation in the melt of bimodal polyethylenes. Kornfield et al.24-26 have developed a
channel flow geometry apparatus and the group of Hsiao27 uses a rotary shear cell to investigate
morphology and structure development during flow. Kumarswamy et al. showed that the
formation of oriented nuclei due to shear depends on the distribution of relaxation time in the
melt.28 They further showed that shear orientation of polymer chains eliminates the free energy
barrier to facilitate nucleation. Seki et al.29 used solution blended bimodal blends of low and high
molecular weight polypropylenes to indicate that the role of long chains in shear induced
oriented crystallization is cooperative, enhanced by long chain-long chain overlap. Yang et al.30
studied shear induced precursor structures using binary blends of high (Mw= 250 kg/mol) and
low (Mw= 50 kg/mol) molecular weight polyethylene. They showed that the high molecular
weight component dominated the formation of crystallization precursor structure under shear,
which can act as template for shish-kebab structures. Zuo et al.31 used solution blended high
(Mw= 1500 kg/mol) and low (Mw= 50 kg/mol) molecular weight polyethylene to demonstrate
the thermal stability of shear induced shish-kebab precursors from high molecular weight
polyethylene chains. These polyethylene blends described above are solution blended and then
precipitated to obtain model systems for investigation of flow induced crystallization precursors.
63
Chapter 4
Solution blending is well known to achieve homogeneous blends; however, removal of traces of
solvent and improper mixing at molecular scale is sometimes observed. Due to the very high
melt viscosity of high molecular weight polyethylene, its extrusion or melt blending with low
molecular weight polyethylene is not possible. Considering the possible drawbacks in post
synthesis blending of polymers, the preparation of miscible reactor blends of high and low
molecular weight polyethylenes was considered to be an effective approach to study flow
induced crystallization.
In this study, a miscible reactor blend of high and low molecular weight is used to investigate the
effects of high molecular weight chains on crystallization induced by flow.
4.2 Experimental
4.2.1 Materials
An intimate blend of high molecular weight (H-MW, Mw = 1.1 x 106 g/mol) and low molecular
weight (L-MW, Mw= 5.5 x 104 g/mol) polyethylene (bimodal PE) was synthesised using the
procedure described in Chapter 3. For comparison as a standard sample, low molecular weight
(L-MW, Mw= 5.5 x 104 g/mol) PE was also synthesised. Both bimodal and monomodal L-MW
samples were stabilised with 0.25 wt % Irganox 1010, added as a solution in acetone. The
acetone was evaporated and the samples were dried in a vacuum oven at 60 °C for about 16
hours.
4.2.2 GPC characterization
Molecular weight and molecular weight distributions of the resulting polymers were determined
by means of size exclusion chromatography on a PL-GPC210 at 135 °C using 1,2,4trichlorobenzene as solvent.
4.2.3 X-ray characterization
Polyethylene samples (beads obtained from the reactor) were cryo-ground to fine powder and
melt pressed in a compression press at 180 °C and 200 bar pressure. Discs of about 1 mm
thickness and 25 mm diameter were cut from the compression moulded sheets and used for
SAXS and WAXD studies.
A Linkam shear cell CCS450, equipped with a heating system capable of controlled heating, was
used to apply step shear to the molten sample. The sample disc was placed in between two
Kapton windows of the shear cell. One plate is a fixed stage and the other, which is equipped
with a motor, is a rotating stage. This shear cell assembly was then placed on an X-ray beamline
64
Flow induced crystallization in bimodal PE
to perform further experiments. The experimental setup is shown in Figure 4.2 and the shear cell
details are depicted in Figure 4.3.
A
B
Figure 4.2. Experimental setup for WAXD experiments at ESRF, Grenoble. ‘A’ denotes shear
cell aligned with x-ray beam and ‘B’, the detector.
Sample
X-Ray
Figure 4.3. Schematic representation of a Linkam shear cell with a sample in an X-ray beam.
(Reproduced from Somani et al. Macromolecules, 2000, 33, 9385)
WAXD experiments, performed in two sets of experiments, were carried out on beamline ID 11
in the European Radiation Synchrotron Facility in Grenoble. For the first set of experiments, the
wavelength of the synchrotron radiation was 0.427 Å, the sample to detector distance was 280
mm and the 2D WAXD patterns were recorded on a Brucker X-ray detector which had a
resolution of 1024x1024 pixels. For the second set of WAXD experiments, the wavelength of the
synchrotron radiation was 0.495 Å, the sample to detector distance was 259 mm and the patterns
were collected by a Frelon X-ray detector with a resolution of 1024x1024 pixels. The diffraction
angle was calibrated with Lab6 and the images were corrected for background and air scattering.
4.2.4 Differential Scanning Calorimetry (DSC)
Thermal analysis was performed on a Q 1000 DSC from TA Instruments. In order to fractionate
two different crystalline structures, namely extended chain crystals and folded chain crystals, a
65
Chapter 4
slow thermal protocol was applied. A polymer sample (2-3 mg) was cut from the disc, on which
shear experiments had been performed, and in a DSC heated to 90 °C at a rate of 10 °C/min.
Thereafter, a heating rate of 1 °C/min was applied up to 160 °C, followed by cooling to room
temperature at a rate of 10 °C/min. Similar experimental conditions were applied in the second
heating cycle.
4.2.5 Experimental Protocol for flow induced crystallization
Before starting the experiments, the shear cell was calibrated for the experimental temperature
range. The samples were heated to 230 °C, at a heating rate of 30 °C/min, and held there for 2
minutes in order to erase the memory effects from previous thermal and processing history. The
sample was cooled first to 142 °C at a rate of 20 °C/min and then to 137 °C at a rate of 10 °C/
min. A slow cooling rate was applied in the second stage to avoid undercooling, keeping in mind
that polyethylene crystallizes very rapidly. At 137 °C a WAXD pattern of the melt was collected
before applying shear, after which the development of crystallinity was followed with WAXD
under isothermal conditions at 137 °C. The experimental protocol is featured in Figure 4.4. The
flow conditions were chosen considering the limitations of the Linkam Shear Cell modified with
Kapton windows. For the series of experiments the total strain was kept constant at 60 units. The
strongest shear applied was 60 s-1 for 1 s and the weakest flow was 1 s-1 for 60 s. The data
acquisition time was 1 s for each frame of scattering pattern.
Figure 4.4. Experimental protocol.
4.3 Results and Discussion
To establish fundamental concepts on the influence of molar mass on crystallization and the
resultant morphology during shear flow, a study on a polymer having two different molar masses
with distinct differences in the relaxation times, mixed at the molecular length scale, is a
requisite. There are different views on the role of long chains in the formation of shishes.
66
Flow induced crystallization in bimodal PE
As discussed in Chapter 3, the synthesis route makes use of an immobilised dual catalyst system.
The resulting intimate blends of PE are composed of distinctly different high and low molar
mass, due to the synthesis carried out using a co-immobilized catalyst combination. These
intimate blends are ideal model systems to study the effect of long chains in shear induced
crystallization. It is apparent from the results discussed in Chapter 3 that changing the catalyst
loading on the support facilitates the production of bimodal polyethylene differing in the
proportions of the high and low molecular weight components.
In order to perform the shear induced crystallization experiments, a bimodal blend was
synthesized by co-immobilizing two precatalysts: 0.18 µmol [1-(8-quinolyl)indenyl]CrCl2 and 1
µmol of a bis(imino)pyridyl iron catalyst {2,6-[ArN=C(Me)]2C5H3N}FeCl2, where Ar = 2chloro-4,6-dimethylphenyl, on 100 mg of a MgCl2/AlEtn(OEt)3-n support. Details of the
immobilization procedure and ethylene polymerization conditions are reported in Chapter 3. The
resulting reactor blend, here onwards denoted as BM-PE, comprised of 7 wt % of H-MW (Mw =
1.1 x 106 g/mol, PDI = 2.0) and 93 wt % L-MW polyethylene (Mw = 5.5 x 104 g/mol, PDI = 3.8).
The molecular weights of the H-MW and L-MW components of BM-PE reported here are a little
lower than those given in Chapter 3. These numbers of molecular weight and molecular weight
distribution were obtained by fitting the GPC data using a log normal distribution model. This
blend, containing only 7 wt % of H-MW, was chosen taking into account the fact that high melt
viscosity of the high molar mass component hinders efficient stress transfer during shearing in a
Linkam shear cell. The concentration of high molecular weight chains in this sample is
sufficiently higher than the critical overlap concentration, which is estimated to be 0.5 wt%.32,33
As a reference sample in which the effect of long chains can be excluded, a low molecular
weight monomodal polyethylene similar to the L-WM component in the bimodal blend was
synthesized by singly immobilizing the iron catalyst on the MgCl2/AlEtn(OEt)3-n support.
Small amplitude oscillatory shear experiments were performed on the thus synthesized bimodal
blend and the monomodal low molar mass PE to determine their viscoelastic response in the melt
state. As discussed in Chapter 3, expected trends in the storage modulus and phase angles
confirmed the miscibility of the H-MW and L-MW components in the bimodal blends.
To investigate the influence of shear on molecular orientation, a crystallization temperature
higher than the normally observed melting temperature (135 °C) for linear polyethylenes
crystallized from an isotropic melt was chosen. What follows are the time resolved X-ray
experiments performed at temperatures higher than 135 °C.
67
Chapter 4
4.3.1 Flow Induced Crystallization at 137 °C
The molecularly miscible bimodal blend of polyethylene having two components differing in
their relaxation times provides a valuable opportunity to follow the influence of flow on
crystallization behaviour. This is followed by performing time resolved X-ray synchrotron
experiments. Because of the nearly 20-fold difference between the molecular weights of the two
components in the blend, the relaxation time of the H-MW chains is approximately 10000 times
higher than that of the L-MW chains. This “bimodal” relaxation time generates different
constraints to the same long molecule. The constraints provided by physical entanglements
between L-MW and H-MW chains will be more dynamic and short lived, compared to the
entanglements between H-MW and H-MW chains. Thus the coil to stretch transition during flow
will be influenced by the molar mass between the long and short lived entanglements. The stress
transferred to the long molecules during shear will vary because of the different constraints
provided by the long and short lived entanglements. The difference in constraint release
mechanism is proposed by Marrucci34 and is in agreement with the theoretical concepts proposed
by Muthukumar.35-37
A large degree of undercooling, which surpasses the nucleation barrier, promotes the formation
of crystals in an isotropic melt. The nucleation barrier can also be suppressed by the application
of constraints, especially on coil to stretch transition under flow. Constrained conformations in
the stretched chains will favour nucleation even at higher temperatures than normally observed
in an isotropic melt. This results in the origin of shishes, which provides nucleation sites for the
crystal growth.
According to the classical thermodynamics,38 the driving force for nucleation is the reduction of
Gibbs free energy (G): ∆G < 0. For quiescent crystallization, crystallization leads to the folded
chain
lamellae
and
the
corresponding
free
energy
change
is
where ∆H is the latent heat and
is the equilibrium
melting temperature for the unconstrained extended chain crystals (141.5 °C for polyethylene).
The thickness of the folded chain lamellae (l) can be determined by the Gibbs-Thomson
equation.
(4.1)
T is the melting temperature determined by DSC and
is the surface energy of the fold plane.
From the Gibbs-Thomson equation it is known that above
chain folding is not possible, 141.5
°C being the upper temperature limit for chain folding.
Considering these facts, temperature conditions for shearing experiment were chosen such that
only the H-MW fraction of the blend could crystallize, while the L-MW fraction would remain in
68
Flow induced crystallization in bimodal PE
the melt state. Thus the biphasic anisotropic system would resemble the dilute polymer solution,
in which crystallisable longer polymer chains were suspended in the non crystallisable low
molecular weight polymer matrix. In order to satisfy this condition, two temperature conditions
were defined for these experiments: 1) 137 °C, which is higher than the normal melting
temperature of polyethylene and less than
; 2) 142 °C, which is much higher than the normal
melting temperature of PE and a little above . In this section the experiment carried out at 137
°C is dealt with and experiments performed at 142 °C will be discussed in the following section.
In order to define the experimental temperature, below the melting temperature of extended
chain crystals, several shearing experiments were performed at different temperatures. The
experimental protocol was the same as described in the experimental section, while temperatures
at which the samples were subjected to flow were varied. Figure 4.5 compares the crystallinity
obtained after applying a shear flow of 30 s-1 for 2 s under isothermal conditions at 123, 133 and
137 °C. Without going into depth on the effect of temperature, some of the key findings are
reported here.
50
Crystallinity (%)
40
30
20
10
0
0
50
100
150
200
250
Time (sec)
Figure 4.5. Crystallinity development under isothermal conditions after shearing at 123 °C (●),
133 °C (■) and 137 °C (▲).
From Figure 4.5 it is apparent that the crystallinity, in a sample where flow was applied at 137
°C, is much lower than the other two instances at 133 °C and 123 °C. As discussed earlier, upon
flow at higher temperature (near ) generation of the isotropic nucleation sites is unfavourable,
while on decreasing the temperature the possibility of the formation of isotropic nucleation sites
increases. Thus at 137 °C, crystallinity is observed mainly due to the shish type of oriented
structures, whereas at 133 °C and 123 °C higher crystallinity is observed due to the combination
of anisotropic and isotropic crystallization. It is to be noted that in the sample at 137 °C, the
anisotropic structure arises first followed by the isotropic X-ray pattern. The two dimensional
WAXD patterns collected at t = 150 s is shown in Figure 4.6. The crystalline WAXD reflections
arising from the (110) and (200) orthorhombic crystal planes show the presence of orientation.
The oriented structure once obtained does not change within the experimental time. This
69
Chapter 4
suggests the formation of long lived crystalline structures that are stable even at temperatures as
high as 137 °C after the application of flow. Unlike crystallization at 137 °C, isotropic
crystallization occurs, prior to the application of shear, while cooling the sample from melt to the
lower isothermal crystallization temperatures, 133 °C and 123 °C. On the application of shear at
these low crystallization temperature, partially oriented structures are obtained. In Figure 4.6,
weak but narrow equatorial reflections, for the sample sheared at 137 °C, suggest lower
crystallinity arising due to oriented structures. On the other hand, for the other two samples,
strong reflections appear due to high crystallinity; however, the high span of the arcs is indicative
of isotropic crystallization. Crystallization prior to flow complicates the desired study on the
influence of high relaxation times provided by the high molar mass content in the blend.
137 °C
133 °C
123 °C
(110)
(200)
Figure 4.6. WAXD patterns showing inner (110) and outer (200) crystalline reflections.
These preliminary observations prompted the selection of 137 °C as a suitable temperature, as a
temperature condition below , to investigate the effect of H-MW chains on the crystallization
induced by flow. The experimental protocol for FIC was applied as described in the experimental
section, 4.2.4. Crystallinity was determined by subtracting the amorphous halo from the total
WAXD intensity, leaving two peaks due to (110) and (200) crystalline reflections. Figures 4.7
and 4.8 summarise the WAXD results of experiments performed on bimodal PE at 137 °C and at
different shear rates. Figure 4.7 compares the evolution of crystallinity under isothermal
conditions after shearing at 137 °C at varied applied shear rates. No evolution of crystallinity was
observed, independent of the shear rate applied, for monomodal low molecular weight
polyethylene at 137 °C. Figure 4.8 shows two dimensional WAXD patterns for the experiments
described in the Figure 4.7.
70
Flow induced crystallization in bimodal PE
4.0
3.5
Crystallinity (%)
3.0
2.5
2.0
1.5
1.0
0.5
0.0
0
50
100
150
200
250
300
350
400
Time (sec)
Figure 4.7. Evolution of crystallinity at 137 °C obtained after shearing at 60 s-1/ 1 s ( ), 30 s-1/
2 s (■), 15 s-1/4 s (▲) and 1 s-1 for 60 s (●).
1 s-1 / 60 s
15 s-1 / 4 s
30 s-1 / 2 s
60 s-1 / 1 s
88 s after step shear
220 s after step shear
(110)
(200)
Figure 4.8. WAXD patterns collected after shearing at varied shear rates at 137 °C.
As described in the experimental section, the shear rate was varied, nevertheless keeping the total
strain constant for a set of experiments. Four different shear rates were applied: a shear rate of 60
s-1 was applied for 1 s, 30 s-1 for 2 s, 15 s-1 for 4 s and 1 s-1 for 60 s. In Figure 4.7 evolution of
crystallinity at 137 °C due to four different flow conditions is compared as a function of time.
After applying a step shear when the rotating motor of the shear cell comes to a halt, that time is
denoted as time t=0. Diffraction patterns were collected at the fixed temperature at which the
step shear is applied. The experimental temperature being higher than the normal melting
temperature of polyethylene, the observed crystallinity appears to be quite low, the highest being
3.5 % for the strongest shear flow. From Figure 4.7, it is obvious that for a constant strain the
crystallinity increases with the shear rate, while it remains independent of the shear time. This
suggests that the shear rate is more dominant than the shear time in the formation of oriented
structures. As mentioned earlier, the crystallinity was determined from the integrated intensity of
crystalline reflections arising from (110) and (200) orthorhombic crystal planes. With increasing
71
Chapter 4
shear rate, increased intensity of the equatorial reflections is apparent from Figure 4.8. From
these observations it is evident that, after cessation of the flow, formation of the anisotropic
nuclei and growth of the oriented structures occurs. However, WAXD is less sensitive for the
investigation of the nature of oriented nuclei and to follow the growth of kebab.
To follow the formation of the shish-kebab structures, SAXS experiments were performed
keeping all experimental conditions the same as were applied during the WAXD experiments.
Figure 4.9 shows the SAXS fingerprints of the FIC experiments for all four shear conditions.
Figure 4.9 showing the presence of only equatorial signals at t = 0, just after the step shear
suggests the immediate formation of shish structures. After some time the meridional reflections
have appeared indicating the growth of kebab structures on top of the shishes. With time the
intensity of the meridional reflections tend to increase, which is suggestive of the growth of
kebabs on shishes.
Figure 4.9. SAXS fingerprints for two flow conditions at initial stages of the experiments.
For these two flow conditions, 15 s-1 for 4s and 60 s-1 for 1s, the difference in the intensities of
both equatorial and meridional reflections is observed. In the case of strong shear conditions,
more intense reflections are observed. This difference can be attributed to the more number of
shish and kebab structures generated due to the stronger flow.
These results are in accordance with the observations reported by Hsiao et al.31 In their bimodal
blend prepared by solution blending of 2 wt% of crystallizable H-MW PE with 98 wt% of non
crystallizable L-MW PE, development of anisotropic structure upon shearing was observed.
Their findings suggest formation of shishes followed by kebabs, during isothermal crystallization
after the application of similar step shear at 127 °C. However, there are also some differences in
the observations. Hsiao et al. investigated the effect of H-MW chains on the FIC by applying
72
Flow induced crystallization in bimodal PE
step shear at 127 °C, while in the present study isotropic crystallization was apparent even at 133
°C before applying the step shear.
4.3.2 Degree of orientation
The degree of orientation of lamellar crystals can be expressed in terms of Herman’s orientation
function. Taking the flow direction as a reference direction, crystalline orientation can be
characterised by the average orientation of normal to the crystalline plane. Herman’s orientation
function (f) can be mathematically calculated by following equation.
ff =
3< cos2 θ > - 1
2
(4.2)
where θ is the azimuthal angle of scattering intensity.
Based on this equation three major ‘f’ values can be deduced: When all the crystals are oriented
in the direction parallel to the flow direction, then θ = 0° making f = 1. When the crystals are
oriented perpendicular to the flow direction (θ = 90°), f has a value of -0.5. Random orientation
of crystals gives f = 0. These three possibilities are illustrated in Figure 4.10.
f=1
f = -0.5
f=0
Figure 4.10. Illustration of crystal orientation and Herman’s orientation function.
The average intensity was calculated from the integration of intensities of (110) and (200)
reflections in the WAXD pattern. Figure 4.11 compares Herman’s orientation functions for BMPE at four different flow conditions as a function of time. As described in the experimental
section, time zero is the time when the flow is stopped; in other words, when the rotating motor
of the shear cell comes to a halt.
The first point to be noted is that the data points are very much scattered due to the lower
crystallinity at this temperature. Under these experimental flow conditions, the highest
crystallinity obtained was approximately 4% by the strongest flow of 60 s-1/1s. In the SAXS
patterns at these conditions both equatorial streaks and meridional lobes were observed; this
indicated the presence of both shishes and kebabs at 137 °C.
73
Chapter 4
1.4
1.2
1.0
0.8
f
0.6
0.4
0.2
0.0
0
50
100
150
200
250
300
350
Time (s)
Figure 4.11. Herman’s orientation function at 137 °C at four different flow conditions, Shear
Rate / ts: (●) 1s-1/ 60 s, (▲) 15 s-1 /4 s, ( ) 30 s-1 /2 s, () 60 s-1 /1 s.
As shown in Fig. 4.11, for the weakest flow of 1 s-1/60 s, f values remain around 0.5 throughout
the experiment and upon increasing the shear f reaches 1. Herman’s orientation function f ≈ 0.5
indicates that several crystals are oriented in the direction of flow and many of the crystals are
still randomly oriented. So, although the weakest of flows has resulted in generation of crystals,
these crystals are not perfectly oriented along the flow direction. The strongest flow of 60 s-1/1 s
shows fewer fluctuations in the data points as they lie in the range of f = 0.9 to 1.0. This
observation suggests that almost all crystals, which are formed due to a flow of 60 s-1/1 s, are
oriented along the shear direction.
4.3.3 Thermal Analysis
From the FIC experiments, it is evident that the H-MW chains are responsible for the generation
of shishes in the BM-PE sample. The absence of any shear induced orientation in L-MW-PE
confirmed the absence of extended chain crystals. Due to the higher melting temperature of
extended chain crystals than folded chain crystals, DSC evaluation of sheared and un-sheared
samples is expected to show differences in the melting thermograms. As reported in the
experimental section, DSC evaluation was performed at lower heating rates in order to resolve
the melting peaks of folded and extended chain crystal and to minimize the superheating effect
due to higher heating rates. In Figure 4.12, DSC melting thermograms of quiescently crystallized
and shear induced crystallized L-MW PE and BM-PE are depicted.
In Figure 4.12, L-MW-PE and BM-PE samples are the quiescent crystallized samples while the
sheared samples experienced the shear of 60 s-1 for 1 s at 137 °C. For the L-MW-PE sample,
there was not a considerable difference in the DSC thermograms of sheared and un-sheared
samples.
74
Flow induced crystallization in bimodal PE
Figure 4.12. DSC thermograms of quiescently crystallized and shear induced crystallized L-MW
PE and BM-PE.
For both samples only one melting peak was apparent at 133 °C, which is the normal peak
melting temperature of low molecular weight melt crystallized PE. This observation confirms the
presence of only one type of crystal population, folded chain crystals. In the case of un-sheared
BM-PE also one melting peak was observed. The observed peak melting temperature was 132.6
°C. However, the melting thermogram of the sheared sample showed two well resolved melting
peaks, at 132.8 °C and 138 °C. The end melting temperature was 139.5 °C. This second peak,
which is close to the equilibrium melting temperature of unconstrained extended chain crystals,
can be attributed to the formation of extended chain crystals arising on the application of flow
from the crystallization of the H-MW PE component present in the bimodal blend.
The formation of shear induced shish-kebab structures from H-MW chains at 137 °C can be
translated to the extensional flow to form the oriented tapes. By drawing the melt crystallized
and compression pressed bimodal PE strip at elevated temperature, highly oriented high-strength,
high-modulus tapes can be prepared. This possibility is investigated in detail in Chapter 5.
4.3.4 Suspension of only shishes in the matrix of low molecular weight PE
Another temperature condition defined for the shear induced crystallization was 142 °C, just
above the equilibrium melting temperature, where chain folding is not possible. SAXS and
WAXD experiments were performed the similar way as were done at 137 °C. From the SAXS
patterns obtained, it was apparent that immediately after cessation of the flow a streak of
intensity appeared at the equatorial region. This presence of streak and absence of meridional
kebabs pointed out the formation of only shishes and the absence of kebabs. Under the same
conditions WAXD images show an orthorhombic lattice with a high degree of orientation. In
Figure 4.13, SAXS intensity at equatorial region is plotted against time and it is compared with
75
Chapter 4
0.4
1.0
0.3
0.8
0.2
0.6
0.1
0.4
0.0
0.2
-0.1
0
200
400
600
800
1000
1200
1400
Crystallinity (%)
Intensity Equator (a.u.)
the crystallinity obtained from WAXD. In Figure 4.14, the two dimensional WAXD and SAXS
images obtained are shown.
0.0
1600
time (s)
Figure 4.13. SAXS intensity at the equatorial region (●) and the crystallinity obtained from
WAXD (○) as a function of time.
The integrated intensity of the equatorial reflection in SAXS showed an overshoot immediately
after flow was stopped. After the overshoot the integrated intensity decayed exponentially. After
160 s the integrated intensity was nearly constant. However, compared to SAXS, crystallinity
obtained from the WAXD pattern showed a different trend. WAXD crystallinity built up
gradually and after around 160 s reached a plateau value of 0.4 %.
Equatorial streak
(200)
(110)
Figure 4.14. SAXS and WAXD images at different time intervals after application of shear of
120 s-1 for 1 s at 142 °C.
76
Flow induced crystallization in bimodal PE
Analysis of the SAXS images revealed that the intensity of the equatorial streak varied over the
experimental time, while the shape remained the same. These observations show that the flow
triggered the formation of needle-like precursors which varied in their size. The crystallinity
apparent from theses precursors, which are aligned along the flow direction, is quite low.
Decrease in the intensity to half of the original value suggests that half of the precursors, which
are too small to be stable at such a high temperature, have melted during the first 160 s of the
experiments.
As pointed out earlier, according to the tube model, which describes the linear chain dynamics,
because of the difference in the molecular weights, H-MW chains possess relaxation time
times longer than the L-MW chains. In the present bimodal blend,
due to fast chain dynamics, the lifetime of entanglements between L-MW and H-MW chains is
short. Release of the topological constraints is fast, leading to the reduction of relaxation time of
H-MW chains. This disengagement time of a chain with Z number of entanglements can be
calculated using the Doi-Edwards model including contour length fluctuations.
(4.3)
where
is the equilibrium time (7· 10-9 s at 190 °C). The value of
for H-MW PE is 140 s and
for L-MW PE is 0.01 s.
The time scale for their dissolution (t = 160 s) resembles the disengagement time (
=140 s) of
the high molecular weight chains. This analysis strongly suggests that the initial needle-like
precursors, which are part of the high molecular weight backbone, dissolve by a reptation
mechanism. However, the initial build up in the crystallinity suggests that the remaining
precursors refine their structure to a proper crystal lattice. For the following 400 s, these
scaffolds of shishes appear to be stable as suggested by the plateau observed in both SAXS
intensity and WAXD crystallinity. These shishes, which are stable for a sufficiently long time,
suggest that they are extended chain crystals. However, after approximately 600 s, these shishes
start melting which is apparent from the decrease in the SAXS intensity and WAXD
crystallinity. The process of crystallization and/or dissolution of precursors formed due to flow is
depicted in Figure 4.15.
alliza
Cryst
Metastable Precursor
D is
tion
solu
tion
Figure 4.15. Crystallization and dissolution of metastable precursors.
77
Chapter 4
4.4 Conclusion
Miscible bimodal blends of high and low molecular weight polyethylene are model systems for
investigation of the effect of long molecules on crystallization induced by flow. The application
of shear flow to a bimodal polyethylene melt, at a temperature above the equilibrium melting
temperature of unconstrained extended chain crystals of polyethylene, leads to the formation of a
suspension of shish structures. This is the first demonstration of the formation of only shish
structures in a matrix of low molecular weight polyethylene. During flow, metastable needle-like
precursors are formed; of these, the precursors which are too small dissolve on a timescale that
correlates with their reptation time, whereas precursors exceeding a critical size crystallize to
form the shish structure. On cooling, the shish structure can act as nucleant for further
crystallization, giving a shish kebab crystalline morphology. Shear rate is more dominant than
shear time in the formation of oriented structure, which results from flow-induced orientation of
the high molecular weight chains present in blends of high- and low-molecular weight
polyethylene.
One-pot synthesis of bimodal PE via co-immobilization of two different catalysts on a single
support provides a realistic alternative to the solution blending approach hitherto used in most
previous investigations of the effect of bimodality on the formation of shish-kebab structure in
polyethylene.
4.5 References
1.
2.
3.
4.
Wunderlich, B. Macromolecular Physics Vol 2 Academic Press 1976.
Treolar, L. R. G. Trans. Faraday. Soc. 1941, 37, 84.
Gent, A. N. Trans. Faraday. Soc. 1954, 49, 521.
Keller, A; Kolnaar, H. W. H. In: Meijer, editor. Processing of Polymers Vol 18 Chapter
4.
5. Ward, I. M. Structure and Properties of Oriented Polymers, Wiley: New York, 1975.
6. Pennings, J. ; V.d. Mark, J. M. A. A.; Kiel A. M. Kolloid ZZ Polym. 1970, 237, 336.
7. Pennings, J. ; Kiel A.M. Kolloid ZZ Polym, 1970, 205, 160.
8. Keller, A.; Hikosaka, M.; Rastogi, S.; Toda, A.; Barham, P. J.; Goldbeck-Wood, G. J.
Mat. Sci. 1994, 29, 2579.
9. Keller, A.; Hikosaka, M.; Rastogi, S. Phys. Sci. 1996, T66, 243.
10. Bassett, D. C.; Vaughan, A. S. Polymer 1985, 26, 717.
11. Varga, J. J. Mat. Sci. 1992, 27, 2557.
12. Abo el Maaty, M. I.; Hosier, I. L.; Bassett, D. C. Macromolecules, 1998, 31, 153.
13. Keith, H. D.; Padden, F. J. J. Appl. Phys. 1963, 34, 2409.
14. Pennings, J. A.; Kiel, A. M. Kolloid-Z. Z. Polym. 1965, 205, 160.
15. Somani, R.; Yang, L.; Zhu, L.; Hsiao, B. Polymer, 2005, 46, 8587.
78
Flow induced crystallization in bimodal PE
16. Kumaraswamy, G. J Macromol. Sci. Part C: Polym. Rev. 2005, 45, 375.
17. Keller, A.; Machin, M. J. J. Macromol. Sci. Phys. 1967, B1(1), 41.
18. Hill, M. J., Keller, A. J. Macromol. Sci. Phys. 1969, B3(13), 153.
19. Hill, M. J., Keller, A. J. Macromol. Sci. Phys. 1971, B5(3), 591.
20. Keller, A. Mackley, M. R. Pure Appl. Chem. 1974, 39, 195.
21. Miles, M. J.; Keller, A. Polymer, 1980, 21, 1295.
22. Dukovski, I.; Muthukumar, M. J. Chem. Phys. 2003, 118, 6648.
23. Somani,r.; Yang, L.; Hsiao, B.; Sun, T.; Pogodina, N. V.; Lustiger, A. Macromolecules,
2005, 38, 1244.
24. Kumaraswamy, G.; Verma, R. K.; Kornfield, J. A Rev. Sci. Instrum. 1999, 70, 2097.
25. Kumaraswamy, G.; Issaian, A. M.; Kornfield, J. A Macromolecules. 1999, 32, 7537.
26. Kornfield, J. A; Kumaraswamy, G.; Issaian, A. M. Ind. Eng. Chem. Res. 2002, 41, 6383.
27. (a) Somani, R.; Hsiao, B.; Nogales, A.; Srinivas, S.; Tsou, A. H.; Sics, I.; Balta-Kalleja,
F. J.; Ezquerra, T. A. Macromolecules, 2000, 33, 9385. (b) Somani, R.; Yang, L.; Hsiao,
B.; Agarwal, P.; Fruitwala, H.; Tsou, A. H.; Macromolecules 2002, 33, 9096.
28. Kumaraswamy, G.; Kornfield, J. A.; Yeh, F.; Hsiao, B. S. Macromolecules 2002, 35,
1765.
29. Seki, M.; Thurman, D. W.; Oberhauser, J. B.; Kornfield, J. A. Macromolecules, 2002, 35,
2583.
30. Yang, L.; Somani, R.; Sics, I.; Hsiao, B.; Kolb, R.; Fruitwala, H.; Ong, C.
Macromolecules 2004, 37, 4845.
31. Zuo, F.; Keum, J. K.; Yang, L.; Somani, R.; Hsiao, B. Macromolecules 2006, 39, 2209.
32. de Gennes, P. G. Scaling Concepts in Polymer Physics 1979, Cornell University Press,
NY.
33. Ballard, D. G. H.; Cheshier, P.; Longman, G. W.; Schelten, J. Polymer, 1978, 19, 379.
34. Marrucci, G. J. J. Non-Newtonian Fluid Mech. 1996, 62, 279.
35. Muthukumar, M.; Baumgärtner, A. Macromolecules, 1989, 22, 1937.
36. Muthukumar, M.; Baumgärtner, A. Macromolecules, 1989, 22, 1941.
37. Baumgärtner, A.; Muthukumar, M. Adv. Chem. Phys. 1996, 94, 625.
38. Muthukumar, M. Adv. Chem. Phys. 2004, 128.
79
Chapter 4
80
Chapter 5
Towards ultimate properties of bimodal polyethylene
Abstract
In this chapter, which is the final link of the chain of knowledge in this thesis, the previously
gained knowledge of tailored polymer composition and flow induced crystallization is linked
together to achieve the ultimate mechanical properties. Several synthesized grades of bimodal
polyethylenes, differing in the proportions of the high and low molecular weight components,
were used for solid state drawing in order to generate strong and stiff tapes. During solid state
drawing at 137 °C, orientation of the high molecular weight chains was found to be assisted by
the low molecular weight matrix. A strong effect of the amount of high molecular weight
component on the mechanical properties of bimodal PE was established. Preliminary
experiments indicated improved mechanical properties for the solid state drawn PE tapes.
5.1 Introduction
Linear polymers, due to the strong covalent bonding, are potentially strong and stiff in nature.
This intrinsic strength and stiffness can be effectively utilized to improve the mechanical
properties of the flexible chain polymers, such as polyethylene, by chain extension and
alignment. When all the molecules in a polymeric chain are well aligned, so the covalent bonds
lie along the fibre direction, then the macroscopic fibre, along the fibre direction, can have the
properties approaching those of the covalent bonds. Treloar1 in 1960 introduced the concept of
extended chains and he calculated the stiffness of PE chains to be 182 GPa. Since then several
researchers used quantum chemistry, X-ray and neutron diffraction techniques to arrive at the
theoretical numbers for the stiffness of PE of up to 340 GPa and a strength of 19 GPa.2 However,
due to the scale difference between an individual chain used in simulations and the real fibre, the
calculated theoretical values differ greatly from the experimental values. Nevertheless, these
predictions were successful in triggering research in the area of oriented polymers.
In practice, to orient flexible polymer systems, the polymer can either be deformed in the solid
state or crystallized in the oriented form after shear or elongational flow to retain extended chain
81
Chapter 5
morphology. Crystallization of polymers in the oriented form is achieved by spinning either in
the melt state or solution; whereas solid-state deformation is achieved by drawing, (ram)
extrusion or a combination of both.5 Since the 1970s, a tremendous amount of research has been
done in the area of oriented polyethylene and several routes have been adopted to obtain highstrength and high-modulus fibres. Pioneering work in solid state drawing was reported by the
group of Ward,3 while high strength fibres from melt processing originated from Keller’s group
and were developed by Odell and Bashir.4 The commercially available solution-spun fibres,
which were based on the solution process reported initially by Pennings and co-workers,5 were
developed by Smith and Lemstra.6Before reviewing earlier work, we may recall a few basic
concepts on deformation behaviour of polymers.
5.1.1 Deformation of polymers
Load
It is well established that thermoplastic polymers can be drawn at room temperature (cold
drawing). However, not all polymers show cold drawing. The solid line in Figure 5.1(a)
represents a typical stress-strain curve of cold drawn polymer and the dotted line represents the
stress-strain curve of a redrawn polymer. Figure 5.1(b) shows a schematic representation of
alignment of crystals effected due to the applied load in a semicrystalline polymer.
B
D
C
A
A
Extension
(a)
B
C
D
(b)
Figure 5.1 (a) A typical stress-strain curve of cold-drawn ( ) and a redrawn ( ) polymer.
(b) Alignment of chains in polymer crystals.7 The notations A-D show the regions on the stressstrain curve in Figure 5.1(a).
82
Towards ultimate properties of bimodal polyethylene
Initially, from point A, the stress rises linearly with strain. At point B, the nominal stress reaches
a maximum and then it starts to fall and the sample thins, forming a neck. Point B is the yield
point, after which the deformation is non-recoverable. Point C shows a minimum stress point, at
which the strain equals its natural draw ratio. Then the neck propagates and occupies its whole
length before the specimen fractures. In a semi-crystalline polymer, the unfolding of chains
begins in the amorphous regions between the lamellae of the crystals. This is followed by
breaking-up and alignment of crystals, as shown in Figure 5.3(b). In amorphous polymers,
yielding occurs by molecular uncoiling. At the yield point, a neck forms which is followed by an
overall drop in stress. In the neck region, the folded chains become aligned. Macroscopically,
because of the thinning down in cross section, the stress rises locally and any deformation occurs
preferentially there. This helps the neck to propagate along the waist of the specimen under a
steady load, in a process known as cold drawing. When the elongated part of a cold drawn
sample is extended again, the stiffness increases remarkably, apparent from the greater slope of
the dashed line in Figure 5.3(a). Redrawn specimens show brittle failure, which is a
characteristic fracture mode of oriented polymers. This increase in stiffness due to orientation of
chains is the basis of oriented fibre technology, which is used in the present chapter.
5.1.2 Orientation, desired morphology and the ultimate properties
It is known that, in a flexible polymer, the strength and stiffness comes as a consequence of
oriented crystals. Lamellar polyethylene is associated with low stiffness and various drawing
techniques are used to increase the modulus by extending the polymer chains.8 Upon flow in the
molten state and subsequent crystallization followed by hot drawing, oriented structures such as
shish and interlocked kebabs are reported in the literature.4 Bashir et al. used a PE blend having 3
wt % of ultra high-MW PE and 97 wt % of low MW PE to promote shish-kebab formation.
Further extrusion of this blend at a critical temperature and winding up the extruded filament
gave rise to interlocked shish-kebab morphology of the plugs in the continuously spun fibres. It
is possible to generate this structure by forming fibrous extended chain nuclei and then growing
lamellar crystals on them. The lamellar crystals, due to interpenetrated structures, assist in
preventing fibrillation. A schematic representation of interlocked shish-kebab structures is
depicted in Figure 5.2 The most important requirement to achieve this morphology is the
minimum distance between shishes. If the shishes are spaced far apart the lamellar crystals start
twisting, leading to a lower modulus in the direction normal to the fibre direction.5 So, by
optimizing the proportion of the oriented fibres and the interlocked kebab (formed from L-MWPE), the spacing between the shishes can be optimized.
83
Chapter 5
Figure 5.2 Shish and inter-locked kebab structures. (Shading is done for differentiation)
In this chapter, several bimodal reactor blends, differing in the proportions of the high and low
molecular weight components, are subjected to extension in order to achieve orientation and
consequently high stiffness of the oriented tapes.
5.2 Materials and methods
Bimodal polyethylenes, details of the synthesis of which are described in Chapter 3, were cryoground to obtain a fine powder. The PE powder was compression moulded at 180 °C and 150 bar
pressure to obtain plate of 1 mm thickness. From this plate, dumbbell shaped tensile bars of 30
mm length, 5 mm width and 11 mm gauge length were cut out using a mechanical stamp. Tensile
tests were performed using a Zwick 100 tensile tester equipped with a thermostat. The testing
conditions and parameters will be discussed in the following section. A load-cell of force limit
100 N was used when the sample was elongated at elevated temperature, while a load-cell of
force limit 10 KN was used when the specimens were redrawn at room temperature.
Thermal characterization (DSC) was carried out using a Q1000 differential scanning calorimeter.
The samples (2-3 mg), which were cut from the oriented tensile bars, were heated to 90 °C at a
heating rate of 10 °C/min followed by a lower heating rate of 1 °C/min up to 160 °C. The heating
rate of 1 °C/min was applied to resolve the melting peaks obtained due to oriented and unoriented structures. For the cooling cycle, a rate of 10 °C/min was applied. Second heating and
cooling cycle observed the rate of 10 °C/min.
5.3 Results and Discussion
The synthesized bimodal polyethylene blends possess a broad molecular weight distribution due
to the 10-fold difference in the molecular weight of the two components. As discussed in Chapter
3, the melt viscosity of these blends is lower, as compared to the monomodal H-MW PE,
depending upon the concentration of H-MW component in the blends. The broad MWD of the
blends thus provides an opportunity to utilise conventional processing techniques, while the HMW component promises better physical properties. In the following sections, the possibilities of
84
Towards ultimate properties of bimodal PE
attaining orientation using conventional techniques and the mechanical testing of the oriented
polyethylene will be discussed.
5.3.1 Possibilities of generating oriented structures
Two possible processing routes were considered to prepare a strong and stiff fibre composite
from the bimodal blends, the synthesis of which is discussed in Chapter 3.
(1) Sold state drawing: This route will effect deformation of the melt crystallized bimodal PE
samples at elevated temperature but below the melting temperature of unconstrained
extended chain crystals giving rise to oriented tapes.
(2) Melt Spinning/ Extrusion: Extrusion of PE melt through a capillary die and post
extrusion extension followed by crystallization will generate oriented wires or filaments.
5.3.2 Solid state drawing
Ward and co-workers provided the first instance of stretching of a molecular network by
stretching the polymers at temperatures sufficient to allow molecular mobility (between Tg. and
Tm).9 They discovered that the stiffness of the stretched polymer in the fibre direction is related to
the draw ratio. By controlling initial morphology, molecular weight and tensile drawing
conditions, high draw ratios of PE were reported with Young’s moduli of 70 GPa and tensile
strength of 1.5 GPa. 10,11 Later they investigated the effect of molecular weight,3d molecular
weight distribution and thermal treatment3d on the drawing behaviour of PE and on the
mechanical properties3a of oriented PE.3b
Solid state drawing differs from cold drawing in the temperature at which the sample is drawn.
Solid-state deformation is performed above the α-relaxation temperature but below the melting
temperature of PE, while cold drawing is performed at room temperature. The design of
experiments was selected based on four parameters. Two parameters, the PE composition and the
temperature condition for drawing, will be discussed first as they have to be considered for initial
design of experiments before performing the actual drawing experiments. The remaining two
parameters, deformation rate and total deformation, will be discussed later as they were evident
during the drawing experiments.
Six bimodal blends, differing in the composition of high and low molecular weight, were
selected for the series of experiments. The composition of the bimodal blends is noted in Table
5.1. The samples PE8, PE12 and PE 13 are the same samples reported in Table 3.4 (Chapter 3) as
PE1, PE6 and PE7, respectively. However, for fluency in the discussion the samples are labelled
in a sequence of increasing high MW content in the present series.
85
Chapter 5
Table 5.1 Composition of bimodal blends.
Sample
PE Composition
(wt %)
M
w
M
n
(kg/mol)
(kg/mol)
M w/ M
H-MW
L-MW
PE8
0
100
97
16
5.9
PE9
7
93
205
16
13.1
PE10
10
90
240
16
14.50
PE11
15
85
289
11
26.9
PE12
80
20
686
81
8.5
PE13
100
0
1324
824
1.6
n
In this series of the samples, PE 8 and PE 13 are the monomodal low and high molecular weight
samples, respectively. These two samples are the two extremes in the series and will be treated as
the reference samples for comparison. PE9 to PE12, which have increasing amounts of the HME component in the blend, complete the series. The appropriate composition of PE blends, as
the proportion of high and low molecular weight content, for this investigation is important as
the ultimate properties will be governed by the content of oriented H-MW component in the
blend. However, as discussed earlier, a high amount of the H-MW component will make drawing
difficult and decrease the orientation due to an increased number of entanglements. In the wake
of this fact, PE9, PE10 and PE 11 are important for these experiments in order to follow the
effect of H-MW chains on orientation and stiffness of the oriented tapes.
The drawing temperature is another important parameter for attaining orientation and desired
ultimate properties. Jarecki and Meier12 investigated the effect of the drawing temperature on the
drawing behaviour of PE and the properties of ultra drawn HDPE. They found that the effective
temperature for drawing depends on molecular weight and molecular weight distribution. The
temperature range for effective drawing shifted to higher temperature for polyethylene having
broad MWD and high MW. In this present work, to choose the drawing temperature, earlier
results of FIC on bimodal blends can be revisited.
As discussed in Chapter 4, the present bimodal blends provided a unique opportunity to generate
a suspension of only shishes in a non-crystallizable PE matrix. These experiments were
performed at a temperature (142 °C) near the equilibrium melting temperature (141.5 °C) of
unconstrained extended chain crystals. Shear flow, which is relatively weak as compared to the
extensional flow, was sufficient to orient the H-MW chains along the flow direction. It was
observed that at 137 °C, where most of the L-MW component remained in the random coil state,
the H-MW component oriented to generate shish like structures. From the observations evident
from Chapter 4, for drawing experiments 137 °C appears to be the appropriate temperature. At
86
Towards ultimate properties of bimodal PE
this temperature the specimen, which experiences elongation in the tensile tester, remains in the
solid state. At 137 °C, due to fast chain dynamics, the constraints provided by physical
entanglements between L-MW and H-MW chains are more dynamic and short lived, compared
to the entanglements between H-MW and H-MW chains. Thus the coil to stretch transition
during flow is influenced by the molecular weight difference between the long and short lived
entanglements. The stress transferred to the long molecules during elongation varies because of
the different constraints provided by the long and short lived entanglements. This will result in
the dilution of physical entanglements and, upon deformation, the chains will tend to experience
the elongation flow. This elongation flow will assist in orienting the H-MW chains in the
direction of the applied force.
5.3.3 Solid state deformation and mechanical testing
To orient PE chains, the test specimen was clamped in the tensile testing machine equipped with
a thermostat, which was equilibrated at 137 °C. After leaving the specimen for 10 minutes at 137
°C, it was elongated to 300 % of its length. Two deformation rates, 0.5 mm/s and 2.0 mm/s, were
applied for each specimen for the same final deformation. During the shear induced
crystallization experiments, which are discussed in Chapter 4, the strain rate was found to be
more effective to induce orientation rather than the shear time. However, the test specimens
reported in this section were deformed in solid state and experienced the thermal treatment from
room temperature to 137 °C, whereas in the shear experiment the specimen was heated to 230 °C
and then cooled to 137 °C. So in the samples which were subjected to solid state deformation at
137 °C, not all PE chains were in the random coil state at a given point of time. The fact that
elongation flow is stronger than the shear flow will also give rise to differences in the orientation
behaviour of chains in shear and elongation.
Figure 5.3 shows a photograph of a tensile specimen, which was cut from a compression
moulded plate, stretched at 137 °C to 300 % of its length and redrawn at room temperature. The
redrawing experiment will be discussed later; however, to have better understanding of the
experimental procedure, a fractured sample is also featured in the same figure. The stress-strain
curves of the bimodal PE samples (PE8-PE13) at 137 °C are shown in Figure 5.4. The
deformation rate for these experiments was 2.0 mm/s. The graph is split into two parts to have a
clear separation in the curves. The stress-strain behaviour for the 0.5 mm/s deformation rate was
similar to that apparent in Figure 5.4 and therefore is not included in the discussion.
87
Chapter 5
Figure 5.3 A photograph showing elongation of a specimen at 137 °C and a redrawn specimen at
room temperature. The fractures apparent at the sides indicate fibrillation.
0.008
0.008
PE13
0.007
Engineering Stress (GPa)
Engineering Stress (GPa)
0.007
PE12
0.006
0.005
0.004
0.003
PE8
0.002
0.006
PE9
0.005
PE11
0.004
0.003
PE10
0.002
0.001
0.001
0.000
0.000
0
50
100
150
200
250
300
Engineering Strain (%)
0
50
100
150
200
250
Engineering Strain (%)
Figure 5.4 Stress-strain curves of bimodal blends obtained by performing tensile tests at 137 °C
at a deformation rate of 2 mm/s.
The dumbbell shaped tensile specimen, when stretched at 137 °C, elongates by formation of a
neck, which propagates until it achieves the specified deformation. As the specimen thins
between the two grips of the dumbbell it becomes inhomogeneous, creating some weaker points
at the grips of the dumbbell. So, for the redrawing experiments, the specimens were cut to obtain
tapes, one of which is apparent in Figure 5.3.
The stress-strain curves in Figure 5.4 show a typical behaviour, for samples varying in molecular
weight and molecular weight distribution, in tension influenced by the entanglements present in
the H-MW chains. PE8, which is monomodal low molecular weight PE, could not be stretched to
300 % at 137 °C. It is known that the fewer the entanglements, higher is the draw ratio.
However, for higher drawability and strength, the presence of entanglements or tie chains is
required. Due to the low molecular weight and fast chain dynamics at 137 °C, which is discussed
in detail in to the low molecular weight and fast chain dynamics at 137 °C, which is discussed in
88
Towards ultimate properties of bimodal PE
detail in Chapter 4, the number of effective entanglements is low. Thus the strength of PE8 is
low at 137 °C, which led to the failure before the specimen reached the specified deformation.
Due to the premature failure, PE8 was not subjected to the redrawing experiments. In the case of
PE12 and PE13, which have 80 wt % and 100 wt % H-MW component respectively, strain
hardening is apparent immediately after the yield point. This strain hardening can be attributed to
the physical cross-links in the H-MW chains. The higher slope of strain hardening for PE13 is
again due to the higher number of entanglements, as there is no dilution effect, which is present
in PE12. PE9, PE10 and PE11, due to the considerable amount of effective entanglements, are
strong enough for the current testing conditions and do not fail until the specimens reach the
specified deformation of 300 %. The effect of entanglements on the strain hardening behaviour is
observed in PE11, which contains 15 wt % of the H-MW component. As these tests were
performed at elevated temperature, the resistance to deformation (Young’s modulus) was low as
expected.
Redrawing experiments were performed at room temperature to obtain the modulus of the
oriented tapes. For the comparison of moduli of the oriented and the unoriented bimodal blends,
tensile tests were performed on unstreched specimens of each bimodal blend. The stress-strain
curves obtained from redrawing experiments are depicted in Figure 5.5 (a), while Figure 5.5 (b)
compares the E-moduli of oriented and unoriented samples.
0.18
8
PE12
PE9
0.14
PE11
PE13
0.12
0.10
E-modulus (GPa)
Engineering Stress (GPa)
0.16
PE10
0.08
0.06
6
4
2
0.04
0.02
0.00
0
0
5
10
15
20
25
PE9
PE10
PE11
PE12
PE13
Specimen
Engineering Strain (%)
(a)
(b)
Figure 5.5 (a) Stress-strain curves obtained upon redrawing the oriented specimens at room
temperature. (b) Comparison of E-moduli of unoriented (■), oriented with a deformation rate of
0.5 mm/s (●) and 2.0 mm/s (▲) samples. Tensile tests were performed at room temperature.
It is clearly evident, from the greater slopes in the stress-strain curves, that the moduli of the
oriented samples are higher than those of the unoriented samples. This observation has been
reported by Ward et al. some 35 years ago; nevertheless, there are several interesting features
apparent from Figure 5.5. For PE9, PE 10 and PE11, the modulus shows an approximately 20
89
Chapter 5
times increase upon inducing orientation in the sample. However, the modulus showed only a 9
times increase for PE12 and in the case of PE13 the modulus of the oriented sample was similar
value to that of the unoriented sample. These observations point out the effect of entanglements
in polyethylene on the drawing behaviour. The bimodal PE blends PE8, PE9 and PE11 contain
low amounts of H-MW chains and therefore they show a dilution effect of entanglements, which
assists in stretching H-MW chains in the matrix of L-MW chains. It is analogous to the chain
orientation in a dilute solution, the L-MW PE matrix acting as a solvent for the longer chains.
Due to the fewer physical cross links, the orientation of longer chains is more effective. This is
evident from the relatively high values of the E-modulus. On the other hand, in PE 12 the
dilution effect of entanglements is less as the amount of H-MW (80 wt %) is quite high. Due to a
greater number of physical entanglements, the drawing becomes difficult and instead of
orientation of long chains it appears to be the stretching of a physical network of PE chains.
However, some parts of the chains get oriented in the direction of the stretching, apparent from a
small increase in the modulus of the elongated specimen. Due to a large number of long lived
entanglements, orientation of chains is hindered in sample PE 13, H-MW PE with narrow MWD,
though the specimen is stretched. In figure 5.4 the strain hardening shown by PE 12 and PE 13 is
a consequence of a large number of entanglements.
In Figure 5.3 a photograph of a tensile tested specimen (PE9) is shown. The stretched sample
when redrawn showed fibrillation, suggesting orientation in the direction of stretching but low
strength in the transverse direction. On the other hand samples PE12 and PE13 did not fibrillate,
the low modulus confirming less orientation of the chains.
It is well known that the draw ratio governs the stiffness of the fibres and oriented tapes. Several
methods of attaining high draw ratio and its influence on modulus are described in a review by
Bastiaansen. 13 Full extension of the polymer chains can be achieved by high draw ratio and high
modulus can be achieved as a consequence of chain orientation. However, entanglements in
polymer chains do not allow high draw ratios. In the present experimental series, PE9 and PE10,
due to less number of entanglements, were subjected to high draw ratio. For these two samples, a
draw ratio of 5 was achieved at a deformation rate of 0.5 mm/s at 137 °C. Upon redrawing, for
both samples, modulus of 8GPa was obtained. Thus for these samples, increase in the draw ratio
did not show considerable increase in the modulus.
Single shot synchrotron X-ray experiments were performed on the samples which were subjected
to deformation at 137 °C. The deformed samples were held in the X-ray beam and the diffraction
patterns were recorded. The SAXS fingerprints of the samples are shown in Figure 5.6. In the
case of PE 13, which is a high molecular weight PE, a four lobe pattern was observed. The
absence of equatorial streaks suggests the absence of oriented structures in the stretch direction.
Nevertheless, the four lobes pattern is an indication of a stretched network. These lobes which
are tilted at an angle to the stretch direction suggest the lamellae are aligned to an angle to the
90
Towards ultimate properties of bimodal PE
PE13
PE12
PE11
PE9
Figure 5.6 SAXS fingerprints of samples deformed at 137 °C.
preferred direction of orientation. As discussed earlier, due to the presence of a large number of
long lived entanglements, even at 137 °C, the pulling of individual chains leading to orientation
along the stretch direction becomes difficult. The deformation therefore results in a stretched
network exhibiting the properties of isotropic HDPE. In contrast, PE12, in which the high
molecular weight PE is ‘diluted’ by 20 wt % of the low molecular weight PE, showed equatorial
reflections in addition to the four lobes. The addition of 20 wt % of the low molecular weight PE
changes the chain dynamics, which in turn assists the orientation of a certain weight fraction of
high molecular weight chains. However, a number of entanglements still exist, giving rise to a
stretched network. The combined effect of the stretched network and oriented high molecular
weight results in the equatorial reflections and four crossed lobes. PE9 and PE11, which contain
7 wt % and 15 wt%, respectively, of high molecular weight chains, showed identical SAXS
fingerprints but different from those of PE12 and PE13. The equatorial reflections suggest
orientation along the direction of deformation. As discussed earlier, the matrix of low molecular
weight gives rise to faster chain dynamics at 137 °C. The constraints provided by physical
entanglements between low molecular weight and high molecular weight chains are more
dynamic and short lived, compared to the entanglements between high molecular weight and
high molecular weight chains. This faster dynamics assists in the orientation of high molecular
weight chains along the stretch direction. The equatorial reflections obtained in the case of PE9
and PE11 are broad and long indicating that the oriented crystals are short in length with low
thickness. In fully extended chain crystals or fibres these reflections thin to give a pointed
pattern. This is due to the infinite length of the extended chain crystals and high thickness of the
crystals as a result of close packing of the crystals in the fibre.
5.3.4 Melt extrusion of bimodal blends
As discussed earlier, the interlocked shish-kebab structure provides the strength and stiffness in
both the fibre direction and in the direction normal to it. Solid state drawing may not be able to
generate this morphology; however, from SSD experiments it is evident that melt drawing may
produce this morphology for the strong and stiff fibres. The effect of long chains on the
morphology of oriented structures obtained by flow induced crystallization also supports these
arguments. These arguments and observations form the basis of future work on melt extrusion of
bimodal blends and parallel and subsequent crystallization of the same to form strong and stiff
91
Chapter 5
fibres having interlocked shish-kebab morphology.
Due to the unavailability of appropriate instrumentation at the university, melt spinning of the
bimodal blends was not performed. During the course of time it was realised that the melt
spinning of these blends is not straightforward. Proper investigation of melt processability of
bimodal blend is beyond the scope of this thesis.
5.4 Conclusion
Solid state drawing of a bimodal polyethylene blend can result in a 20-fold increase in Young’s
modulus, compared to the modulus of unoriented bimodal PE. In contrast, solid state drawing
has little effect on the modulus of monomodal, high molecular weight PE. It was observed that,
in a bimodal blend containing up to 15 wt % of the high molecular weight fraction, the
orientation of high molecular weight chains at 137 °C is assisted by the low molecular weight
fraction, which remains in the random coil state. Increasing the content of high molecular weight
in a blend increases the number of long lived entanglements. The deformation of blends
containing higher amounts of high molecular weight chains therefore results in the stretching of a
network rather than orientation of predominantly high molecular weight chains. This study
provides a basis for future investigation of the possibility of obtaining high modulus fibres via
the melt spinning of bimodal polyethylene blends.
5.5 References
1. Treloar, L. R. G. Polymer, 1960, 37, 95.
2. (a) Sakurada, I.; Nukushina, Y. K.; Ito, T. J. Poly. Sci. 1962, 57, 651. (b) Strobl, G. R.;
Eckel, R. J. Polym. Sci. 1976, 14, 913. (c) Schauufele, R. F.; Shimanouchi, T. J. Chem.
Phys. 1967, 47, 9.
3. (a) Capaccio, G.; Ward, I. M. Polymer 1974, 15, 233. (b) Capaccio, G.; Ward, I. M. Polym.
Eng. Sci. 1975, 15, 219. (c) Capaccio, G.; Ward, I. M. Polymer 1975, 16, 239. (d) Capaccio,
G.; Crompton, T. A.; Ward, I. M. J. Polym. Sci. Polym. Phys. 1976, 14, 1641. (e) Capaccio,
G.; Crompton, T. A.; Ward, I. M. Polymer 1976, 17, 645. (f) Capaccio, G.; Crompton, T. A.;
Ward, I. M. J. Polym. Sci. Polym. Phys. 1980, 18, 301.
4. (a) Keller, A.; Odell, J. A. J. Polym. Sci. Polym. Symp. 1978, 63, 155. (b) Odell, J. A.; Grubb,
D. T.; Keller, A. Polymer, 1978, 19, 617. (c) Bashir, Z.; Odell, J, A.; Keller, A. J. Mat. Sci.
1984, 19, 617. (d) Bashir, Z.; Keller, A. Colloid. Polym. Sci., 1989, 267, 116.
5. (a) Pennings, A. J.; Lagaven, R.; Devries, E. S. Colloid. Polym. Sci. 1977, 255, 532. (b)
Zwijneburg, A.; Pennings, A. J. Colloid. Polym. Sci. 1975, 253, 452. (c) Zwijneburg, A.;
Pennings, A. J. Colloid. Polym. Sci. 1976, 254, 868.
6. (a) Smith, P.; Lemstra, P. J. J. Mat. Sci. 1980, 15, 505. (b) Smith, P.; Lemstra, P. J. Polymer
1980, 21, 1341. (c) Smith, P.; Lemstra, P. J.; Kalb, B.; Pennings, A. J. Polym. Bull. 1979, 1,
92
Towards ultimate properties of bimodal PE
733.
7. Ward, I. M. ‘An introduction to the mechanical properties of solid polymers’ 1993.
8. Barham, P. J.; Keller, A. J. Mat. Sci. 1985, 20, 2281.
9. Allison, S. W.; Ward, I. M. Br. J. Appl. Phys. 1967, 18, 1151.
10. Andrews, J. M.; Ward, I. M. J. Mater. Sci. 1970, 5, 411.
11. Capacciao, G.; Ward, I. M. Nature Phys. Sci. 1973, 243, 143.
12. Jarecki, L.; Meier, D. L. Polymer, 1979, 20, 1078.
13. Bastiaansen, C. W. M. Mater. Sci. Techn., 1997, 18, Chapter 11.
93
Chapter 5
94
Chapter 6
Investigation of the synergistic effect on catalyst
activity of a co-immobilized dual catalyst system
containing vanadium and iron catalysts
Abstract
Co-immobilization of [PhC(NSiMe3)2]VCl2(THF)2 and [2,6-{(2-chloro-4,6-dimethylphenyl)N=C(Me)}2C5H3N]FeCl2 on a MgCl2/AlEtn(OEt)3-n support generated bimodal polyethylene, as
these two precatalysts produced polyethylenes differing in molecular weight. The precatalysts
retained their active centre nature upon immobilization, but the combined activity of the two
catalysts was found to be much higher than expected for particular catalyst loadings. This
unexpected increase in catalyst activity prompted the investigation of possible reasons behind
this
observation
and
a
model
catalyst
[2,6-{(2-chloro-4,6-dimethylphenyl)N=C(Me)}2C5H3N]VCl3 was synthesized and used in ethylene polymerization. Two new
vanadium complexes, namely [2,6-{(2-chloro-4,6-dimethylphenyl)-N=C(Me)}2C5H3N]VCl3 and
[2,6-{(2,4,6-trimethylphenyl)- N=C(Me)}2C5H3N]VCl3, were not active for ethylene
polymerization under homogeneous conditions, were active after immobilizing on a
MgCl2/AlEtn(OEt)3-n support.
6.1 Introduction
Recently, one pot synthesis of bimodal polyethylene was effected by co-immobilizing 1-(8quinolyl)indenyl CrCl2
and [2,6-{(2-chloro-4,6-dimethylphenyl)N=C(Me)}2C5H3N]FeCl2
catalysts on a MgCl2/AlEtn(OEt)3-n support.1 Thus synthesized polyethylene consisted of two
molecularly mixed components having a large difference in their molecular weights. The
chromium catalyst generated PE with high molecular weight (H-MW, Mw = 1 × 106 g/mole) and
the iron catalyst produced low molecular weight (L-MW, Mw = 1 × 105 g/mole) PE. These
bimodal polyethylenes provided a model system for studying oriented morphologies generated
due to flow-induced crystallization; nevertheless, a possibility of improper mixing of the two
95
Chapter 6
components due to a large difference in their relaxation times was not totally ruled out,
particularly with a high proportion of high molecular weight polyethylene. In order to reduce the
difference in the relaxation times of the two components, an alternative to the chromium catalyst
was sought, ideally a catalyst capable of producing PE with Mw in the range 0.5 - 1 × 106 g/mole.
Severn et al. have reported the heterogenization and activation of a range of transition metal
catalysts.2 Several catalysts were found to retain their single-centre nature even upon
immobilization on the support. A vanadium amidinate catalyst,2e which produced moderately
high molecular weight PE (Mw = 7 × 105 g/mole) with narrow polydispersity index (Mw/Mn =
2.0), appeared to be the most attractive and suitable choice to replace the chromium catalyst in
the binary catalyst system. The iron catalyst, which is known to have limited ability to
incorporate higher alpha-olefins3,4 and is able to produce linear, low-molecular weight PE upon
immobilization on MgCl2/AlEtn(OEt)3-n supports,5 was retained as the other catalyst in the binary
catalyst system.
6.2 Materials and Methods
6.2.1 Materials
All manipulations were performed under an argon atmosphere using glove box (Braun MB-150
G1 or LM-130) and Schlenk techniques. Light petroleum (b.p. 40-60 °C) and dichloromethane
were passed over a column containing Al2O3 and stored over 4Å molecular sieves. All the
solvents were freeze-thaw degassed at least twice prior to use. Pentane was dried by refluxing
over potassium and stored over molecular sieves.
AliBu3 (1 M solution in hexane) was purchased from Akzo Nobel and AlEt3 was purchased
from Acros. AlEt2Cl, 2,6-diacetylpyridine (99 %) and 2-chloro-4,6-dimethylaniline (96 %) and a
silica-alumina support (grade 135) were purchased from Aldrich.
Ethylene (3.5 grade supplied by Air Liquide) was purified by passing over columns of 4Å
Molecular Sieves and BTS copper catalyst.
6.2.2 Catalysts
Catalyst (7) and catalyst (13) are reported in Chapters 2 and 3, respectively.
Si
thf
N
Cl
V
N
Si
thf
Cl
N
N
Cl Cl
Fe
N
Cl
Cl
7
13
Scheme 6.1. Precatalysts used for immobilization on MgCl2 support.
96
Synergistic effect in dual catalyst system
Details
of
the
synthesis
procedure
of
[2,6-{(2-chloro-4,6-dimethylphenyl)N=C(Me)}2C5H3N]VCl3 and [2,6-{(2,4,6-trimethylphenyl)-N=C(Me)}2C5H3N]VCl3 catalysts
are given in the results and discussion section.
6.2.3 Catalyst Immobilization and Polymerization
The details of catalyst immobilization and polymerization have appeared in previous chapters.
Ethylene polymerization under homogeneous conditions was performed in a 200 mL Büchi glass
reactor equipped with a mechanical overhead stirrer and inlets for monomer and solvent. Solvent
(petroleum ether), co-catalyst and catalyst were introduced to the reactor under inert atmosphere
and then pressurized with gaseous monomer to start the polymerization.
6.2.4 Polymer Characterization
Molecular weight and molecular weight distributions of the resulting polymers were determined
by means of size exclusion chromatography on a PL-GPC210 at 135 °C using 1,2,4trichlorobenzene as solvent. The particle morphology was examined using a Philips S-250 MK3
SEM EDX.
6.3 Results and Discussion
Previous studies on immobilization and activation of several early- and late-transition metal
catalysts2,5 offered an opportunity to select one from an array of precatalysts. Ideally, if both the
components of a dual catalyst system have similar activities under a given polymerization
condition, the ratio of two metals would approach unity for equal proportions of high and low
molecular weight polyethylene. This may lead to a probability of statistical distribution of
catalytically active centres. On the other hand, if there is a large difference in the activities of
two precatalysts, very different metal contents would be required. In the present study, a singlecentre vanadium amidinate catalyst, which was reported to produce moderately H-MW PE (Mw
= 7 × 105 g/mole) was selected to replace catalyst 4, although it was less active than the iron
catalyst. Other catalysts from our catalyst library were not able to satisfy the ideal conditions of
easy immobilization and acceptable activity, as well as the capacity to produce PE molecular
weights in the target range.
Earlier investigation on the effect of co-catalyst in ethylene polymerization performed with a
single-centre catalyst revealed a broadening of molecular weight distribution, although only to a
moderate degree, when AlEt3 rather than AliBu3 was used as a co-catalyst.6 However, during the
present study AlEt3 was nevertheless used as co-catalyst, as the activity of the vanadium catalyst
was relatively low when AliBu3 was used. The polymerization results with singly immobilized
and co-immobilized vanadium and iron catalysts are reported in Table 6.1.
97
Chapter 6
Table 6.1. Effect of the catalyst composition on activity and polymer molecular weight.
Entry
1
2
3
4
5
6
7
Catalyst
Loading
(µmol/100 mg)
7
13
0
2
1
1
1
0.5
2
0.2
2
0.1
3
0.1
2
0
Combined Activity
kg.(mol M)-1 (bar)1
(h)-1
Mw
Mn
(kg/mol)
(kg/mol)
8000
6700
5866
4000
4666
3354
3000
95
825
912
1512
1680
1541
2 491
13
20
19
24
68
239
1106
Mw/Mn
6.8
41.6
48.0
64.0
24.6
6.4
2.3
Support Composition: MgCl2·0.11AlEt1.32(OEt)1.68
Catalyst loadings were selected such as to keep the total metal content at around 1.5-2.2 µmol
per 100 mg support. Only in the case of entry 6 was the total metal content was increased. The
amount of AlEt3 was varied according to the catalyst loading, keeping an Al/M molar ratio of
1000. It is immediately obvious from the results in Table 6.1 that the polyethylene molecular
weight obtained with the vanadium catalyst (entry 7) was far in excess of the target Mw, differing
greatly from the Mw value of 762,000 obtained in previous studies.2e Possible reasons for this
could be differences in support composition, the use of a different co-catalyst (AlEt3 rather than
AliBu3) and poor reproducibility in Mw determination using GPC. The rheological and
mechanical properties of these polymers were therefore not investigated. However, a surprising
observation in the activity column of Table 6.1 justified further investigation of this dual catalyst
system. Decrease in the loading of catalyst 13, which is the more active component of the
catalyst system, should have decreased the combined catalyst activity. This argument was valid
throughout the series except for entry 5, which represents the metal ratio V/Fe = 20/1. All
experiments reported in Table 6.1 were reproducible, showing similar synergistic effects within
the limits of experimental error.
6.3.1 Synergistic effect
In order to investigate the reason behind the increase in catalyst activity, a closer look at
theoretical yield, observed yield and PE composition was required. The findings are reported in
Table 6.2, in which the sample sequence is the same as reported in Table 6.1.
98
Synergistic effect in dual catalyst system
Table 6.2. Comparison of calculated and observed PE yield and composition.
Entry
1
2
3
4
5
6
7
Polymer Yield (g)
Excess Yield
Calculated
Obtained
(%)
80
55
35
38
34
49
30
80
67
44
44
49
52
30
18
25
13
45
6
-
PE Composition PE Composition
(wt %) Calculated (wt %) Observed
L-MW H-WM L-MW H-MW
100
72
57
21
12
8
0
0
28
43
79
88
92
100
100
67
64
37
17
4
0
0
33
36
63
83
96
100
The theoretical polymer yield composition was calculated on the basis of the activities of the
singly immobilized catalysts (entries 1 and 7 in Table 6.1). The actual PE composition, i.e.
relative proportions of low- and high-molecular weight fractions, was obtained by deconvolution
of GPC curves. From Table 6.2 it is apparent that all co-immobilized catalyst systems produced
more polymer than calculated. Considering the margin in errors in weighing and catalyst
injection, an error bar up to 15% can be placed. The excess of 45% in the case of entry 5, which
was reproducible, can therefore be regarded as a significant result. This type of behaviour was
not observed in another dual catalyst system involving 1-(8-quinolyl)indenyl CrCl2 along with
catalyst 13.
In the studies with combined iron and chromium catalysts, retention of the nature of the active
centres upon co-immobilization on MgCl2/AlEtn(OEt)3-n was observed. However, the
observations in Table 6.2 indicate the presence of a definite synergistic effect, responsible for an
increase in catalyst activity, for the combination of catalysts 7 and 13. At the V/Fe ratio of 1/1, it
is apparent that the proportion of V-catalyzed H-MW polymer obtained exceeded that predicted
on the basis of the relative catalyst activities. As the V/Fe ratio increases up to 20/1, however, the
proportion of Fe-catalyzed L-MW PE is seen to exceed the predicted level. These observations,
taking into account the higher actual as opposed to calculated polymer yields, suggests the
presence of a synergistic effect which increases the activity of either one or both catalysts.
In order to obtain a clearer picture of the synergistic effects present in this system, the actual
weights of L-MW and H-MW polyethylene produced in each blend were calculated, allowing the
contribution of each individual catalyst to the excess yield to be determined. In Table 6.3 the
calculated yields of L-MW and H-MW polyethylene are compared with the observed yields
determined from deconvolution of the GPC plots.
99
Chapter 6
Table 6.3. Comparison of calculated and observed yield of polyethylene.
Entry
1
2
3
4
5
6
7
L-MW Yield (g)
Calculated Observed
80
80
40
45
20
28
8
16
4
8
4
2
0
0
H-MW Yield (g)
Calculated Observed
0
0
15
22
15
16
30
28
30
40
45
50
30
30
From Table 6.3 it is apparent that the iron catalyst produced a higher yield than calculated for all
combinations of dual catalyst loadings except entry 6, in which a particularly high loading of the
vanadium catalyst was used. In the case of a 1:1 ratio of V and Fe (entry 2), both precatalysts
contribute equally to the excess yield. As the series progressed, the loadings of iron catalyst were
decreased, leading to a doubling in Fe activity at V/Fe ratios of 10 and 20 (entries 4 and 5). A
possible reason for the increase in activity of the iron catalyst component when co-immobilized
in the present system could be related to the effect of catalyst loading. Decreasing the loading of
precatalyst on a magnesium chloride support to very low levels can lead to significant increases
in catalyst activity. This has been demonstrated for Cp2TiCl2,2b a MgCl2/TiCl4-type Ziegler-Natta
catalyst,7 and very recently also for (n-PrCp)2ZrCl2 immobilized on MgCl2/AlEtn(OEt)3-n.8 In the
latter studies, particularly large increases in activity were observed when the zirconocene loading
was decreased to 1 µmol/g support. It might be expected that a similar effect could be obtained
with the iron catalyst, which would provide an explanation for the increased productivities of
catalyst 13 in entries 2-5. The lower productivity of the iron catalyst in entry 6 may be due to the
high vanadium loading applied in this experiment.
The data in Table 6.3 also indicate, in most cases, some increase in the productivity of the
vanadium component when catalysts 7 and 13 are co-immobilized. In contrast to the iron
catalyst, the increased productivity of the vanadium catalyst can not be ascribed to large
variations in loading on the support. A possible explanation for the observed increases in the
yields of the V-catalyzed H-MW component could be a more rapid catalyst/polymer particle
fragmentation and growth in the presence of the (more active) iron catalyst, although this effect
was not observed (Chapter 3) in ethylene polymerization with co-immobilized iron and
chromium catalysts.
In order to obtain more insight into catalyst immobilization and to support the proposed
hypothesis that the synergistic effect could at least in part be related to the low loadings of the
iron complex 7 on the support, a series of UV experiments was performed.
100
Synergistic effect in dual catalyst system
6.3.2 UV/VIS Analysis
As described in Chapter 3, the kinetics of immobilization of different precatalysts on a support
can be followed using UV/VIS spectroscopy, measuring the rate of disappearance of the
precatalyst complex from solution in dichloromethane when contacted with the support. Several
UV/VIS experiments were therefore performed with catalysts 7 and 13 in order to determine and
compare their rates of immobilization on the support. The results of UV/VIS experiments
involving immobilizations of each single catalyst are depicted in Figure 6.1.
0.8
1.2
0.6
Complex 7
0.8
Absorbance (AU)
Absorbance (AU)
1.0
Complex 13
0.6
0.4
0.2
0.4
0.0
0.2
300
400
500
600
700
0
800
100
200
300
400
500
600
Time (min)
Wavelength (nm)
(a)
(b)
Figure 6.1. (a) UV/VIS spectra of precatalysts 7 and 13; (b) Immobilizations profile with time
for precatalysts 7 (○) and 13 (●).
As apparent from Figure 6.1 (a), complex 7 showed a strong absorption peak at 455 nm and
complex 13 showed a peak at 715 nm. In Figure 6.1 (b), individual immobilization profiles of
both complexes are compared. From earlier observations reported in Chapter 3, it was apparent
that complex 13, the iron precatalyst, immobilized faster than the chromium complex 4.
However, Figure 6.4 (b) shows that upon individual immobilization, the vanadium precatalyst
(complex 7) immobilized faster than the iron precatalyst (complex 13). During the initial 30
minutes of immobilization, both complexes show similar profiles, whereas complex 7 then
continues its rapid immobilization to completion, the iron complex 13 shows slower
immobilization after the initial 30 minute period.
To investigate the immobilization rates for co-immobilized vanadium and iron precatalysts, a set
of experiments was performed by varying the ratio of V to Fe. The combinations of V: Fe chosen
were 1:1, 20:1 and 30:1, which were featured in Table 6.1 as entries 2, 5 and 6 respectively. The
immobilization profiles of these experiments are shown in Figure 6.2.
101
Chapter 6
0.8
0.8
0.6
Absorbance (AU)
Absorbance (AU)
0.6
0.4
0.2
0.4
0.2
0.0
0.0
0
100
200
300
400
500
0
600
100
200
300
400
500
Time (min)
Time (min)
(a)
(b)
1.0
Absorbance (AU)
0.8
0.6
0.4
0.2
0.0
-0.2
0
100
200
300
400
500
600
Time (min)
(c)
Figure 6.2. Co-immobilization profiles of precatalysts 7 (○) and 13 (●) at V/Fe ratios of (a) 1:1
(b) 20:1 and (c) 30:1.
There are two common features in all three figures in Figure 6.2; namely, similar immobilization
profiles of the two catalysts during the first stage of immobilization and faster immobilization of
the iron precatalyst in the later stages of heterogenization. Interestingly, this result is the opposite
to what was observed when each catalyst was immobilized individually, when the vanadium
complex appeared to be supported faster than the iron complex in the later stages of
immobilization. These observations also differ from those of the co-immobilized iron and
chromium catalyst system reported in Chapter 3, in which a significantly faster immobilization
of the iron complex compared to the chromium complex was apparent both in the initial and later
stages of the immobilization process.
The more rapid disappearance from solution of the iron complex in the co-immobilization of 7
and 13 indicates its preferential immobilization. This might be one of the reasons of the
synergistic effect observed in this system, if it can be assumed that MgCl2-based supports
contain only a limited number of surface sites for effective precatalyst immobilization and
activation, and that these sites are the first to be occupied when the support and precatalyst(s) are
102
Synergistic effect in dual catalyst system
brought into contact. However, the lack of a synergistic effect in the (Fe + Cr) system casts some
doubt on the hypothesis that preferential immobilization of the iron catalyst and an inverse
relationship between the iron loading and polymerization activity may explain the synergy
observed in the (Fe + V) system.
In addition to this preferential immobilization of the iron catalyst and the effects of catalyst
loading, two other possible explanations for the synergistic effects observed in the present
system were considered and investigated. As described in the following sections, these were the
possibility of branched polymer formation, which may overcome a monomer mass transfer
limitation in the early stages of polymerization, and ligand exchange reactions between the iron
and the vanadium catalyst.
6.3.3 Chain Branching
It is well known that, in ethylene homopolymerization with heterogeneous catalysts, slow
monomer diffusion through crystalline PE formed on the surface of the catalyst particle can
retard the polymerization rate.9,10 The monomer diffusion limitation can be overcome by
introducing a co-monomer, which eases the monomer diffusion through less crystalline polymer.
Higher catalyst activity obtained by this technique is called the comonomer activation effect,
which is observed with both heterogeneous Ziegler-Natta catalysts and immobilized early- and
late-transition metal systems. In ethylene homopolymerization, incorporation into the catalyst
system of a nickel complex giving branched polyethylene similarly alleviates the monomer mass
transfer limitation, giving increased catalyst productivity.11 Prepolymerization with propylene or
other alpha-olefin also leads to easier particle fragmentation and growth in ethylene
polymerization.12,13
It is known that the bis(imino)pyridyl iron catalysts produce polyethylene with unsaturated chain
ends.14 If in a co-immobilized dual catalyst system the other catalyst is able to incorporate
comonomer, there is the chance of insertion of vinyl-terminated polyethylene chains generated
from the iron catalyst into the main chain, resulting in chain branching. Due to the broad
polydispersity of iron-catalyzed PE and the generation of vinyl-terminated oligomers,15 there can
be the possibility of both short- and long-chain branching, and an associated co-monomer
activation effect.
Nuclear magnetic resonance (NMR) and differential scanning calorimetry (DSC) experiments
are usually performed to detect branching in semi-crystalline polymers. In order to investigate
the branching level in the polyethylene synthesized using the dual catalyst system, melt state
NMR experiments were performed. The details of the experimental procedure are described in
Chapter 3. The polyethylene sample reported as entry 5 in Table 6.1, which showed the
103
Chapter 6
maximum synergistic effect, was first selected for NMR analysis. The spectrum is depicted in
Figure 6.3(A).
A
B
Figure 6.3. NMR spectrum of PE samples A (entry 5) and B (entry 2) confirming very low
branching.
As apparent from the NMR spectrum A, a small peak appearing at 34 ppm, which corresponds to
three carbon atoms at the α-position relative to the branching point, confirms that there is some
branching in the sample. Two other peaks, which can be attributed to penultimate and prepenultimate chain-end carbon atoms, were also apparent in the spectrum, but the peak for carbon
atoms at the β-position relative to the branch point (normally appearing at 27 ppm with the same
peak area as that of α-carbon) was absent. If the presence of impurities and solvent traces in the
sample is excluded, the absence of the peak for β-carbon atom can be attributed to its longer
relaxation time. Quantification of the peak areas showed that the degree of branching was only
one branch per 8000 carbon atoms. This level of branching is too low to affect the polymer
crystallinity and to cause a co-monomer activation effect, excluding the possibility of easy
monomer diffusion through less crystalline polymer generated from branched PE. NMR
spectrum B in Figure 6.3 is of the PE sample prepared with equal loadings of 7 and 13 and
therefore having a greater probability of branching. However, quantification of the α-C peak
showed only one branch per 15000 carbon atoms, a further indication of very low levels of
branching in these polymers.
6.3.4 Ligand exchange reaction
Another possibility of increased catalyst activity, as mentioned earlier, was ligand exchange
reaction between two complexes when in solution, resulting in a highly active complex. Ligand
exchange reactions in transition metal complexes are well documented, usually involving monoanionic groups such as halide, alkoxide or amide, etc.16-19 Gibson et al. have performed a series
of exchange reactions involving monoanionic and dianionic ligands for pseudo-tetrahedral
complexes. They observed that monoanionic ligands exchange faster than their dianionic
104
Synergistic effect in dual catalyst system
counterparts as a consequence of greater metal-ligand bond strength of dianionic ligand systems.
Additionally, exchange of less bulky ligands (e.g. oxo, chloride) goes faster than those of bulkier
ligands such as arylimido and tert-butylamide.20 There are several reports of bis(imino)pyridyl
vanadium catalysts used in ethylene polymerization.5,21,22 Therefore, although considering the
stability of the complexes the possibility of ligand exchange reaction in dichloromethane solution
appeared to be unlikely, the synthesis of a V(III) complex with the bis(imino)pyridyl ligand from
catalyst 13 was undertaken. In the case of a possible ligand exchange reaction, an iron amidinate
complex will be formed. However, there is no report on such kind of iron complex and therefore
the synthesis of an iron amidinate complex was not undertaken.
Since the discovery of highly active iron and cobalt complexes with bis(imino)pyridyl
ligands,14,23 several groups have investigated homogeneous and supported catalysts bearing this
ligand system.24-26 The discovery of these tridentate systems, having aryl substituted α-diimine
systems with bulky substitution on aryl rings at the ortho position, triggered an extensive
research on similar class of ligands with different substitutions involving different metals.
Although tridentate bis(imino)pyridyl ligands have been widely used in the synthesis of late
transition metals for olefin oligomerization and polymerization, there are fewer reports involving
early transition metals. Under homogeneous conditions bis(imino)pyridyl chromium27-29 and
vanadium21,22 complexes have been found to be active in ethylene polymerization. Esteruelas et
al.28 reported the formation of waxes and low molecular weight polyethylene using {2,6[ArN=C(Me)]2C5H3N}CrCl2 or {2,6-[ArN=C(Me)]2C5H3N}CrCl3 (Ar = 2,6-diisopropylphenyl).
Evidence for the generation of similar types of active species derived from Cr(II) and Cr(III),
apparent from the formation of mixtures of oligomers and polymers with a range of
bis(imino)pyridyl chromium complexes, was reported by Small et al.29 Gambarotta and
coworkers21 prepared {2,6-[ArN=C(Me)]2C5H3N}VCl3 (Ar = 2,6-diisopropylphenyl), which
after activation by MAO produced polyethylene with broad molecular weight distribution,
suggesting the formation of different active species. Schmidt et al.22 synthesized a series of
bis(imino)pyridyl V(III) complexes, most of which gave low molecular weight oligomers.
Here in this section, the synthesis of two new V(III) complexes bearing bis(imino)pyridyl
ligands, is reported. These two complexes were immobilized on a MgCl2/AlEtn(OEt)3-n support
and used in ethylene polymerization. As discussed earlier, the synthesis of [2,6-{(2-chloro-4,6dimethylphenyl)-N=C(Me)}2C5H3N]VCl3 complex was carried out to determine whether its
possible formation could explain the increased activity observed in the dual catalyst system
involving
[PhC(NSiMe3)2]VCl2(THF)2 (7)
and
[2,6-{(2-chloro-4,6-dimethylphenyl)N=C(Me)}2C5H3N]FeCl2 (13).
105
Chapter 6
6.3.4.1 Synthesis of [2,6-{(2-chloro-4,6-dimethylphenyl)-N=C(Me)}2C5H3N]VCl3
The 2,6-{(2-chloro-4,6-dimethylphenyl)-N=C(Me)}2C5H3N ligand was synthesized by following
a procedure reported in the literature.30 1.23 g (7.5 mmol) of 2,6-diacetyl pyridine and 2.80 g (18
mmol) of 2-chloro-4,6-dimethylaniline were transferred to a round bottom flask containing about
40 mL toluene. A silica alumina support (grade 135; 0.6 g) and 4Å molecular sieves (1.5 g) were
added and the mixture was heated at 40 °C for 72 hours. Anhydrous methanol was added to
obtain a solid compound, A (Scheme 6.2). Complex A (0.25 g ≡ 0.57 mmol) was then reacted
with a solution of VCl3 · 3THF (0.190 g ≡ 0.57 mmol) in 100 mL anhydrous THF by heating the
mixture at 70 °C for 1 hr and then stirring overnight at room temperature. The dark brown
coloured solid was filtered and washed with THF. The yield of the complex was around 65 %.
Elemental analysis of 14 (C25H25N3VCl5): calculated C, 50.4; H, 4.2; N, 7.05 and found C, 46.2;
H, 3.88; N, 5.73.
N
+
N
O
N
+ VCl
VCl33 · 3THF
THF
N
N
N
Cl
O
V
Cl
NH2
Cl
Cl
Cl
Cl
A
N
Cl
Cl
14
Scheme 6.2. Synthesis of complex 14.
6.3.4.2 Synthesis of [2,6-{(2,4,6-trimethylphenyl)-N=C(Me)}2C5H3N]VCl3
The 2,6-{(2,4,6-trimethylphenyl)-N=C(Me)}2C5H3N ligand was synthesized as reported in the
literature.4 The vanadium complex was synthesized using the same procedure as described for
complex 14. The yield obtained was 82 % and the elemental analysis of 15 (C27H31N3VCl3) was:
calculated C, 58.5; H, 5.6; N, 7.6 and found C, 56.7; H, 7.7; N, 6.5.
N
N
N
+ VCl3 · 3THF
N
N
V
Cl
N
Cl
Cl
15
Scheme 6.3. Synthesis of complex 15
6.3.4.3 Polymerization results
The results of ethylene polymerizations performed using 14 and 15 are reported in Table 6.4,
from which it is apparent that the activities of both catalysts are much lower than those of the
106
Synergistic effect in dual catalyst system
corresponding iron complexes.5 However, independent of the co-catalyst and temperature, these
complexes produced high molecular weight polyethylene upon immobilization. In contrast, no
activity was obtained when these complexes were used along with aluminium trialkyls or
methylaluminoxane under homogeneous polymerization conditions. The moderate activity and
high molecular weight obtained upon heterogenization indicates a significant effect of
immobilization on the activation of these precatalysts. The activity obtained was dependent on
cocatalyst and temperature. In the case of 14, triethylaluminium co-catalyzed polymerization at
50 °C produced PE with very high molecular weight and narrow polydispersity. A doubling in
activity was obtained at the higher polymerization temperature of 70 °C, without any significant
effect on molecular weight and polydispersity. Similar results have earlier been reported by
Huang et al.,5 who obtained PE with high molecular weight and narrow polydispersity using a
bis(imino)pyridyl V(III) complex with 2,6-diisopropyl substitution in the aryl rings. It is apparent
from Table 6.4 that, when AlEt2Cl and AliBu3 were used as co-catalysts, the activity of 14
dropped considerably and a lowering in PE molecular weights was also observed. A higher
activity with AlEt3 than with other aluminum alkyls was expected from previous studies, but the
very low activity with AliBu3 was surprising. Repeated polymerizations confirmed this low
activity, and the broad PE polydispersity and lower molecular weight indicated some anomalous
behavior using AliBu3 as cocatalyst with complex 14.
Table 6.4 Effect of co-catalyst and temperature in ethylene polymerization of [2,6-{(2-chloro4,6-dimethylphenyl)-N=C(Me)}2C5H3N]VCl3
(14)
and
[2,6-{(2,4,6-trimethylphenyl)N=C(Me)}2C5H3N]VCl3 (15).
Entry Catalyst
Co-catalyst Temperature Activity
kg.(mol M)-1
(°C)
Mw
(kg/mol)
Mn
(kg/mol)
Mw/Mn
1872
1685
1084
657
960
750
866
911
619
144
310
225
2.2
1.9
1.8
4.6
3.1
3.3
(bar)-1(h)-1
1
2
3
4
5
6
14
14
14
14
15
15
AlEt3
AlEt3
AlEt2Cl
AliBu3
AlEt3
AliBu3
50
70
50
50
50
50
600
1400
140
15
200
150
The activity of [2,6-{(2,4,6-trimethylphenyl)-N=C(Me)}2C5H3N]VCl3 (15), using AlEt3 as
cocatalyst, was lower than that of both 14 and its iron counterpart. In contrast to the results with
complex 14, however, the use of AliBu3 as co-catalyst gives an activity approaching that
obtained with AlEt3.
These observations confirmed the relatively low activity of bis(imino)pyridyl vanadium catalysts
compared to their iron counterparts, ruling out the possibility that ligand exchange in the dual
catalyst system could contribute to the observed synergistic effect on catalyst activity.
107
Chapter 6
6.3.4.4 Polymer Morphology
During earlier investigations of catalyst heterogenization,2,5 the immobilization of several early
and late transition metal catalysts on MgCl2/AlEtn(OEt)3-n supports has been shown to produce
polyethylene with spherical morphology. The polyethylene synthesized from immobilized
bis(imino)pyridyl vanadium(III) catalysts (14 and 15) also had spherical polymer particle
morphology, with no evidence of reactor fouling. Scanning electron micrographs of polymers
synthesized from these two catalysts are shown in Figure 6.4, demonstrating that the spherical
particle morphology of the support material had been retained and replicated throughout catalyst
immobilization and polymerization.
It is apparent from Figure 6.4 that the highly active iron catalyst (13), upon immobilization on a
MgCl2/AlEtn(OEt)3-n support, produced PE having regular spherical morphology. However, the
immobilized vanadium catalysts, though they did not show any evidence of reactor fouling,
generated PE with less regular morphology. The inferior particle morphology of the vanadium
catalyzed polyethylene is due to the relatively low catalyst activity, as evident from somewhat
better particle morphology shown in Figure 6.4(b) compared to the one apparent in Figure 6.4(c).
(a)
(b)
(c)
Figure 6.4 Particle morphology of polyethylene synthesized using immobilized (a) [2,6-{(2chloro-4,6-dimethylphenyl)-N=C(Me)}2C5H3N]FeCl2
(13);
(b)
[2,6-{(2-chloro-4,6dimethylphenyl)-N=C(Me)}2C5H3N]VCl3 (14, entry 1, Table 6.3); (c) [2,6-{(2,4,6trimethylphenyl)-N=C(Me)}2C5H3N]VCl3 (15, entry 5, Table 6.3).
6.4 Conclusion
Co-immobilization of [PhC(NSiMe3)2]VCl2(THF)2 and [2,6-{(2-chloro-4,6-dimethylphenyl)N=C(Me)}2C5H3N]FeCl2 on a MgCl2/AlEtn(OEt)3-n support indicated a synergistic effect on the
catalyst activity. A relatively low loading of the iron catalyst produced a greater than expected
amount of polyethylene. A faster immobilization rate of the iron catalyst compared to the
vanadium complex indicated its preferential immobilization. The possibility of increased activity
108
Synergistic effect in dual catalyst system
being due to chain branching or a ligand exchange reaction having been refuted, the synergistic
effect might be attributed to the low catalyst loading. However, there was no evidence of such an
effect in the previously investigated co-immobilized iron and chromium system, raising some
doubts that the synergistic effect is related to the decrease in loading of the iron catalyst. These
observations leave the study inconclusive. Two new vanadium complexes, namely [2,6-{(2chloro-4,6-dimethylphenyl)-N=C(Me)}2C5H3N]VCl3 and
[2,6-{(2,4,6-trimethylphenyl)N=C(Me)}2C5H3N]VCl3, were not active for ethylene polymerization under homogeneous
conditions, but were active after immobilizing on a MgCl2/AlEtn(OEt)3-n support.
6.5 References
1. (a) Kukalyekar, N.; Rastogi, S.; Chadwick, J. C. Polymer Preprints, 2007, 48(1), 280. (b)
This thesis, Chapter 3.
2. (a) Severn, J. R.; Chadwick, J. C. Macromol. Rapid Commun. 2004, 25, 1024. (b) Severn, J.
R.; Chadwick, J. C Macromol. Chem. Phys. 2004, 205, 1987. (c) Severn, J. R.; Chadwick, J.
C.; Van Axel Castelli, V. Macromolecules 2004, 37, 6258. (d) Severn, J. R. ; Kukalyekar, N.;
Rastogi, S ; Chadwick, J. C. Macromol. Rapid Commun. 2005, 26, 150. (e) Severn, J. R.;
Duchateau, R.; Chadwick, J. C. Polym. Int. 2005, 54, 837.
3. Small, B. L.; Brookhart, M. Macromolecules, 1999, 32, 2120.
4. Britovsek, G. J. P.; Bruce, M.; Gibson, V. C.; Kimberley, B. S.; Maddox, P. J.; Mastroianni,
S.; McTavish, S. J.; Redshaw, C.;Solan, G. A.; Strömberg, S.; White, A. J. P.; Williams, D. J.
J. Am. Chem. Soc. 1999, 121, 8728.
5. Huang, R.; Kukalyekar, N.; Koning, C. E.; Chadwick, J. C. J. Mol. Catal. A: Chem. 2006,
260, 135.
6. Kukalyekar, N.; Huang, R.; Rastogi, S.; Chadwick, J. C. Macromolecules, Submitted.
7. Echevskaya, L. G.; Matsko, M. A.; Mikenas, T. B.; Nikitin, V. E.; Zakharov, V. A. J. Appl.
Polym. Sci. 2006, 102, 5436.
8. Huang, R.; Duchateau, R.; Koning, C. E.; Chadwick, J. C., Manuscript in preparation.
9. Floyd, S.; Mann, G. E.; Ray, W. H. In Catalytic Polymerization of Olefins; Keii, T., Soga, K.,
Eds.; Elsevier: Amsterdam, 1986; p 339.
10. Soga, K.; Yanagihara, H.; Lee, D. Makromol. Chem. 1989, 190, 995.
11. Huang, R.; Koning, C. E.; Chadwick, J. C. Macromolecules 2007, 40, 3021.
12. Kou, B.; McCauley, K. B.; Hsu, J. C. C.; Bacon, D. W. Macromol. Mater. Eng. 2005, 290,
537
13. Zakharov, V. A.; Bukatov, G. D.; Barabanov, A. A. Macromol. Symp. 2004, 213, 19.
14. Britovsek, G. J. P.; Gibson, V. C.; Spitzmesser, S. K.; Tellmann, K. P.; White, A. J. P.;
Williams, D. J. J. Chem. Soc. Dalton Trans. 2002, 1159.
15. Huang, R.; Koning, C. E.; Chadwick, J. C. J. Polym. Sci.: Part A: Polym. Chem. 2007, 45,
4054.
16. Arney, D. J.; Wexler, P. A.; Wigley, D. E. Organometallics, 1990, 9, 1282.
109
Chapter 6
17. Chrisholm, M. H.; Folting, K.; Huffmann, J. C.; Kirkpatrick, C. C. Inorg. Chem., 1984, 23,
1021.
18. Weingarten, H.; Van Wazer, J. R. J. Am. Chem. Soc. 1966, 88, 2700.
19. Weingarten, H.; Van Wazer, J. R. J. Am. Chem. Soc. 1965, 87, 724.
20. Gibson, V. C.; Graham, A. J.; Jolly, M.; Mitchell, J. P. Dalton Trans. 2003, 4457.
21. Reardon, D.; Conam, F.; Gambarotta, S.; Yap, G.; Wang, Q. J. Am. Chem. Soc. 1999, 121,
9318.
22. Schmidt, R.; Welch, M. B.; Knudsen, R. D.; Gottfried, S.; Alt, H. G. J. Mol. Catal. A: Chem.
2004, 222, 17.
23. Small, B. L.; Brookhart, M.; Bennett, A. M. A. J. Am. Chem. Soc. 1998, 120, 4049.
24. Wang, Q.; Yang, H.; Fan, Z.; Xu, H.; J. Poly. Sci. A: Polym. Chem. 2004, 42, 1093.
25. Barabanov, A. A.; Bukatov, G. D.; Zakharov, V. A.; Semikolenova, N. V.; Echevskaja, L.
G.; Matsko, M. A. Macromol. Chem. Phys. 2005, 206, 2292.
26. (a) Hlatky, G. G. Chem. Rev. 2000, 100, 1347. (b) Severn, J. R.; Chadwick, J. C.; Duchateau,
R.; Friederichs, N. Chem. Rev. 2005, 105, 4073.
27. Devore, D. D.; Feng, S. S.; Frazier, K. A.; Patton, J. T. 2000 WO00/69923.
28. Esteruelas, M. A.; López, A. M.; Mèndez, L.; Olivan, M.; Onate, E. Organometallics, 2003,
22, 395.
29. Small, B. L.; Carney, M. J.; Holman, D. M.; O’Rourke, C. E.; Halfen, J. A. Macromolecules,
2004, 37, 4375.
30. Qian, C.; Gao, F.; Chen, Y.; Gao, L. Synlett, 2003, 10, 1419.
110
Technology Assessment
Outstanding scientific and technological developments have made the polyolefin industry one of
the most successful growth industries in the area of large volume materials. During the course of
the past 50 years polyethylene, labelled as a commodity plastic, has extended itself to several
specific applications, such as fishing lines and bulletproof vests, transforming it into a specialty
polymer. This transformation was the result of lateral and out of the box thinking relating to
engineering aspects and the physical and chemical properties of polyethylene.
A balance between the processability of high molecular weight polyethylene and the ultimate
properties is a longstanding question. In past few decades, this issue has been addressed by
proposing several synthesis routes and processing techniques. In the present thesis, advances in
catalysis, polymer crystallization and processing have been brought together to make an attempt
to achieve the golden mean between melt processability of high molecular weight polyethylene
and the ultimate properties.
In order to achieve this golden mean between melt processability and ultimate properties,
emphasis has been placed on bimodal polyethylene. It is estimated that the market for bimodal
HDPE is expected to grow by approximately 6 million tons between 2005 and 2015. The most
common industrial process for the production of bimodal polyethylene, which is known as the
cascade reactor process, utilizes two rectors in series as has been discussed in Chapter 3. In
contrast, the present work concerns the synthesis of bimodal polyethylene in a single reactor,
using a combination of two different catalysts immobilized on a single support. Compared to the
cascade process, this approach has both advantages and disadvantages. The use of a single
reactor is clearly an advantage in terms of capital investment. However, whereas in the cascade
process the molecular weights of the two components of bimodal PE are controlled by applying
different hydrogen concentrations in each reactor, in a single reactor process molecular weight
and bimodality are primarily dependent on the choice of catalysts. The single reactor concept
therefore lacks the flexibility provided by hydrogen in controlling the molecular weight.
Nevertheless, considerable interest in both industry and academia is being paid to ethylene
polymerization using combinations of catalysts on a single support. One example of the use of
two or more catalysts working in tandem is the preparation of linear low density polyethylene
from ethylene as single monomer source, employing for example a trimerization catalyst
(generating 1-hexene) together with a polymerization catalyst. Another example is the use of two
polymerization catalysts which differ in their ability to incorporate an α-olefin comonomer into
111
Technology assessment
the polyethylene chain. Bimodal polyethylenes comprising a relatively low molecular weight
homopolymer component and a high molecular weight copolymer component offer significant
advantages for pipe and film applications, in particular with respect to stress cracking resistance.
A bis(imino)pyridyl iron catalyst, such as that used in the present work, would be a particularly
suitable component for the synthesis of the relatively low molecular weight homopolymer
component, as a result of the very low copolymerization capacity of such catalysts. A singlecentre chromium or vanadium catalyst, such as those described in Chapters 2, 3 and 6, could be
used for the synthesis of the (high molecular weight) copolymer fraction. Combinations of silicasupported metallocenes differing in their ability to incorporate an α-olefin comonomer in
ethylene polymerization have been described by Univation Technologies. They have developed a
catalyst system, termed ProdigyTM, which is capable of producing film grade bimodal
polyethylene in a single reactor. Compared to silica supports, the MgCl2-based system discussed
in this thesis has the advantage of better fragmentation and easier particle growth. MgCl2
supports are well established for Ziegler-Natta catalysts and have the further advantage that
many MgCl2-immobilized catalysts can be activated using simple aluminium alkyl cocatalysts
such as AlEt3 or AliBu3, whereas methylaluminoxane (MAO) is typically required with silicasupported systems. The MgCl2-based systems give high catalyst activity expressed as PE
produced per mol of transition metal immobilized on the support, but relatively low catalyst
loadings result in relatively low productivity expressed as kg PE produced per gram of support.
The present work has been directed toward bimodal polyethylene homopolymers rather than
copolymers, with special emphasis on the effects of bimodality on flow-induced crystallization.
Since the early 1970s, a tremendous amount of research has been done in the area of oriented
polyethylene and several routes have been adopted to obtain high-strength and high-modulus
fibres. The process for the production of commercially available solution-spun fibres, which
were developed by Smith and Lemstra, utilizes a semi-dilute solution of ultra high molecular
weight polyethylene. The solution spinning (commonly known as gel-spinning) process involves
a tremendous amount of solvent and the cost involved in the recovery of the solvent adds to the
production cost of the finished product. A melt processing route, if it can achieve comparable
ultimate properties, can be more economical and cost effective in the production of strong and
stiff fibres or tapes. The use of intimate bimodal blends, produced with co-immobilized catalysts,
offers the prospect of obtaining strong and stiff fibres or tapes at low process operating costs.
The discovery in this study that a suspension of only shishes can be generated in a matrix of low
molecular weight polyethylene represents a fundamental step in the journey towards the ultimate
goal of producing strong and stiff fibres by the melt processing of high molecular weight
polyethylene. The initial findings reported in Chapter 5, which describe the improved mechanical
properties of solid state drawn bimodal polyethylenes, can be taken further in achieving high
modulus for the melt processed fibres. As it was realised in Chapter 5, a choice of an appropriate
composition of high and low molecular weight in a bimodal blend and the processing
112
Technology assessment
temperature are the key parameters influencing the ultimate mechanical properties. Bimodal
blends consisting of up to 20 wt% of high molecular weight fraction appear to be within in the
processable range.
The fundamental studies reported in this dissertation are thus potentially significant with respect
to industrial development and application. The results with the polyethylenes obtained with an
up-scalable catalyst system and polymerization process, clearly suggest that the ultimate goal of
improved ultimate properties via a melt processing route is coming closer.
113
Technology assessment
114
Samenvatting
De ultieme eigenschappen van een polymeer worden niet alleen door de chemische structuur
bepaald, maar zijn ook afhankelijk van de verwerkingscondities die worden toegepast tijdens
fabricage van het eindproduct met name ten gevolge van de oriëntatie van de lange ketens.. De
intrinsieke eigenschappen van een polymeer worden vooral bepaald door het moleculair gewicht;
hoe hoger de zogenaamde gewichtgemiddelde molmassa (Mw), des te beter zijn de fysische
eigenschappen van het polymeer. Een hoog molgewicht is essentieel voor goede mechanische
eigenschappen zoals de treksterkte en taaiheid alsmede de levensduur. Echter, conventionele
verwerkingstechnieken kunnen niet worden gebruikt voor polymeren met een te hoog
molgewicht vanwege een prohibitief hoge smeltviscositeit. Tot aan een kritisch molgewicht van
een polymeer (Mc, typisch twee maal het gemiddelde molgewicht tussen verstrengelingen) neemt
de viscositeit evenredig toe met molgewicht (Mw<Mc ; η0 ~ Mw), maar boven Mc neemt de
viscositeit snel toe (Mw>Mc ; η0 ~ Mw3.4). De slechte verwerkbaarheid van een polymeer met een
hoog molgewicht beperkt de wijze van verwerking en de toepassingsgebieden van het polymeer.
Als gevolg van de hoge kosten van speciaal ontworpen verwerkingstechnieken worden de
eindproducten te duur. Dus voor een succesvol product moet er in gelijke mate aandacht worden
besteed aan chemie, fysica en verwerkingsmethoden. In de praktijk bij verwerking van
polymeren (plastics) is er dus vaak een compromis tussen de eigenschappen (hoog molgewicht
gewenst) en de (snelle) verwerkbaarheid, viz. een lager molgewicht.
Vaak wordt een polymeer met een laag molgewicht gemengd met een polymeer met een hoog
molgewicht. Maar mengen (blenden) in de gesmolten fase van twee polymeren is moeilijk, met
name wanneer een component hoog-moleculair is. Mengen via oplossing is geen gerede optie op
industriële schaal vanwege de hoge kosten. Op industriële schaal wordt gebruik gemaakt van
sequentiele reactoren waarbij in de eerste reactor polymeer wordt gemaakt, bijv. laagmoleculair, en in de volgende reactor hoog-moleculair polymeer wordt vervaardigd.
In dit proefschrift worden polyethylenen met hoog en laag molgewicht gesynthetiseerd door twee
verschillende katalysatoren op een enkele drager te immobiliseren, om zo het probleem van
mengbaarheid te overwinnen. Deze route omzeilt dus het gangbare gebruik van meerdere
reactoren en geeft de gewenste homogene distributie van een polymeer met een hoog
molgewicht (Mw > 1 x 106 g/mol) in een matrix van een polyethyleen met een laag molgewicht .
Behalve de synthese is er ook de nodige aandacht besteed aan de morphologie en rheologie
(vloeigedrag) van de gemaakte produkten. Onder bepaalde vloeicondities leiden verschillende
115
Samenvatting
relaxatietijden van ketens met een laag en hoog molgewicht tot kristallisatie van het hoogmolgewicht materiaal en tot de vorming van een zogenaamde shish structuur. Dit betekent dat de
lange ketens worden verstrekt/georiënteerd in de vloeirichting en als kiem kunnen fungeren voor
het la(a)g(er) molgewicht in het mengsel, hetgeen resulteert in de bekende shish-kebab
morphologie; een gestrekte kern van georienteerde (hoog-moleculaire) ketens met daarop als
decoratie de gevouwen ketenkristallen (naar analogie van het bekende voesel produkt , de shishkebab) . De aanwezigheid van de shish-structuur verlaagt de barrière tot kiemvorming voor de
vorming van gevouwen ketenkristallen waardoor het polymeer met laag molgewicht op relatief
hoge temperaturen kan kristalliseren. De gewenste ultieme eigenschappen zoals de E- Modulus
en treksterkte zijn gecorreleerd met deze structuurontwikkeling tijdens vloei.
Er werden verschillende katalysatoren geïmmobiliseerd op dragers van het type
MgCl2/AlRn(OR’)3-n
om
polyethylenen
met
verschillende
molgewichten
en
molgewichtsverdelingen te synthetiseren. Smelt rheometrie werd gebruikt om onderscheid te
maken tussen zogenaamde “single-centre” en “multi-centre” katalysatoren. Onderzoek van de
rheologische eigenschappen van gesmolten polyethyleen, gesynthetiseerd met chroom, titaan,
vanadium en zircoon katalysatoren geïmmobiliseerd op MgCl2 dragers, bevestigen dat er een
smalle (Schulz-Flory) molgewichtsverdeling ontstaat hetgeen indicatief is voor single-centre
katalyse. Dit geldt voor met Cr-, V- en Zr-gebaseerde systemen, maar niet met Ti. In het geval
van polymeren gemaakt met MgCl2-geimmobiliseerde Ti- complexen en met een smalle
molgewichtsverdeling (Mw/Mn 2-3), wordt een afname van de storage modulus met een
afnemende trillings frequentie geconstateerd, hetgeen indicatief is voor een afwijking van een
Schulz-Flory distributie. Echter, polyethylenen gesynthetiseerd met soortgelijke vanadium
complexen laten een constante (plateau) modulus over een breed frequentiegebied zien. De
aanwezigheid van een plateaumodulus is karakteristiek voor polyethyleen met een smalle
molgewichtsverdeling. Hierdoor kan gebruik worden gemaakt van smelt rheologie om de
aanwezigheid van een Schulz-Flory distributie te bewijzen of te weerleggen in gevallen waar de
gebruikelijke meet techniek om molgewichtsverdelingen te bepalen, Size-ExclusionChromatography (SEC) geen definitief antwoord geeft.
Co-immobilisatie
van
1-(8-quinolyl)indenyl
CrCl2
en
[2,6-{(2-chloro-4,6dimethylphenyl)N=C(Me)}2C5H3N]FeCl2 op een MgCl2/AlEtn(OEt)3-n drager werd gebruikt om
bimodaal polyethyleen te maken in een enkele reactor, het zogenaamde “one pot” proces. Op
deze manier werden mengbare PE-mengsels van hoog (Mw = 106 g/mol, Mw/Mn = 2.0) en laag
(Mw = 105 g/mol, Mw/Mn = 5.5) molgewicht gesynthetiseerd. De relatieve verhouding van de
hoog en laag molgewicht componenten kon worden gestuurd door de beladingen van de twee
katalysatoren te variëren. Rheologische studies van bimodale polyethylenen wijzen op een
homogene menging van de hoog en laag molgewicht componenten binnen het tijdsbestek van de
proeven met een lage deformatie.
116
Samenvatting
Homogeen gemengde polyethylenen met een zeer hoog en laag molgewicht, gesynthetiseerd op
deze manier, zijn modelsystemen voor de studie van de invloed van molgewicht op kristallisatie
onder invloed van afschuifspanning. Vanwege de duidelijke verschillen in relaxatietijden werd
de invloed van molgewicht op kristallisatie en de resulterende kristallijne morfologie bestudeerd
door middel van tijdsafhankelijke röntgen-technieken. De resultaten zijn dat zeer stabiele shish
structuren worden gevormd van ketens met een hoog molgewicht, waardoor kiemen voor verdere
kristallisatie ontstaan Deze shish structuren zijn stabiel boven het evenwichtssmeltpunt van vrije
gestrekte keten-kristallen van polyethyleen, bij afwezigheid van vloei.
De structuurontwikkeling tijdens vloei is gerelateerd aan de mechanische eigenschappen van het
bimodale polyethyleen. Enkele van de gesynthetiseerde bimodale polyethylenen, met
verschillende verhoudingen tussen de hoog en laag molgewicht componenten, werden verstrekt
in de vaste fase. Tijdens het verstrekken in de vaste fase bij 137 °C, werd de oriëntatie van hoog
molgewicht ketens mede beïnvloed door de aanwezigheid van de laag molgewicht matrix. Dit
werd bevestigd door het resultaat dat de mechanische eigenschappen van bimodaal PE
afhankelijk zijn van de hoeveelheid hoog molgewicht polymeer.
Ter afsluiting van het chemisch/katalytisch onderzoek werden nog co-immobilisatie
experimenten uitgevoerd met [PhC(NSiMe3)2]VCl2(THF)2 en [2,6-{(2-chloro-4,6dimethylphenyl)-N=C(Me)}2C5H3N]FeCl2 op een MgCl2/AlEtn(OEt)3-n drager en dit liet een
synergetisch effect op de katalysator activiteit zien. Dit effect was niet zichtbaar in het geval van
de geïmmobiliseerde chroom en ijzer katalysatoren. Verschillende mogelijke verklaringen voor
dit effect werden onderzocht, maar een definitieve verklaring kan niet worden gegeven.
Desalniettemin werden tijdens dit onderzoek twee nieuwe vanadium complexen gesynthetiseerd,
namelijk
[2,6-{(2-chloro-4,6-dimethylphenyl)-N=C(Me)}2C5H3N]VCl3
en [2,6-{(2,4,6trimethylphenyl)-N=C(Me)}2C5H3N]VCl3. Deze complexen zijn niet actief in ethyleen
polymerisatie onder homogene condities, maar zijn actief na immobilisatie op een
MgCl2/AlEtn(OEt)3-n drager. [2,6-{(2-chloro-4,6-dimethylphenyl)-N=C(Me)}2C5H3N]VCl3 gaf
een hoog molgewicht polyethyleen met een smalle (Schulz-Flory) molgewichtsverdeling.
117
Samenvatting
118
Acknowledgements
I would like to deeply thank several people who, during my stay in Eindhoven, provided me with useful
and helpful assistance. Without their care and consideration, this thesis would likely not have matured.
It has been a great experience working at SKT and I enjoyed the last 4 years here. First of all, I would like
to thank prof. P.J. Lemstra for giving me an opportunity to work on a project, spanning a broad spectrum
of polymer chemistry and technology. His inspiring and motivating (?) remarks, especially during the
later stages of the project, helped me a lot. It would have been difficult to bring in the flavour of polymer
physics in this thesis without the help of Prof. Sanjay Rastogi. His contribution towards structuring the
thesis is highly acknowledged. Sanjay thanks a lot for everything. I express my gratitude to Dr John
Chadwick for his guidance, inputs and encouragement since my day one in the Netherlands. Over the
period of four years, under his able guidance, I got trained not only as a researcher but also as a good
scientific writer, though John may not agree on the latter part.
I would like to extend my thanks to prof. Vincenzo Busico, Dr. Markus Gahleitner and prof. Cor Koning
for the constructive reading which helped me in improving my thesis. Vincenzo’s critical remarks during
most of the DPI meetings made me to think on the broader perspective. I am especially thankful to
Markus for his quick and constructive comments on each and every aspect in the thesis. I am thankful to
Prof. Sivaram for agreeing to be in the defence committee in spite of his busy schedule.
I am very thankful to John Severn for his help in introducing me to this project. John it was nice to have
you around. I thank Rob Duchateau for several intriguing discussions and for giving me a feeling of
organometallic chemistry. I thank Gerrit Peters for several discussions on rheology and flow induced
crystallization. I acknowledge DPI PO cluster community for their inputs and intriguing and lively
discussions (especially in the bar) during the PO cluster meetings. I would like to acknowledge Tim Kidd,
Gerard van Doremaele, Nic Friederichs, Klaas Remerie, Markus Gahleitner, Wouter van Meerendonk,
Jan Stamhuis and Vincenzo Busico for the support and help during the course of this project. I thank Prof.
Christian Bailly for several discussions and help in rheological measurements. I am thankful to Dr. Ton
de Weijer (Teijin) for his timely help in processing the synthesized materials. The fortnightly discussions
with John C, John S, Madri, Soazig, Rubin and Adelaida were helpful in keeping up the pace of the work.
In these four years several people helped me with experimental techniques, analysis and related
discussions. I would like to acknowledge Andries Jekel (RuG, Groningen) and Valerie Grumel
(Stellenbosch) for GPC analysis. I thank Guido, Giuseppe, Caroline and Luigi for X-ray experiments and
analysis. Collaboration with Luigi proved out to be the most successful, fruitful and productive. I am
thankful to Bing Wang, Nilesh Patil and Han Xu for helping me during my frequent visits to
Loughborough. I acknowledge Yao (Max Plank Institute, Mainz) for helping me with Solid State NMR
119
Acknowledgements
experiments, She-She Yao!! I am thankful to J.F. Vega for introducing me to the experimental rheology. I
thank Bob, Bjorn, Tom Engels, and Chris Reynolds for helping me with processing and mechanical
testing. My special thanks to Bob and Chris for helping me when it mattered the most. I thank Xuejing,
Pauline, Anne, Irina and Charef for nice microscopy images; after all, one picture is better than a
thousand words. I thank Otto for IR measurements, Joost and Weizhen for helping me in thermal analysis.
Special thanks to Raf, who is always ready to help others in all sorts of things in the lab. I thank MK
Singh for the help in Matlab programming. I acknowledge John, Jules, Roy and PLEM for writing and
correcting the Dutch summary of my thesis. Bedankt!!
I am indebted to Elly, the lifeline of SKT, for her support. Her help in filling up the DUTCH forms,
calling IND for residence permit and her ever ready to help attitude is highly appreciable. I extend my
thanks to Ineke for support and will surely miss her reminders.
I thank all SKT and SPC members for making my stay here enjoyable. My Karma-bhoomi (work
place/lab) was one of the most enjoyable places in Eindhoven. The joyous mood, jokes and funny remarks
made the atmosphere in the Cactus lab lighter than hydrogen. I enjoyed my time in the lab with John,
Wouter, Joep, Delphine, Kirti, Rachel, Madri, Soazig, Rubin, Saeid, Simona, Maurice and of course Raf. I
enjoyed my time in the office with Lijing, Cees, Charef, Weizhen and Gosia. The incredible fights with
Lijing (China Queen) and Cees (m-astard) were really refreshing. The coffee time used to be the best
starting point of the day. A nice warm coffee and crispy gossips with Irina, Cees, Joost, Jules, Roy, Bob,
Gosia, Lijing, Merina and Dong used to be the highlight of the day.
Here come the Mafias! The best part of Eindhoven is the incredible number of Indians around.
(Eindhoven is just one example; we are there in all nice places). In the company of our desi-janata I
always felt at home. I am thankful to Sachin, Mano, Rajesh, Sreejit, Keshav, Sreepad, Rajan, Mani,
Vidya, Ankur, Sai, Girish, Anish, MK, Vishal, Shalaka, Vinit, Dilna, Subi,Vinayak, Kirti, Sanjeev,
Prakriti, Ruchi-bhabhi, Merina, Soney, Chaitnaya, Riyaz, Rahul, Amol, Nilesh Rane, Dilip, Chattarbir,
Sharan, Yogesh, Bhushan, Koustubh, Narendra to list a few. Introduction of Gaytri-Rajan, Radha, Manju,
Rupali, Lipika, Sudha, Debu, Ashwini added more flavour to the group.
I am indebted to my parents and my family for their faith and support. Without their wishes and support I
would not have been there, where I am now. Aai, Baba and Mahesh thanks for everything. I am also
grateful to aai and Abhi for their wishes and support. Finally I am grateful to my wife, Gayatri, for her
faith in me, support, patience and understanding, especially during the thesis writing time.
There are more people whom I should thank but it is too difficult to put all names here. My family
members, friends, colleagues and everyone who has helped me directly or indirectly- I extend my word of
thanks to all of them.
120
Curriculum vitae
Nileshkumar Kukalyekar was born in Yellapur, India on March 30th 1978. After completing the
bachelor’s degree in chemistry from Willingdon College Sangli, he opted for the post graduate
course in Polymer Science at PGTC, Shivaji University.
After graduation, in June 2000 he joined Asian Paints India Ltd, where he worked as a quality
assurance chemist in the solvent based paints division. Later, in July 2001 he moved to GE
Plastics, John F Welch Technology Centre Bangalore, where he worked for a little over two
years. His area of interest was synthesis and up-scaling of catalysts for the production of
bisphenol-A.
From December 2003, he started his PhD-study at Eindhoven University of Technology,
Eindhoven, The Netherlands, in the group “Polymer Technology (SKT)”, under the supervision
of Prof. P.J. Lemstra. During his PhD-study, he completed four modules of the course
‘Registered Polymer Technologist’ (RPK, Register Polymeerkundige), organized by the
‘National Dutch Graduate School of Polymer Science and Technology’ (PTN,
‘Polymeertechnologie Nederland’). He was a part of PromoVE, the Ph.D. organization in TUE,
where he worked initially as a foreign Ph.D. contact and later as the chairman of the
organization.
From December 1, 2007 Nilesh will be working with DSM Research in the performance material
division at Geleen, Netherlands.
121
122