. » v mm. .u

7

:7
d 2". |
‘99:
3

5"... not"
”was. ..
‘I

If

. ’1
, .--.t’

.'?1

“f

 

 

’H 58”

{Q

LIBRARE

llUlllUlHllllllllllllllllllllllllllllllllll

3 1293 01566 084

ll “

 

 

 

 

 

 

 

LIBRARY

Michigan State
University

 

 

 

This is to certify that the
thesis entitled

Polymers from Renewable Resources:
Cationic Photopolymerization of Epoxidized
Natural Oils

presented by

Shailender Kumar Moorjani

has been accepted towards fulfillment
of the requirements for

M.S. Chemical Engineering

degree in

 

 

W

Major professor

DateflufifléLsTJflé.

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

 

PLACE IN RETURN BOX to remove thle checkout from your record.
TO AVOID FINES return on or bdore date due.

   
 
  
 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

M—

 

 

 

 

 

 

 

C

 

 

 

 

 

 

 

 

 

 

1
1%

[—_l

MSU le An Nflnnetive Action/Equei Opportunity lnetltuion
W

 

 

m1

-———-———'=——

POLYMERS FROM RENEWABLE RESOURCES:
CATIONIC PHOTOPOLYMERIZATION OF
EPOXIDIZED NATURAL OILS

By

Shailender Kumar Moorj ani

A DISSERTATION

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

MASTER OF SCIENCE

Department of Chemical Engineering

1996

ABSTRACT

POLYMERS FROM RENEWABLE RESOURCES:
CATIONIC PHOTOPOLYMERIZATION OF EPOXIDIZED NATURAL OILS

By

Shailender Kumar Moorjani

The interest in producing polymeric materials from renewable feedstocks is
motivated by several factors including environmental concerns, economic development of
rural areas, and the diversification of the resource base away from the traditional
petroleum feedstocks. Cationic photopolymerization of epoxidized oils has tremendous
potential for the development of polymer films as well as composite structures. Cationic
polymerizations form highly crosslinked polymers which exhibit excellent mechanical
and thermal properties. In spite of these advantages, these chain polymerizations have
not been well characterized. In this contribution, differential scanning photocalorimetry
has been used to obtain important kinetic information on the photopolymerization of
epoxidized soybean oil. Anthracene was used as a photosensitizer with a diaryliodonium
hexafluoroantimonate salt to initiate the polymerizations using UV light. The effect of
temperature and light intensity on the effective propagation rate constant for
polymerization has been characterized. Steady state fluorescence monitoring was
employed to study the effect of viscosity on the diffusion controlled kinetics, and

correlative model has been developed to describe this effect.

To God and my Parents.

iii

ACKNOWLEDGEMENTS

I would like to express my gratitude and appreciation to all the people who have
contributed to this thesis and made my period of stay at Michigan State University most
rewarding.

First of all, I am grateful to my mentor at MSU, Professor Alec Scranton, for his
guidance, encouragement and support (financial and otherwise) throughout our period of
interaction. I cannot thank Dr. Bharath Rangarajan enough for sharing his insight and
knowledge in fields as diverse as chemical engineering and the Wall Street transactions.

I would like to acknowledge the encouragement and the amusement provided by
Arvind Mathur and Lindsay Coons, almost respectively. I appreciate the help I received
from my peer, Julie Jessop, both inside and outside the laboratory.

1 would also like to thank Dr. Kris Berglund and Dr. Robert Ofoli for their interest
in my work, and for taking the time and effort to be on my thesis committee. I am also
thankful to Julie Caywood, Candy McMaster, and Faith Peterson, at the Department of

Chemical Engineering for their administrative assistance.

iv

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................ viii
LIST OF FIGURES ............................................................................................................ ix
CHAPTER I.

INTRODUCTION AND MOTIVATION ............................................................................ 1
1.1. Issues for the Development of UV Curable Polymers .............................................. l
1.2. UV-Initiated Photopolymerizations ......................................................................... 3
1.3. UV-Sensitive Initiators for Cationic Photopolymerizations ..................................... 5
1.4. Motivation to Use Agricultural Feedstocks as Starting Materials ............................ 6
1.5. List of References ..................................................................................................... 8

CHAPTER II.

BACKGROUND ............................................................................................................... 10
H. 1. Overview of Polymerization Mechanisms ............................................................ 10
11.2. Properties of Cationic Polymerizations. ................................................................ 11
11.3. Curing of Epoxy Resins ......................................................................................... 12

11.3.1. Conventional Methods ................................................................................... l2
H.3.2. Ring Opening Polymerizations ....................................................................... 13
11.3.3. Kinetics of Ring Opening Polymerizations .................................................... 15
11.3.4. Effect of temperature on reaction rate and degree of polymerization ............ 15
11.4. Photopolymerizations ............................................................................................ 17
11.4.1. Free Radical Photopolymerizations ................................................................ 17
11.4.2. Cationic Photopolymerizations. ..................................................................... 18
H5. UV-Sensitive Initiators for Cationic Photopolymerizations .................................. 19
11.6. Advantages of Cationic Photopolymerizations ..................................................... 19
11.7. Photopolymerizations of Epoxides and Vinyl Ethers ............................................. 20
11.8. Cationic Photoinitiators ......................................................................................... 22
11.9. General Characteristics of Onium Salts ................................................................ 25
H. 10. Photosensitization ................................................................................................ 26
11.11. Polymerization Characteristics of Onium Salts ................................................... 29
11.12. Epoxidized Soybean Oil: An Agri-based Source of Polymerizable Epoxides...30
II.12.1. Motivation to Use Agricultural Feedstocks as Starting Materials ................ 30
1112.2. Traditional Methods of Producing Polymeric Films and Coatings
from Natural Oils. ......................................................................................... 31
H. 12.3. Reactions in Natural Drying Oils. ................................................................ 32
H.12.4. Reactions of Functionalized Oils .................................................................. 33
11.13. Need for New Polymerization Methods Based Upon Natural Oils ..................... 36

V

11.14. Cationic Polymerizations of Epoxidized Oils. .................................................... 37

H.15. List of References ................................................................................................ 42
CHAPTER 111.
OBJECTIVES OF RESEARCH ........................................................................................ 45
CHAPTER IV.
KINETIC STUDIES OF PHOTOPOLYMERIZATION OF EPOXIDIZED NATURAL
TRIGLYCERIDES ............................................................................................................ 47
1V. 1. Introduction .............................................................................................................. 47
1V.2. Experimental ............................................................................................................ 53
1V.2.1. Materials ........................................................................................................... 53
IV .2.2. Absorbance and Calorimetric Experiments ...................................................... 53
1V .3. Results and Discussion ............................................................................................ 55
IV.3.1. Proposed Mechanism ........................................................................................ 55
IV.3.2. Absorbance Results ........................................................................................... 57
IV.3.3. Calculation of Termination Rate Constants ...................................................... 58
IV.3.4. Calculation of Initiation Kinetic Constants ...................................................... 59
IV.3.5. Effect of Light Intensity .................................................................................... 6O
IV.3.6. Effect of Temperature ....................................................................................... 62
IV.4. Conclusions .............................................................................................................. 67
1V .5. List of References .................................................................................................... 85
CHAPTER V.
THE EFFECT OF VISCOSITY ON THE RATE OF PHOTOSENSITIZATION OF
DIARYLIODONIUM SALTS BY ANTHRACENE ........................................................ 87
V1 Introduction ................................................ 87
V2. Experimental ............................................................................................................. 90
V21. Materials ............................................................................................................. 90
V.2.2. Fluorescence Measurements. ............................................................................. 91
V3. Results and Discussion .............................................................................................. 92
V3. 1. Steady State Fluorescence Measurements .......................................................... 92
V.3.2. Analysis of the Photosensitization Mechanism .................................................. 94
V33. Simulation of the Photosensitization Mechanism .............................................. 95
V34. Diffusion Limited Reactions - The Smoluchowski Theory ............................... 98
V4. Conclusions ............................................................................................................. 102
V5. List of References ................................................................................................... 114
CHAPTER V.
CONCLUSIONS AND RECOMMENDATIONS .......................................................... 116
V11.1. Summary of Results .............................................................................................. 116
V11.1.1. Differential Scanning Photocalorimetry Studies ........................................... 117
V1.1.2. Steady State Fluorescence Studies .................................................................. 118
V1.2. Recommendations for Future Work ...................................................................... 119
V1.2. 1. Alternate Non toxic Photosensitizers .............................................................. 120

vi

V1.2.2. Fluorescence and Phosphorescence lifetime experiments. ............................. 122

V1.2.3. Epoxidation and subsequent photopolymerization of other natural oils ......... 122
V1.2.4. Hydrolysis of epoxidized oils or epoxidation of hydrolyzed oils before
polymerization. .............................................................................................. 123
V1.2.5. Mechanical and thermal characterization of polymers obtained from
photopolymerizations of epoxidized oils. ...................................................... 123
V1.2.6. Accelerated weathering tests on polymers ...................................................... 125
V1.2.7. Photooxidation of polymers produced by UV light ........................................ 125
V1.3. List of References .................................................................................................. 129

vii

LIST OF TABLES

Table V1.1. Comparison of Soybean and Corn Oil ......................................................... 127

viii

Figure 1.1.

Figure 1.2.

I Figure IV.1

Figure 1V.2.

Figure 1V.3.
Figure IV.4.

Figure IV.5.

Figure IV.6.

Figure 1V.7.

Figure 1V.8.

Figure IV.9

LIST OF FIGURES

Chapter 1
Average chemical structure of the epoxidized soybean oil molecule .......... 4O
Schematic of the crosslinking polymerization of epoxidized soybean oil. .41
Chapter 1]
Experimental setup for Differential Scanning Photocalorimetry Studies. ..71

Typical Exotherm Obtained During Polymerization of Epoxidized Soybean
Oil (ESO) ..................................................................................................... 72

Corrected exotherrn obtained from the DSC after subtracting the baseline.73
Effect of epoxidation on the UV and visible absorbance of soybean oil.....74

Spectral absorbance of diaryliodonium initiator and anthracene photo-
sensitizer as a function of wavelength ......................................................... 75

Emission scan from the Hg(Xe) lamp source used for DSP experiments. ..76

Initiation rate constant as a function of temperature at different UV light
intensities. .................................................................................................... 77

DSP exotherms for the cationic photopolymerization of ESO at different
incident light intensities ............................................................................... 78

Overall rate of propagation versus time for E80 polymerization with 1%
diaryliodonium hexafluoroantimonate photosensitized by 0.01%
anthracene .................................................................................................... 79

Figure IV. 10.Conversion profile of epoxy moieties in ESO during photopolymerization at

various temperatures .................................................................................... 80

Figure 1V.1 1. kp[M+] as a function of conversion for E80 photopolymerization at various

temperatures. ............................................................................................... 8 1

Figure IV. 12. Apparant propagation rate constant l<p as a function of conversion for E80

photopolymerization at various temperatures .............................................. 82

Figure IV.13. Viscosity of epoxidized soybean oil as a function of temperature. ............. 83

ix

Figure IV. 14. Viscosity of epoxidized soybean oil as a function of inverse absolute

Figure V.l.

Figure V.2

Figure V.3.

Figure V.4.

Figure V.5.

Figure V.6.

Figure V.6.

Figure V.7.

Figure V.8.

Figure V.9.

Figure V. 10.

Figure 6-1.

temperature. ................................................................................................. 84
Chapter V

Absorbance of the photoinitiator (UV9310C, GE Silicones) and the
photosensitizer (anthracene) as a function of wavelength ......................... 103

Chemical Structure of the photoinitiator, bis (4-dodecylphenyl) iodonium
hexafluoroantimonate and the photosensitizer, anthracene. ...................... 104

Experimental setup for measuring the steady state fluorescence of
anthracene during the photosensitization of a diaryliodonium
hexafluoroantimonate salt. ........................................................................ 105

Typical fluorescence profile of anthracene during the photosensitization of
diaryliodonium salts by exciting at 363.8 nm ............................................ 106

Comparison of computer simulation and experimental photosensitizer
fluorescence intensity for reaction of anthracene and diaryliodonium
hexafluoroantimonate. ............................................................................... 107

The effect of glycerol content on solution viscosity and reaction rate
constant ...................................................................................................... 108

Electronic energy level diagram for anthracene fluorescence and
phosphorescence ........................................................................................ 109

Simulation results for the concentration profiles for various chemical
species in the photosensitization reaction ................................................. 110

Simulated decay profiles of anthracene during photosensitization at
different viscosities .................................................................................... 111

Comparison of experimental data with numerical solutions of the
photophysica] equations using SSE modified kinetic constants ................ 112

Comparison between experimental data on reaction rate coefficient and
simulation with adjusted constants. ........................................................... 113

Chapter VI

Chemical structures of thioxanthones employable as photosensitizers in
conjunction with diaryliodonium initiators. .............................................. 128

CHAPTER I.

INTRODUCTION AND MOTIVATION

1.1. Issues for the Development of UV Curable Polymers.

An important issue to consider with many traditional polymer formulations is
their contribution to atmospheric pollution. The emission of volatile organic components
(VOCs) from the drying of inks, films, and coatings is a leading cause of atmospheric
pollution. Volatile organic solvents have traditionally been used to impart desirable
viscosity and cure rate properties to coating formulations. In order to ensure complete
coverage of a surface, a coating formulation must be sufficiently fluid to be easily
applied, be able to penetrate inner recesses, but must cure rapidly to a protective film.
However, meeting the coating formulation demands through the use of a rapidly
evaporating organic solvent results in the formation of photochemical smog and other air
pollutants. In light of increasingly stringent clean air regulations, there is substantial
motivation for the inclusion of environmental concerns in the development of new high
performance coatings.

An improved, pollution-free formulation for coatings must meet several stringent
requirements. The cured ink or coating should exhibit excellent adhesion to the substrate,
abrasion resistance, and solvent resistance. For applicability in a coating process, the

formulation should have a long shelf life and be stable during the application process, but

2

should cure very rapidly when initiated by the energy source.1 To satisfy safety and
environmental concerns, the formulations should be non-toxic, and should emit no
volatile organic substances. The abrasion and chemical resistance is provided by a highly
crosslinked network. For this reason, many traditional formulations consist of volatile
solvents and diluents dispersed in multifunctional oligomers that form highly crosslinked
networks upon curing. Improved formulations could be based upon highly reactive
monomers of low enough viscosity so that no organic solvents may be necessary. If the
low viscosity monomer also has low vapor pressure, no organic vapors are emitted during
cure. Formulations used for specific applications have specific needs. For example, the
coating systems developed for wood substrates must not cause swelling which may distort
the wood grain.2 A concern with external coatings is the fact that the coating must be
sufficiently elastic to expand or contract with the substrate. Coatings that are too rigid
will crack under thermal or mechanical expansion of the building. Architectural roofs are
often subjected to extreme temperature and pressure gradients, and these forces result in
significant deformation of the structure.

The curing of a polymerizable formulation to form a highly crosslinked network
could be initiated by a variety of methods including heat, electromagnetic radiation, and
electron beam irradiation. Unfortunately, the elevated temperatures in heat initiated
systems lead to high energy costs, in addition to causing significant distortions in the final
polymer product.l High energy radiation methods lead to very rapid curing, but may
cause degradation of the substrate. In contrast, ultraviolet and visible light induced
polymerizations have many potential advantages for ink and coating applications in terms

of rapid reaction rates at ambient temperatures, low energy requirements, wide range of

3
polymerizable monomers and also excellent control over the reaction. With a proper

choice of intensity and wavelength of initiating light, the degradation of the substrate can

be avoided.

1.2. UV-Initiated Photopolymerizations.

Photopolymerizations initiated by ultraviolet (UV) light have gained prominence
in recent years for rapid, pollution-free curing of polymer films (for reviews, see
references 3-5). These solventless polymerizations proceed very rapidly with a fraction
of the energy requirements of thermally cured systems, and create films with excellent
properties. Ultraviolet light is a convenient energy source for photopolymerization
because a variety of readily available compounds will initiate chain polymerizations upon
absorption of UV/visible light.“5 UV-sensitive photoinitiators are currently available for
free—radical or cationic polymerizations. These photoinitiators are typically effective for a
variety of incident wavelengths. This is useful for UV-curable coatings because
commonly used pigments may be strong absorbers of light in the visible and ultraviolet
wavelengths, and an initiator which will be effective at a wavelength outside this window
must be chosen.

Light induced polymerizations carried through the free radical mechanism were
first reported nearly fifty years ago.8 The most widely used classes of monomers for UV-
initiated free-radical photopolymerizations are multifunctional acrylates and
methacrylates. These monomers polymerize very rapidly, and are easily modified on the

ester group, allowing materials with a variety of properties to be obtained. However, the

4
acrylates are relatively volatile, and have an unpleasant odor.4 Moreover, recently there

has been a growing concern over potential health hazards associated with the

15

acrylates.3'5' Several researchers have also shown that free radical polymerizations of

multifunctional acrylates and methacrylates exhibit unusual kinetic behavior, including

. . a
autoacceleration, formation of heterogeneous polymers,4’9"0'l "'2""‘4

and final conversion
considerably lower than a hundred percent.4 Finally, the free radical photo-
polymerizations are inhibited by oxygen and must be carried out under an inert
atmosphere such as nitrogen, increasing the application costs.”8

UV-initiated cationic photopolymerizations exhibit several advantages when
compared to the free-radical photopolymerizations discussed above. First, the cationic

photopolymerizations are not inhibited by oxygen.5'8'”5

This feature provides an
important practical advantage for the coating of large or complex parts since it is not
necessary to blanket the system with nitrogen to achieve the rapid cure rates needed for
quick application rates. Secondly, in contrast to free-radical polymerizations which
experience a rapid decrease in polymerization rate when the light source is removed (due
to radical—radical termination reactions), the cationic polymerizations will proceed long
after the irradiation has ceased, consuming nearly all of the monomer.8’l7 This is a key
feature for the application of cationically photopolymerized coatings for complex shapes,
since light need not penetrate into all corners and crevasses of the substrate to achieve a
completely cured coating. Finally, cationic photopolymerizations are a very versatile
technique, and may be used to polymerize important classes of monomers, including

15-19

epoxides and vinyl ethers. These classes of monomers exhibit many desirable

properties, including low volatility, good rheological properties and negligible toxicity.S

5
Furthermore, the cured polymer films associated with these monomers exhibit excellent
chemical resistance, adhesion, abrasion resistance, and clarity.8'15 Although fully
formulated mixtures containing the monomer and initiator can react within seconds of
exposure to UV light, they are stable for times on the order of months when they are

stored protected from light.

1.3. UV-Sensitive Initiators for Cationic Photopolymerizations.

Applicable photoinitiators for cationic polymerizations were not developed until
the late 19705.27’30 Crivello and Lam reported two classes of thermally stable
photoinitiators for cationic polymerizations: diaryliodonium and triarylsulfonium salts in
1976 and 1979, respectively.20 Upon photolysis, these compounds undergo irreversible
photo-fragmentation in which the carbon—iodine or carbon-sulfur bond is cleaved to
produce an aryliodonium or an arylsulfonium cation-radical capable of initiating cationic
polymerization.7'8 While the diaryliodonium and triarylsulfonium salts actively initiate
cationic polymerization in the presence of UV light, they are remarkably latent in the
absence of light. In fact, fully formulated solutions of these salts in epoxide monomers
have shelf lives of several years.7

The spectral absorbance peaks of diaryliodonium and triarylsulfonium salts are
mostly between 225 nm and 275 nm, depending on the substituents attached to the phenyl
rings.7 Diaryliodonium salts of hexafluoroantimonate produce the fastest rates of reaction
of the onium salts due to their size and acidity of the anion. In spite of their strong

absorbance in the deep ultraviolet region of the spectrum, the absorbance of these

6
initiators is very low in the near UV region.7 Most of the medium and high pressure
mercury vapor lamps used in the industry have their maximum emission bands in the
region above 300 nm and the lack of absorbance in this region limits the efficacy of the
onium salts. However, the spectral region over which the initiators are effective may be
expanded by the addition of a variety of photosensitizers, including hydrocarbons,
ketones, and heterocyclic compounds.7 These compounds actively sensitize the

21. 2 2
2 or energy transfer, 3

diaryliodonium and triarylsulfonium salts by electron transfer
thus rendering them useful in the near UV and visible regions of the light spectrum.
Despite the promise shown by UV light initiated polymerizations, the reactions
have received only limited attention so far. This fact may be attributed to the lack of
efficient initiators until recent years. In fact, most of the work reported in the literature
focuses on the initiation steps of the reaction.:"724'25’26 The development of thermally
stable photoinitiators for cationic chain polymerizations has led to significant progress in
the study of photopolymerizations of epoxides and vinyl ethers in the last decade.27'28‘29‘30

However, the field is significantly less developed compared to that of free radical

photopolymerization.

1.4. Motivation to Use Agricultural Feedstocks as Starting Materials

The use of renewable agricultural feedstocks for the production of chemicals and
materials has been identified as a national priority. The interest in renewable feedstocks

is motivated by several factors, including environmental concerns, economic

7
development of rural areas, and diversification of the resource base from the traditional
petroleum feedstocks.

Epoxides and vinyl ethers are the most conducive monomers for cationic chain
polymerizations. The development of the sensitive onium salt photoinitiators in the last
two decades has made it possible to polymerize virtually all types of cationically
polymerizable monomers. This development has led us to consider what potential
monomers can be obtained from natural resources that could lend themselves to cationic
polymerization and provide a usable product. Glycerol triesters are produced by most
plants as a means of energy storage. These natural oils such as soybean oil, linseed oil,
and corn oil can serve as a potentially diverse and inexpensive class of renewable
substrates from which cationically polymerizable monomeric species can be generated.
The high amount of unsaturation present in these oils can be exploited to introduce
moieties such as epoxides which may be polymerized.

While the use of monomers derived from natural feedstocks clearly meets the
objectives of a renewable feedstock, the use of cationic polymerization also offers
significant practical processing advantages. Since this process is directly based on an
agricultural commodity, traditional materials and methods for growing, harvesting and
collecting the raw material can be used. In addition, cationic polymerizations are
amenable to continuous processing schemes since they exhibit rapid reaction rates and
low energy requirements. These processing advantages are important since many
proposed schemes based upon renewable feedstocks are limited or disqualified because
they are not easily adaptable to large-scale production. This method could make a

tremendous impact on the substitution of synthetic polymers by agri-based polymers.

10.

11.

12.

13.

14.

15.

8

1.5. List of References

E.L. Graminski, Fundamental Chemical and Engineering Issues in Paper
Currency Production, National Science Foundation, Washington DC, 1, (1991).

Wicks, Z. W. Jr.; Jones, F. N.; Pappas, S. P., Organic Coatings Science and
Technology, Vol. 2: Applications, Properties, and Performance, Wiley, New
York, ( 1994).

SP. Pappas, UV Curing, Science and Technology, Vol. 2, Technology Marketing
Corporation, Norwalk, CT, (1985).

J.G. Kloosterboer, Adv. Polym. Sci, 84, l, (1988).

CG. Roffey, Photopolymerization of Surface Coatings, Wiley, New York, (1981).
A. Ledwith, S. Al-Kass, A. Hulme-Lowe, Cationic Polymerization and Related
Processes, E.J. Goethals, Ed., International Union of Pure and Applied Chemistry,
Oxford, UK, 275, (1984).

J.V. Crivello in Cationic Polymerization and Related Processes, E.J. Goethals,
Ed., International Union of Pure and Applied Chemistry, Oxford, UK, 289,
(1984).

A. Reiser, Photoreactive Polymers, Wiley, New York, NY, (1989).

H.M.J. Boots, J.G. Kloosterboer, and G. Van de Hei, Brit. Polym. J., 17, 219,
(1985).

H. Galina, B. Kolarz, P. Wieczorek, and M. Wojczynska, Brit. Polym. J., 17, 215,
(1985).

W. Funke, Brit. Polym. J., 21, 107, (1989).
J. Bastide, and L. Leiber, Macromolecules, 21, 2649, (1988).

A. Matsumoto, H. Matsuo, H. Ando, and M. Oiwa, Eur. Polym. J., 25, 237,
(1989).

J. Baselga, M. Llorenta, I. Hemandez-Ruentes, and I. Pierola, Eur. Polym. J., 25,
471, (1989).

SP. Lapin in Radiation Curing of Polymeric Materials, C.E. Hoyle and IF.
Kinstle, Eds., ACS Symposium Series, 417, Washington DC, 363, (1989).

l6.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

9

J.V. Crivello and J .H.W. Lam, in Epoxy Resin Chemistry, R.S. Bauer, Ed., ACS
Symposium Series, Vol. 114, American Chemical Society, Washington DC, 1,
(1979).

G. Odian, Principles of Polymerization, 2nd Edition, Wiley, New York, NY,
(1981).

PH. Plesch, in Cationic Polymerization and Related Processes, E.J. Goethals,
Ed., International Union of Pure and Applied Chemistry, Oxford, UK, 1, (1984).

J .V. Crivello in Organic Coatings, Science and Technology, Vol. 5, GD. Parfitt
and A.V. Patsis, Eds., Marcel Dekker, Inc., New York, NY, 35, (1983).

J.V. Crivello and J .H.W. Lam, in Epoxy Resin Chemistry, R.S. Bauer, Ed., ACS
Symposium Series, Washington DC, 114, 1, (1979).

E.W. Nelson, T.P. Carter, and AB. Scranton, J. Polym. Sci., Part A: Polym.
Chem, 33, 247, (1995).

J .L. Dektar, and N .P. Hacker, J. Org. Chem, 55, 639, (1990).

LR. Gatechair, and SP. Pappas, Proc. Org. Coating App. Polym. Sci. Div., 46,
707, (1982).

J.V. Crivello, Adv. Polym. Sci., 62, 1, (1984).

J .V. Crivello, and J.L. Lee, J. Polym. Sci., Polym. Chem, 27, 3951, (1989).
Ledwith, A. et al, Polymer, 22, 143, (1981).

SP. Pappas, Prog. Org. Coat, 13, 35, (1985).

J.V. Crivello, Organic Coatings, Science and Technology, Vol. 5, GD. Parfitt and
A.V. Patsis, Eds., Marcel Dekker, Inc., New York, 35, (1983).

J .V. Crivello, and R. Narayan, Chem. Mater., 4, 692, (1992).

W.R. Watt, in Epoxy Resin Chemistry, R.S. Bauer Ed., ACS Symposium Series,
Vol. 114, American Chemical Society, Washington DC, 17 (1979).

CHAPTERIL

BACKGROUND

11.1. Overview of Polymerization Mechanisms

Broadly speaking, there are two different types of polymerization mechanisms-
step polymerization and chain polymerization.l In step polymerization, a monomer reacts
with another monomer to form a dimer which may again react with another monomer to
form a trimer and so on. As a result, the reaction rates are pretty slow and high molecular
weights are attainable only near the end of the reaction."2

In contrast, chain polymerizations are much faster since large chains of polymer

are formed almost immediately after the initiation. The initiator is essential because it

creates the initiating species or reactive center R' which may be a free radical, an anion,

or a cation. The monomer species react with this active center on the forming polymer
rather than other monomers, as in the case of step polymerization. The polymer chains
are terminated when the reactive center is transferred to another monomer, solvent or
other foreign material. In the case of free radicals, they may also terminate chain growth
by combination with each other.

Free radical chain polymerizations are the most commonly used method in the
industry. There are a number of radical initiating systems available, depending on the

initiating method - the initiator can be decomposed into radicals by exposure to heat,

10

11
radiation, or a redox reaction. The free radicals generated can be transferred to a
monomer molecule allowing this new reactive center to react with other monomer
molecules. Every time this reacts with a monomer, the free radical is transferred to the
chain end, thus aiding the propagation. Termination of these chains can occur by two
ways: i) combination, where two radicals react to produce a dead site, or by ii)
disproportion, where a hydrogen is transferred from one chain to the other to consume the

radical and produce a double bond.

11.2. Properties of Cationic Polymerizations.

Cationic polymerizations are chain reactions in which a propagating cationic
center successively reacts with many monomer units to form long polymer chains.
Classes of monomers which will undergo cationic polymerizations include oc-olefins, 1,3-
dienes, vinyl ethers, and epoxides. Cationic initiators are typically protonic or Lewis
acids, and the resulting reactive cations are sufficiently stable to have a reasonable
lifetime for chain growth by propagation. The polymer chain growth may be terminated
by chain transfer of the reactive center, or occasionally by counterion combination.
Compounds such as water, alcohols, and esters are particularly effective chain transfer
agents. Although chain transfer results in a decrease in the average primary polymer
chain length, it typically will not affect the polymerization rate because the resulting
cation will continue to propagate.l'3‘4

While cationic polymerizations are not used extensively for the commercial

synthesis of long polymer chains, they show considerable promise for polymerizations of

l2
highly crosslinked polymer coatings as well as thick parts. Due to the predominance of
transfer processes during the course of the reaction, linear cationic polymerizations tend
to yield polymers of moderate molecular weight and relatively broad molecular weight
distribution. These facts tend to limit the usefulness of the reactions for the synthesis of
linear high polymers. However, for polymerizations of multifunctional monomers, a high
linear primary chain length is not necessary because the resulting highly crosslinked
polymers derive their excellent properties from the network structure, not from long
linear chains. For these systems, the insensitivity to oxygen and high reaction rates
exhibited by cationic polymerizations represent significant advantages. Several classes of
monomers exhibiting desirable properties and rapid polymerization rates have been
reported, including epoxides,5'6 novel silicon—containing epoxy resins7 and a variety of
bisvinyl ethers.S The selection of commercially available monomers and initiators is now

reasonably broad.

11.3. Curing of Epoxy Resins

11.3.1. Conventional Methods

The epoxide group is the smallest possible cyclic ether moiety. The strain in the
three membered carbon-carbon-oxygen ring can be exploited to conduct ring-opening
polymerizations. Epoxy resins are monomers which typically contain one or more
epoxide moieties in their molecular structure. Curing of epoxy resins is being

increasingly carried out using either Lewis acids or Lewis bases. Also, Lewis acids being

13
generated photochemically are becoming more important because of their unique
advantages.

Most epoxy formulations contain diluents (reactive or non-reactive), fillers or
other reinforcement materials, and toughening agents. The toughening or flexibilizing
agents aid the crosslinking process and usually consist of polyesters or aliphatic
diepoxides. Polymers made from epoxy resins possess high chemical and corrosion
resistance, toughness and flexibility, excellent mechanical and electrical properties and
superb adhesion to various substrates. For specific uses, proper selection of additives and
curing agents is required.

More than 500 million pounds of epoxy plastics are produced and used every year
in the United States alone. End uses include structural, non-structural and coating
applications. Epoxy composites are in use in military, automotive, aircraft, and piping
industry. Laminates made from epoxies are used in the electric and electronic industry
whereas other coating applications include roof coatings, coats for roads and bridges, and
also for marine and storage drums. Epoxy coatings are used for beer and beverage cans-
for decorative purposes on the outside and to prevent reaction between metal and

beverage on the inside.

11.3.2. Ring Opening Polymerizations

The polymerization of epoxides is often classified as “ring opening
polymerization” because it involves the cleavage of the oxirane or epoxide ring (also

known as the cyclic ether ring) to form crosslinking intermolecular chains. Other

14
examples of cyclic monomeric species that undergo ring opening polymerizations are
acetals, esters, amides, and siloxanes.

Epoxides or oxiranes are the smallest form of cyclic ether moieties. While the
higher cyclic ethers can undergo ring opening polymerizations only when initiated by
cationic initiators, epoxides are polymerizable by both anionic and cationic initiators.
This is attributable to the great amount of strain in the 3 membered C-O-C ring. The
ionic initiators used for the ring opening polymerizations are the same as the ones used
for polymerizing monomers with C=C or C=O bonds, including the photochemical
initiators. Most cationic polymerizations proceed through the formation and propagation
of oxonium ions. Each propagation step involves the nucleophilic attack of a monomer
on the cationic site at the end of the growing polymer chain, thus adding itself to the
chain and attaining a positive charge on its own oxygen atom. In contrast, anionic
polymerizations proceed through an attack on the monomer molecule by the anionic site
on the polymer chain.

Ring opening polymerizations are very similar to chain polymerizations because
only the monomeric species adds to the polymer chain and species larger than the
monomer do not react among themselves. On the other hand, the molecular weight of
polymers formed has been found to increase slowly with conversion, which is typical of
step polymerizations. Also, the rate constants have been found to be closer to the values
obtained for step polymerizations rather than chain polymerizations. Irrespective of these
discrepancies, it is easier to describe the kinetics of polymerization in a manner
resembling chain polymerizations because the polymers grow monomer by monomer.

Unless the mechanism is complicated by the presence of polymerization-

15
depolymerization equilibria, the kinetics are very well modeled. Because of the high
strain involved in the 3-membered oxirane ring, the probability of depolymerization is

very less.8

11.3.3. Kinetics of Ring Opening Polymerizations

The kinetics of ring opening polymerizations of epoxides can be represented by
the same expressions used in describing chain polymerizations of alkenes and vinyl
ethers. In polymerizations where there is little or no termination involved, the kinetic
expression pertaining to living polymerizations can be utilized:

R, = k,[M1[M*1
where [M] is the instantaneous monomer concentration, and [M] is the concentration of
the propagating cationic species. Since ring opening polymerizations are relatively
uncommon, few reports of kinetic parameters are available. For various oxirane
polymerizations, kp has been reported in the range of 10'1 to 10'3 liters/mol-sec.9‘m'll The

k,, values are much closer to the polyesterification values than to chain polymerizations.

II. 3.4. Efiect of temperature on reaction rate and degree of polymerization

The effect of temperature on the reaction rate and the degree of polymerization
attained is very system dependent in case of cationic ring-opening polymerizations. As in
ionic chain polymerizations of alkenes and vinyl ethers, all system components like the
types and concentrations of monomer, solvent, initiator, and sensitizer play an important

part in the reaction mechanism and kinetics. The activation energy of the forward

l6

propagation reaction is always positive and an increase in temperature almost always
results in an increase in polymerization rate. The activation energies of propagation are
less than the corresponding energies for vinyl ethers.12

The effect on the degree of polymerization or percent conversion is much more
complex. For most ring opening polymerizations, increasing the temperature results in
lower polymer molecular weight and also less conversion of the cyclic ether moieties.
This has been propounded to be due to increase in the rates of chain transfer and
termination relative to propagation.l3 This leads to formation of cyclic polymers at higher
temperatures due to the intramolecular chain transfer. A more reasonable explanation
seems to be that the propagation rate keeps increasing while the termination and transfer
rates do not show much increase as the temperature rises. The resulting effect is that the
polymer molecular weight increases with increasing temperatures up to a certain value,
and then starts decreasing. This could be due to the fact that termination is relatively
unaffected at lower temperatures but increases with temperature at higher temperatures.

Another reason for this effect, suggested by Rose”, who conducted experiments
on the oxetane ring in Tetrahydrofuran(THF), is that the initiator is being thermally
destroyed or decomposed. Another approach which assumes the propagation reaction to
be reversible suggests that the propagation-depropagation equilibria shifts towards the

monomer at higher temperatures. This causes a temperature ceiling effect.

17

11.4. Photopolymerizations

Photopolymerizations initiated by ultraviolet (UV) light have gained prominence
in recent years for the rapid, pollution-free curing of polymer films.""'15‘16 These
solventless polymerizations proceed very rapidly with a fraction of the energy
requirements of thermally cured systems, and create films with excellent properties.
Ultraviolet light is a convenient energy source for photopolymerization because a variety
of readily available compounds will initiate chain polymerizations upon absorption of
UV/visible light.5"7"8"9'25 UV-sensitive photoinitiators are currently available for free-
radical or cationic polymerizations and are typically effective at a variety of incident
wavelengths. This feature is useful for UV-curable coatings because commonly used
pigments may be strong absorbers of light in the visible and ultraviolet wavelengths, and

an initiator which will be effective at a wavelength outside this window must be chosen.

[1.4.1. Free Radical Photopolymerizations

Free radicals can be produced by photoinitiation in two ways. Either the initiating
species itself absorbs the energy from the ultraviolet light or electron beam and
dissociates, or a species known as a photosensitizer may be added which is excited by the
radiation and reacts with the initiator to dissociate it into radicals. The most widely used
classes of monomers for UV-initiated free-radical photopolymerizations are
multifunctional acrylates and methacrylates. These monomers polymerize very rapidly,
and are easily modified on the ester group, allowing materials with a variety of properties

to be obtained. However, the acrylates are relatively volatile and have an unpleasant

18
odor.l5 Moreover, there has been a growing concern over potential health hazards

5.16.25

associated with the acrylates. Finally, the free radical photopolymerizations are

inhibited by oxygen and must be carried out under an inert atmosphere, such as nitrogen,

increasmg the application costs.l '6 '9

11.4.2. Cationic Photopolymerizations.

UV-initiated cationic photopolymerizations exhibit several advantages when
compared with the free-radical photopolymerization discussed above. First, the cationic
photopolymerizations are not inhibited by oxygen.”19 This feature provides an important
practical advantage for the coating of large or complex parts since it is not necessary to
blanket the system with nitrogen to achieve the rapid cure rates needed for quick
application rates. Secondly, in contrast to the free-radical polymerizations which
experience a rapid decrease in polymerization rate when the light source is removed (due
to radical-radical termination reactions), the cationic polymerizations will proceed long

after the irradiation has ceased, consuming nearly all of the monomer.'9'2'

This is a key
feature for the application of cationically photopolymerized coatings for complex shapes,
since light need not penetrate into all comers and crevasses of the substrate to achieve a
completely cured coating. Finally, cationic photopolymerizations are a very versatile
technique, and may be used to polymerize important classes of monomers, including

epoxides and vinyl ethers.5‘20

These classes of monomers exhibit many desirable
properties, including low volatility, good rheological properties and negligible toxicity.‘6

Furthermore, the cured polymer films associated with these monomers exhibit excellent

. . . . . . 5,19
chemical resrstance, adheSion, abraSion res1stance, and clarity.

19

11.5. UV-Sensitive Initiators for Cationic Photopolymerizations.

Despite the advantages of UV-initiated photopolymerizations discussed above, the
technique has received considerably less attention than the analogous free-radical
reactions. This fact may be attributed to the lack of suitable UV-sensitive cationic

- - . - 20,21
photomitiators until recent years.

In fact, applicable photoinitiators for cationic
polymerizations were not developed until the late 19705, more than thirty years after free-
radical initiators had been reported. Crivello and Lam32 reported two classes of thermally
stable photoinitiators for cationic polymerizations: diaryliodonium and triarylsulfonium
salts in 1976 and 1979, respectively. Upon photolysis, these compounds undergo
irreversible photo-fragmentation in which the carbon-iodine or carbon-sulfur bond is
cleaved to produce an aryliodonium or an arylsulfonium cation-radical capable of
initiating cationic polymerization.'8'19 While the diaryliodonium and triarylsulfonium
salts actively initiate cationic polymerization in the presence of UV light, they are

remarkably latent in the absence of light. In fact, fully formulated solutions of these salts

. . . 18
in epox1de monomers have shelf lives of several years.

11.6. Advantages of Cationic Photopolymerizations

UV cured coatings are widely used in the industry, with applications ranging from
furniture, compact discs and car headlamp reflectors to printed circuit boards. All these

coatings are cured by UV light with wavelengths between 200-400 nm. Most industrial

20
UV lamps emit in this region, so it is important to have photoinitiators that absorb light in

this region. Some of the advantages of using radiation curing are:

1. Lower energy consumption compared to thermal curing.

2. Compact operating environment.

3. Very fast cure times, very amenable to mass production of articles.
4. Instant start and shut down of process.

5. Low overall equipment and maintenance costs.

6. Freedom from volatile solvents.

7. Heat sensitive substrates are coated without degradation.

11.7. Photopolymerizations of Epoxides and Vinyl Ethers

Laboratory studies of the cationic photopolymerizations of epoxides and vinyl
ethers have provided information about the salient features of the reactions. The
observed polymerization rate depends on several variables, including the type and
concentration of the initiator, the intensity and wavelength of the UV light source,
temperature, and the structure of the monomer. When diaryliodonium and
triarylsulfonium salts are used as the photoinitiators, the optimum initiator concentrations
are typically 23 wt%.2"32

The first patentszzand journal articles23 on cationic photopolymerizations appeared
in the mid-sixties. In the early to mid seventies, patents were granted to American Can,

covering the use of diazonium salts as photoinitiators for the polymerization of epoxies. 24

In brief, the reaction mechanism can be described as:

21

Ar—N+=N.BF4. ——> ArF + N2 + 31:3

The Lewis acid (BF3) then reacts with the oxirane rings to form the polymer.
Although this chemical mechanism had certain advantages, the commercialization of the
technology met with some difficulties. The aryldiazonium salts tend to react with the
epoxy resins in the dark and the storage stability of the formulated system was
questionable due to its susceptibility to moisture and temperature. As the reaction
mechanism shows above, nitrogen gas is released during the dissociation of the initiator,
which leads to practical difficulties for obtaining bubble-free cured films. Thus, the
technique was restricted to films of thickness less than 10 microns. Towards the end of
the decade, several triarylsulfonium and diaryliodonium salts were developed by General
Electric and 3M. These salts undergo photolysis under the influence of UV light to
generate cationic active centers. Various commercially available triarylsulfonium salt
photoinitiators have been studied comparatively and the results have been tabulated in
literature available from the manufacturers.

A major reason for the delay in the implementation of photoinitiation technology
in the industry was the patent right disputes between General Electric and 3M. It was
only in 1984 that a court ruling clarified the situation and decided that General Electric’s
base patents prevailed over the others. A lot of academic and industrial work25 has been
conducted using cycloaliphatic epoxides with most of the applications being in the
coatings area. So far, no attempt to polymerize thick parts from epoxies using
photopolymerization has been reported. Coatings made with just the resins have been

found to be quite hard and brittle, rendering them unsuitable for most end uses. Apart

22
form increasing the flexibility of the coatings, use of flexibilizers improves the chemical

resistance, weatherability, and the cure response of the coating.

11.8. Cationic Photoinitiators

During the past couple of decades, there has been a lot of academic and industrial
research on photoinitiated polymerizations as people have begun to foresee the numerous
possible applications of the technique. Although photopolymerization by way of free
radicals has been used since the 19505, the field of photoinduced cationic polymerizations
is relatively nascent. This disparity could be attributed to the fact that while there were
many methods of producing free radicals using radiation, the number of methods of
generating stable cationic active centers were much less. Also, cationic polymerizations
are known to proceed very rapidly and hence are very difficult to study and model. The

methods currently used to generate cationic species are by irradiating charge transfer

6 27

complexes,2 aryl onium salts,27 inorganic salts/complexes,

organometallic
compounds,28 and alkyl iodides.”30 Of these, only the first two are mostly used. The
remaining categories are limited by either their relative insolubility, specificity with
respect to the monomeric species, or by their modest initiating capabilities.

Photoreactive onium salts of the Group V, VI, and VH elements of the periodic
table are the most effective cationic initiators. Primary efforts have been to use
diazonium salts, and their mechanistic and kinetic behavior has been well characterized

22,24,31
d.

and utilize Lately, halogen and sulfur based onium salts have displaced the

azonium salts27 and have been utilized in our research. Cationic chain polymerizations

23
induced by onium salts (excited directly or photosensitized) are actually “dark” i.e.
proceed unaffected by light like the conventional polymerizations caused by initiators
such as Lewis or Bronsted acids.

Bronsted acids are much stronger than Lewis acids and polymerize oxirane
containing molecules at much higher cure speeds. The proposed photolysis also
generates free radicals, which suggests a possibility of amalgamating cationic
photopolymerization with radical polymerization.

The photo-generated Bronsted acids are long-living. This means that the active
centers generated during the illumination remain active even after the radiation is stopped
and will continue to induce propagation. This “dark cure” phenomenon has been
evidenced by previous researchers in the case of both epoxies and vinyl ethers by
observing the decrease in the intensity of the absorbance bands representative for the
oxirane structure. Solvent resistance of the cured films has been seen to vary with time,
providing more evidence. Furthermore, the dark cure is heat-catalyzed and therefore heat
producing lamps may be beneficial to use. Preheating of the wet coating or the substrate
will also help the curing. This dark cure phenomenon also facilitates the cure of remote
areas of coated objects which are shielded from the ultraviolet light. The advancing
polymerization propagation front, combined with the high diffusion characteristics of the
Bronsted acid are responsible for this. One significant advantage of using epoxies for
coatings applications instead of urethanes is that they exhibit excellent adhesion to a
variety of substrates, including metal. The contributing factor for this excellent adhesion
is the high wetting characteristics exhibited by epoxies on metals and most plastics.

Another important characteristic is that the cationic polymerization of the epoxies is

24
associated with extremely low volume shrinkage, l to 2 percent, as compared to the 15-
20 percent volume shrinkage for the acrylate systems that undergo free radical
polymerizations.

The primary requirement of photoinitiators is that they must possess
“chromophores” which absorb the incident light and undergo photochemical changes so
as to generate species capable of initiating cationic polymerization. It should have a long
shelf life when not exposed to light and be completely dormant when mixed with the
polymerizable monomeric species. On irradiation, it should rapidly produce active
centers with high quantum efficiency and not produce byproducts which could retard the
initiation or polymerization. Also, it should have an absorbance spectra well resolved
from other components in the system (monomer, solvents) and also, the absorbance peaks
should coincide well will emission peaks in the irradiating source. The absorbance of the
products and byproducts formed should be away from the absorbance of the initiator. To
enhance the solubility of photoinitiators in the monomers, certain moieties are added onto
the initiator molecules. Care has to be exercised such that these don’t retard the
efficiency of the initiator. For commercial applicability of the initiating system, toxicity,
stability of active centers, color, and cost should be taken into consideration.

Diaryliodonium and triarylsulfonium salts undergo photoinduced fragmentations
to generate aryl radicals and either aryliodinium or diarylsulfinium cation radicals. These
onium salts can be photosensitized to respond to the longer wavelength UV and visible
light. This photosensitization can be caused by compounds like anthracene and perylene
by an electron transfer reaction. While the structure of the cations in the onium salts

dictates the photochemistry and the rate at which the initiator fragments are produced, the

25
nature of the anion determines the probability of the active cations to undergo
termination. Photoinitiated cationic chain polymerizations employing these onium salts
are used in a number of commercial processes to produce coatings, adhesives, inks, and

novel photoresists for semiconductor applications.

11.9. General Characteristics of Onium Salts

Diaryliodonium, bromonium and chloronium salts belong to the category of

halogenic onium salts, with a structure like
Ar—Yt—Ar
X-

where Y = 1, Br, Cl and X is a nucleophilic species such as BF4', SbF6'. Typically, the
centrally placed halogen atom possesses a ‘+3’ oxidation state. The synthetic methods for
preparation of bromonium and chloronium salts is difficult and the resulting salts are also
very unstable.32 In contrast, the diaryliodonium salts have long shelf lives and are
therefore preferred for use as photoinitiators. The diaryliodonium salts are generally
colorless to pale yellow crystalline compounds and are soluble in most solvents.
Depending on the nature of the nucleophilic species, they are available as ionic salts or as
covalent ones. The mechanism for generation of cationic active centers by shining UV

light on diaryl onium salts is given below.
Ar,I+X‘ —i”—+[Ar,1+X“]‘

[Ar,1+X"]‘ —> ArI'+X‘ + Ar'

26
In the above scheme, light is absorbed by the molecule leading to an electronically
excited state. This photoexcited species then undergoes rapid decay by cleavage of the
carbon-iodine bond, thus producing an aryliodonium cation radical and an aryl radical.
Triarylsulfonium salts also undergo a similar reaction mechanism. The excited singlet
state of the molecule is responsible for reaction in both cases.

The general chemical structure of the triaryl onium salts can be represented by:

Ar

|,_

Ar—M X
where M = O, S, Se, Te (Group VI of the Periodic Table) and X- is again a nucleophilic
species. Of these, only the triarylsulfonium and triarylselenonium salts have been
investigated for use as photoinitiators. The triarylsulfonium salts are preferred because of
their case of synthesis and their thermal stability. Most of the salts are colorless crystals
and are soluble in organic solvents. Ledwith has showed that the salts are ionic and more
highly dissociated than their diaryliodonium salt counterparts.33 The mechanism of

photogeneration of cationic active centers is shown below.
Ar,s+x- —i”—>[Ar,s+X‘ 1‘

[Ar,s+X-]‘ —> Ar,S°+X— + Ar'

11.10. Photosensitization

Most triarylsulfonium and diaryliodonium salts absorb strongly at wavelengths

between 200 and 300 nm. Chemical substitution of the aryl rings does not substantially

27

change the region of absorbance. The lack of absorbance peaks in the region between
300 and 400 nm causes a major concern because most medium and high pressure
industrial lamps emit in this spectral region and, for optimum light utilization, they
should absorb here. Moreover, the products and byproducts of initiation also absorb in
the same region, thus leading to lower efficiency because of the competitive absorption.
Also, monomers with aromatic moieties absorb around 250 nm, leading to suboptimal
operation.

This problem can be offset in two ways. One way is to synthesize photoinitiators
which incorporate “chromophores” that permit absorbance at higher wavelengths.
Altemately, the photolysis of onium salts can be photosensitized, facilitating the use of
near UV and visible light for initiation of cationic chain polymerizations. In the past,
many aromatic hydrocarbons, aromatic ketones and dyes have been used as
photosensitizers. Three mechanisms have been proposed by different researchers to
explain the phenomenon of photosensitization.

1. Energy transfer photosensitization

2. Electron transfer photosensitization

3. Photosensitization by free radical induced decomposition

It has been shown by many researchers that the photosensitization of onium salts
takes place by the electron transfer mechanism. Analysis of the products of
photosensitization has shown them to be identical to those produced when this
mechanism is assumed. Additional evidence is presented by experimental observations
by Gatechair34 based on thermodynamic arguments. Essentially, the electron transfer

mechanism holds ground because it satisfies both primary requirements - first, the

28
excitation energy of the photosensitizer (E*) is greater than the net energy required to

oxidize the photosensitizer (E°" )and the energy (Ered

sens onium

) to reduce the onium salt.

Secondly, the magnitude of the free energy (AG) of photosensitization is negative, i.e.

the process is exothermic.

1

AG = (13" — 13':d

sens onium

) — E
The general mechanism for photosensitization of onium salts is given below.
P—“1—> P *
P *+Ar21+X’ —) [P*...Ar21+X‘] —-> P+ ~X' + Arzl-

Ar21-—> Ar! + Ar-

nM—i’f—a—(M), —

In the mechanism shown above for diaryliodonium salts, the sensitizer absorbs the
incident radiation and attains an excited state. The excited state responsible for reaction
with the onium salt could be the singlet or triplet excited state of the sensitizer molecule.
For anthracene, it has been shown by Nelson et al35 to be the triplet state. Upon
interaction between the sensitizer and the onium salt, an exciplex is formed which
undergoes a photoredox reaction. The sensitizer undergoes oxidation to form a cation
radical while the onium salt reduces to a diaryliodide radical. It should be noted that
unlike direct excitation of the onium salts, the initiating species is produced from the
sensitizing species.

Diaryliodonium salts have lower oxidation potentials compared to the
triarylsulfonium salts, and hence are easier to photosensitize. Thus, while the former can

be photosensitized by a variety of compounds, the latter is restricted to compounds with

29
low oxidation potentials and high triplet energies. Condensed ring aromatics such as

anthracene and perylene are well suited for the purpose.

11.11. Polymerization Characteristics of Onium Salts

Since the cation part of the initiator molecule is responsible for the initiation and
propagation, the initiating characteristics of the salt is greatly dependent on the
“chromophores” residing in that part of the molecule. The absorbance spectra of the
photosensitizer determines amount of light absorbed and effectively used. Both the direct
and photosensitized photolysis of onium salts generate similar initiating species, namely,
aryliodinium and diarylsulfinium cation radicals.

Since the nature of the anion plays no direct role in the photochemistry of the
onium salts, the quantum yields of salts with different anions but same cations is
essentially the same. Nevertheless, the anions do play a part in determining the reactivity
of the cations towards the monomer as well as influence the propagation and termination
rates.

Photoinitiators bearing strongly nucleophilic anions such as the halogens are very
weak photoinitiators. The onium salts bearing very weakly nucleophilic anions such as

BF," ,PF;,AsF6',SbF6‘ exhibit excellent reactivity when irradiated. Because of the lack of

reactivity of these anions, they remain in ionized form and the cationic active centers
continue with the propagation. 1n the case of epoxide containing monomers, the

reactivity of the diaryliodonium and triarylsulfonium salts has been found to increase in

the following order, depending on the type of anion: BF,“ < PF; < AsF6‘ <SbF6‘ .

 

30
Although the reasons for this trend are not fully explained, the nucleophilicity of the
anion definitely will play a part. As the nucleophilicity of the anion decreases in the order

BF,— > PF; > Ang > Sng , the separation between the ions increases, thus increasing the

reactivity. At the same time, the susceptibility of the anion toward fluorine abstraction
decreases, thus decreasing the tendency of termination.

It has been shown that heterocyclic polymerizations proceeding by a cationic
mechanism will continue for very long times in the absence of terminating impurities.
Such living polymerizations have been reported in the case of THF using diaryliodonium
and triarylsulfonium salts.

Photoinitiated cationic chain polymerizations can be carried out in the presence of
oxygen since it neither quenches the excited states of the initiator, nor does it interact
with the pr0pagating species. In cases where free radical induced decomposition of the
initiator takes place, oxygen may intercept the free radicals. Bases of all types inhibit
chain polymerization and should be removed. Water and other hydroxyl containing
compounds can act as inhibitors or chain transfer agents depending on the amount

present.

11.12. Epoxidized Soybean Oil: An Agri-based Source of Polymerizable Epoxides.

[1.12.]. Motivation to Use Agricultural F eedstocks as Starting Materials

The use of renewable agricultural feedstocks for the production of chemicals and
materials has been identified as a national priority. The interest in renewable feedstocks

is motivated by several factors including environmental concerns, economic development

31
of rural areas, and diversification of the resource base from the traditional petroleum
feedstocks. In this research, we have attempted to meet these criteria by developing a
novel and feasible process for producing polymeric materials from renewable agricultural
feedstocks. Specifically, we have investigated cationic polymerizations of epoxidized
natural oils, and evaluated their potential for use as substitutes to synthetic polymeric
materials for bulk (non-coating) applications. While the use of monomers derived from
natural feedstocks clearly meets the objectives of a renewable feedstock, the cationic
photopolymerization scheme also offers significant practical processing advantages.
Since this process is directly based on an agricultural commodity, traditional materials
and methods for growing, harvesting and collecting the raw material can be used. In
addition, cationic polymerizations are amenable to continuous processing schemes, since
they exhibit rapid reaction rates and low energy requirements. These processing
advantages are important since many proposed schemes based upon renewable feedstocks
are limited or disqualified because they are not easily adapted to large-scale production.
Since the targeted applications are bulk (thick) polymers, this method could make a

tremendous impact on the substitution of synthetic polymers by agri-based polymers.

II.12.2. Traditional Methods of Producing Polymeric Films and Coatings
from Natural Oils.

Natural oils and their derivatives have long been used as precursors to polymeric
components in films, paints and varnishes. These current techniques for using agri-based
oils in films and coatings fall into two major classes. One of the oldest methods for

making films from oils is based on naturally-occurring oxidative crosslinking reactions in

32
unsaturated “drying oils.” These reactions formed the basis for many of the early
lacquers and varnishes. The second class of reactions involve step reactions of oils that
either naturally contain a reactive moiety or are modified with a reactive functionality.
For example, maleated, epoxidized, or esterified oils may be polymerized with
comonomers by a step reaction to produce a polymer product. Both of these current
polymerization methods for natural oils have significant problems and limitations, and are
not well suited for the mass production of polymers, especially for parts of appreciable
thickness. Most of these limitations ultimately arise from the inherently slow

polymerization rates associated with the current curing techniques.

ll.12.3. Reactions in Natural Drying Oils.

Drying oils are liquid vegetable or fish oils which react with oxygen to form solid
films. Such oils were widely used in the 19th and early 20th century as binders for paints,
and are still used as raw materials for alkyd resins, epoxy esters and uralkyds. During
drying, a series of complex free-radical reactions take place and ultimately lead to the
formation of a crosslinked polymer. For example, the crosslinking of linseed oil, a major
drying oil, is thought to consist of the following steps36'37: 1) an induction period during
which naturally present antioxidants (toc0pherols) are consumed; 2) a period of oxygen
uptake and formation of hydroperoxides; and 3) a sequence of catalytic reactions where
the hydroperoxides are consumed and the cross-linked film is formed. These reactions
are catalyzed by heavy-metal “driers” such as cobalt, lead or zirconium salts of octanoic
acid or naphthenic acid. The rate of polymerization depends on the degree of conjugation

of the oil, with oils containing conjugated double bonds drying more rapidly than non-

33
conjugated oils. In addition, free radical polymerization of conjugated oils can lead to
chain-growth polymerization rather than just cross-linking, and therefore the resulting
films typically exhibit better water and alkali resistance compared to films formed from
non-conjugated oils.

Although drying oils have traditionally filled a niche in the coatings industry, they
have lost much of their market share due to a variety of problems and limitations. First of
all, the requirement of oxygen in the reaction scheme leads to relatively long cure times,
and inherently limits the thickness of the cured polymer. In addition to prolonged cure
times (on the order of 50 hours) several other problems associated with drying oils arise
from the complex sequence of reactions which take place in the curing film. For
example, rearrangement and cleavage of the hydroperoxides leads to low molecular
weight volatile by-products, as well as discoloration and embrittlement of the film.
Moreover, the use of the metal catalyst (the drier) not only enhances the rate of the
“drying” reactions, but also catalyzes a number of post-drying reactions which lead to
embrittlement and discoloration of the film. Finally, the presence of the toxic metal
catalysts has raised many environmental concerns and has resulted in severe restrictions

and limitations on their use.

[1.1 2.4. Reactions of F unctionalized Oils.

Alkyd Resins. The most prevalent class of functionalized oils are the alkyd resins,
which are unsaturated polyesters containing carbon double bonds at the chain ends.

Alkyd resins are commonly produced by the condensation reaction of oil-derived acids

34
with polyols (forming ester linkages). Two classes of alkyd resins have been identified.
The first class, oxidizing alkyd resins, cures by the same mechanism as described above
for natural drying oils and are hence sometimes referred to as synthetic drying oils. For
example, soybean oil is commonly combined with glycerol to form a synthetic drying oil
which cures by this mechanism. In contrast, non-oxidizing alkyd resins are used as
polymeric plasticizers or as hydroxy-functional resins which may be cross-linked by
reaction with melamine-formaldehyde, urea-formaldehyde or isocyanate cross-linkers.
This second class of alkyds cures by specific condensation reactions between groups on
the alkyd and the crosslinker. The commercial use of oxidizing alkyd resins is far greater
than that of non-oxidizing alkyd resins. For example, while about 120 million pounds of

soybean oil were used in the manufacture of oxidizing alkyd resins, only 6-8 million

. . . . . . 3
pounds of 011 were used in the manufacture of non-ox1dizmg alkyd resms.3 However,

the use of soybean oil in the manufacture of both types of resins has steadily declined
since 1949.

F rec-radical polymerizations of unsaturated oils. Both conjugated and non-
conjugated unsaturated oils may be copolymerized with vinyl monomers such as styrene,
vinyltoluene, cyclopentadiene and acrylic esters. These polymerizations proceed by the
standard free radical mechanism, however they typically exhibit a high degree of chain
transfer. Due to this chain transfer, the molecular weight of the product is limited, and
the method typically yields a variety of low molecular weight homopolymers, short-chain
graft copolymers, and dimerized drying oil molecules. Therefore, the method has very

limited applicability for the production of high polymers. The only such reaction that is

35

commercially viable is the reaction of cyclopentadiene with linseed oil to form a faster
reacting drying oil. However, the dark color and odor of the oil limit its applications.

Epoxidized oils. Unsaturated oils may be epoxidized by reaction with
peroxyacids to convert the carbon double bonds to oxirane functionalities. Once
epoxidized, these oils may be polymerized by reaction with amine-containing monomers
by the standard epoxy-arnine reactions. Epoxidized oils have found modest application in
specialty coating systems, however the cure rate is typically rather slow. One factor
which prevents the extensive application of these oils is the relatively low reactivity of
internal epoxide groups. This is a drawback for polymerization of epoxidized oils since
these oils primarily contain internal oxirane groups. At ambient temperature, cure times
on the order of a week are common for reactions of these oils with aliphatic amines.
Other limitations of these systems arise from the inevitable trade-off between pot-life and
cure rate. Ketones such as methyl ethyl ketone (MEK) have been used to extend the pot-
life of the resins through an amine-blocking mechanism. Upon exposure to ambient
moisture, the ketone is released, allowing cure to take place. While this ingenious
scheme extends the pot-life of the resins, the use of MEK, a highly volatile organic
compound, is undesirable due to its negative environmental impact.

Due to the problems described above, current utilization of natural oils in bulk
polymers is largely limited to secondary roles such as plasticizers, reactive diluents and

comonomers. Epoxidized soybean oil and to a lesser extent epoxidized linseed oil are

39
used as plasticizers for PVC, and impart excellent heat and light stability. Epoxidized

and other functionalized oils from soybean, vernonia and castor seed have been used to

form two-phase interpenetrating networks, via step reactions with amines, terephthalates

36

40.41.42

and other epoxy monomers. Recently Crivello and Narayan43 demonstrated that

epoxidized oils can be used to form films by cationic photo-polymerization. This
technique has been used to make thin-films which possess good adhesion and mechanical

properties, at relatively high reaction rates.

11.13. Need for New Polymerization Methods Based Upon Natural Oils.

Although there is a compelling motivation to develop polymers based upon
natural oils, all of the techniques reported to date exhibit significant drawbacks and
limitations. The above discussion illustrates many of the concerns associated with the
slow cure and the environmental impact of additives in the current systems. These issues
have limited the use of natural oils, even for polymer film and coating applications.
Moreover, the slow rates exhibited by these methods precludes the possibility of their use
for producing thick parts or bulk polymers. Clearly new, commercially viable
polymerization methods must be developed if natural oils are to be extensively utilized as
a feedstock for the production of polymeric materials. Cationic polymerizations may be
used to rapidly polymerize epoxidized oils in either thin films or thick parts without the
use of environmentally detrimental additives. This technique has significant practical
advantages (high cure rates with an essentially unlimited pot-life) and is amenable to

mass production by continuous process schemes.

37

11.14. Cationic Polymerizations of Epoxidized Oils.

Cationic polymerizations are chain reactions in which a propagating cationic
center successively reacts with many monomer units to form long polymer chains.
Classes of monomers which will undergo cationic polymerizations include a-olefins, 1,3-
dienes, vinyl ethers, and epoxides. Cationic initiators are typically protonic or Lewis
acids, and the resulting reactive cations are sufficiently stable to have a reasonable
lifetime for chain growth by propagation. The polymer chain growth may be terminated
by chain transfer of the reactive center, or occasionally by counterion combination.
Compounds such as water, alcohols, and esters are particularly effective chain transfer
agents. Although chain transfer results in a decrease in the average primary polymer
chain length, it typically will not affect the polymerization rate because the resulting
cation will continue to propagate.‘

Cationic polymerizations have considerable potential for the development of a
novel polymerization scheme for epoxidized oils which avoids the limitations of the
current methods. They require a fraction of the energy of their thermal counterparts, and
may be carried out with no diluting solvent, resulting in low emissions of volatile organic
components. The cationic initiation reaction is very rapid, and the propagation reaction
continues long after the irradiation has ceased, resulting in almost complete conversion of
the monomer to polymer. Finally, unlike their free-radical counterparts, cationic
polymerizations are not inhibited by oxygen, and therefore do not require an inert
atmosphere to achieve rapid polymerization rates. Epoxides and vinyl ethers are

important classes of monomers which undergo cationic polymerizations to yield polymers

38
exhibiting excellent mechanical properties, adhesion, abrasion resistance, and chemical
resistance. Despite the advantages of cationic photopolymerizations discussed above, the
technique has received only limited attention (considerably less than the analogous free-

radical reactions). This fact may be attributed to the lack of stable and suitable cationic
photomitiators until recent years. In fact, applicable photomitiators for cationic

polymerizations were not developed until the late 19708, more than thirty years after free-
radical initiators had been reported.

Epoxidized soybean oil (ESO), which is produced by the reaction of soybean oil
with hydrogen peroxide, accounts for more than half the epoxidized oil market. Soybean
oil consists of triglyceridesters of long chain (16-18 carbon) unsaturated fatty acids
(Figure 11.1). The average molecular weight of the ESO molecule is 976 grams. Upon
titration, it was found that the epoxy moieties per molecule of ESO averaged around 6.
At room temperature, the epoxidized oil is lighter than water (specific gravity = 0.93) and
moderately viscous (340 cP).

Olefinic double bonds in the oil are reacted with peroxy compounds to form
epoxide moieties. Epoxidized soybean oil is widely used as a plasticizer for PVC, and has
an annual production of 70 x 106 kg with a selling price of $1.40/kg (S 0.65/lb).
Although the technology to convert the agricultural feedstock (soybean oil) to the
monomer exists, very little work has been done to convert the monomer to bulk polymer.
This is due in part to the low rates of polymerization using conventional techniques
(curing the epoxide by reaction with a diamine). Recently developed cationic
photoinitiators have been used to polymerize thin films of epoxidized oils at relatively

high reaction rates, and the resulting films were found to possess good adhesion and

39
. . 43 . .
mechanical properties. However, no attempt was made to polymerize bulk speCimens

from epoxidized oils.

The conventional curing of oxirane containing molecules such as cycloaliphatic
epoxides with organic acid anhydrides, organic acids, Lewis acids, and blocked Lewis
acids is widely practiced in the industry and the reaction mechanisms are well

documented and described in the 1iterature."2'45

One way to describe cationic UV
initiated polymerization is to say that it is the polymerization of the oxirane molecules by
use of a “blocked” Lewis or Bronsted acid, which unblocks upon exposure to UV light
and liberates an acid. This liberated acid in turn polymerizes the oxirane molecules
through a mechanism similar to conventional curing of epoxies.

Diaryliodonium salts can be easily employed to photoinitiate cationic ring opening
polymerization of epoxidized soybean oil using anthracene as a photosensitizer. Due to

the multifunctional nature of the ESO molecule, a crosslinked polymer can be obtained

very rapidly. A schematic of the crosslinking polymerization of ESO is shown in Figure

11.2.

40

o o o
CH2-O—&—(CHQ)7—Ci-l—\CH—CI-iz—Ci-l—\CH2—CH—(CHZ)4—CH3

o o o o
(bl-i—o—ii—(CHZ),—-ci—I-—\3H—CI~12——c41—\:i+—CHZ—ciJI——\:H——(CHZ)4—Ci-I3

 

O O
éw-O—i—icmy—c/H—EH—icwz-Cm

Figure 11.1. Average chemical structure of the epoxidized soybean oil molecule.

 

41

\

~~~~~ (CH2)7-C|/'|O—\CH—CH2-Cé)—CH ~~~~~~~ + Initiator
0 CH2 """""
/ \ (E
“W (CH2)7-CH—CH—CH2—(|3H— H
0 CH2 ~~~~~
(CH2)7—éH—?H—CH2—(|3H— H CH2 ~~~~~
o o b

...... (CH2)7-CH—CH—CH2—$H—CEH (EH—CH—CHz—OH—(EH """""

EH2 ..... EH2 ~~~~~

Figure 11.2 Schematic of the crosslinking polymerization of epoxidized soybean oil.

10.

11.

12.

l3.

14.

15.

16.

42

11.15. List of References
G. Odian, Principles of Polymerization, 2nd Edition, Wiley, New York, NY,
(1981).

F.W. Billmeyer, Jr., Textbook of Polymer Science, 3rd Edition, Wiley—Interscience,
New York, NY, (1984).

PH. Plesch, in Cationic Polymerization and Related Processes, E.J. Goethals,
Ed., International Union of Pure and Applied Chemistry, Oxford, UK, 1, ( 1984).

DJ. Dunn, in Developments in Polymerization, R.N. Haward, Ed., Applied
Science Publishers, London, 1, 45, ( 1979).

SP. Lapin, Radiation Curing of Polymeric Materials, C.E. Hoyle and J.F. Kinstle,
Eds., ACS Symposium Series, 417, Washington DC, 363, (1989).

J .V. Crivello in Organic Coatings, Science and Technology, Vol. 5, GD. Parfitt
and A.V. Patsis, Eds., Marcel Dekker, Inc., New York, NY, 35, (1983).

RF. Eckberg and KB. Riding, in Radiation Curing of Polymeric Materials, C.E.
Hoyle and J.F. Kinstle, Eds., ACS Symposium Series, ACS, Washington DC,
417, 382, (1989).

P. Kubisa, in Cationic Polymerizations - Mechanisms, Synthesis, and
Applicaitons, K. Matyjaszewski, Ed., Marcel Dekker, Inc, New York, (1996).

J .C.W. Chien, Y.G. Cheun, C.P. Lillya, Macromolecules, 21, 870, (1988).

P. Dreyfuss, M.P. Dreyfuss, in Comprehensive Polymer Science, Vol. 3, G.C.
Eastmond, A. Ledwith, S. Russo, and P. Sigwalt, Eds., Pergamon Press, London,
(1989).

K. Matyajaszewski, R. Szymanski, , P. Kubisa, S. Pencsek, Acta Polym., 35, 14,
(1984).

E.W. Nelson, J.L. Jacobs, A.B. Scranton, K.S. Anseth, and ON. Bowman,
Polymer, 36(24), 4651, (1995).

J .B. Rose, J. Chem. Soc., 542, 547, (1956).

SP. Pappas, UV Curing, Science and Technology, Vol. 2, Technology Marketing
Corporation, Norwalk, CT, ( 1985).

J .G. Kloosterboer, Adv. Polym. Sci., 84, l, (1988).

CG. Roffey, Photopolymerization of Surface Coatings, Wiley, New York, (1981).

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

43

A. Ledwith, S. Al-Kass, A. Hulme-Lowe in "Cationic Polymerization and Related
Processes," edited by E.J. Goethals, International Union of Pure and Applied
Chemistry, Oxford, UK, 275, (1984).

J.V. Crivello in Cationic Polymerization and Related Processes, E.J. Goethals,
Ed., International Union of Pure and Applied Chemistry, Oxford, UK, 289,
(1984).

A. Reiser, Photoreactive Polymers, Wiley, New York, NY, (1989).

SP. Pappas, Prog. Org. Coat, 13, 35, (1985).

W.R. Watt, in Epoxy Resin Chemistry, R.S. Bauer, Ed., ACS Symposium Series,
Washington DC, 114, 17, (1979).

1.1. Licari and RC. Crepeau, US. Patent 3,205,157, (1965).
V.W. Strohmeier, and C. Barbeau, Mahromol. Chemic., 81, 86, (1965).
8.1. Schlesinger, US. Patent 3,708,296, (1973).

SP. Pappas, UV Curing, Science and Technology, Vol. 2, Technology Marketing
Corporation, Norwalk, CT, (1985).

J .P. Kennedy, E. Marechal, Carbocationic Polymerization, John Wiley and Sons,
New York, (1982).

J .V. Crivello, Advances in Polymer Science, 62, l, (1984).

J.V. Crivello, in UV Curing: Science and Technology, S.P Pappas, Ed.,
Technology Marketing Corporation, Stamford, U.S.A., 24, (1978).

T. Diem, J .P. Kennedy, and R.T. Chou, Polym. Bull, 1, 281, (1979).
J .F. Kinstle, and TA. Tufts, ACS Preprints, 20(2), 661, (1979).
E. Fischer, US. Patent, 3,236,784, (1966).

J .V. Crivello and J .H.W. Lam, in Epoxy Resin Chemistry, R.S. Bauer, Ed., ACS
Symposium Series, Washington DC, 114, 1, (1979).

A. Ledwith, Polymer, 19, 1217, (1978).

LR. Gatechair, and SP. Pappas, Proc. Org. Coating App. Polym. Sci. Div., 46,
707, (1982).

E.W. Nelson, T.P. Carter, and AB. Scranton, J. Polym. Sci., Part A: Polym.
Chem, 33, 247, (1995).

36.

37.

38.

39.

40.

41.

42.

43.

45.

44

Z.W. Wicks, F.N. Jones, and SP. Pappas, in Organic Coatings Science and
Technology, Volume 1: Film Formation, Components and Appearance, Wiley,
New York, (1992).

Z.W. Wicks, Jr., F.N. Jones, S.P. Pappas, Organic Coatings Science and
Technology, Vol. 2: Applications, Properties, and Performance, Wiley, New
York, (1994).

N.O.V. Sonntag, J. Amer. Oil Chem. Soc., 62(5), 928, (1985).

K.D. Carlson, S.P. Chang, J. Amer. Oil Chem. Soc., 62 (5), 934, (1985).

L.W. Barret, G.S. Ferguson, L.H. Sperling, J. Polym. Sci. A, Polym. Chem, 31,
1287, (1993).

L.W. Barret, L.H. Sperling, J. Gilmer, S.G. Mylonakis, J. Applied Polym. Sci., 48,
1035, (1993).

L.W. Barret, L.H. Sperling, G]. Murphy, J. Amer. Oil. Chem. Soc., 70, 523,
(1993).

J.V. Crivello, and R. Narayan, Chem. Mater., 4, 692, (1992).

W.R. Watt, in Epoxy Resin Chemistry, R.S. Bauer Ed., ACS Symposium Series,
Washington DC, 114, 17 (1979).

H. Lee, and K. Neville, Handbook of Epoxy Resins, McGraw Hill Book Company,
(1967).

CHAPTER III.

OBJECTIVES OF RESEARCH

Based upon the motivation and background provided, it is clear that cationic chain
polymerizations of vinyl ethers and epoxides initiated by UV light have considerable
potential for the development of high speed, low cost, pollution free coatings, inks and
thick polymer parts. It is also generally acknowledged that polymers produced from
renewable agricultural resources offer several attractive advantages over traditional
synthetic materials.

The overall goal of this project was to demonstrate that epoxidized oils can be
rapidly and efficiently polymerized to produce usable polymers. Using cationic
polymerizations, we identified suitable processing parameters necessary to establish this
process as a viable technique for producing polymers from renewable feedstocks. A
variety of processing conditions were investigated, including initiator and photosensitizer
formulations, different light intensities, and temperatures. To characterize the reaction,
the epoxide conversion was monitored as a function of time using differential scanning
calorimetry.

Diaryliodonium hexafluoroantimonate salts, photosensitized by anthracene, were
used to perform photoinitiated cationic ring opening polymerizations of epoxidized

soybean oil. During the course of this research, we found that the photosensitization

45

46

reaction between anthracene and diaryliodonium salts needs further investigation.
Although this reaction has been extensively studied by various researchers in the past, the
effect of viscosity on the reaction kinetics has not been investigated. Steady state
fluorescence spectroscopy experiments were employed to study the effects of diffusion on
the reaction. Mathematical modeling of the photosensitization mechanism using a well
known theory for diffusion controlled reactions was used to corroborate the experimental

findings.

Specific objectives of this project were:

i. to establish reaction conditions for the production of bulk polymers from
epoxidized natural triglyceride oils;

ii. to obtain detailed profiles of oxirane conversions during the photopolymerization
of epoxidized soybean oil at different light intensities and reaction temperatures;

iii. to experimentally characterize the effect of viscosity on the rate of
photosensitization of diaryliodonium salts using anthracene.

iv. to provide a correlative model for the effect of viscosity on the rate of

photosensitization of diaryliodonium salts using anthracene.

CHAPTER IV.

KINETIC STUDIES OF PHOTOPOLYMERIZATION OF
EPOXIDIZED NATURAL TRIGLYCERIDES

Abstract

Differential Scanning Photocalorimetry (DSP) was used to investigate a series of
photoinitiated cationic polymerizations at different temperatures, using epoxidized
soybean oil as the monomer. The apparent propagation rate constant exhibited an initial
sharp increase, decreased gradually after reaching a plateau, then decreased rapidly upon
reaching a limiting conversion. The limiting conversion obtained was found to be higher
for the polymerizations which proceeded at lower temperatures, although the initial rate
of polymerization was found to increase with increasing temperatures. These trends were

attributed to diffusional mobility effects in the rapidly polymerizing systems.

IV.1. Introduction

Natural oils such as soybean, linseed, and corn could be used as monomers for
chain polymerization if the unsaturation in these triglyceride esters was utilized.
Epoxidation of these oils is an excellent means of introducing oxirane moieties into the
chemical structure of the oils to make them conducive to cationic polymerization.

Epoxidized Soybean Oil (ESO) has been used for several years as a plasticizer for PVC,l

47

48

and the polymers formed from ESO have been shown to be biodegradable.2 Its source is
in an agricultural crop, and hence it is a renewable source. Lately, BSD and other
epoxidized oils have also been employed as reactive diluents for thermal curing with
other epoxy monomers,3 as epoxy components in the production of interpenetrating
polymer networks,4 and as toughening materials for commercial epoxy resins.5

Whereas almost all monomers containing unsaturation are amenable to radical
polymerization, ionic polymerizations are found to be highly selective.6 This selectivity
is attributed to the unstable nature of the propagating species, hence limiting the
commercial utilization of ionic polymerizations. It also has been extremely difficult to
obtain reproducible kinetic data in the past for ionic polymerizations due to their very
rapid reaction rates, and their sensitivity to the nature of impurities and solvents. Low
temperatures are required in both anionic and cationic polymerizations to suppress
termination and chain transfer reactions whereas free radical polymerizations proceed
comfortably at relatively higher temperatures. Also, the activation energies for ionic
polymerizations are significantly lower than their radical counterparts and are sometimes
found to be negative. In the past, some solvents have been employed to stabilize the
propagating species and increase their effective lifetime. In ionic polymerizations, the
propagating species can be a positively charged cation or a negatively charged anion.
While cationic polymerization requires electron-releasing substituents such as alkoxy,
phenyl, oxirane, or vinyl in the monomer, anionic polymerization requires monomers
with electron-withdrawing groups like nitrile, carboxyl, phenyl or vinyl.

In this chapter kinetic studies of cationic photopolymerizations of epoxidized

soybean oil are described. Differential Scanning Photocalorimetry (DSP) was used to

49

investigate a series of polymerizations performed at a variety of temperatures. The
kinetics of cationic photopolymerizations are extremely complex and system dependent
largely because the reactivity of the propagating cation depends upon the proximity of the
counterion.7 Therefore, depending on the nature of the polymerization medium, the
propagating species may exist in any state between the extremes of closely associated ion
pairs, and free ions. The ion pairs in cationic polymerization tend to be loose ion pairs
even in moderately polar solvents because of the relatively high mobility of the cations
compared to that of the negatively charged counterions which are typically large in size.
The situation is further complicated by the fact that in bulk (solvent-free)
polymerizations, the medium changes from a relatively low viscosity liquid at the start of
the reaction, to an elastic solid at the end of the reaction, with a possible change in the
dielectric constant. Therefore the proximity of the counter ion (and therefore the
reactivity of the cation) may change during the course of the reaction.

The most fundamental description of the kinetics of cationic polymerizations
would be based upon the identification of the nature of the propagating species and the
proximity of the counterion. However, such a determination is practically impossible for
bulk polymerizations in which the nature of the propagating ion may change during the
course of the reaction. Therefore, a useful approach which provides the important
information about the effect of reaction conditions on the kinetics is to evaluate the
effective propagation rate constant defined as the proportionality constant between the
rate of polymerization and the product of the epoxide concentration and the active center
concentration. This effective propagation rate constant contains contributions from all

the types of active centers ranging from ion pairs to separated ions.

50

In general, termination of active centers in polymerizing systems occurs by
combination, disproportionation or by chain transfer to similar or different species.
Termination in cationic polymerizations is limited to chain transfer to solvent, monomer,
or another polymeric chain because, in contrast to free radical polymerizations, the
combination of active centers does not lead to termination in ionic polymerization. Due
to this fact, the initiation and termination velocities are rendered unequal and the steady '
state approximations generally held true for polymerization mechanisms cannot be
employed for the cationic polymerization scheme. In bulk polymerizations without the
use of a solvent, the termination can occur only by chain transfer to a monomer or a
polymeric chain. Although the chain transfer to a monomer causes a decrease in the
average molecular weight, the degree of polymerization or conversion remains
unaffected. Termination by chain transfer in bulk polymerizations with negligible
amounts of impurities can be considered to be a facile (surface) reaction only. Hence, the
polymerizing system can be considered to be a non-terminating or “quasi-living” system.

Differential Scanning Photocalorimetry (DSP) has been used to characterize the

‘ . ' . . . . 3
a189'0” '2 as well as cationic reactions.l

kinetics of photopolymerizations- free radic
This simple analytical method utilizes the enthalpy of polymerization to characterize the
extent of reaction with time. DSP is especially suited to the study of reactions which are
highly exothermic since it measures the extent of reaction by monitoring the heat released
from the polymerizing sample. The rate of heat released from the polymerizing sample
can be profiled with time to provide the instantaneous rate of propagation reaction which

can then be used to calculate the kinetic parameters.'4 Although this technique has been

employed to study photopolymerizations that proceed by the free radical mechanism, it

 

51

has received little attention in studying cationic photopolymerizations. This is because,
typically, cationic photopolymerizations proceed very rapidly and also have large
exotherms. The typical response time of a DSC15 "6 is on the order of 2-3 seconds which
makes it crucial for the calorimeter to remove the heat at a sufficiently rapid rate so as to
maintain the temperature constant in the sample cell. By using samples of smaller size
and also using initiating radiation of low intensity during photoinduced polymerizations,
these problems can be circumvented.

The Differential Scanning Photocalorimetric technique is very well suited to our
application because the enthalpy release is about 20 kilocalories per mole of oxirane ring
reacted. By manipulating the level of radiation and the size of the sample, the reaction
can be made to proceed at rates such that the response of the DSC is rapid enough that the
profile obtained is free of instrumental artifacts due to thermal instability.

1n the studies presented here, photocalorimetry was used to study the
photosensitized cationic polymerization of the oxirane moieties in epoxidized soybean oil
(ESO) to form a crosslinked thermoset polymer network. Although the viscosity of
epoxidized triglyceride oils is considerable (200-300 cP at room temperature), it is
completely non-toxic and also exhibits a very low vapor pressure. An added advantage is
that, since it is based on a natural agricultural feedstock, it is a renewable resource. The
films formed by homopolymerization of ESO have been shown to exhibit excellent
adhesion, abrasion and optical properties.2 Although chain polymerization of ESO has
been utilized and studied,2 a detailed analysis of the photopolymerization mechanism and

kinetics has not yet been reported. A study of the rapid cationic photopolymerization of

ESO is important because of the increasing number of applications for rapid, solvent free

52
curing of polymer films as well as thick geometries. The rapid rate of polymerization
renders the process amenable to mass production.

This contribution includes the proposition of a simple correlative reaction
mechanism for the ring-opening photopolymerization of the oxirane rings much similar to
that employed for the cationic chain polymerization of monomeric species containing
carbon double bonds.13 Experimental DSP results for different values of isothermal
system temperature and intensity of initiating light are presented and analyzed. The
reaction was initiated by a diaryliodonium salt which was photosensitized by anthracene.
The sensitizer was used to shift the initiating spectral window from the deep UV (where
the onium salts absorb light) to the near UV part of the spectrum (300-400 nm).
Isothermal DSP runs were performed to determine the kinetic constants at different
temperatures. The activation energy of propagation is obtained by plotting the
propagation rate constants against the absolute temperature of the experiment. For these
calculations, previously obtained values of initiation rate constants and activation energy
for initiation were utilized.'3 The resulting profiles obtained for the propagation rate
constants and the overall conversion of the epoxy moieties provided an insight into the
nature of the cationic photopolymerization and how it could be used to obtain desirable

properties for the polymer product.

53

1V.2. Experimental

1V.2. 1. Materials

In the photocalorimetric studies described, epoxidized soybean oil, obtained as a
gift from FMC Corporation was employed as monomer. The monomer was dried over
molecular sieves to remove any traces of water. A commercially available photoinitiator
(UV9310C, GE Silicones) was utilized. Its formulation was: 5-10 wt % linear alkylate
dodecylbenzene, ~50 wt % 2-ethyl-l,3-hexanediol, and ~50 wt % bis (4-dodecylphenyl)
iodonium hexafluoroantimonate. The diol in the initiating mixture acts as a solvent and
helps to prevent cyclization during polymerization. The dodecyl chains were added on to
the initiator to impart solubility in the monomer. Initiator concentrations specified
henceforth correspond to the total UV9310C concentration. The photosensitizer,

anthracene, was purchased from Aldrich Chemical Company and was used as received.

I V.2.2. Absorbance and Calorimetric Experiments

Absorbance measurements were conducted using a Hewlett-Packard UV-Visible
Spectrophotometer model 8452A. The DSP experiments were conducted using a Perkin-
Elmer Differential Scanning Calorimeter, model DSC-7. The experimental setup is
shown in Figure 1V .1. The light source used in the study was a 200 watt Hg(Xe) arc lamp
from Oriel Corporation (model number 6291). Magnetic stirbars were used to dissolve
the initiator (~l% by weight) and the anthracene sensitizer (~0.01% by weight) in ESO to
prepare the photopolymerizable samples. Sample sizes were about 10-15 milligrams.

The samples were placed in uncovered aluminum DSC pans obtained from Perkin-Elmer

54
and cured at different reaction conditions in the DSC. Glass slides from VWR Scientific
were placed above the sample and reference cells of the DSC to filter out the deep UV
light from the lamp (<3OO nm) so that the sensitization phenomenon could be studied,
isolated from the direct initiation of the diaryliodonium salts. This was done to simulate
conditions in industrial environments where the low and medium pressure mercury lamps
used do not emit considerably in the deep UV region. Apart from absorbing most of the
light which could directly activate the initiator, the glass slides help in preventing the loss
of thermal energy to surrounding air, thus keeping the system isothermal. Therefore, the
conversion we observe is the result of photosensitized initiation caused by anthracene
only. The contribution of thermal initiation is minimal because of the low range of
temperatures used for studying these reactions. The reactions were maintained isothermal
by a water cooling mechanism. Light intensity measurements were performed using a
High Energy UV Integrating Radiometer (Uvicure Plus, EIT, Inc.) The emission spectra
of the lamp source was obtained using an SDlOOO Dual Channel Fiber Optic
Spectrophotometer from Ocean Optics, Inc. The spectra were calibrated using the
radiometer reading in the range of 320-390 nanometers (UV-A region) by integrating the
area under the curve and setting it equal to the radiation dose from the lamp in this
spectral region. More than 60% of the irradiance in this range is provided by the mercury
line of the lamp at about 363.8 nm. This is very close to one of the anthracene
absorbance peaks and hence the sensitizer is very effective in causing the reaction. The
unique chemical structure of anthracene with three fused aromatic rings helps it to
sensitize the diaryliodonium salt by electron transfer.'9 The triplet state of anthracene

rather than the singlet state is responsible for this excited state photochemical reaction.13

55

IV.3. Results and Discussion

I V.3. I . Proposed Mechanism

Eq. 1V.1 through Eq. 1V.4 describe the proposed mechanism for ring-opening
photopolymerization of epoxidized soybean oil. In these photopolymerizations, the
cationic active centers are generated through a photosensitization reaction in which an
electron is transferred from an excited anthracene in its triplet state to the initiator.l9
Although the reaction mechanism is a composite of several photochemical reactions, it
can still be represented by a simple second order reaction scheme as in Eq. 1V.l. We
assume that there is only one active center produced by the photochemical dissociation of
each anthracene molecule. Also, we assume that all active centers generated are fully and

equally capable of propagating.

k.
A + I + hv ——‘-—-> I+ Eq. IV.1
k
1* +M—"—>M+ Eq. Iv.2
Mn+ +M—E’;>Mm,,)+ Eq. IV.3
. k.
Mp ——>Mp+l Eq. 1V.4

The rate of active center generation is directly proportional to the intensity of light
at the sensitizing wavelength. The concentrations used in the system being investigated
are such that the molar concentration of the diaryliodonium salt (6.08 mM) is about 10
times that of the sensitizer (0.505 mM). Hence, the initiation mechanism can be assumed

to be first order with respect to anthracene and zeroth order with respect to the initiator.

56

Rate of active center generation = ki [A] [1] = k,*[A] Eq. 1V5

where ki is directly proportional to the intensity of incident UV light. k.‘

is the pseudo
first order initiation rate constant and includes the effects due to the diaryliodonium
initiator concentration and the incident light intensity. An unsteady state balance on the
active propagating species gives

d[M+]
dt

 

= k,‘[A] - k,[M+] Eq. IV.6

After a little modification and integration, we get an equation for the temporal

profile of the cationic active centers.

d[M+] = k,‘[A]dt — k,[M+]dt Eq. 1v.7

“W1 = [M {RT (W “"‘)

t i

Eq. IV.8

Figure IV .2 shows a typical exotherm obtained from the calorimeter. As can be
seen, the exotherm does not return to the same value that it started with. This implies the
difference in the specific heats of the initial monomer sample and the final polymer
product. For calculation of the reaction parameters from the experimental data, different
types of base lines can be utilized. A sigmoidal base line is the most effective but it is
cumbersome to devise because of its non-objective nature. A straight base line with a
constant lepe is one of the simplest and most practical. After drawing the straight base
line and subtracting it from the exotherm curve, we get the corrected DSC data shown in
Figure IV.3.

Utilizing the direct proportionality between the heat released by the sample and

the reaction rate of propagation, we can calculate the kinetic parameters for

57

polymerization. The rate of propagation is readily obtainable from the calorimetric

measurements by utilizing the expression below:

R _ d[M] _ Height of exotherm (W / g) x density (g / l)

 

 

 

 

E . IV.
P dt AHp (1) q 9
Using Eq. 1V.8, the propagation rate can be obtained as
kpk,‘ k. k
Rp= kp[M][M*]=[Ao]k k . (e' ' t-e' 't)[M] Eq. IV.10
i ' i
[M] = [M0] _ [Exotherm area x p] Eq. 1V.11
AHP

Using the instantaneous values for [M+] and [M] obtained from Eq. 1V.8 and Eq.
IV.11 respectively, the instantaneous value of the propagation rate constant kp can be
calculated. The total activation energy of the polymerization is equal to the algebraic sum
of the activation energies of the individual steps. The negative sign for the termination

term is because it causes the destruction of the carbocation.

I V.3.2. Absorbance Results

Figure 1V.4 shows the effect of epoxidation on the spectral absorbance of the
soybean oil. The soybean oil has a sharp absorbance peak at 205 nm and diminishing
absorbance up to 250 nm. The epoxidized soybean oil has two absorbance peaks at 205
nm and 240 nm with considerable absorption until 260 nm. Figure 1V.5 contains the
absorbance spectra of the photoinitiator and the photosensitizer as a function of the

wavelength of light. The two important absorbance peaks of the diaryliodonium

58
photoinitiator are located at 205 and 240 nm. As can be seen from these two figures, the
absorbance window of the ESO monomer considerably overlaps that of the initiator.
Hence, the use of a sensitizer like anthracene is necessitated if higher reaction rates are
desired. Apart from the high absorbance in the deep ultraviolet region, the anthracene
molecule exhibits four sharp absorbance bands between 300 and 400 nm. These four
closely spaced peaks in the absorbance of anthracene are well away from the absorbance
bands of the monomer and the initiator and can be used to excite the molecule without
competitive absorption from the other components of the reactive formulation. Also, as
the anthracene dissociates, the products of the dissociation do not absorb in the same
region. Hence, these peaks decrease in size rapidly, creating a self-eliminating window
and making it possible for the light to penetrate further into the sample. This
phenomenon is very conducive for production of thicker polymers and our attempt to

utilize it is published elsewhere.l7

I V.3. 3. Calculation of Termination Rate Constants

Dark cure DSP experiments were performed to determine the termination rate
constant kt for the cationic polymerization of ESO. The polymerizing samples were
exposed to the initiating light for a short period of time and then the light source was
removed. The remaining cure was monitored without any exposure to UV light. The
time of exposure was limited to 3 minutes such that the epoxy conversion in the sample
did not exceed 5% of the total epoxy concentration. This was done to ensure that the
DSC decay profile obtained was due to the reduction in polymerization rate caused by

unavailability of the cationic active centers and not a result of monomer depletion.

59

Upon comparison of the decay profiles obtained with polymerizable samples and
just the ESO monomer, little difference was observed. This led us to believe that the
cationic active centers did not undergo termination at an observable rate. Cationic active
centers have been previously known to exhibit relatively long lifetimes and small
termination rate constants compared to their free radical counterparts. The termination
rate constants for the polymerization of divinyl ethers'3 has been shown to be an order of
magnitude less than the rate constant for photoinitiation. We conclude that the
contribution of the termination reaction is negligible and for all practical calculations, k.

can be equated to zero.

IV. 3.4. Calculation of Initiation Kinetic Constants

The initiation rate constants for different photocalorimetric experiments were
determined by solving the photophysical rate expressions developed by Nelson et al.[9
The incident light intensity on the sample was calculated by correcting the emission scan
(Figure IV.6) from the lamp source for the distance from the lamp source and the
absorbance of the cover glass slides. The intensity of light varies inversely as the square
of the distance from the lamp source. The transmittance through the glass slide can be
calculated from the absorbance data collected for it using the following equation where A
is the absorbance at a specified wavelength and T is the percent transmittance through the

sample.

A = 104%) Eq. IV.13

60
Nelson et all9 have calculated the activation energy for the photosensitization
reaction between diaryliodonium salts and anthracene from fluorescence measurements of

the photosensitization reaction to be E1 = 28.5 i 1.6 kJ mol". Using this value, the effect

of reaction temperature on the kinetics of the photosensitization reaction can be
estimated. The value of k: thus obtained can be used to calculate the instantaneous
reactive center concentration, [M+], during the photopolymerization experiments.

The emission spectrum of the Hg(Xe) lamp is obtained using a fiber optic
spectrophotometer and is shown in Figure 1V .6. The radiation dose between 320 and 390
nm was measured using a Uvicure radiometer and was used to calibrate this emission
spectrum. The data from this spectrum when combined with the extinction coefficient of
anthracene at different wavelengths provide us with the kinetic constants for the
photosensitization reaction. k; obtained as a function of temperature at different light

intensities has been plotted in Figure IV.7.

I V.3.5. Eflect of Light Intensity

To study the effect of light intensity on the propagation reaction profile, DSP
experiments were carried out at different incident light intensities, keeping the
temperature constant. To vary the incident light intensity on the sample and reference
cells, the distance between the light source and the cells was changed by elevating or
lowering the lamp or by changing the condensor and rear reflector position in the lamp
assembly. At very high intensities, the reaction does not remain isothermal and goes out
of control. At very low intensities, the reaction proceeds very slowly, such that the

polymerization is not complete in a reasonable amount of time.

61

Figure 1V .8 profiles the heat released with time for different incident light
intensities. The ordinate is a direct measure of the velocity of the propagation reaction
(see Eq. IV.9). Each of the curves shown in Figure 1V .8 is an average of several profiles
obtained at each light intensity and therefore could be wider and shorter than the original
experimental data. However, they represent the general shape and provide an accurate
representation of the average time of maximum heat flux and the exotherm area. As
expected, an increase in the initiating light intensity resulted in an increased reaction rate,
thus reaching its maximum more rapidly. As calculated above, the complete conversion
of the epoxy groups will release a total heat of 582.13 joules per gram of ESO
polymerized. Therefore, the quotient of the heat released and the above value provides us
with the percent conversion of the epoxides. For example, at an intensity of 0.518

mW/cm2 at 365 nm, the heat flux reached the maximum at 17.97 i 0.43 minutes and had

an average exotherm area of 292.81J/g. The photopolymerization at an intensity of 1.555

mW/cm2 had the maximum heat flux at 4.33 i032 minutes and an average exotherm area

of 315.68 J/g of sample. The polymerization at an intermediate intensity of 1.037

mW/cm2 peaked at 11.65 i 0.51 minutes. Although the rate of propagation increases

rapidly for experiments at higher incident radiation, the final conversion of epoxide
groups does not increase much. This may be because of the “non-terminating” nature of
the cationic active centers. Although the initiation proceeds at higher rates in the
beginning, the total number of active centers that can be generated in the course of the
polymerization is approximately the same for both cases. The slightly higher conversion

obtained in the case of higher light intensity is due to the fact that the more rapid

62
generation of active centers led to the active centers reaching the epoxide groups earlier

and hence leading to more conversion.

1 V.3. 6. Effect of Temperature

The photopolymerizations of ESO, initiated by UV9310C and photosensitized by
anthracene, were carried out at a series of temperatures from 50°C to 110°C in order to
characterize the activation energy over a large temperature range. Preliminary studies
revealed that no appreciable thermal initiation of the samples occurs at these timescales
for temperatures below 200°C. The polymerization rate profiles obtained from the DSP
experiments conducted at different temperatures keeping the incident light intensity
constant have been plotted in Figure IV.9. As can be expected, the initial reaction rate
increases with increasing temperature as illustrated by the fact that the exotherms reach a
maximum value earlier and also the peak value of heat flux is higher. Surprisingly, on
integration and comparison of the areas under the exotherm at different temperatures, we
found that the reactions at lower temperatures reach a higher final conversion, though
they start very slowly.

The conversion of the epoxy moieties in the monomer was calculated using the
theoretical value of the enthalpy of oxirane ring opening. This value is 94.5 kJ/mol of
oxirane group. This translates to about 582.13 joules/gram of ESO since the average
ESO molecule has six oxirane rings with a molecular weight of 974 grams. The
conversion at any point of time can be calculated by division of the area under the DSP

exotherm by this number.

63

Exotherm area(yg)9(%l

AH, [M01

Conversion of epoxide moities =

 

Eq. IV.14

As can be seen from Figure IV. 10, the final limiting conversion obtained is higher
for lower experimental temperatures. This phenomenon of limiting conversions is not
due to thermodynamic equilibrium effects because the high ring strain of the epoxides
causes the polymerizations to proceed irreversibly.‘8

The effect can be explained in terms of viscosity and gelation effects. Our
explanation of this phenomenon is that the decrease in rate is due to the lack of mobility
of the cationic active centers at higher conversions of the epoxide group. At
polymerizations proceeding at lower temperatures, the viscosity of the system is high and
the reaction progresses at a moderate pace such that the local viscosity rises slowly until
solidification (vitrification) of the system at which point the polymerization rates are
drastically reduced. At higher temperatures, the reaction starts to occur very rapidly
because of the low initial viscosity of the monomer. The rapid polymerization causes an
enormous increase in the local viscosity of the system, severely restricting the movement
of the cationic sites. The oxonium ions are hence “trapped”, causing a phenomenon of
limiting conversion at higher temperatures.

For each exotherm obtained for the photopolymerization reaction, a profile of the
quantity kp[M+] may be obtained using the empirical profiles of RI) and [M]. Based upon

the DSP exotherm, this product may be determined using the following equation.

 

k M+ R” E IVlS
,[ l—[M] q- -

64

Profiles of kp[M+] obtained in this manner are shown in Figure 1V.11. As shown
in this figure, the value of kp[M+] increases rapidly initially. The maximum value
attained by kp[M+] is directly proportional to the reaction temperature. After it reaches a
maximum value, it decreases slowly with time because it reaches a limiting conversion.
The shape of the profiles indicates that the apparent propagation rate constant kp changes
as the reaction proceeds. If kp had remained constant throughout the reaction, the value
of kp[M+] would have been dictated solely by the steady increase in the active center
concentration. That is, the value of kp[M+] would have increased monotonically till it led
to the complete consumption of the photosensitizer.

Further information about the behavior of kp may be obtained by coupling the
information on kp[M+] with that of k,‘. We can utilize the photosensitization kinetic
parameters for anthracene proposed by Nelson et all9 for the cationic
photopolymerization of epoxidized oils. Using the initiation expression obtained,
assuming that the polymerization reaction is non-terminating, and using the values
calculated for Rp and [M+], we can calculate the intrinsic propagation rate constant
profile. Figure IV .12 shows kp as a function of the oxirane ring conversion in the
photopolymerizable sample. The curves exhibit a similar characteristic shape as the
curves for kp[M+] of Figure IV.11 although the plateau value of the rate coefficient is
higher, owing to the small numerical value of the active center concentration. The
similarity in the shapes indicates that the effects observed in kp[M+] are relatively more
dependent upon kp, rather than [M1].

The initial large rapid increase in the propagation rate coefficient can be explained

in terms of diffusional mobility effects. Initially, the propagation reaction caused by the

 

65
cationic active centers is slow because of the close proximity of the negatively charged
counterions. As polymerization proceeds, it leads to rapid increase in the local viscosity
of the reaction system. On account of their larger size compared to the cationic active
centers, the negatively charged counterions are unable to diffuse rapidly and thus are
unable to keep up with the considerable mobility of the cations attained through
propagation by reaction with the monomeric molecules. This ‘reaction diffusion’

mobility has been mentioned in the literature,”"3’2°

and can be a very probable
contributing factor in crosslinking polymerizations. The large hexafluoroantimonate
counterions are thus separated from the carbocations because of their limited diffusivity.
Since separated ions are much more reactive compared to ion pairs,°‘7 this leads to a large
increase in the propagation rate coefficient.

The steady and comparatively slower decrease in the rate coefficient after it
reaches the maximum (Figures IV.11 and IV.12) can be explained in terms of trapping of
active centers in the crosslinked polymer formed. This is the reason that a limiting
conversion is observed for the cationic photopolymerizations for epoxides and vinyl

3
ethers. '

As previously mentioned, this analysis is based upon the assumption that one
carbocationic active center is produced for each anthracene molecule consumed, and that
all reactive centers thus generated are equally and fully capable of propagating the
polymerization reaction.21

For calculation of activation energy of the propagation reaction, we need the
kinetic constant unhindered by the diffusion parameters. An average initial reaction rate

was determined for each temperature by measuring the time required to reach 20 %

conversion of the monomer. A low value of the conversion was chosen because the

66

constants will represent the kinetic parameters in case of excess monomer. At higher
conversions, the reaction kinetics might be dictated by the availability of the monomer in
the vicinity of the initiator. Using the relationships presented in Eq. IV.9 through Eq.
IV.13, the apparent propagation rate coefficient was calculated at different temperatures.
An Arrhenius fit of the intrinsic propagation rate coefficients at the specified conversion
provides us with the activation energy for the propagation step (F1, = 4.3 kJ/mol).

As the complexity of the chemical structure of an organic compound increases, its
viscosity becomes a more acute function of the temperature. The effect of temperature on

viscosity of liquids can be estimated by the Guzman-Andrade equation.22

it = AeB” Eq. IV.16

As Figure IV .13 shows, the viscosity of the epoxidized soybean oil is higher than

that of soybean oil and is a stronger function of temperature. It follows an exponential
curve and in the range of temperatures used for calorimetric experiments, it decreases
from 160 CF to below 30 CF. Figure 1V.l4 shows the close agreement of the ESO
viscosity to that predicted by the Andrade equation. As the temperature is increased, the
intrinsic viscosity of the system decreases. We know that the diffusivity is an inverse
function of the viscosity and hence the propagation proceeds relatively rapidly at higher
temperatures. Because of this accelerated reaction, higher local conversions are reached,
hence causing local increases in viscosity. These increases cause microscopic gelation
and entrapment of the cationic active centers. On the other hand, polymerizations
conducted at lower temperatures proceed at a much slower pace, causing the viscosity to
increase more uniformly across the whole system. The gelation can hence be called more

“macroscopic” in nature. For this reason, we see the increase in the propagation rate

67
coefficient continue for longer times at lower temperatures. If our supposition of gelation
effects is true, we should be able to see a definite cross-over for the intrinsic propagation
constants for the various temperatures at a definite percentage conversion. From Figure
IV. 12, this appears to be around 23% conversion of the epoxide moieties. The theory of a
diffusion controlled reaction is also corroborated in the low value calculated for the

activation energy.

IV.4. Conclusions

In this contribution, Differential Scanning Photocalorimetry (DSP) has been used
to study the photopolymerization kinetics of epoxidized soybean oil (ESO). The DSP
technique offers a direct method for evaluating the heat generated during a
polymerization reaction. The reaction exotherms were used to characterize the
polymerization kinetics and evaluate the kinetic constants. The DSP, experiments were
challenging to design because of the rapid polymerization rates coupled with the highly
exothermic nature of the reaction. Using small sample sizes and low light intensities, we
could increase the reaction time so that the kinetic parameters could be easily
characterized.

Isothermal DSP experiments were conducted for a series of unsteady state ESO
polymerizations at different temperatures and incident light intensities. The initiation
system used was a diaryliodonium hexafluoroantimonate salt photoinitiator
photosensitized by anthracene. This system has been studied by several investigators and

the photosensitization mechanism and kinetics are well characterized. Using the

68

sensitization rate constants from these, the kinetic constants for the propagation of the
ring—opening polymerization of ESO were obtained. As expected, an increase in the
initiating light intensity resulted in an increase in the reaction rates. However, the final
conversion obtained remained almost constant. The effect of temperature on the
polymerization was much more complicated. An increase in the isothermal reaction
temperature led to an increase in the rate of propagation but a decrease in the final
conversion obtained. This can be explained as an effect of the lack of mobility of the
cationic active centers at higher conversions of the epoxide group. With polymerizations
proceeding at lower temperatures, the initial viscosity of the system is high and the
reaction progresses at a moderate pace such that the local viscosity rises slowly until
solidification (gelation) of the system. At this point, the polymerization rates are
drastically reduced. At higher temperatures, the reaction starts to occur very rapidly
because of the low initial viscosity of the monomer. The rapid polymerization causes an
enormous increase in the local viscosity of the system, severely restricting the movement
of the cationic sites. The oxonium ions are hence “trapped” thus causing a lower limiting
conversion.

To determine the termination rate constant k(, dark cure experiments were
conducted using the DSC. The decay profile of the exotherm from the point at which the
initiating light source was removed, seemed to be very similar for a polymerizing and a
non-polymerizing sample. This led us to believe that the cationic active centers possess
very long lifetimes and relatively small termination rate constants compared to their free

radical counterparts. This can be attributed to the fact that the cationic active centers do

69
not terminate by combination with other cations. Moreover, the polymerizing system is
free of moisture and solvent thus leading to a decrease in chain transfer.

From the DSP reaction profiles, an apparent propagation rate constant kp was
calculated. As the reaction temperature was increased, a higher value of kp was observed.
The apparent propagation rate constant for an isothermal photopolymerization reaction
underwent a rapid initial increase, leveled to a plateau and then underwent a gradual
decrease. This implied that the propagation rate constant changed drastically as the
reaction proceeded. The initial large increase can be explained to be due to changes in
the reactivity of the cationic species due to the proximity of the counterion. During the
polymerization, the large hexafluoroantimonate counterion of the initiator experiences a
rapid decrease in mobility as the viscosity increases. Meanwhile, the active carbocation
retains its diffusional mobility by reacting with neighboring oxirane rings, thus leading to
separation of the two species. This increases the reactivity of the cationic active center.
The final decrease in kp is due to trapping of the active centers in the rapidly formed
crosslinked polymer network. Therefore, a limiting conversion is observed in spite of the
presence of “live” cationic species.

The profiles obtained for the propagation rate constant have provided valuable
insight into the nature of crosslinking photopolymerizations. Understanding these kinetic
aspects is important because of the rapidly increasing number of applications for rapid,
solvent free curing polymerizations. Conventional methods of polymerization have been
discounted for production of polymeric materials from agricultural feedstocks. This

research on cationic photopolymerization of epoxidized oils has provided proof that

70
useful polymers can be obtained from renewable natural resources by a process amenable

to rapid, mass production.

71

  
 
 

Dichroic Mirror
280-400 nm

 

Computer
controller

 

Figure 1V.1 Experimental setup for Differential Scanning Photocalorimetry Studies.

72

1000 -

 

Reaction Exotherm

300 , --------- Corrected Baseline

-<—_--
..--~~--.---..
__....‘____”
""-~-...-...._

 

600

400

Heat flux, watts

200

 

o a a a I n I L I n l m I L I

40 50 60 70 80 90

 

Time, minutes

Figure IV.2. Typical Exotherm Obtained During Polymerization of Epoxidized
Soybean Oil (ESO).

73

25

N
O

15

10

Heat Removed, watts/gm

 

 

 

Time, minutes

Figure IV.3. Corrected exotherm obtained from the DSC after subtracting the baseline.

L4-T

L2<+

Absorbance

 

74

— Soybean Oil

2. ------ Epoxidized Soybean on

 

 

 

190 240 290 340 390 440 490

Wavelength, nm

Figure IV.4. Effect of epoxidation on the UV and visible absorbance of soybean oil.

75

 

 

 

""""" Anthracene
3.5 .. "" -- UV9310C
3.0 'i I: \ 1‘\
g 2.5 "’ . :: ‘\‘.' ‘2 ' . \
(U . .. “ d. ' ‘ r ‘
E 2.0 .. /\ ~ .' ‘ - '. .' '.
8 I \ I. ‘\ : ‘3 t
g 1.5 i' I, \"/:f\\ I“ :' t:
1.0 1' I ' \ ‘\ o " |
I 'r \\ \‘ ’. “
0.5 J' I " \\ X. '.‘;' \‘
n— I " \ ‘ :L‘ ' - \‘
0.0 ............ : ?=-‘ —————— -: $1“
190 240 290 340 390
Wavelength (nm)
Figure 1V.5.

Spectral absorbance of diaryliodonium initiator and anthracene photo-
sensitizer as a function of wavelength.

76

.o .o .o
(a) h 01

Intensity, mWIcm2

F’
to
If

 

 

 

 

 

 

0A *

 

0 l l r
190 240 290 340 390 440 490
Wavelength, nm

~0-
...

Figure IV.6. Emission scan from the Hg(Xe) lamp source used for DSP experiments.

77

 

 

0.0030 —
(A
00025 - Light Intensities at 365 nm
—I— 0.5183 mW/cm2 / "
0.0020 - —-e—— 1.0366 mW/cm2 . "
/- ,e
A1555 mW/cm2 A" /
'5. 0.0015 - //
e .. i / /
x._ .. ,/‘ /
/ / /”.’
0.0010 - it // e
d ‘ '1' I I. "l/
/'/./ ./
0.0005 - “
0.0000 T - l . i ' I ' ' ' 7
20 40 50 so 100 120

Temperature, 0C

Figure IV.7. Initiation rate constant as a function of temperature at different UV light
intensities.

78

00
01
I

--—- 0.518 mW/cm2
------ 1.037 mW/cm2
---------- 1.555 mW/cm2

8
L

A

5‘: 8 8
i i i

J

Heat Flux, watts per gram of sample
‘1’

m
1

 

 

 

Time in minutes

Figure 1V.8. DSP exotherms for the cationic photopolymerization of ESO at different
incident light intensities.

79

 

,_ 0.00030 +

g . p ——5000
U)

1 0.00025 - ----------- 70°C
32 ‘ -------- 900C
3 0.00020- i‘. ---—--110°C
0

E

m“ 0.00015

.9“

(U

m 0.00010

C

.9

‘8, 0.00005

(0

O.

9

0. 0.00000

 

 

 

 

Time, minutes

Figure IV.9 Overall rate of propagation versus time for ESO polymerization with 1%
diaryliodonium hexafluoroantimonate photosensitized by 0.01%
anthracene.

80

80-

Light Intensity (365 nm): 0.518 mW/cm2

Percent Conversion of Oxirane in ESO

 

 

O ' -' 1 I l I n I A I a I L I A I 4L I

0 10 20 30 40 50 60 70 00 90
Time of Polymerization, minutes

Figure IV. 10. Conversion profile of epoxy moieties in ESO during photopolymerization
at various temperatures.

81

0.1

0.01

 

k [M+], see:1

p

 

 

80

 

Percent Conversion

Figure IV.11. kp[M+] as a function of conversion for ESO photopolymerization at
various temperatures.

82

1000

'. _-......-._
.0 s.

.1

 

 

 

 

 

 

g r Temperature \ ‘ ‘.

‘ l

E; ' 50W: \ a '
O 1‘ '1'

E 1 .- ........ 70°C “‘3, i

E E .............. 90°C

.0. i- I

x .. ' 7 T 1 1 0°C | i
0.1 E- i

0‘01 1 I 1 I 1 I i 1 I
0 20 40 60 80

Percent Conversion

Figure IV.12. Apparant propagation rate constant kp as a function of conversion for E80
photopolymerization at various temperatures.

Viscosity, cP

Figure IV.13.

83

 

 

.
350- Q
\‘I
300. \ 0 Epoxidized Soybean Oil
\\ A Soybean Oil
250- is
'l \\
\
200- \\
150- \Ex
._\
100‘ \0\
i A. ‘o\0
5°" A.
--A.. ‘
0 r l r 1 l r 1 I I 1 I ' l
0 10 20 30 40 50 00 70 80 90 100

Temperature, 0C

Viscosity of epoxidized soybean oil as a function of temperature.

84

 

 

//
//
/'/‘
/
/
5.5 -4
//t
/
/
u //
y
/
5.0 - /
"1(11). /{
u in GP . //
/// I Experimental Data
4.5 q ./ ------- Fit to Andrade Equation
/
//
4.0 .. /
/ r ' f r r r 1 r 1
0.0029 0.0030 0.0031 0.0032 0.0033
1/T, T in Kelvin

Figure 1V.14. Viscosity of epoxidized soybean oil as a function of inverse absolute
temperature.

10.

ll.

12.

13.

14.

15.

16.

17.

85

1V.5. List of References

R. Gachter, H. Muller, Eds., Plastics Additives Handbook, 2“(1 Edition, 282,
(1987).

J.V. Crivello, and R. Narayan, Chem. Mater., 4, 692, (1992).

S. Qureshi, J.A. Manson, J.C. Michel, R.W. Hertzberg, L.H. Sperling, in
Characterization of Highly Crosslinked Polymers, ACS Symposium Series 243,
SS. Labana, R.A. Dickie, Eds., ACS, Washington, DC, 8, (1984).

AM. Fernandez, J .A. Manson, L.H. Sperling, in Renewable Resource Materials,
L.H. Sperling, Ed., Plenum Press, New York, 177, (1986).

I. Frischinger, S. Dirlikov, Polym. Mater. Sci. Eng. Preprints, 65, 178, (1991).
G. Odian, Principles of Polymerization, 3rd Edition, J. Wiley and Sons, (1991).

J .P. Kennedy, E. Marechal, Carbocationic Polymerization, John Wiley and Sons,
New York, (1982).

T. Doornkamp, and Y.Y. Tan, Polymer Commun., 31, 362, (1990).

J .G. Kloosterboer, Adv. Polym. Sci., 84, 1 (1988).

J .G. Kloosterboer, and G.F.C.M. Lijten, Polymer, 28, 1149, (1987).

K5. Anseth, C.M. Wang, and C.N. Bowman, Macromolecules, 27, 650 (1994).
K.S. Anseth, C.M. Wang, and C.N. Bowman, Polymer , 35, 3243 (1994).

E.W. Nelson, J.L. Jacobs, A.B. Scranton, K.S. Anseth, and C.N. Bowman,
Polymer, 36(24), 4651, (1995).

T. Doornkamp, and Y.Y. Tan, Polymer Commun., 31, 362, (1990).
C. Decker, and K. Moussa, J. Coatings Technology, 62, 55, (1990).

J .G. Kloosterboer, and G.F.C.M. Lijten, in Cross-linked Polymers, R.A. Dickie,
S.S. Labana, and RS. Bauer, Eds., ACS, Washington DC, 409, (1987).

S.K. Moorjani, B. Rangarajan, A.B. Scranton, Characterization of the Axial Light
Intensity Gradient in Photosensitized Cationic Polymerizations, Polymer, under
preparation.

l8.

19.

20.

21.

22.

86

P. Kubisa, in Cationic Polymerizations - Mechanisms, Synthesis, and
Applicaitons, K. Matyjaszewski, Ed., Marcel Dekker, Inc, New York, (1996).

E.W. Nelson, T.P. Carter, and AB. Scranton, J. Polym. Sci., Part A: Polym.
Chem, 33, 247, (1995).

K.S. Anseth, and C.N. Bowman, Polym. React. Eng., 1, 499, (1992).
E.W. Nelson, T.P. Carter, and AB. Scranton, Macromolecules, 27, 1013, (1994).

J. Andrade, Nature, 125, 309, 582, (1930).

CHAPTER V.

THE EFFECT OF VISCOSITY ON THE RATE OF
PHOTOSENSITIZATION OF DIARYLIODONIUM SALTS BY
ANTHRACENE

Abstract

An attempt has been made to characterize the effect of viscosity on the reaction
rate of photosensitization of diaryliodonium salts by anthracene. A well-established
theory for diffusion controlled reactions has been used to qualitatively explain the effect,
and a comparison of the predicted behavior with experimental data has been presented.
Upon addition of an adjustable parameter, the model has been shown to reasonably
accurately describe the photosensitization rate at any viscosity. In conjunction with the
knowledge of the activation energy of the reaction, the model can be used to predict the

photosensitization rate at any given combination of temperature and viscosity.

V.1. Introduction

The utilization of light to induce polymerization reactions results in more than a
few advantages. In addition to attaining rapid reaction rates, it also helps in exercising
superior spatial and temporal control over the reaction. Photopolymerizations require an

order of magnitude less energy than conventional thermal means of polymerization.‘

87

88
Since the process precludes the need for solvents, it leads to a decrease in the emission of
volatile organic chemicals, thus rendering the process environmentally benign.
Cationic photopolymerizations of epoxides and vinyl ethers offer tremendous

potential in high-speed, solvent-free curing of films and coatings. The polymers formed

exhibit excellent clarity, adhesion, abrasion and chemical resistance?4 In addition, UV-

initiated cationic polymerizations offer several processing advantages. These reactions
are not inhibited by oxygen and can be carried out without an inert atmosphere. In
addition, cationic species do not terminate by combination with each other, as do free
radicals. The cations generated are able to diffuse much faster than the large, negatively

charged counterions. Because of these reasons, cationic polymerizations proceed long

after the irradiation has ceased and may penetrate areas where the light does not reach.5’6

Finally, several vinyl ether monomers are available, which exhibit low vapor pressures,

low viscosities and low toxicity, but polymerize rapidly to form films exhibiting excellent
properties-“4 The excellent polymer properties and processibility have resulted in a

surge in research devoted to characterizing the mechanism and kinetics of these reactions,
as well as the physical properties of the polymeric films obtained. Because of the
advantages listed above, photoinduced polymerizations are being increasingly utilized in
making a large variety of products such as coatings for various substrates, inks,
photoresists used in semiconductor applications and also recently, fiber-reinforced
composites.15"°

Interest in UV-initiated cationic photopolymerizations was fueled by the

development of two classes of thermally stable photoinitiators. Diaryliodonium and

89

triarylsulfonium salts have been used in the last decade to rapidly photoinitiate

l7

polymerizations in solvent free systems like epoxides and vinyl ethers.3’ While these

salts are most efficiently initiated by deep-UV (225-275 nm), their applicability is
severely challenged by the fact that most monomers also absorb in the same region of the
ultraviolet spectrum and may compete for the incident radiation. The absorbance range of
these initiators may be expanded by combined use with photosensitizers or co-initiators
such as anthracene and perylene which shift the effective spectral window to well above
300 nm, the region where most industrial medium and high pressure mercury lamps have
strong emission bands. The absorbance of a typical diaryliodonium photoinitiator and
anthracene as a function of wavelength is shown in Figure V.1. Also, if the products of
photosensitization do not absorb in the same region as the photosensitizer, this fact can be
used to produce thick polymer products.‘8 Understanding the mechanism of
photosensitization is critical to the optimal use of photopolymerization techniques to
produce adhesive coatings as well as thick polymer composite parts.

Photosensitization reactions are very rapid, with the life times of the excited
singlet and triplet states of anthracene being on the order of nanoseconds and
microseconds, respectively.17 As a result, reactions involving these species, like several
other chemical reactions, are limited by the rate of diffusion of reactants and/or products.
The time taken for any species to diffuse is directly proportional to the square of the
distance to be traveled. On a molecular scale, these distances are normally on the order of
a few nanometers and as a result, only very rapid reactions tend to be affected by
diffusion. It is our hypothesis that the photosensitization reactions are diffusion limited

and factors such as viscosity play an important role in determining the rate of reaction.

90

Smoluchowski theoryl9 explains the behavior of rapid chemical reactions which
are hindered by diffusion. This theory assumes that equally sized, spherical, reacting
molecules diffuse by a simple random walk. When combined with Stokes law and the
Einstein diffusion equation, it results in the Smoluchowski-Stokes-Einstein (SSE)
equation which states that the rate constant of a diffusion controlled reaction is inversely
proportional to the solution viscosity.

In this contribution, we have studied the effects of viscosity on the rate of the
photosensitization reaction. Using steady state fluorescence spectroscopy, we have
monitored the rate of reaction between anthracene (photosensitizer) and diaryliodonium
hexafluoroantimonate salt (photoinitiator), in propanol/glycerol solutions of different
viscosities. The results have been analyzed by numerical solutions of the photophysical
kinetic equations in conjunction with the mathematical relationships provided by the
Smoluchowski theory. The applicability of a modified Smoluchowski theory by the

addition of an adjustable parameter has also been investigated.

V.2. Experimental

V.2.1. Materials.

The 1-propanol, glycerol, and anthracene were obtained from Aldrich Chemical
Company and were used as received. The initiator (UV9310C; GE Silicones) had a
composition of 5-10 wt % linear alkylate dodecylbenzene, ~50 wt % 2-ethyl-l,3-
hexanediol, and ~50 wt % bis (4-dodecylphenyl) iodonium hexafluoroantimonate. The

chemical structure of the photoinitiator and photosensitizer are presented in Figure V.2.

91
In this initiator, various dodecyl isomers have been attached to the phenyl rings of the
diphenyliodonium salt to impart solubility in the monomer and do not have any effect on
the reactivity of the initiator.I The hexanediol has been added to hinder the formation of

cyclic polymers.

V.2.2. Fluorescence Measurements.

The fluorescence experiments to measure the effect of system viscosity on the
sensitization rate were performed in the LASER Laboratory at Michigan State University.
A schematic diagram of the experimental setup is shown in Figure V3. The excitation
source used was a Coherent Innova 200 Argon ion laser operating at 363.8 nm. The same
light source serves to excite the anthracene to cause the photosensitization as well as help
us to study the fluorescence caused by its excitation to the singlet state. To ensure the
simultaneous start of illumination and fluorescence data acquisition, a Newport 845HP01
digital shutter system was opened with an electronic pulse from the detector controller.
Unfocused laser radiation of 15 milliwatts in a beam of about 3 mm diameter was
directed upon the quartz capillary tube (1 mm id, 2 mm o.d., 25 mm long) placed
perpendicular to the beam direction. A Scientech 362 power/energy meter was used to
measure the illumination. The fluorescence was collected at a right angle from the
incident beam and right angle from the longitudinal axis of the capillary tube. A Spex
1877 Triplemate spectrometer was used to analyze the fluorescence signal. The
spectrometer had two stages - a subtractive dispersion filter stage and a spectrograph
stage. A diffraction grating of 150 gratings per inch was employed. A gated EG&G

Princeton Applied Research Model 1530 C/CUV CCD, cooled with liquid nitrogen to -

92
120°C to minimize dark charge levels, was used to detect the signal, and was interfaced
with an EG & G Princeton Applied Research Model 1463 OMA 4000 detector controller.
The OMA 4000 software was used to analyze the fluorescence data.
All steady state fluorescence experiments were conducted with the sample placed

in a thermo-statted cell with temperature maintained at 30°C. The concentrations of

anthracene and initiator used were 0.000505 and 0.00608 moles per liter respectively.
The relative quantities of solvents (n-propanol and glycerol) were adjusted to achieve
solutions of different viscosities, while maintaining the same molar concentration of the

reactive solutes.

V.3. Results and Discussion

V. 3. 1. Steady State Fluorescence Measurements

Figure V4 is a time overlay chart and shows the decay in the fluorescence
spectrum of anthracene in a representative sample. Since the photosensitization reaction
results in the consumption of anthracene, monitoring the fluorescence decay of
anthracene provides a convenient method for measuring the rate of reaction. The initial
increase in the fluorescence can be attributed to an instrument artifact and is not a
consequence of the reaction mechanism. The intensifier in the detector operates in a
fashion similar to a photomultiplier tube. When a photon strikes the intensifier surface, a
number of photons are generated. This process requires incident light to initiate and takes

a finite length of time to reach its full value of enhancement after multiple reflections of

93
the photons. This artifact does not affect our technique of data analysis and the important
kinetic information can still be deduced.

During the reaction, as the anthracene reacts to form free radicals and radical

cations,2

0 a majority of it degrades to dihydroanthracene. This product does not possess
appreciable absorbance in the region between 300 and 400 nm. As time proceeds, the
slight shift in the fluorescence peak is due to partial degradation of the anthracene to 9,10
divinyl anthracene and 9,10 diphenyl anthracene. By monitoring the steady state
anthracene fluorescence peak at 425 nm with reaction time, we can obtain a measure of
the concentration changes of anthracene in the sample. After fitting the fluorescence
profile to a mathematical model, normalizing the fluorescence intensity with respect to
the initial molar concentration of anthracene in the sample allows us to obtain the
anthracene concentration as a function of time (Figure V.5). Although the
photosensitization reaction is a second order reaction, the concentration profile of
anthracene closely follows an exponential decay which is typical of a first order reaction.
This is due to the fact that the initiator concentration exceeds that of the photosensitizer
by more than a factor of ten, resulting in a pseudo-first order reaction. This steady state
fluorescence experiment can be repeated for photosensitizer and initiator solutions of
various viscosities and the concentration profile obtained can be used to arrive at a decay
rate coefficient for anthracene at different viscosities.

The effect of viscosity on the rate of reaction is shown in Figure V.6, where it can
be seen that the rate coefficient decreases as the viscosity of the system is increased. As

we move from a solution of the reactants in pure propanol to a solution in pure glycerol,

the viscosity of the system rises by three orders of magnitude. Steady state fluorescence

94
experiments of these solutions show that the reaction rate coefficient decreases as we
increase the viscosity. This is commonly observed in diffusion-controlled reactions, and
therefore we propose to use Smoluchowski theory for diffusion-limited reactions to

understand and explain this behavior.

V. 3.2. Analysis of the Photosensitization Mechanism

The photosensitization reaction is the sum of many individual reactions between
various reaction species. The kinetics of the various photophysical reactions occurring
during the process of photosensitization in an inert solvent have been described by Nelson
et al.,2° who have modeled the photoinitiation process as a set of nine reactions. These
reactions, which are shown below, represent the electronic state transitions of anthracene
(A) as well as reaction of excited state anthracene (S—singlet, T-triplet) with the initiator
(I). The symbols Rs and Rt denote the reactive centers formed by reaction with the onium

salt by the singlet and the triplet state anthracene respectively.

 

 

 

 

 

 

k
A+hv< abs >S Eq.V.1
k
S flmr>A+hv Equ
k-
S——J£—9T Equ
k
S "onflum >A Eq. v.4
k h
T "Mp as >A Eq.V.5
kdimer \
S + A , 2A Eq. V.6
kh
T pas >A+hv Eq.V.7

 

k
S+I ”"fluor >Rs Eq.V.8

95

ksenphos

 

T + I > Rt Eq. v.9

This set of photochemical reactions can be well understood with the help of a
energy level diagram for anthracene (Figure V.7). As shown in the figure, the ground
state anthracene absorbs the incident light energy to reach the excited singlet state. The
anthracene in the singlet state may take one of four reaction paths. It can revert back to
the ground state with or without the simultaneous release of light energy in the form of
fluorescence. It can transfer an electron to a neighboring onium salt molecule to generate
a cationic active center. Altemately, it can undergo an intersystem crossing to form the
triplet excited state. The anthracene in the triplet state can go back to the ground state
with or without simultaneous release of phosphorescence. The triplet state is also capable
of donating an electron to an initiator molecule to form a reactive cationic species. The
reaction in which an excited singlet state anthracene interacts with a ground state
anthracene (Eq. V.6) has not been shown in the energy level diagram. This reaction has
very little bearing on the overall kinetics and can be neglected for all practical

calculations.

V.3.3. Simulation of the Photosensitization Mechanism

To simulate and model the photosensitization of diaryliodonium
hexafluoroantimonate by anthracene, the kinetic parameters for the equations presented
above were either obtained from the literature or independently established. Based upon
the known values of the laser power and the beam cross sectional area, the photon flux

onto the sample was calculated. Using the value of the photon flux and the extinction

96
coefficient of anthracene at 365 nm, the value of kabs was calculated to be 5.979 s'1 for
our experiments. Using fluorescence lifetime experiments, the sum of the rate constants
. 17,20 8 -1
kfluor and knonfluo, was determined by Nelson et al to be equal to 1.38 x 10 s .
Similar experiments with phosphorescence lifetime experiments have shown the sum of

5 -l .
kphos and kmmphos to be 1.0 x 10 s . By conducting fluorescence and phosphorescnce
lifetime experiments in the presence of initiator, the values of kmfluo, and ksenphos have

been obtained to be 8.20 x 107 lit mol'1 8'1 and 2.50 x 106 lit mol'1 5’1 respectively. The

kinetic constant for intersystem crossing between the singlet and triplet states of

anthracene in solution has been reported to be between 7.38 x 107 to 1.12 x 108.21’22 An

average of the reported values (0.929 x 108 s") was utilized in our calculations. The
reaction between the singlet and the ground state anthracene has a kinetic constant of 4.0
x 109 lit mol'] s". These kinetic parameters were used in conjunction with the mass

balance and the resulting set of differential equations were solved simultaneously using
the Rosenbrock Method for stiff systems. 23

Figure V.8 contains the simulation results in the form of concentration profiles
for the various chemical species in the reaction system. The anthracene concentration
decreases exponentially with time, while the active centers generated from both the
singlet and the triplet states of anthracene increase monotonically. The singlet and triplet
state concentrations remain very low compared to those of the active centers generated
from them. This can be attributed to their very short lifetimes. At the start of
illumination, their concentrations increase with time, reach a maximum and then begin

decreasing. The triplet state concentration is always 2-3 orders of magnitude greater than

97
the singlet state. This fact, coupled with the fact that the lifetime of the triplet is higher
than the singlet, proves that under the reaction conditions, the triplet is responsible for

more than 90% of the photosensitization reaction.l7

 

 

 

d A
g : kubs (I IA] — [A]) + (k/luur + kmmrud )[ IA] + (khhns + knmirad )[3A]
Eq. V.10
d lA
—L—l : kabs ([A] _[1A]) — (kflunr + kmmrad )[IA] — kisc[lA] - kSm/IWVPAHI]
dt Eq. V.ll
3
dl: A] : kisc[1A] — (kphos + knonrad )[3A] - k5€"PhO-‘[3A][I]
Eq. v.12
fl]- : "ksenflmHIlAIIII - k.reriphns[3A][I]
dt ‘ Eq. V.l3
d[RS] : k.retijlimr[lA ][1]
dt Eq. v.14
“R” = k....,,....[3AlUl
dt Eq. v.15

The values of the rate constants suggest that the photosensitization reactions are
extremely rapid, and may be diffusion controlled. In a polymerizing system, as the
reaction proceeds, the viscosity of the reaction system increases rapidly, thus drastically
affecting the photosensitization reaction velocities. Although the increase in viscosity has
no effect on the single order reactions in the scheme presented above, the bimolecular
reaction rates will be adversely affected. The second order reactions of interest to us are
those that lead to generation of active centers from the singlet and triplet states, i.e. Eq.
V8 and Eq. V.9. Equation V6 is also a bimolecular reaction leading to reaction between
the singlet and the ground state anthracene. To simulate the changes in the reaction

coefficients, we need to use a theory which describes diffusion-limited reactions.

98

V.3.4. Diffusion Limited Reactions - The Smoluchowski Theory

A chemical reaction between two reactive solutes in an inert solvent can be
pictured to proceed in two stages - the approach of the reactants from a large separation
distance to a distance at which the chemical forces become dominant and the subsequent

reaction while in this close configuration. This can be schematically described as:

k k
A + B a——'d—“— [AB] —-§—> Products
k-d

Depending on the nature of interaction between A and B, the species [AB] can be
described by different terms. If [AB] is at the saddle point of the potential energy surface
between A and B, it can be described as an activated complex. When A and B are in
contact, but little or no interaction takes place between them, then [AB] can be termed as
a collision pair. Usually, the type of interaction is hard to decipher, but will most
probably be between these two extremes. The separation distance between the chemical
species may extend from essential contact to where they are separated by about one radius

of the solvent molecule. The rate of reaction can be calculated as follows:

 

kakd
Rate = ka[AB] = k [A][B]
'4 Eq. 16

Assuming a steady state reaction and that the complex [AB] is in very small
concentration, the overall rate coefficient is kdka/(k-d+ka). When the reaction of the
encounter pair to form products becomes rapid compared to the diffusive forward and
backward processes, the rate of formation of products becomes more dependent on the

rate at which the reactants can diffuse together to form the encounter pair.

99

Smoluchowski") suggested that the reactants A and B have to diffuse together until they
are close enough to interact chemically. For statistical and mathematical simplicity, this
theory assumes that A is the limiting reactant and B diffuses towards A to try and deplete
it by reaction. Because A and B disappear upon reaction, there is a concentration gradient
of B around A. From this concentration gradient of B about A, the flux and hence the
current of B towards A can be calculated. The rate of reaction is simply the current of B
diffusing towards A multiplied by the molar concentration of A.

A simple transport equation

a_P_ 2 _ fill Zip
8t_DV p—D[ar2+r8r Eq.V.l7

can be written for the diffusion of B, assuming a spherically symmetric distribution of B

about A. The density distribution of species B around A can be given by:

= [310)
[310

 

p(r,t) Eq. V.18

The initial density distribution at time t=0, i.e. the initial condition can be written
as:
p(r,0) = 0 (r S R) Eq. V.19

l (r>R)

The boundary conditions during reaction suggested by Smoluchowski and co-
workers are:
p(r —-> 00,0 =1 t 2 0 Eq. V.20

p(R,t) =0 :2 0 Eq. v.21

100
To solve Eq. V.17 subject to the density distribution given by Eq. 18, the
boundary conditions given by Eq. 20 and Eq. 21 and the random initial distribution
represented by Eq. 19 is straightforward. Using Laplace transforms and few

mathematical maneuvers, we can arrive at the expression for the rate coefficient:

 

a R
k(t)=41tR2D—p =41tRD 14.—3; Eq. v.22
3’ R (ED!) '

Typical values of R and D are 0.5 nm and 10.9 mzs'], respectively. For these

values and at times of the order of seconds, the first term in Eq. V.22 becomes relatively
more important.
k(oo) = 41tRD Eq. V.23

The Einstein equation for diffusion in non-reactive solvents is:
D = — Eq. V.24

where C, is the friction coefficient, k3 is the Boltzmann constant and T is the absolute
temperature. Stokes law shows that
Q = 61ma Eq. V.25
where a is the radius of the diffusing particle and n is the viscosity of the solution. This
equation may be off by as much as 50% since Stokes law is not exactly valid on a
molecular scale, but the general trend should remain valid.
From the above equations, we see that the reaction rate coefficient is inversely
proportional to the viscosity of the reaction system. Hence, according to Smoluchowski-
Stokes-EinsteinI9 theory, the reaction rate constants should correlate with the reciprocal

viscosity. However, in the reaction mechanism described above, only bimolecular

101
reactions will be diffusion controlled, and therefore affected by viscosity. We have
therefore adjusted the kinetic constants kdimer, ksenfluor and ksenphos to account for the

increased viscosity.

 

 

111
k = k -- E .V.26
senfluor 112 senfluor I '12 q
ksenphos : ksenphos 71—1 Eq. V.27
‘12 Ill 112

 

 

The differential equations were then solved using these viscosity-adjusted
constants. The results of the simulations are plotted in the form of anthracene decay
profiles in Figure V.9. At low viscosities, the decay rate constant is very high and leads
to rapid exponential decrease of anthracene concentration. At higher viscosities, the
decay rate is much slower. The rate constants obtained from the slopes of these profiles
are compared with experimental data in Figure V.10. It can be seen that while the
simulations describe the trend seen in the experimental data, they are not quantitatively
accurate. The differences are attributable to several reasons. First of all, the
Smoluchowski analysis assumes that the reacting molecules are spherical and of
approximately equal size. In reality, the anthracne molecule with its fused aromatic rings
has a hard, planar structure whereas the initiator molecule having a linear chain of 20
carbon atoms and two phenyl rings has a long molecular structure.

By modifying the simplistic inverse behavior suggested by the Smoluchowski
theory and inserting an adjustable parameter as the exponent for the viscosity term, we
can better describe the trend seen in the experiment. Figure V.11 compares the

experimental results with the simulation trend having the bimolecular rate constants

102

5

varying as n'°'° . This implies an almost inverse square root dependence of the kinetic

constant on the viscosity.

V.4. Conclusions

The viscosity of the solvent has a significant effect on the overall rate of the
reaction. As the viscosity of the solvent. increases, the overall rate of the
photosensitization reaction decreases, suggesting that these reactions are diffusion
controlled. The decrease in the rate of reaction is qualitatively explained using the
Smoluchowski-Stokes-Einstein equation and numerical solutions of the photophysical
equations. Based on the studies presented here, we would expect a profound effect on the
rate of initiation in a polymerizable mixture. As the reaction proceeds, the viscosity of
the mixture increases due to polymerization, and this increase in viscosity will reduce the
rate of initiation. It is therefore expected that the rate of initiation will become extremely
low after a certain degree of polymerization, and that any further reaction will be almost

exclusively from propagation of active centers.

Absorbance

Figure V. 1. Absorbance of the photoinitiator (UV9310C, GE Silicones) and the

103

 

 

 

------- Anthracene
3.5 ., - - - — uvsaioc
3e0 II ... “ O.
2.5 u : .“‘. ‘I
2_0 u ,\ ..o- .' x :0“ .0 '. . \1
I \ .. ‘0 .' v I I
1.5 u ' \"’—.:\ I. .l 2
l ,' \ - . .
1.0 1 h ' l \ “ . e I. l
j .' \ ‘. . ',
0.5 u I : \ ‘ x x ‘
_’ .. \“-‘.;...' e“
00 ~"vu- ----- ' :4 :-=..____-, ,1,
190 240 290 340 390

Wavelength (nm)

photosensitizer (anthracene) as a function of wavelength.

104

 

 

Figure V.2. Chemical Structure of the photoinitiator, bis (4-dodecylphenyl) iodonium
hexafluoroantimonate and the photosensitizer, anthracene.

105

Intensified Diode Array Spectrometer

 

 

 

 

 

Electronic
Shutter

M. .w......, o N. ..wme
Sample containing
0.505 mM Anthracene

  
 

0.00708 mM Photoinitiator

Photoexcitation by
363.8 nm Ar+ Laser

Figure V.3. Experimental setup for measuring the steady state fluorescence of
anthracene during the photosensitization of a diaryliodonium
hexafluoroantimonate salt.

106

220000 '
200000

1 80000

[IIITI

160000
140000
120000

1 00000

Fluorescence

 

80000

 

60000

 

 

40000

 

20000

 

 

 

20

 

. , . .
420 425 430 435 440
Wavelength, nm

Figure V.4. Typical fluorescence profile of anthracene during the photosensitization of
diaryliodonium salts by exciting at 363.8 nm.

107

 

 

 

 

 

0.0006- —80000
2
E
5 o'ooosd Fl ' l t425
o uorescence 5: na a nm
8 . , _ 9 -60000 I
9 Simulation Results ‘3
E 0.0004-1 2
a) j 8
g o
0 0.0003- -40000 ‘3”
0 o g
Q) ‘ o _
5 00002 L 9-
U - 1? g
(3 ~20000 (D
5 5'
: 0.0001 -
<
0.0000 . , e , . , . . . , . , . , o
0 2 4 6 8 10 12 14

time, sec

Figure V.5. Comparison of computer simulation and experimental photosensitizer
fluorescence intensity for reaction of anthracene and diaryliodonium
hexafluoroantimonate.

108

1000

  
 
  

35
0 Reaction Rate Constant
win—- Viscosity ' 30

r

l
N
01

 

 

 

:n
(D
n)
9.
a.
100- 3
e s "202
e 8
a) i o
O
a ”52.
">- 101 ' 9’
= -10_3
' E
-5 3
e
" \
1 (D
0 20 40 60 80 100

Percent Glycerol

Figure V6. The effect of glycerol content on solution viscosity and reaction rate
constant.

109

 

[‘A]

 

   

senphos

 

[3A]

kabs

kfluor + knonfluor kphos + knonphos

Absorption

Fluorescence

 

 

Phosphorescence

 

 

[0A] v '
Ground State Anthracene

Figure V.7. Electronic energy level diagram for anthracene fluorescence and
phosphorescence.

110

 

 

 

WM ' W v v v v
. 0
0.0004 -
C ..
8 —-—Ground State Anthracene
g 0.0003 - -~-—--—- Singlet State Anthracene
C .
8 , . Triplet State Anthracene
§ 00002 _’ ----v~— Active Centers from Singlet
a —-+---Active Centers from Triplet
T3
2 0.0001 -.
'VVV' V V V V - '._....'_._..
0.0000 eeWe-e-e-e-e-e-‘e-e- -
0 5 10 15 20

time, sec

Figure V.8. Simulation results for the concentration profiles for various chemical species
in the photosensitization reaction. The initial anthracene concentration was
0.505 mM and the initiator concentration was 6.08 mM.

 

lll

 

05 . . e. ...... - . . H- ............................
N ‘0... +—+—+—+—+—+~s-+—s.+-+-+_+_ _
2 \\\ + +—+—+_e—+-+—+_H_+_ e—-
E. 0 4 - \ Dijkskw 56.25 cP
C . V» “o“
.9 1 \ \‘V Mk
E \ ‘VVW N“.-
E 03 d \\ \“k'w New.
\ w
§ \\ ‘w 11.25 0P
8 0.2 ‘ kkk WW‘www‘v
Q) ] kkk “*N‘k
: k k 5.625 cP
“g 0.1 - 2.25 cP ka,‘ &
h M
5 Hukk
& 0.0 1 0.9 0P
1 r T 1 l
0 5 10 15 20

Time, seconds

Figure V.9. Simulated decay profiles of anthracene during photosensitization at different
viscosities.

112

100-

 

 

 

 

(D

h

o

E

E 10_ I

E" ' Simulation
*3 f - Experiment
c 1‘

o

0

..‘1.’

«1

02 0.1-

C

.9

‘6

8 001

0'5 ' 0.0 0.5 1.0 1.5

Reciprocal Viscosity, 1/cP

Figure V. 10. Comparison of experimental data with numerical solutions of the
photophysical equations using SSE modified kinetic constants.

113

100

#4 l llll]

   

l

10

l llllll

 

Results from Simulation
I Experimental Data

 

Apparent Reaction Rate Constant

 

0.1 f I I T I 1' I V 1
0D 05 10 15 20

Reciprocal Viscosity, 1/cP

Figure V.11. Comparison between experimental data on reaction rate coefficient and
simulation with adjusted constants.

10.

11.

12.

13.

14.

15.

l6.

17.

114

V.5. List of References

J .V. Crivello, and R. Narayan, Chem. Mater., 4, 692 (1992).

SP. Pappas, UV Curing, Sci. and Tech, Technology Marketing Corp., Norwalk,
CT, 2, (1985).

J.V. Crivello, Adv. Polym. Sci., 62, 1 (1984).
A. Reiser, Photoreactive Polymers, Wiley, New York, (1989).

SP. Lapin, in Radiation Curing of Polymeric Materials, C.E. Hoyle and J.F.
Kinstle, Eds., ACS Symp Ser., 417, ACS, Wash DC, 363 (1989).

SP. Pappas, Prog. Org. Coatings, 13, 35 (1985).

J.V. Crivello and J.H.W. Lam, in Epoxy Resin Chemistry, R.S. Bauer, Ed., ACS
Symposium Series, Vol. 114, ACS, Wash DC, 1 (1979).

J .V. Crivello, Organic Coatings, Science and Technology, G.D. Parfitt and A.V.
Patsis, Eds., Dekker, New York, 5, 35 (1983).

W.R. Watt, in Epoxy Resin Chemistry, R.S. Bauer, ed., ACS Symposium Series,
114, ACS, Wash DC, 17 (1979).

F. Lohse and H. Zweifel, Adv. Polym. Sci., 78, 61 (1986).

J.V. Crivello and DA. Conlon, J. Polym. Sci. Part A: Polym. Chem, 21, 1785
(1983).

J .V. Crivello and J.L. Lee, J. Polym. Sci. Part A: Polym. Chem, 27, 3951 (1989).

P. Sundell, S. Jonsson, A. Hult, J. Polym. Sci., Part A: Polym. Chem, 29, 1525
(1991).

G. Manivannan, J. P. Fouassier, J. Polym. Sci., Part A: Polym. Chem, 29, 1113
(1991).

AB. Scranton, L.S. Coons, B. Rangarajan, US. Patent Application Serial No.
08/467,729, Filed June 5, 1995.

S. K. Moorjani, B. Rangarajan, and A. B. Scranton, “Characterization of the Axial
Light Intensity Gradient in Three Dimensional Polymerizations”, in preparation
for J. Polym. Sci.

E. W. Nelson, T. P. Carter, A. B. Scranton, Macromolecules, 27, 1013 (1994).

18.

19.

20.

21.

22.

23.

115

S. K. Moorjani, B. Rangarajan, and A. B. Scranton, “Characterization of the Axial
Light Intensity Gradient in Three Dimensional Polymerizations”, in preparation
for J. Polym. Sci.

C.H. Bamford, C.F.H. Tipper, R.G. Compton, eds., Comprehensive Chemical
Kinetics, Vol. 25, Elsevier, New York, NY, 1985.

E. W. Nelson, T. P. Carter, A. B. Scranton, J. Poly. Sci., A: Poly. Chem, 33, 247
(1995)

G. J. Kavamos, in Photoinduced Electron Transfer, M. J. S. Dewar, J. D. Dunitz,
K. Hafner, S. Ito, J. M. Lehn, K. Niedenzu, K. N. Raymond, C. W. Rees, and F.
Vogtle, Eds., Topics in Current Chemistry, 156, Springer, New York, NY, 27
(1990).

N. J. Turro, Molecular Photochemistry, W. A. Benjamin, New York, NY, 86
(1965).

R. L. Borrelli and C. S. Coleman, Differential Equations - A Modeling Approach,
Prentice-Hall, Englewood Cliffs, NJ, 1987.

CHAPTER VII.

CONCLUSIONS AND RECOMMENDATIONS

The work reported in this thesis has contributed to the understanding of the
kinetics and mechanisms of cationic photopolymerizations of epoxides and vinyl ethers.
It will lead to a better understanding of the photosensitization mechanism and hopefully
to an increased use of renewable agricultural resources like natural oils in the production
of polymers. The fundamental experimental and theoretical investigations presented in
this thesis complement the more macroscopic studies currently available in the literature.
Furthermore, novel techniques for cure monitoring such as differential scanning
photocalorimetry and steady state fluorescence monitoring have been highlighted and
their utility for studying reactions of widely varying kinetic characteristics has been
demonstrated. In this chapter, some general conclusions and suggestions for future work

will be presented.

VII.1. Summary of Results

This research has provided considerable insight into the mechanism and kinetics
of cationic ring—opening photopolymerization of epoxidized natural oils. Several distinct
aspects of the photopolymerizations were investigated. A series of isothermal differential

scanning photocalorimetric studies were performed at various light intensities and

116

ll7

temperatures to characterize the polymerization reaction kinetics. A series of studies
using steady state fluorescence monitoring were conducted to gain better insight into the
diffusion controlled behavior of the photosensitization reaction between anthracene and
diaryliodonium salts. The photophysical equations describing the photosensitization
reaction were solved in conjunction with relationships provided by a theory for rapid,
diffusion-limited, chemical reactions and the simulation results was compared with

experimental results.

VII.1.I. Differential Scanning Photocalorimetry Studies

Differential Scanning Photocalorimetry (DSP) was employed to investigate a
series of photoinitiated cationic polymerizations at different temperatures using
epoxidized soybean oil as the monomer. The apparent propagation rate was found to
increase sharply in the beginning, but after reaching a peak, it decreased gradually. Also,
the overall conversion obtained was found to be higher for the polymerizations which
proceeded at lower temperatures, although the initial rate of polymerization was higher
for reactions at increased temperatures. We have tried to explain both of these
phenomena using arguments pointing to diffusion effects in a rapidly polymerizing
system.

During the polymerization, the large hexafluoroantimonate counterion of the
initiator experiences a rapid decrease in mobility as the viscosity increases. Meanwhile,
the active carbocation retains its diffusional mobility by reacting with neighboring
oxirane rings, thus leading to separation of the two species. This increases the reactivity

of the cationic active center. The final decrease in kp is due to trapping of the active

118

centers in the rapidly formed crosslinked polymer network. Therefore, a limiting

conversion is observed despite the presence of “live” cationic species.

V1.1 .2. Steady State Fluorescence Studies

The photosensitization mechanism was investigated using in-situ, steady state
fluorescence spectroscopy. The direct dependence of the anthracene fluorescence
intensity at 425 nm on the concentration of anthracene in the sample was utilized to
monitor the reaction kinetics. The experimental fluorescence profiles show that the
anthracene concentration follows an exponential decay, suggesting kinetics similar to a
first order reaction. This can be attributed to the fact that the concentration of
diaryliodonium hexafluoroantimonate initiator in the sample was an order of magnitude
higher than that of anthracene. As the viscosity of the sample was increased, the rate of
photosensitization appeared to decrease, suggesting a diffusion limited reaction
mechanism. Photophysical equations were formulated for the different steps in the
reaction scheme and a well known theory for diffusion controlled reactions was employed
to model the behavior. Using an adjustable parameter, an empirical relationship was
arrived at between the viscosity of the reaction system and the photosensitization kinetic

constant.

ll9
V1.2. Recommendations for Future Work

Cationically polymerizable systems have received limited attention in the past as
viable techniques to produce solvent-free coatings, films and composites. The research
described in the previous chapters has proven the usefulness and applicability of this
technique to various applications. Also, the experimental techniques described in this
thesis can be employed to study other systems with alternate initiators, photosensitizers,
and monomers. Since we have already proved that polymers with excellent physical
properties can be produced from renewable agricultural resources, further research in this
area is warranted to further enhance the curing properties as well as the final mechanical
and thermal properties of these polymers.

As described in detail below, future research conducted in the following areas will
aid in our understanding of cationic photopolymerizations:

i) Investigation of alternate photosensitizers such as thioxanthones.

ii) Fluorescence and phosphorescence lifetime experiments on anthracene at different
viscosities.

iii) Epoxidation and subsequent photopolymerization of other natural oils.

iv) Hydrolysis of epoxidized oils or epoxidation of hydrolyzed oils before
polymerization.

v) Mechanical and thermal characterization of polymers obtained from
photopolymerizations of epoxidized oils.

vi) Accelerated weathering tests on polymers to test feasibility for different

applications.

120

vii) Succeptibility to photooxidation of polymers produced by UV light.

V1.2.]. Alternate Non toxic Photosensitizers

Conventionally, the photosensitizers used in conjunction with diaryliodonium
salts are anthracene, perylene and pyrene. Benzophenone has also been employed as a
photosensitizer. Although the conjugated nature provided by the fused aryl rings in these
compounds makes them very photoreactive and suitable for use as free-radical initiators
and cationic photosensitizers, these compounds are classified as toxic. The
concentrations of these sensitizers in coating formulations is on the order of hundredths
of a percent, but they may still be restricted from use in consumer products.

Thioxanthones are non-toxic radical photoinitiators being increasingly used in the
polymer coating industry for pigmented systems (See Figure 6-1 for chemical structures).
Isopropyl thioxanthone (ITX) has been used as a free radical initiator for
photopolymerizations in clear as well as pigmented coatings. The ability of the
thioxanthone structure to generate radical moieties as a result of photochemical reactions,
can be exploited to photosensitize diaryliodonium salts for conducting cationic
polymerizations. These compounds exhibit maximum absorbance in the range from 380
to 420 nm ( e=104 litres/mol cm) and give reduced yellowing in the polymer product
compared to other initiators. These initiators have thus ingratiated themselves with
manufacturers of printing inks and white pigmented lacquers.

Thioxanthones however have significant disadvantages. Since an amine co-
initiator is needed, the thioxanthones can only be used in acrylate based systems, and not

in systems containing unsaturated polyesters and styrene.'2 Further, the amines cause

121

yellowing of the coating which increases with exposure to light. Flash photolysis of a

wide variety of differently substituted thioxanthone derivatives revealed that the OL-

aminoalkyl radicals act as the initiating species for these systems.3 Since the parent
thioxanthone is poorly soluble in most formulations, substituted derivatives are used.
Alkyl- and chloro-substituted derivatives are the most prominent and commercially
available. Isopropylthioxanthone (ITX) has an absorbance maxima at 383 nm and is
widely used in pigmented systems. The initiating performance of thioxanthone
derivatives has been continually improved. 1-Chloro-4-propoxythioxanthone (CPT X)
which has been recently introduced in the commercial market has been shown to be twice
as reactive for certain printing ink applications.4 The nature and position of substituents
on the thioxanthone nucleus is closely related to the absorbance characteristics of these
compounds.5 Electron donating substituents in the 2 or 7 position as well as electron
withdrawing substituents in the 3 or 6 position cause shifts in absorbance bands toward
longer wavelengths. A methyl group in the 1 position deactivates the molecule probably
via intramolecular hydrogen abstraction while a chlorine group in the same position leads
to a pronounced activating effect. Methyl groups in the 3 and 4 position also activate the
molecule. Ether groups in the 2 and even more in the 4 position result in more reactive
photoinitiators than those with alkyl groups at these positions. CPTX is the most
promising derivative so far, because of its high reactivity as well as its ease of synthetic
preparation. It has a strong absorbance peak at 387 nm besides the short wavelength
maximum at 254 nm. An additional peak is observed at 314 nm because of the 4-position

substitution. This absorbance peak is ideally placed to accept photons from the 313 nm

122

radiation band from medium pressure mercury lamps and probably is a significant
contributor to the high photo-reactivity exhibited by this molecule.

Upon illumination with ultraviolet light, all thioxanthones exhibit fluorescence in
the 450-500 nm region and this property can be exploited to study polymerizations caused

by these compounds.

VI.2.2. Fluorescence and Phosphorescence lifetime experiments.

To further enhance our understanding of the effect of viscosity on the
photosensitizaton reaction mechanism and kinetics, we need to conduct in situ, time
resolved fluorescence and phosphorescence lifetime studies on anthracene at different
vsicosities. This will provide us with information on the effect of viscosity on the
lifetimes of the singlet and triplet excited state anthracene, respectively. The relative
contribution of these two species toward photosensitization is a function of their
lifetimes. Hence, it is quite possible that the change in viscosity of the formulation may

affect the sensitization kinetics.

VI.2.3. Epoxidation and subsequent photopolymerization of other natural oils.

Other natural triglyceride oils have chemical structures very similar to that of
soybean oil. For example, the composition and different properties of corn oil are
compared with soybean oil in Table 6-1. The iodine value of an oil is an accurate
measurement of reactive unsaturation present in the oils. This unsaturation can be
exploited to introduce epoxide moieties and study polymerization. The similar iodine

values of soybean and corn oil and their similar chemical composition suggest that com

123

oil could also be epoxidized and subsequently photopolymerized like soybean oil. Corn

oil is also available at a price close to that of soybean oil ($0.26/lb).

VI.2.4. Hydrolysis of epoxidized oils or epoxidation of hydrolyzed oils before
polymerization.

The average representation of a soybean oil molecule is a triglyceride ester with
three branches of 16-18 carbons each. If the triglyceride molecule is hydrolyzed, we
obtain the consituent saturated and unsaturated acids as shown in Table 6-1. The mixture
of acids could be epoxidized using hydrogen peroxide and acetic acid to provide a
polymerizable mixture. Photopolymerization of this mixture of oils has definite
advantages over photopolymerization of epoxidized soybean oil. First, the effective size
of the polymerizable monomer is reduced while maintaining the multifunctional nature of
the molecule. The polymers thus obtained will definitely have better physical properties.
To make the process even more efficient, the unsaturated components of the oil mixture
can be separated from the saturated components using standard techniques described in
the literature. An alternate route is to epoxidize the natural oils first before hydrolyzing

them.

VI.2.5. Mechanical and thermal characterization of polymers obtained from
photopolymerizations of epoxidized oils.

The thermal and mechanical properties of the crosslinked polymers formed by
cationic polymerizations of epoxidized oils can be characterized using standard techniques
such as differential scanning calorimetry (DSC), thermal gravimetric analysis (TGA), and

thermal mechanical analysis (TMA). These experiments will allow the glass transition

I24

temperature (if any), the degradation temperature, and the expansion coefficient to be
established. The flow properties of the polymers can be investigated using a Dynamic
Mechanical Analyzer. Using standard procedures, this technique can be used to measure
the loss and storage moduli of the resulting polymers for a spectrum of frequencies, using
standard time-temperature superposition principles. Similar measurements can be
performed on systems which contain short chain epoxide moieties, and the system will be
adjusted to obtain superior mechanical properties.

Preliminary studies have shown that the polymers formed from epoxidized

soybean oil had a coefficient of thermal expansion of about 210 urn/m °C. These

polymers were quite stable at elevated temperatures and showed negligible weight loss at
temperatures below 200°C. In fact, at 300°C, the weight loss (as measured by TGA) was
less than 2%. Above 350°C the ESO polymers rapidly degrade and lose most of their
weight. Differential scanning calorimetry of a cured ESO polymer did not show a definite
glass transition. These polymers appeared to be rubbery but, upon stressing, failed in
brittle mode.

As expected, preliminary studies have shown that the addition of an epoxy resin
as a comonomer resulted in an increased modulus and a considerably tougher material.
Commercial resins such as DER 331 (Dow Chemicals) can be employed to obtain
enhanced mechanical and thermal properties.

Our preliminary swelling studies indicate that ESO polymers swell only very

slightly in water, swell to about 20% in acetone and partially degrade in benzene. Again,
the solvent resistance may be enhanced by copolymerization with the aforementioned

epoxy resins. Extensive swelling studies could be performed using other solvents. Any

125

change in the properties of the polymer upon exposure to the solvents should be

thoroughly characterized.

VI.2.6. Accelerated weathering tests on polymers

The weatherability and chemical resistance of the polymers also need to be
characterized. ASTM weatherability studies can be performed using Weatherometer Q
Panel Testers. This state-of-the-art instrument allows samples to be maintained at
controlled environmental conditions (temperature, UV exposure, humidity, etc.). These
accelerated aging tests can be performed to identify superior formulations for potential

use in automotive and exterior coating applications.

V1.2. 7. Photooxidation of polymers produced by UV light.

The photoaging of polymers is mainly caused by solar radiation reaching the earth
surface (mostly with wavelengths greater than 300 nm) via complex oxidative processes.6
This photo-oxidative process is often accompanied by thermal oxidation depending on
the environmental and weather conditions. Photo- and photo-thermal oxidative
degradation cause chemical changes in the polymer structure with the build—up moieties
such as carbonyls, acids, esters and alcohols, apart from causing chain scissions and
crosslinking. In addition to severe drops in the mechanical and thermal properties of the
polymers, the polar functional groups generated also lead to decrease in performance by

causing a drop in water repellence, transparency, and dielectric properties.

126

Photo-oxidation is a radical chain process which can be initiated in various ways.
First, the photoactive functional groups present in the polymer structure, the residual
initiator, or the impurities may absorb light and give rise to free radical active centers.
The propagation of the oxidation commonly takes place by hydroperoxidative cycles
involving peroxiradicals as intermediates.6 Since our polymerizations are conducted with
carbocationic active centers and not free radicals, the photodegradation should not be a
very critical issue. Nevertheless, the investigation of this could lead us to design better

formulations for various applications.

127

Table V1.1. Comparison of Soybean and Corn Oil

 

 

 

Acids Chemical Formula % in Soybean Oil % in Corn Oil
Saturated Acids
Palmitic C15H3ICOOH 6.5 6.0
Ar achidic C 19H39COOH 0.7 1 .0
Stearic C17H35COOH 4.5 2.0
Unsaturated Acids
Oleic C17H33COOH 33.5 37.0
Linoleic C .7H31COOH 52.5 54.0
Linolenic CnHngOOH 2.3 0.0
Relative Unsaturation 145.4 145.0
Iodine value 122 - 134 111 - 128

 

 

 

 

 

128

O—CHz-CHz—CH3

l-Chloro-4-propoxythioxanthone (CPTX)

Figure V1.1. Chemical structures of thioxanthones employable as photosensitizers in
conjunction with diaryliodonium initiators.

129

V1.3. List of References

L. Gothlich, R. Osterloh, and M. Jacobi, in Conf. Proc. FATIPEC XIII, 244,
(1976).

P. Hauser, R. Osterloh, and M. Jacobi, in Conf. Proc. FATIPEC XIV, 241,
(1978).

J .-P Fouassier, and D.-J.J. Lougnot, J. Appl. Poly. Sci., 34, 477, (1987).

A.W. Green, A.W. Timms, and RN. Green, in Proc. Conf. Radtech Europe,
Edinburgh, 636, (1991).

K. Meier, and H. Zweifel, J. Photochem, 35, 353, (1986)

A. Faucitano, A. Buttafava, G. Camino, and L. Greci, Trends in Polymer Science,
4(3), 92. (1996).