f

g

p arflwmr.

fl

 

A",

2

 

‘5
.k

,

2 L 4
awn.
2 .2.
A.
2
1.2.. 2...
M .
. an at}!
27 V ,
an? . . .. .
“hawks . . gumfiwfi
, £3:
. “‘1. V2Ho‘o 5m
Luna.” (:4 .1...
335%.”:qu
$2. .2
am
2 .
~W‘
has.
h.“
h
2 3?
Mn».
. 22.2..“ 2
cu.
ugqfi

Tut-fir)
4... 2.
.13 .2

. 12...).
ahf: Hunt)...
L 2V.§2

v

_ , _ _ __V __ _ ,, . _ ___J

This is to certify that the

thesis entitled

Reactive Extrusion with Star-Polycaprolactone

presented by

Ryoko Yamasaki

has been accepted towards fulfillment
of the requirements for

M'8 degreein CHE '

 

 

C’N‘Wv 9"”: /\‘ C’K/Wfl “A”

Major professor kg]

 

 

Date M 0% H71 0207;" .

0-7639 MS U is an Affirmative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State
“University

 

 

 

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

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cleIRC/DatoDue.p65-p.15

REACTIVE EXTRUSION WITH STAR-POLYCAPROLACTONE
By

Ryoko Yamasaki

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Department of Chemical Engineering

2000

ABSTRACT
REACTIVE EXTRUSION POLYMERIZATION
By

Ryoko Yamasaki

There are many advantages in carrying out a reaction within the extruder, called a
reactive extrusion, such as a lower processing cost, better control over residence time
distribution and the temperature profile. In this study, an extruder was utilized to carry
out different types of reactions: reactive extrusion polymerization of e-caprolactone to
produce a star-polycapro lactone (REX-PCL); reactive extrusion polymerization of REX-
PCL and diisocyanate to produce polyurethane; and transesterification ofREX-PCL and
polyethylene terephthalate. By studying the reactive extrusion processes with different
reaction chemistries, limitations with the current reactive extrusion system was also
determined. The results of this study showed that the range of the molecular weight of
REX-PCL achievable was found to be 20,000 to 210,000 for the particular system used in
this study due to the physical limitation with the REX-PCL of different molecular weight.
Also, a polyurethane was produced continuously from the production of REX-POL;
however, further analysis are necessary in order to evaluate the usefulness of this product.
Finally, the titanium (IV) isopropoxide and dibutyltin dilaurate was found to work
effectively in carrying out the transesterification of REX-PCL and polyethylene

terephthalate.

To my family
who has always been there for me...

iii

ACKNOWLEDGMENTS

I want to thank my advisor, Dr. Ramani Narayan, for giving me the opportunity to
work on this project, as well as his guidance and patience throughout my work.

I also want to thank Dr. Carl Lira and Dr. Ruben Hernandez for agreeing to be my
committee members, and being patience while I am finishing my thesis.

I also would like to thank the member of past and present Narayan Research
Group. I want to thank Chris Shuster for training me and working with me during the
most struggling and messiest time of this study, and Yudith Opel for working overtime to
help me finish this project on time. I also want to thank Dale Smith for training me and
Sunder Rajan Balakrishnan for changing the screw configurations.

Also, I would like to thank Mike Rich and staff of Composite Materials Structural
Center for their support and valuable advice.

iv

TABLE OF CONTENTS

LIST OF TABLES
LIST OF FIGURES
CHAPTER 1: INTRODUCTION
1.1 Reactive Extrusion Polymerization
1.2 Process Conditions
1.3 Project Overview
CHAPTER 2: BACKGROUND
2.1 Twin Screw Extruders
2.2 Screw Configuration

2.3 Characterization of Polymers

2.3.1 Measurement of Intrinsic Viscosity
2.3.1.1 Standard Procedure

2.3.2 Determination of Mark-Houwink-Sakurada

Constants

2.3.3 Thermogravimetric Analysis
2.3.3.1 Standard Procedure

2.3.4 Differential Scanning Calorimeter
2.3.4.1 Standard Procedure

2.3.5 Fractional Conversion
2.3.5.1 Standard Procedure

CHAPTER 3: REACTIVE EXTRUSION POLYMERIZATION OF
EPSILON-CAPROLACTONE

3.1 Introduction

3.2 Background

3.3 Outline of Chapter

3.4 Reactive Extrusion Polymerization of s-Caprolactone to

Produce REX-PCL of Various Molecular Weight
3.4.1 Objective

xi

xiii

11
13
14
15
17
19
21
22
25
27
27

28

28

28

31

32

32

3.4.2 Materials

3.4.3

3.4.4
3.4.5

Procedure

3.4.3.1 Flowrate Calculation

3.4.3.2 Preparation of ATSB Solution
3.4.3.3 Flow Scheme

3.4.3.4 Start-up

3.4.3.5 Collecting Samples and Shut Down
Results

Discussion

3.5 Characterization of Star-Polycaprolactone

3.5.1

3.5.2

3.5.3

3.5.4

3.5.5

Determination of Fractional Conversion
3.5.1.1 Objective

3.5.1.2 Materials

3.5.1.3 Procedure

3.5.1.4 Results

3.5.1.5 Discussion

Determination of Intrinsic Viscosity
3.5.2.1 Objective

3.5.2.2 Materials

3.5.2.3 Procedure

3.5.2.4 Results

3.5.2.5 Discussion

Determination of Mark-Houwink-Sakurada
Constants

3.5.3.1 Objective

3.5.3.2 Procedure

3.5.3.3 Results

3.5.3.4 Discussion
Thermogravimetric Analysis

3.5.4.1 Objective

3.5.4.2 Materials

3.5.4.3 Procedure

3.5.4.4 Results

3.5.4.5 Discussion

Differential Scanning Calorimeter
3.5.5.1 Objective

3.5.5.2 Materials

3.5.5.3 Procedure

3.5.5.4 Results

3.5.5.5 Discussion

3.6 Reactive Extrusion Polymerization of s-Polycaprolactone
fi'om Monomer with Impurities

3.6.1
3.6.2

Objective
Materials

32
34
34
37
38
42
42
43

46

46
46
47
47
48
48
48
48
49
50
5 1

51
51
52
53
54
54
54
54
54
56
56
56
56
57
57
59

59

59
59

CHAPTER 4:

3.6.3 Procedure
3.6.4 Results
3.6.5 Discussion

3.7 Monomer Conversion and Reproducibility
3.7.1 Monomer Conversion
3.7.1.1 Objective
3.7.1.2 Materials

3.7.1.3 Procedure
3.7.1.4 Results
3.7.1.5 Discussion
3 .7 .2 Reproducibility
3.7.2.1 Objective
3.7.2.2 Materials
3.7.2.3 Procedure
3.7.2.4 Results
3.7.2.5 Discussion

3.8 Conclusions

REACTIVE EXTRUSION POLYMERIZATION OF STAR-
POLYCAPROLACTONE AND DIIOCYAN ATE

4.1 Introduction

4.2 Background
4.2.1 Effect of Catalyst
4.2.2 Effect of Temperature
4.2.3 Effect of Water

4.3 Outline of Chapter

4.4 Reactive Extrusion Polymerization of Diisocyanate and
REX-PCL with High Molecular Weight Star-
Polycapro lactone

4.4.1 Objective
4.4.2 Materials
4.4.3 Procedure
4.4.3.1 Flowrate Calculation
4.4.3.2 Preparation of ATSB and HMDI
Solutions
4.4.3.3 Flow Scheme
4.4.3.4 Start Up
4.4.3.5 Collecting Samples and Clean Up
4.4.4 Results

vii

60
60

6 1
61
61
61
62
62
63
63
63
63
63

65

66

67

67

69
69
69
70

71

71

71
72
72
72
75

76
77
78
78

CHAPTER 5:

4.4.5 Discussion

4.5 Characterization of Polyurethane made with High
Molecular Weight Star-Polycaprolactone
4.5.1 Objective
4.5.2 Materials
4.5.3 Procedure
4.5.4 Results
4.5.5 Discussion

4.6 Reactive Extrusion Polymerizatoin of Diisocyanate and
REX-PCL with Low Molecular Weight Star-
Polycaprolactone

4.6.1 Objective
4.6.2 Materials
4.6.3 Procedure
4.6.4 Results
4.6.5 Discussion

4.7 Characterization of Polyurethane made with Low
Molecular Weight Star-Polycaprolactone
4.7.1 Objective
4.7.2 Materials
4.7.3 Procedure
4.7.4 Results
4.7.5 Discussion

4.8 Conclusions

TRANSESTERIFICATION OF STAR-
POLYCAPROLACTONE AND POLYETHYLENE
TEREPHTHALATE

5.1 Introduction

5.2 Background
5.2. 1 PET
5.2.2 Screw Configuration

5.3 Outline of Chapter

5.4 Transesterification of REX-PCL and PET with
Triphenylphosphine Solution
5.4.1 Objective
5.4.2 Materials
5.4.3 Procedure

viii

79

79

79
80
80
80
86

86

86
87
87
89
89

90
90
91
91
91
94
94

95

95
97
97
98
98
99
99

99
100

5.4.3.1 Preparation of Triphenylphosphine
Solution

5.4.3.2 Determination of Feeding Location

5.4.3.3 Extrusion Temperature

5.4.3.4 Reactive Extrusion Process

5.4.4 Results
5.4.5 Discussion

5.5 Transesterification of REX-PCL and PET with
Triphenylphosphine Pellets

5.5.1 Objective

5.5.2 Materials

5.5.3 Procedure
5.5.3.1 Reactive Extrusion Process
5.5.3.2 Characterization

5.5.4 Results

5.5.5 Discussion

5.6 Transesterification of REX-PCL and PET with

Dibutyltin Dilaurate

5.6.1 Objective

5.6.2 Materials

5.6.3 Procedure
5.6.3.1 Reactive Extrusion Process
5.6.3.2 Characterization

5.6.4 Results

5.6.5 Discussion

5.7 Transesterification of REX-PCL and PET with
Titanium (IV) Isopropoxide

5.7.1 Objective

5.7.2 Materials

5.7.3 Procedure
5.7.3.1 Reactive Extrusion Process
5.7.3.2 Characterization

5.7.4 Results

5.7.5 Discussion

5.8 Conclusions

CHAPTER 6: CONCLUSION AND RECOMMENDATIONS

6.1 Reactive Extrusion Polymerization of e-
Caprolactone

100

100
101
101

102

103

104

104
105
105
105
105
106
107

108

108
108
108
108
109
109
111

111

111
111
112
112
112
113
114

115
116

116

REFERENCES

6.2 Reactive Extrusion Polymerization of Star-
Polycapro lactone and Diisocyanate

6.3 Transesterification of Star-Polycaprolactone and
Polyethylene Terephthalate

6.4 Limitation with Reactive Extrusion

117

117

117

119

Table 2-1 :

Table 2-2 :

Table 3-1 :

Table 3-2 :

Table 3-3 :

Table 3-4 :

Table 3-5 :
Table 3-6 :
Table 3-7 :
Table 3-8 :

Table 3-9 :

Table 3-10 :

Table 4-1 :

Table 4-2 :
Table 4-3 :
Table 4-4 :

Table 4—5 :

Heating condition of TGA used in this study

Three types ofheating conditions of DSC used in this study: (a)

LIST OF TABLES

Type 1; (2) Type 2; and(c) Type 3

An example of the calculation sheet used to determine the

monomer and initiator flowrate

An example of the various monomer and the initiator flowrate to
yield different theoretical molecular weight of REX-PCL

Molecular weight of REX-PCL produced (T=180°C, l 101-pm)

Fractional conversion and actual molecular weight of REX-PCL

produced

Intrinsic viscosity of linear PCL

Intrinsic viscosity of REX-PCL

Mark-Houwink-Sakurada constants

Degradation temperature of linear and REX-PCL

Molecular weight of REX-PCL produced with the monomer with
moisture content

Intrinsic viscosity of REX-PCL collected at two different

locations within the extruder

An example of the calculation sheet used to determine the

flowrate of reactants

An extruder temperature profile used in this experiment

Condition used to produce polyurethane

Intrinsic viscosity of polyurethane

MI-IS constants of three polymers

xi

22

36

37

50

50

53

55

60

62

74

77

79

80

82

Table 4—6 : The extruder temperature profiles corresponding to the two types 88
of feeding schemes

Table 4-7 : Feed conditions used to produce polyurethane (T=120°C, 89
llOrpm)

Table 5-1 : Six feeding combinations tested 100

xii

Figure 2-1 :

Figure 2-2 :

Figure 2-3 :

Figure 2-4 :

Figure 2-5 :

Figure 2-6 :

Figure 2-7 :
Figure 2-8 :
Figure 2-9 :

Figure 2-10 :

Figure 2-11 :

Figure 2-12:

Figure 3-1 :

Figure 3-2 :
Figure 3-3 :
Figure 3-4 :
Figure 3-5 :

Figure 3-6 :

LIST OF FIGURES

A schematic of co-rotating and counter-rotating twin screw
extruder with various degree of intermeshing

A schematic of conjugated and non-conjugated co-rotating twin
screws

A schematic of (a) conical twin screw extruders and (b)
cylindrical extruder of unequal screw length

A schematic of (a) matched screw configuration and (b)
staggered screw configuration

Photograph of Werner Pfleiderer twin screw extruder ZSK-3O

Screw configurations used in this study: (a) Screw
Configuration #1; (b) Screw Configuration #2

An example of graphs used to determine the intrinsic viscosity
A schematic of typical output of Thermogravimetric Analysis

A photograph of TA Instruments Hi-Res TGA 2950
Thermogravimetric Analyzer

A schematic of output from Differential Scanning Calorimeter

A schematic of how DSC output changes before (a) and after (b)
the transesterification

A photograph of TA Instruments DSC 2920 Modulated DSC

Ring-opening polymerization of s-caprolactone to produce REX-
PCL

The structure of linear (a) linear PCL and (b) REX-PCL
A structure of e-caprolactone
A structure of aluminum tri-sec butoxide

Flow scheme of e-caprolactone monomer

Flow scheme of ATSB solution7

xiii

10

12

32

33

39

4o

Figure 3-7 :
Figure 3-8 :

Figure 3-9 :

Figure 3-10 :

Figure 3-11 :

Figure 3-12 :

Figure 3-13 :

Figure 3-14 :

Figure 3-15 :

Figure 3-16 :

Figure 4-1 :

Figure 4-2 :

Figure 4-3 :

Figure 4-4 :

Figure 4-5 :
Figure 4-6 :
Figure 4-7 :
Figure 4-8 :

Figure 4-9 :

Flow scheme of nitrogen gas

An example of plot used to determine the intrinsic viscosity of
(a) linear and (b) REX-PCL

Determination of Mark-Houwink-Sakurada constants for linear
PCL

Determination of Mark-Houwink-Sakurada constants for REX-
PCL

The output of TGA for linear and REX-PCL

The output of DSC for REX-PCL with the molecular weight of
30,000

The output of DSC for REX-PCL with the molecualr weight of
60,000

The output of DSC for REX-PCL with the molecualr weight of
1 00,000

A plot used to determine the intrinsic viscosity of REX-PCL
made on three different days

A plot used to determine the intrinsic viscosity of REX-PCL
made with two different flowrate

Polyurethane synthesis by the reaction of diisocyanate and
dihyroxy-group

A reaction mechanism of the formation of polyurethane with
hexamethylene diisocyanate and REX-PCL

Flow scheme of HMDI solution

Comparison of the intrinsic viscosity of the three types of
polymers

Output of TGA for the polyurethane and REX-PCL (30,000)
Output of TGA for the polyurethane and REX-PCL (60,000)
Output of TGA for the polyurethane and REX-PCL (100,000)
Output of TGA for three polyurethanes

Output of DSC for the polyurethane and REX-PCL (30,000)

xiv

41
49

52

53

55

57

58

58

65

67

68

76
81

82

83

83

84

84

Figure 4-10 : Output of DSC for the polyurethane and REX-PCL (60,000) 85

Figure 4-11 : Output of DSC for the polyurethane and REX-PCL (100,000) 85

Figure 4-12 : Output of TGA for REX-PCL and polyurethane with the 92
molecular weight of 3000

Figure 4-13 : Output of DSC for the REX-PCL and polyurethane with the 92
molecular weight of 3000

Figure 4-14 : Output of TGA for the REX-PCL and polyurethane with the 93
molecular weight of 7500

Figure 4-15 : Output of DSC for the REX-PCL and polyurethane with the 93
molecular weight of 7500

Figure 5-1 : A reaction mechanism of the transesterification of REX-PCL and 96
PET

Figure 5-2 : A structure of polyethylene terephthalatc 97

Figure 5-3 : A structtn‘e of triphenylphosphine 99

Figure 5-4 : A schematic of possible feeding location 100

Figure 5-5 : Output of TGA for the REX-PCL/PET mixture with 102
triphenylphosphine solution as a catalyst

Figure 5-6 : Output of DSC for teh REX-PCL/PET mixture with 103
triphenylphosphine solution as a catalyst

Figure 5-7 : Output of TGA of REX-PCL/PET mixture with 106
triphenylphosphine pellets as a catalyst

Figure 5-8 : Output of DSC of REX-PCL/PET mixture with 107
triphenylphosphine pellets as a catalyst

Figure 5-9 : A structure of dibutyltin dilaurate 108

Figure 5-10 : Output of TGA for the REX-PCL/PET mixture with dibutyltin 110
dilaurate as a catalyst

Figure 5-11 : Output of DSC for REX-PCL/PET mixture with dibutyltin 1 10
dilaurate as a catalyst

Figure 5-12 : A structure of titanium (IV) isopropoxide 112

XV

Figure 5-13 : Output of TGA for REX-PCL/PET mixture with titanium 1 13
isopropoxide as a catalyst

Figure 5-14 : Output of DSC for REX-PCL/PET mixture with titanium 114
isopropoxide as a catalyst

xvi

Chapter I

 

INTRODUCTION

 

Growing environmental concerns worldwide demand the reduction of
environmentally unfavorable waste, as well as the addition of the biodegradability or
recyclability of the product as design criteria”. In order to produce environmental
fiiendly materials, various types of biodegradable plastics have been developed,
including poly-s-caprolactone, polylactide, and polyethylene succinate. Although the
mechanism and rate of degradation of each polymer varies, they were all found to be
converted to carbon dioxide with the presence of adequate temperature, humidity and
microbial activity”. However, due to their high cost, many plastic products that tend to
be disposed in short service life, such as trash bags and disposable utensils, are still being
made with traditional (non-biodegradable) plastics.

One of the ways to reduce the production cost of biodegradable plastics is to make
the polymer synthesis to be continuous. Traditionally, a synthesis of plastics involves
batch polymerization, which is a big stirred pot versions of beakers or flasks9. This
method is rather costly since it involves series of several steps: first, a container is filled
with reactant, followed by the polymerization, then emptying, and sometimes even
cleaning, the container. On the other hand, when a polymerization is carried in an
extruder, called reactive extrusion polymerization, polymer can be produced
continuously, which not only reduces the processing cost, but also provides several

advantages over batch polymerization, which will be discussed in the following section.

1.1 Reactive Extrusion Polymerization

In addition to the continuous process that reactive extrusion polymerization can
provide, use of extruder is ideal in carrying out a polymerization since it can provide
adequate mixing to achieve homogeneous material, as well as good control over
residence time distribution and temperature profile. Also the same extruder can be used
to combine the polymerization and blending processes which eliminates the extra steps
that are typically involved in the production of blended materials9’30.

There are many types of twin screw extruders commercially available. Their
difference can be categorized in terms of the degree of intermeshing, the sense of rotation
(co- or counter-rotating), and the fimctions that the screws are designed to perform”.

Different types of twin screw extruder will be discussed in Chapter 2.

1.2 Process Conditions
Use of extruder allows one to customize process conditions for each application

easily. Parameters that can be adjusted include:

Reaction temperature

Screw configuration
Screw speed

Number of vent ports 2

Die size

Most of the extruders are equipped with a temperature control cooling system so
the reaction can be carried isothermally. Also the extruders are usually divided into

number of “zones” and their temperatures are controlled individually. Therefore the

temperature of each zone can be varied to design a temperature profile suitable for the
process of interest. This feature became convenient in the polyurethane production,
which will be discussed in Chapter 4.

By varying screw configuration and screw speed, level of mixing and residence
time distribution can be controlled. Detail of screw configuration will be discussed in
Chapter 2.

Some of the “zones” are equipped with vent ports, which can be opened or closed
for a particular application. For example, when the starch is being processed, one of the
vent ports must be open to release the steam generated by heating the starch. On the
other hand, vent ports have to be tightly closed for the production of star-
polycaprolactone since it must be carried under nitrogen environment. Vent ports can
also be used to feed additional material at the middle of the extruder. This feature was
also useful in the case of polyurethane production. Although the same die type was used
throughout this study, it can be changed depending on the application.

Many features of extruder makes the polymerization in extruder more
advantageous, compared to the traditional batch process. However, there are also some
unique factors that have to be considered when planning a reactive extrusion
polymerization.

One of the examples of the factors to be considered is the viscosity of the
materials to be rmde or processed. In the case of batch polymerization, viscosity of
materials can be as low as liquid solvent or as high as viscous polymer (in latter case, the
solvent can be used to dilute the polymer for further processing). However, in an

extruder, a material of low viscosity may cause “extruder underload” and may result in

inadequate mixing of the materials. On the other hand, if a material has extremely high
viscosity, the force required to turn the screws in the extruder (which is usually indicated
as “torque %”) exceeds the acceptable level and results in an automatic shut-off of the
extruder.

Also the time required to complete a polymerization have to be considered when
it is carried out in an extruder. If the reaction takes more than 1000 seconds, the use of

extruder is unlikely to be economic, because of high capital cost of an extruder9.

1.3 Project Overview

By choosing the appropriate reaction chemistry, the process conditions, and the
extruder types, various types of materials can be produced by reactive extrusion process.
In this study, three types of firlly or partially biodegradable plastics were produced with
reactive extrusion polymerization.

The first reaction conducted was reactive extrusion polymerization of e-
caprolactone to produce star-polycaprolactone (REX-PCL). Since the procedure to make
REX-PCL was already developed by Molnn Krishnan in 199830, the focus of the study of
this topic was to reproduce REX-PCL by following his procedure, and characterize the
product. The methods used to characterize REX-PCL were measurement of intrinsic
viscosity, thermogravimetric analysis (TGA), and differential scanning calorimeter
(DSC). Additionally, linear polycaprolactone was purchased and same characterizations
were performed and compared with REX-PCL. Also, Mark-Houudnk-Sakurada (MHS)
constants were determined fi'om the intrinsic viscosity of REX-PCL and linear PCL. This

topic will be discussed in Chapter 3.

The second reaction involves the production of polyurethane by reacting star-
polycaprolactone with diisocyanate. This was done by conducting polymerization of
REX-PCL, followed by the reaction of REX-PCL and diisocyanate continuously within
the same extruder. The diisocyanate used in this experiment was hexamethylene
diisocyanate. The characterization of polyurethane produced were done by measurement
of intrinsic viscosity, TGA and DSC. This topic will be discussed in Chapter 4.

The last reaction conducted was the transesterification of REX-PCL with
polyethylene terephthalate (PET). Catalysts used in this experiment were
triphenylphophine, dibutyltin dilaurate, and titanium (IV) isopropoxide. Products of
transesterification were analyzed with TGA and DSC to compare the effect of three
difi‘erent catalysts. This topic will be discussed in Chapter 5.

Since some of the topics, such as the extruder type, screw configurations, and
method of polymer characterizations, are applicable to multiple experiments, those topics

are discussed in Chapter 2.

Chapter 2

 

BACKGROUND

 

2.1 Twin Screw Extruders

Twin screw extruders have been used for processing of viscous materials for over
a century, and their applications in the plastic industry started in the 19308 and 408 with
the efforts of Colombo, Pasquetti, Meskat and Erdmenger, Leistritz, Fuller, and otherss’Z'.
Since then, wide varieties of applications have been found for twin screw extruders, and a
number of different extruder geometries have been developed to meet each application
requirements5 .

Currently, the geometry of a twin screw extruder mainly depends on the following

specifications:

co-rotating or counter-rotating
- fully, partially, or non-intermeshing
- conjugated or non-conjugated screw profile
- conical or cylindrical geometry
- equal or unequal screw length
- matched or staggered screw configuration
A twin screw extruder is called co-rotating if the screws rotate in the same
direction, and it is called counter—rotating if the screws rotate in opposite direction. Both
types of extruders can have various degrees of intermeshing, which can be distinguished

in three levels: fully intermeshing, partially intermeshing, and non-intermeshing. Non-

intermeshing can be further subdivided into non-intermeshing with distance between two
screws, and non-intermeshing without distance between two screws (tangential). Figure
2-1 illustrates the co-rotating and counter-rotating twin screw extruder with various

degrees of intermeshingss.

ea ”co,

' (a) (b)

_
(d)

(C)

(e) (f)

Figure 2-1: A schematic of co-rotating and counter-rotating twin screw extruder with
various degree of intermeshing: (a) counter-rotating, fully intermeshing; (b)
co-rotating, fully intermeshing; (c) counter-rotating, partially intermeshing:
(d) co-rotating, partially intermeshing; (e) counter-rotating, tangential; (1) co-
rotating, tangential.

Fully intermeshing extruder can be either closely fitting with wide flights

(conjugated screws) or open with narrow flights (non-conjugated screws). Extruders with

conjugated screws are generally operated at low screw speeds, around 10 to 20 rpm, to

prevent high pressure that can build up in the intermeshing region. On the other hand,

extruders with non-conjugated screws can be operated at high speeds, as high as 600 rpm,
because of its rather open structure. One of the disadvantages of the extruders with non-
conjugated screws is the possibility of back-flow at a low screw speed due to the poor
sealing of the channels by the flights. Figure 2-2 shows geometry of the conjugated and

non-conjugated co-rotating twin screwss.

 

 

 

(a) (b)

Figure 2-2: A schematic of conjugated and non-conjugated co-rotating twin screws: (a)
conjugated; (b) non-conjugated

Although a cylindrical geometry with equal screw length, which has the same diameter
throughout the extruder, is the most popular type of twin screw extruder, conical
geometry and an extruder with unequal screw length are also available. A conical
extruder has a large diameter at the feed end and it is gradually reduced toward the
discharge end. This kind of extruder is particularly useful in the case of a low bulk
density feed stock or a product with high viscous heat generation. Extruders with
unequal screw length has a characteristic of generating high pressure at the single screw
area and it is particularly useful in the case of non-intermeshing twin extruder. Figure 2-
3 shows the conical twin screw extruders and the cylindrical extruder of unequal screw

length5 .

 

 

(a)

 

 

 

 

 

 

 

 

 

 

 

 

(b)

Figure 2-3: A schematic of (a) conical twin screw extruders and (b) cylindrical extruder
of unequal screw length.

Non-intermeshing twin screw extruders can also be matched or staggered screw

configurations as illustrated in Figure 2-4.

 

Figure 2-4: A schematic of (a) matched screw configuration and (b) staggered screw
configuration.

Twin screw extruders can be designed for any particular applications by
combining above specifications, though not all combinations are commercially available.
For example, all twin screw extruders are either co-rotating or counter-rotating, and they
are firlly, partially or non-intermeshing. However, a conical geometry is currently
available only for a counter-rotating extruder and an extruder of unequal screw length is
only available for a non-intermeshing counter-rotating extruders.

In additional to the extruder specifications, a combination of screw elements,
called screw configuration, can be varied to provide various levels of mixing, different

conveying rates, and different residence distribution and time.

In this study, Werner Pfleiderer twin screw extruder ZSK—30 was used. The type
of this extruder is: 30 mm co-rotating, fully intermeshing screws with barrel length of

approximately 960 mm. A photograph of this extruder is shown in Figure 2-5.

 

Figure 2-5: Photograph of Werner Pfleiderer twin screw extruder ZSK-30.

2.2 Screw Configuration

Screw configuration in a twin screw extruder primarily consists of two types of
elements: screw elements and kneadling blocks.

The screw configuration mainly consists of right-handed screw elements, also
called conveying elements, and they are primarily used to convey a material through an
extruder with little mixing. They are available in various axial length, pitches, and
different flight types such as single, double or triple flight to provide different level of
conveying capacity. The screw element is also available in reverse flight (left-handed),
which is used for sealing purpose: the flow of a material is stopped at a left-handed screw
element until the upstream screw sections over a certain distance are completely filled to
build enough pressure to overcome left-handed screw. The type of screw element is
usually represented by a set of two numbers, such as X/Y, where X represents the pitch of
the screw in millimeter (mm) and Y is the axial length of the element (in mm). In the
case of left-handed elements, “LH” follows the two numbers”).

Another type of element, a kneadling block, is used to provide various levels of
mixing with less conveying capability than screw elements. They are available in various
staggered angles and pitches to provide different level of mixing. Kneadling blocks can
also be reverse staggered to provide greater level of mixing. In this case, the type of
element is usually represented by a set of three numbers, such as X/Y/Z, where X is an
angle representing the degree of stagger, Y is the axial length of the element (in mm), and
Z is the pitch (in mm). Just as screw elements, reverse staggered element is represented

with “LI-I” following three numbers”.

11

In this study, two types of screw configurations were used. One of them consists
of about 50% kneadling block and the remaining of screw elements, (screw
configuration#l), and another one consists only of screw elements, (screw configuration

#2). Schematics of those two screw configurations are shown in Figure 2-6.

PKR/IO
20/ 10
2 — 42/42

PKR/l 0
20/10
2 - 42/42

6 — 2808 4 —- 28/28

2—20/10

KB 45/5/20
KB 90/5/28
KB 45/5/42
20/10

KB 90/5/28
20/10 LH
20/20

KB 45/5/14
20/10

KB 45/5/14
20/20

KB 45/5/14 LH
20/20

KB 45/5/14
20/20

KB 45/5/14 LH
20/20

KB 45/5/14
20/10

KB 45/5/14
20/20

KB 45/5/14
KB 45/5/14 LH
20/20

KB 45/5/14
KB 90/5/28
KB 45/5/14 LH
20/10 LB

2 — 42/42

21 -—20/20

28- 14/14

28/14
6 — 20/20

 

(a) (b)

Figure 2-6: Screw configurations used in this study: (a) screw configuration #1; (b)
screw configuration #2.

12

Screw configuration #1 provides good combination of conveying and mixing, and
it can be used in wide variety of applications. On the other hand, screw configuration #2
provides large conveying capacity with little mixing. Screw configuration #2 was
developed by Krishnan during his study of REX-PCL3o in order to resolve the processing
difficulty encountered with the use of screw configuration #1.

The results of Krishnan’s study show that, due to the extra hold-up time in the
kneadling blocks in screw configuration #1, viscosity of polymer was increased
significantly and resulted in complete stoppage of material flow within the extruder. On
the other hand, since there is no kneadling block used in screw configuration #2, the high
viscosity problem found in screw configuration #1 was not observed in the experiments

with screw configuration #2.

2.3 Characterization of Polymers

Once a new material is developed, its properties must be evaluated and usually
compared with the properties of already known materials to verify the proposed reaction.
Analysis of properties of newly developed material is also important in determining the
applications for which the materials can be used.

There are various analytical and evaluative methods currently available. Many of
them are equipped with high technology device with computer programs (such as
spectroscopic, thermal, and mechanical methods) while some still require more manual
processes (such as the determination of intrinsic viscosity using glass viscometer and

determination of fi'actional conversion with the extraction method). Although there is no

13

single test that can provide all the answers needed, one can obtain a good picture of the
type of material the person is dealing with by combining the results of various tests.
The methods used in this study were measurements of intrinsic viscosity using
glass viscometer, thermogravimetric analysis (TGA), differential scanning calorimeter
(DSC), and the determination of fi'actional conversion. The following sections discuss

each type of analytical method and their standard procedure used in this study.

2.3.1 Measurement of Intrinsic Viscosity
Intrinsic viscosity or limiting viscosity number, [1]], reflects the contribution of
the polymeric solute to the difference between the viscosity of the mixture and that of the

solvent', and it is defined as:

lim

misc- l 1110.

no
The intrinsic viscosity is constant for a particular polymer/solvent combination at a given
temperatureI . Use of glass viscometer, such as Ubbelohde and Cannon-Fenske, is one of
the simplest and widely used technique to determine intrinsic viscosity, using dilute
polymer solutions”.

Each type of viscometer is available in different sizes, and the type and size to be
used depend on the range of viscosity of particular polymer/solvent combination to be

measured. Ubbelohde viscometer was used in this study since it is one of the most

widely used viscometer, and the size 0C, which is the second smallest size of Ubbelohde

14

commercially available, was selected because of the low viscosity of polymer/solvent
combination that was used in this study.

Following section discusses the procedure used to determine intrinsic viscosity
using Ubbelohde viscometer. Although similar procedure can be used with other types of
viscometers, it is recommended to read the publication by American Society for Testing

and Materials (ASTM) for the specifications applicable to the viscometer of interest.

2.3.1.1 Standard Procedure

First, the dilute polymer solution of known concentration was made by dissolving
pre-weighted polymer (between 0.1 to 1.5 g depending on the type of polymer) into
known volume of solvent, usually between 50 to 100 ml. There is no particular
concentration that can be used for all types of polymers. In other words, any
concentration can be used as long as the following specifications are met: efilux time
(will be discussed later in this section) is over five minutes and under ten minutes (latter
is merely for a convenience, and not a requirement). Toluene was used as a solvent
throughout this study: 1 g of polymer per 100 ml of toluene was usually used as an initial
concentration.

Both Ubbelohde and solution to be tested must be particle-flee, since the diameter
of capillary inside the Ubbelohde is very small that even a small particle can be trapped
to give inaccurate result.

In order to clean Ubbelohde, the solvent was passed through all three tubes,
followed by dry air to remove the final trace of the solvent. Then, enough dilute polymer

solution to be tested was introduced through The largest tube (tube A), so level of the

15

solution is between two lines on the bulb at the bottom of the tube A. The Ubbelohde
was placed into a constant-temperature water bath, so the water level is above the bulb on
the tube consisting the capillary (tube B). Then the Ubbelohde was left for least 20
minutes to bring water and solution temperature to be at equilibrium. A temperature of
26 °C was used throughout this study. After approximately 20 minutes, the Ubbelohde
was vertically aligned inside the water bath, then a finger was placed over the third tube
(tube C) and suction was applied to tube B until the solution reaches to the bulb on the
tube B. Suction was then removed fi'om tube B, followed by removing the finger fiom
tube C and the solution was allowed to flow freely. The efflux time was determined by
measuring the time required for the level of solution in tube B to drop from one line to
the otherus.

Above procedure was repeated at least three times for each solution and the
average efflux time was determined.

Determination of intrinsic viscosity can be done by applying either the Huggins
equationl’zgz
c"(n/no — 1) = [n] + knlnlzc

or the Kraemer equation” ':

c"ln(n/no) = [n] - k: [1.1%
where c is the concentration of the polymer solution tested, 11 is the solution viscosity, 1].,
is the solvent viscosity, [1]] is the intrinsic viscosity and 1(1-1 and k1 are the proportional
constants associated with each equation. Both constants depend on the type of polymers

being tested. For dilute solutions, it can be said that:

n/no = t/to

16

where t is the efilux time of dilute polymer solution and t0 is the efflux time of solvent
alone‘. Both t and t0 were determined as discussed above.

In order to determine the intrinsic viscosity, lefi-hand-side of Huggins equation
and the Kraemer equation were plotted against c and straight lines were drawn through
the points for each equation. Lines of both equations yield at the same y-intercept, which

represents the intrinsic viscosity of particular polymerl as shown in Figure 2-7.

 

 

c"(n/n. - I)

 

 
   

c"In(n/n..)

[no]

0-.
n.-
0..
.0
0..
o
......
'-
co.
0..
on.
an...

 

 

Lefi-Hand-Side of Equations

 

Concentration

 

 

 

Figure 2-7: An example of graphs used to determine the intrinsic viscosity

Determined intrinsic viscosity were then used for the comparison of different

materials and to determine the Mark-Houwink-Sakurada (MHS) constants.

2.3.2 Determination of Mark-Houwink-Sakurada (MHS) Constants
Mark-Houwink-Sakurada (MHS) equation relates intrinsic viscosity ([11]) of a
polymer to a viscosity average molecular weight (M.-) with the equation:

[11] = K My,

17

in which K and a are called Mark-Houwik-Sakurada constants. MHS constants depend
on the polymer, type of solvent used and the solution temperature used for testingl’zs.
Since MHS constants representing wide range of polymers, with different types of
solvents and testing temperature, have been published in various literaturesl"9’28'3°, Mv
can be determined if the intrinsic viscosity of a polymer is determined at the condition in
which MHS constants are known.

If MHS constants are unknown for a particular combination of polymer, solvent,
and testing temperature, they can be calibrated by determining [n] and plotting ln[n]
versus lnMw or lnM“, where Mw is the weight average molecular weight and Mn is the
number average molecular weight. Since 1n [n] is linearly related to lnMw or lnM.1 by the
equation:

ln[n] = an + a lnM
MHS constants can be determined fiom the slope and intercept of the line passing
through the plotted points. Mv then lies between those of the corresponding Mw and Mn,
though it is rather closer to Mw than M... This method, however, is merely an estimation
and determination of exact Mv involves fiactionation to divide the whole polymer sample
into subspecies of relatively narrow molecular weight distributions, then averaging
measured Mn and MW to calculate le’zs.

In general, MHS equation can be applied to estimate MW or M of unknown
samples only if a fi'action of a molecular weight distribution is very similar to the sample
used for the calibration. However, this restriction does not apply to the polymers with

random molecular weight distribution, such as linear polyamides and polyesters

18

polymerized under equilibrium conditions. For these types of polymers, Mw = 2Mn and
the relation

[11] = KMn‘
can be applied'.

In this study, MHS constants of linear and REX-PCL were determined by
applying the equation above based on two assumptions: (1) a fraction of a molecular
weight distribution is very similar between the same type of polymer of different
molecular weight; and (2) both linear and REX-PCL have random molecular weight
distribution. The latter assumption was based on the report provided by Daisel Chemical
Company, which showed that the polydispersity index (MW/Mn) is close to 2 for both
linear and REX-PCL”.

MHS constants has been reported in literatures for linear polycaprolactone with
various solvents and temperatures, as well as for some types of star-polymers, such as
star-polystyrene'g’”. However, the values of MHS constants could not be found for the

star-polycapro lactone (REX-PCL).

2.3.3 Thermogravimetric Analysis

Thermogravimetric analysis (TGA) is primarily used to determine the thermal
stability of polymer by continuously measuring the weight of polymer as the sample
temperature is increased (non-isothermal TGA). The increase in the sample temperature
results in a weight loss of the sample. At a low temperature, the weight loss may be due
to the evaporation of moisture and solvent. However at higher temperatures, it is a result

of a polymer decomposition. Therefore, by monitoring sample weight vs. temperature,

19

2.28

degradation temperature and the rate of degradation of the sample . A schematic of

TGA output is shown in Figure 2-8.

 

 

Weight (%)

 

 

 

 

Temperature

Figure 2-8: A schematic of typical output of Thermogravimetric Analysis

The determination of degradation temperature and the rate of degradation arenot
only useful to learn about the thermal stability of the material, but also useful in
characterizing the unknown materials. This is because the degradation temperature and
the rate of degradation are different for many materials, due to the difi‘erence in their
degradation mechanism. By comparing the results of TGA before and after the reaction

can provide some degree of information about the unknown materials.

20

In this study, TGA was utilized to determine the thermal stability of the materials
produced, and to observe the result of reaction attempted, such as in the case of
polyurethane production and transesterification. A standard procedure was developed for

these purposes and used throughout this study.

2.3.3.1 Standard Procedure
TA Instruments Hi—Res TGA 2950 Thermogravimetric Analyzer was used

throughout this study. A photograph of this device is shown in Figure 2-9.

 

Figure 2-9: A photograph of TA Instruments Hi-Res TGA 2950 Thermogravimetric
Analyzer '

Heating the sample and measuring the sample weight was entirely controlled by the

computer program; however, the initial and final temperature, as well as the rate of

heating was specified for the type of materials being used in this study and those

conditions are shown in Table 2-1.

21

 

 

 

 

 

 

Initial temperature 25 °C
Final tempareture 500 °C
Rate of heating 25 °C/min

 

Table 2-1: Heating condition of TGA used in this study.

Results of TGA were plotted in weight % vs. temperature (in °C). Also, the
degradation point (in °C) and the rate of weight loss (in weight %/°C) were determined,

using the computer program

2.3.4 Differential Scanning Calorimeter (DSC)

In DSC, a polymer sample and an inert reference, usually an empty pan, are
heated and thermal transitions in the sample are measured and recorded. This process is
usually carried in a nitrogen atmosphere and the same type of pans must be used for
polymer sample and inert reference. Aluminum pan is most commonly used; however,
other materials such as gold, graphite, and platinum are also available.

With DSC, the sample and the reference are heated separately, with separate
heating source so the temperature increases with the rate specified by an operator. Since
energy required to increase the temperature of the sample is different fi'om the energy
required to increase the temperature of reference, the difl‘erence between two electrical
power used for heaters (dAQ/dt) is measured and recorded'z’zs. Figure 2-10 shows a

schematic of DSC output.

22

 

  

Exothermic

 
 
 

Heat Flow (Wig)

Endothermic

   

 

 

 

 

Team (‘C)

Figure 2-10: A schematic of output from Difl'erential Scanning Calorimeter

Just as the thermal stability of the materials, thermal transition characteristics such
as glass-transition temperature and melting temperature, are also different for different
materials. Therefore, by comparing the output fi'om DSC, one can determine whetheran
attempted reaction had taken place. This feature is particularly useful when analyzing the
result of transesterification. This is because when two materials are mixed without any
reaction in between, there are two melting temperatures associated with each material
being mixed. However, once appropriate reaction has taken place between the mixed

materials, there will be one melting temperature, which is most likely at the temperature

23

between the melting temperatures of individual materials. Figure 2-1 1 shows a schematic

of this phenomenon.

 

 

 

 

 

 

 

 

 

 

 

 

 

3
2
2‘:’
Temperature
(8)
3
u-
i i
Temperature
0))
Figure 2-11: A schematic of how DSC output changes before (a) and after (b) the

transesterification

24

In this study. three types of heating schemes were programmed and utilized for different

materials.

2.3.4.1 Standard Procedure
TA Instruments DSC 2920 Modulated DSC was used throughout this study. A

photograph of this device is shown in Figure 2-12.

ii»
’ 5‘

»'.~III"

“4...“..-A’wauap ' ‘

 

Figure 2-12: A photograph ofTA Instruments DSC 2920 Modulated DSC

Heating the sample and measuring the difference in energy were entirely controlled by a
computer program; however, initial and final temperatures, as well as the rate of
temperature increase were specified for the type of polymers used in this study and they
are shown in Table 2-2. Three types of heating conditions were designed and chosen

depending on the materials to be analyzed.

25

(a) Type 1 heating condition

 

 

 

 

 

Initial temperature -60 °C
Final tempareture 100 °C
Rate of heating 10 °C/min

 

 

(b) Type 2 heating condition

 

 

 

 

 

 

 

 

 

 

Initial temperature -60 °C

Final tempareture 300 °C

Rate of heating 10 oC/min
(c) Type 3 heating condition

Initial temperature 40 °C

Final tempareture 300 °C

Rate of heating 10 °C/min

 

 

 

 

Table 2-2: Three types of heating conditions of DSC used in this study: (a) Type 1; (b)
Type 2; and (c) Type 3.

Type 1 heating condition was primarily used for the analysis of both linear and
REX-PCL, since melting temperature of those two materials are approximately 60 °C.
On the other hand, higher final temperature was chosen for Type 2 heating condition
since melting temperature of polyethylene terephthalate (PET) is approximately 250 °C.

For both Type 1 and 2, low initial temperature (-60 °C) was used in order to
observe the thermal transitions that might occur at low temperature.

Type 3 was used for the analysis of products from transesterification. In this case,
the initial temperature was chosen to be at 40 °C since the focus of this analysis was to
compare different transesterification products, and since there was no significant thermal
transition observed below that temperature for either PET or REX-PCL.

Results of DSC were plotted for heat flow (in W/g) vs. temperature (in °C). Also,

some characteristic peaks were determined by using the computer program

26

2.3.5 Fractional Conversion

When a product seems different from reactants, it does not mean the reaction went
to completion. Therefore, it is important to determine the fiactional conversion of
reactants as a part of the characterization of product.

There are various methods that can be used to determine the fractional conversion
such as Gas Chromatography method and dissolution-precipitation gravimetric method”.
The latter method was used in this study to determine the fiactional conversion of REX-

PCL.

2.3.5.1 Standard Procedure

In dissolution-precipitation gravimetric method, unreacted monomer is separated
from the polymer by precipitation and the difl‘erence in weight before and after the
extraction can be used to determine the fi'actional conversion of monomer.

This experiment was done, first, by dissolving a known amount of REX-PCL
(approximately 5 g) in 50 m1 of toluene, leaving it for 24 hours at room temperature, then
precipitating dissolved REX-PCL with approximately 500 ml of heptane. The precipitate
was then filtered and left inside a chemical hood for a few days to allow heptane to
evaporate. The remaining polymer was then weighted and compared with the weight .
before the precipitation. The fiactional conversion of e-caprolactone monomer was

determined by the equation:

Fractional conversion = Weight after extraction/ Weight before extraction

27

Chapter 3

 

REACTIVE EXTRUSION POLYMERIZATION OF

EPSILON -CAPROLACTON E

 

3.1 Introduction

Polycaprolactone (PCL) is a semi-crystalline polymer and it belongs to the family
of polyesters. Because of its hydro lyzable ester group, polycaprolactone and other
aliphatic polyesters are biodegradable, which makes them environmental fi'iendly. Other
attractive properties of PCL include its miscibility with number of polymers (such as
polycarbonate, polyvinyl chloride, and cellulose propionate) and its compatible
mechanical properties to more common polymers (such as polyethylene)”. However,
commercially available linear PCL also has some drawbacks, such as its high cost, low
melting temperature (about 60 °C), and low melt strength. In order to overcome these
disadvantages, a process of synthesizing high molecular weight star-polycaprolactone

(REX-PCL) in extruder was developed.

3.2 Background

A process of making REX-PCL was developed by Mohan Krishnan in 1998, and
the detail of the process was documented in his doctoral dissertation, titled Engineering
of Compositions Based on Star-Polycaprolactone Derived By Extrusion Polymerization

of Epsilon-Caprolactone3 0. This process involves a coordination insertion reaction with a

28

ring-opening polymerization of e-caprolactone monomer, using aluminum tri-sec

butoxide as an initiator. The reaction mechanism is shown in Figure 3-1.

on
0 03‘3““ on
+ A1(OR)3 —— \C 0

OR
0 3A]:- OR
9" 9 <‘°
o - - -
R “(o (0H,)5 coon

 

OR

‘EH3

OR = OCHCHZCH3

Figure 3-1: Ring-opening polymerimtion of e-caprolactone to produce REX-PCL
The structmeofREX-PCLisso-called“star” shapeduetothepresenceofaluminmn.

atom which connects three PCL molecules. The structures of linear and REX-PCL are

shown in Figm'e 3-2.

29

0
II
(a) ‘( O - (CH2)5 - C Tn—

0

II
o - (CH2)5 - C )rr—OH
O 0

11
HO—-( O - (CH2)5 '- C )E‘Al_|_( O _ (CH2)s - ('3 )a—OH

(b)
Figure 3-2: The structures of (a) linear PCL and (b) REX-PCL

In his study, wide range of extrusion temperatures were evaluated, ranging from
150 to 220 °C, with two grades of e-caprolactone, TONE EC and TONE ECEQ. The
difference between TONE EC and ECEQ are their purity: purity of ECEQ is higher with
lower free acid contents”. The results of his study showed that polymerization with
ECEQ at the extrusion temperature of 190 °C was found to give best monomer
conversion (98 % for the molecular weight of both 80,000 and 120,000). Additionally,
screw configuration #2, which consists only of conveying elements, was developed for
the production of REX-PCL. The process was developed using Werner Pfleiderer twin
screw extruder ZSK-30, and the theoretical molecular weight of REX-PCL produced .
range from 80,000 to 137,000.

In this study, production of REX-PCL was started by reproducing the procedure
developed by Krishnan, then some adjustment in the process conditions were made in

order to produce the REX-PCL with wider range of molecular weight.

30

3.3 Outline of Chapter

This chapter focuses on the production and characterization of REX-PCL of
various molecular weight, and it is subdivided into four sections: (1) Reactive Extrusion
Polymerization of e-Polycaprolactone to Produce REX-PCL of Various Molecular
Weight (section 3.4); (2) Characterization of Star-Polycaprolactone (section 3.5); (3)
Reactive Extrusion Polymerizatoin of e-Polycapro lactone from Monomer with Impurities
(section 3.6); and (4) Monomer Conversion and Reproducibility (section 3.7).

The first topic discusses the process of making REX-PCL by reproducing the
method developed by Krishnan""”3 1, and the determination of upper and lower limits of
molecular weight that can be produced. Various observations made during the process of
making REX-PCL will also be discussed in this section.

Next section discusses the characterization of REX-PCL of various molecular
weight and compare the result with the one of linear PCL. Method used to characterize
the polymers were: determination of fiactional conversion, measurement of intrinsic
viscosity, Thermogravimetric Analysis (TGA), and Differential Scanning Calorimeter
(DSC). Also, Mark-Houwink-Sakurada (MHS) constants for both linear and REX-PCL
were determined at temperature of 26 °C with toluene as a solvent.

Additionally, in order to determine the effect of impurities on the formation of ,
star-polycaprolactone, the same procedure was used to produce REX-PCL with the
monomer with some water contents. A process of making REX-PCL with the monomer

with impurities is discussed in section 3.6.

31

Finally, the monomer conversion within the extruder, as well as the
reproducibility of the process of making REX-PCL were determined from the intrinsic

viscosity and they will be discussed in section 3.7.

3.4 Reactive Extrusion Polymerization of e-Caprolactone to Produce REX-PCL
of Various Molecular Weight
3.4.1 Objective
Objectives of this experiment were: (1) to make REX-PCL with wide range of
molecular weight; (2) to determine the lower and upper limit of molecular weight that can

be produced from the same system.

3.4.2 Materials
e-caprolactone monomer (TONE ECEQ) was purchased fiom Union Carbide and
used throughout this study, except the experiments discussed in section 3.6 in which

contaminated monomer was used instead. A structure of e-caprolactone is shown in

Figure 3-3.

Figure 3-3: A structure of s-caprolactone

32

Aluminum tri-sec butoxide (ATSB) was used as an initiator and its structure is

shown in Figure 3-4.

CH3
' l
o CHCH,CH3
I

Al .
CH3CH2 CH 0 ’ ‘o CHCH2CH3
I l
CH3 CH3

Figure 3-4: A structure of aluminum tri-sec butoxide

Although the polymerization can take place without the presence of any solvent,
ATSB was mixed with anhydrous toluene (ATSB solution hereafter). Purpose of mixing
ATSB with anhydrous toluene was to increase the flowrate, since required flowrate of
ATSB by itself is too small (less than 0.1 g/min in most cases) to be pumped with
commercial pump. Also the addition of anhydrous toluene resulted in the better
distribution of ATSB throughout the monomer. This was important because the screw
configuration used in this experiment lacked mixing characteristics. However, the use of
anhydrous toluene also had negative effects. For example, use of toluene increased the
production cost, and is also environmentally unfavorable. Furthermore, some amounts of
toluene remained in the product had to be evaporated. Therefore, the dilution of ATSB
with anhydrous toluene was limited to 5 wt% ATSB.

Note that anhydrous toluene was used instead of a “regular” toluene also to

prevent a contact of water with ATSB. Since both monomer and ATSB are sensitive to

33

moisture, preparation of ATSB solution as well as the production of REX-PCL were

carried under dry nitrogen.

3.4.3 Procedure
Procedure discussed in this section was applied to the production of REX-PCL

throughout the study.

3.4.3.1 Flowrate Calculation

In most cases, production of REX-PCL starts with the objective of making a
REX-PCL of certain molecular weight. For this reason, calculation of monomer and
ATSB flowrate also starts with choosing desired theoretical molecular weight. For a
fixed theoretical molecular weight, degree of polymerization (DP) was calculated with
the equation:

DP = (theoretical MW) x 3 / (MW of monomer)

A term “3” in the above equation came from the fact that ATSB has three arms that can
react with a monomer.

Since DP is a molar ratio of monomer to initiator, it can be represented as:

DP = (molar flowrate of monomer) / (molar flowrate of ATSB)

in which molar flowrate of monomer can be calculated by choosing volumetric flowrate
of a monomer. A choice of volumetric monomer flowrate depends on the size of pump
head and tube to be used, as well as the rate of production of REX-PCL to be achieved.

Latter is especially important in order to make this process cost effective.

34

Once volumetric flowrate of monomer is chosen, mass flowrate of monomer can
be calculated with the equation:
mass flowrate of monomer = (volumetric flowrate of monomer) x pmmm,
where pmmm, is the density of monomer. The molar flowrate of monomer was then
calculated as:
molar flowrate of monomer = (mass flowrate of monomer) / (MW of monomer)
With a known DP and a molar flowrate of a monomer, a molar flowrate of the
ATSB can be calculated as:
molar flowrate of ATSB = (molar flowrate of monomer) / DP
which can be converted to mass flowrate by:
mass flowrate of ATSB = (molar flowrate of ATSB) x (MW of ATSB)
Since the mass flowrate of the ATSB itself is very small (less than 1 g/min for most of
the experiment done in this study), ATSB was mixed with anhydrous toluene to make the
ATSB solution. Then, the flowrate of ATSB solution was calculated as:
mass flowrate of solution = (mass flowrate of ATSB) / (fiaction of ATSB in solution)
which can be converted into volumetric flowrate of a solution by the equation:
volumetric flowrate of ATSB = (mass flowrate of solution) / pmimion
where psoluu‘on is the density of ATSB solution, which can be calculated as:
psoimion = 1/[(frac of ATSB/puss) + (the of Toluene/pm...”

where pwim is the density of anhydrous toluene.

An example of the calculation sheet used in this study is shown in Table 3-1.

35

 

 

Flowrate Calculation for REX-PCL ' I pick value

.1.

Date:
Run#:

Theoretical Mw = "

e-Caprolactone Monomer

Volumetric Flowrate of Monomer = 2 i ""13 ml/min

a. xvi
pm,= 1.030 g/ml

Mass Flowrate of Monomer = Vol F lowrate of Monomer "‘ pmm = g/min
MW of Monomer = 114.14 g/mol
Molar F lowrate of Monomer = Mass F lowrate / MW of Monomer = mol/min
Aluminum tri-sec Butoxide
Molar Flowrate of ATSB = Molar Flowrate of Monomer / DP = mol/min
MW of ATSB = 246.33 g/mol

Mass Flowrate of ATSB = Molar Flowrate * MW = g/min
Fraction of ATSB in initiator solution =3; ' 1‘
Mass Flowrate of Solution = Mass F lowrate of ATSB / frac of ATSB =11 I g/min
p of ATSB = 0.967 g/ml
p of Toluene = 0.865 g/ml
p of Solution = l/[(frac of ATSB/plugs) + (free of Toluene/gdmfl = g/ml

Volumetric Flowrate of Solution = Mass Flowrate of Sol. / p of Sol. = ml/min

 

Table 3-1: An example of the calculation sheet used to determine the monomer and
initiator flowrate

36

 

An example of the various monomer and the initiator flowrate to yield different

theoretical molecular weight are shown in Table 3-2.

 

 

 

 

 

 

 

 

 

 

Theoretical Monomer ATSB Solution % ATSB in the
Moleclar Weight Flowrate (m1/ min) Flowrate (ml/min) Solution (wt%)
30,000 40 2.6 5
30,000 80 5.2 5
30,000 20 1.3 5
30,000 80 1.7 15
60,000 50 1.6 5
60,000 1 50 1 .6 1 5
100,000 80 1.6 5
100,000 80 0.8 10
100,000 40 0.8 5

 

 

 

 

 

 

Table 3-2: An example of the various monomer and the initiator flowrate to yield
different theoretical molecular weight of REX-PCL
3.4.3.2 Preparation of ATSB Solution

A preparation of the ATSB solution was carried under nitrogen environment in a
glove box/bag. In this section a preparation of ATSB in glove bag will be discussed,
however, similar procedure can be followed in the case of using a glove box.

ATSB content of 5 and 15 wt% in anhydrous toluene were mostly used in this
experiment; however, these values were chosen merely for the convenience, and any
concentration can be used depending on the molecular weight to be made and the type of
the pump to be used.

In order to make the ATSB solution, a tube was first connected between the
nitrogen cylinder and the glove bag, where all the required materials were placed. Then,

the nitrogen flow was started and left for a few minutes to purge inside the bag.

37

Once purging was done, the glove bag was closed and the bag was filled with
nitrogen. Then a scale was used to prepare a catalyst solution of a desired concentration
(in weight percent) in a 2 L flask. The ATSB and the anhydrous toluene bottles as well

as the flask containing the catalyst solution were tightly closed with parafilrn prior to

opening the glove bag.

3.4.3.3 Flow Scheme
Flow scheme of e-caprolactone monomer, the ATSB solution, and the nitrogen

gas are discussed in this section.

a-Caprolactone Monomer

Schemcatic of monomer flow is shown in Figure 3-5.

38

 

 

   

seale

feed throat T LC;

 

monomer ATSB

N2

 

 

 

 

Figure 3-5: Flow scheme of s-caprolactone monomer

The monomer was first pumped out fi'om the monomer tank (55 gal drum was
usually used) and entered a transfer vessel (2 L flask) on the scale. The use of a transfer
vessel served two purposes: monitoring the flowrate of the monomer, and to prevent from
ruining large amount of monomer (in drum) by accident. From the transfer vessel, the
monomer was pumped into the extruder through the feed throat. A level of monomer in
transfer vessel was kept at about 1.5 L throughout the process. Both the monomer tank
and the transfer vessel were tightly closed and only the dry nitrogen was allowed to flow
through them to prevent the air from entering the system.

Masterflex” L/So Easy-Load” pumps and the Masterflex” L/SO precision tubing

size 15 were used for the monomer flow. The size of the pump head as well as the tubing

39

size were determined by considering the typical monomer flowrate (20 mein to 220

ml/min) used in this study.

ATSB Solution

Flow scheme of ATSB solution is shown in Figure 3-6.

feed throat

 

monomer ATSB

N2

1

’-

 

{\gflg

_____________________________ extruder .1

Figure 3-6: Flow scheme of ATSB solution

 

 

ATSB solution was made in a 2 L flask and was pumped out and directly fed to
the extruder (through feed throat) without using a transfer vessel. The flowrate of ATSB
solution was monitored by measuring the weight of 2 L flask. Just as the monomer tank
and the transfer vessel, a container of ATSB solution was closed tightly and dry nitrogen
was continuously passed through to prevent the air from entering.

Masterflex® L/S® Easy-Load® pumps and the Masterflex” L/S‘” precision tubing
size 14 were used to feed the ATSB solution. The size of the pump head and the tubing
size were determined by considering the typical flowrate of ATSB solution (1.4 ml/min

to 5 ml/min) used in this study.

40

Nitrogen

Flow scheme of nitrogen gas is shown in Figure 3-7.

flowmeters

iii}.

l l

 

linder to extruder t0 initiator outlet tra
flask (mineral 011)
to monomer to monomer i i

 

 

tank transfer vessel ,5" I:

Figure 3-7: Flow scheme of nitrogen gas

Nitrogen was separated into four streams and entered flowmeters. Flowmeters
were necessary in order to monitor the nitrogen flowrate. After passing through the
flowmeters, each nitrogen stream entered into the four separate containers: the extruder,
the initiator flask, the monomer tank, and the monomer transfer vessel. The nitrogen
output from the initiator flask, the monomer tank, and the monomer transfer vessel were
then led into the separate outlet trap containing mineral oil to prevent a back flow of air.

Tygon® tubes were used for the nitrogen streams, and overall nitrogen flowrate

was kept at about 10 psi.

41

3.4.3.4 Start-Up

An extruder was started first and the temperature of all six zones were set to be at
180 - 190 °C. The water bath was filled while waiting.

The nitrogen flow was started before both the monomer drum and the flask
containing ATSB solution were opened. The tubes were carefirlly connected to each of
the containers with the nitrogen flowing through it. Then, the monomer and the initiator
pumps were both started and the flowrate was adjusted to the desired value. The screw
speed was then gradually increased to 110 rpm. The product coming out from the die
was collected in a metal trash can and discarded until the system became stable. The
system was determined to be stable when “torque %” of the extruder stayed in the same
value i2 % for over 20 minutes. The metal trash can was especially important at the
beginning of the experiment since the unreacted feed tend to come out from the die for

the first 10 to 30 seconds.

3.4.3.5 Collecting Sample and Shut Down

When the system stabilized, the product fi'om the die was passed through a water
bath to cool down and pelletized. Air was blown just before the pelletizer to blow the
water off the polymer string. The pelletized polymer was then dried in air for a few days
before characterization.

When the REX-PCL of different molecular weight were collected continuously
without purging (cleaning inside extruder), the flowrate of the monomer and the ATSB

solution were gradually changed, with at least 5 minutes at each flowrate. This process

42

was particularly important in the case of going from a high molecular weight to a lower
one (see section 3.4.5 for explanation).

When the desired amount of samples were collected, the monomer and the
initiator pumps were stopped. The tubes were disconnected, and the containers were
closed. All these steps were done before the nitrogen was turned off, so the air would not
enter the monomer or the ATSB containers. Also at the end, a tube used for ATSB
solution was cleaned by passing toluene through it. The remaining toluene inside the
tube was dried by blowing air through. This last step was necessary in order to prevent

the tube from clogging.

3.4.4 Results

The molecular weight of the REX-PCL produced, as well as the feed rate of the
monomer and the ATSB solution, a fi'action of the ATSB in the solution used, and the
resulted torque % were recorded as shown in Table 3-3. The extruder temperature and
the screw speed were kept constant at the value indicated in the Procedure section

(section 3.4.3.4) for all the molecular weight produced.

 

 

 

 

 

 

 

 

 

 

Theoretical Monomer ATSB Solution % ATSB in Observed
Molecular Weight F lowrate Flowrate Solution Torque %
(each arm) (ml/ min) (ml/min)
20,000 20 2 5 15
30,000 40 2.6 5 21
60,000 50 1.6 5 35
80,000 70 1.7 5 37
100,000 80 1.6 5 41
110,000 80 1.4 5 42
120,000 100 1.6 5 45
150,000 130 1.7 5 48
210,000 150 1.4 5 55

 

 

 

 

 

 

 

Table 3-3: Molecular weight of REX-PCL produced (T = 180 °C, 110 rpm)

43

3.4.5 Discussion

As shown in Table 3-3, the REX-PCL of molecular weight ranging fiom 20,000
to 210,000 were produced only by varying the flowrate of reactants. The melt fiacture of
REX-PCL clearly showed the increase in viscosity as the molecular weight of the
polymer was increased. The increase in the torque % also indicated the same
phenomenon.

Note that in some extrusion experiments, the temperature of zone 1 (the zone
closest to the feed throat) is kept low, such as 60 °C. However, in the production of
REX-PCL, it was important to keep the temperature of zone 1 high, such as 180 °C used
in this experiment, in order to prevent an overflow of reactants at the feed throat. When
the temperature of zone 1 is low, the reaction between the monomer and the ATSB is too
slow that the material of low viscosity in the zones closer to the feed throat can not push
the viscous material at the later zones out fiom the die.

Also by varying the flowrate of the monomer and the ATSB solution, the lowest
and the highest molecular weight that can be produced with current system was
determined. For the lowest limit, a molecular weight of 20,000 was achieved. Although
even the lower molecular weight could be obtained, it is not recommended for this
particular system since “extruder underload” warning appears every few minutes.

Numerous attempts were made to achieve the high molecular weight. However,
the highest molecular weight that could be achieved steadily was found to be 210,000. A
molecular weight higher than 210,000 was unable to be achieved because of the poor
initiator distribution. To make a polymer with high molecular weight, the monomer to

the initiator ratio has to be very high. However, as the ratio increases, the number of

initiator available for the monomer decreases, resulting in some monomer to be left
unreacted. This phenomenon was observed in terms of torque % of the extruder. As the
monomer to initiator ratio hits a certain level, torque % stops increasing or even start to
decrease. In some cases, “oscillation” of torque % was also observed - that is, the torque
% fluctuating between 55% and 35% as the polymer coming out of the die showing
different melt fiacture every few inches. One of the possible solutions to this problem, in
order to make even higher molecular weight, is to add a few mixing elements at the feed
area to increase the initiator distribution.

Note that the color of commercial linear polycaprolactone is white; however, the
color of REX-PCL was found to be light yellow when it is flesh out of the extruder. This
yellowish color faded away as the polymer was sitting in the storage. The yellowish
color was also found to be darker for the REX-PCL of lower molecular weight, and the
color was more like orange than yellow in the case of REX-PCL with molecular weight
of 30,000. Additionally, the fieshly made REX-PCL smelled like a toluene; however,
this too, faded away as the polymer was left in storage for a few days.

These observations suggest that the difference in colors between linear and REX-
PCL was most likely due to the toluene used in the ATSB solution. This also explains
why the color of REX-PCL of lower molecular weight is darker: for the same
concentration of ATSB solution, increasing the flowrate of the solution result in greater

toluene content.

45

3.5 Characterization of Star-Polycaprolactone

Characterization of Star-Polycapro lactone produced were done with four
methods: determination of fi'actional conversion, measurement of intrinsic viscosity,
thermogravimetric analysis (TGA), and diffi'action scanning calorimeter (DSC). The
same characterizations were also done with linear PCL purchased and compared with the
results of REX-PCL. Additionally, Mark-Houwink-Sakurada constants were determined
for both the linear and the REX-PCL from their intrinsic viscosity and the results were

compared.

3.5.1 Determination of Fractional Conversion
3.5.1.1 Objective
Objective of this experiment was to determine the actual molecular weight of the

REX-PCL by determining the fractional conversion of e-caprolactone.

3.5.1.2 Materials

The REX-PCL with the molecular weight of 30,000, 60,000, 80,000, 100,000,
120,000, 150,000, and 210,000 were used in this experiment.

Toluene was used to dissolve, and heptane was used to precipitate the polymer

from the solution.

3.5.1.3 Procedure

Standard procedure described in Chapter 2 was followed.

46

3.5.1.4 Results
Fractional conversion of the monomer was determined for the each REX-PCL
collected. The results were tabulated as shown in Table 3-3. Also, the actual molecular

weight was calculated for each polymer based on the fiactional conversion and also

 

 

 

 

 

 

 

 

shown in Table 34.

Theoretical Mll Fractional Conversion Actuaan standard
(%) deviation

30,000 97.4 29,220 1 .46

60,000 96.3 57,780 0.78

80,000 98.8 79,040 0.22

100,000 98.3 98,300 0.50

120,000 97.5 1 17,000 0.29

150,000 95.3 142,950 0.85

210,000 88.5 185,850 2.30

 

 

 

 

 

 

Table 3-4: Fractional conversion and actual molecular weight of REX-PCL produced

3.5.1.5 Discussion

As shown in Table 3-4, high monomer conversion was achieved for the molecular
weight up to 120,000. This indicates that, for the theoretical molecular weight up to
120,000, the polymerization had taken place almost to completion and the theoretical
molecular weight is very close to the actual molecular weight.

On the other hand, the fiactional conversion of 150,000 is slightly lower than the
ones of lower molecular weight, resulting in the actual molecular weight to be
approximately 5 % lower than the actual molecular weight. The reduction in fractional

conversion is even lower for the theoretical molecular weight of 210,000, though this

47

result was expected since a longer reaction time is required to produce a polymer of
higher molecular weight.

One of the possible ways to increase the fractional conversion of REX-PCL with
high molecular weight is to increase the residence time by decreasing the screw speed.
Increasing the extruder temperature also increases the rate of reaction and can result in

higher fractional conversion.

3.5.2 Determination of Intrinsic Viscosity
3.5.2.1 Objective

Objective of this experiment was to determine the intrinsic viscosity of linear and
REX-PCL of various molecular weight with Ubbelohde viscometer and compare the

results.

3.5.2.2 Materials

Three different molecular weight of linear polycaprolactone (Mn = 10,000,
42,500, and 80,000) were purchased to compare with the REX-PCL. The theoretical
molecular weight of REX-PCL used in this experiment were 30,000, 60,000, 100,000,

110,000, and 120,000. Toluene was used as a solvent.

3.5.2.3 Procedure

Intrinsic viscosity of linear and REX-PCL were determined by following the

standard procedure discussed in Chapter 2.

48

3.5.2.4 Results

Examples of a plot used to determine the intrinsic viscosity of REX-PCL and

linear PCL are shown in Figure 3-8.

 

 

 

 

 

 

 

 

 

 

 

 

 

(3)
Determination of Intrinsic Viscosity
(Linear PCL with MW = 80,000)
140
8 . 100 ~ R =0.9983
-- r:
m e I I I . ------ I
.2 ,6 80 'i I f
w 8- 60 «
3.3 m 40 . y = 627.98x + 78.857
% 2 _
0 I I I I
0 0.002 0.004 0.006 0.008 0.01
Concentration (g/ml)
0))
Determination of Intrinsic Viscosity
(REX-PCL with MW =110,000)
300
250 4 y=9639.5x+ 143.45
E R2=0.9419 \
:3 r: 200
"3 2
'U "' _
£3150 .n-n-.--.------.
,3 m 100 ~ /'
-‘ 50 1 y=-2198.7x+ 144.85
R2=0.84l
o I I I l I
0 0.002 0.004 0.006 0.008 0.01 0.012
Concentration (g/ml)

 

 

 

Figure 3-8: An example of plot used to determine the intrinsic viscosity of (a) linear and
(b) REX-PCL.

49

The intrinsic viscosity of the linear and the REX-PCL tested are tabulated in Table 3-5

and 3-6 respectively.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Mn [nlfluggins [anraenier [nlaverage Standard
deviation
80,000 73 79 76 3.0
42,500 60 61 60.5 0.5
10,000 28 27 27.5 0.5
Table 3-5: Intrinsic viscosity of linear PCL
M“ [Tllfluggins [anraemer [nlaverage standar C1
deviation
120,000 144 148 146 2.0
110,000 143 145 144 1.0
100,000 116 122 119 3.0
60,000 82 77 79.5 2.5
30,000 66 62 64 2.0

 

 

 

 

 

 

 

Table 3-6: Intrinsic viscosity of REX-PCL

3.5.2.5 Discussion
Figure 3-8 shows that plotting the left-hand-side of Huggins equation
(c'l (n/no - 1)) and the left-hand-side of Kraemer equation (c'l ln(n/n°)) against
concentration yields linear relationship for both the linear and the REX-PCL, as expeCted.
As also expected, the PCL with higher molecular weight has higher intrinsic
viscosity, as shown in Table 3-5 and 3-6. By comparing the intrinsic viscosity shown in
Table 3-5 and 3-6, it was found that the intrinsic viscosity of the REX-PCL is higher than

the one of linear PCL. This result was also expected since the molecular weight of REX-

50

PCL used were actually for “each arm”, and the fact that the REX-PCL has a star-shape
makes it to have more bulky structure with longer chain than linear PCL.

As shown in Figure 3-8, the intrinsic viscosity determined from Huggins equation
and Kraemer equation did not exactly agree. This was due to the fact that the equations

used were truncated versions of the power series expression of the form]:

e'm/n, — 1) = [n] + kgmizc + k’H[n]3c2 +
and

chum/n.) = [n] — kilnlzc - 16.111138 -

In this study, the average of the intrinsic viscosity calculated from the two equations

(Huggins and Kraemer equations) were calculated for a comparison purpose.

3.5.3 Determination of Mark-Houwink-Sakurada Constants

3.5.3.1 Objective

Objective of this experiment was to determine the Mark-Houwink-Sakurada
(MHS) constants for both linear and REX-PCL, using toluene as a solvent with the

testing temperature of 26 °C.

3.5.3.2 Procedure
Intrinsic viscosity determined in section 3.5.2 were used to plot mm] vs lnMn for

both the linear and the REX-PCL separately. A linear regression was the used to draw a

51

straight line passing through the data points. The equations of the lines were determined

by using a computer program

3.5.3.3 Results
Plot of ln[n] vs lnMn for linear and REX-PCL are shown in Figure 3-9 and 3-10

respectively. Finally, the MHS constants were determined fi'om the plots and the results

are tabulated in Table 3-7.

 

 

 

 

 

 

5

4 .. A”

k1
F. 3 —
5
E 2 _ y=0.5012x-1.2914
R2=0.9921
1 r
o
8 8.5 9 9.5 10 10.5 11 11.5
lnMu

 

 

 

Figure 3-9: Determination of Mark-Houwink-Sakurada constants for linear PCL

52

 

 

 

 

 

 

 

6
5 '— ‘w
4 . ...i '
'2‘-
: 3 “
2 _ y = 0.622x - 2.33
1 R2 = 0.9342
0 L
10 10.5 11 11.5 12
lnM.

 

 

 

Figure 3-10: Determination of Mark-Houwink-Sakurada constants for REX-PCL

 

a K
Linear PCL 0.5012 0.2749
REX-PCL 0.6220 0.0973

 

 

 

 

 

 

 

Table 3-7: Mark-Houwink-Sakurada constants

3.5.3.4 Discussion

Both Figure 3-9 and 3-10 show that the measured intrinsic viscosity fits to the
linear relationship of MHS equation fairly well. By using the linear regression, one of
the MHS constants, a, for the linear and the REX-PCL were found to be 0.5012 and
0.6220 respectively. These values are reasonable since the range of “a” is between 0.5 to
1, in which 0.5 represents the randomly coiled polymer in a theta solvent and 1 represents

the polymer has rod-like extended chain”.

53

Another MHS constant, K, for the linear and the REX-PCL were found to be
0.2749 and 0.0973 respectively. These values are also within the range of various

reported values of K: the K values generally vary between 10'3 to 0.5 28.

3.5.4 Thermogravimetric Analysis
3.5.4.1 Objective

Objectives of this experiment was: (1) to determine the thermal stability of REX-
PCL and to compare it with the one of linear PCL; and (2) compare the thermal stability

of the same type of PCL with different molecular weight.

3.5.4.2 Materials
Linear PCL with the molecular weight of 80,000 and the REX-PCL with the

molecular weight of 30,000, 60,000, and 100,000 were tested.

3.5.4.3 Procedure
Standard procedure discussed in Chapter 2 was followed to test both linear and
REX-PCL. Also a degradation temperature were determined by using the computer

program and recorded.

3.5.4.4 Results
Output of TGA of the REX-PCL with the molecular weight of 30,000, 60,000,
and 100,000 are plotted together and shown in Figure 3-11. The output of TGA of linear

PCL with the molecular weight of 80,000 was also plotted and shown in Figure 3-11.

54

 

 

 

 

 

 

 

 

 

 

 

.20
— 60.000
—. 30.000
_._. 100.000
mm
80-
E
is °°‘
3
‘0—
mi
0 I I I ' I
1 100 200 300 400 500

 

Temperature ('C)

Figure 3-11: The output of TGA for linear and REX-PCL

Also the degradation temperature determined using the computer program were recorded

and tabulated in Table 3-8.

 

Table 3-8: Degradation temperature of linear and REX-PCL

55

3.5.4.5 Discussion

As shown in Table 3-8, a degradation temperature of linear PCL was found to be
372 °C which is much higher than the degradation temperature of REX-PCL which is at
about 330 °C for all three molecular weights. This means the REX-PCL is thermally less
stable than the linear PCL.

By comparing the results of REX-PCL of different molecular weight, it was found
that the REX-PCL of higher molecular weight is more thermally stable than the REX-
PCL of lower molecular weight. Additionally, it was found that the weight loss of linear
PCL occurs in small range of temperature with one steep slope; however, the weight loss
of REX-PCL was found to occur in wider range of temperature, and the slope became
rather bumpy as the molecular weight of REX-PCL increases. These results suggests that
the REX-PCL has broader molecular weight distribution than the linear PCL, which also

confirms the result obtained fiom the determination of Nfl-IS constants.

3.5.5 Differential Scanning Calorimeter
3.5.5.1 Objective
Objective of this experiment was to determine the thermal transition

characteristics of REX-PCL and compare the results between the REX-PCL of different

molecular weights.

3.5.5.2 Materials

REX-PCL with the molecular weight of 30,000, 60,000, and 100,000 were tested.

56

3.5.5.3 Procedure

Standard procedure discussed in Chapter 2, with the type 1 heating condition, was

used to test both linear and REX-PCL. Also observed peaks were labeled by using the

computer program.

3.5.5.4 Results

DSC output for REX-PCL with the molecular weight of 30,000, 60,000 and

100,000.are shown in Figure 3-12 to 3-14 respectively.

 

 

 

 

 

 

 

 

0.5
0.0 d
53.47'C
f 85ng
l
. ' i
29-53‘0 74.69‘6
Ea -o.5 ~
g r
u.
7:7
I -1.0 -*
-1.5 ~ U
”.490
4459111119
'2-0 ' If r ' 1 ' - T L fi r
-80 £0 0 20 40 60 80
E” ”P Temperature ('0)

Figure 3-12: The output of DSC for REX-PCL with the molecular weight of 30,000

57

 

 

0.5

 

 

Heat Flow (W/g)

 

 

 

 

 

 

 

 

 

 

0.0 - \
52.83'0
70.88JIg
25.40°c 79.17°c
-0.5 —
-1 .0 -
60.98'0
4.053ng
-1.5 v r v r ' 1 Y 1 ' I V I ' r ' r
-80 -60 ~40 -20 O 20 4O 60 80
5.0 Up Temperature (°C)
Figure 3-13: The output of DSC for REX-PCL with the molecular weight of 60,000
0.5
0.0 ~\
more
i . nearly
31.87'0 72f43°c

1;. cs .
.3,
11'.
ii
I -1.0 '1

-1.5 ~

i 00.51‘0
4.me
'10 * t ' r ' T ' I ' I . 1 fi r Y
en -,4o -20 o 20 40 so 00
EM” Temperature (‘C)

Figure 3-14: The output of DSC for REX-PCL with the molecular weight of 100,000

58

3.5.5.5 Discussion

As shown in Figure 3-12 to 3-14, the melting temperature of REX-PCL was found
to be approximately 53 °C, which is close to the reported melting temperature of linear
PCL.

The glass-transition temperature was not observed, although the test was
conducted from —60 °C. This was also not surprising since the reported glass-transition
temperature of linear PCL is approximately at —60 °C.

These results conclude that linear and REX-PCL go through very similar thermal

transitions.

3.6 Reactive Extrusion Polymerizatoin of e-Polycaprolactone from Monomer
with Impurities

3.6.1 Objective
Objective of this experiment was to determine the effect of moisture content in

the monomer on the production of REX-PCL.

3.6.2 Materials

In order to determine the effect of moisture content, TONE ECEQ was left open
' to air for approximately 5 minutes.

The same ATSB solution as discussed in section 3.4.2 was used, and the entire
process was carried under dry nitrogen to prevent monomer and initiator from further

exposure to the moisture.

59

3.6.3 Procedure
Exactly the same procedure as discussed in section 3.4.3 was followed, except the

monomer with moisture content was used instead of the “regular” monomer.

3.6.4 Results

Molecular weight of REX-PCL produced, as well as the feed rate of monomer and
ATSB solution, fraction of ATSB in the solution used, and the resulted torque % were
recorded and shown in Table 3-9. Same extruder temperature and screw speed as

discussed in section 3.4.3.4 were used.

 

 

 

 

 

 

 

 

 

Theoretical Monomer ATSB Solution % ATSB in Observed
Molecular Weight Flowrate Flowrate Solution Torque %
(ml/min) (ml/min)
30,000 80 1.7 15 23
60,000 100 1.6 10 32
80,000 70 1.7 5 38

 

 

Table 3-9: Molecular weight of REX-PCL produced with the monomer with moisture
content

3.6.5 Discussion

With monomer containing moisture, the highest molecular weight of REX-PCL
tlmt could be produced was found to be 80,000. Above this value, the product coming
out from the extruder became jelly-like with clear yellow color, and changing the process
conditions did not change the result. This result suggests that the moisture content of
monomer (or anywhere in the system) result in the poor production of REX-PCL possibly
due to the reaction of initiator and the moisture which decreases the number of initiator

available for the monomer.

60

3.7 Monomer Conversion and Reproducibility

Fractional conversion of monomer was determined in section 3.5.1 and they were
found to be close to 100 % for the most of the molecular weight. However, this does not
tell where in the extruder that the reaction reached near completion. Determination of
where the reaction reaches completion is important when designing a process in which
multiple reactions take place in series, such as in the case of the production of
polyurethane, which will be discussed in next chapter.

Therefore, monomer conversion within the extruder was evaluated and it will be
discussed in the first part of this section.

Additionally, successfiil production system requires a production of desired
material not once, but over and over again. In other words, a production system must be
able to reproduce the product as long as the required conditions are met.

Therefore the reproducibility of REX-PCL was evaluated and it will be discussed

in the second part of this section.

3.7.1 Monomer Conversion
3.7.1.1 Objective
Objective of this experiment was to determine the monomer conversion within the

extruder during the process of making REX-PCL with the molecular weight of 120,000.

3.7.1.2 Materials

TONE ECEQ monomer, ATSB solution, and nitrogen as discussed in section

3.4.2 were used to make REX-PCL with molecular weight of 120,000.

61

3.7.1.3 Procedure

Same procedure as making the REX-PCL was followed, except when the system
stabilizes, the extruder was stopped and the material at the middle of extruder (zone 3)
and at the die were immediately removed from the extruder for evaluation.

The REX-PCL collected from two locations were cooled in air and their intrinsic

viscosity were determined by following the standard procedure discussed in Chapter 2.

3.7.1.4 Results

Intrinsic viscosity of two samples collected at different locations, one at die and
another one at the middle of the extruder, were measured and tabulated in Table 3-10.
Also, by applying these intrinsic viscosity to the MHS equation determined in section

3.5.3, molecular weight of these two materials were estimated as shown in Table 3-10.

 

 

 

 

Location Collected Intrinsic Viscosity Molecular Weight
Middle of Extruder 131.03 107,389
At Die 140.28 119,836

 

 

 

 

Table 3-10: Intrinsic viscosity of REX-PCL collected at two different locations within the
extruder
3.7.1.5 Discussion
As shown in Table 3-10, the intrinsic viscosity of the sample collected at the die
was found to be higher than the intrinsic viscosity of the sample collected at the middle of
the extruder. Also by applying these result in MHS equation, the molecular weight of the

sample collected at the die was found to be approximately equal to the theoretical

62

molecular weight; however, the molecular weight of the sample collected at the middle of
the extruder was found to be about 10 % lower than the expected value.
These results indicate that the REX-PCL with the theoretical molecular weight of

120,000 yields 90 % monomer conversion at the zone 3.

3.7.2 Reproducibility
3.7.2.1 Objective

Objective of this experiment was to determine the reproducibility of the process
of making REX-PCL. This was done by comparing the intrinsic viscosity of REX-PCL
of same molecular weight produced with same procedure but in different days.
Molecular weight of 100,000 was evaluated in emeriment.

Additionally, REX-PCL with the molecular weight of 30,000 were produced with
different overall flowrate (same monomer to ATSB ratio) on different days and their

intrinsic viscosity were also determined to evaluate the reproducibility of the process.

3.7.2.2 Materials

Just as before, TONE ECEQ monomer, ATSB solution, and nitrogen were used as
specified in section 3.4.3.
3.7.2.3 Procedure

REX-PCL with the molecular weight of 100,000 were produced on different days
with different ATSB solutions (same ATSB concentration in the solution but prepared on
different days). Process condition and the procedure used were the same for both

samples.

63

Also REX-PCL with the molecular weight of 30,000 were produced on two
separate days. The difference in the process between those two days was the flowrate of
monomer and initiator; a flowrate of the monomer and the ATSB solution were doubled,
but the ratio of monomer to ATSB were kept constant for the experiment on second day.

The intrinsic viscosity of all the samples collected were then determined by

following the standard procedure discussed in Chapter 2 and their results were compared.

3.7.2.4 Results

A left-hand-side of Huggins equation vs. concentration were plotted for the REX-

PCL made on two different days (Mn = 100,000) and it is shown in Figure 3-15.

 

 

 

 

 

Repeatability of Experiment
200
X
150 4 M
.15
3 100 e
’2‘
3 50 _ dayl ......... y = 9884.3x + 119.75
day 2 -— y = 10370x + 115.5
0 T T T I I l
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007
Concentration

 

 

 

Figure 3-15: A plot used to determine the intrinsic viscosity of REX-PCL made on three
different days.

Also, left-hand-side of Huggins equation vs. concentration were plotted for the

REX-PCL made with two different flowrates (Mn = 30,000) and it is shown in Figure 3-

16.

 

 

 

 

Repeatability of Experiment
100
M- ' x

80 ‘ - ‘ . I .
5‘: x X
{3 60 4
5 4o «
3 day 1 ......... y = 900.28x + 65.628

20 I day 2 y = 839.6x + 71.266

0 I I I I I I

 

 

0 0.005 0.01 0.015 0.02 0.025 0.03 0.035

Concentration

 

 

 

Figure 3-16: A plot used to' determine the intrinsic viscosity to REX-PCL made with two
different flowrate

3.7.2.5 Discussion

As Figure 3-15 shows, two lines representing the REX-PCL of same molecular
weight produced on different days almost coincides. Also the Figure 3-16 shows that the
two lines representing the REX-PCL of same molecular weight produced with different

reactants flowrate were found to be very close. These results indicate that the process of

making REX-PCL has very high reproducibility.

65

3.8 Conclusions

The REX-PCL of the molecular weight ranging fi'om 20,000 to 210,000 were
produced successfully only by varying the monomer to initiator ratio. The REX-PCL
with the molecular weight higher than 210,000 was unable to be produced due to the
large monomer to initiator ratio which prevent some of the monomer fiom involving in
the polymerization. On the other hand, the lowest limit was determined as 20,000 not
because of the poor monomer to initiator reaction but because of the extruder which
could not handle such a low viscosity. In either case, the addition of a few mixing
elements may help to achieve even broader range of molecular weight.

By determining the intrinsic viscosity of linear and REX-PCL, the MHS constants
were determined to be 0.5012 and 0.2749 for the linear PCL and 0.6220 and 0.0973 for
the REX-PCL. These values indicates that the linear PCl is more randomly coiled
polymer than REX-PCL.

The difference between linear and REX-PCL was also evident from the results of
TGA. The output of TGA showed that REX-PCL has lower thermal stability than linear
PCL. Also the rate of degradation REX-PCL was found to be slower than the one of
linear PCL, and the rate decreased as the molecular weight of REX-PCL increased.

By producing the REX-PCL with the monomer containing some moisture, it was
found that the range of the REX-PCL producible was narrowed down to between 30,000
and 80,000. One of the possible explanations to this is the reaction between ATSB and
the moisture within the monomer, which reduced the amount of ATSB available for the

polymerization.

66

Chapter 4

 

REACTIVE EXTRUSION POLYMERIZATION OF STAR-

POLYCAPROLACTONE AND DIISOCYANATE

 

4.1 Introduction

Applications of polyurethane widely range fi'om foam materials, fibers and foils
to the high elastic coating materials, solvent adhesives and the construction materials of
various hardness. This is due to the wide range of properties of polyurethane that can be
made by selecting the types and the ratio of reactants”. However, a production of
polyurethane of desired properties also requires a greater degree of control over the
chemical reaction than many other polymer productions”. This is due to the high
reactivity of diisocyanate commonly used to produce the polyurethane.

One of the widely used ways to synthesize polyurethane is by the addition of
dihydroxy compounds, such as low-molecular-weight hydroxyl-terminated polyesters, to

diisocyanates28 as shown below:

0 0
II II
OCN—R—NCO + HO—R’— OH —-) —(— CNH—R—NHCO—R’O )—

Figure 4-1: Polyurethane synthesis by the reaction of diisocyanate and dihyroxy-group

One of the earliest commercial polyurethane was developed in Germany by reacting 1,6-

hexanediisocyanate and 1,4-butanediol. Although this particular process is no longer

67

used in commercial, use of diisocyanate in the formation of polyurethane is still widely
used to manufacture polyurethane fibers, plastics, elastmers, and coatings”.

The production of polyurethane is currently done in various ways. One of the
widely used technique is the batch polymerization, such as in the case of making
polyurethane foams. The reactive extrusion has also been developed for the production
of linear polyurethane ”’39.

In this study, a production of polyurethane was also conducted by utilizing
extruder. The polyurethane was produced by reacting hydroxyl end-group of REX-PCL

with hexamethylene diisocyanate. A reaction mechanism of the formation of

polyurethane from hexamethylene diisocyanate and REX-PCL is shown in Figure 4-2.

OCN (CH2), NCO + 0

ll
0 O - (CH2)5 - C )u—OH

I
-(0-(CH2)s-C)a- Al—l 0

ll
—( O - (CH2)s - C )a—OH

——>
0

ll
0 O 0 O - (CH2)5 " C )r-OH

0

II II
—-(-CNH— (CH2)6 -NH("3 -( O - (CH2)5 - C )5— Al —I

ll
—( O - (CH2)5 - C )a—OH

Figure 4-2: A reaction mechanism of the formation of polyurethane with hexamethylene
diisocyanate and REX-PCL

68

4.2 Background
There are many factors influencing the formation of polyurethane, such as the
presence of catalysts and water, and the process temperature. This sections addresses

how these factors influence the formation polyurethane.

4.2.1 Effect of Catalyst

Synthesis of polyurethane by the addition of hydroxy end-group and isocyanate
can take place without the presence of a catalyst; however, tertiary amines and metal
compounds are widely used as a catalyst to increase the selectivity of isocyanates
reacting with hydroxyl groups rather than waterm’zs. Some of the examples of a catalyst
that can be used include stannous: octanone, N-methylmorpholine, triethylamine and
dibutyltin diacetate20'39.

In this study, polyurethane formation was carried without the use of a catalyst
mainly due to the fact that the reaction of polyurethane formation is carried immediately

after (or at the same time in some cases) the polymerization of REX-PCL. The presence

of a strong catalyst may ruin the formation of star-shaped PCL.

4.2.2 Effect of Temperature

Just as any other reactions, temperatm'e has a significant influence on
polyurethane synthesis. For example, at a low temperature, say up to 50 °C, a reaction of
isocyanates with hydroxyl group, as well as water, proceed relatively easily. However, at
a higher temperature, between 50 to 150 °C, forrmtion of allophanates, biurets and

isocyanurates occur at a significant rate in the absence of a catalyst. Therefore, the low

69

temperature (below 50 °C) should be used if a linear polyurethane, without any cross-
links or branch, is desired. On the other hand, if cross-linked or branched polyurethane is
desired, the temperature between 50 to 150 °C is recommended. It is important, however,
to keep the reaction temperature below 150 °C since some of the less stable links start to
break down at the temperature over that temperature, and results in partially degraded
polymer". In addition to the above reactions, the residual carboxylic acid ends in
polyesters rmy react with isocyanates to give amides. However, this reaction is of
relatively minor importance”

In this study, the extruder temperature of 120 °C was mostly used since it is low
enough to prevent the thermal degradation of polyurethane, but high enough to still carry

out the polymerization of REX-PCL and keep the product in molten state.

4.2.3 Effect of Water

Water plays a significant role in the reaction of polyurethane. One of the
examples of water being used as an advantage is the production of flexible foams. In this
case, a reaction involving water results in the formation of carbon dioxide, which is used
for blowing the foam”.

On the other hand, a presence of moisture can easily upset the stoichiometric
balance of isocyanate to alcohol which determines the structure and molecular weight of
the product. For example, 18 parts of water can react with as much isocyanate as 2000
parts of linear polyester. Therefore, water should be eliminated fi'om the system in order

to produce linear polyurethane.

70

In this study, the reactions were carried under nitrogen environment since the
formation of polyurethane is carried inside the extruder where the polymerization of
REX-PCL also takes place. However, moisture is available in atmosphere as soon as the
product exits from the die, thus excess isocyanate or the isocyanate end-group may react

with the moisture to form crosslinkage.

4.3 Outline of Chapter

This chapter discusses the reactive extrusion polymerization process to make
polyurethane with REX-PCL. Two types of REX-PCL were used for this purpose: REX-
PCL of high molecular weight (30,000, 60,000, and 100,000) and the REX-PCL of low
molecular weight (3000 and 7500). Each process is discussed in section 4.4 and 4.6
respectively.

In addition, the polyurethane produced were characterized and discussed in
section 4.5 for polyurethane made with REX-PCL of high molecular weight and in

section 4.7 for polyurethane with REX-PCL of low molecular weight.

4.4 Reactive Extrusion Polymerization of Diisocyanate and REX-PCL with High
Molecular Weight Star-Polycaprolactone
4.4.1 Objective
Objective of this experiment was to produce polyurethane by reacting hydroxyl
end-group of REX-PCL and diisocyanate. The two reactions in series, starting with
polymerization of REX-PCL followed by the reaction of REX-PCL and diisocyanate,

were conducted continuously within the same extruder.

71

4.4.2 Materials

The same TONE ECEQ monomer and ATSB solution as discussed in Chapter 3
were used. Additionally, hexamethylene diisocyanate (I-IMDI) was used to react with the
hydroxyl end-group of REX-PCL to form polyurethane. This particular diisocyanate was
chosen because of its wide availability and less toxicity compared to some other types of
diisocyanate.

As in the case of ATSB solution, HMDI was mixed with anhydrous toluene to
increase the flowrate in order for the commercial pump to be used. The calculation to
determine the flowrate of HMDI will be discussed in section 4.4.3.1. Since HMDI is also
sensitive to moisture, anhydrous toluene and not “regular” toluene was used as a solvent.

The molecular weight of REX-PCL used in this experiment were 30,000, 60,000,
and 100,000. From the results of the determination of the monomer conversion in section
3.7.1, polymerization of the REX-PCL of all three molecular weights were expected to be

at near completion at zone 3, where HMDI solution is added.

4.4.3 Procedure
Polyurethane was produced with the similar method as the production of REX-
PCL. The only difference between the production of REX-PCL and polyurethane was

the additional feed (HMDI) at the middle of the extruder.

4.4.3.1 Flowrate Calculation

Same procedure was followed as discussed in the case of production of REX-PCL

to determine the flowrate of the monomer and the ATSB solution. Then, the flowrate of

72

HMDI feed was calculated so the number of OH end-group is equal to the number of

NCO group. The equation used to determine this amount is as follows:

Number of OH end-group = 3 x Number of ATSB fed
and
Number of NCO group = 2 x Number of Hit/DI fed
To have equirnolar of OH and NCO groups,
Number of OH end-group = Number of NCO group
then
Number of HMDI fed = 3/2 x Number of ATSB fed

An example of calculation sheet used in this experiment is shown in Table 4-1.

73

 

 

Reactive Extrnaion Polymerization of Polyurethane " pick value

Date:
Run#:
Theoretical MW = h .
DP=Theoretical MW *3/114.14=H 1
Monomer
Volumetric F lowrate of Monomer =. 3 ml/min
pumaner = 1.030 g/ml
Mass Flowrate of Monomer = Vol F lowrate of Monomer "‘ pm = g/min
MW of Monomer = 114.14 g/mol
Molar F lowrate of Monomer = Mass Flowrate / MW of Monomer = moi/min
Initiator
Molar F lowrate of ATSB = Molar Flowrate of Monomer / DP = monin
MW of ATSB = 246.33 g/mol
Mass Flowrate of ATSB = Molar Flowrate * MW = g/min
Fraction of ATSB m ATSB solution — :00. 210’;
Mass Flowrate of Solution= Mass Flowrate of ATSB / Fraction of ATSB - l I , .. g/min
p of ATSB = 0.967 g/ml
p of Toluene = 0.865 g/ml
p of Solution = 1/[(fiac of ATSB/puss) + (fiac of Toluene/admfl = g/ml
Volumetric Flowrate of Solution = Mass F lowrate of Sol. / p of Sol. = ml/min
HMDI
Molar Flowrate of HMDI = Molar F lowrate of ATSB * 3/2 = monin
MW of HMDI = 168.20 g/mol
Mass F lowrate of HMDI = Molar Flowrate * MW g/min
Fraction of HMDI in HMDI solution =:w I '13
Mass Flowrate of Solution = Mass Flowrate of HMDI / Fraction of HMDI = g/min

p of HMDI = 1.04 g/ml

p of Toluene = 0.865 g/ml

p of Solution = 1/[(Frac of HMDI/pmm) + (F rac of Toluene/puma,» = g/ml
Volumetric F lowrate of Solution = Mass F lowrate of Sol. / p of Sol. = Ml/min

 

Table 4-1: An example of the calculation sheet used to determine the flowrate of
reactants

74

 

4.4.3.2 Preparation of ATSB and HMDI Solutions

The HMDI solution refers to the mixture of hexamethylene diisocyanate and
anhydrous toluene. Just as the ATSB solution, the weight percentage of HMDI in the
HMDI solution was determined merely for the convenience. Typically, 6 and 20 wt% of
HMDI in the solution was used.

Preparation of both ATSB and HIVHDI solution were done in the same glove bag
under nitrogen environment, and the same procedure as in the case of preparing ATSB

solution (discussed in section 3.4.3) was followed.

4.4.3.3 Flow Scheme

For the production of polyurethane, three streams were added to the system of
producing REX-PCL. They were: (1) HMDI solution to extruder (2) nitrogen to
container of HMDI solution; and (3) to the extruder where HMDI solution is fed.
Streams of HMDI solution and nitrogen were introduced to the extruder from zone 3.

Flow scheme of HIVHDI solution and two additional nitrogen streams are shown in

Figure 4-3:

75

 

     

 

   
   
  
 

 

to extruder outlet trap
zone 3 (mineral oil)
to HMDI 7 3
solution 11.x
N
feed throat for HMDI solution
HMDI solution Monomer
and

N2 Initiator

 

.1

  

 

 

 

Figure 4-3: Flow scheme of HMDI solution

Just as the nitrogen used in the production of REX-PCL, the nitrogen streams
added in this system was divided into two streams, then entered to the flowmeter to
monitor the nitrogen flow. The nitrogen was then fed into the extruder (at zone 3) and

the container of HMDI solution. The nitrogen flowing out from the container of HMDI

76

solution was then fed into outlet trap containing mineral oil to prevent back flow of the
air.

HIVHDI solution was made in 2 L flask, and it was pumped out and directly fed
into the extruder at zone 3 without using a transfer vessel. Flowrate of HIVflDI solution
was monitored by measuring the weight of 2 L flask. Just as the monomer and the ATSB
containers, container of the HMDI was tightly closed and dry nitrogen was continuously
passed through to prevent the solution from being exposed to moisture.

Mas’terflex® L/S® Easy Load® pumps and the Masterflex® L/S® precision tubing
size 14 were used for this system. The size of the pump head and the tubing size were
determined by considering the typical flowrate of the HMDI solution used in this study

(approximately the same range as the flowrate of the ATSB solution).

4.4.3.4 Start Up
Extruder was started first and the temperature of extruder zones were set to the

desired temperature as shown in Table 4-2.

 

Zone 1

Zone 2

Zone 3

Zone4

Zone 5

Zone6

 

 

T (°C)

 

180

 

150

 

120

 

120

 

120

 

120

 

 

Table 4-2: An extruder temperature profile used in this experiment

Nitrogen flow was started prior to opening any of the reactant containers. Then

all tubes were carefully connected to avoid the solutions from being exposed to moisture

while nitrogen flowing through them.

77

Once the extruder reached the desired temperatures, pumps of the monomer and
ATSB solution, but not HMDI solution, were started and adjusted to produce REX-PCL
of desired molecular weight. As before, a product coming out from the die was collected
in a metal trash can until the system stabilized.

When the system of producing REX-PCL was stabilized, the pump for the HMDI
solution was started and adjusted to desired flowrate. The screw speed of 110 rpm was

used throughout the process.

4.4.3.5 Collecting Samples and Clean Up

After letting the system run steadily for approximately 20 minutes, the product
was collected through the water bath and dried in air for a few days. Approximately 1 lb
each of polyurethane, as well as the REX-PCL before the addition of HMDI, were

collected for characterization.

4.4.4 Results

Molecular weight of REX-PCL used as well as the feed rate of the monomer,
ATSB solution and HIVfl)I solution, fiaction of ATSB in the solution used, fraction of
HMDI in solution used, and the observed torque % for REX-PCL andpolyurethane were
recorded and is shown in Table 4-3. The same extruder temperature and screw speed as

discussed in Chapter 3 were used.

78

 

 

 

 

 

 

 

 

 

 

 

 

Theoretical Molecular Weight 30,000 60,000 100,000
Monomer Flowrate (ml/min) 4O 50 8O
ATSB Solution Flowrate (ml/min) 2.6 1.6 1.6
% ATSB in Solution 5 5 5
HMDI Solution Flowrate (ml/mini 2.2 1.4 1.3
% HMDI in Solution 6 6 6
Observed % Torque

(for REX-PCL) 21 35 41
Observed % Torque

Q‘or polyurethane) 26 36 38

 

Table 4-3: Condition used to produce polyurethane

4.4.5 Discussion

Addition of HMDI solution to the process of making REX-PCL was done without

any major problems. The only difference observed before and after the addition of

HMDI is the melt fiacture, which seemed to be thicker when the HMDI was added.

However, a significant increase in torque % was observed only with the molecular weight

of 30,000.

The next section discusses the characterization of polyurethane produced in this

experiment.

4.5 Characterization of Polyurethane made with High Molecular Weight Star-

Polycaprolactone

4.5.1 Objective

Objective of this experiment was to determine whether the reaction between

REX-PCL and HMDI has taken place. This was done by determining the intrinsic

viscosity of polyurethane produced, and comparing with the one of REX-PCL. The

79

 

polyurethane was also tested with TGA and DSC, and the results were compared with the

results of REX-PCL.

4.5.2 Materials
Polyurethane produced fi'om three different molecular weight of REX-PCL (Mn =

30,000, 60,000, and 100,000) were used.

4.5.3 Procedure
Analysis of polyurethane with TGA and DSC, as well as the determination of

intrinsic viscosity were done by following the standard procedures discussed in Chapter

2.

4.5.4 Results
Intrinsic viscosity of polyurethane produced were determined and tabulated in

Table 4-4.

 

 

 

 

 

 

Molecular Weight Intrinsic Viscosity standard

of REX-PCL of Polyurethane deviation
300,000 47 1.32
600,000 73 1.58
100,000 105 0.68

 

 

 

Table 4-4: Intrinsic viscosity of polyurethane

80

Then, the intrinsic viscosity of polyurethane, as well as the intrinsic viscosity of
the linear and REX-PCL previously determined, were plotted for ln[n] vs. lnMn and

shown in Figure 4-4.

 

 

 

 

 

 

 

 

 

6

5 4 ‘

4 ' ‘7‘“ a

. .‘,.-"" -'-Linear(linearPCL)
5 1 — —Linear (polyurethane)
5 3 WWW“, = 0.6704x- 3.0684 —Linear (“mm

2 1 ylinearPCL=O-5012x' 1.2914

yREX-PCL = 0.6220X - 2.33
1 3
0 T I I I T
9 9.5 10 10.5 11 11.5 12
lnMn

 

 

 

Figure 4-4: Comparison of the intrinsic viscosity of the three types of polymers

The molecular weight of the original REX-PCL were used to plot polyurethane, since the
whole purpose of this plot was to do the comparison before and after the reaction. From
the Figure 4-4, MHS constants of polyurethane was determined and tabulated with the

MHS constants of REX-PCL and linear PCL previously determined in Table 4-5.

81

 

 

 

 

 

 

 

 

a K
Linear PCL 0.5012 0.2749
REX-PCL 0.6220 0.0973
Polyurethane 0.6704 0.0465

 

Table 4-5: MHS constants of three polymers

The results of TGA of polyurethane and the original REX-PCL the polyurethane
was made with were plotted together in Figure 4-5 to 4-7 for the molecular weight of
30,000, 60,000, and 100,000 respectively. The results of TGA of three polyurethanes
produced were also plotted on the same plot in Figure 4-8 for comparison.

The results of DSC of polyurethane and the original REX-PCL the polyurethane
was made with were plotted together in Figure 4-9 to 4-11 for the molecular weight of

30,000, 60,000, and 100,000 respectively.

 

120

 

 

 

 

 

 

 

 

 

 

 

Polyurethane
J -—-- REX-PCL
100-
804
g
E 60‘
40"
l
l
I
l
I
20~ I
I
l
I
I
L ___
0 V I ' r ' r ' ’ r '
O 100 200 300 400 500

Figure 4-5: Output of TGA for the polyurethane and REX-PCL (Mn = 30,000)

82

120

 

 

 

 

 

 

 

 

 

 

 

 

 

__ Polyurethane
............. REX-PCL
100«
60 ~ I
I
I
A ‘ l
as I
E m4 ‘l
g I
I I
I
I
.0- I
I
I
I
I
I
20- I
I
I
‘ I
I
|~R¥
0 ' T V r . #— v 1 - - ‘ - v
o 100 zoo 300 460 500
Temperature (°C)
Figure 4-6: Output of TGA for the polyurethane and REX-PCL (M. = 60,000)
120
. Polyurethane
-— REX-PCL
10m
I
”-4
g 4
E .0.
i .
‘o-I
1 I
I
204 |\
\
\
0 v a ;:\\‘&-_‘— -
o 150 in - 360 - «So 7 500
Temperature ('C)

 

Figure 4-7: Output of TGA for the polyurethane and REX-PCL (M. = 100,000)

83

 

 

Weight (%I

 

 

 

 

 

260
Tm" ('C)

o 160

Figure 4-8: Output of TGA for three polyurethanes

 

 

0.5

0.0 -‘

4.5“

 

 

 

 

 

-2.o . 1
ac

EDUP

.zo ' 6 2'0
TWCC)

Figure 4-9: Output ofDSC for theipolyurethane and REX-PCL (M, = 30,000)

 

 

 

 

 

 

 

 

 

0.5
Polyurethane
J —— REX-PCL
0.0..
g -o.5
E
7:7
I -1.0~
1.5a
~20 V r 1 r ' 1 ' r ' I ' I J r . r .
-80 ~60 ~40 —20 O 20 4O 60 80 100

Figure 4-10: Output of DSC for the polyurethane and REX-PCL (Mn = 60,000)

 

0.5
__.._ Polyurethane
“mm” REX-PCL

0.0 ~

 

Heat Flow (Wig)
.6
(II

t

r‘

O
l

4.5"

 

 

 

 

.2.o ' I ' I ' I v I ' I ' I T '
~80 -60 -40 -2o 0 20 40 so 8'0 100
E” U" Temperature ('C)

Figure 4-11: Output of DSC for the polyurethane and REX-PCL (Mu = 100,000)

85

4.5.5 Discussion

By comparing the intrinsic viscosity of REX-PCL and polyurethane that was
produced from the same REX-PCL, it can be seen that the intrinsic viscosity of
polyurethane is consistently lower than the intrinsic viscosity of REX-PCL. Additionally
the Figure 4-4 shows that the intrinsic viscosity of polyurethane are very close the one of
linear PCL.

One of the possible explanations to this observation is the broken linkage between
aluminum at the center and the polycaprolactone arms. If this is the case, the polymer is
no longer in the “star” shape, and each of the resulted linear PCL have two OH end-
groups. This means that the ratio of OH end-group to NCO group is 2 to 1 instead of 1 to
1, and the product would have OH end-group with the molecular weight of approximately
two times that of each “arm”. There will be no crosslinkage in this product.

The results of TGA and DSC show very small difference between the
polyurethane and REX-PCL. This was no surprise since the diisocyanate contents in
polyurethane were very small. Therefore the reactive extrusion polymerization of
polyurethane was done with low molecular weight of REX-PCL in order to observe the

reaction between diisocyanate and REX-PCL better.

4.6 Reactive Extrusion Polymerizatoin of Diisocyanate and REX-PCL with Low
Molecular Weight Star-Polycaprolactone

4.6.1 Objective
In this experiment, polyurethane was produced fi'om REX-PCL of a low

molecular weight, which is more commonly used for the production of polyurethane28

86

than higher molecular weight used in the previous experiment. The two different
amounts of diisocyanate were added to produce different types of polyurethane: excess
OH end-group to prevent crosslinkage, and excess NCO groups to have greater amount of
crosslinkage.

By using the REX-PCL of low molecular weight, NCO content was increased.
This was advantageous in terms of the characterization of polyurethane since the

difference between the polymers before and after the reaction will be more significant.

4.6.2 Materials
The TONE ECEQ monomer was used, and ATSB solution and HMDI solution

were prepared by following the procedure discussed in section 4.4.3.2

4.6.3 Procedure

The same procedure as discussed in section 4.4 was followed to produce
polyurethane, except REX-PCL with the molecular weight of 3000 and 7500 were used
instead of 30,000+. Screw Configuration #1 was used opposed to the Screw
Configuration #2 used in the past experiments. This was because the molecular weight
used in this experiment was so low that the use of Screw Configuration #1 would not
result in the flow stoppage. Furthermore the use of Screw Configuration #1 provided an
additional mixing which would be desirable to produce homogenerous material.

Two types of feeding locations were evaluated. They are: (1) feeding monomer

and initiator at feed throat (before zone 1) and feeding diisocyanate at zone 3; and (2)

87

feeding monomer and initiator at feed throat and feeding diisocyanate at zone 1.

Temperature profile of two cases are shown in Table 4-6.

 

 

 

 

Zone 1 Zone 2 Zone 3 Zone 4 Zone 5 Zone 6
Case 1: T (°C) 180 150 120 120 120 120
Case 2: T (°C) 120 120 120 120 120 120

 

 

 

 

 

 

 

 

Table 4-6: The extruder temperature profiles corresponding to the two types of feeding
schemes

In case 1, the temperature of zone 1 was kept at 180 °C, just as the temperature of
a regular REX-PCL production. Then, from zone 2 to 3, the temperature was gradually
decreased to 120 °C to prepare for the reaction with diisocyanate. With this condition,
block co-polymer of PCL linked with diisocyanate was expected.

On the other hand, the temperature of case 2 was kept at 120 °C throughout the
extruder to prevent degradation of polyurethane at the earlier zones in the extruder. With
this condition, the mixture of REX-PCL and diisocyanate links was expected to be more
random than the one produced in case 1.

Since the molecular weight of REX-PCL selected were very low, its
polymerization was assumed to be completed even at the low temperature profile used in
this experiment. For each molecular weight of REX-PCL, the flowrate of HMDI solution
were calculated by following the same equation as discussed in section 4.4.3.1. Amount
of HMDI solution fed was then varied above and below the stoichiometric amount

calculated to produce the different types of polyurethane.

88

Approximately 1 lb of each polyurethane were collected for characterization. The

REX-PCL with the molecular weight of 3000 and 7500 were also collected to be

compared with the polyurethane produced.

4.6.4 Results

The molecular weight of REX-PCL used, the amount of monomer, ATSB

solution and HMDI solution fed, and torque % were recorded and tabulated in Table 4-7.

The last column indicates whether the flowrate of HMDI was above or below the

stoichiometric amount.

 

 

 

 

 

 

Theoretical Monomer ATSB Solution HMDI Solution Torque Above or
Molecular Flowrate F lowrate Flowrate % Below
Weight Stoich.
(each arm) (ml/min) (ml/min) (ml/min)
3000 20 3.2 1.3 21 below
3000 20 3.2 5.4 26 above
7500 50 3 .2 1.3 19 below
7500 50 3.2 5.2 21 above

 

 

 

 

 

 

Table 4-7: Feed conditions used to produce polyurethane (T = 120 °C, 110 rpm)

All the samples collected were produced from the case 2 of the feed schemes.

 

The feed schemes did not work due to the overflow at the zone 3. Detail of this

observation will be discussed in the next section.

4.6.5 Discussion
REX-PCL with the molecular weight of 3000 and 7500 were found to have low

viscosity, as expected, and it was impossible to be pelletized. Also, the production of

89

those materials caused a constant “extruder underload” warning due to their low
viscosities.

Despite the numerous attempts, case 1 of feed schemes did not work due to the
overflow at the zone 3. The overflow was caused by the large viscosity difference
between the materials before and alter the zone 3. Since diisocyanate was added at zone
3, the polymer after zone 3 suddenly becomes very viscous that the polymer of low
viscosity (because of its low molecular weight) coming from earlier zones could not push
the material out from the zone 3. On the other hand, case 2 worked without causing any
major problems. Therefore the polyurethane listed in Table 4-6 were collected by adding
the HMDI solution at zone 1.

As shown in Table 4-7, the torque % indicated the viscosity of the polymer
increased as the HMDI was added. The increase in torque % was especially significant
before and after the addition of HMDI, since the typical torque % of REX-PCL with the

molecular weight of 3000 and 7500 were 9 - 10 %.

4.7 Characterization of Polyurethane made with Low Molecular Weight Star-
Polycaprolactone
4.7.1 Objective
In order to characterize the polyurethane produced from the REX-PCL of low
molecular weight, REX-PCL before the reaction and polyurethane were tested with TGA

and DSC.

90

4.7.2 Materials

Polyurethane produced in section 4.6 were used in this experiment.

4.7.3 Procedure
Analysis of polyurethane with TGA and DSC were done by following the

standard procedures.

4.7.4 Results

The results of TGA and DSC of polyurethane produced fi'om the REX—PCL with
the molecular weight of 3000 were plotted in Figure 4-12 and 4-13 respectively. Also the
REX-PCL with the molecular weight of 3000 was plotted on the same graphs for the
comparison. The results of TGA and DSC of polyurethane produced from the REX-PCL
with the molecular weight of 7500 were plotted in Figure 4-14 and 4-15 respectively.
The REX-PCL with the molecular weight of 7500 was also plotted on the same graphs

for the comparison.

91

120

 

100<

 

 

 

 

Weight (96)

I

I

I
I

I
I
l
I
I
I
I
l
I
I
I
I
I
I
I
I
I

.\“

. - I
0 . r - . . "'7 .
4W

 

 

 

o 100 260 360

Temperature ('C)

Figure 4-12: Output of TGA for REX-PCL and polyurethane with the molecular weight
of 3000

 

500

0.5

 

0.0 4

 

Heat Flow (Wig)

 

 

 

-2.5

 

 

4o - 9b 130 V 150

5100‘? Temperature ('C)

Figure 4-13: Output of DSC for REX-PCL and polyurethane with the molecular weight
of 3000

230 ' zéo

92

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

120
IUD-t
w-
g
a w-
£
‘0‘
I
204
0 v I
0 1m SIX)
TW (‘0)
Figure 4-14: Output of TGA for REX-PCL and polyurethane with the molecular weight
of 7500
1
E
g .
’33
I
i I]
.3 - T , I v . v , . .
40 90 140 1% 240 290
5"“ Team ('6)
Figure 4-15: Output of DSC for REX-PCL and polyurethane with the molecular weight
of 7500

93

4.7.5 Discussion

Both Figure 4-12 and 4-14 show a significant decrease in thermal stability as the
amount of diisocyanate contents increase. This was true for the REX-PCL with the
molecular weight of 3000 and 7500; However, the reduction in thermal stability is more
significant for the REX-PCL with the lower molecular weight.

A similar trend was observed from the results of DSC. Figure 4-13 and 4-15
show the reduction in melting temperature when the diisocyanate was added. This too

was significant for the REX-PCL with lower molecular weight.

4.8 Conclusions

The results of TGA and DSC, as well as the melt fi'acture observed during the
experiments indicates that there had been some reaction taken place between REX—PCL
and diisocyanate. However, fiIrther analysis of the material properties are necessary in
order to determine the type of polymer as well as the application that this type of

materials can be used.

94

Chapter 5

 

TRANSESTERIFICATION OF STAR-POLYCAPROLACTONE AND

POLYETHYLENE TEREPHTHALATE

 

5.1 Introduction

Blending process, such as transesterification, plays an important role in the
current polymer industry. By blending two or more materials, one can obtain a material
with the properties that was not available from any of the original materials. However,
the production of thermodynamically stable polymer blend with high molecular weight
polymers is usually difficult, since the chain connectivity makes the entropy of mixing of
these systems very small”.

The transesterification is a reaction of esters and alcohols in an acid- or base-
catalyzed transformation“. This type of blending process has been studied by a large
number of researchers with various types of polymers, catalysts, and process conditions.
One of the examples is the transesterification of polyethylene terephthalate (PET) and
polycaprolactone (PCL) in a twin screw extruder, using stannous octoate as a catalyst.
This process was developed by Weirning Tang et al“). In this process, the PET was
initially fed to the extruder to melt. When the PET was molten, the catalyst and the
caprolactone monomer were added to carry out the transesterification. This scheme
would yield low PCL formation. On the other hand, the monomer and the catalyst can be
added at the beginning of the extruder to carry out sorrre degree of polymerization. Then,

another ingredient such as PET, can be added at the middle of the extruder. This scheme

95

would result in a more firm control of PCL polymerization. However, whether the PET
melts fast enough to react effectively within the extruder remains a question. In addition
to the feeding scheme, the choice of a catalyst to be used is an important issue in planning
the transesterification process.

In this study, four types of catalyst, triphenylphosphine solution,
triphenylphospho ine pellets, dibutyltin dilaurate, and titanium (IV) isopropoxide, were
evaluated for the transesterification of REX-PCL and PET. The reaction mechanism is
shown in Figure 5-1. The various feeding schemes were tested in order to carry out the

process successfully.

 

0 O
a II
—(C-@-c-O(CH2)209_ + (u)
o o - (CH2)5 - c )a—OH
II o
.. .. n A] ' '
Flo—(O (CH2)5 C)— _L(O-(CH2)5-C)rOH
o ‘ o. E.) .‘
e —(C-@--C-O(CH2)20-C(CH2)509._

Figure 5-1: A reaction mechanism of the transesterification of REX-PCL and PET

96

With above mechanism, mixed ester product has no ordered structure, therefore it does
not crystallize (amorphous polymer).
The focus of this study was to determine the catalyst and the process conditions

that can be used to successfully carry out the transesterification of REX-PCL and PET.

5.2 Background
5.2.1 PET

Polyethylene terephthalate (PET) is one of the most widely used plastics. Its
applications range from food storage and container to the electronics packaging and
natural gas distribution pipes. PET is a semi-crystalline thermoplastic, and it turns white
when heated. Some of the advantage of PET over other thermoplastics, such as nylon
and acetal, are its high wear and abrasion resistance, as well as the better resistance to
acidic solutions and chlorinated aqueous solutions. Latter property is especially useful
for the application of PET in food packaging, because acidic solutions and chlorinated
aqueous solutions are commonly used to sanitize food processing equipment”. A

structure of PET is shown in Figure 5-2.

o I?
4&-@-c-O(CH2)20%—

Figure 5-2: A structure of polyethylene terephthalate (PET)

97

5.2.2 Screw Configuration

In the process of transesterification, different materials are to be mixed randomly,
and a catalyst has to be equally distributed throughout the materials. For this reason,
screw configuration #1, which consist of approximately equal amount of kneadling and

conveying elements, was used for the transesterification process.

5.3 Outline of Chapter

This chapter discusses the steps that took place in order to determine the catalyst
and process conditions for the transesterification of REX-PCL and PET.

Section 5.4 discusses the transesterification of REX-PCL and PET, using
triphenylphosphine solution (TPP solution) as a catalyst. TPP solution refers to the
solution in which triphenylphosphine was dissolved in toluene. Various feeding schemes
were tested. This topic will also be discussed in section 5.4.

Due to the problem caused by the solvent used to make the TPP solution, TPP
pellets was used as a catalyst, without being dissolved, to carry out the transesterification
of REX-PCL and PET. This topic will be discussed in section 5.5.

In sections 5.6 and 5.7, transesterification of REX-PCL and PET with two more
catalysts, dibutyltin dilaurate (section 5.6) and titanium (IV) isopropoxide (section 5.7),
were done, and the results will be discussed.

Finally in the Conclusion (section 5.8), the recommended process conditions and

type of catalyst, as well as the topics to be improved will be discussed.

98

5.4 Transesterification of REX-PCL and PET with Triphenylphosphine Solution
5.4.1 Objective

Purposes of this experiment were to determine whether triphenylphosphine
solution can be used as a catalyst in the transesterification of REX-PCL and PET, and to
determine the optimum process conditions for the process. Latter part mainly focused on

varying extrusion temperatures and feeding locations.

5.4.2 Materials

REX-PCL with the molecular weight of 100,000 was produced by following the
procedure discussed in Chapter 3. The REX-PCL was dried at room temperature for a
few days before further processing. The commercial grade PET was obtained fi'om
Eastman Chemical Company and used throughout this study. Triphenylphosphine was
chosen as a catalyst in this study because it is one of the commonly used catalyst for the
transesterifrcations. It was purchased fiom Aldrich Chemical Company in the pellet form

and used for the first part of this experiment. A structure of triphenylphosphine is shown

00

Figure 5-3: A structure of triphenylphosphine

in Figure 5-3.

99

5.4.3 Procedure
5.4.3.1 Preparation of triphenylphosphine Solution

Triphenylphosphine pellets were dissolved in toluene (20 wt%
triphenylphosphine) under a chemical hood. The fraction of triphenylphosphine in the

solution was chosen as 20 wt%, which was chosen merely for the convenience.

5.4.3.2 Determination of Feeding Location

A schematic of the possible feeding locations are shown in Figure 5-4.

I I II

feed throat I 2

 

5 6 die

 

 

 

 

 

 

 

 

 

 

flow direction p

 

Figure 5-4: A Schematic of possible feeding location

There was no vent port available for additional feed at zones 2 and 5, and zone 6
was not considered as a choice of feeding location since it was too close to the die.

In this experiment, six feeding combinations were tested as shown in Table 5-1:

 

 

 

 

 

 

 

Combination # PET REX-PCL catalyst
1 feed throat feed throat feed throat
2 feed throat feed throat zone 1
3 feed throat feed throat zone 3
4 feed throat feed throat zone 4
5 feed throat zone 1 zone 4
6 feed throat zone 3 zone 4

 

 

 

 

 

 

Table 5-1: Six Feeding Combinations Tested

PET was always fed at the feed throat since it has higher melting temperature than

REX-PCL.

100

5.4.3.3 Extrusion Temperature
Since the melting temperature of PET is about 250 °C, the temperature at this
range was used in this experiment. This temperature range was still in the “safe” range

where REX-PCL does not degrade (based on the results of TGA shown in Chapter 3).

5.4.3.4 Reactive Extrusion Process

Equal ratio of REX-PCL and PET with various amounts of catalyst were fed into
the extruder at a various feeding locations discussed in section 5.4.3.2.

The screw speed of the extruder was varied between 30 rpm to 150 rpm to
determine the speed that works best for each of the feeding combinations. In addition,
the feeding rate of REX-PCL and PET was kept at 20 g/min (total) to minimize the
“overflow” at the open zones, which will be discussed later in this chapter. The two
different ratios of polymer mixture to catalyst were tested: they were 20 to 1 (5 wt%
catalyst) and 40 to 1 (2.5 wt% catalyst).

The mixture of REX-PCL and PET with different amounts of triphenylphosphine
solution were collected and tested with TGA and DSC. Also, the mixture of PCL and
PET without the catalyst was processed in an extruder at the same processing condition

and tested with TGA and DSC for comparison.

101

5.4.4 Results
The output of TGA of the product with no catalyst and 5 wt% catalyst were
plotted in Figure 5-5. The output of DSC of products with no catalyst, 5 wt% catalyst,

and 2.5 wt% catalyst were plotted in Figure 5-6.

 

 

 

 

 

 

 

 

............ no catalySt
_ 5 wt%
120
100 .4
80—
‘2":
E 60‘
4
40..
I
20“
o a . ' 1 T ' j t
0 1m 2m am 400

Temperature (‘0)

Figure 5-5: Output of TGA for the REX-PCL/PET mixture with triphenylphosphine
solution as a catalyst

102

----- no catalyst

 

 

 

 

 

 

-—— 2.5 wt%
5 wt%
0.2
0.0«
3
E
3
LL
ii
E
1.340 9'0 ' 150 1930 230 ' 290
5’0 UP Temperature (°C) .
Figure 5-6: Output of DSC for the REX-PCL/PET mixture with triphenylphosphine
solution as a catalyst.

5.4.5 Discussion

By evaluating the various fwding combinations, it was found that the feeding
combination #4 (PET and REX-PCL were fed at the feed throat and catalyst solution was
fed at the zone 4) was the only one that worked in this experiment. This was due to two
reasons: when REX-PCL was fed into the extruder from the zones other than feed throat,
they tended to stick on the top of molten polymer and accumulates as more REX-PCL
was fed; and because the PET did not completely melt until zone 4, feeding the catalyst at

the locations other than zone 4 caused the solid polymers to stick to each other and start

103

to accumulate at the zone. Also, the feed rate of polymer mixture was kept at 20 g/min
and the screw speed of 100 rpm was used. This combination was found to work best
without having any overflow at the open zones.

The polymers made in this experiment had a rough surface and they looked
slightly brown. The brownish color was slightly darker for the one with a higher catalyst
content. Additionally, the results of DSC showed that the transesterification did not
occur. Furthermore, the result of TGA shows that the polymer with 5 wt% catalyst
content started to degrade at low temperature (approximately at 100 °C). These
observations may be indicating that the 5 wt% catalyst sample was partially degraded.

One of the possible explanations to this result is the poor choice of the solvent.
Toluene was chosen as a solvent because of the fact that both PCL and catalyst can be
dissolved in it, and it evaporates quickly at a high temperature. Latter led to the
assumption that the immiscibility of toluene and PET has a negligible effect on the
transesterification. However, this assumption turned out to be bad since the product
polymer did not look well-mixed.

A use of another solvent was considered; however, because the common solvents
do not dissolve PET, the triphenylphosphine pellets without being dissolved in a solvent

was chosen as a next step. This topic will be discussed in the next section.

5.5 Transesterification of REX-PCL and PET with Triphenylphosphine Pellets
5.5.1 Objective
Since the transesterification of the REX-PCL and PET with triphenylphosphine

solution was unsuccessful, possibly due to the use of a toluene as a solvent,

104

triphenylphosphine pellets (without use of solvent) was used as a catalyst to carry out the
transesterification process in this experiment.
Objective of this experiment was to determine the effect of triphenylphosphine as

a catalyst in carrying out the transesterification of REX-PCL and PET.

5.5.2 Materials

The same REX-PCL and PET as used in the previous experiment was used. Also,
the same triphenylphosphine as used in the previous experiment was used, except in this
case, it was fed to the extruder in the form of pellets instead of being dissolved in a

toluene.

5.5.3 Procedure
5.5.3.1 Reactive Extrusion Process

Since the solid was only successful to be fed into the extruder from the feed
throat, triphenylphosphine pellets were fed into the extruder with REX-PCL and PET at
the feed throat. Because all the mterials were fed into the extruder from the feed throat,
all the other vent port were kept closed. As before, equal weight ratio of REX-PCL to
PET was used, and 5 wt% and 2.5wt% catalysts were fed. The same temperature profile

and screw speed as used in previous experiment was applied.

5.5.3.2 Characterization
Product obtained from the transesterification of REX-PCL and PET with

triphenylphosphine pellets as a catalyst was analyzed with TGA and DSC.

105

5.5.4 Results

The output of TGA of the product with no catalyst and 5 wt% catalyst are plotted

in Figure 5-7. The output of DSC of products with no catalyst, 5 wt% catalyst, and 2.5

wt% catalyst were plotted in Figure 5-8.

........... no catalyst
__ 5 wt%

 

120

 

100*

80‘

Weight (%)

20‘

 

 

 

1 ' I

0 f I V I V . v
0 100 200 . 300 400 . 500
Temperature ('0)

 

Figure 5-7: Output of TGA of REX-PCL/PET mixture with u'iphenylphosphine pellets as
a catalyst

106

 

0.2

 

 

 

 

————— no catalyst
—-- 2.5 wt%
0.0“ — 5 wt%
’23
E
E
‘6
0
I
1.6 T I I ' l l T
40 90 140 190 240 290
3° ”9 Temperature (°C)

Figure 5-8: Output of DSC for REX-PCL/PET mixture with triphenylphosphine pellets as
a catalyst

5.5.5 Discussion

The polymers made in this experiment had smooth appearance, though the ones
with catalyst added seemed slightly brown with the one with more catalyst content being
darker. The results of DSC showed a reduction in the size of REX-PCL melting point
peak; however, no change can be observed with the peak of PET. Also, the result of
TGA shows the similar trend as the one observed in the previous experiment (Figure 5-
3). These results indicate that the triphenylphosphine is not suitable for the catalyst of
transesterification between REX-PCL and PET. Therefore, another type of catalyst was

evaluated in the next section.

107

5.6 Trausesterif'rcation of REX-PCL and PET with Dibutyltin Dilaurate
5.6.1 Objective
Objective of this experiment was to determine the effect of dibutyltin dilaurate as

a catalyst in carrying out the transesterification of REX-PCL and PET.

5.6.2 Materials
Same REX-PCL and PET as previous experiments were used.
Dibutyltin Dilaurate was purchased from Aldrich Chemical Company to be used

as a catalyst. A structure of dibutyltin dilaurate is shown in Figure 5-9.

0
II
I
o —cl:I —c H,(CH,),CH3
0

Figure 5-9: A structure of dibutyltin dilaurate
5.6.3 Procedure
5.6.3.1 Reactive Extrusion Procees

As before, both REX-PCL and PET (1:1 weight ratio) were fed through the feed

throat. However, since dibutyltin dilaurate was purchased in a liquid form, it was fed into

108

the extruder through zone 4, as it was determined to work best in the previous experiment
with TPP solution. Also, the same weight fi'action of polymer mixture to catalyst as used
in the previous experiments were applied. The same temperature profile and the screw

speed used in previous experiments were also applied in this section.

5.6.3.2 Characterization
The product obtained from the transesterification of REX-PCL and PET with
dibutyl dilaurate as a catalyst was analyzed with TGA and DSC. Standard procedure as

discussed in Chapter 2 was followed.
5.6.4 Results

The output of TGA and DSC of products with no catalyst, 5 wt% catalyst, and 2.5

wt% catalyst were plotted in Figure 5-10 and 5-11 respectively.

109

 

 

 

 

 

 

 

 

 

 

 

   

 

 

 

120
100 « fimw‘ \
‘ x \
so“ no catalyst
----- 2.5 wt%
3 ............ ..._. 5 wtofi,
g so
i
40 -
‘ 20 r
o o 7 $00 j 250 . 350 . 460 . soc
Temperate! (’C)
Figure 5-10: Output of TGA for the REX-PCL/PET mixture with dibutyltin dilaurate as a
‘ catalyst
0.2
—-—-— no catalyst I
0.01 —— 2.5 Wt%
.......___ 5 wt%
‘6 “’24:- ‘_‘:::/< .f' .
5
.5
.‘t
3
I
-o.a« I»;
4.0 f r ' r ' I ' I x r
. 40 so 140 too 240 290
5“” Temperan- ('C)
Figure 5-11: Output of DSC for REX-PCL/PET mixture with dibutyltin dilaurate as a
catalyst

llO

5.6.5 Discussion

The product seemed less brownish than the ones with triphenylphosphine, and it
had a smooth surface. The result of DSC showed the shift in melting temperature and the
reduction in the size of the peak occured for the product with 5 wt% catalyst. Although
the change is small, the reduction in the size of the peak and the shift of the melting
temperature was also observed for the product with 2.5 wt% catalyst. Additionally, the
result of TGA showed a significant reduction in degradation temperature as the amount
of catalyst was increased.

These observations indicate that the dibutyltin dilaurate can be used as a catalyst
for the transesterification. Also, with the fact that the reduction and a shift in melting
point peak was significant when the catalyst content was doubled, indicates that
increasing the catalyst content above 5 wt% added may yield even better

transesterification between the REX-PCL and the PET.

5.7 Transesterifrcation of REX-PCL and PET with Titanium (IV) Isopropoxide
5.7.1 Objective
Objective of this experiment was to determine the effect of titanium (IV)

isopropoxide as a catalyst in carrying out the transesterification of REX-PCL and PET.

5.7.2 Materials
The same REX-PCL and PET discussed in the previous experiments were used.
Titanium (IV) isopropoxide was purchased fiom Aldrich Chemical Company to

be used as a catalyst. A structure of titanium isopropoxide is shown in Figure 5-12.

111

‘st

CH, C— CHCH,
CH,C H —o -— Ti —0 — CHCH,
CH, CH —0 CH,

CH,

Figure'5-12: A structure of titanium (IV) isopropoxide

5.7.3 Procedure
5.7.3.1 Reactive Extrusion Process

Since titanium (IV) isopropoxide was also purchased in a liquid form, it was fed
into the extruder through zone 4. Also, as before, the equal weight ratio of REX-PCL
and PET were fed into the extruder through feed throat with the feeding rate of 20 g/min.
The same screw speed and temperature profile used in the previous sections were applied.
As before, 5 wt% and 2.5 wt% of catalyst was added for evaluation. The entire process
was carried under nitrogen environment since the titanium isopropoxide is sensitive to a

moisture.

5.7.3.2 Characterization
The product obtained from the transesterification of REX-PCL and PET with
titanium isopropoxide as a catalyst was analyzed with TGA and DSC. Standard

procedure as discussed in Chapter 2 was followed.

112

5.7.4 Results
The output of TGA and DSC of products with no catalyst, 5 wt% catalyst, and 2.5

wt% catalyst were plotted in Figure 5-13 and 5-14 respectively.

---— no catalyst
—— 2.5 wt%
......._.... 5 wt%

 

120

 

100«

Weight (%)
f

 

 

 

 

0 r r f r ' I ". V

II ' too zoo zoo - 400 V 50.0.
Temperatum (‘6)

 

Figure 5-13: Output of TGA for REX-PCL/PET mixtm'e with titanium isopropoxide as a
catalyst. ’ *

113

-—-- no catalyst
__ 2.5 wt%
............... 5 Wto/o

 

0.2

00"

 

 

024W ' ‘ Irv—w.

‘
_‘ -‘————fi.h
\ fl—n——————————-_ .\
~- _—_.— ———‘.__ ”I.
—‘
-_—

 

43.4 ~

Heat Flow (Wig)

-0.6 . ‘

-o.a« I,“

 

 

 

‘1.0 fl T V T ' I V Y 7 I
40 90 140 190 240 290

an Up Temperature ('C)

Figure 5-14: Output ofDSC for REX-PCL/PET mixture with titanium isopropoxide as a
catalyst.

5.7.5 Discussion

The product that came out from the extruder had a dark reddish brown color,
somewhat clear, and was very soft. This was true forthe products ofdifl‘erent catalyst
contents, and the color stayed the same even after being left to be dried for a few days,
with the same soft, stickiness remaining. 3

The results of DSC showed the addition of catalyst resulted in the significant
reduction of two melting point peaks, and a formation of one broad peak at
approximately 120 °C. The result of the polymer with 5 wt% catalyst and 2.5 wt%

“4

catalyst also showed very similar results. These observations indicate that the increase in
catalyst content does not carry the transesterification further.

The result of TGA showed slight reduction in degradation temperature; however,
the amount of reduction was very small compared to the result of the polymer with
dibutyltin dilaurate as a catalyst.

Although the increase in catalyst content did not carry the transesterification
further, the increase in temperature residence time may. This can be done by the addition

of a few mixing elements in the screw configuration.

5.8 Conclusions

Four types of catalysts were evaluated to conduct the transesterification of REX-
PCL and PET. However, only titanium isopropoxide and dibutyltin dilaurate was found
to be effective as a catalyst for the transesterification of REX-PCL and PET.

However, because the titanium isopropoxide was found to be most effective does
not mean it can be recommended for every application. The mixture of REX-PCL/PET
with titanium isopropoxide was found to be very sticky with reddish brown color.
Therefore, this type of catalyst cannot be used for the applications where optical

properties of the nmterials is important.

”5

Chapter 6

 

CONCLUSIONS AND RECOMMENDATIONS

 

6.1 Reactive Extrusion Polymerization of e-Caprolactone

By reproducing the method developed by Krishnan in 1998, with a few
adjustments, REX-PCL with the wide range of molecular weight were produced. Due to
the large monomer to initiator ratio, the highest molecular weight of REX-PCL produced
was limited to 210,000. The lowest molecular weight of REX-PCL was determined as
20,000 due to the low viscosity of the material that resulted in "extruder underload".

The effect of moisture content in the monomer to the formation of REX-PCL was
also evaluated. The results showed the range of molecular weight that can be produced
was narrowed to 30,000 to 80,000 in the case of the monomer with the moisture contents.
As long as the process conditions were met and the moisture was kept away from the
system, the production of REX-PCL shows excellent reproducibility.

Also, the results of the characterization of REX-PCL indicated that the monomer
conversion is near 100 % with the particular system for the molecular weight up to
120,000. Furthermore, the results of characterizations indicated that the REX-PCL is
thermally less stable but more viscous than the linear PCL.

Additional property analysis, such as mechanical tests and determination of the
material density, are recommended in order to determine the applications that the REX-

PCL can be utilized.

116

6.2 Reactive Extrusion Polymerization of Star-Polycaprolactone and
Diisocyanate
The REX-PCL was reacted with hexamethylene diisocyanate to produce
polyurethane. Although the melt fi’actures of the materials, as well as the results of the
characterization, indicated that the reaction between the REX-PCL and the diisocyanate
took place, further analysis are required to determine the exact structure of the material as

well as the application that this particular material can be used.

6.3 Transesterification of Star-Polycaprolactone and Polyethylene Terephthalate

Transesterifrcation of REX-PCL and PET were performed. Four types of
catalysts were evaluated: triphenylphosphine dissolved in toluene; triphenylphosphine
pellets; dibutyltin dilaurate; and titanium isopropoxide. Only titanium isopropoxide and
dibutyltin dilaurate was found to be effective.

Although above two catalysts were found to be effective to carry out the
transesterification of REX-PCL and PET, it is recommended to perform additional
experiments with higher catalyst content for dibutyltin dilaurate and longer residence

time or higher reaction temperature for titanium isopropoxide.

6.4 Limitation with Reactive Extrusion

Although an extruder can be utilized to perform various types of reactions, there
also has been various limitations to the reactive extrusion process that was encountered
during this study. For example, solid was unable to be fed at the middle of the extruder

since they tend to stick on the top of a polymer melt and accumulates at the feeding zone.

ll7

Additionally, the liquid could not be fed into the zones where the solid materials were

present.
If these limitations could be overcome, the application of reactive extrusion

process will be significantly broaden.

118

10.

11.

I2.

13.

REFERENCES

Alfied Rudin, The Elements of Polymer Science and Engineering: An
Introductory Text for Engineers and Chemists, Academic Press Inc., California,
1982

Arthur E. Newkirk, Thermogravimetric Measurements, Anal. Chem, vo 1.32,
1558-1563, 1960

Bjorksten Research Laboratories, Inc., Polyesters and Their Aapplications,
Reinhold Publishing Corporation, New York, 1956

C. Maier and M. Lambla, Esterification in Reactive Extrusion, Polymer
Engineering and Science, vol.35, no.15, 1197-1205, 1995

Chris Raudendaal, Comparison of Twin Screw Extruder, J SW Technical Review,
no. 1 6

D.M. Miley and J. Runt, Transesterification in Poly (Butylene Terephthalate) /
Polyarylate Blends, Polymer, vol.33, no.2], 1992

Daisel Chemical Company, A report produced by Ohtake Plant., March 1998
Dan Mielcarek, Developing, Scaling Up, Implementing and Improving the Twin-
Screw Extrusion Process, Processing Lines, vol.9, no.1, Krupp Werner &
Pfleiderer, 1999

David B. Todd, Reactive Extrusion — Rules of Thumb, Polymer Processing
Institute, 1992

Denise Carlson, Li Nie, Ramani Narayan, and Philippe Dubois, Maleation of
Polylactide (PLA) by Reactive Extrusion, Journal of Applied Polymer Science,
vol.72, 477-485, 1999

EC. Kraemer, Ind. Eng. Chem. Vol.30, 1200, 1938

ES. Watson, M.J. O’NeilL Joshua Justin, and Nathaniel Brenner, A Difl'erential
Scanning Calorimeter for Quantitative Differential Thermal Analysis, Anal. -
Chem, vol.36, 1233-1238, 1964

Echols et al., Extruder Polymerization of Polyurethanes, United States Patent,
Patent Number 4,948,859, 1990

119

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

Eric Martz, Mike Sierocki, Liz Stanley, and Lesley Veresh, Copolyesters by
Transesterification of Two Homopolyesters,
http://www.chem.mtu.edu/classes/ch462/angel/webpage.htm, 2000

Eui-Jun Choi, Chang-Hyeon Kim, and Jung-Ki Park, Synthesis and
Characterization of Starch-g-Polycaprolactone Copolymer, Macromolecules,
vol.32, 7402-7408, 1999

6.6. Risch, G.L. Wilkes, and J.M. Warakomski, Crystallization Kinetics and
Morphological Features of Star-Branched Nylon-6: Effect of Branch-Point
Functionality, Polymer, voL34, no.11, 2330-2343, 1993

GS. Shapoval, T.E. Lipatova, L.S. Sheinina, V.A. Zhabenko, and V.A. Morozov,
The Reactivity of Diisocyanates in the Polyurethane Formation Process, Synthesis
and Physical Chemistry of Urethanes, Soviet Progress in Polyurethanes Series,
vol.2, 197 5

Instructions for the Use of the Ubbelohde Viscometer, Cannon Instrument Co.,
1998

J. Brandrup and E. H. Immergut, Polymer Handbook, 3rd Ed., Wiley-Interscience
Publication, New York, 1989

J. M. Buist and H. Gudgeon, Advances in Polyurethane Technology, John Wiley
and Sons Inc., New York, 1968

J .L.White, Twin Screw Extrusion, Hanser Publishers, New York, 1990

Jack F. Douglas, Yi Feng, Bradford J. Factor, Frederick I. Mopsik, Alamgir
Karim, and Charles C. Han, The Compatibilization of Polymer Blends by Strong
Interactions and Reactions,
http://www.msel.nist.gov/structure/polymer/techactv95/compatplblnd.htm

Jacob John, Jian Tang, Zhihong Yang, and Mrinal Bhattacharya, Synthesis and
Characterization of Anhydride-Functional Polycaprolactone, John Wiley & Sons,
Inc., 1997

K. Peter C. Vollhardt and Neil E. Schore, Organic Chemistry, 2"d ed., W.H.
Freeman and Company, New York, 1994

Kevin Bastian, PET & PETG Drive a ‘Hard’ Bargain, Plastics Machining &
Fabricating, http://www.plasticmachining.com/199802/petg.htm

Krishnan et al., Biodegradable Mufti-Component Polymeric Materials Based on

Unmodified Starch-Like Polysaccharides, United States Patent, Patent Number
5,500,465, 1996

120

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

Ludwig Bottenbruch et al., Engineering Thermoplastics:Polycarbonates,
Polyacetals, Polyesters, Cellulose Esters, Hanser Publishers, Cincinnati, 1996
Malcolm P. Stevens, Polymer Chemistry: An Introduction, 2"d Ed., Oxford
University Press, New York, 1990

ML. Huggins, J. Am. Chem. Soc. Vol.64, 2716, 1942

Molmn Krishmn, Engineering of Compositions based on Star-Polycaprolactone
Derived by Extrusion Polymerization of Epsilon-Caprolactone, Dissertation,
Michigan State University, 1998

Narayan et al., Bulk Reactive Extrusion Polymerization Process Producing
Aliphatic Ester Polymer Compositions, United States Patent, Patent Number
5,801,224, 1998

Neil P. Stuber and Matthew Tirrell, Continuous Polymerization Studies in a
Twin-Screw Extruder, Polymer Process Engineering, 3 (1 & 2), 71-83, 1985

Philippe Dubois, Mohan Krishnan, and Ramani Narayan, Aliphatic Polyester-
Grafted Starch-Like Polysaccharides by Ring-Opening Poolymerization, Polymer,
vol.40, 3091-3100, 1999

RA. Sanayei, S. Pang, and A. Rudin, A New Approach to Establishing Universal
Calibration Curves for Size Exclusion Chromatography, Polymer, vol.34, no.11,
2320-2323, 1993

Ramani Narayan, Commercializing Technology: A Case Study of Starch-Based
Biodegradable Plastics Technololgy, AOCS Press, 1998

Richard M. Taylor, Liquid Flow Metering for Extruder Feeding, Processing
Lines, vol.8, no.2, Krupp Werner & Pfleiderer

S.F. Sun, Temperature and Solvent Effects on the Intrinsic Viscosities of a Star-
Shaped Polystyrene, Polymer, vol.28, no.2, 283-287, 1987

Shiow C. Lin, New Polymers Prepared fiom N-Cyanourea Conrpound, New
Monomers and Polymers, Polymer Science and Technology, vol.25, 1984

Twin-Screw Extruder Reactors for the Continuous Production of Solid
Polyurethane Products, Plastic Technology, Werner & Pfleiderer

Weirning Tnag, N. Sanjeeva Murthy, Frank Mares, Milton E. McDonnell, Sean A.
Curran, Poly(ethylene terephthalate)-Poly(capro lactone) Block Copolymer. I.
Synthesis, Reactive Extrusion, and Fiber Morphology, Journal of Applied
Polymer Science, vol.74, 1858-1867, 1999

121

41. William Thiele, Reactive Compounding — With Your Extruder, Plastics:
Formulating & Compounding, vol.2, no.6, 1996

I22

 

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIII