View metadata, citation and similar papers at core.ac.

uk brought to you by CORE

provided by Stellenbosch University SUNScholar Repository

Synthesis and characterization of

unsaturated polyesters for use in multi-
vesiculated particles (MVPs)

by

Jaylin Mitch Simpson

Thesis presented in partial fulfilment of the requirements for the degree

Master of Science (Polymer Science) at the University of Stellenbosch

Supervisor: Prof. P. E. Mallon

Co-supervisor: Dr. J. B. McLeary
Faculty of Science
Department of Chemistry and Polymer Science

December 2010
 Declaration

By submitting this thesis/dissertation electronically, I declare that the entirety of the work
contained therein is my own, original work, and that I have not previously in its entirety or in part
submitted it for obtaining any qualification.

December 2010

Copyright © 2010 University of Stellenbosch

All rights reserved

 ABSTRACT

Unsaturated polyesters resins (UPRs) of maleic anhydride (MA), phthalic anhydride (PA) and
propylene glycol (PG) were synthesized using the fusion polycondensation process for use in
Multi-vesiculated Particles (MVPs). The UPRs were synthesized using different MA:PA mole
ratios and process parameters, including heating rates, agitation speed, exotherm rate and
maximum processing temperature. Design of Experiments (DoE) software (Design Expert 7)
was employed to find the optimum experimental space, i.e. least amount of experiments, but
covering all the factors. The variations in the formulation and process parameters had a
significant effect on the molecular structure and physical properties of the UPRs. The
molecular structure and physical properties of the UPRs was successfully determined using
various techniques including viscometry, acid-base titration, Proton Nuclear Magnetic
Resonance (1H NMR) spectroscopy, Size-Exclusion Chromatography (SEC), Fourier
Transform Infrared (FTIR) spectroscopy and Differential Scanning Calorimetry (DSC).
Furthermore, the carboxyl and hydroxyl end-group concentrations of the UPRs were
determined by derivatization of the groups with trichloroacetyl isocyanate (TAI) and analysis
by 1H NMR.

As a second part of the study, the effect of the above-mentioned monomer ratio and process
factors of the UPRs, on the properties of the MVPs (produced from the UPRs), was
investigated. Results showed that these factors had a significant effect on the particle size and
degree of vesiculation of the MVPs. The degree of vesiculation of the MVPs was determined
by Scanning Electron Microscopy (SEM). Attempts were also made to determine the relative
hardness of the MVPs by AFM and microhardness testing to determine a relationship with
UPRs properties (e.g. molecular weight, degree of unsaturation and chain branching). These
techniques were however found to be unsuitable due to the physical nature of the MVPs.
 OPSOMMING

Onversadigde poliësters (OPs) van maleïensuuranhidried (MA), fataalsuuranhidried (PA) en

propileen glikol (PD) is berei deur die fusie polikondensasie-proses vir gebruik in multi-
vesikulerende partikels (MVPs). Die OPs is berei deur gebruik te maak van verskillende
MA:PA molverhoudings en reaksie faktore wat verhittingstempo‟s, roerspoed, eksoterm
tempo en maksimum reaksie temperatuur, insluit. As gevolg van die groot hoeveelheid
faktore is eksperimentontwerp sagteware (Design Expert 7) gebruik om die aantal
eksperimente te verminder, maar ook waardevolle afleidings van die data te maak. Die
verskille in die formulasie en reaksie faktore het „n merkwaardige effek op die molekulêre
struktuur en fisiese eienskappe van die OPs gehad. Hierdie eienskappe is bepaal m.b.v.
verskeie tegnieke, naamlik viskometrie, suur-basis titrasie, Proton Kern-Magnetiese
Resonansie (1H KMR) spektroskopie, Grootte-Uitsluitings-Chromatografie (SEC), Fourier
Transform Infrarooi (FTIR) spektroskopie en Differensiële Skandeerings Kalorimetrie (DSC).
Die hidroksiel en karboksielgroep konsentrasies van die OPs is bepaal deur hul reaksie met
trichloroasetiel-isosianaat (TAI) gevolg deur 1H KMR analise.

In die tweede deel van hierdie studie is die invloed van bogenoemde faktore op die
eienskappe van die MVPs ondersoek. Die resultate het getoon dat hierdie faktore „n
merkwaardige effek op die partikel grootte en graad van “vesiculation” van die MVPs gehad
het. Die graad van “vesiculation” van die MVPs is bepaal met behulp van Skandeer Elektron
Mikroskopie (SEM). Pogings is ook gemaak om die relatiewe hardheid van die MVPs te
bepaal deur middel van AFM en Mikro-hardheid toetsing. Dit is gedoen om die verhouding
van die OPs eienskappe (molekulêre gewig, graad van onversadigheid en sy-kettings) vas te
stel. Hierdie tegnieke is egter ongeskik bevind as gevolg van die inherente fisiese aard van die
OPs.
 ACKNOWLEDGEMENTS

I would first like to thank God for the health and strength to carry out this study.

Secondly, I would like to thank my parents for all their support during my studies.

I would also like to thank my promoters Prof. P. E. Mallon and Dr. J. B. McLeary for their

help and support, and the opportunity to carry out this MSc study.

I would also like to thank the following people for their contributions:

1. Mr. Mohammed Jaffer (University of Cape Town) for the microtoming assistance

2. Mrs. M. R. Franzenburg for the SEM assistance

3. Dr. M. Meincken (University of Stellenbosch) for the AFM assistance

4. Dr. J. Brand and Elsa Malherbe (University of Stellenbosch) for the NMR assistance

5. Dr. N. Mautjana for the SEC assistance.

Lastly, I would like to thank my Plascon colleagues, especially Bertus Smit and Stephen

Clarke, for all their support.

Dedicated to my mother and father
 LIST OF CONTENTS

LIST OF CONTENTS i

LIST OF FIGURES iv

LIST OF TABLES vii

LIST OF ABBREVIATIONS viii

LIST OF SYMBOLS x

CHAPTER 1: INTRODUCTION AND OBJECTIVES 1

1.1 Introduction 1

1.2 Objectives 2

1.3 References 3

CHAPTER 2: HISTORICAL AND THEORETICAL BACKGROUND 4

2.1 Polycondensation: Polyesters 4

2.1.1 Introduction 4

2.1.2 Polycondensation 6

2.1.3 Reactions 9

2.2 Suspension polymerization 16

2.2.1 Introduction 16

2.2.2 History 17

2.2.3 Process of suspension polymerization 17

2.2.4 Control of particle size in suspension polymerization 21

2.2.5 Bead-suspension polymerization kinetics 23

2.3 Suspension polymerization – Vesiculated particles 24

2.3.1 Introduction 24

i
 2.3.2 Single vesicle particles 24

2.3.3 Multi-vesiculated particles (MVPs) 25

2.4 Design of Experiments (DoE) 30

2.4.1 Introduction 30

2.5 References 32

CHAPTER 3: EXPERIMENTAL 39

3.1 Introduction 39

3.2 Materials 40

3.3 Unsaturated polyester (UPR) synthesis 40

3.4 Design of Experiments (DoE) 42

3.5 Multi-vesiculated particle (MVP) synthesis 43

3.6 Analytical techniques and measurements 44

3.6.1 Size-Exclusion Chromatography (SEC) 45

3.6.2 Dynamics Scanning Calorimetry (DSC) 45

3.6.3 Nuclear Magnetic Resonance (NMR) Spectroscopy 45

3.6.4 Scanning Electron Microscopy (SEM) 45

3.6.5 Microtoming 46

3.6.6 Laser diffraction (LD) 46

3.6.7 Acid-value 46

3.6.8 Solids content 47

3.6.9 Viscosity 47

3.6.10 Contrast ratio (opacity) 48

3.6.11 Hardness – microhardness and Atomic Force Microscopy (AFM) 48

ii
CHAPTER 4: RESULTS AND DISCUSSION 50

4.1 Design of Experiments (DoE) analysis 50

4.1.1 Introduction 50

4.1.2 Results and discussion – UPRs 50

4.1.2.1 Brookfield viscosity 54

4.1.2.2 Acid-value 61

4.1.2.3 Number- and weight-average molecular weights (Mn and Mw) 64

4.1.2.4 Polydispersity index (PDI) 68

4.1.2.5 Preliminary conclusion: UPRs 70

4.1.3 Results and discussion – MVPs 71

4.1.3.1 SMD-values 73

4.1.3.2 VMD-values 74

4.1.3.3 Preliminary conclusion: MVPs 79

4.2 Detailed analysis of selected DoE runs 80

4.3 Effect of UPR maximum processing temperature on properties of the UPR

and the resultant MVPs 85

4.4 Determination of hardness of MVPs 92

4.5 Chain-branching 94

4.6 End-group analysis by isocyanate derivatization and 1H NMR 97

4.7 References 101

CHAPTER 5: CONCLUSIONS 103

5.1 Conclusions 103

5.2 Recommendations for future work 105

iii
 LIST OF FIGURES

CHAPTER 2
Figure 2.1: Generalized reaction scheme of esterification. 6
Figure 2.2: Reaction scheme of polyesters via self-condensation of ω-hydroxy acids. 6
Figure 2.3: Reaction scheme of polyesters via condensation of diols with diacids. 7
Figure 2.4: Reaction scheme of polyesters via ester exchange with a diol. 7
Figure 2.5: Reaction scheme of polyesters via ring-opening of lactones. 7
Figure 2.6: Reaction scheme of polyesters via alcoholysis. 7
Figure 2.7: Formation of the UPR, poly (propylene fumarate phthalate). 8
Figure 2.8: Reaction scheme of monoester formation. 9
Figure 2.9: Reaction scheme of polyesterification. 9
Figure 2.10: Reaction scheme of cis-trans isomerization (maleate to fumarate). 10
Figure 2.11: Reaction scheme of the Ordelt saturation of monoester by diol (short chain
branches) 12
Figure 2.12: Reaction scheme of the Ordelt saturation of monoester by monoester (long
chain branches). 13
Figure 2.13: Reaction scheme of transesterification side reaction during
polyesterification. 14
Figure 2.14: Illustrative scheme of breakage and coalescence mechanisms involved in
suspension polymerization. 20
Figure 2.15: Schematic diagram of the Ropaque opaque polymer particle. 25
Figure 2.16.1: Scanning electron micrograph of Spindrift particles. 26
Figure 2.16.2: Scanning electron micrograph of a cross-section of a Spindrift particle. 26
Figure 2.17: Micelle formation. 28
Figure 2.18: Crosslink reaction of unsaturated polyester and styrene. 29

CHAPTER 4
Figure 4.1: Two component mix graph for the formulation factors affecting the viscosity
of the UPR-STY mixtures. 55
Figure 4.2: Contour mix-process graph of the interaction effect (BC) of heating rate 1
and %MA in MA:PA mole ratio on the Brookfield viscosity of UPR-STY mixtures. 57
Figure 4.3: Contour mix-process graph of the interaction effect (BD) maximum
temperature and %MA in MA:PA mole ratio on the Brookfield viscosity of UPR-STY
mixtures. 58

iv
Figure 4.4: Pertubation graph for the process factors affecting viscosity of the UPR-STY
mixtures (fixed MA:PA ratio at 26.19: 13.18). 60
Figure 4.5: Two component mix graph for the formulation factors affecting the acid-
value of UPRs. 62
Figure 4.6: Contour mix-process graph of the interaction effect (AF) of %PA in MA:PA
mole ratio and agitation speed on the acid-value of UPRs. 63
Figure 4.7: Contour mix-process graph of the interaction effect (BF) of %MA and
agitation speed on acid-value of UPRs. 64
Figure 4.8: Two component mix graph for the formulation factors affecting the Mw-value
of UPRs. 66
Figure 4.9: Contour mix-process graph of the interaction effects AD and ABD, i.e.
interactions between %PA, MA:PA mole ratio and maximum temperature on Mw-value of
UPRs. 67
Figure 4.10: Two component mix graph for the formulation factors affecting the PDI-
value of UPRs. 69
Figure 4.11: Contour mix-process graph of the interaction effect ABD, interaction
between MA:PA ratio and the maximum temperature on PDI-value of UPRs. 70
Figure 4.12: : Two component mix graph for the formulation factors of the UPRs
affecting SMD-value of the MVPs. 74
Figure 4.13: Contour mix-process graph of the interaction effect BD, i.e. interaction
between %MA and the maximum temperature of the UPRs on SMD-value of MVPs. 75
Figure 4.14 Two component mix graph for the formulation factors of the UPRs affecting
VMD-value of the MVPs. 77
Figure 4.15: Contour mix-process graph of the interaction effect BD, i.e. interaction
between %MA and the maximum temperature of the UPRs on VMD-value of MVPs. 78
Figure 4.16: Contour mix-process graph of the interaction effect BE, i.e. interaction
between %MA and the heating rate 2 of the UPRs on VMD-value of MVPs. 79
Figure 4.17: SEM images of (a) MVPs of RUN 6 at 100× magnification and (b)
microtomed MVPs of RUN 6 at 1790 × magnification. 82
Figure 4.18: SEM images of (a) MVPs of RUN 26 at 100× magnification and (b)
microtomed MVPs of RUN 26 at 735 × magnification. 82
Figure 4.19: Surface area- and cumulative distributions of the particle sizes of MVPs of
DoE run 6 and DoE run 26. 84
Figure 4.20: SEC chromatograms of UPRs prepared at different maximum processing
temperatures. 86

v
Figure 4.21: Schematic diagram of the effect of chain branching on the acid-value of the
UPRs. 87
Figure 4.22: DSC curves of the UPRs prepared at different maximum processing
temperatures. 88
Figure 4.23: SEM micrograph of microtomed MVPs of a) RUN 29 at 622× magnification
and (b) RUN 33 at 913× magnification. 89
Figure 4.24: Surface area- and cumulative distributions of the particle sizes of MVPs of
runs 29 (red) and 33 (blue). 91
Figure 4.25: AFM 3D image (a) and force-distance curve (b) of a sub-micron MVP of
RUN 29. 93
Figure 4.26: AFM 3D-image (a) and force-distance curve (b) of a sub-micron MVP of
RUN 33. 94
Figure 4.27: 1H NMR spectrum of polyester with phthalic anhydride. 95
1
Figure 4.28: H NMR spectrum of polyester with maleic anhydride. 95
1
Figure 4.29: H NMR spectrum of UPR of RUN 29. 96
Figure 4.30: Derivatization of OH-groups of UPRs with TAI. 98
Figure 4.31: Derivatization of COOH-groups of UPRs with TAI. 98
Figure 4.32: FTIR spectrum of the UPRs of RUN 29 and RUN 29 derivatized with TAI. 99
Figure 4.33: 1H NMR spectrum of UPR of RUN 29 derivatized by TAI. 100

vi
 LIST OF TABLES

CHAPTER 2
Table 2.1: Isomeration vs glycol type. 11
Table 2.2: Isomerization vs reaction temperature. 11
Table 2.3: Level of unsaturation. 15
Table 2.4: Typical properties of Ropaque. 24

CHAPTER 3
Table 3.1: Standard unsaturated polyester formulation. 41
Table 3.2: Formulation and process factors for unsaturated polyester synthesis. 42
Table 3.3: Standard MVPs formulation. 44

CHAPTER 4
Table 4.1: Experimental design space for UPR synthesis. 52
Table 4.2: Significant factors affecting the Brookfield viscosity of UPR-STY mixtures. 54
Table 4.3: Significant factors affecting the acid-value of UPRs. 61
Table 4.4: Significant factors affecting the Mw-value of UPRs. 65
Table 4.5: Significant factors affecting the PDI-value of UPRs. 68
Table 4.6: Particle size results of MVPs. 72
Table 4.7: Significant factors affecting SMD-value of the MVPs. 73
Table 4.8: Significant factors affecting VMD-value of the MVPs. 76
Table 4.9: Formulation and process parameters of DoE runs 6 and 26. 80
Table 4.10: Properties of UPRs of DoE runs 6 and 26. 80
Table 4.11: Properties of MVPs with UPRs of DoE runs 6 and 26. 80
Table 4.12: Effect of UPR maximum processing temperature on properties of the UPRs of
runs 29 – 33. 85
Table 4.13: Effect of UPR maximum processing temperature of runs 29 – 33 on
properties of the MVPs. 88
Table 4.14: Degree of branching results of UPRs of runs 29 and 33 via 1H NMR and
SEC. 96
Table 4.15: Relative hydroxyl -and carboxyl group percentages of UPRs of runs 29 to 33. 100

vii
 LIST OF ABBREVIATIONS

AV Acid-value
AFM Atomic Force Microscopy
ANOVA Analysis of variance
CHP Cumene hydroperoxide
DB Degree of branching
DCPD Dicyclopentadiene
DEG Diethylene glycol
DSC Differential Scanning Calorimetry
DoE Design of Experiments
DETA Diethylene triamine
EG Ethylene glycol
FTIR Fourier Transform Infra-red spectroscopy
HEC Hydroxy ethyl cellulose
HPLC High Performance Liquid Chromatography
IPA Isophthalic acid
KOH Potassium hydroxide
LD Laser Diffraction
LMA Lauryl methacrylate
MA Maleic anhydride
MALDI-MS Matrix-Assisted-Laser-Desorption-Ionization-Mass-
Spectrometry
MEKP Methyl-ethyl ketone peroxide
MMA Methyl methacrylate
MVPs Multi-vesiculated particles
NMR Nuclear magnetic resonance
OH Hydroxyl
PA Phthalic anhydride
PET Poly (ethylene terephthalate)
PG Propylene glycol
PVA Poly (vinyl acetate)
PVOH Poly (vinyl alcohol)

viii
RSM Response surface methodology
SEM Scanning Electron Microscopy.
STY Styrene
TAI Trichloroacetyl isocyanate
TBC Tert-butyl catechol
Tg Glass transition temperature
THF Tetrahydrofuran
TiO2 Titanium dioxide
UPR Unsaturated polyester resin
ZnO Zinc oxide

ix
 LIST OF SYMBOLS

D NMR intensity of branched unit in polyester chain

DPn Degree of polymerization
L NMR intensity of linear unit in polyester chain
Mn * Growing polymer chain
Mn Number average molecular weight
Mw Weight average molecular weight
N Normality of KOH solution
PDI Polydispersity index
SMD Sauter Mean Diameter
SUPR-STY Solids content of UPR-STY mixture
VKOH Volume of KOH solution used in acid-base titration
VMD Volume Mean Diameter
wi Sample weight
WUPR-STY Weight of UPR-STY mixture
R* Radical generated by dissociation of the initiator
Pn Polymer with n repeat units
d Average particle size
k Parameters e.g. apparatus design, type of stirrer
Dv Diameter of vessel
Ds Diameter of stirrer
R Volume ratio of droplet phase / suspension medium
N Stirring speed (or power of mixing)
va Viscosity of the droplet phase
vm Viscosity of the suspension medium
ε Interfacial tension between two immiscible phases
Cs Stabilizer concentration

x
 CHAPTER 1

INTRODUCTION AND OBJECTIVES

1.1 Introduction

Titanium dioxide (TiO2) is the primary opacifying pigment in the coatings industry. The
cost of the TiO2 pigment have soared in the past few decades due to increased
consumption in a variety of applications, therefore alternatives are attractive 1. It is
known that opacity of a coating can be increased by incorporating microvoids or micro-
vesicles into a coating as they scatter light1. Micro-voids can be incorporated into a
coating by introducing organic opacifying particles, such as the commercial products
Multi-vesiculated particles® (MVPs) produced by Plascon and single-vesiculated
particles like Ropaque Opaque Polymer® produced by Rohm & Haas.

The present study focuses on MVPs which are dispersions of crosslinked

(thermosetting) non-film forming particles in water. The particles are formed by the
reaction of the unsaturated polyester resin (UPR) with reactive diluent styrene (STY)
and diethylene triamine (DETA) in a continuous aqueous phase containing water, poly
(vinyl alcohol) (PVOH) and hydroxyl ethyl cellulose (HEC), via suspension
polymerization 2. These particles contain air voids which form by the reaction between
the UPR and DETA 2. The voids/vesicles are filled with water in the wet-state. The
water from the voids diffuses through the particle-wall upon drying, resulting in the air
voids/vesicles. These vesicles have the ability to scatter light (and thus impart opacity
and “hiding power” to a surface coating) due to the difference in the refractive index
between the particle shell and the air voids 1, 3.

MVPs are used in surface coating formulations as matting agents and can be used to
partially replace the primary but expensive opacifying pigment, TiO2, depending on
their particle size1, 3. Generally, two versions of MVPs are commercially available
which have average particle sizes of 5µm (used in silk/satin finish paints) and 25µm
(used in matt finish paints) 1.

Numerous processing- and formulation factors affect the properties and performance of
the MVPs. These include process parameters such as temperature, addition rate, mass
1
Chapter 1: Introduction and objectives
4-5
and heat transfer, agitation, stabilizer system, and reaction kinetics . The effect of the
properties of the UPR is also a major factor affecting the properties of MVPs such as
particle size and particle size distribution, viscosity, and degree of particle vesiculation
4-5
.
The UPR used in the synthesis of the MVPs is commercially produced by the
polycondensation reaction between maleic anhydride (MA), phthalic anhydride (PA)
and propylene glycol (PG) at 220°C. A reactive diluent (styrene) is added which reduce
the viscosity of the resins to assist processing, and copolymerize with the resin in order
to cure the material 6.

1.2 Objectives

Despite the commercial production of the particular UPR and the MVPs, little is
understood about the effect of the UPR molecular structure on the properties of the
MVPs (e.g. particle size or degree of vesiculation). It is known that the molecular
structure and inherent properties of the UPRs (e.g. molecular weight and degree of
chain branching) are greatly affected by the ratio of the monomers and process
7-8
parameters such as reaction temperature, agitation speed and heating rate . The
present study attempts to relate the UPR molecular structure with the MVP properties
by synthesizing MVPs from UPRs with varying monomer ratios and process
parameters. Due to the large number of factors affecting the properties of the UPRs,
statistical Design of Experiments (DoE) is employed to reduce the number of
experiments and isolate the most significant factors affecting the UPR and MVP
properties.

The major objectives of this study are as follows:

1.2.1 Investigation of the effects of varying the MA:PA mole ratio and process
parameters, e.g. heating rates, agitation speeds, etc. on the properties of the
UPR, via DoE.
1.2.2 Synthesis of MVPs using the different synthesized UPRs to investigate the
effects of varying the PA: MA ratio and process parameters on the properties of
the MVPs.
1.2.3 Determination of the molecular composition/ structure of the UPR and relate to
specific properties of MVPs.

2
Chapter 1: Introduction and objectives
1.3 References

1. Goldsbrough, K.; Simpson, L. A.; Tunstall, D. F. Progress in Organic Coatings

1982, 10, 35.
2. Engelbrecht, J. F.; De Wet-Roos, D.; Smit, A. C.; Cooray, B. Vesiculated
polymer particles. US 2006/0111474 A1, 2006.
3. Ritchie, P., Spindrift Pigmented Vesiculated Beads, In Surface Coatings. Raw
materials and their usage, The New South Wales University Press: Kensington,
NSW, Australia, 1993; pp 351 - 355.
4. Terblanche, J. C. The development of vesiculated beads. MScEng, University of
Stellenbosch, 2002.
5. Gous, K. Continuous processing of vesiculated beads. MScEng, University of
Stellenbosch, 2003.
6. Johnson, K. G.; Yang, L. S. Preparation, Properties and Applications of
Unsaturated Polyesters, In Modern Polyesters: Chemistry and Technology of
Polyesters and Copolyesters, Scheirs, J.; Long, T. E., Eds. John Wiley & Sons:
West Sussex, England, 2003; p 699.
7. Processing Unsaturated Polyesters based on Amoco Isophthalic acid; IP-43b;
Amoco Chemical Co.: 1989; pp 1 - 19.
8. How Ingredients Influence Unsaturated Polyester properties; IP-70; Amoco
Chemical Co.: pp 1 - 27.

3
 CHAPTER 2

HISTORICAL AND THEORETICAL BACKGROUND

2.1 Polycondensation: Polyesters

2.1.1 Introduction

Polyesters are a class of polymers that can be formulated to be hard or soft, brittle or
flexible depending on the structure and can therefore be used in various applications.
These applications include fibers ( e. g. polyethylene terephthalate (PET)) 1-3, coatings4-
5
, plasticizers 6, adhesives 5-6, polyurethane based resins 6, films 6, laminates 5, industrial
construction 5, insulation5, molding compounds 5, etc. Polyesters are used for these
applications because of their low cost 2-3, 5, 7, ease of process 7, ease of combination with
2-3, 7
reinforcements . In addition, these polymers cure rapidly without any volatile
2-3, 7-8
products , have excellent dimensional stability 7, have high impact resistance 5,
transparency 5 and good weathering resistance 5.

Polyesters do however have disadvantages including poor adhesion, high cure shrinkage
and inhibition of the curing reaction by air and certain fillers 5, as well as poor
resistance to hydrolysis by water 8.

History suggests that the first saturated polyesters were synthesized by Berzelius in
9-10
1847 from tartaric acid and glycerine , Berthelot in 1853 from glycerol and
camphoric acid 11, and Lorenzo in 1863 from ethylene glycol and succinic acid 12. UPRs
were first prepared by Vorlander in 1894 who studied glycol maleates 9.

6, 8, 13
In the 1920s, Wallace Carothers studied linear polyesters among which were
unsaturated derivatives from ethylene glycol and unsaturated carboxylic acids and
anhydrides such as MA and fumaric acid. The resins synthesized by Carothers were
14
either extremely viscous or solid . Carlton Ellis discovered that a diluent monomer
such as styrene (STY) can be added to the UPR to produce a low viscosity mixture that
will also copolymerize (under certain conditions) at rates much faster than that of the
14
homopolymerization . In 1947, during the Second World War, styrenated polyesters
were reinforced with glass fibers to yield high strength materials 8, 14.
4
Chapter 2: Historical and theoretical background
The first major commercial polyesters were PET discovered by Whinfield and Dickson
8, 15
and polyadipates (for polyurethane resins) 16.

Commercial polyester resins are classified as follows 17:

General-purpose orthophthalic resins – PA, MA and glycol resins.
Isophthalic resins – isophthalic acid (IPA), MA and glycol resins.
Dicyclopentadiene (DCPD)-capped resins.

These polyesters are generally produced via four main processes:

Fusion process: In the fusion process, monomers are polymerized at very high
temperatures. The main advantage is that the process is robust but the process is
disadvantageous because great difficulties are experienced due to slow-down of
polycondensation reaction in the final reaction phase due to increase in viscosity of
reaction mixture 18. In addition, the use of esterification catalysts is not possible because
18
the reaction is carried out at very high temperatures . Large excesses of acidic or
alcoholic components are also not possible, since these components cannot be removed
from the reaction mass in an economic manner 18.

Azeotropic solvent process: The azeotropic solvent process involves the addition
of an aromatic hydrocarbon, e.g. xylene (concentration = 3 – 6 wt. %), to the reaction
mixture during polymerization to increase the rate of water removal by increasing the
vapor pressure of the volatile compounds 18. It also results in a decrease in the solubility
18
of water in the system and a reduction of the overall viscosity to some extent .
Aromatic hydrocarbons are water-insoluble, so that separation of the water phase from
the organic phase occurs very rapidly and the organic phase can be recycled
18
immediately to the reactor or to the distillation element above the reactor . The main
advantages of the azeotropic solvent process include reducing the polyesterification
time by 30-50 % and producing polyesters with improved color, and in some cases
narrower molecular weight distribution 18.

Inert gas transport process: In the inert gas transport process, inert gas (usually
nitrogen or carbon dioxide) is sparged through the reaction mass during the fusion
19-21
process to aid water removal . Reduction in polyesterification is comparable to that
18
of the azeotropic solvent process . The main disadvantage of the inert gas transport
process is that other volatile components e.g. PA are also removed from the reaction

5
Chapter 2: Historical and theoretical background
mass during this process 18.

Vacuum process: The vacuum process involves the dehydration of the reaction
18
mass by reducing the interior pressure of the reactor . The main advantage of the
vacuum process is the fact that water, the azeotropic component, and other volatile
18
compounds are converted into the gas phase more readily . The vacuum process also
results in an increase in the rate of acid-value reduction18. The rate at which the acid
value decreases is proportional to the rate at which the reaction mass is dehydrated. The
main disadvantage of vacuum polyesterification is the formation of additional molecular
structures, which are not present when polyesterification is carried out at atmospheric
pressure.

2.1.2 Polycondensation

Polyesters are condensation polymers with ester/ carboxylate repeating units in the main
chain. Figure 2.1 shows the generalized reaction scheme of esterification, where N: is
nucleophile eg. OR‟, where R‟ is an alkyl group. The rate of reaction is dependent on R,
X and N and whether or not the reaction is catalyzed 1, 22-25.

O O: O

+ N: R N
R N
+ X:
R X
X

Figure 2.1: Generalized reaction scheme of esterification (R = alkyl).

Polyesters can be synthesized via a variety of pathways, these include:

1. Self-condensation of ω-hydroxy acids 8
O
O
R O O R O
R + HO R R O R
HO OH
OH O

Figure 2.2: Reaction scheme of polyesters via self-condensation of ω-hydroxy acids

(R = alkyl).

6
Chapter 2: Historical and theoretical background
2. Condensation of a diol with a dicarboxylic acid 8
O O

R OH
O OH O R1 O O
R OH R
HO
R
OH
+ HO R1 O + O
HO
OH
+ H2O

Figure 2.3: Reaction scheme of polyesters via condensation of diols with diacids (R,
R1 = alkyl groups).

3. Ester exchange 8
O O O
O O

R1 + R2 OH
R2 OH + R1 OH
O R O O
O R O
HO OH
R1

Figure 2.4: Reaction scheme of polyesters via ester exchange with a diol (R, R1, R2
= alkyl groups).

4. Ring opening of a lactone, e.g. of e-caprolactone with dihydroxy or

trihydroxy initiators 8:
O
O
R
O

O R

Figure 2.5: Reaction scheme of polyesters via ring-opening of lactones (R = alkyl

group).

5. Alcoholysis of the acid chloride of a dicarboxylic acid with a polyhydroxy

alcohol 8
O O O O

O + R1
HO R O
R1
OH
Cl R HO OH

Cl

Figure 2.6: Reaction scheme of polyesters via alcoholysis (R, R1 = alkyl groups).

6
Chapter 2: Historical and theoretical background
UPRs, which are the focus of this study, are polymers obtained by the polycondensation
reaction of a difunctional unsaturated acid or anhydride, e.g. MA and a difunctional
saturated acid or anhydride, e.g. PA with a diol, e.g. propylene glycol (PG) 14,26-28 at 150
- 280 °C, predominantly at atmospheric pressure due to the disadvantages of reduced
pressure reactions (e.g. vacuum process) (see Section 2.1.1). Water is the main by-
product of the condensation reaction and is removed from the reaction to drive it to
14, 26
completion . Figure 2.7 shows the formation of the UPR, poly (propylene fumarate
phthalate).
O
OH

+ O + O 150 - 280 deg.C

O O
HO CH3
O
Propylene Glycol Maleic Anhydride
Phthalic Anhydride

O COOH

O O
O O
H2O
HO CH3
+
O

CH3
poly (propylene fumarate phthalate)
Figure 2.7: Formation of the UPR, poly (propylene fumarate phthalate).

As mentioned previously in Section 2.1.1, UPRs are usually diluted with a reactive
2, 18
diluent, e.g. styrene (STY) or methyl methacrylate (MMA) . The STY or MMA
reduces the viscosity of the resins to assist processing, and copolymerizes with the resin
in order to cure the material 17.
UPRs are produced by the formation of monoesters or ring-opening of the anhydrides in
the first reaction step which reacts to form di-esters after the temperature is increased.
These diesters react further through polycondensation to produce the final polyester and
water as by-product. The reaction mechanisms throughout polyesterification depend on
the reaction temperature, monomer ratio, catalyst system, and processing variables such
27
as water removal and agitation . Detailed description of the reactions involved in the
synthesis of UPRs, i.e. ring-opening or monoester formation, polyesterification and side
reactions, are discussed in the Section 2.1.3.

7
Chapter 2: Historical and theoretical background
2.1.3 Reactions

2.1.3.1 Ring-opening (monoester formation)

Monoesters are formed by the reaction between the anhydride (A) and the glycol (G) to
form acid and alcohol end groups (COOH and OH) and an ester bridge with no water
formation 27-28. Figure 2.8 shows the monoester formation during polyesterification.
HO OH
R
O R
O
+ HO OH
O O
O
O

Figure 2.8: Reaction scheme of monoester formation (R = alkyl group).

The monoester formation reaction normally occurs in the temperature range of

60 – 130 °C. The kinetics of this reaction suggests that it is a second order reaction and
irreversible. The formation of monoesters depends on the reactivities of the glycols
towards the anhydrides, for example, ethylene glycol (EG) and PG are assumed to have
the same reactivity towards ring-opening, but anhydrides have different reactivities. The
rates at which MA react with EG and PG during ring-opening are equal, but different
for when they react with PA 29. In most cases, monoester formation is followed by the
main polyesterifcation reaction between the formed esters and acids and alcohols, with
the elimination of water as by-product 18.

2.1.3.2 Polyesterification

The esterification reaction proceeds by reaction of a hydroxyl end-group of one

29
molecule with a carboxylic acid end group to form an ester and water . Figure 2.9
illustrates the polyesterification which occurs in the temperature range of 160 to 220
°C28. A high degree of polyesterification is obtained upon efficient removal of water.
The removal of water is, however, hampered due to an increase of its solubility in
organic substances with increasing temperature 18, 28.
HO OH HO

n O R O R + (n - 1) H2O

O OH
O

O O
n
Figure 2.9: Reaction scheme of polyesterification (R = alkyl group).

8
Chapter 2: Historical and theoretical background
A number of side-reactions take place during the polyesterification stage. These
reactions are discussed in Section 2.1.3.3.

2.1.3.3 Side reactions

2.1.3.3.1 Isomerization

Side reactions in the later stages of the polyesterification reaction prevent high degrees
30 30
of polymerization . These side reactions are related to a cis-trans isomerization .
Figure 2.10 illustrates the cis-trans isomerization reaction which involves the rapid
isomerization of maleate (cis) to fumarate (trans) (from the MA component) at a
reaction temperature of approximately 200 °C17, 31.
H H O

H OH

O OH

OH O HO H

Cis Trans

Figure 2.10: Reaction scheme of cis-trans isomerization (maleate to fumarate).

The kinetics suggest that the isomerization reaction is acid catalyzed and obeys a
second-order rate law with respect to the acid. The activation energy of isomerization is
63.2 kJ/mol 32.

Furthermore, it is known that more than 90 % of the cis maleate (highly strained
structure) isomerizes to the trans-fumarate isomer (less strained, lower energy level)7, 14.
The fumarate isomer is also much more (about 20 times) reactive to the crosslinking
7, 14, 36-38
reaction with styrene than the maleate isomer . The increased reactivity of the
17
fumarate isomer is due to lesser “crowding” of chains in the trans-configuration ,
33-34 35
which leads to products with improved mechanical , (improved tensile strength
36 33-34
greater hardness, high moduli or stiffness, lower elongation) and thermal
properties (higher heat distortion temperature, reduced gel and propagation time and
higher exotherms) 36 and better chemical resistance 35.

9
Chapter 2: Historical and theoretical background
Table 2.1 shows the effect of the glycol type on the percentage isomerization of the
maleate groups UPRs. It can be seen from Table 2.1 that the structure of the glycol
starting material has a marked influence on the maleate-fumarate isomerization 30, 34, i.e.
the reaction is favored by secondary glycols e.g. PG 14. Primary glycols such as EG are
14
less favorable . It has been shown experimentally that isomerization decreased by
around 10-15 % when replacing ethylene glycol (EG) with 1,3-propylene glycol 37, due
to the steric hindrance between hydroxyl groups, the glycol chain-length and the
“intramolecular spatial interaction between ester groups” which are closely arranged 38.

Isomerization is also affected by reaction temperature and time, i.e. greater

14, 34
isomerization is favored by high temperatures and long reaction times . Other
39
factors influencing isomerization include the nature of the starting materials , the
33, 40-41
concentration of catalyst (e.g. piperidine and morpholine) and stereochemical
33
agents . The addition of aromatic acids (e.g. phthalic acid) also has a favorable effect
on the degree of isomerization33-34.

Table 2.1: Isomerization vs glycol type# 32.

Glycol Type Isomerization, %
1,2-propylene Secondary 96
Ethylene Primary 64
Diethylene Primary 53
1,6-Hexamethylene Primary 36
#
Reaction Temperature constant = 180 ˚C

Table 2.2 shows the effect of reaction temperature on the percentage of isomerization of
the maleate groups of UPRs. It can be seen from Table 2.2 that the percentage of
isomerization increases with an increase in the reaction temperature as in the case of
poly (propylene glycol maleate) at temperatures ranging from 105-180 ˚C 32.

Table 2.2: Isomerization vs reaction temperature 32.

Reaction temperature (˚C) Isomerization to fumerate (%)
105 34
125 56
140 75
180 96

10
Chapter 2: Historical and theoretical background
2.1.3.3.2 Double bond saturation (Ordelt reaction)

The Ordelt reaction is the main side reaction in the polyesterification reaction. The
Ordelt reaction involves the double bonds of the maleate or fumarate component
becoming saturated by the addition of hydroxyl groups via electrophilic addition
42-45
(hydroxyl-alkoxy addition) from the glycol. The kinetics of this electrophilic
addition were studied by Fradet and Marechal (1982) 46. It was shown that the reaction
is acid catalyzed and follows a first order rate law with respect to the acid and the
alcohol 46.

28
The saturation reaction occurs during the isomerization of maleate to fumarate and
29, 34
may reach 10-20 % . Paci et. al. studied this saturation reaction of UPRs from
47
diethylene glycol (DEG), PG, MA and PA . The saturation reaction effectively leads
to branched structures (side chains) and deviation from the stoichiometry of the
reactants and is known to affect the mechanical properties of the final cured polyester
28, 34
resin . The side chains may consist of one diol molecule or more than one diol
28
molecule or long ester chains . Theoretically the chain end number is 2 for linear
chains, although in practice it is 2 plus the sum of chain ends due to short and long
chain branches 28.

The probability of monoester saturation by diol is greater than saturation by monoester.

The greater probability of saturation by diol is due to the fact that hydroxyl group of the
28
diol monomers is more reactive than that of the monoester . It has been shown that a
polyesterification system only shows Ordelt saturation when the diol consists of more
48
than two carbons . Figures 2.11 and 2.12 show the Ordelt saturation reactions via
monoester and diol (short chain branches), and monoester and monoester (long chain
branches), respectively.

O
O
HO R

HO
O
R
OH + HO
R
OH
O OH

O O
O R

OH
Figure 2.11: Reaction scheme of the Ordelt saturation of monoester by diol (short
chain branches) 28 (R = alkyl group).

11
Chapter 2: Historical and theoretical background
O O

O OH O OH
HO R
HO R +
O O

HO O

O O

R
O O

O OH
HO R

Figure 2.12: Reaction scheme of the Ordelt saturation of monoester by monoester

(long chain branches) 28 (R = alkyl group).

2.1.3.3.3 Transesterification

Transesterification occurs when the polyester chains undergo alcoholysis or acidolysis

by the hydroxyl or carboxyl groups of the diols and acids or the oligomers at high
temperature. Transesterification results in a statistical distribution of repeating units and
28
functional end-groups . Figure 2.13 illustrates the transesterification side reaction
during polyesterification.

12
Chapter 2: Historical and theoretical background

O H
O H
O R
HO O R
+ n
O
HO O
m O
O

O H

O R
O H

HO O R +
p HO O
O q
O
O
HO

O
O

or O O
O
+ HO O O H
HO O R
O s
r
O

Figure 2.13: Reaction scheme of transesterification side reaction during

polyesterification (R = alkyl group).

2.1.3.3.4 α-Diol dehydration

Higher molar mass diols may be formed during α-diol dehydration such as DEG and
49
dipropylene glycol from 1,2-ethanediol and 1,2-propanediol . The α-diol dehydration
50
reaction also results in low boiling point compounds such as THF . The structure of
the diol is effectively modified and the reactant ratio of diol to acid changes, although
the linear structure is maintained 28.

2.1.3.3.5 Formation of cyclic structures

The formation of cyclic structures during polyesterification is heavily affected by the

flexibility of the main chains. These structures can be formed in two major ways:
The structures can be formed at the beginning of the polycondensation reaction when
the functional end-groups of the same chain react (ring closure)51.
The carboxylic end-group can react with the ester component of the same chain (back-
biting)51.

13
Chapter 2: Historical and theoretical background
2.1.4 Curing of unsaturated polyester-styrene solution

The crosslinking reaction of UPRs with STY has been extensively studied. Crosslinking
takes place between the double bond of the unsaturated acid component (maleate or
fumarate) in the polyester chains and the double bond of STY (reactive diluent) through
a free-radical reaction 2, 8-9, 14, 36, 52-55.

Molecularly, the UPR-STY solution may be described as “coiled polyester chains in

styrene solution”. The copolymerization reaction commences by the decomposition of
the initiator molecules to produce free radicals in the system. The UPR molecules are
linked to the styrene molecules by the free radicals to create long chain molecules,
which forms spherical-type structures as a result of the intra-molecular cross-linking
between the maleate / fumarate groups of the pendant polyester chains 52.

7, 54
The crosslinking reaction normally takes place at room temperature with a
combination of catalyst/initiator usually methyl-ethyl ketone peroxide (MEKP) and
accelerator, usually cobalt napthenate (4 % solution in styrene) 7. During the accelerated
crosslinking reaction, vinylene, STY-STY and polyester-vinylene-polyester-vinylene
52
bonds can be formed . Linear polystyrene chains form at the unsaturation site of the
polyester chain, which extends to the unsaturation site of another polyester chain,
essentially crosslinking the polyester chains 56. The length of the polystyrene chains has
been estimated to be 2 – 3 monomer units long 57-58 and the styrene to double bond ratio
in general purpose resins are 2:3 14.

The degree of crosslinking or crosslink density (controlled by the concentration of

5, 36
unsaturated acid –or glycol) and the length of these polystyrene chains (can be
5
controlled by the type and concentration of the monomer, initiators, and catalyst)
affect the properties of the cured resin59.
36
Table 2.3 illustrates the calculation of the level of unsaturation in UPRs . It is
generally accepted that higher levels of unsaturation results in a faster crosslinking
reaction, although a decrease in the unsaturation level leads to increase in the impact
resistance 36.

14
Chapter 2: Historical and theoretical background
Table 2.3: Level of unsaturation of a UPR.
Reactant mol Mass (kg)
IPA 3.0 0.498
MA 2.0 0.196
PG 5.5 0.418
Less water -8.0 -0.144
Yield 2.5 0.968
Unsaturation = 2.0 / 0.968 = 2.07 mol/kg

UPRs can be crosslinked with styrene via suspension polymerizations resulting in a

60-61
dispersion of micro-particles or beads in water . Depending on the production
process, these particles may contain single or multiple air voids or vesicles, which are
62
used as organic opacifying particles . These particles can be used as a partial
62
replacement for pigments, extenders and additives in surface coatings . Suspension
polymerization theory and the production of the vesiculated micro-particles are
discussed in Section 2.2.

2.2 Suspension polymerization

2.2.1 Introduction

63
Suspension polymerization, also called pearl or bead polymerization , refers to a
system in which droplets of monomer (s) are suspended in a medium (usually water) by
means of controlled agitation and a suspension stabilizer (suspending agent or
protective colloid or dispersant) 63. These materials are usually water-soluble polymers,
e.g. PVOH 66-68 or cellulose derivatives 63-64 and inorganic salts e.g. Talc.

Suspension polymerization can be considered a bulk polymerization in small monomer

droplets, typically 10-1000 µm in diameter 63, 65. Monomer-soluble initiators are used 63,
65-66
and the reaction is carried out above room temperature but below the boiling point
65-66
of the water phase . Under these conditions, the monomer droplets are converted to
66
polymer particles . Suspension polymerization obeys the kinetics of solution
polymerization if the particles are not too small and polymerization is initiated in the
monomer phase 63, 67. The kinetics becomes more like that of emulsion polymerization

15
Chapter 2: Historical and theoretical background
if the particles become too small in size, so that polymerization is only initiated by a
few free radicals in the particles 63.
The size of the suspension particles, typically 50-200 µm in diameter, depends on the
65
monomer type, suspending agent, and agitation in the reactor . Micro-suspension
processes can yield particles in a range of 20-30 µm 65.

The applications of products made by suspension polymerization includes,

68 65, 68
chromatographic separation media , ion-exchange resins supports for enzyme
68
immobilization and poly(vinyl chloride) (PVC) and poly (vinylidene chloride-co-
vinyl chloride) for molding applications 65.

2.2.2 History

69-71
Hofman and Delbruck developed the first suspension polymerization in 1909 . In
1931, Bauer and Lauth performed the first suspension polymerization which led to the
formation of beads from acrylic monomers 72. Hohenstein and Mark, Trommsdorff and
73
coworkers discussed the basic aspects of suspension polymerization in early papers .
The first commercial resin obtained by suspension polymerization was poly (vinyl
chloroacetate) 74.

The major challenge experienced by Voss et. al. was coalescence of the droplets which
led to large polymeric masses 76. Cooling of these polymeric masses to remove the heat
of polymerization was difficult. Water-soluble polymers were used to overcome this
problem 75. Other dispersants e.g. talc, barium sulfate, magnesium hydroxide and other
materials were discovered in later years 76.

2.2.3 Process of suspension polymerization

2.2.3.1 General

Suspension polymerization involves the mixing of two immiscible phases by

mechanical agitation. Droplets are formed and their size is determined by the balance
68, 77
between breakup and coalescence between the droplets . Suspension droplets are
not thermodynamically stable, as these droplets do not remain dispersed for long time
periods like in emulsion systems 65. Suspension droplets (50 – 200 µm) tend to coalesce

16
Chapter 2: Historical and theoretical background
65
rapidly when agitation of the monomer phase is discontinued . The process of
coalescence is controlled by the balance of agitation and stabilizer system 65. The size of
the droplets are affected to a large degree by the properties of the continuous (water)
68, 78 68, 78-79
and dispersed (oil/ monomer) phases like interfacial tension , density 78 and
viscosity 78-79, type and concentration of stabilizer 68, 77-79, monomer-phase hold-up 68, 77,
68, 77-79 66
impeller type and agitation speed , apparatus design and the polymerization
kinetics 68.

80 81 82
Arshady and Ledwith , Hopff et. al. , Kavarov and Babanov , Mersmann and
83 84
Grossman , and Sculles reported the relationship between particle size and the
factors mentioned in the following equation:
Dv R vd
d k (2.1)
Ds N v m Cs
Where d = average particle size, k = parameters such as apparatus design, type of stirrer,
self stabilization, etc., Dv = diameter of vessel, Ds = diameter of stirrer; R = volume
ratio of the droplet phase to suspension medium; N = stirring speed (or power of
mixing); vd = viscosity of the droplet phase; vm = viscosity of the suspension medium; ε
= interfacial tension between the two immiscible phases; and Cs = stabilizer
concentration.

The agitation speed is the most convenient way to control the particle size among the
66
parameters in Equation 2.1 . Generally, the size of suspension polymerized particles
ranges between 10 µm and 5 mm, although much smaller particles, which are
unintended, are also present. Suspension based particles are known to display broad or
bimodal particle size distributions 68.

2.2.3.2 Droplet formation

The formation and the maintenance of individual monomer droplets in the continuous
phase throughout the polymerization process is the most important feature in suspension
85
polymerization . The dispersion of the monomer droplets is usually achieved by
mechanical agitation 85. The main problems in suspension polymerization are reduction
of droplet coalescence and the formation of suspension droplets of equal sizes 65.

17
Chapter 2: Historical and theoretical background
65
In order to form droplets, the monomer must be water-insoluble . Typically these
monomers have a lower surface tension than water 65. Dispersion of these monomers in
the continuous phase without surfactants or stabilizers results in an unstable dispersion
as a result of the break-up-coalesce process of the monomer droplets 65. The monomer-
water system separates into the individual phases once agitation is discontinued 65.
The coalescence of monomer droplets or agglomeration of particles varies during the
65
reaction . The dispersion of monomer droplets is stable in the initial stages of
65
polymerization . At conversion levels of about 20 %, agglomeration of the particles
can occur if they collide, as the surfaces of the droplets (which contain polymer and
monomer) become “tacky”/ sticky 65. The droplets become tacky due to the fact that the
monomer is soluble in the polymer or the polymer is swollen by the monomer (e.g. PVC
65
in vinyl chloride) . As polymerization continues, the particles become less “tacky”,
resulting in reduced particle agglomeration 65.

2.2.3.3 Droplet/ particle stabilization

Reduction of the surface tension of the droplets and minimization of the force at which
the droplets collide can prevent the large coagulation during the “tacky” stage 64, 85. The
addition a small of amount droplet stabilizer and choosing the correct apparatus design
85
, may prevent the coagulation. The droplet stabilizer is usually a water-insoluble
inorganic salt (e.g. talc, bentonite, calcium sulfate, etc.) or an organic polymer (e.g.
64, 85 63-64, 85 64
poly(vinyl pyrolidone) , PVOH , salts of acrylic acid polymers , cellulose
63-64 64
ethers and natural gums . The inorganic salt reduces coagulation of the droplets
by forming a thin layer around the droplets 85. The organic polymer must be insoluble in
the monomer phase and have low solubility in the water phase 85. Organic polymers can
be removed more easily from the polymer bead particles, and are therefore preferred to
inorganic salts 85.

2.2.3.4 Particle size

The process of coalescence – breakup/ dispersion in the droplets dictates the final
particle size 65, 68, 77-78. The breakage of the droplets occurs as a result of the high-shear
stresses at the impeller blades or pressure fluctuations and turbulence at the surface of
78 78
the droplets . The “turbulent flow field” dictates the coalescence process . The

18
Chapter 2: Historical and theoretical background
coalescence process is assumed to be negligible in dilute dispersions with very high
stabilizer concentrations 86-87.

Figure 2.14 shows the breakage and coalescence mechanisms involved in suspension
88
polymerization . It can be seen from Figure 2.14(a) when drop breakage occurs as a
89
result of “viscous shear forces”, the droplet becomes elongated to a threadlike form .
The deformed droplet subsequently splits into two drops of equal size and small
droplets due to the breakage of the fluid thread between the two fluid lumps 78.

Figure 2.14(b) shows that erosive breakage of the droplets can also occur due to
pressure and relative velocity fluctuations 78.

78
Coalescence of droplets can occur by two different mechanisms . Coalescence can
occur by collision of two droplets with the continuous phase being trapped initially
between the drops. Figure 2.14(c) shows that the continuous phase is eventually drained
from between the droplets by the attractive forces, resulting in complete drop
coalescence 86, 90.

Figure 2.14(d) shows that immediate coalescence can also occur when the turbulent
energy of the collision between droplets is greater than that of the surface energy of the
droplets 91.

19
Chapter 2: Historical and theoretical background

Figure 2.14: Illustrative scheme of breakage and coalescence mechanisms involved

in suspension polymerization 88.

2.2.3.5 Particle morphology

Suspension polymer particles or beads or pearls, may be hard or soft in nature. The
relative hardness of the beads depend on the monomer composition and whether
miscible diluents are present 64. The morphology of suspension polymerized particles is
largely affected by the coalescence and dispersion of droplets before and during
polymerization, the change in specific gravity between the monomer and the polymer
phases, and the solubility of the monomer in the polymer 65.

Suspension polymer particles have smooth surfaces and relatively uniform texture (non-
porous) when the polymer is soluble or swellable in its monomer (e.g. poly (styrene)
64, 66
dissolve in its own monomer and smooth and translucent beads are produced) . On
the other hand, porous beads with a rough surface are expected when the polymer is
insoluble or not swellable in its monomer (e.g. PVC is insoluble in vinyl chloride) 64, 66.

Suspension polymerized particles, or, “beads” can be made porous by the addition of
inert diluents or porogens, to the monomer phase during polymerization. The porogen
20
Chapter 2: Historical and theoretical background
diluents are generally alkanes, aromatics (e.g. toluene), alcohols and carboxylic acids 64.
64, 77
The porogen is extracted at the end of polymerization . The size of the pores
depends on the mutual solubility of monomer and diluents, chemical structure of the
diluents, chemical nature of the monomer phase, type of stabilizer, and the
concentration of the crosslinking monomer 64.

67
Particles may also collapse inward (e.g. PVC particles ) and can be attributed to “the
change in the specific gravity between the monomer and the polymer” 65. If the density
of the polymer is higher than that of the monomer and the shape of the particle is not
maintained by crosslinking, then less volume must be occupied by the final polymer
particle relative to the original monomer droplet 65. The collapse of the particle inwards
relieves the strain induced by the difference in specific gravity between the polymer and
monomer, when the surface area of the particle is fixed upon surface skin formation 65.

2.2.4 Control of particle size in suspension polymerization

2.2.4.1 Stabilizer

The type and concentration of the stabilizer plays a vital role in the particle formation
68, 78, 92
stage of suspension polymerization . The stabilizer must prevent aggregation or
coagulation of the droplet-particle by maintaining complete coverage of the droplet-
particle surface by physical adsorption and/or adsorption by anchoring 69, 84, 96.

As mentioned previously, the stabilizers used in suspension polymerization are usually

68
either water-soluble polymeric materials or inorganic materials . Water-soluble
64, 68
stabilizers aids droplet break-up by dropping the interfacial tension . One of the
most commonly used water-soluble stabilizers is partially hydrolyzed poly (vinyl
acetate) (PVA) or PVOH. The best grade of PVOH for use as a suspension stabilizer is
one with a high degree of hydrolysis (80-90 %) and molecular weight in excess of
93
70000 g/mol . The mentioned PVOH grade provides a thicker and stronger layer
68
around the monomer droplet/particle and the tendency to be desorbed is less . The
polymer particles coalesce and agglomerate forming lumps when PVOH grades with a
lower degree of hydrolysis (less than 80 %) or lower molecular weight is used 68. On the
other hand, unstable dispersions are obtained when a PVOH grade with a very high
degree of hydrolysis (>90 %) is used 68. The ability of PVOH to stabilize suspensions is

21
Chapter 2: Historical and theoretical background
also influenced by the stereochemistry, degree of branching and distribution of acetate
94
and hydroxyl groups . Generally an increase in the concentration of the stabilizer,
64, 68 68
results in a more stable dispersion due to steric and Marangoni effects . An
increase in the PVOH concentration also leads to smaller average droplet/ particle size
with a narrow distribution 68.

2.2.4.2 Agitation

Droplets are formed when two immiscible phases are mixed 68. The size of the droplets
68, 77
is affected by the agitation speed . The role of agitation is to balance coalescence/
dispersion of the droplets/particles, suspension of the droplets/ particles and the effects
of heat transfer65. The rate at which droplets are broken-up is enhanced by an increase in
the agitation speed, and therefore the formation of smaller droplets is therefore
favored68, 95. However, there is a U-shaped relationship between the average droplet size
and agitation speed as the droplet size may increase at very high agitation speeds due to
68
an increase in the rate at which the droplets coalesce . The increased droplet
coalescence can be explained by the large surface area of these droplets and the reduced
effectiveness of the stabilizer molecules at the interface 68 at high agitation speeds.

2.2.4.3 Monomer hold-up

The kinetics of the reaction are affected by the solubility of the monomer in the water
phase. When the monomer is insoluble in water (e.g. styrene) the kinetics of suspension
polymerization is similar to that of bulk polymerization. In cases where the monomer is
moderately to highly soluble in water, the kinetics deviate from that of bulk
polymerization as some monomer molecules reside in the water phase, which means
that monomer transfers from the water phase to replace monomer in the polymerizing
droplets 68.

2.2.4.4 Temperature

Monomer and oil phase properties are affected by the variation in reaction
temperature68. It is known that the rate of polymerization can be increased by two-to-
three fold by increasing the reaction temperature by 10 °C 96.

22
Chapter 2: Historical and theoretical background
PVOH shows a decrease in its ability to stabilize droplets with increasing temperature97.
An increase in the agitation speed, monomer hold-up and temperature leads to an
95
increase in the ratio of final particle/droplet size D32,s/ D32,d . On the other hand, an
increase in the stabilizer and initiator concentrations leads to a decrease in the final
particle/ drop size ratio 95.

2.2.5 Bead-suspension polymerization kinetics

Suspension polymerizations are considered bulk polymerizations in the monomer

droplets when the “initiators and chain transfer agents are soluble only in the monomer
phase”. The particle size and type of stabilizer used in the suspension polymerization of
beads do not greatly influence the rate of the bead-polymerization. The following
reactions occur in the monomer phase:

kd
Initiation I 2R (2.2)

ki
R M R M (2.3)

kp
Propagation Mn M Mn 1 (2.4)

k tr,m
Chain transfer (to monomer) Mn M Pn M (2.5)

Termination
k tc
(Coupling) Mm Mn Pm n (2.6)
k td
(Disproportionation) Pm Pn (2.7)

Where I is the initiator, R* is the radical generated by dissociation of the initiator (with
rate constant kd), M is the monomer concentration, Mn* is growing polymer chain
having n repeat units, Pn is the polymer with n repeat units. The rate constants kd, ki, kp,
ktr,m, ktc, and ktd refer to the rate equations for Eqs. 1-6 65.

23
Chapter 2: Historical and theoretical background
2.3 Suspension polymerization – Vesiculated particles

2.3.1 Introduction

92
Titanium dioxide (TiO2) is the primary opacifying pigment in the coatings industry .
The amount of opacity it imparts to a coating depends on its crystal and particle size,
92
relative refractive index, and the degree of which it is dispersed in the dry film . The
cost of the TiO2 pigment have soared in the past few decades due to increased
consumption in a variety of applications (e.g. paints), therefore alternatives are
attractive 92. The opacity of a coating can also be increased by incorporating microvoids
into a coating as they also scatter light. Willkie produced a microvoid coating in 1923
98
. As in the case of TiO2, the size, concentration, etc. of microvoids determines its
light-scattering efficiency. Furthermore, combinations of microvoids and the TiO2 are
62
effective in achieving and/ or increasing opacity of a coating . Microvoids increase
opacity due to the air contained in the matrix which increases the refractive index
difference between the pigment and the matrix 92.

The major developments in the field of air-void technology include the Spindrift process
which is used by Dulux Australia to commercially produce multi-vesiculated particles
(MVPs) and single vesiculated particles (Ropaque opaque polymer) which are
manufactured by Rohm & Haas Company. The processes involved in the production of
these particles are discussed in Section 2.3.2.

2.3.2 Single vesicle particles

Single-void polymer micro-particles were successfully developed by the Rohm & Haas
Company in the early 1980s. These micro-particles, commercially known as Ropaque
opaque polymer, are produced via emulsion polymerization where water is encapsulated
in the core of the polystyrene particles. Figure 2.15 shows the Ropaque particle with a
polymer shell and the air void or vesicle. When these particles are dried, the water in the
core irreversibly evaporates and is replaced with air. The resultant opaque particles have
the ability to scatter light, which is due to the refractive index difference between the
polymer shell and the air filled void. Table 2.4 shows the typical properties of
Ropaque61.

24
Chapter 2: Historical and theoretical background

Figure 2.15: Schematic diagram of the Ropaque opaque polymer particle61.

Table 2.4: Typical properties of Ropaque 61.

Property Value

Solids content (%) 37 - 38

Dry relative density 0.741
Wet relative density 1.038
Average particle size (µm) 0.4
Viscosity (cPs) 100

2.3.3 Multi-vesiculated particles (MVPs)

In 1976, Kershaw filed a patent describing the process of producing vesiculated

99
polymer beads with an average particle size of 10 µm . The vesicles or small air
pockets/ bubbles encapsulated in the beads were approximately 0.7 µm in diameter
which displaced 50 % of the total bead volume. A double emulsion technique, water-in-
oil-in-water emulsion, was used to prepare the beads. Water was dispersed in a mixture
of UPR and STY, and this dispersion was added to an aqueous solution of polymeric
stabilizers, eventually forming the water-in-oil-in water emulsion. The organic phase
droplets were cured using a redox-free radical system to form the crosslinked polymer
particles or beads. Interfacial tensions, phase volumes and agitation were optimized to
control the bead and vesicle sizes 99.

25
Chapter 2: Historical and theoretical background
101
Kershaw then used an aqueous TiO2 dispersion in the first stage of the process . In
this way, the TiO2 pigment is introduced into the beads, which resulted in an increased
scattering efficiency of this pigment. The beads can also be incorporated into emulsion
paints without sacrificing the integrity of the paint film due to the presence of air. These
92, 100
pigmented vesiculated beads, marketed under the tradename Spindrift , contain
TiO2 pigment and vesicles (filled with water) encapsulated in crosslinked UPR-STY
matrix. Figures 2.16.1 and 2.16.2 show the scanning electron micrographs of Spindrift
pigmented vesiculated particles and the microvoids or vesicles (indicated by the holes or
indentations), respectively 100. As with Ropaque opaque polymer, when the paint film is
dried, water in the vesicles evaporates by diffusion through the matrix wall, due to the
difference in the vapor pressure between that inside the vesicle and the surface of the
paint film. When the difference in vapor pressure diminishes, the process is
92
discontinued . During this process, micro-void cells are formed which can be
monitored by an increase in the opacity of the paint film. Furthermore, no water diffuses
back into the vesicles of the beads when water is applied to the paint film 92.

Figure 2.16.1: Scanning electron Figure 2.16.2: Scanning electron

micrograph of Spindrift particles 100. micrograph of a cross-section of a
Spindrift particle 100.

26
Chapter 2: Historical and theoretical background

The vesicles of the beads contain a high partial pressure of water, although it has become
common practice to assume that the vesicles are filled with air when the paint is dry.
92
Approximately 50 % of the dry bead volume is occupied by air . An increase in the
volume of the voids can increase light-scattering power, but leads to a decrease in the
mechanical strength of the beads 92.

The average particle size of beads can be controlled and two versions are commercially
available based on the size distributions, namely beads with average particle size of 5 μm
(suitable for paints with silk/ satin finish) and 25 μm (suitable for paints with matt finish)92.

In 2006, Engelbrecht et. al. filed a patent for the production of MVPs via a single-
emulsification process in the presence of a hydrophobic monomer, for the beneficial
replacement of TiO2, extenders and additives 101. Vesiculated beads or MVPs are currently
being manufactured and used in coatings by Plascon South Africa.

The process steps for the manufacture of the MVPs are as follows:
Pigment particles are in a UPR-STY solution at high speed,
The UPR-STY mixture (with dispersed pigment) is dissolved in a suitable monomer
(e.g. STY) in the presence of a water-soluble base, e.g. diethylene triamine (DETA),
The UPR-STY-DETA mixture is then added to a mixture of water, stabilizer, e.g.
PVOH and thickener, e.g. hydroxyl ethyl cellulose (HEC), forming a stable emulsion
of oil droplets in water,
A hydrophobic monomer, e.g. laurel methacrylate (LMA) is added to the emulsion,
after which UPR is polymerized with the co-polymerizable monomer (e.g. STY),
producing a dispersion of opaque cross-linked vesiculated particles in water.

The proposed mechanisms for the vesicle formation and the crosslink-reaction between the
UPR and STY are as follows:

4
Chapter 2: Historical and theoretical background

2.3.3.1 Vesicle formation

The addition of the polyamine, DETA, to the mixture of UPR, STY and LMA, results in
the neutralization of the main and side-chain carboxylic groups of the polyester chains
(Equation 2.8).

R-COOH + R‟-NH2 → R-COO- + R‟-NH3+ (2.8)

The neutralization reaction in Equation 2.8 causes the negatively charged carboxylic
groups of multiple polyester chains to orientate themselves toward the positively charged
amine groups of DETA, forming stable macro-molecular structures, also referred to as
micelles. Figure 2.17 shows that the formed micelles possess hydrophilic character due to
the neutralized groups in the centre, i.e. when the organic phase is added to the aqueous
phase (comprising of water, HEC, PVOH and a small amount of DETA), the organic phase
becomes “soluble” in the aqueous phase. Water molecules from the aqueous phase can
migrate or diffuse to the centre of the micelles, resulting in the aqueous voids or entrapped
water molecules in the organic phase droplets.
R

R
R

COO-
-OOC
+ +
NH 3 H3N
COO- R
+ +
NH 2 NH 2 -OOC
R
+ H2O
H3N
COO- H2O
+
H3N H2O +
H3N
COO- -OOC
+
+ NH 3
NH 2
R -OOC
COO- R

Figure 2.17: Micelle formation by interaction between carboxyl groups of UPR and
ammonium groups of DETA 102.

40
Chapter 2: Historical and theoretical background

Gous 102 showed that the amount of DETA plays an important role, as an increased amount
of amine added to the organic phase results in an increased amount of neutralized groups,
which in turn may lead to increase in the amount of water molecules attracted to the
organic phase droplets. The small amount of DETA added to the aqueous phase prevents
the system from experiencing pH-shock when the organic phase is added.

The steric stabilizer PVOH in the aqueous phase also plays an important role in the
formation of vesicles of the MVPs. The PVOH chains orientate themselves so that the
hydrophilic acetate groups are orientated toward the aqueous phase and hydrophobic
groups are orientated toward the surface of the organic phase droplets, to ensure that no
water molecules leave the micelles.

2.3.3.2 Curing reaction of UPR and STY.

Polymerization is initiated by the reaction of the organic peroxide cumene hydroperoxide

(CHP) with ferrous sulphate (FeSO4.7H2O). The peroxide molecules react with the ferrous
ions via single-electron transfer, leading to the splitting of the peroxide molecules into a
103
stable anion and a reactive free radical . The initiation reaction believed to take place is
illustrated by Equation 2.9.

ROOH + Fe2+ → RO· + OH- + Fe3+ (2.9)

The free radicals formed by the reaction in Equation 2.9 initiate the crosslinking reaction
between the fumarate (and maleate) double bonds of the UPR chains and STY, leading to
the MVPs. Figure 2.18 shows the reaction between the UPR chains and styrene 78.

41
Chapter 2: Historical and theoretical background

O O
CH3
HO O
O
CH2 + O
O OH
O
H3C

O O
Styrene
CH3
HO O
O
OH
O
O H3C

O
O
CH3
O
HO O OH
O
O
CH3
O

CH3
O
O
HO O

O
O O Styrene

OH
H3C

Figure 2.18: Crosslink-reaction of unsaturated polyester (UPR) and styrene (STY)78.

Numerous factors influence the properties of the MVPs which include process factors such
as temperature, addition rate, mass and heat transfer, agitation, stabilizer system, properties
102, 104
of the UPR and reaction kinetics . One of the main objectives of this study is to
determine the effect of the polyester chemistry on the properties of the MVPs. Due to the
large number of factors that affect the properties of the UPR (and ultimately the properties
of MVPs), statistical design of experiments was used to minimize the amount of
experiments and draw meaningful conclusions from the results. Statistical design of
experiments (DoE) is discussed in Section 3.

42
Chapter 2: Historical and theoretical background

2.4 Design of Experiments (DoE)

2.4.1 Introduction

Statistical design was led by work of Sir Roland A. Fisher in the agricultural industry in the
1920s and the early 1930s. Fisher introduced several experimental design principles which
include factorial design and analysis of variance (ANOVA) concepts107-109.

Statistical design, first applied in industry in the 1930s, was promoted by Box and Wilson
105
(1951) with the development of the response surface methodology (RSM) . RSM can be
defined as a group of mathematical and statistical techniques used to model and analyze
results/data in which the response/property of interest is affected by numerous factors and
106
the aim is to obtain an optimum response . RSM and other designs spread throughout
several industries over the next 30 years.

Statistical design for quality improvement began in the 1970s. The use of experimental
107-109
design was expanded by Genichi Taguchi and others . Today, statistical design is
applied in many areas including industries such as the chemical, automotive and electronics
industry and the service sector of business and financial services 106.

Statistical design of experiments refers to the process of planning a series of experiments so

that appropriate data can be analyzed by statistical methods, to draw meaningful
106, 110
conclusions from the data . Statistical and experimental design is applied in product
and process development to improve process yields, and minimize variability, development
time and overall cost 106, 110.

Two approaches (apart from statistical design) exist when conducting an experiment:
Best-guess approach: The best-guess approach involves the selection and testing of
an arbitrary combination of factors. The best-guess approach often works well when
the experimenters have advanced technical or theoretical knowledge as well as
considerable practical experience. The best guess approach does not always produce

43
Chapter 2: Historical and theoretical background

the desired results and there is no guarantee that when an acceptable result is found, it
is the best solution 106.
One-factor-at-a-time approach: In the one-factor-at-a-time approach, a starting point
(or a baseline of set levels) is selected for each factor, whereby each factor is varied
successively while the other factors are held constant. The disadvantage of the one-
factor-at-a-time approach is that no interaction effects are considered. An interaction
is when one factor fails to reproduce an effect on the response at different levels of
another factor 106.

These approaches are always less efficient than statistical design methods. The most
common statistical design is the factorial statistical design in which all combinations of
106
levels of factors are varied together, instead of one at a time . Other statistical designs
include factorial screening designs, response surface methods, mixture designs (the factors
are the ingredients of a product (e.g. paint formulation) and the levels of the
factors/ingredients are dependent on each other, i.e. the levels of ingredients add up to 100
percent) 106 and combinations thereof.

A screening design, which is used in the present study, is a fraction of a full design and is
often employed first when a process or product is affected by a large number of factors, to
determine the major factors that affect a particular response, the range of the factors and the
curvature of the response. The major advantages of screening designs are the fact that a
minimal number of experiments are required, and that information obtained can be used in
subsequent designs to optimize or fine-tune a process or product.

2.5 References

1. Goodman, I., In Encyclopedia of Polymer Science and Technology, 1969; Vol. 11, p
62.
2. Ahamad, A.; Lubic, M.; Mohan, A.; Safeer, M.; Thachil, E. T. Designed Monomers
and Polymers 2001, 4, 3, 261.

44
Chapter 2: Historical and theoretical background

3. Thachil, B. C.; Thachil, E. T. Polymer - Plastics Technology and Engineering 2005,

44, 931.
4. von Boenig, H., In Encyclopedia of Polymer Science and Technology, 1969; Vol.
11, p 129.
5. Tawfik, S. Y.; Asaad, J. N.; Sabaa, M. W. Polymer Testing 2003, 22, 747.
6. Sandler, S. R.; Karo, W., Polyesters. In Polymer Syntheses, 2nd ed.; Academic
Press, Inc.: Boston, 1992; Vol. 1, pp 68 - 86.
7. Cherian, B.; Thachil, E. T. International Journal of Polymeric Materials 2004, 53,
829.
8. Brydson, J. A., Polyesters. In Plastics Materials, 7th ed.; 1999; pp 694 - 743.
9. von Boenig, H., Unsaturated Polyesters. Elsevier: New York, 1964.
10. Berzelius, J. Rapport Annuel sur les Progres des Sciences Physiques et Chimiques
1847, 26, 1.
11. Berthelot, M. M. Comptes Rendus 1853, 37, 398.
12. Lorenzo, A. V. Annales De Chimie Et De Physique 1863, 67, 2, 293.
13. Bjorksten, J., Polyesters and Thier Applications. Reinhold: New York, 1956.
14. Zaske, O. C.; Goodman, S. H., Unsaturated Polyester and Vinyl ester resins. In
Handbook of Thermoset Plastics, 1999; pp 97-168.
15. Whinfield, J. R., In Nature, London, 1946; Vol. 158, p 930.
16. Bayer, O. Justus Liebigs Annalen Der Chemie 1941, 549, 286.
17. Johnson, K. G.; Yang, L. S., Preparation, Properties and Applications of
Unsaturated Polyesters. In Modern Polyesters,: Chemistry and Technology of
Polyesters and Copolyesters, Scheirs, J.; Long, T. E., Eds. John Wiley & Sons:
2003.
18. Kaska, J.; Lesek, F. Progress in Organic Coatings 1991, 19, 283.
19. United States Patent 3, 109, 831, 1961.
20. United States Patent 3, 109, 832, 1962.
21. United States Patent 3, 109, 834, 1962.
22. Hawthorne, J. M.; Heffelfinger, C. J., In Encyclopedia of Polymer Science and
Technology, 1988; Vol. 11, p 1.

45
Chapter 2: Historical and theoretical background

23. Goodman, I., In Encyclopedia Polymer Science and Engineering, 1988; Vol. 12, p
1.
24. von Boenig, H., In Encyclopedia of Polymer Science and Technology, 1969; Vol.
11, p 729.
25. Korshak, V. V.; Vinogradova, S. V., Polyesters. Elsevier: Amsterdam, 1953.
26. Korbor, J.; Golob, J.; Sebenik, A. Polymer Engineering and Science 1993, 33, 18,
1212.
27. Jedlovcnik, R.; Sebenik, A.; Golob, J.; Korbar, J. Polymer Engineering and Science
1995, 35, 17, 1413.
28. Yang, Y. S.; Pascault, J. P. Journal of Applied Polymer Science 1997, 64, 133.
29. Shah, M.; Zondervan, E.; de Haan, A. B. Journal of Applied Sciences 2010, 10, 21,
2551.
30. Larez, C. V.; Perdomo, M. G. A. Polymer Bulletin 1990, 23, 577.
31. Lehtonen, J.; Immonen, K.; Salmi, T.; Paatero, E.; Nyholm, P. Chemical
Engineering Science 1996, 51, 11, 2799.
32. Vanso-Szmercsanyi, I.; Marcos, L. K.; Zahran, A. A., Journal of Applied Polymer
Science 1948, 10, 5, 513.
33. Matynia, T.; Worzakowska, M.; Tarnawski, W. Journal of Applied Polymer Science
2006, 101, 3143.
34. Grobelny, D.; Kotas, A. Polymer 1995, 36, 7, 1363.
35. Gamstedt, E. K.; Skrifvars, M.; Jacobsen, T. K.; Pyrz, R. Composites: Part A 2002,
33, 1239.
36. Processing Unsaturated Polyesters based on Amoco Isophthalic acid; IP-43b;
Amoco Chemical Co.: 1989; pp 1 - 19.
37. Larez, C. J.; Mendoza, G. A. P., Polymer Bulletin 1990, 23, 577.
38. Larez, C. J.; Mendoza, G. A. P. Journal of Applied Polymer Science 1991, 43,
1605.
39. Vancso-Szmercsanyi, I.; Manx, L. K.; Zahran, A. A. Journal of Applied Polymer
Science 1966, 10, 5, 513.
40. Grobelny, J. Polymer 1995, 36, 22, 4215.

46
Chapter 2: Historical and theoretical background

41. Savicheva, O. I.; Sedov, L. N.; Khramova, T. S.; Urman, Y. G.; Slonim, I. Y.,
Vysokomelkulyarnye Soedineniya Seriya A & Seriya B 1973, A15, 96, 1.
42. Ordelt, Z. Vysokomelkulyarnye Soedineniya Seriya A & Seriya B 1962, 4, 1110.
43. Ordelt, Z. Chemicky Prumsyl 1966, 16, 22, 97.
44. Ordelt, Z. Makromolekulare Chemie 1963, 63, 153.
45. Ordelt, Z. Makromolekulare Chemie 1963, 68, 166.
46. Fradet, A.; Marechal, E. Makromolekulare Chemie 1984, 188, 319.
47. Paci, M.; Crescenzi, V.; Supino, N. Makromolekulare Chemie 1982, 183, 377.
48. Piras, A. Universite´ Pierre et Marie Curie (Paris VI), France, 1988.
49. Herrmann, D.; Kimmer, W.; Sauer, W. Acta Polymer 1984, 35, 277.
50. Robins, R. G. Australian Journal of Applied Science 1954, 5, 187.
51. Ahjopaloa, L.; sterholm, H. O.; Ja¨rvinen, H.; Pietila¨, L. O. Polymer 2000, 41,
8283.
52. Hsu, C. P.; Lee, L. J. Polymer 1991, 32, 12, 2263.
53. Eisenberg, P.; Lucas, J. C.; Williams, R. J. J. Unsaturated polyesters, John Wiley &
Sons, Inc. 1997, 755 - 760.
54. Parkyn, B., Chemistry of polyester resins. Composites 1972, 29 - 33.
55. Rieckmann, T. H.; Volker, S., Poly (ethylene Terephthalate) Polymerization -
Mechanism, catalysis, kinetics, mass transfer and reactor design. In Modern
Polyesters: Chemistry and Technology of Polyesters and Copolyesters, J. Scheirs,
T. E. L., Ed. John Wiley & Sons Ltd.: West Sussex, England, 2003; pp 31-115.
56. Goodman, I., John Wiley: New York, 1972; Vol. 11.
57. Brydson, J. A. Plastic Materials. Butterworths: USA, 1973.
58. Jadhav, J. Y.; Kantor, S. W., In Encyclopedia of Polymers Science and
Engineering, 3rd ed.; John Wiley: New York, 1964; Vol. 12, pp 256 - 290.
59. Auria, B. D. Journal of Polymer Materials 1988, 25, 125.
60. Narkis, M., Size distribution of Suspension-Polymerized Unsaturated Polyester
Beads. Journal of Applied Polymer Science 1979, 23, 2043 - 2048.
61. Stewart, D. M. D., Opacifiers for Latex Paints. Microvoid Polymer Particles. In
Surface Coatings. Raw materials and their usage, The New South Wales University
Press: Kensington, NSW, Australia, 1993; pp 343-350.

47
Chapter 2: Historical and theoretical background

62. El-Aasser, M. S., Polymer News 1977, 3, 4, 209.

63. Polymerization Conditions. In Experiments in Polymer Science, E. A. Collins, J. B.,
F. W. Billmeyer, Ed. John Wiley & Sons: 1973; p 33.
64. Dowding, P. J.; Vincent, B. Colloids and Surfaces A: Physicochemical and
Engineering Aspects 2000, 161, 259.
65. Grulke, E. A., Suspension Polymerization. In Encyclopedia of Polymer Science and
Engineering, Mark, H. F., Ed. John Wiley & Sons: 1988, Vol. 16, pp 443 - 473.
66. Arshady, R., Leading Contribution: Colloid Polymer Science 1992, 270, 717.
67. Yuan, H. G.; Kalfas, G.; Ray, W. H. Journal of Macromolecular Science - Reviews
in Macromolecular Chemistry & Physics 1991, C31, 215.
68. Jahanzad, F.; Sajjadi, S.; Brooks, B. W. Chemical Engineering Science 2005, 60,
5574.
69. Hofman, F.; Delbruck, K. German Patent, 250,690, 1909.
70. Hofman, F.; Delbruck, K. German Patent, 254,672, 1912.
71. Hofman, F.; Delbruck, K. German Patent, 255,129, 1912.
72. Bauer, W.; Lauth, H. German Patent, 656.134, 1931.
73. Hohenstein, M., Journal of Polymer Science 1946, 1, 127.
74. Voss, A.; Eisfeld, K.; Freundenberger, H. United States Patent, 210,9981, 1938.
75. Crawford, J. W. C.; McGrath, J. United States Patent, 210,8044, 1938.
76. Rohm, O.; Trommsdorff, E. German Patent, 735,284, 1943.
77. Dowding, P. J.; Goodwin, J. W.; Vincent, B. Colloids and Surfaces A:
Physicochemical and Engineering Aspects 1998, 145, 263.
78. El-Nahas, H. H.; Khalil, F. H.; Ibrahim, G. M.; El-Gammal, B. Journal of Applied
Polymer Science 2007, 104, 1149.
79. Winslow, F. H.; Matreyek, W. Industrial Engineering Chemistry 1951 1951, 43, 5,
1108.
80. Arshady, R.; Ledwith, A., Reactive Polymers 1983, 1, 159.
81. Hopff, H.; Lussi, H.; Hammer, E., Makromolekulare Chemie 1965, 82, 175, 184.
82. Kavarov, V. V.; Babanov, B. M., Journal of Applied Chemistry USSR (Engl
Translation) 1959, 32, 810.

48
Chapter 2: Historical and theoretical background

83. Mersmann, A.; Grossman, H., Chemical Engineering and Technology 1980, 52,
621.
84. Sculles, D. B., Journal of Applied Polymer Science 1976, 20, 2299.
85. Arshady, R. Journal of Chromatography 1991, 586, 181.
86. Kotoulas, C.; Kiparissides, C. Chemical Engineering and Technology 2006, 61,
332.
87. Chatzi, E. G.; Gavrielides, A. D.; Kiparissides, C. Industrial and Engineering
Chemistry Research 1989, 28, 1704.
88. Niwas, R.; Khan, A. A.; Varshney, K. G. Colloids and Surfaces A 1999, 7, 150.
89. Taylor, G. Proceedings of the Royal Society A (London) 1934, 1, 46A, 501.
90. Shinnar, R.; Church, J. M. Industrial and Engineering Chemistry Research 1960,
35, 253.
91. Howarth, W. Chemical Engineering Science 1964, 19, 33.
92. Goldsbrough, K.; Simpson, L. A.; Tunstall, D. F. Progress in Organic Coatings
1982, 10, 35.
93. Castellanos, J. R.; Mendizabal, E.; Puig, J. E. Journal of Applied Polymer Science
1991, 49, 91.
94. Dawkins, A., Chain Polymerization II. In Comprehensive Polymer Science, A.
Ledwith, S. R. a. P. S., Ed. Pergamon Press: Oxford, 1989; Vol. 4, p 231.
95. Jahanzad, F.; Sajjadi, S.; Brooks, B. W. Industrial and Engineering Chemistry
Research 2005, 44, 11, 4112.
96. Odian, G., Principles of Polymerisation. 3rd ed.; Wiley: New York, 1991.
97. Lazrak, N.; Bolay, N. L.; Ricard, A. European Polymer Journal 1998, 34, 1637.
98. Willkie, H. F. United States Patent, 144,9157, 1923.
99. Kershaw, R. W., Australian Oil and Colour Chemists Association Proceedings
1971, 4.
100. Ritchie, P., Spindrift Pigmented Vesiculated Beads. In Surface Coatings. Raw
materials and their usage, The New South Wales University Press: Kensington,
NSW, Australia, 1993; pp 351 - 355.
101. Engelbrecht, J. F.; De Wet-Roos, D.; Smit, A. C.; Cooray, B. Vesiculated polymer
particles. US 2006/0111474 A1, 2006.

49
Chapter 2: Historical and theoretical background

102. Gous, K. Continuous processing of vesiculated beads. MScEng, University of

Stellenbosch, 2003.
103. Kolthoff, I. I.; Aiedalia, A. I., Determination of Organic Peroxides by Reaction with
Ferrous Iron. Analytical Chemistry 1951, 23, 4, 595.
104. Terblanche, J. C. The development of vesiculated beads. MScEng, University of
Stellenbosch, 2002.
105. Box, G. E. P.; Wilson, K. G. Journal of the Royal Statistical Society B 1951, 13, 1.
106. Montgomery, D. C., Design and Analysis of Experiments. 6th ed.; John Wiley &
Sons, Inc.: 2005.
107. Taguchi, G.; Wu, Y., Introduction to Off-line Quality Control. In Central Japan
Quality Control Association, Nagoya, Japan, 1980.
108. Taguchi, G., System of Experimental Design: Engineering Methods to Optimize
Quality and Minimize Cost. In UNIPUB, New York, 1987.
109. Taguchi, G., Introduction to Quality Engineering. In Asian Productivity
Organization, UNIPUB, New York, 1991.
110. Rekab, K.; Shaikh, M., Statistical Design of Experiments with Engineering
Applications. Taylor & Francis Group, LCC: London, 2005.

50
 CHAPTER 3

EXPERIMENTAL

3.1 Introduction

In this chapter the experimental methods used to synthesize and characterize a series of
UPRs (poly (propylene fumarate phthalate)), which vary in terms of molecular structure,
are described. UPRs with different molecular structures were obtained by polymerizing PG
with varying MA:PA mole ratios and different processing parameters (e.g. heating rates,
stirring rates and maximum temperatures). The resultant polymers were added to a mixture
of STY and a polymerization inhibitor, tertiary butyl catechol (TBC).

The resultant UPR-STY mixtures were then used in the synthesis of MVPs by suspension
polymerization, to determine the effect of the UPR molecular structure on the properties
(e.g. particle size and degree of vesiculation) of the MVPs.

Due to the large number of factors that affect the molecular structure and properties of the
UPR, DoE software (Design Expert 7) was employed. The Design Expert software was
used to find the optimum experimental space, i.e. least amount of experiments, but
covering all the factors. The software allows the results to be analyzed in a statistical
manner, i.e. the percentage contribution of each factor can be determined and their
interaction with other factors.

The properties of the UPRs including molecular weight and molecular weight distribution,
molecular structure (end-group concentration, % chain branching) and glass-transition
temperature (Tg) was determined by SEC, NMR and DSC respectively.
SEM was used to directly visualize the MVPs morphology, while SEM with microtoming
was used to determine the degree of vesiculation of the MVPs. Laser diffraction was used
to determine the particle size– and distributions of the MVPs. In the following sections the

4
Chapter 3: Experimental

formulations, experimental procedures and characterization of the UPRs and MVPs are
described.

3.2 Materials

The UPRs were synthesized using maleic anhydride (MA), phthalic anhydride (PA),
propylene glycol (PG), styrene (STY) and tertiary butyl catechol (TBC). The MVPs raw
materials included the UPRs obtained from DoE series, STY, lauryl methacrylate (LMA),
diethylene triamine (DETA), poly (vinyl alcohol) (PVOH) (10.7% solution in water),
hydroxy ethyl cellulose (HEC) (2.5% solution in water), distilled water (DI), cumene
hydroperoxide (CHP), ferrous sulphate (FeSO4.7H2O), anionic surfactant (sulfosuccinate
based on a fatty alcohol ether, disodium salt), ammonia (12.5% solution in water),
hydrophobically modified anionic thickener (50% solution in water), biocide based on 2-
bromo-2-nitropropandiol and zinc oxide (ZnO) dispersion. All the above mentioned raw
materials were used as received from Plascon SA.

Trichloroacetyl isocyanate (Aldrich 96%) and deuterated chloroform (CIL 99.8%) were
used as received.

3.3 UPR Synthesis

PG, MA and PA are charged into a 1L reactor under continuous agitation. The mixture is
heated from room temperature to 120 °C (heating rate 1), when the exotherm sets in.
(nitrogen is blown into the reactor before heating starts for approximately 15 minutes to
displace any oxygen, and then the N2 is blown throughout the whole process). A “slow”
heating rate involves increasing the temperature of the heating mantle in increments of
10 °C from room temperature to 120 °C, after which heating is discontinued when the
exotherm sets in. A “fast” heating rate involves setting the heating mantle at 120 °C from
room temperature. The maximum exotherm temperature is generally between
160 °C – 175 °C depending on heating rate 1.

40
Chapter 3: Experimental

The reactor is allowed to cool back to 160 °C for a slow exotherm rate, whilst a fast
exotherm rate involves heating the reactor to the maximum reaction temperature as soon as
the maximum exotherm temperature is reached.

After the exotherm, the mixture is heated to the maximum reaction temperature, heating
rate 2. Fast heating rate 2 means heating immediately after the maximum exotherm
temperature is reached, whilst a slow heating rate 2 means setting the heating mantle in
increments of 10 °C from the exotherm to the maximum reaction temperature.

The distillation column temperature is approximately 100 °C while water and a small
amount of glycol are continuously removed from the reaction by distillation. As soon as the
column temperature subsides to below 60 °C (no water is condensed), vacuum is applied
(average of 3 times for 1 minute) using a Vacuubrand MZ 2C diaphragm vacuum pump at
9.0 mbar together with a 1L glass solvent vapor trap, to remove residual water and un-
reacted monomers from the system, effectively lowering the acid-value and increasing the
viscosity of the polyester until the specifications are reached.

The mixture is allowed to cool to approximately 160 °C, at which it is slowly discharged
into a mixture of STY and TBC, maintaining the polyester/styrene mixture at
approximately 60 °C. Once the discharge is complete, the mixture is allowed to cool to
room temperature. A typical unsaturated polyester resin formulation is tabulated in Table
3.1.

Table 3.1: Standard UPR formulation.

Reagents Weight (%)
PG 30.74
MA 26.19
PA 13.18
STY 27.25
Additional STY 2.62
TBC 2.00×10-3

41
Chapter 3: Experimental

3.4 Design of Experiments (DoE)

A Combined response surface and mixture design was chosen to determine the effects of
the formulation and processing factors on the properties of the resins. The Design Expert 7
software determined that 27 runs (which include multiple replicate runs) were sufficient to
evaluate the effects of the tested factors. The following factors were varied:

Formulation factors
MA:PA mole ratio

Process factors
Heating rates
Exotherm time/ rate
Maximum temperature
Agitation speed.

Table 3.2 illustrates the experimental design space used to synthesize the resins. Each
factor was varied over three levels (lower, intermediate and high level)

Table 3.2: Formulation and process factors for UPR synthesis.

Factor Lower level Intermediate High level
(-1) (0) (+1)
% Phthalic anhydride 9.24 13.18 17.12
% Maleic anhydride 22.25 26.19 30.13
Average heating rate 1 (±18 °C to 120 °C) (°C/min) 3.20 6.20 5.90
Exotherm rate (120 °C to max. exotherm temp) slow - fast
Average heating rate 2 (±160 °C to max.) (°C/min) 1.00 1.20 1.30
Maximum temperature (°C) 240 250 260
Agitation speed (rpm) 300 400 500
Notes: %Propylene glycol is kept constant at 30.74%.
All process parameters are affected by the formulation.
Exotherm rate depends heavily on heating rate 1.

42
Chapter 3: Experimental

The resultant UPR samples were used in the synthesis of Multi-vesiculated particles
(MVPs) via suspension polymerization to determine the effect of UPR structure and
properties on those of the MVPs.

3.5 MVP synthesis

The unsaturated polyester resin – styrene (UPR-STY) mixture is mixed with STY, followed
by LMA and DETA. The LMA acts as a hydrophobic monomer and DETA is added to
neutralize the main and side-chain carboxylic groups of the UPR and to increase the
1-2
solubility of water of the mixture . The UPR-STY-LMA-DETA mixture ( hereafter the
organic phase) is added to an aqueous phase made up of DI water, PVOH and HEC
solutions and DETA over ± 10 minutes at room temperature (±23 °C) and whilst stirring at
330 rpm in a 2L round-bottom metal reactor suspended in a temperature-controlled water-
bath (Dispersion stage). The PVOH acts as a polymeric stabilizer to ensure a stable oil-in-
water emulsion and the HEC is added to increase the viscosity of the aqueous phase and
prevent settling of the oil droplets 1-2.

The dispersion stage is followed by the “Emulsification stage” (± 20 minutes) where the
organic phase droplets undergo the break-up-coalescence process to form stable droplet
sizes and aqueous voids are formed in the droplets.

The crosslinking reaction of the UPR and STY is initiated by the addition of free-radical
initiator CHP and catalyzed with FeSO4.7H2O. After the initiation stage, the stirring is
discontinued for 30 minutes.

After the crosslinking reaction, the system is stirred again and heated up to 50 °C in the
water-bath. After 60 minutes at 50 °C, the system is heated to 60 °C and kept at that
temperature for 3 hours. The suspension is left to cool to room temperature over-night after
which it is post treated with the surfactant, ammonia, anionic thickener, biocide and ZnO
dispersion. Table 3.3 shows a typical MVPs formulation and the relative weight
percentages of the reagents.

43
Chapter 3: Experimental

Table 3.3: Standard MVPs formulation.

Reagents Weight (%)
UPR – STY mixture 15.07
STY 5.79
LMA 0.64
DETA 0.21
Water 51.26
PVOH solution (10.7%) 13.08
HEC solution (2.5%) 10.30
DETA 0.06
Water 0.21
FeSO4.7H2O 0.01
CHP 0.12
Anionic surfactant 0.99
Water 0.19
25 % Ammonia solution 0.19
Water 0.19
Anionic thickener 0.19
Biocide 0.24
ZnO dispersion 0.38

3.6 Analytical techniques and measurements

Various analytical techniques and measurements were used to characterize both the UPRs
(no styrene) and MVPs. All analytical methods and the purpose of using them are listed in
sections 3.6.1 – 3.6.11.

3.6.1 Size-Exclusion Chromatography

Size-Exclusion Chromatography (SEC) was used to determine the molecular weight and
molecular weight distributions of the UPRs. The UPR was dried at ambient temperature
and dissolved in HPLC-grade THF (1 mg/mL) for 24 h before being filtered through a

44
Chapter 3: Experimental

0.45 µm nylon filter. The SEC system used for the analyses consisted of a Waters 717 auto-
sampler, Waters 600E system controller (run by Breeze software), Waters 610 fluid unit
and a Waters 410 differential refractometer at 35 °C. THF (HPLC-grade) spurged with IR-
grade helium was used as an eluent at a flow rate of 1 mL/min. Two PLgel 5-µm mixed-C
columns and a pre-column PLgel 5-µm guard were used. The system was calibrated with
narrow polystyrene standards ranging from 800 to 2 × 106 g. mol-1. All SEC results are
given relative to polystyrene standards.

3.6.2 Differential Scanning Calorimetry (DSC)

Differential Scanning Calorimetry was used to determine the glass-transition temperatures
(Tg) of the UPRs. A Perkin Elmer Pyris 1 DSC was used and the sample (±12 mg) was
heated under a nitrogen atmosphere (gas flow = 20 mL/min) from -50 °C to 50 °C at a
heating rate of 20 °C/min.

3.6.3 Nuclear Magnetic Resonance (NMR) spectroscopy

Proton Nuclear Magnetic Resonance Spectroscopy (1H NMR) was used to resolve the
molecular structure of the UPRs. Dried UPR was dissolved in deuterated chloroform (with
TMS as reference) (300 mg/mL) for 24h. The 1H NMR analyses were performed on Varian
VXR 300MHz Spectrometers (300 and 400 MHz).

3.6.4 Scanning Electron Microscopy (SEM)

Scanning Electron Microscopy (SEM) was used to determine the particle size and
morphologies of the MVPs. A Leo® 1430VP Scanning electron microscope was used to
obtain images of the MVPs at 7kV at a working distance of ±10 mm, using secondary
electron imaging.
The samples were prepared by drying the MVPs at ambient temperature, placing the dried
powder on SEM sample stubs and gold-plating the stubs.

45
Chapter 3: Experimental

3.6.5 Microtoming
Microtoming in conjunction with SEM was done to determine the internal morphology
(degree of vesiculation) of the MVPs. Dried MVPs were mixed with a crosslinkable resin
and dried in an oven at 50 °C for 24h. Thin layers (±0.1 µm thickness) of the cured resin-
MVPs were sliced off or microtomed using a diamond blade, until the sliced/ cut MVPs
were exposed at the resin surface. These microtomed resins were placed on stubs and the
internal structure of the MVPs viewed via SEM.

3.6.6 Laser Diffraction (LD)

Laser Diffraction (LD) was used to determine the particle size and distribution of the
MVPs. A Sympatec SUCELL HELOS/BF-OM laser diffraction sensor was used for the
particle size and distribution analysis. The HELOS is equipped with a high precision
optical bench, 31 channel multi-element detector and auto-alignment unit (auto-focus), on
board data acquisition electronics, automatic laser beam diameter adaptation, 5 mW HeNe-
laser with fibre optic light transmission and a high speed fibre optic data communication
interface.

The method requires that a small amount of sample material is added in a well chosen
liquid into the basin of the SUCELL. The resulting suspension or emulsion is homogenised
by a speed-controllable, double-stirrer and dispersed by ultrasonication. It is transported
through a cuvette with the help of a peristaltic pump. The suspension is examined by the
HELOS laser beam. The SUCELL is a universal wet dispersion system which can detect
particle sizes from 0.1 μm to 875 μm, representing ranges R1 up to R5 of the HELOS
system. The R3 lens was used as it can detect particle sizes in the range of 0.5 to 175 µm.
All operations are controlled and monitored by the WINDOX software via an RS485
interface.

3.6.7 Acid-value
The main chain and unreacted carboxylic acid concentration (acid-value) of the UPR-STY
mixtures was determined visually by acid-base titration using a 4.95 N potassium

46
Chapter 3: Experimental

hydroxide (KOH) in a methanol solution and phenolphthalein as indicator. The acid-value

(mg KOH per gram UPR) was calculated using the following equation:

VKOH N
Acid value (3.1)
S UPR STY
WUPR STY
100
Where:
VKOH : Volume (mL) of KOH solution used in titration
N : Normality of KOH solution
WUPR-STY : Weight (g) of UPR-STY mixture sample
SUPR-STY : Solids content of UPR-STY mixture

3.6.8 Solids content

The solids content of the UPR-STY mixture was determined by drying ~ 1 g of the mixture
at 105 °C in an oven for 2 h. The solids content of the MVPs dispersion was determined by
drying ~ 1.5 g of the dispersion at 180 °C in a Mettler Toledo HR73 Halogen Moisture
Analyzer for 11 min 40 sec. The solids content of both the UPR-STY mixture and the MVP
dispersion was calculated using the following equation:

w1 w2
Solids Content (3.2)
w1

where w1 and w2 are the sample weight before and after drying respectively.

3.6.9 Viscosity
Viscosity: The viscosities of the UPR-STY mixtures and MVP dispersions were
determined at ambient temperature using a Brookfield DV-II+ Pro Viscometer with a
LVDV-II+ Pro using spindles #3 and #4, at 50 and 30rpm respectively.

47
Chapter 3: Experimental

3.6.10 Contrast ratio (opacity)

The contrast ratio or opacity of material is a measure of the dry hiding power (or ability to
completely cover/ obliterate a substrate) it possesses. The contrast ratio of a paint/ coating
is usually determined by the reflectance ratio of a dried film applied to a black substrate
relative to a white substrate:
reflective index of black surface
Opacity (3.3)
reflective index of white surface

Where opacity = 1, constitutes complete hiding/ obliteration of the substrate by the test
material.

Test samples were prepared by drawing a thin film (200 µm) of the MVP dispersion, using
an automatic film applicator, onto an opacity chart consisting of black and white colored
surface. The thin film is left to dry in an oven at 50 °C for 15 minutes.
A reflectometer (at a 45° angle) was used to determine the reflective indices of the black
and white colored surfaces, whereafter Equation 3.3 was used to determine the contrast
ratio or opacity imparted by the MVPs.

3.6.11 Hardness of MVPs

The hardness of the MVPs was determined using two techniques, namely:
 Microhardness: The UHL VMH-002 microhardness tester was used to determine the
hardness of MVPs at 25 °C, with indentation load of 0.01 N, indentation speed of
25 μm/s and dwell time of 15 seconds.

The samples were prepared by sprinkling oven-dried MVPs onto a disk-shaped mold of
semi-crosslinked alkyd resin, after which the resin was allowed to cure completely in
an oven at 50 °C for 24h. The test was performed on the cured alkyd resin with dried
MVPs exposed at the surface, to ensure that the indenter is in direct contact with the
MVPs and to ensure that the MVPs remain stationary when they are indented. The
microhardness of the MVPs would be equal to the sum of the hardness values of the
MVPs embedded in the resin, minus the hardness value of the resin itself.

48
Chapter 3: Experimental

 Atomic Force Microscopy (AFM): The Veeco Multimode AFM in contact mode was
the second technique used to determine the hardness of the MVPs.

The test samples were prepared by sprinkling oven-dried MVPs on a thin layer semi-
dry PVOH on a glass plate, after which the PVOH was allowed to dry completely. The
MVPs were sprinkled onto the PVOH to ensure that the MVPs are directly exposed to
the cantilever tip of the AFM, and to ensure that the MVPs remain stationary when the
cantilever moves across them.

3.7 References

1. Terblanche, J. C. The development of vesiculated beads. MScEng, University of

Stellenbosch, 2002.
2. Gous, K. Continuous processing of vesiculated beads. MScEng, University of
Stellenbosch, 2003.

49
 CHAPTER 4

RESULTS AND DISCUSSION

4.1 Design of Experiments (DoE) analysis

4.1.1 Introduction

A series of UPRs with varying chemical composition were synthesized by varying the
process parameters as well as the MA:PA mole ratio. Due to the large number of
variables that affect the polymer structure, statistical design of experiments (DoE) was
used to minimize the number of experiments, but still allow for meaningful conclusions
to be drawn from the data. Design Expert 7 software was used to generate a combined
response surface and mixture design which is sufficient to determine the effects of the
MA:PA mole ratio and processing factors on the properties of the UPRs. A total of 27
UPRs were synthesized to evaluate the effects of the factors. These UPRs (in styrene)
were subsequently used to synthesize MVPs to indirectly determine the effect of the
MA:PA mole ratio and processing factors in the synthesis of the UPRs on the particle
size and particle size distribution on the MVPs.

4.1.2 Results and discussion – UPRs

The effects of the following parameters were considered in the DoE:

A: %PA,
B: %MA,
C: Heating rate 1 (°C/min),
D: Maximum process temperature (°C),
E: Heating rate 2 (°C/min),
F: Agitation speed (rpm),
G: Exotherm rate (°C/min).

50
Chapter 4: Results and discussion

These parameters were evaluated to determine the effect on the Brookfield viscosity,
acid-value, number and weight average molecular weights (Mn and Mw) and
polydispersity index (PDI) of the UPRs. These properties were investigated as it is
known that the viscosity of UPRs, which is a function of the polymer molecular weight
and PDI (degree of polymerization, DPn) and degree of chain branching, affects the
droplet/particle formation during the suspension polymerization (Section 2.2.3) of the
MVPs. The acid-value is also of great importance due to the interaction between the
carboxylic groups of the UPRs with the base, DETA, resulting in the water-filled vesicles
of the MVPs during the suspension polymerization (see Section 2.3.3.1).

Table 4.1 shows the experimental design space and summarizes the results/ properties of
the 27 UPRs. These results were entered into the Design Expert 7 program to determine
the significance of the parameters on the specific properties of the UPRs, via analysis of
variance (ANOVA).

A factor (or interaction between factors) with a p-value (Prob > F) less than 0.05,
indicates that the factor is significant at a 95 % and higher confidence level. Futhermore,
the lower the p-value the more significant the particular factor is. Factors or interactions
with p-values more than 0.05 indicate that the model terms are significant at much lower
confidence levels.

51
Chapter 4: Results and discussion

Table 4.1: Experimental design space for UPR synthesis.

RUN PA MA Heating Heating Exotherm Maximum Agitation Brookfield Acid-value Mn Mw PDI
(%) (%) rate 1 rate 2 rate temperature speed viscosity (mgKOH/g) (Dalton) (Dalton)
(°C) (rpm) (cPs)
1 9.24 30.13 -1 -1 fast 260 300 1612 40.37 1172 4751 4.05
2 13.18 26.19 1 1 slow 240 300 1370 38.16 1407 3153 2.24
3 13.18 26.19 0 0 fast 250 400 1735 25.31 1329 6247 4.70
4 17.12 22.25 1 1 slow 260 300 2000 32.91 1025 3046 2.97
5 13.18 26.19 1 1 fast 251 500 3311 25.94 1596 17068 10.69
6 17.12 22.25 -1 1 slow 240 500 1509 28.15 1351 4979 3.68
7 13.18 26.19 1 -1 fast 260 300 2199 27.59 1319 7661 5.81
8 11.21 28.16 -0.7 -1 slow 257 475 1507 27.56 1309 6431 4.91
9 13.18 26.19 -1 1 slow 260 500 830.2 28.65 1270 3716 2.92
10 17.12 22.25 1 -1 slow 260 500 1430 20.13 1547 5552 3.59
11 13.18 26.19 0 0 fast 250 400 1728 23.44 1230 4956 4.03
12 9.24 30.13 -1 1 slow 260 500 1452 29.08 1102 3849 3.49
13 17.12 22.25 -1 -1 slow 260 300 746.2 32.07 1055 3079 2.92
14 17.12 22.25 1 -1 fast 240 300 1111 32.46 1015 2896 2.85
15 13.18 26.19 -1 1 fast 247 300 1564 35.18 1253 4715 3.76

52
Chapter 4: Results and discussion

RUN PA MA Heating Heating Exotherm Maximum Agitation Brookfield Acid-value Mn Mw PDI

(%) (%) rate 1 rate 2 rate temperature speed viscosity (mgKOH/g) (Dalton) (Dalton)
(°C) (rpm) (cPs)
16 17.12 22.25 -1 1 slow 240 500 1668 24.31 1141 4156 3.64
17 9.24 30.13 -1 -1 slow 240 300 1207 33.07 1093 3476 3.18
18 13.18 26.19 0 0 slow 250 400 1365 27.43 1423 3682 2.59
19 13.18 26.190 0 0 slow 250 400 1658 30.17 1149 4026 3.50
20 13.18 26.19 -1 -1 fast 240 500 1651 24.09 1101 4279 3.89
21 17.12 22.25 -1 1 fast 260 500 1821 19.28 1232 5497 4.46
22 9.24 30.13 1 -1 fast 240 500 2318 29.16 1158 4489 3.88
23 9.24 30.13 1 -1 slow 260 300 5075 31.99 1146 4482 3.91
24 9.24 30.13 0.04 -0.2 fast 249 384 1879 39.1 1388 5384 3.88
25 17.12 22.25 -1 1 fast 260 500 1908 23.47 1210 4970 4.11
26 17.12 22.25 1 -1 fast 240 300 1207 28.92 1246 3288 2.64
27 9.24 30.13 1 1 fast 240 300 1411 47.64 992 2659 2.68

53
Chapter 4: Results and discussion

4.1.2.1 Brookfield viscosity

Table 4.1 shows the Brookfield viscosity data determined for the UPR-STY mixtures.
The results show that the values ranged from ± 700 cPs to ± 5000 cPs (mean = 1732.46
cPs ± 332.49). Table 4.2 shows the significant factors (and interaction between factors)
affecting the Brookfield viscosity of the UPR-STY mixtures.

Table 4.2: Significant factors affecting the Brookfield viscosity of UPR-STY

mixtures.
Factor Significance (p-value: Prob > F)
AB 0.0029
AE 0.0295
BC < 0.0001
BD 0.0007
BE 0.0058
BG 0.0062
ABD 0.0322
ABG 0.0023

Table 4.2 shows that the most significant factors and interactions that affect the
Brookfield viscosity of the UPR-STY mixtures are AB which is the MA:PA mole ratio,
BC and BD, i.e. interactions between %MA and heating rate 1; and maximum process
temperature.

Figure 4.1 shows a two component mix graph where the effect of varying the MA:PA
mole ratio on the Brookfield viscosity of the UPRs or UPR-STY mixtures at fixed
process parameters is presented. In other words, all other parameters are at their
intermediate levels values given in Table 3.2. Figure 4.1 clearly shows that an increase in
the %MA leads to an increased viscosity of the UPR-STY mixtures. The increased
viscosity of the UPR-STY mixtures can be explained by the fact that the maleic groups in
the UPR chains undergo a side reaction with PG (Ordelt reaction) leading to chain

54
 Chapter 4: Results and discussion

branches, which influence the viscosity. An increase in the degree of unsaturation

(maleic groups) in the UPR chains leads to an increase in the probability of the Ordelt
reaction which in turn leads an increase in the degree of branching, leading to an increase
in the viscosity of the UPR-STY mixtures. However, the DB-values of the UPRs are also
a function of the degree of polymerization (DPn), i.e. number-average and weight-
average molecular weights (Mn and Mw) and the polydispersity index (PDI). The
relationships between the viscosity and molecular weight and molecular weight
distribution of the UPRs are discussed in sections 4.1.2.3 and 4.1.2.4.
Software

sity
ss %)
ss %) 2600

1 = 0.00
250.00
2 = 0.00
ed = 400.00
e = slow 2200
Brookfield Viscosity (cPs)

1800

1400

1000

Actual PA (mass %) 9.243 11.212 13.180 15.149 17.117

Actual MA (mass %) 30.127 28.158 26.190 24.221 22.253

Figure 4.1: Two component mix graph for the formulation factors affecting the
viscosity of the UPR-STY mixtures.

55
Chapter 4: Results and discussion

Figure 4.2 shows a “contour mix-process” graph where the interaction effect of %MA
and heating rate 1 (BC) on the Brookfield viscosity of the UPR-STY mixtures, is
presented.
It should be noted that a contour mix-process graph illustrates the effect of varying the
MA:PA mole ratio together with a single process parameter (e.g. heating rate 1), on the
properties of the UPRs and UPR-STY mixtures. All other parameters are kept constant at
the default values given by the Intermediate level (0) in Table 3.2 with a “slow”
exotherm rate (unless stated otherwise). An interaction between factors increases the
value of a particular property following the color spectrum from blue to red, i.e. a blue
region in the contour plot represents a low value of a particular property, whereas a red
region represents a high value of that particular property. The values of a given property
are displayed on the contours.

Figure 4.2 shows that an increase in the %MA and heating rate 1 leads to an increase in
the Brookfield viscosity of the UPR-STY mixtures. It can also be seen that a slow
heating rate 1 together with low percentages of MA leads to very low Brookfield
viscosities of the UPR-STY mixtures. Heating rate 1 influences the melting and ring-
opening of MA (by reaction with PG), i.e. a fast heating rate 1 (20 – 120 °C) allows for
improved ring-opening of MA. Thus, a higher percentage MA is incorporated in the UPR
chains, leading to an improved probability of Ordelt branching and an increase in the
viscosities of the UPR-STY mixtures.

56
 Chapter 4: Results and discussion
Design-Expert® Software

Brookfield Viscosity 1.00

3500

1200 3176.81

X1 = B: MA (mass %)
X2 = A: PA (mass %)
X3 = C: Heating Rate 1 2670.81
0.50
Actual Factors

Heating Rate 1 (coded units)

D: Max. Temp. = 250.00
E: Heating Rate 2 = 0.00 2164.80
F: Agitation Speed = 400.00
G: Exotherm time = slow
1658.80

0.00 1152.80

-0.50

-1.00

Actual MA (mass %) 22.253 24.221 26.190 28.158 30.127

Actual PA (mass %) 17.117 15.149 13.180 11.212 9.243

Figure 4.2: Contour mix-process graph of the interaction effect (BC) of heating rate
1 and %MA in MA:PA mole ratio on the Brookfield viscosity of UPR-STY
mixtures.

Figure 4.3 shows a contour mix-process graph where the interaction effect of %MA and
maximum temperature (BD) on the Brookfield viscosity of the UPR-STY mixture is
presented. Figure 4.3 shows that an increase in the %MA and maximum reaction
temperature leads to an increase in the viscosity of the UPR-STY mixtures. A high
concentration of MA together with a high maximum temperature leads to an increased
probability of the Ordelt side reaction, leading to an increased degree of branching and
increased viscosity of the UPR-STY mixtures.

57
 Chapter 4: Results and discussion
Design-Expert® Software

Brookfield Viscosity 260.000

3000 2892.40

1400

X1 = B: MA (mass %)
X2 = A: PA (mass %)
X3 = D: Max. Temp.
255.000
2521.21

Maximum Temperature (deg.C)

Actual Factors
C: Heating Rate 1 = 0.00
E: Heating Rate 2 = 0.00
F: Agitation Speed = 400.00
G: Exotherm time = slow 2150.03

250.000 1407.66
1778.84

245.000

240.000

Actual MA (mass %) 22.253 24.221 26.190 28.158 30.127

Actual PA (mass %) 17.117 15.149 13.180 11.212 9.243

Figure 4.3: Contour mix-process graph of the interaction effect (BD) maximum
temperature and %MA in MA:PA mole ratio on the Brookfield viscosity of UPR-
STY mixtures.

Figure 4.4 shows a pertubation graph where the individual effects of the process
parameters on the Brookfield viscosity of the UPR-STY mixtures (at a fixed UPR
formulation) is presented.

It should be noted that a pertubation graph compares the effect of the process factors (at
fixed MA:PA mole ratio) at a particular point in the design space. All the pertubation
results are for the main effects only and do not consider any interaction effects. For
example, the effect of the combination of maximum temperature and agitation speed on
the Brookfield viscosity of the UPR-STY mixtures. These results represent the effect of
varying each parameter on a one-factor-at-a-time basis, e.g. C = heating rate 1 will be

58
Chapter 4: Results and discussion

varied from minimum to maximum level whilst the other parameters are kept constant.
The default values of the parameters used to construct the pertubation graph are given by
the Intermediate level (0) in Table 3.2 with a “slow” exotherm rate. A steep gradient of
the line indicates that the particular property is sensitive to this factor. A relatively flat
line indicates that the factor has very little influence on the particular property. The x-
axis of the Pertubation graph is labeled “process parameter” in coded units and not the
actual units of the parameters to account for the fact that multiple parameters are
analyzed simultaneously.

Figure 4.4 illustrates that an increase in heating rates 1 and 2 as well as the agitation
speed leads to an increase in the Brookfield viscosity of the UPR-STY mixtures.

It can also be seen that the maximum reaction temperature appears to have no significant
effect on the Brookfield viscosity (indicated by the relatively flat line), although its
interaction with other factors is significant, i.e. only a high maximum process
temperature and high %MA in the MA:PA mole ratio, significantly affect the Brookfield
viscosity of the UPR-STY mixtures (see Figure 4.3).

59
 Chapter 4: Results and discussion
Software

osity

ents 1900
= 13.180
= 26.190 C

1 = 0.00
= 250.00
2 = 0.00 1575
ed = 400.00 F
Brookfield Viscosity (cPs)

e = slow

rs E
D
1250 D
E

F
925

C
600

-1.000 -0.500 0.000 0.500 1.000

Process Parameter (Coded Units)

Figure 4.4: Pertubation graph for the process factors affecting viscosity of the UPR-
STY mixtures (fixed MA:PA ratio at 26.19: 13.18).

60
Chapter 4: Results and discussion

4.1.2.2 Acid-value

As previously mentioned in Section 3.6.7, the acid-value or concentration of the carboxyl

end-groups of the UPRs was determined by an acid-base titration method. The titration
method is effective in determining the carboxyl end-group concentration of the UPRs in
the DoE space, but does not allow the determination of the hydroxyl end-groups. Section
4.6 describes the determination of both end-groups of selected experimental runs by
isocyanate derivatization coupled with 1H NMR.

Table 4.1 shows that the acid-values determined for the UPRs by acid-base titration,
ranged from 19 – 47 mgKOH/g resin (mean = 29.84 mgKOH/g resin ± 3.44). Table 4.3
shows the factors (and interaction between factors) affecting the acid-value of the UPRs.

Table 4.3: Significant factors affecting the acid-value of UPRs.

Factors Significance (p-value: Prob > F)
AB 0.0002
AF 0.0028
AG 0.0388
BE 0.0329
BF 0.0023

Table 4.3 suggests that the most significant factors affecting the acid-value of UPRs, are
the MA:PA mole ratio and interactions AF and BF between %MA, %PA and agitation
speed.

Figure 4.5 shows a two component mix graph where the interaction effect of the MA:PA
mole ratio on the acid-value of the UPRs is presented. It can be seen from Figure 4.5 that
a decrease in the %MA (or increase in %PA) leads to a lower acid-value of the UPRs.
The two anhydrides, MA and PA, have different reactivities towards PG, i.e. MA has a
higher reactivity towards PG compared to PA. Thus, a UPR formulation with a high
concentration of MA relative to PA, results in an increased concentration of carboxyl

61
 Chapter 4: Results and discussion

end-groups relative to hydroxyl end-groups as more MA reacts with PG. A UPR

formulation with a high concentration of PA relative to MA results in a higher
concentration of hydroxyl end-groups (or lower concentration of carboxyl groups/ acid-
value) due to a decreased reactivity of PA towards PG (see Section 2.3.1.1).
Software

ss %)
ss %) 31.7

1 = 0.00
250.00
2 = 0.00
ed = 400.00
e = slow
Acid-value (mg KOH/g resin)

30.625

29.55

28.475

27.4

Actual PA (mass %) 9.243 11.212 13.180 15.149 17.117

Actual MA (mass %) 30.127 28.158 26.190 24.221 22.253

Figure 4.5: Two component mix graph for the formulation factors affecting the
acid-value of UPRs.

Figures 4.6 and 4.7 show the contour mix-process graphs where the interaction effects of
% PA and % MA with the agitation speed on the acid-value of the UPRs are presented.
It can be seen in Figures 4.6 and 4.7 that an increase in the %PA and agitation speed
leads to a decrease in the acid-value of the UPRs. The agitation speed greatly affects the
reaction rate of PA and MA with PG, i.e. a high agitation speed with a high concentration
of PA in the UPR formulation leads to an improved reaction between PA and PG, which

62
 Chapter 4: Results and discussion

effectively leads to a higher concentration of hydroxyl end-groups or lower concentration

of carboxyl end-groups (acid-value). On the other hand, a high %MA and high agitation
speed, may lead to a higher concentration of carboxyl end-groups due to the improved
reaction between MA and PG compared to PA and PG.
Design-Expert® Software

Acid-value 500.00
35

25
25.87
X1 = A: PA (mass %)
X2 = B: MA (mass %)
X3 = F: Agitation Speed
450.00 27.99
Actual Factors
C: Heating Rate 1 = 0.00
D: Max. Temp. = 250.00
Agitation Speed (rpm)

E: Heating Rate 2 = 0.00

G: Exotherm time = slow

400.00
30.11

32.24
350.00

34.36

300.00

Actual PA (mass %) 9.243 11.212 13.180 15.149 17.117

Actual MA (mass %) 30.127 28.158 26.190 24.221 22.253

Figure 4.6: Contour mix-process graph of the interaction effect (AF) of %PA in
MA:PA mole ratio and agitation speed on the acid-value of UPRs.

63
 Chapter 4: Results and discussion
Design-Expert® Software

Acid-value 500.00
35

25
25.87
X1 = B: MA (mass %)
X2 = A: PA (mass %)
X3 = F: Agitation Speed
450.00
Actual Factors 27.99
C: Heating Rate 1 = 0.00
D: Max. Temp. = 250.00

Agitation Speed (rpm)

E: Heating Rate 2 = 0.00
G: Exotherm time = slow

400.00 30.11

32.24
350.00

34.36

300.00

Actual MA (mass %) 22.253 24.221 26.190 28.158 30.127

Actual PA (mass %) 17.117 15.149 13.180 11.212 9.243

Figure 4.7: Contour mix-process graph of the interaction effect (BF) of %MA and
agitation speed on acid-value of UPRs.

4.1.2.3 Number and weight-average molecular weights (Mn and Mw)

Statistical analysis of the Mn values of the UPRs, revealed that the MA:PA mole ratio
and process parameters do not affect the Mn value significantly, as all the parameters or
interaction between parameters have p-values greater than 0.05, which is important as it
implies that the DPn of the UPRs are relatively similar.

Table 4.1 shows the Mw values of the UPRs determined by SEC. The results show that
the values ranged from 2659 Dalton to 17068 Dalton (mean = 4725.94 Dalton ±
2446.39). Table 4.4 shows the factors (and interaction between factors) affecting the Mw-
value of the UPRs.

64
Chapter 4: Results and discussion

Table 4.4: Significant factors affecting the Mw-value of UPRs.

Factors Significance (p-value: Prob > F)
AB 0.0013
AD 0.0009
AE 0.0039
AF 0.0011
AG 0.0006
BD 0.042
BG 0.0434
ABD 0.0006
ABE 0.0015
ABF 0.0025
ABG 0.0012
ACD 0.0020
ACE 0.0008

Table 4.4 shows that the most significant factors affecting Mw-value of the UPRs are the
MA:PA mole ratio (AB), and interactions AG, AD and ABD, i.e. interactions between
%PA and exotherm rate, %PA and maximum process temperature and the interaction
between the MA:PA mole ratio and maximum process temperature.

Figures 4.8 and 4.9 shows the two component mix and contour mix-process graphs where
the effect of the MA:PA mole ratio and its interaction with the maximum process
temperature on the Mw-value of UPRs, is presented. Figure 4.8 shows that a decrease in
the %PA (or increase in %MA) leads to an increase in the Mw-value of the UPRs. Figure
4.9, on the other hand, shows that an increase in the maximum process temperature (at
MA:PA mole ratio of 26.19 : 13.18) leads to increase in the Mw-value of the UPRs.

The Mw-value of polymers is sensitive not only to the number of molecules (as with Mn-
value), but also to the size or weight of the molecules. Thus the M w-value is sensitive to

65
 Chapter 4: Results and discussion

the degree of polymerization and branching extent, i.e. a branched molecule has a greater
size or weight compared to a linear molecule of the same length.

Due to the fact that the DPn of the UPRs is fairly similar, the Mw-value, as with the
Brookfield viscosity of the UPRs is greatly affected by the degree of chain branching
which is affected by the MA:PA mole ratio (AB) and its interaction with maximum
reaction temperature (ABD). A high concentration of MA relative to PA or high MA
concentrations with high maximum reaction temperatures lead to an increase in the
degree of chain branching, and hence a higher Mw-value (and Brookfield viscosity) of a
given UPR.
® Software

ss %)
ss %) 4100

1 = 0.00
= 250.00
2 = 0.00
eed = 400.00
me = slow 3500
Mw-value (Daltons)

2900

2300

1700

Actual PA (mass %) 9.243 11.212 13.180 15.149 17.117

Actual MA (mass %) 30.127 28.158 26.190 24.221 22.253

Figure 4.8: Two component mix graph for the formulation factors affecting the Mw-
value of UPRs.

66
 Chapter 4: Results and discussion
Design-Expert® Software

Mw-value (microns) 260.00

17068 14658.07
2659

X1 = A: PA (mass %)
X2 = B: MA (mass %) 11004.30
X3 = D: Maximum Temperature
255.00

Maximum Temperature (deg.C)

Actual Factors
C: Heating Rate 1 = 0.00 8145.50
E: Heating Rate 2 = 0.00
F: Agitation Speed = 400.00
G: Exotherm time = slow

4608.78
250.00
2659.00

245.00
2659.00

4608.78

240.00

Actual PA (mass %) 9.243 11.212 13.180 15.149 17.117

Actual MA (mass %) 30.127 28.158 26.190 24.221 22.253

Figure 4.9: Contour mix-process graph of the interaction effects AD and ABD, i.e.
interactions between % PA, MA:PA mole ratio and maximum temperature on Mw-
value of UPRs.

The effect of the interaction AG between % PA and the exotherm rate cannot be shown
graphically in a contour mix-process graph as the exotherm rate was not varied at
multiple levels, i.e. the UPRs were only synthesized with a “slow”- and “fast” exotherm
rate. However, the interaction effect between the % PA and exotherm rate is important,
as a slow exotherm rate would improve the reaction between PA and PG, effectively
increasing the incorporation of PA in the UPR chains. The increased incorporation of PA
would indirectly influence the MA incorporation into the UPR chains, which influences
chain-branching and therefore the Mw-value of the UPRs.

67
Chapter 4: Results and discussion

4.1.2.4 Polydispersity index (PDI)

Table 4.1 shows the PDI values of the UPRs determined by SEC. The results show that
the values ranged from 2.21 to 10.69 (mean = 3.79 ± 1.43). Table 4.5 shows the factors
(and interaction between factors) affecting the PDI values of the UPRs.

Table 4.5: Significant factors affecting the PDI-value of UPRs.

Factors Significance (p-value: Prob > F)
AB 0.0035
AC 0.0445
AD 0.0033
AE 0.0096
AF 0.0043
AG 0.0013
ABD 0.0017
ABE 0.0041
ABF 0.0092
ABG 0.0034
ACD 0.0036
ACE 0.0024

It can be seen from Table 4.5 that the PDI-value of the UPRs is significantly affected by
the MA:PA mole ratio (AB), and the interactions AG and ABD, i.e. interactions between
%PA and exotherm rate and MA:PA mole ratio and maximum temperature.

Figures 4.10 and 4.11 show the two component mix and contour mix-process graphs
where the effect of the MA:PA mole ratio (and its interaction with the maximum process
temperature) on the PDI-value of UPRs, is presented. Figure 4.10 shows that a decrease
in the %PA (or increase in %MA) leads to an increase in the PDI-value of the UPRs.
Figure 4.11, on the other hand, shows that an increase in the maximum process

68
 Chapter 4: Results and discussion

temperature (at MA:PA mole ratio of 26.19 : 13.18) leads to increase in the PDI-value of
the UPRs.

The PDI-value of polymers is the ratio of Mw- and Mn-values, and is therefore a measure
of the degree of polymerization (DPn) and degree of branching of the UPR chains. The
ANOVA evaluation of the Mn-values revealed that the DPn of the UPRs is relatively
similar, which means that the PDI value of the UPRs is significantly affected by the M w
contribution. Thus, the PDI value (as with the Mw value) is significantly affected the
MA:PA mole ratio, interactions between %PA and exotherm rate; and MA:PA mole ratio
and the maximum reaction temperature.
Software

ss %)
ss %) 3.6

1 = 0.00
= 250.00
2 = 0.00
ed = 400.00
e = slow 3.075
PDI

2.55

2.025

1.5

Actual PA (mass %) 9.243 11.212 13.180 15.149 17.117

Actual MA (mass %) 30.127 28.158 26.190 24.221 22.253

Figure 4.10: Two component mix graph for the formulation factors affecting the
PDI-value of UPRs.

69
 Chapter 4: Results and discussion
Design-Expert® Software

PDI 260.00
10.69

2.24

X1 = A: PA (mass %) 7.58
X2 = B: MA (mass %)
X3 = D: Maximum Temperature
255.00

Maximum Temperature (deg.C)

Actual Factors 5.92
C: Heating Rate 1 = 0.00
E: Heating Rate 2 = 0.00
F: Agitation Speed = 400.00
G: Exotherm time = slow
3.98

250.00
2.24

245.00 2.24

240.00

Actual PA (mass %) 9.243 11.212 13.180 15.149 17.117

Actual MA (mass %) 30.127 28.158 26.190 24.221 22.253

Figure 4.11: Contour mix-process graph of the interaction effect ABD, interaction
between MA:PA ratio and the maximum temperature on PDI-value of UPRs.

4.1.2.5 Preliminary conclusion: UPRs

The MA:PA mole ratio and its interaction with the maximum process temperature
were found to be the most significant factors affecting the degrees of unsaturation
and chain branching, which in turn affects the viscosity, acid-value, Mw- and PDI-
value of the UPRs.

70
Chapter 4: Results and discussion

4.1.3 Results and discussion – MVPs

Following the synthesis of the MVPs with the UPR-STY mixtures, the particle size and
particle size distribution of the MVPs were determined using the Sympatec SUCELL
HELOS/BF-OM laser diffraction sensor. The SUCELL machine determines the particle
size of the MVPs by calculating the Sauter mean diameter (SMD) and the De Brouckere
mean diameter (VMD) values, which are independent of the number of particles. The
SMD (Equation 4.1) and VMD (Equation 4.2) values are calculated based on the surface
area and mass or volume of the particles, respectively 1:

(13 23 33 ) d3
SMD : D 3,2 2.57 (4.1)
(12 22 32 ) d2

(14 24 34 ) d4
VMD : D 4,3 2.72 (4.2)
(13 23 33 ) d3

where d = diameter.

The SMD and VMD-values of the 27 MVPs runs are given in Table 4.6. These results
were also entered into the Design Expert 7 program (as with the data of the UPRs) to
determine the significance of the polyester synthesis parameters (and resulting
properties) on the particle sizes of the MVPs, via ANOVA.

It is known that a small particle has a greater surface area per unit volume compared to a
larger particle. Thus, due to the fact that the SMD-value is calculated based on the
surface area of the particles, relatively smaller diameter particles will contribute more
toward this value.

The VMD-value, on the other hand, is calculated based on the volume or mass of the
particles, which means that relatively large diameter particles will contribute more
toward this value. Thus, the VMD-value will always be greater than the SMD-value for a
given group of particles.

71
Chapter 4: Results and discussion

Table 4.6: Particle size results of MVPs.

RUN SMD (µm) VMD (µm)
1 21.03 67.43
2 23.23 48.32
3 30.15 79.13
4 29.93 57.97
5 20.03 50.57
6 40.33 61.15
7 25.03 69.45
8 21.14 65.49
9 28.47 57.06
10 28.62 79.00
11 20.17 55.13
12 15.87 56.35
13 17.84 41.47
14 17.75 40.17
15 26.64 71.76
16 29.35 70.49
17 31.61 98.36
18 18.94 67.06
19 22.10 63.00
20 25.28 68.25
21 26.89 84.01
22 34.07 117.44
23 21.09 67.70
24 29.92 78.91
25 33.58 71.06
26 12.72 30.31
27 22.28 64.59

72
Chapter 4: Results and discussion

4.1.3.1 SMD-values

Table 4.6 shows that the SMD-values of the MVPs ranged from 12.72 – 40.33 μm (mean
= 24.97 μm ± 4.67). Table 4.7 shows the significant factors (and interaction between
factors) affecting the SMD-value of the MVPs.

Table 4.7: Significant factors affecting SMD-value of the MVPs.

Factors Significance (p-value: Prob > F)
AE 0.0198
BD 0.0467
BE 0.0414

Table 4.7 shows that the interactions AE, BD and BE i.e., interactions between %PA and
Heating rate 2, %MA and maximum process temperature, and %MA and exotherm rate
are the significant terms affecting the SMD-value of the MVPs.

Figures 4.12 and 4.13 show the two component mix and contour mix-process graphs
where the effects of the MA:PA mole ratio and the interaction between %MA and
maximum process temperature of the UPRs, on the SMD-value of the MVPs are
presented. It can be seen from Figure 4.12 that an increase in the %PA of the UPRs leads
to an increase in the SMD-value of the MVPs. Figure 4.13 shows that an increase in the
%PA and decrease in the maximum process temperature of the UPRs lead to an increase
in the SMD-value of the MVPs.

73
 Chapter 4: Results and discussion
Software

ss %)
ss %) 27.1

1 = 0.00
= 250.00
2 = 0.00
ed = 400.00
e = slow 25.05
SMD-value (micrometers)

23

20.95

18.9

Actual PA (mass %) 9.243 11.212 13.180 15.149 17.117

Actual MA (mass %) 30.127 28.158 26.190 24.221 22.253

Figure 4.12: Two component mix graph for the formulation factors of the UPRs
affecting SMD-value of the MVPs.

74
 Chapter 4: Results and discussion
Design-Expert® Software

SMD 260.00
27

18
19.00
X1 = A: PA (mass %)
X2 = B: MA (mass %)
X3 = D: Max. Temp.
20.77
255.00

Maximum Temperature (deg. C)

Actual Factors
C: Heating Rate 1 = 0.00
E: Heating Rate 2 = 0.00
F: Agitation Speed = 400.00 22.53
G: Exotherm time = slow

250.00 26.07

24.30

245.00

240.00

Actual PA (mass %) 9.243 11.212 13.180 15.149 17.117

Actual MA (mass %) 30.127 28.158 26.190 24.221 22.253

Figure 4.13: Contour mix-process graph of the interaction effect BD, i.e. interaction
between %MA and the maximum temperature of the UPRs on SMD-value of
MVPs.

The droplet or particle size during the suspension polymerization of MVPs is determined
by the droplet break-up – coalescence equilibrium which is affected by the viscosity and
surface tension of the oil droplet dispersed phase. In the suspension polymerization of
MVPs, the dispersed phase comprises the UPR-STY mixture, LMA and DETA. The
viscosity and surface tension of the dispersed phase is determined by the inherent
viscosity of the UPR-STY mixture and the neutralization reaction between the carboxyl
end-groups of the UPRs and DETA, i.e. a high viscosity mixture is achieved when UPRs
with a high concentration of carboxyl end-groups are neutralized by the ammonium
groups of the DETA. Thus a high SMD-value (high surface area particles with small
diameter) will be the result of a UPR-STY mixture with a low viscosity and low acid-

75
Chapter 4: Results and discussion

value, which can be achieved with a high %PA (or low %MA) in the MA:PA mole ratio,
and low maximum reaction temperature, which has been shown in the Section 4.1.2.

4.1.3.2 VMD-values

Table 4.6 also shows that the VMD-values of the MVPs ranged from 30.31 – 117.44 μm
(mean = 64.599 μm ± 11.12). Table 4.8 shows the factors (and interaction between
factors) affecting the VMD-value of the MVPs.

Table 4.8: Significant factors affecting VMD-value of the MVPs.

Factors Significance (p-value: Prob > F)
AB 0.005
AF 0.0466
BD 0.0071
BE 0.0035

Table 4.8 shows that the most significant factors affecting the VMD-values of the MVPs
are the MA:PA mole ratio (AB), and the interactions BE and BD, i.e. interactions
between %MA and heating rate 2 and interaction between %MA and maximum process
temperature.

Figures 4.14 and 4.15 show the two component mix and contour mix-process graphs
where the effects of the MA:PA mole ratio and the interaction between the %MA and
maximum process temperature of the UPRs are presented. It can be seen from these
figures that an increase in %MA and decrease in the maximum reaction temperature of
the UPRs leads to a high VMD-value of the MVPs.

A high VMD-value will be the result of a UPR-STY mixture with a high viscosity and/
or high acid-value, which can be achieved by a high % MA with a low maximum
reaction temperature and agitation speed. Thus, the MA:PA mole ratio and interaction

76
 Chapter 4: Results and discussion

between % MA and maximum reaction temperature of the UPRs are highly significant
factors affecting the VMD-value of the resultant MVPs.
® Software

ss %)
ss %) 71

1 = 0.00
= 250.00
2 = 0.00
eed = 400.00
me = slow 67.25
VMD-value (micrometers)

63.5

59.75

56

Actual PA (mass %) 9.243 11.212 13.180 15.149 17.117

Actual MA (mass %) 30.127 28.158 26.190 24.221 22.253

Figure 4.14: Two component mix graph for the formulation factors of the UPRs
affecting VMD-value of the MVPs.

77
 Chapter 4: Results and discussion
Design-Expert® Software

VMD 260.00
79

50
60.83
X1 = B: MA (mass %)
X2 = A: PA (mass %)
X3 = D: Max. Temp.
255.00

Maximum Temperature (deg.C)

Actual Factors
C: Heating Rate 1 = 0.00
E: Heating Rate 2 = 0.00
F: Agitation Speed = 400.00
G: Exotherm time = slow
60.83
250.00

66.81

245.00
72.79
54.85

78.77

240.00

Actual MA (mass %) 22.253 24.221 26.190 28.158 30.127

Actual PA (mass %) 17.117 15.149 13.180 11.212 9.243

Figure 4.15: Contour mix-process graph of the interaction effect BD, i.e. interaction
between %MA and the maximum temperature of the UPRs on VMD-value of
MVPs.

Figure 4.16 shows a contour mix-process graph where the interaction effect of %MA and
heating rate 2 is presented. It can be seen from Figure 4.16 that an increase in the %MA
and decrease in the heating rate 2 of the UPRs lead to an increase in the VMD-value of
the MVPs. The slow heating rate 2 leads to improved reaction of the MA with PG, which
increases the degree of chain branching and viscosity of the UPRs (and UPR-STY
mixtures), and subsequently high VMD-values of the MVPs.

78
 Chapter 4: Results and discussion
Design-Expert® Software

VMD 1.000
79

50

X1 = B: MA (mass %)
X2 = A: PA (mass %)
X3 = E: Heating Rate 2
0.500
Actual Factors

Heating rate 2 (coded units)

C: Heating Rate 1 = 0.00
D: Max. Temp. = 250.00
F: Agitation Speed = 400.00
G: Exotherm time = slow
62.72

0.000

68.43
57.01

-0.500 74.14

79.85

-1.000

Actual MA (mass %) 22.253 24.221 26.190 28.158 30.127

Actual PA (mass %) 17.117 15.149 13.180 11.212 9.243

Figure 4.16: Contour mix-process graph of the interaction effect BE, i.e. interaction
between %MA and the heating rate 2 of the UPRs on VMD-value of MVPs.

4.1.3.3 Preliminary conclusion: MVPs

The significant factors affecting the particle sizes (SMD and VMD-values) of the MVPs
include the MA:PA mole ratio and its interaction with the maximum process temperature.
The particle size of suspension-based polymers is controlled by the droplet/particle
coalescence-break-up equilibrium during the polymerization. The coalescence-break-up
equilibrium is determined by the viscosity and surface tension of the oil droplets, which
are affected by the MA:PA mole ratio and maximum process temperature.

79
Chapter 4: Results and discussion

4.2 Detailed analysis of selected DoE runs

Detailed analysis of experimental runs 6 and 26 (UPRs and resulting MVPs) was
performed. These polymers were selected since the particle sizes of MVPs of these runs
are vastly different, although they have been synthesized from UPRs according to the
same formulation, but slightly different process parameters, i.e. heating rates 1 and 2,
exotherm rate and agitation speed. Table 4.9 shows the MA:PA mole ratio and process
parameters used in the synthesis of the UPRs of runs 6 and 26. Tables 4.10 and 4.11
show the properties of the two UPRs and the corresponding MVPs, respectively.

Table 4.9: Formulation and process parameters of DoE runs 6 and 26.
RUN PA MA Actual Exotherm Maximum Actual Maximum Agitation
(%) (%) heating rate exotherm heating temperature speed
rate 1 temperature rate 2 (°C) (rpm)
(˚C/min) (°C) (˚C/min)
6 17.12 22.25 4.0 slow 147.4 1.1 240 500
26 17.12 22.25 6.0 fast 174.0 0.8 240 300

Table 4.10: Properties of UPRs of DoE runs 6 and 26.

RUN Viscosity Acid-value Mn Mw PDI Tg Degree of
(cPs) (mgKOH/g) (Dalton) (Dalton) (°C) branching#

6 1509 28.15 1351 4979 3.68 -5.59 0.47

26 1207 28.92 1246 3288 2.64 -6.50 0.47
#
Degree of chain branching was determined by 1H NMR (Section 4.5).

Table 4.11: Properties of MVPs with UPRs of DoE runs 6 and 26.
RUN SMD (µm) VMD (µm)
6 40.33 61.15
26 12.72 30.31

Table 4.10 shows that RUN 6 has a greater viscosity and molecular weight and molecular
weight distribution compared to RUN 26. RUN 6 was processed with a higher agitation

80
Chapter 4: Results and discussion

speed and heating rate 2 which leads to an increase in the overall reaction rate, and
therefore an increase in the viscosity, molecular weight and molecular weight
distributions.

It can also be seen from Table 4.10 that a higher Tg-value is obtained for RUN 6, which
is possibly due to the increase in the concentration of bulky PA groups incorporated in
the chains due to a (1) slower exotherm rate and (2) faster agitation speed, which lead to
improved reaction between PA and PG, compared to RUN 26. The improved PA
incorporation results in reduced main-chain motion and hence a higher Tg-value of the
UPR compared to RUN 26.

Table 4.10 also shows that the UPRs of runs 6 and 26 have very similar acid-values and
degree of chain branching, which means that the carboxylic acid distribution of the
polyester chains is similar (see Section 4.5). Consequently, when DETA is added to the
UPRs (during the synthesis of the corresponding MVPs) they will have similar
concentrations of neutralized carboxylic groups and therefore similar micelle
concentrations in the organic phase (see Section 2.3.3.1), which in turn leads to similar
concentrations of the aqueous voids (and degrees of vesiculation) in the organic phase
droplets in the dispersion stage (see Section 3.5).

Figures 4.17(b) and 4.18(b) show the SEM images of the microtomed MVPs of runs 6
and 26, respectively, illustrating the degree of vesiculation of the MVPs. The MVPs of
runs 6 and 26 have similar degrees of vesiculation due the fact that the UPRs have
similar acid-values and degrees of chain branching as they were synthesized at the same
maximum temperature (maximum process temperature and its interaction with MA:PA
mole ratio determine the degree of chain branching and acid-value) (see Section 4.1.2).

81
Chapter 4: Results and discussion

a) b)

Figure 4.17: SEM images of (a) MVPs of RUN 6 at 100× magnification and (b)
microtomed MVPs of RUN 6 at 1790 × magnification.

a) b)

Figure 4.18: SEM images of (a) MVPs of RUN 26 at 100× magnification and (b)
microtomed MVPs of RUN 26 at 735 × magnification.

82
Chapter 4: Results and discussion

Table 4.11 also indicates that the high molecular weight/ viscosity of the UPR of RUN 6
results in larger particle sizes (SMD and VMD-values) of the MVPs compared to RUN
26. The SEM images in Figures 4.17(a) and 4.18(a) confirm the larger particle size of the
MVPs of RUN 6.

Figure 4.19 shows the bimodal particle surface area distributions (“Density”
distributions) and cumulative distributions of the particle sizes of the MVPs of runs 6 and
26 determined by light scattering. It can be seen from Figure 4.20 that the MVPs of RUN
6 have a greater percentage of larger particles, indicated by a greater intensity of the
larger particle size region of the distribution. The SEM images in figures 4.17 and 4.18
also indicate the bimodal nature of the particle size distributions of runs 6 and 26.

The MVPs of RUN 6 have larger particle sizes due to the high viscosity of the polyester/
organic phase (during the suspension polymerization of the MVPs) which reduces the
droplet break-up, resulting in larger dispersed droplets in the aqueous phase, and
ultimately larger MVPs.

The effect of processing the UPRs at different maximum process temperatures on the
properties of the UPRs and MVPs is investigated in Section 4.3.

83
Chapter 4: Results and discussion

100.0 0.84
97.5 0.82
95.0 0.80

92.5 0.78

90.0 0.76

87.5 0.74
0.72
85.0
0.70
82.5
0.68
80.0
0.66
77.5
0.64
75.0
0.62
72.5
0.60
70.0
0.58
67.5
0.56
65.0
0.54
62.5 0.52
60.0 0.50
Cumulative distribution Q2 / %

57.5 0.48

Density distribution q2*

55.0 0.46
52.5 0.44
50.0 0.42
47.5 0.40
45.0 0.38
42.5 0.36

40.0 0.34

37.5 0.32
0.30
35.0
0.28
32.5
0.26
30.0
0.24
27.5
0.22
25.0
0.20
22.5
0.18
20.0
0.16
17.5
0.14
15.0
0.12
12.5 0.10
10.0 0.08
7.5 0.06
5.0 0.04
2.5 0.02
0.0 0.00
0.4 0.6 0.8 1.0 2 4 6 8 10 20 40 60 80 100 200
X / µm

Figure 4.19: Surface area and cumulative distributions of the particle sizes of MVPs of DoE run 6 (red) and DoE run 26 (blue).

84
Chapter 4: Results and discussion

4.3 Effect of UPR maximum processing temperature on properties of the UPR

and the resultant MVPs.

Fairly anomalous results were obtained with regards to the particle size values (SMD and
VMD) of the MVPs in Section 4.1.3. ANOVA evaluation of particle sizes of these MVPs
revealed that a high SMD-value can be achieved with a low UPR maximum reaction
temperature, whilst a high VMD-value can be achieved with a high UPR maximum
temperature. To investigate the individual effect of the UPR maximum temperature on
the particle size (and other properties) of the MVPs, a series of UPRs and MVPs were
synthesized with a variation in the maximum reaction temperature while the other
parameters are kept constant.

The UPR series were synthesized according to the standard formulation (See Table 3.1)
and experimental procedure (i.e. standard heating rates, agitation, and exotherm time/
temperature) (see Section 3.3) but with different maximum temperatures, i.e. 220, 230,
240, 250, and 260 °C. These UPRs were also analyzed for viscosity, acid-value,
molecular weights (Mn, Mw), PDI and Tg-values. The five resultant UPR-STY mixtures
were used to synthesize MVPs to determine the effect of the maximum process
temperature of the UPRs on the particle sizes of the MVPs. The properties of the five
UPR-STY mixtures are given in Table 4.12.

Table 4.12: Effect of UPR maximum processing temperature on properties of the

UPRs of runs 29 – 33.
RUN Maximum Viscosity Acid-value Mn Mw PDI Tg Degree of
temperature (cPs) (mgKOH/g) (Dalton) (Dalton) (°C) branching*
(°C)
29 220 791.8 65.09 746 1556 2.08 -8.77 0.323
30 230 566.3 57.83 821 1783 2.17 -6.00 0.299
31 240 777.4 50.84 992 2406 2.43 0.01 0.387
32 250 861.4 41.82 1081 3328 3.08 4.56 0.407
33 260 1346 39.96 1273 6950 5.46 6.96 0.413
*Degree of chain branching (DB) of the UPRs was determined by 1H NMR (see Section 4.5).

85
Chapter 4: Results and discussion

Table 4.12 shows that an increase in the polyester maximum process temperature results
in an increase in the viscosity, Mn-, Mw-, PDI-, Tg-values, degree of chain branching and
a decrease in the acid-value of the UPR-STY mixtures.

The increase in the viscosity can be attributed to an increase in molecular weight, PDI
value and the degree of chain branching of the UPR chains (see Section 4.5). Figure 4.20
shows the molecular weight distributions of the five UPRs obtained by SEC.

RUN 29
RUN 30
1.0 RUN 31
RUN 32
RUN 33

0.8
Normalized RI signal

0.6

0.4

0.2

2 3 4 5 6 7
Log M

Figure 4.20: SEC chromatograms of UPRs prepared at different maximum

processing temperatures.

The decrease in the acid-value of the UPRs can be explained by the fact that the
carboxylic acid concentration decreases as the polyesterification reaction progresses.
However, the degree of chain-branching increases as the polyesterification reaction
progresses at a higher process temperature, which means that the carboxylic acid
concentration per chain increases. In other words, the concentration of carboxylic acid

86
Chapter 4: Results and discussion

groups of the individual chains increases due to the acid end groups of the branches/side
chains. The schematic diagram in Figure 4.21 illustrates this phenomenon.

COOH

HOOC
COOH

COOH

Low degree of chain branching

HOOC COOH

HOOC
COOH

HOOC COOH

High degree of chain branching

Figure 4.21: Schematic diagram of the effect of chain branching on the acid-value of
the UPRs.

The increase in the Tg values of the UPRs may be attributed to the increased molecular
weight, degree of chain-branching as well as the increased incorporation of the bulky
phthalate groups in the chains at higher temperatures. These factors decrease the main-
chain mobility of the UPRs, which leads to an increase the Tg value. Figure 4.22 shows
the DSC thermograms with the Tg values of the UPRs.

87
Chapter 4: Results and discussion

35
RUN 29
RUN 30
RUN 31
RUN 32
RUN 33
Heat Flow (mW/g)

30

25

20
-40 -20 0 20 40

Temperature ( C)

Figure 4.22: DSC curves of the UPRs prepared at different maximum processing
temperatures.

Table 4.13: Effect of UPR maximum processing temperature of runs 29 – 33 on

properties of the MVPs.

RUN SMD (µm) VMD (µm) Contrast ratio

29 11.76 53.24 0.17
30 8.47 41.64 0.30
31 13.93 47.35 0.24
32 28.87 56.28 0.57
33 37.58 93.65 0.61

It is known that an increase in the maximum processing temperature of the polyester

leads to an increase in the percentage chain branching of the polyester. An increase in the
chain branching leads to an increase in the concentration of carboxylic groups per chain

88
Chapter 4: Results and discussion

and therefore an increase in the concentration of neutralized carboxylic groups per main
chain with the addition of DETA to the polyester mixture (see Section 2.3.3.1). The
increased concentration of neutralized groups leads to an increase in the concentration of
micelles in the organic phase, which in turn leads to an increase in the concentration of
the aqueous voids in the organic phase droplets, and eventually an increase in the
concentration of vesicles or degree of vesiculation of the dried MVPs. In other words, the
concentration of vesicles or degree of vesiculation increases as the polyester chain
branching increases, which in turn increases as the polyester maximum processing
temperature increases. Figure 4.23 shows the SEM micrographs of microtomed MVPs of
experimental runs 29 and 33, indicating the difference in the degree of vesiculation of the
runs.

Figure 4.23(a) clearly shows that the MVPs of RUN 29 have a lower degree of
vesiculation compared to those of RUN 33 in Figure 4.23(b). RUN 29 has a low degree
of vesiculation due to the lower degree of chain-branching of the polyester chains,
resulting in a low concentration of neutralized carboxylic groups (micelles). On the other
hand, the MVPs of RUN 33 has a higher degree of vesiculation due to a high percentage
of chain-branching, as a result of a high polyester processing temperature.

a) b)

Figure 4.23: SEM micrograph of microtomed MVPs of a) RUN 29 at 622×

magnification and (b) RUN 33 at 913× magnification.

89
Chapter 4: Results and discussion

Table 4.13 indicates that RUN 33 has a higher contrast ratio/ opacity value (at the same
film wet-film thickness) than RUN 29. The opacity of RUN 33 is higher due to the higher
degree of vesiculation and light scattering ability of the MVPs (see Section 2.3.1).

Table 4.13 also shows that the UPR maximum temperature indirectly affects the particle
size of MVPs. As previously concluded from Section 4.1.3, a low maximum reaction
temperature with a high %PA or a high maximum temperature with a high %MA of the
UPRs, can lead to a large particle size (SMD or VMD-values) under certain
circumstances. Table 4.13, however, shows that an increase in the UPR maximum
temperature leads to an overall increase in the particle size (SMD and VMD values) of
MVPs at a MA:PA mole ratio = 26.19:13.18.

Figure 4.24 illustrates the bimodal particle surface area distributions of MVP runs 29 and
33. It can be seen from Figure 4.24 that the MVPs of RUN 33 have a greater percentage
of larger particles compared to smaller particles, indicated by a greater intensity of the
larger particle size region of the distribution.

The larger particle size of the MVPs of RUN 33 can be explained by the high viscosity of
the UPR processed at a higher maximum temperature. The UPR of RUN 33 has a higher
viscosity than RUN 29 due to increased DPn and DB-values at the high maximum
process temperature. The higher viscosity of the UPR reduces the droplet break-up and
increases the probability of droplet coalescence, resulting in larger dispersed droplets in
the aqueous phase, and ultimately larger MVPs.

90
Chapter 4: Results and discussion

100.0 1.075
97.5 1.050
95.0 1.025

92.5 1.000
0.975
90.0
0.950
87.5
0.925
85.0
0.900
82.5
0.875
80.0
0.850
77.5
0.825
75.0 0.800
72.5 0.775
70.0 0.750
67.5 0.725
65.0 0.700
62.5 0.675

60.0 0.650
Cumulative distribution Q2 / %

0.625
57.5

Density distribution q2*

0.600
55.0
0.575
52.5
0.550
50.0
0.525
47.5
0.500
45.0
0.475
42.5
0.450
40.0 0.425
37.5 0.400
35.0 0.375
32.5 0.350
30.0 0.325

27.5 0.300

25.0 0.275
0.250
22.5
0.225
20.0
0.200
17.5
0.175
15.0
0.150
12.5
0.125
10.0
0.100
7.5 0.075
5.0 0.050
2.5 0.025
0.0 0.000
0.4 0.6 0.8 1.0 2 4 6 8 10 20 40 60 80 100 200
X / µm

Figure 4.24: Surface area and cumulative distributions of the particle sizes of MVPs of runs 29 (red) and 33 (blue).

91
Chapter 4: Results and discussion

4.4 Determination of hardness of MVPs

Attempts were made to determine the hardness/ stiffness of the MVPs to quantitatively
determine the effect of the crosslink density, degree of vesiculation, and particle size on
the hardness of the MVPs to confirm the following postulates:
The crosslink density of the MVPs is determined by the reaction between the MA
of the UPR and STY, i.e. a high MA:PA mole ratio in the UPR leads to a high
crosslink density, which in turn results in harder MVPs. The impact resistance of
the MVPs is however compromised.
The degree of vesiculation (size and concentration of vesicles) of the MVPs is
determined by the interaction between the COOH-groups of the UPR and DETA
during the synthesis of the MVPs. Thus a high concentration of COOH-groups
leads to a high degree of vesiculation due to increased interaction with the DETA.
The degree of vesiculation would affect the hardness and impact resistance of the
MVPs, e.g. highly porous MVPs will have lower impact resistance compared to
those with low porosity.
MVPs with different particle sizes will have different hardness values.

The microhardness tester and Atomic Force Microscopy (AFM) in contact mode were
used to determine the hardness of the MVPs.

As previously mentioned in Section 3.6.11, the microhardness testing of the dried MVPs
involved indenting the individual MVPs with the micro-indenter. The MVPs were
sprinkled onto the surface of a liquid alkyd resin before it was allowed to cure
completely. The resin was cured to keep the MVPs stationary during testing. The
sprinkling of the MVPs onto the resin surface meant that not all the MVPs were at the
same depth at the resin surface. Thus, when the micro-indenter was applied to the MVPs,
erratic results were obtained. In other words, erratic microhardness values were obtained
as the surfaces of the MVP-resin disks were not homogeneous.

92
Chapter 4: Results and discussion

AFM was found to be effective to determine the hardness of the MVPs smaller than 1
micron, although the mean particle size of the MVPs is approximately 25μm. In other
words, AFM is not a suitable technique to determine the hardness of MVPs due to their
relatively large particle size. AFM was however performed on the sub-micron sized
MVPs of runs 29 (low degree of vesiculation) and 33 (high degree of vesiculation) to
determine the effect of degree of vesiculation on the hardness of the MVPs. For example,
Figures 4.26 and 4.27 show the AFM software-generated 3D-images and “Force-
Distance” curves of sections of individual MVPs of runs 29 and 33, respectively. The
force-distance curves approaches from -1μm towards 20nm on the z-axis as the cantilever
tip moves towards the MVPs. Adhesion between the cantilever tip and the MVP is
indicated by the sharp dip in the retracting curve. The cantilever is thereafter deflected
from the MVP, where the angle of the deflection slope indicates the hardness of the
MVP.

It can be seen from Figures 4.25(b) and 4.26(b) that the deflection angle (of the slope) of
the approaching curve of the MVP from RUN 29 (45.94°) is greater than that of RUN 33
(33.89°), indicating that the MVP from RUN 29 is relatively harder than that of RUN 33.
The MVP of RUN 29 is possibly harder due to the lower degree of vesiculation of the
MVPs.

approaching

retracting

Figure 4.25: AFM 3D image (a) and force-distance curve (b) of a sub-micron MVP
of RUN 29.

93
Chapter 4: Results and discussion

Figure 4.26: AFM 3D-image (a) and force-distance curve (b) of a sub-micron MVP
of RUN 33.

4.5 Chain-branching

As mentioned in Section 4.1, the Ordelt reaction is the main side reaction in the
polyesterification reaction. The Ordelt reaction involves the double bonds of the maleate
or fumarate component becoming saturated and leading to branched structures or side
chains. The degree of branching can be determined by size-exclusion chromatography
using the PDI as a measure of the branching extent, whereby the greater the PDI, the
greater the branching extent 2. Zetterlund et. al. 3, on the other hand, studied the Ordelt
reaction by 1H NMR by using the work of Paci et. al. 4 and Judas et. al. 5 together with
the NMR spectra of polyesters with only PG and PA (Figure 4.27) as well as polyesters
with only PG and MA (Figures 4.28). It was concluded that the shoulder of the resonance
peak at 5.2 ppm in Figure 4.28 was related to the Ordelt reaction, as no shoulder was
observed for the resonance peak in the spectrum of PA in Figure 4.27.

94
Chapter 4: Results and discussion

Figure 4.27: 1H NMR spectrum of Figure 4.28: 1H NMR spectrum of

polyester with phthalic anhydride 3. polyester with maleic anhydride 3.

Holter et. al. 6 derived an equation to determine the degree of branching based ratio of the
1
H NMR signal intensities of branched (due to the Ordelt reaction) units and the linear
units of the unsaturated polyester chains. The ratio is as follows:

2D
Degree of Branching (DB) (4.3)
2D L
Where
D = NMR intensity of a branched unit in the polyester chains
L = NMR intensity of a linear unit in the polyester chains

The DB values were determined by using the 1H NMR intensities of the shoulder of the
methine peak at 5.2 – 5.4 ppm as D and the methine peak at 5.1 – 5.2 ppm as L in
Equation 4.3 (e.g. 1H NMR spectrum of the UPR of RUN 29 in Figure 4.29). Table 4.14
shows the DB- and PDI-values obtained from 1H NMR and SEC of runs 29 to 33,
respectively.

95
Chapter 4: Results and discussion

l
j
c

L
b D f
a g h
e d
i k

8 6 4 2 0
(ppm)

Figure 4.29: 1H NMR spectrum of UPR of RUN 29.

Table 4.14: DB results of UPRs of runs 29 and 33 via 1H NMR and SEC.
1
RUN H NMR SEC
Intensity of Intensity of DB PDI
methine unit branched unit
(g) (e)
DoE 6 11.877 5.160 0.465 3.68
DoE 26 11.779 5.188 0.468 2.64
29 2.040 0.486 0.323 2.08
30 1.488 0.318 0.299 2.17
31 2.779 0.877 0.387 2.43
32 2.976 1.022 0.407 3.08
33 0.577 0.203 0.413 5.46

96
Chapter 4: Results and discussion

Table 4.14 clearly shows that the DB values of the UPRs of DoE runs 6 and 26 are very
similar, because they were synthesized at the same maximum temperature (240 °C).
However, the DB values of the UPRs from runs 29 – 33 increase from run 29 (0.323) to
33 (0.413) due to the increased probability of the Ordelt reaction (chain-branching) at
higher maximum processing temperatures of the UPRs.

The DB results of runs 29 to 33 are in good agreement with the PDI values obtained from
SEC. The PDI- and DB-values were not in good agreement for runs 6 and 26, because
these UPRs were synthesized according to different process protocols, i.e. different
exotherm rates, agitation speeds and heating rates 1 which affect the molecular weights
(and therefore PDI-values) of the UPRs. These parameters (unlike the maximum process
temperature) do not significantly affect the degree of branching of the UPRs.

4.6 End-group analysis by isocyanate derivatization and 1H NMR

Historically, numerous methods have been employed to analyze the end-groups of

7-8
polyesters, including classical titration methods , Fourier Transform Infra-red
9-10
spectroscopy (FTIR) , Matrix-Assisted-Laser-Desorption-Ionization-Mass-
Spectrometry (MALDI) 11 and High Performance Liquid Chromatography (HPLC) 12.

NMR spectroscopy is an attractive method for the quantitative and qualitative

determination of end-groups of polyesters, providing these resonances can be resolved
13-14 13
from those attributed to the backbone signals . C NMR spectroscopy have been
used to determine the end-groups of PET 13, 15, but the technique is complex and does not
have general applicability. Derivatization of the polyester chains with various chemical
19 16 31 17
compounds followed by F NMR and P NMR has also been used, but these
methods require complex experimental protocols.

The method used to determine the end-groups of the UPRs synthesized in this study is an
18
efficient and quantitative method developed by Donovan and Moad . The method

97
Chapter 4: Results and discussion

entails the characterization of the COOH and OH end-groups by 1H NMR after

derivatization of these groups by trichloroacetyl isocyanate (TAI). The TAI-derivatized
end-groups have imidic NH resonances in a clear region of the 1H NMR spectrum, i.e.
the COOH-groups have resonances at 10–11.5 ppm for C(O)–O–C(O)–NH–C(O)CCl3
groups, and the OH-groups at 8–9 ppm for O–C(O)–NH–C(O)CCl3 groups. The reaction
schemes involved are given in Figures 4.30 and 4.31.
O O
C H
O N
R OH + N
O R
O CCl 3
CCl 3

Figure 4.30: Derivatization of OH-groups of UPRs with TAI.

O O O
O
C
+ N H3C O N
H
R OH
O
CCl 3 O CCl 3

Figure 4.31: Derivatization of COOH-groups of UPRs with TAI.

The derivatization reaction is performed by dissolving a sample of the polyester (300 mg)
in deuterated chloroform CDCl3 (2 mL) in a NMR tube, after which an excess of
trichloroacetyl isocyanate (TAI) (0.5 mL) is added and the 1H NMR spectrum is recorded
for 32 scans. The derivatization reaction is instantaneous and any excess TAI, being
aprotic, causes no additional resonance signals in the spectra. Figure 4.32 shows the
FTIR spectrum which confirms the derivatization reaction of the UPR of RUN 29 with
TAI.

98
Chapter 4: Results and discussion

RUN 29
RUN 29 - derivatized
1.00
0.95
Normalized %Transmission

0.90
0.85
0.80
0.75
0.70
0.65
0.60
0.55
0.50
0.45
0.40
0.35

4000 3500 3000 2500 2000 1500 1000 500

-1
wavenumbers (cm )
Figure 4.32: FTIR spectrum of the UPRs of RUN 29 (black) and RUN 29
derivatized with TAI (red).

Figure 4.32 clearly shows that the UPR chains of RUN 29 were derivatized by TAI due
to the presence of N-H stretch vibrational frequencies at 3500-3100 cm-1 and the absence
of O-H vibrational frequency at 3500cm-1 which is present in the spectrum of the un-
derivatized sample.

Figure 4.33 shows the 1H NMR spectrum of TAI-derivatized UPR of RUN 29. It can be
seen from Figure 4.33 that the TAI-derivatized OH-groups at 9.0 ppm -and COOH-
groups at 11 ppm are in a clear region of the NMR spectrum. Integration of these
resonance signals corresponding to the OH- and COOH end-groups relative to the linear
units (e.g. fumarate unit signal at 6.8 – 7.0 ppm) provides the relative percentages of the
COOH -and OH-groups. Table 4.15 shows the relative percentages of the OH- and
COOH-groups of runs 29 to 33.

99
Chapter 4: Results and discussion

COOH
OH

12 10 8 6 4 2 0
(ppm)

Figure 4.33: 1H NMR spectrum of UPR of RUN 29 derivatized by TAI.

Table 4.15: Relative OH- and COOH-group percentages of UPRs of runs 29 to 33.
RUN Integrated - Integrated Fumarate %COOH %OH
COOH peak –OH peak

29 0.90 2.30 11.70 7.69 19.66

30 0.00 1.10 36.90 0.00 2.98
31 0.90 2.90 18.90 4.76 15.34
32 1.00 4.50 32.20 3.11 13.98
33 1.20 3.80 26.00 4.62 14.62

Table 4.15 shows that the OH- and COOH-group concentrations of the UPRs decrease as
the maximum processing temperature increases. In other words, a higher conversion is
achieved at higher processing temperatures. The latter is supported by the acid-value
results in Section 4.3.2.

100
Chapter 4: Results and discussion

Table 4.15 also indicates that the UPRs have higher concentrations of OH-groups
compared to COOH-groups, because they were synthesized with excess PG. In other
words, the probability of an OH-terminated polyester chain is higher than COOH-
terminated chain. It should be noted that the derivatization technique was not successful
for all UPR samples, e.g. the %COOH of RUN 30 could not be determined, as the
COOH-peak had an integration-value of zero in multiple 1H NMR spectrums.

4.7 References

1. Rawle, A. Basic Principles of Particle Size Analysis; Malvern Instruments

Limited: Worcestershire, pp 1 - 8.
2. Podzimek, S.; Hanus, J.; Klaban, J.; Kitzler, J. Journal of Liquid Chromatography
1990, 13, 1809.
3. Zetterlund, P. B.; Gosden, R. G.; Weaver, W.; Johnson, A. F. Polymer
International 2003, 52, 749.
4. Paci, M.; Crescenzi, V.; Supino, N. Makromolekulare Chemie 1982, 183, 377.
5. Judas, D.; Fradet, A.; Marechal, E. Makromolekulare Chemie 1983, 184, 1129.
6. Holter, D.; Burgath, A.; Frey, H. Acta Polymer 1997, 48, 30.
7. Ubranski, J.; Czerwinski, W.; Janicka, K.; Majewska, F.; Zowall, H. In Handbook
of analysis of synthetic polymers and plastics, 1st ed.; Cameron, G. G., Ed.
Halsted Press: London, 1977; pp 223 - 272.
8. Price, G. F. In Techniques of polymer characterisation, Allen, P. W., Ed.
Butterworths Scientific: London, 1959; pp 209 -229.
9. Koenig, J. L. In Spectroscopy of polymers, Elsevier: New York, 1999.
10. Al-AbdulRazzak, S.; Lofgren, E. A.; Jabarin, S. A. Polymer International 2002,
51, 174.
11. Guittard, J.; Tessier, M.; Blais, J. C.; Bolbach, G.; Rozes, L.; Marechal, E.
Journal of Mass Spectrometry 1996, 31, 1409.
12. Rissler, K. Journal of Chromatography A 1997, 786, 85.

101
Chapter 4: Results and discussion

13. Fox, B.; Moad, G.; Diepen, G. V.; Willing, R. I.; Cook, W. D. Polymer 1997, 38,
3035.
14. Bovey, F. A.; Mirau, P. NMR of polymers. In Academic Press: New York, 1996;
pp 199 - 212.
15. Petiaud, R.; Waton, H.; Pham, Q. T. Polymer 1992, 33, 3155.
16. Kenwright, A. M.; Peace, S. K.; Richards, R. W.; Bunn, A.; MacDonald, W. A.
Polymer 1999, 40, 2035.
17. Spyros, A.; Argyropoulos, D. S.; Marchessault, R. H. Macromolecules 1997, 30,
327.
18. Donovan, A. R.; Moad, G. Polymer 2005, 46, 5005.

102
 CHAPTER 5

CONCLUSIONS AND RECOMMENDATIONS

5.1 Conclusions

The first objective of the study was to investigate the effects of varying MA:PA mole
ratio and process parameters, e.g. heating rates, agitation speed, etc. on the properties of
the UPRs, via Design of Experiments (DoE).

Statistical analysis of the DoE results revealed that the process and formulation
parameters/factors (as well as interactions between the factors) had significant effects on
the basic properties (e.g. viscosity, acid-value) of the UPRs. In particular, the MA:PA
mole ratio and its interaction with the maximum process temperature significantly affect
most of the UPR properties as these factors control the degree of chain branching,
concentration of carboxyl groups and viscosity.

The second objective was to determine the molecular composition/ structure of the UPRs
and relate to specific properties of MVPs.

The molecular weights (Mn and Mw values) and molecular weight distribution (PDI) of
the UPRs were successfully determined by SEC; glass-transition temperature (Tg) by
DSC; degree of chain branching by 1H NMR; and the percentage carboxyl- and hydroxyl
end-groups via isocyanate derivatization coupled with 1H NMR.

The particle sizes (SMD and VMD-values) of the MVPs were found to be significantly
affected by viscosity, molecular weights and molecular weight distribution of the UPRs.
The properties mentioned are affected by the MA:PA mole ratio and its interaction with
the maximum process temperature, due to the fact that the particle size of suspension-
based polymers is controlled by the droplet/particle coalescence-break-up equilibrium
during the polymerization. The coalescence-break-up equilibrium is determined by the

103
Chapter 5: Conclusions and recommendations

viscosity and surface tension of the oil droplets, which are affected by the MA:PA mole
ratio and maximum process temperature.

It was found that UPRs synthesized with a low %MA and low maximum reaction
temperature can result in MVPs with a high SMD-value, and those synthesized with a
high %MA and high maximum reaction temperature can result in MVPs with a high
VMD-value.

To determine the effect of maximum reaction temperature (at fixed MA:PA mole ratio)
on the properties of the UPRs and MVPs, a detailed evaluation was conducted. The
evaluation revealed that an increase in the magnitude of the maximum process
temperature results in an increase in the viscosity, molecular weight (Mn and Mw values)
and molecular weight distribution (PDI), Tg value and degree of chain branching; and a
decrease in the acid-value of the UPRs.

The evaluation also revealed that an increase in the maximum process temperature (at
fixed MA:PA mole ratio) of the UPRs resulted in larger MVPs in terms of their SMD and
VMD-values. Furthermore, it was found that the degree of vesiculation (or opacity) of
the MVPs is dependent on the MA:PA mole ratio and maximum reaction temperature in
the UPR synthesis. These parameters control the distribution of the carboxylic groups
(and degree of branching) of the UPR chains, and therefore their interaction with DETA,
and the formation of the MVP vesicles.

In addition, an attempt was made to determine the hardness of the MVPs using the
microhardness tester and AFM. The relative hardness of the MVPs were determined to
establish a relationship between the properties of the hardness and the crosslink density,
particle size and degree of vesiculation of the MVPs.

Unfortunately, the microhardness testing was found to be unsuccessful due to the fact
that not all the MVPs were homogeneously distributed at the surface of the cured resin,
which led to erratic results during testing.

104
Chapter 5: Conclusions and recommendations

AFM was only found to be effective to determine the hardness of MVPs smaller than
1μm, which means that the hardness of MVPs with an average particle size e.g. 25μm
could not be determined.

5.2 Recommendations for future work

As mentioned in Section 4.1, the screening design involved a minimum of experimental

points to determine the main effects affecting the properties of the UPRs and MVPs.
Future work may include augmenting the design, i.e. broadening the experimental space
by adding additional experiments to the existing data to be able to optimize the synthesis
and properties of the UPRs and MVPs.

As mentioned in Section 4.5, the end-group analysis via isocyanate derivatization

coupled with 1H NMR was not successful for all samples. Future work may include
19 31
investigating alternative techniques e.g. F NMR and P NMR to quantitatively
determine the hydroxyl -and carboxyl end-groups of the UPRs.

Great difficulty was experienced in determining the physical properties (e.g. hardness) of
the MVPs. AFM and microhardness testing were unsuccessful, but other mechanical tests
may be attempted in the future to determine the physical properties of MVPs.

105