«as:

15...:
a...¢.l...Pr3.
has“: a...

It}; .
3......

..

. . 2..

i... H»

3...»:

.5..." H

v...
. . . ;
munflmn.

.rdxifix

f
I?

V 1.3.3le
.91 3.15.31).
1 Ii: :33835.)
.98. lqvia 1.52:,
:- '1‘

23 .6

Roman“. i5..n:..m.
dc . . “cu...

.v

v Ru!
lat—01:10:52..“z. A
Fl. .- z :A.

 

3p! xv. t.
v\\t->1
\

14...; :1.
Runawh. .
O Q. :-

Macaw».

q :3!
K

.. r:

2:: .1.
7
~ . ‘

A.
.. v; ...t\.;
V .35‘

 

THESIS

(14%)

SITYLI

Milli:lililiiiitllliilil

3 1293 01421 0680

This is to certify that the

dissertation entitled

Kinetics and Mechanisms for Cationic

Photopolymerizations of Divinyl Ethers

presented by

Eric W. Nelson

has been accepted towards fulfillment
of the requirements for

Ph - D degree in CM:

W

Major professor

 

Date 5/25/95

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

LIBRARY
Michigan State
University

 

 

 

 

PLACE I! RETURN BOX to remove We checkout from your record.
To AVOID FINES return on or More dete due.

DATE DUE DATE DUE DATE DUE

 

T

5‘1

UN ‘0 1399

I
l
‘ i
I
I.
);.
"l i
I
J.
|
I
I

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

l

 

MSU le An Afflrmetive ActionlEqml Opportunity Inetnuion
Wanna-m

KINETICS AND MECHANISMS FOR CATIONIC
PHOTOPOLYMERIZATIONS OF DIVINYL ETHERS

By

Eric W. Nelson

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Chemical Engineering

1995

 

ABSTRACT

KINETICS AND MECHANISMS OF CATIONIC
PHOTOPOLYMERIZATIONS OF DIVINY L ETHERS

By

Eric W. Nelson

This research provided detailed pictures of the kinetics and mechanisms of a
model cationic photopolymerization system, and resulted in the development of a host of
analytical techniques which may now be used to characterize a broader set of cationic
photopolymerizations.

In situ, time resolved fluorescence spectroscopy provided a means to characterize
cationic photopolymerizations of vinyl ethers by monitoring the photosensitize intensity
as a function of time. An observed reduction in fluorescence intensity was attributed to
consumption of the photosensitizer, providing a means to monitor the production of
active cationic centers.

The photosensitization mechanism for cationic polymerizations initiated by
diaryliodonium salts photosensitized by anthracene was investigated using fluorescence
and phosphorescence spectroscopy. Transient phosphorescence studies demonstrated that
electron transfer occurred from the triplet state of anthracene to the initiator. The
numerical solution of the photophysical kinetic equations indicated that ~95 % of the

active cationic centers are produced by reaction of initiator with the triplet state.

In situ Raman spectroscopy was used to obtain complete conversion profiles as a
function of time. The apparent kinetic constant for propagation, which comprises all
types of propagating centers was determined. The Raman studies provided a direct
measure of limiting conversion which is less than unity in these highly crosslinked
polymerizations due to active center trapping.

Photo-differential scanning calorimetry (PDSC) experiments were used to
determine effective kinetic constants for propagation and termination for unsteady-state
divinyl ether polymerizations. The profile for the propagation rate constant exhibited a
dramatic increase at the start of the reaction, plateaued, and then decreased as the reaction
reached limiting conversion. The effective termination rate constant was found to be very
low, with active center lifetimes approaching twenty minutes. Excellent agreement for
the values of the kinetic rates constants was found with the Raman and PDSC

experiments.

The temperature-dependent luminescence of tris(B-diketone) chelates of europium

was used for in situ temperature measurements during cationic photopolymerizations of
vinyl ethers. These molecular-level luminescent probes provided a real-time, noninvasive
method for monitoring temperature during these high-speed polymerizations. In situ
temperature profiles obtained using this novel technique verified the thermal effects

during these highly exothermic photopolymerizations.

ACKNOWLEDGMENTS

I would like to express my appreciation to the many people that made
contributions to this thesis.

First, I would like to thank my advisor, Professor Alec Scranton, for all his
inspiration, encouragement and friendship during my stay at Michigan State University.
Secondly, I am grateful to Jeffry Jacobs for all his assistance and the motivation he
provided for completion of this work. Many of his contributions were the result of
numerous discussions during the last three years.

The fluorescence and phosphorescence work in this thesis was performed in the
LASER Laboratory, MSU Department of Chemistry. I would like to give special thanks
Dr. Thomas Carter for all his assistance in experimental set-up and interpretation.

The Photo-Differential Scanning Calorimetry experiments were performed at the
Department of Chemical Engineering, University of Colorado, Boulder in Professor Chris
Bowman’s laboratory. I am grateful for the assistance and encouragement of Dr.
Bowman and Dr. Kristi Anseth with the PDSC work.

I thank professors Kris Berglund, Gary Blanchard, and Dennis Miller for serving
on my thesis committee, and for making useful suggestions.

I would like to thank Arvind, Bharath, Julie, Kevin, Glen, Roger, Scott and Shail
for making the laboratory and office both enjoyable and interesting. Special thanks go to
Czarena Crofcheck and Brenda Becker for their work on the luminescence temperature

probes and thermocouple studies.

iv

 

I am grateful to the National Science Foundation (Grant No. CT S 9216939) for
providing support of this research. I would also like to thank the Upjohn company for the
fellowship which supported me during the last year of the project.

Finally, I would like to thank my family, friends and especially Beth Curtiss for

all her encouragement, support and companionship during the completion of this work.

TABLE OF CONTENTS

LIST OF FIGURES .................................................................................................... xii
CHAPTER I. INTRODUCTION AND MOTIVATION ............................................ l
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. Need for Research on Cationic Photopolymerizations ............................ 7

1.5. List of References .................................................................................... 9
CHAPTER 11. BACKGROUND ................................................................................ 13
11.1. Cationic Polymerizations ....................................................................... 13

11.1.1. Properties of Cationic Polymerizations ........................................ 13

11.1.2. Cationic Photopolymerizations of Epoxides and Vinyl Ethers.... 15

11.1.3. Photosensitization of Cationic Polymerizations ........................... 17
11.2. Fluorescence Cure Monitoring .............................................................. 18
11.3. Photosensitization Mechanism .............................................................. 23
11.4. Infrared and Raman Cure Monitoring ................................................... 26

11.4.1. Introduction to Infrared and Raman Spectroscopy ....................... 26

11.4.2. Physical Basis and Background for IR Spectroscopy ................... 30

vi

11.4.3. Instrumentation for Infrared Spectroscopy ................................... 33

11.4.4. Sampling Techniques for Infrared Spectroscopy .......................... 39
11.4.5. 1R Application to Polymers .......................................................... 43
11.4.6. Physical Basis and Background for Raman Spectroscopy ........... 47
11.4.7. Hardware Requirements for Raman Spectroscopy ....................... 51
11.4.8. Sampling Techniques for Raman Spectroscopy ........................... 53
11.4.9. Raman Application to Polymers ................................................... 55
11.4.10 Raman Cure Monitoring ............................................................. 60
11.5. Cure Monitoring Using Photo-Differential Scanning Calorimetry ....... 61
H6. Luminescence Temperature Monitoring ............................................... 64
11.6.1. Need for Temperature Monitoring ............................................... 64
11.6.2. Probe Requirements ..................................................................... 68
11.7. Modeling of the Formation of Highly Crosslinked Polymers ............... 69
11.8. List of References .................................................................................. 73
CHAPTER 111. OBJECTIVES .................................................................................. 83
CHAPTER 1V. IN SIT U FLUORESCENCE SPECTROSCOPY:

CURE MONITORING TECHNIQUE .......................................................... 85
IV. 1. Introduction .......................................................................................... 85
IV.2. Experimental ........................................................................................ 87
IV.2.]. Materials ...................................................................................... 87
IV.2.2. Absorption and Fluorescence Measurements .............................. 87
IV.3. Results and Discussion ......................................................................... 9O
IV.3. 1. Absorption Experiments .............................................................. 9O

vii

IV .3.2. Fluorescence Experiments ........................................................... 90

IV.3.3. Initiation and Polymerization Kinetics ........................................ 96
IV.3.4. Effect of Initiator Concentration .................................................. 99
IV.3.5. Effect of Photosensitizer Concentration ...................................... 99
IV.3.6. Effect of Anthracene Functionalization .................................... 102
IV.3.7. Fluorescence Experiments Without Monomer ......................... 106
IV.3.8. Initial Increase in Fluorescence Intensity .................................. 108
NA. Conclusrons ..................... 113
IV .5 . List of References ............................................................................... 1 16

CHAPTER V. PHOTOSENSITIZATION MECHANISM: EXPERIMENTS

AND MODELS ............................................................................................ 1 17
V.1. Introduction ......................................................................................... 117
V2. Experimental ........................................................................................ 118
V2]. Materials ..................................................................................... 118
V.2.2. Laser Fluorescence Measurements ............................................. 119
V.2.3. Phosphorescence Lifetime Measurements ................................. 119
V.2.4. Fluorescence Lifetime Measurements ........................................ 120
V5 .2. Absorption and Fluorescence Measurements ............................. 121
V3. Results and Discussion ........................................................................ 122
V.3.1. Photosensitization Mechanism .................................................... 122
V.3.2. Fluorescence and Phosphorescence Lifetime Studies ................ 129
V.3.3. Triplet Quencher Studies ............................................................ 139
V.3.4. Competing Reactions .................................................................. 140

viii

V.3.5. Fluorescence of the Products from Photosensitizing

Diaryliodonium Salts with Anthracene ...................................... 143
V4 Conclusions ......................................................................................... 151
V5. List of References ............................................................................... 151

CHAPTER VI. IN SIT U RAMAN SPECTROSCOPY: CURE MONITORING

AND KINETICS .......................................................................................... 154
V1.1. Introduction ........................................................................................ 154
V1.2. Experimental ...................................................................................... 155
V1.2.]. Materials .................................................................................... 155
V1.2.2. Raman Experiments .................................................................. 156
V1.3. Results and Discussion ....................................................................... 158
V1.3.l. Raman Cure Monitoring ............................................................ 158
V1.3.2. Effect of Light Intensity ............................................................ 167
V1.3.3. Effect of Temperature ................................................................ 168
V1.3.4. Initiation Rate Constant ............................................................. 170
V1.3.5. Propagation Rate Constant ........................................................ 172
V1.4. Conclusions ........................................................................................ 180
V1.5. List of References ............................................................................... 182

CHAPTER VII. IN SIT U PHOTO-DIFFERENTIAL SCANNING

CALORIMETRY: CURE MONITORING AND KINETICS .................... 184
VII.1. Introduction ...................................................................................... 184
VII.2. Experimental ..................................................................................... 186

V11.2.1. Materials .................................................................................. 186

ix

VII.2.2. PDSC Experiments .................................................................. 186

 

VII.3. Results and Discussion ..................................................................... 187
V11.3.1. Effect of Light Intensity ........................................................... 187
V11.3.2. Effect of Temperature .............................................................. 189
V11.3.3. Termination Rate Constant ...................................................... 191
V11.3.4. Initiation Rate Constants .......................................................... 192 r.
V11.3.5. Propagation Rate Constant ....................................................... 194 L
V11.4. Conclusions ....................................................................................... 201 ._
VII.5. List of References ............................................................................. 203
CHAPTER VIII. IN SIT U LUMINESCENCE TEMPERATURE PROFILES ..... 205
VIII.1. Introduction ..................................................................................... 205
V1112. Probe Requirements ......................................................................... 207
V1113. Experimental .................................................................................... 208
V111.3.1. Materials ................................................................................ 208
VIII.3.2. Synthesis of Temperature Probes .......................................... 209
VIII.3.3. Absorption and Luminescence Measurements ...................... 209
VIII.4. Results and Discussion .................................................................... 211
VIII.4.1 Probe Selection ........................................................................ 211
VIII.4.2. Temperature Studies ............................................................... 214
VIII.4.3. Viscosity Independence .......................................................... 221
VIII.4.4. Inherent System Limitations ................................................... 225

VIII.4.5. Temperature Profiles for the Cationic Photopolymerizations. 225

VIII.4.6. Further Applications ............................................................... 23 l

VIII.5. Conclusions ..................................................................................... 233
V1116. List of References ........................................................................... 235
CHAPTER 1X. CONCLUSIONS AND RECOMMENDATIONS ........................ 237
IX.1. Summary of Results ........................................................................... 237
IX.1.1. Fluorescence Cure Monitoring .................................................. 238
D(.l.2. Characterization of the Photosensitization Mechanism ............ 239
Ix.1.3. In Situ Raman Studies ........................ 240
D1. 1 .5. Photo—Differential Scanning Calorimetry Studies ..................... 240
IX.1.4. Temperature-Sensitive Luminescence Studies .......................... 241
IX.2. Suggestion for Future Work ............................................................... 242
IX.2.1. Characterization of the Photosensitization Mechanism ............ 243
IX.2.2. Characterization of the Photopolymerization Reactions ........... 245
IX.2.3. Relationships between the Monomer Structure and the
Reaction Kinetics ...................................................................... 246
IX.2.4. Effect of Multifunctional Monomer Structure on Limiting
Conversions ............................................................................... 247
IX.2.5. Relationships between the Monomer Structure and Polymers
Properties .................................................................................. 249
IX.2.6. Investigations of the Propagation Reactions Based upon
Linear Polymerizations ............................................................. 250
IX.2.7. Effect of Light Source and Intensity ......................................... 251
IX.2.8. Swelling Studies ........................................................................ 251

xi

 

 

 

LIST OF FIGURES

 

Chapter 11.
Figure 11.1. Temperature profile for a thin film reaction of DVE-3 on a
glass substrate - ........................... 66
Chapter IV.
Experimental Setup for Fluorescence Cure Monitoring ................. 89

Figure IV.l.

Figure IV.2.

Figure IV .3.

Figure 1V.4.

Figure IV.5.

Figure IV.6.

Figure 1V.7.

Figure IV.8.

Absorption Spectrum of Monomer (DVE-3), Initiator
(UV9310C) and Photosensitizer (Anthracene) ........................ 91

Polymerization of DVE-3 monomer (0.5 wt % initiator) produced
by exciting anthracene (2.8 x10’2wt %), at 363.8 nm, and
monitored by measuring the intensity of the anthracene fluore-
scence at time intervals of 17 ms... . - - . - ................ 92

 

Photosensitization mechanism based on electron transfer from
triplet state anthracene to initiator .............................................. 94

Fluorescence intensity at 425 nm for polymerization of DVE-3
monomer with varying concentrations of initiator (UV9310C)
and 10'2 wt % of anthracene as the photosensitizer ............... 100

Fluorescence intensity for polymerization of DVE-3 monomer
with varying concentrations of anthracene photosensitizer
and 1 wt % initiator (UV9310C).... - - ........... 101

Fluorescence intensity for polymerization of DVE-3 monomer
with UV9310C and 10'3 wt % of 9-vinylanthracene and
anthracene as the photosensitizers . ........ 103

Polymerization of DVE-3 monomer with UV9310C and 10'2
wt % of 9,10-diphenylanthracene and 9,10-dimethyl-
anthracene as the photosensitizers, monitored by photo-
sensitizer fluorescence intensity -- ............ 105

xii

 

Figure IV .9. Fluorescence intensity at 425 nm for photosensitization
reaction of 2.2 x 10'2 wt % of anthracene as the photo-
sensitizer and 1.2 wt % of diaryliodonium hexafluoro-
antimonate (UV9310C) in methanol .......................................... 107

Figure IV. 10. Fluorescence monitoring of DVE-3 with 2.4 x 10'2 wt %
anthracene and 1 wt % initiator. The sample was moved during
the reaction to a new unreacted section ...................................... 109

Figure IV.1l. Initiation Started at time equal zero, but the fluorescence
emission was not allowed to reach the detector for the first
250 ms of the reaction ................................................................ 1 1 1

Figure 1V . 12. Different time delays of the initiation shutter and the detector
to determine the cause of the initial build-up ............................. 112

Chapter V.

Figure V. 1. Proposed photosensitization mechaniSm based on electron transfer
from triplet state anthracene to initiator ..................................... 123

Figure V.2. Hypothetical pathway and energetics of the photosensitization
reaction ....................................................................................... 126

Figure V.3. Electronic energy level diagram for anthracene fluorescence and
phosphorescence ......................................................................... l 27

Figure V.4. Transient phosphorescence profiles of 6.6 x 10'3 wt % anthracene
in methanol with and without 2.0 wt % initiator ........................ 130

Figure V.5. Transient fluorescence profiles for 6.6 x 10‘3 wt % anthracene in
methanol with and without 2 wt % initiator ................................ 133

Figure V.6. Simulation results for the concentration profiles of the reaction
species. The initial [A] was taken to be 0.5 mM with
[I] = 14.5 mM ............................................................................. 136

Figure V.7. Comparison of computer simulation and experimental photo-
sensitizer fluorescence intensity for reaction of anthracene
(0.5 mM - 1.0 x 10'2 wt %) and diaryliodonium hexafluoro-
antimonate (14.5 mM - 1 wt %) in methanol ............................. 138

xiii

Figure V.8. Polymerization of DVE-3 monomer with 2.1 wt % initiator and
10'2 wt % anthracene as the photosensitizer in the presence
of 0.21 wt % of the triplet quencher 9,10-dihydroanthracene.. 141

Figure V.9. Relative absorption of anthracene and its derivatives .................... 145
Figure V.10. Relative fluorescence intensity of anthracene and its derivatives. 146

Figure V.11. Fluorescence intensity of anthracene before and after
photosensitization of diaryliodonium salts ................................. 148

Figure V. 12. Fluorescence profile during polymerization of DVE-3 monomer
(1 wt % initiator) produced by exciting 9-phenylanthracene
(1.1 x 10'2 wt %), at 363.8 nm, and monitoring the

fluorescence intensity at time intervals of 70 ms ........................ 150
Chapter VI.
Figure V1.1. Experimental setup for Raman cure monitoring ........................... 157

Figure V1.2. Raman spectra of DVE-3 before and after cationic
photopolymerization ................................................................... 160

Figure V1.3. Cis and Trans isomers which result in the ~1630 cm'l doublet
in the monomer Raman spectrum ............................................... 161

Figure V1.4. Raman spectra during a cationic photopolymerization of
DVE-3 with 1 wt % initiator and 10'2 wt % anthracene .............. 163

Figure V1.5. Double bond conversion during a cationic photopolymerization
of DVE-3 calculated from the 1322 cm'1 Raman band ............... 164

Figure V1.6. Comparison of fluorescence and Raman monitoring techniques. 166

Figure V1.7. Raman conversion profiles for isothermal reactions ranging
from 20 to 50 °C .......................................................................... 169

Figure V1.8. Rate of reaction for several different temperatures versus time
for a cationic photopolymerization of DVE-3 with 1 wt %
initiator and 10'2 wt % anthracene ............................................... 174

Figure V1.9. Overall rate of propagation versus conversion for DVE-3
photosensitized with 10'2 wt % anthracene and
1 wt % initiator ............................................................................ 175

xiv

 

\

\\

\

 

Figure V1.10. kP[M'] versus conversion for the cationic photopolymerization

of DVE-3 at temperatures ranging from 20 to 50 °C ............... 177
Figure V1.11. 1tp versus conversion of the photopolymerization of DVE-3 at
various temperatures .................................................................. 178
Chapter VII.

Figure V111. PDSC exotherms for the cationic polymerization of DVE-3
at different light intensity values ................................................ 188

Figure V112. PDSC exotherms for the cationic polymerization of DVE—3
at temperatures ranging from 20 to 50 "C .................................. 190

Figure V113. Overall rate of propagation versus conversion for DVE-3
photosensitized with 1.0 x 10'2 wt % anthracene and 1 wt %
initiator ....................................................................................... 197

Figure V114. kP [1W] versus conversion for the cationic photopolymerization

of DVE-3 at temperatures ranging from 20 to 50 “C ................... 198
Figure V115 kp versus conversion for the photopolymerization of DVE-3
at various temperatures ............................................................... 200
Chapter VIII.

Figure V1111. The structure for the two lanthanide B-diketone chelates,
Eu(btfa)3 and Eu(hfa)3 ................................................................. 212

Figure V1112. Absorption spectra of Eu(btfa)3 (0.001 wt %), Eu(hfa)3
(0.01 wt %), and DVE-3 in methanol ......................................... 213

Figure V1113. Luminescence spectra of Eu(btfa)3 (0.001 wt %) and
Eu(hfa)3 (0.01 wt %) in DVE-3 and of DVE—3 ........................... 215

Figure V1114. Luminescence spectra at varying temperatures for
Eu(btfa)3 at 0.001 wt % in DVE-3, excited at 364 nm ................ 216

Figure V1115. The luminescence intensity of Eu(btfa)3 at an emission
of 617 nm plotted at different temperatures, resulting in
a linear calibration curve ............................................................. 217

Figure V1116. Luminescence spectra at varying temperatures for
Eu(hfa)3 at 0.01 wt % in DVE-3 excited at 340 nm .................... 219

XV

Figure V1117. The relative intensity, referring to the intensity of the
618 peak divided by the modified intensity of the 622 peak,
of Eu(hfa)3 in DVE-3 (0.01 wt %) plotted versus reciprocal
temperature ................................................................................. 220

Figure V1118. High temperature calibration curves for Eu(btfa)3 and Eu(hfa)3
excited at 364 nm and monitored at an emission of 620 nm ...... 222

Figure V1119. Temperature calibration curves for 0.01 wt % Eu(hfa)3
excited at 351 nm in liquid and solid systems ............................ 224

Figure V11110. The luminescence of Eu(hfa)3 (0.01 wt %) in methanol,
methylene chloride, and DVE-3 ................................................ 226

Figure V1111 1. Temperature calibration for 1.6 x 10'2 wt % Eu(hfa)3
in DVE-3 with 1.4 x 10'2 wt % anthracene excited with
15 mW at 351 nm using an argon ion laser ................................ 229

Figure V11112. Temperature profile of the cationic polymerization of
DVE-3 with 1.6 x 10'2 wt % anthracene and 1.0 wt % initiator
monitored with 1.0 x 10‘2 wt % Eu(hfa)3 .................................... 230

Figure V111. 13. The temperature profile generated from the luminescence
intensity of Eu(hfa)3 in a free radical polymerization of
I-IEA/DMF (40/60) with an initiator concentration of 0.2 wt %..232

xvi

q. .,

9““

 

'\

t.
‘1

 

 

 

'e
‘k '.
e ’o
b
‘.
‘0
c
._i

CHAPTER I. INTRODUCTION

1.1. Issues for the Development of UV Curable Polymers

The emission of volatile organic components (VOCs) from curing 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 ink and
coating formulations. Typically, an ink or coating formulation must be sufficiently fluid
to be easily applied (often onto a rapidly moving substrate through a series of transfer
rollers), but must cure rapidly to a hard film. Historically, these processing demands
have been met through the use of a rapidly evaporating organic solvent in the ink and
coating formulations. Numerous studies have demonstrated that when these volatile
organic solvents enter the atmosphere, they result in the formation of smog and air
pollution. There is an urgent need for the development of high performance inks and
coatings that will not emit VOCs to the atmosphere.

An improved, pollution-free formulation for inks or coatings must meet several
stringent requirements. The cured ink or coating should exhibit good abrasion resistance,
chemical resistance, flexibility, and excellent adhesion to the substrate. For applicability
in a continuous web process, the formulation should have a shelf life of at least six

mounts,l should be stable on the press, and cure very rapidly when initiated by the energy

.4.

 

 

a,

.
‘e.u '
i.

v
'\

“".

2

source. Finally, to satisfy safety and environmental concerns, the formulations should be
nontoxic, and should emit no volatile organic components. The required abrasion
resistance and chemical resistance can be readily provided by a highly crosslinked
polymer network. For this reason many traditional formulations consist of pigments and
volatile solvents dispersed in multifunctional oligomers that form highly crosslinked
networks upon curing. The improved formulations could be based upon highly reactive
monomers of low enough viscosity that no organic solvent is necessary. If the monomers
used in 100% reactive systems also have a low vapor pressure, they will emit no organic
vapors during cure.

The curing of an ink or coating formulation to form a highly crosslinked polymer
film could potentially be initiated by a variety of energy sources, including heat.
electromagnetic radiation, and electron beam irradiation. Unfortunately, heat-initiated
thermal systems typically require high temperatures to attain reasonable cure rates. The
elevated temperatures not only lead to high energy costs, but could also result in
significant distortions in substrate dimensions in many cases.' High-energy irradiation
such as ‘y-radiation and electron beam radiation can lead to rapid cure rates in a variety of
systems“ however, these techniques will most likely have very limited applicability
because they would result in degradation of the substrate. Light-induced photo-
polymerizations have many potential advantages for ink and coating formulations
including very high reaction rates at room temperature, low energy requirements, and
versatility since a wide variety of monomers may be initiated photochemically. With
proper choice of wavelength and light source, 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 the rapid, pollution-free curing of polymer films (for reviews, see
references 6-9). 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

6-12

absorption of UV light. UV-sensitive photoinitiators are currently available for free-

radical or cationic polymerizations. These photoinitiators are typically effective for a

“'2 This feature is useful for UV-curable inks and

variety of incident wavelengths.
coatings because the 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.

Free-radical photopolymerizations were first reported in the literature nearly fifty
years ago,'3 and are currently receiving considerable attention (for reviews see references
6-8,13). By far, 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 odor.6 Recently there has been growing concern over potential

14

health hazards associated with the acrylates?“ Several recent investigations have

4
demonstrated that free-radical polymerizations of multifunctional acrylates and meth—

acrylates exhibit unusual kinetic behavior. including immediate onset of autoacceleration.

6.19-24

the formation of heterogeneous polymers, and the attainment of a maximum

conversion significantly less than unity.°""28

Finally, the free radical photo-
polymerizations are inhibited by oxygen and must be carried out under an inert
atrnosphere,"‘7‘13 such as nitrogen.

UV-initiated cationic photopolymerizations differ greatly from free-radical
photopolymerization and exhibit several advantages when compared with the free-radical
photopolymerization discussed above. First of all, the cationic photopolymerizations are
not inhibited by oxygen.7"3""”‘0 This feature provides a significant practical advantage for
industrial processes since it is not necessary to blanket the system with nitrogen to
achieve rapid cure 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.”33 Finally.
cationic photopolymerization is a very versatile technique, and may be used to
polymerize important classes of monomers, including epoxides and vinyl ethers.”'w"
Although these classes of monomers cannot be cured by free-radical polymerizations,
they exhibit many desirable properties, including low volatility, good rheological
properties and negligible toxicity. Cationic photopolymerizations of divinyl ethers
proceed to completion in less than a minute when exposed to ultraviolet (UV) lightwm‘s

However, fully formulated mixtures containing monomer and initiator are completely

stable in the absence of UV light, eliminating the need for mixing immediately before

 

5
use. Furthermore, the cured polymer films associated with these monomers exhibit

excellent clarity, adhesion, abrasion resistance, and chemical resistance.”""""’

1.3. 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

photoinitiators 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. Early photoinitiators for cationic polymerizations,
first introduced in 1965, were aryldiazonium salts of complex metal halide anions."
Upon photolysis, these species produce an aryl halide, molecular nitrogen, and a strong

Lewis acid capable of initiating cationic polymerization."“‘"

While the aryldiazonium
salts were highly photoactive, they were also thermally unstable, which resulted in
unacceptably short shelf lives for fully formulated systems.

Crivello and Lam reported two classes of thermally stable photoinitiators for
cationic polymerizations: diaryliodonium and triarylsulfonium salts in 1976 and 1979.
respectively (for a review of the development of these initiators, see reference 10). Upon
photolysis, these compounds undergo irreversible photofragmentation in which the
carbon-iodine or carbon-sulfur bond is cleaved to produce an aryliodinium or an

arylsulfinium cation-radical capable of initiating cationic polymerization.” While the

diaryliodonium and triarylsulfonium salts actively initiate cationic polymerization in the

6
presence of UV light, they are remarkably latent in the absence of light. In fact. fully

formulated solutions of some salts in epoxide monomers have shelf lives of several
years.‘2

The wavelength at which the diaryliodonium and triarylsulfonium salts exhibit
their maximum UV absorbance depends somewhat upon substituents attached to the
ring, but typically falls between 225 and 275 nm“33 (some sulfonium salts absorb near 300
nm”). Diaryliodonium salts of hexafluoroantimonate produce the fastest rates of reaction
of the onium salts due to the size and acidity of the anion. However, since absorption is
governed by the aromatic portion of the molecule, their absorption maximum is usually
lower than other onium salts, near 250 nm.

Although the initiators absorb strongly at wavelengths near 250 nm, their

9.32

absorption diminishes at longer wavelengths. This fact could limit the efficiency of the

initiators for photopolymerizations driven by mercury lamps which provide most of their

9'” However, the spectral region over which the

emission at wavelengths above 300 nm.
initiators are effective may be expanded by the addition of a variety of photosensitizers,
including hydrocarbons, ketones and heterocyclic compounds.9 These compounds
actively sensitize the iodonium and sulfonium initiators by an electron transfer process,

9.32.33

rendering them effective in the long UV and visible wavelengths of light. Andersen et
al.34 showed that the radical polymerization of styrene could be initiated by photoexci-
tation of anthracene added to the styrene monomer. The anthracene initially absorbs the
light, which rapidly populates its metastable excited triplet state. This excited anthracene

then can interact with a stryene molecule to produce anthracene and styrene radicals

which lead to styrene polymerization.

7

Anthracene and its derivatives can be used to photosensitize cationic polymerization
reactions. In these reactions, the anthracene absorbs light and transfers that energy to the
diaryliodonium salt which then fragments and initiates the cationic polymerization. In
this way cationic photOpolymerizations can be initiated using near-UV or even visible
wavelengths. However, little is known about the kinetics and mechanism of the

photosensitation for these cationic reactions.
1.4. Need for Research on Cationic Photopolymerizations

UV-initiated cationic photopolymerizations have considerable potential for the
development of improved coatings and inks. These reactions may be used to rapidly
form highly crosslinked polymer films exhibiting excellent adhesion, abrasion resistance,
and chemical resistance. Furthermore, these 100% solid systems emit no volatile organic
Components and cure with a fraction of the energy requirements of thermal systems.
Unlike free radical photopolymerizations, the cationic reactions are not inhibited by
atmospheric oxygen, and may be used to polymerize epoxides and vinyl ethers. These
monomers exhibit many desirable properties, including low volatility, good rheological
CharaCtel-istics and negligible toxicity.7‘33 Fully formulated systems containing monomer,
p igment and initiator exhibit shelf lives of more than a year,'2 and initiators may be
Chosen to be effective for wavelengths at which the pigments have low extinction
Coefficients. Finally, the UV irradiation will not significantly degrade the substrate
(llnlike . . . . . .

electron beam curing) and the cationic polymerization can Withstand moderate

“‘8 of water without a significant decrease in reaction rate."“33

8
Despite the promise of UV-initiated cationic photopolymerizations, these

reactions have received only limited attention. This fact may be attributed primarily to
the lack of suitable initiators until recent years. In fact, most of the work on cationic
photopolymerizations reported in the literature focuses on the initiation step of the
reaction (for reviews, see references 8-12,30-33,35). The development of thermally stable
UV-sensitive cationic photoinitiators has lead to increased interest in the cationic

photopolymerizations of epoxides and vinyl ethers in the last ten years; 3"”‘35'3"

however,
the field is significantly less developed than that of free-radical photopolymerizations.
For example, the reaction kinetics of free radical photopolymerizations of multifunctional

acryl ates and methacrylates have been extensively investigated,"""29

revealing anomalous
kinetic behavior, including immediate onset of autoacceleration, the formation of
heterogeneous networks, and the attainment of a maximum conversion significantly less
than unity. However, similar experiments have not been reported for cationic
Photopolymerizations.

The reaction kinetics of cationic photopolymerizations of diepoxides and bisvinyl
ethers typically have been characterized only macroscopically in terms of the tack—free
time Of the polymer film. Similarly, the resulting highly crosslinked polymer networks
have t3’13ically received only a cursory characterization by the pencil hardness test and
solvent rub tests. In addition, the cationic photopolymerizations reported in the literature
are usually initiated using mercury lamps. There have been few studies as to the effect of
the W21\nflength of the light source on the polymerization rate (especially for wavelengths

be]
ow 300 nm). Fundamental investigations of the reaction and the resulting polymer

DCtW
011:5 are needed if the variables that effect the properties of the polymer films are to

7.

9
be fully understood. A thorough characterization of the reaction kinetics would allow the

optimum curing procedure to be established, and would reveal whether phenomena such
as limiting conversions, and trapped active centers are important in these systems. Char-

acterization of the resulting networks would allow the homogeneity of the network to be

determined, and could resulting in the establishment of important structure—property

relations.

1.5. List of References

13-1... Graminski, Fundamental Chemical and Engineering Issues in Paper Currency
Production, National Science Foundation, Washington DC, 1, 1991.

P-E. Sundell, S. Jonsson, A. Hult, “High-Energy-Radiation-Induced Cationic
Polymerization of Vinyl Ethers in the Presence of Onium Salt Initiators,” Radiation
Curing of Polymeric Materials, C.E. Hoyle and J.F. Kinstle Eds., ACS Symposium
Series 417, ACS, Washington DC, 459, (1989).

H-C. Kim, A.M. El-Naggar, G.L. Wilkes, Y. Yoo, 1E. McGrath, “Structure-Property
Behavior of Caprolactone-Ally Glycidyl Ether Copolymers Cross-Linked by
EleCtron-Beam Irradiation,” Radiation Curing of Polymeric Materials, C.E. Hoyle

311d J .F. Kinstle Eds., ACS Symposium Series 417, ACS, Washington DC, 474
(1989).

° V- Stannett and A. Deffieux, “Cationic Polymerization and Grafting of Initiated by

High Energy Radiation,” Cationic Polymerization and Related Processes, E.J.
Goethars, Eds., IUPAC, Oxford, UK, 307, 1984.

1' F’acansky and RJ. Waltman, “Electron-Beam Exposure of Organic Materials,”
Sadiation Curing of Polymeric Materials, C.E. Hoyle and J.F. Kinstle Eds., ACS
yrIlposium Series 417, ACS, Washington DC, 498 (1989).

J‘G~ Kloosterboer, “Network Formation by Chain Crosslinking Photopolymerization
and its Applications in Electronics,” Adv. Polym. Sci, 84, 1 (1988).

C'Q - Roffey, Photopolymerization of Surface Coatings, Wiley, New York, 1981.

10

8. S.P. Pappas, UV Curing: Science and Technology, Vol. 2, Technology Marketing
Corporation, Norwalk, CT, 1985.

9. 1V. Crivello, “Cationic Polymerization-Iodonium and Sulfonium Salt Photo-
initiators,” Adv. Polym. Sci, 62, 1, (1984).

10. J.V. Crivello and IL. Lee, “Alkoxy-Substituted Diaryliodonium Salt Cationic
Photoinitiators,” J. Polym. Sci, Polym. Chem, 27, 3951, (1989).

11. A. Ledwith, S. Al-Kass, A. Hulme-Lowe, “Mechanisms for Photoinitiation of
Cationic Polymerization,” Cationic Polymerization and Related Processes, E.J.
Goethals, Ed., IUPAC, Oxford, UK, 275, 1984.

12. J- V. Crivello, “Recent Progress in the Design of New Photoinitiators for Cationic
Polymerization,” Cationic Polymerization and Related Processes, E.J. Goethals, Ed.,

IU PAC, Oxford, UK, 289, 1984.
13. A- Reiser, Photoreactive Polymers, Wiley, New York, NY, 1989.

14. S -P.Lapin, “Radition-Induced Cationic Curing of Vinyl-Ether—Functionalized
Urethane Oligomers,” Radiation Curing of Polymeric Materials, C.E. Hoyle and J.F.
Kinstle, Eds., ACS Symposium Series 417, ACS, Washington DC, 363, (1989).

15. P-E.M. Allen, G.P. Simon, D.R.G. Williams and EH. Williams, “Dynamic-
Mechanical Properties and Cross-Polarized, Proton-Enhanced, Magic Angle Spinning
”C NMR Time Constants of Poly[oligo(ethylene glycol) dimethacrylates],” '
Macromolecules, 22, 809, (1989).

16. K- Miyazaki and T. Horibe, J. Biomedical Mater. Res, 22, 1011, (1988).

l7.G.H- Hubca. C.R. Oprescu, G.H. Dragan, and M. Dimonie, “Synthesis and
P01)r’l‘nerization of Polyethylene Glycol Dimethacrylates. IV, Study of the
POlyrnerization Process of Polyethylene Glycol Dimethacrylates,” Rev. Roum. Chem,
27. 435, (1982).

18‘ J'G- IQoosterboer and C.F.C.M. Litjen, “Chain Cross-linking Photopolymerization of
Tetl‘aethyleneglycol Diacrylate,” Cross-linked Polymer Chemistry, R. Dickie, S.
LabElna and R. Bauer,Eds., ACS Symposium Series 367, ACS, Washington DC,
409. (1988).

19.
H‘MJ. Boots, 1G. Kloosterboer, and G. van de Hei, “Inhomogeneity During Bulk

PolYmerization of Divinyl Compounds,” Brit. Polym. J., 17, 219, (1985).

11

20. H. Galina, B. Kolarz, P. Wieczorek, and M. Wojczynska, “ Homogeneity of Gel and
Microsyneresis in Porous Styrene-Divinylbenzene C0polymers,” Brit. Polym. J., 17,
215, (1985).

21. W. Funke, “Reactive Microgels-Polymers Intermediate in Size Between Single
Molecules and Particles,” Brit. Polym. J., 21, 107, ( 1989).

22. J. Bastide, and L. Leibler, Macromolecules, 21, 2649, (1988).

23. A. Matsumoto, H. Matsuo, H. Ando and M. Oiwa. “Solvent Effects in the
Copolymerization of Methyl Methacrylate with Oligoglycol Dimethacrylate,” Eur.
Polym. J., 25, 237, (1989).

24. J. Baselga, M. Llorente, 1. Hernandez-Ruentes, and I. Pierola, “Network Defects in
Polymacrylamide Gels,” Eur. Polym. J, 25, 471, (1989).

25. D. Turner, Z. Haque, S. Kalachandra and T. Wilson, “Structure and Properties of
Polydimethacrylates: Dental Applications,” Polym. Mater. Sci. Eng. Proceed, 56,
769, (1987).

26. GP. Simon, P.E.M. Allen, D]. Bennett, D.R.G. Williams and EH. Williams,
“Nature of Residual Unsaturation During Cure of Dimethacrylates Examined by
CPPEMAS l3C NMR and, Simulation Using a Kinetic Gelation Model,”
Macromolecules, 22, 3555, (1989).

27. LG. Kloosterboer, G.F.C.M. Lijten and C.P.G. Zegers, “Formation of Densely
Crosslinked Polymer Glasses by Photopolymerization,” Polym. Mater. Sci. Eng.
Proceed, 60, 122, (1989).

28. P.E.M. Allen, D]. Bennett, S. Hagias, A. Hounslow, G. Ross, G.P. Simon, D.G.R.
Williams and EH. Williams, “The Radical Polymerization of Dimethacrylate
Monomers - The Use of NMR Spectrometry to Follow the Development of Network
Structure,” Eur. Polym. J., 25, 785, (1989).

30. J.V. Crivello and J .H.W. Lam, “The Photoinitiated Cationic Polymerization of Epoxy
Resins,” Epoxy Resin Chemistry, R.S. Bauer, Ed., ACS Symposium Series, 114,
ACS, Washington DC, 1, (1979).

31. SP. Pappas, “Photoinitiation of Cationic and Concurrent Radical-Cationic Poly-
merization. Part V.,” Prog. Org. Coat, 13, 35, (1985).

32. J.V. Crivello, “Photoinitiated Cationic Polymerization of Coatings,” Organic
Coatings, Science and Technology, Vol. 5, GD. Parfitt and AN. Patsis, Eds., Marcel
Dekker, Inc., New York, NY, 35, 1983.

12

33. W.R. Watt, “Photosensitized Epoxides as a Basis for Light-Curable Coatings,” Epoxy
Resin Chemistry, R.S. Bauer, Ed., ACS Symposium Series 114, ACS, Washington
DC, 17, (1979).

34. VS. Anderson and R.G.W. Norrish, “The Polymerization of Styrene Sensitized by
Molecules in the Triplet State,” Proc. Roy. Soc. (London), 251, l, (1959).

35. F. Loshe and H. Zweifel, “Photocrosslinking of Epoxy Resins”, Adv. Polym. Sci, 78,
61, (1986).

36. J.V. Crivello and DA. Conlon, “Aromatic Bisvinyl Ethers: A New Class of Highly
Reactive Therrnosetting Monomers,” J. Polym. Sci, Polym. Chem, 21, 1785, (1983).

38. G. Manivannan, J.P. Fouassier, “Primary Processes in the Photosensitized Poly-
merization of Cationic Monomers,” J. of Polym. Sci. Polym. Chem, 29, 1113, (1991).

39. P. Sundell, S. Jonsson, A. Hult, “Photo-Redox Induced Cationic Polymerization of
Divinyl Ethers,” J. Polym. Sci, Polym Chem, 29, 1525, (1985).

CHAPTER II. BACKGROUND

11.1. Cationic Polymerizations

11.1 .1 . 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 wolefins, 1,3-dienes,
vinyl ethers, and epoxides. Cationic polymerizations are not as well characterized as
radical polymerizations partially because they proceed at very rapid rates, and may be
sensitive to small concentrations of impurities.l 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 should not affect the polymerization rate because the resulting
cation will continue to propagate.' The decrease in the average primary polymer chain

length will not effect the properties of the polymer due to the high level of crosslinking.

l3

14
The average molecular weight will not be effected greatly, also due to the highly cross-

linked network formed in the reaction.

Kinetic data for cationic polymerizations are difficult to interpret for several
reasons."3 First of all, the reactions often proceed so rapidly that the steady-state
assumption is not valid for the reactive centers.1 The analysis is further complicated by
the fact that the nature of the propagating center and the counterion are often unclear.”
For kinetic modeling of these reactions, it is common to assume several types of
propagating centers ranging from completely free (unpaired) ions to covalent species.
The unpaired species typically exhibit much higher kinetic constants for propagation than
pair species. Cationic polymerization rates are often higher than those for free radical
polymerizations for a variety of reasons. For cationic polymerizations of unsaturated
monomers, the kinetic constants for propagation are usually higher than those for free
radical reactions, while the rate constants for termination are lower. Both of these trends
contribute to higher reaction rates for the cationic reactions. Furthermore, the
concentration of propagating species is usually two orders of magnitude higher for
cationic polymerizations than for free radical polymerizations.

While cationic polymerizations are not used extensively for the commercial
synthesis of long polymer chains, they show considerable promise for
photopolymerizations of highly crosslinked polymer films. 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 distributions. These
facts tend to limit the usefulness of the reactions for the synthesis of linear high polymers.

However, for photopolymerizations of multifunctional monomers, a high linear primary

15
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.

Fundamental investigations of the reaction and the resulting polymer networks are
needed if the variables that effect the properties of the polymer films are to be fully
understood. A thorough characterization of the reaction kinetics would allow the
optimum curing procedure to be established, and would reveal whether phenomena such
as limiting conversions, and trapped active centers are important in these systems. Char-
acterization of the resulting networks would allow the homogeneity of the network to be
determined, and could result in the establishment of important structure-property
relations. Finally, the effect of the wavelength and light source on the kinetics and the

reaction/structure of the polymer films should be thoroughly investigated.
11.1.2. Cationic Photopolymerizations of Epoxides and Vinyl Ethers

The development of thermally stable cationic photoinitiators in the late 19705 has
lead to increased interest in cationic photopolymerizations. Most of the work reported in
the literature has focused on the synthesis and development of appropriate monomers and
initiators for cationic photopolymerizations of high performance polymer films."23

Several classes of monomers exhibiting desirable properties and rapid polymerization

N

rates have been reported, including epoxides,""’ novel silicon-containing epoxy

17.18.23 4.l5.16

resins and a variety of urethane and bisvinyl ethers. The selection of appropriate

16
monomers and initiators is now reasonably broad; however, the characteristics of the

cationic photopolymerization reactions have received only limited attention. There is a
need for fundamental studies of the kinetics of the reactions, as well as the structure and
physical properties of the resulting highly crosslinked polymer networks. Such studies
would allow the optimum formulations and curing procedures to be established.
Experimental 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 upon 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 triaryl-
sulfonium salts are used as the photoinitiators, the optimum initiator concentrations are

10.13

typically 2-3 wt%. Until now, most investigators have used mercury lamps for the UV

9.10.13

light source. Many monomers exhibit a substantial increase in cure rate as

temperature is increased. “”3””

This fact presents potential complications for kinetic
studies of extremely rapid cationic photopolymerizations. Polymerization reaction of
epoxides and vinyl ethers are highly exothermic and can liberate significant amounts of
heat in a very short period of time. Therefore, it is important to monitor the temperature
of the film during the polymerization process.

The cure rate of cationic photopolymerizations is typically characterized by the
tack-free time of the polymer film. Cycloaliphatic epoxides generally exhibit shorter
tack-free times than bisphenol-A epoxy resins due to enhanced epoxy ring strain." Vinyl

ethers and silicon-containing epoxides exhibit reaction rates that are an order of

magnitude higher than these epoxides?“ Reported tack-free times range from less than

17
one second to several minutes?” however, the meaning of the tack-free time in terms of

the degree of cure is unclear. Differential scanning calorimetry experiments9 indicate that
the reaction proceeds long after the tack-free time. In fact, films with tack-free times of a
few seconds often exhibit exothermic polymerization reactions for several minutes. The
studies also indicate that UV cured epoxides typically exhibited a cure rate which passed
through a maximum, then rapidly decreased to smaller values. This behavior was
attributed to diffusion limited propagation upon vitrification in the highly crosslinked
films?” This interpretation is supported by the fact that tack free time may be decreased

by the addition of chain transfer agents”2 which delay vitrification. Unfortunately the
chain transfer agents cause a reduction in crosslink density, undermining the mechanical

properties and the chemical resistance of the cured film.9

11.1.3. Photosensitization of Cationic Polymerizations

Although there have been relatively few systematic studies as to the effect of the
wavelength of the light source on the characteristics of cationic polymerizations, several
experimental observations suggest that wavelength could be an important operating
variable. For example, the cure rate of a given monomer depends upon which initiator is
used for the photopolymerization. When medium pressure mercury lamps are used as the
light source (these lamps provide most of their emission in the 300-340 nm region of the
spectrum‘), the cure rate is highest for diazonium salts, followed by triarylsulfonium, and
diaryliodonium salts.l3 This trend can be attributed to more efficient use of the mercury

arc radiation by the diazonium salt since this initiator exhibits a maximum absorbance at

18
a wavelength of approximately 320 nm, '3 while the triarylsulfonium and diaryliodonium

salts exhibit their maxima at lower wavelengths of around 260 and 235 nm, respectively.
A similar effect is observed when photosensitizers are added to a reaction mixture.“"'2'
The photosensitizers lead to an increased cure rate due to more efficient use of the
mercury arc radiation. These results illustrate the importance of the initiation step in the
determination of the cure rate and results suggest that the effect of light sources operating
in shorter wavelengths should be investigated. One method of overcoming short
wavelength absorption of onium salts is to use photosensitizers to expand the spectral
region over which onium salts are effective. Photosensitizers make it possible to initiate
polymerization in the near UV or visible wavelengths of light as a result of a direct
interaction between an excited state of a photosensitizer and the initiator. Heterocyclic
and fused ring aromatic compounds often show activity as photosensitizers. Anthracene
has been used to photosensitize both free radical” and cationicm" polymerizations.
Anthracene absorbs strongly at wavelengths between 320-390 nm, and therefore expands
the initiating window to the near visible region of the spectrum (overlapping with

mercury lamps emission in the 300-340 nm region of the spectrum‘).

11.2. Fluorescence Cure Monitoring

To date, most of the work on cationic photopolymerizations reported in the
literature focuses on the initiation step of the reaction, and the polymerizations
themselves are not well characterized. It is clear that cationic photopolymerizations of

Vinyl ethers proceed very rapidly; however, few detailed kinetic studies have been

 

_‘

19
performed for these reactions. For example, the kinetics of these photopolymerizations

have been characterized in a very cursory manner in terms of the tack-free time of a
curing polymer film, but detailed profiles of conversion versus time are not available.

A major reason for the lack of kinetic studies is the extremely rapid reaction rates
exhibited by cationic photopolymerizations of divinyl ethers. Because these reactions
proceed to completion in a matter of seconds, there are few experimental techniques with
sufficient time resolution to characterize these polymerizations. Differential scanning
calorimetry (DSC) has been used extensively to study free radical photopolymerizations
of acrylates” as well as cationic photopolymerizations of epoxides,”28 however the
technique’s relatively slow response-time limits its applicability for vinyl ether
photopolymerizations. Since the DSC response time is approximately 2-3 seconds, 2”"
the technique is only applicable for vinyl ether photopolymerizations that are initiated
with relatively low intensity light16 when the cure time is relatively long (tens of seconds).
Real-time infrared spectroscopy has recently been used to monitor high-speed, laser-

induced radical photopolymerizations. 293'

This technique has a response time of 0.02
seconds and therefore could be useful for characterizing cationic photopolymerizations.
The advantages and limitations of this technique have recently been addressed. 29
Furthermore, polymerizations of multi-functional monomers are complicated by the fact
that highly crosslinked structures are formed, which makes it difficult to measure the
primary kinetic chain length. For this reason, the photopolymerizations of monoepoxides

m" Information obtained

have received more detailed analysis than that of diepoxides.
from linear polymerizations is of limited value for highly crosslinked polymerizations of

multi-functional monomers. New techniques such as fluorescence spectroscopy need to

20
be developed as a time-resolved, in situ technique for monitoring the extent of cure in

these high speed polymerizations.
Fluorescence methods are well suited to these experiments because they provide a
highly sensitive and selective means of non-invasive, real—time measurement of reactant

concentration?”

Fluorescence spectroscopy is particularly attractive as an in situ, time-
resolved technique for monitoring these high speed polymerizations because it has an
extremely short intrinsic timescale (typically 10'9 seconds). The response time for the
monitoring technique is invariably determined by the speed of the detection system. In

recent years, several authors”37

have reported the use of fluorescence techniques for
monitoring polymerizations. Most of these fluorescence techniques monitor polymeriza-
tion indirectly through the corresponding increase in viscosity. For example, several
fluorescence monitoring schemes are based upon fluorescence probes which exhibit
enhanced fluorescence intensity with increasing viscosity due to decreased radiationless

34.37

energy transfer. Wang et al. have used a similar technique to monitor viscosity and
chemical changes during polymerization of epoxy and methyl methacrylate systems.
They determined that for certain probe molecules, the fluorescence intensity increases as
the system viscosity increases due to the reduced ability of the probe to relax through
vibrational and rotational modes.” Their reaction system also included a probe molecule
which has constant fluorescence intensity with viscosity changes as an internal standard”
This technique was not very useful for monitoring the degree to cure above the gel point
and was subject to uncertainty due to variations in background fluorescence. Alter-

natively, Stroeks et al.” and Wang et al.35 investigated the use of viscosity-sensitive

excimer formation to monitor the cure of epoxy resins. However, Stroeks et al.

 

21
concluded that the apparent intensity changes in the excimer peak were actually

dominated by changes in the intensity of the overlapping monomer peak (which was
enhanced due to reduced radiationless transfer). Finally, Scarlata and Ors‘” have
monitored the cure of an epoxy resin by measuring an increase in fluorescence
polarization which they attributed to a viscosity-induced decrease in the rotational
mobility of the probe molecule.

One drawback of the viscosity-sensitive techniques for monitoring crosslinking
polymerizations is their limited utility for conversions above the gel point where there is
relatively little change in the local viscosity with time. A novel fluorescence monitoring

39.41 1

method not based upon viscosity, was recently reported by Sung et al. Their scheme,
which is applicable to amine/epoxy systems, is based upon a reactive chromophore which
mimics the reactivity of the amine curing agent. The fluorescence quantum yield of the
chromophore increases as the primary amine is converted to a tertiary amine, thereby

41.42

allowing the reactive state of the amine to be monitored. In more recent work, Sung
and collaborators have used intrinsic fluorescence and phosphorescence to characterize
cure in epoxy composites. Bathochromic shifts in emission spectra were observed as a
function of cure time giving a sensitive method for evaluating extent of cure.

The kinetics of cationic photopolymerizations of vinyl ethers can be investigated
using in situ laser-induced fluorescence spectroscopy. The fluorescence intensity of the
photosensitizer can be monitored during the reaction to obtain a direct measure of the rate
of initiation, and provide indirect information on the rate of propagation. The response

time for the monitoring technique is determined by the speed of the detection system. In

our studies, a high-speed diode array detection system was employed, and fluorescence

22
spectra were collected in intervals as short as 2 milliseconds. Therefore, this fluorescence

monitoring method has a considerably faster response time than DSC and real-time IR
methods, and provides ample time-resolution to characterize the cationic photo-
polymerizations. In addition, since the fluorescence method allows a relatively wide
window of wavelengths to be monitored (as opposed to a single wavelength), spectral
changes arising from more than one functionality can be monitored simultaneously.

The fluorescence monitoring scheme reported here is considerably different than
the viscosity-sensitive methods reported in the literature. First, because cationic photo-
polymerizations proceed to completion in a matter of seconds, our studies require
considerably faster sampling methods than the aforementioned epoxy resin studies. A
more fundamental difference between our fluorescence monitoring scheme and the
viscosity-sensitive probe approach arises from the fact that we are monitoring the
fluorescence of one of the reactants (the photosensitizer), and not that of an inert probe.
This technique provides a more direct measurement of the reaction itself, and does not
lose sensitivity at the gel point. Laser induced fluorescence spectroscopy can be used to
monitor divinyl ether photosensitization reactions because the intensity of many of the
aromatic hydrocarbons used as the photosensitizers changes as the reaction proceeds.
Therefore, laser induced fluorescence was chosen to monitor the photosensitization rate
of cationic polymerization of divinyl ethers photosensitized with varying concentrations

of anthracene and its derivatives.

23

11.3. Photosensitization Mechanism

In order to expand the spectral region over which cationic initiators are effective,
a variety of photosensitizers may be added to the system. In general, photosensitization
results from a direct interaction between an excited photosensitizer molecule (the first
excited singlet state or the lowest energy triplet state) and the initiator. Photosensitization
is often attributed to an energy transfer mechanism in which the excited-state
photosensitizer interacts with a ground state initiator molecule to form an exciplex. 1n the
exciplex. the excited-state energy is transferred to the initiator, which fragments to create
active centers. According to this mechanism, the photosensitizer functions essentially as
a catalyst and is regenerated in its ground-state while the initiator is promoted to an
excited state, and fragments to create active centers. For energy transfer to be thermo-
dynamically feasible, the excited-state energy of the photosensitizer must be greater than
that of the acceptor.‘ The energy transfer mechanism is not feasible for photo-
sensitization of onium salts by anthracene derivatives due to the relatively low excited
state energy of the anthracene derivatives. Pappas and Gatechair“ reported that this
energy transfer process may occur for onium salts with photosensitizers such as
acetophenone and benzophenone which have relatively high triplet energies (> 290
kJ/mol) and high oxidation potentials, but the energy transfer results in a triplet state
initiator molecule which is unreactive to bond cleavage.“ In more recent work, Dektar
and Hacker demonstrated that triplet energy transfer photolysis does take place for

diaryliodonium salts sensitized by xanthone and benzophenone. However, the triplet

24
energy transfer mechanism is not thermodynamically feasible for photosensitization of

iodonium salts by anthracene due to the relatively 10w excited state energy of the
anthracene.”
It has been generally accepted that the photosensitization of diaryliodonium

'25 Electron transfer

initiators most often proceeds by the electron transfer mechanism."ll
from the excited state of the photosensitizer to the initiator within an exciplex results in a
photoredox reaction in which the photosensitizer is oxidized to form a radical-cation and
the initiator is reduced to a radical species.‘3 The thermodynamic feasibility of this
reaction is governed by the oxidation potential of the excited state photosensitizer and the
reduction potential of the initiator. Although there is general agreement that
photosensitization of diaryliodonium salts by anthracene occurs by electron transfer, there
is disagreement on the identity of the anthracene excited state (singlet or triplet) from
which the electron transfer occurs. In early work Andersen et al.“ demonstrated that the
radical polymerization of styrene could be photosensitized by anthracene. Based upon
lifetime considerations, the authors concluded that it was triplet state anthracene that
interacted with styrene to produce anthracene and styrene radicals which lead to
polymerization. Crivello4 reported that the photosensitization of iodonium salts by aryl
ketones, condensed ring aromatic hydrocarbons and dyes takes place by an electron
transfer mechanism from the lowest energy triplet state. Based upon anthracene emission
quenching studies, Pappas and collaborators"'” concluded that the electron transfer occurs
from the first excited singlet state of anthracene. Similarly, Schnabel et. al."'” have

concluded from flash photolysis studies, that anthracene reacts with N-ethoxy-Z-methyl-

pyridinium ions from the excited singlet state. However, none of these authors

25
demonstrated their proposed mechanisms with definite experiments. Finally, based on

CINDP (Chemically Induced Dynamic Nuclear Polarization) studies, DeVoe et. a1" “
concluded the electron transfer to the iodonium salts occurs from the anthracene excited
singlet state.

A series of studies will be performed to establish the photosensitization
mechanism for cationic photopolymerizations of vinyl ethers photosensitized by
anthracene derivatives. Results obtained using several different anthracene derivatives as
the photosensitizer will provide considerable insight into the photosensitization
mechanism. Furthermore, a series of studies will be performed to determine which
anthracene excited state (singlet or triplet) is involved in photosensitization of diaryl-
iodonium salts. Anthracene phosphorescence and fluorescence lifetime studies were
performed to determine the effect of the initiator on the lifetimes of the triplet and singlet.
The phosphorescence measurements will be used to examine the triplet lifetime of the
anthracene when used as a photosensitizer. The role of the triplet state will be
investigated further using studies performed using an anthracene triplet quencher and
from computer simulations of the photophysical-state populations. Other photo-induced
reactions of anthracene which could interfere with the photosensitization will also be
investigated. Dimerization and oxidation will be examined as competing reactions, in
order to eliminate the possibility of the reduction in anthracene fluorescence being
attributed to these side reactions.

Both the reactivity and relative population of the singlet and triplet state need to
be investigated to determine the relative number of active centers generated from each

state. The relative population of the triplet and singlet state depends primarily on the

26
intersystem crossing rate constant. The reaction pathway for photosensitization can be

represented by a series of coupled differential equations. The differential equations are
first or second order and depend strongly on the rate constants for the reaction. Solving
these differential equations by numerical methods using rate constants from experiments
or literature can provide a model of the photosensitization process. The Rosenbrock
method for stiff systems with variable step size will be used to solve the set of equations,
using developed or commercially prepared programs. The predictions from the models
will be matched with experimental data to aid in determining the relative number of

active centers formed from the singlet and triplet states.

11.4. Infrared and Raman Cure Monitoring

[1.4.]. Introduction to Infrared and Raman Spectroscopy

Due to technological advances, infrared (IR) and Raman spectroscopy are
becoming increasingly important tools for the investigation of polymers. With the
development of Fourier transform infrared (FTIR) spectroscopy, spurred by the discovery
of the fast Fourier transform (FFT) algorithm by Cooley and Tukey in 1965,” the
usefulness of IR spectroscopy was greatly expanded. Even today as sampling techniques,
lasers, detection systems, computers and software become more sophisticated previous
limitations continue to be overcome. For example, it is now possible, and desirable in
many cases, to examine polymer systems in situ.” High speed FTIR instruments are now

capable of collecting an entire spectrum in times as short as 5 ns." Until recently, IR

27
techniques have been used far more extensively than Raman techniques. However. in the

late 1960’s as laser technology started to come of age, Raman spectroscopy developed as
a complementary procedure.’2 Once regarded solely as a laboratory technique, Raman
spectroscopy can now be used for on-line process analysis, and environmental
monitoring. For example, there are now commercially available, portable, on-line Raman
systems containing various novel lasers, fiberoptic sampling probes, CCD detectors and
complete computer analysis systems.’3

IR and Raman are practical techniques for hydrogels, since the infrared absorption
spectra for polymers are surprisingly simple considering the large number of atoms
involved.” The simplicity results from the fact that many normal vibrations have almost
the same frequency and appear as one band. 9 Furthermore, the strict IR and Raman
selection rules prevent many of the vibrations from producing absorptions,$4 therefore
simplifying the spectrum.

The usefulness of IR and Raman spectroscopy ultimately lies in the variety of
qualitative and quantitative information they can provide about the nature of the polymer
under investigation. Information that can be obtained using these techniques includes:
the chemical nature (structural units, branching, end groups, additives, and impurities),
steric order (isomerism and stereoregularity), conformational order (physical arrangement
of the chain), state of aggregation (crystalline, mesomorphous, amorphous phases,
number of chains per unit cell, etc.), and orientation of the polymer chain and side
groups.’2

IR and Raman spectra provide complementary rather than duplicate information.

For example, IR spectroscopy is generally better suited for the identification of polar

28
groups, while Raman is valuable for the investigation of the homonuclear polymer

backbone.’ 2 Raman spectroscopy is well suited for the study of low concentration
aqueous systems such as hydrogels, since water is an ideal solvent due to its poor Raman
scattering. This is a distinct advantage over 1R techniques since water absorbs strongly in
the IR region and overwhelms the spectrum.” Another advantage of Raman spectroscopy
arise from the fact that Raman scattering from molecular vibrations can be measured in
the visible region of the spectrum. In the this region, optics can be made of relativity
inexpensive and robust materials such as glass or quartz, and sensitive detectors with high
signal-to-noise ratios are readily available.’6 Since Raman spectroscopy is a visible light
technique, remote sampling with fiberoptic probes gives excellent signal throughput and
precise positioning of the sampling beam. Intrinsic limitations of Raman spectroscopy
arise from the requirement of an intense monochromatic (preferably columnated) light
source, and the possible interference from highly colored or fluorescence materials.”

As discussed in detail later in this chapter, IR and Raman spectroscopy are based
upon distinct physical processes. Fundamentally, Raman is a scattering technique while
IR spectroscopy is based upon light absorption, therefore the sampling techniques of the
two methods are significantly different. In transmission IR spectroscopy the light is
required to pass through the sample to reach the detector, therefore severe constraints are

imposed on the sample thickness (typically the thickness must be less than about 40 um).

Methods such as attenuated total reflectance (ATR) may be used with thick samples, but
present thier own problems and limitations associated with the contact to the ATR
crystal, and the limited depth of penetration. In contrast, since Raman spectroscopy is

based upon light scattered by the sample thick samples are not a problem since

29
backseattered light may be collected and analyzed. Therefore Raman spectroscopy

generally requires very little, if any, sample preparation.

The most significant limitations of Raman spectroscopy arise from interference
from samples that are highly fluorescent or colored. This limitation ultimately arises
from the relatively low intensity of the scattered Raman light, which necessitates intense
light sources. In colored samples, the incident light may lead to significant heating of the
sample. If the incident light induces fluorescence, the Raman signal must be resolved
from the fluoresced light (fluorescene is typically several orders of magnitude more
intense than Raman scattering). Both techniques are good for analysis of microscale
sample sizes. Although, IR spectroscopy has the advantage of better resolutions with
resolutions down to approximately 0.05 cm", while Raman instruments are limited to
about 0.25 cm".” In addition, both techniques are fairly easily adapted for high and low
temperature work.’6

There are two general types of infrared spectrometers. The older of the two types
is referred to as a dispersive spectrometer. This type of spectrometer uses a mono-
chromator and slits to separate the infrared radiation into individual frequencies?7 The
more recently developed type of instrument is known as a Fourier transform infrared
spectrometer (FTIR). These instruments are based on the Michelson interferometer and

require no monochromators or slits.’7

The areas of continuing improvement in FTIR
instrumentation include increased scan speed and resolution, improved signal-to-noise
ratios, expansion of the range of wavelengths over which the instrument operates, data

handling software, and automation. Examples of these improvements include FTIR

instruments which are capable of resolutions down to 1 cm" and signal-to-noise ratios of

30
2,000:1 with high speeds able of obtaining spectra in intervals as short as 5

nanoseconds."

The most significant area of improvement in Raman spectroscopy has been the
development of relatively low cost, high power lasers. Some of the more recent
improvements include enhanced detection systems, such as charge coupled devices

(CCD), and the use of fiber-optics for remote process analysis.”

Other improvements in
IR and Raman instrumentation make use of the increased power of personal computers
and computer software. It is now possible to collect, analyze, annotate, and even “cut and
paste” selected spectrographs into reports and presentations all within the same computer.
In addition, computerized libraries of known IR and Raman spectra continue to grow,
allowing for more extensive computer matches of characteristic group frequencies."
There are many applications of IR and Raman spectroscopy to the investigation of
hydrogels. Specific applications include investigations of controlled drug release

delivery systems, and contact lens materials; identification of biocompatible polymers;

and characterization of synthesis products.
II.4.2. Physical Basis and Background for IR Spectroscopy

A brief introduction to the basic principles of infrared spectroscopy is presented in this
section. The infrared (IR) region of the spectrum consists of light with wavenumbers

between 10,000 cm‘I and 10 cm'I (wavelengths between 1.0 pm and 1,000 um), with the
range from 10,000 to 4,000 cm’l (1,000 - 2.5 um) referred to as the near infrared; 4,000 to
400 cm‘I (2.5 - 25 pm) the mid-infrared; and 400 to 10 cm'I (25 - 1,000 um) the far

infrared regions.” The basic premise behind IR spectroscopy is that when infrared light

31
is passed through a sample, certain frequencies are absorbed while others are transmitted,

resulting in an absorption spectrum that is dependent on the molecular vibrational
frequencies.

In polymers, molecular bonds have different vibrational frequencies which
correspond to characteristic IR absorption bands. Therefore an IR spectrum may be used
for identification of the types of bonds and functional groups present in a specific
polymer sample.’8 IR absorption bands due to rotational transitions may be observed for
gases;52 however, the rotation of the molecule is highly restricted in solids, and rotational
transitions are not observed. For the study of polymer systems we are only concerned
with molecular vibrations posessing characteristic absorption bands throughout the IR
spectrum.

In an infrared spectrometer, the sample under investigation is exposed to the light
which spans the entire 1R region (either simultaneously as in an FT IR or one wavelength
at a time as in a dispersive 1R). Both the classical and quantum mechanical viewpoints
are useful in describing what fundamentally occurs on the molecular level. The sample
absorbs the light only at the specific frequencies that match the natural vibrational
frequencies of its constituent molecules.” In order for a molecular vibration to absorb in
the IR region it must cause a change in the magnitude of the dipole moment.”"‘"‘” The
intensity of this absorption is dependent on how effective the molecule transfers the
infrared photon energy to molecular vibrational energy. This is dependent on the change
induced in the dipole moment.” It has been shown that the intensity of an IR absorption

59.60.62

band is proportional to the square of the change in its dipole moment. If a molecule

32
has a center of symmetry in its equilibrium configuration, then any vibrations for which

this symmetry is retained will be infrared inactive.”
The frequency of a photon, V”, is determined by its energy, Er, according to
Equation 1.”
EP = hvp Eq. (1)
Quantum mechanical considerations dictate that molecular energy assumes only specific
discrete values. The values of the vibrational energy levels, E...» are given by,”
E“, = (n+ ’/,)hv Eq. (2)
Where h is Planck’s constant, v is the classical vibrational frequency of the oscillator and
n is a quantum number. If this equation is applied to the infrared absorption of a photon
where the vibrational energy of the molecule is Em, and the frequency of the induced
vibration is v", then by substitution,
E, = (n+ 7,)th Eq. (3)
The only variable in this equation is the quantum number, which can only have integer
values. If the quantum number changes by +1 or -1, then the molecule gains or loses
energy, respectively. For an unsymmetrical diatomic molecule this amount of energy is
AEM, which is equal to the difference between the energy of the n quantum level the
energy of the n+1 level. This gives the result,”
AEM = hvm Eq. (4)
Since the change in energy, AB", is due to the transfer of energy from the photon to the
molecule, AB", = En and subsequently vp = v”. This states that the frequency, vp, of a
photon which has the proper energy to cause the vibrational quantum number to jump up

one level is equal to the classical vibrational frequency of the molecule, vm.” For

33

harmonic vibration the vibrational quantum number may only change by :1, all other

transitions are forbidden. This selection rule corresponds to conditions where the electric
field of the photon causes the molecule to vibrate at the field’s frequency, and induces a
molecular dipole moment. For an anharmonic quantum mechanical oscillator the

selection rule allows for integer changes in the quantum number of i1, i2, i3, etc. In

this case, the excited molecule vibrates at the corresponding integer multiple of the
molecular frequency.” At room temperatures most molecules are in the ground
vibrational state, therefore, the allowed transition that dominates infrared spectroscopy
and is of most interest is from n = 0 to n = 1. This is referred to as the fundamental
transition. The relative intensities of other allowed transitions are low since the
populations of the higher energy levels are low compared to the ground state at room
temperatures.” Although additional bands, known as overtones and combination bands,
may result from anharmonic terms in the potential energy of the molecule.M Since
anharmonic oscillator energy levels are not equally spaced, the frequency is no longer
completely independent of amplitude.” The additive effect of the harmonic and
anharmonic terms results in the appearance of absorption bands that are shifted from their

characteristic absorption wavelength.
II.4.3. Instrumentation for Infrared Spectroscopy
The original type of infrared technique is known as a dispersive IR. A dispersive

1R instrument consists of prisms or gratings that separate or disperse the IR radiation into

its component wavelengths."l Generally modern dispersive infrared spectrometers are

34
made using gratings rather than prisms with a series of diffraction gratings used to

. . 59
separate the entire 1R region.

The diffracted light is then reflected with an off-axis
paraboloid through an exit slit into a detector. A movable mirror allows a progressive
scanning of the individual wavelengths. As a whole, this setup acts as a monochromator
thus focusing one wavelength at a time onto the detector. Before reaching the
monochromator, the IR radiation is first passed through the sample. In a double beam
instrument the 1R radiation is split into two beams, one of which is passed through the
sample; the other is passed through a reference cell. A rotating sector mirror, called a
chopper, alternately reflects or transmits each beam to the monochromator.” By
comparing the two intensities, the absorption of the sample can be determined.

In contrast to the traditional dispersive instruments, the modern Fourier transform
infrared (FTIR) spectrometers do not require a system of slits or a monochromator.
Instead, light containing all IR frequencies is passed through the sample simultaneously
and those not absorbed reach the detector. The advantage is an improved signal-to-noise
ratio and a higher throughput compared to the dispersive IR method.”'””3 The key
component of the FTIR instrument is the Michelson interferometer, which was first
developed around the turn of the century, but was not applied to infrared spectroscopy
until the development of digital computers that could quickly and accurately derive the IR
spectra from the interferograms by taking the inverse Laplace transform.”

In the Michelson interferometer, a beamsplitter is used to transmit half of the
source radiation to a movable mirror and to reflect the other half to a fixed mirror. The

two beams are then reflected back to the beamsplitter and recombined either

constructively or destructively depending on the position of the movable mirror.” In an

35
FT IR spectrometer, this recombined beam is then passed through the sample to a

detector,” where the interferogram is obtained. In the simplest case of a monochromatic
radiation source the intensity (amplitude) of the detector signal is merely the cosine
function of the mirror position.”‘” For polychromatic radiation, however, the
interferogram is a summation of all of the constructive and destructive interferences
resulting from the interactions between all of the wavelengths.” The cosine Fourier
transform provides the mathematical relationship between the detected interferogram
intensity, I(x), which is a function the mirror position, and the infrared absorption

intensity, I( v), as a funciton of frequency,”63
I (x) = J'Mv) cos(21txv)dv Eq. (5)

In order to obtain the infrared spectrum from this interferogram quickly, a computer is

required to calculate the following inverse transform,”"‘3

I (v) = J! (x) cos(21txv)dx Eq. (6)

The optical path difference, x, is a function of the position of the moveable mirror and is
very accurately determined through the use of an internal standard typically a helium-
neon laser.“3

Ideally, the light source for infrared spectroscopy should be continuously tunable

5 .
9“ One source

over the entire infrared region at power levels above the microwatt range.
that provides an ideal continuous high energy output is a blackbody slit cavity; however,

this source is impractical since it requires high temperatures.” There are several common

types of infrared light sources that come close to matching the ideal characteristics of the

36
blackbody radiator in the infrared region. These include the Nersnst Glower and the

Globar.” The Nemst Glower consists primarily of rare earth oxides such as zirconium.
yttrium, and thorium oxide which are fashioned into rods about 20 mm long and 1 mm in
diameter. They are normally operated at 1900 °C since they are nonconducting at room
temperature and must be heated to reach a conducting state.” The Globar is a silicon
carbide rod, 50.8 mm long and 4.8 mm in diameter that operates at about 1200 ”C.” The
disadvantage of the Globar is that it requires a water-cooled jacket to protect the silvered
electrical contacts. In comparison, the Globar has similar radiator properties as the
Nemst Glower, although the Nemst Glower gives better preformance in the region below

15 um.” Unless high performance in this region is specifically needed, the Globar is

recommended due to its advantage of being more rugged.

A recently developed source of infrared light is a laser pumped nonlinear optical
parametric device, such as an optical parametric oscillator (OPO).” Optical parametric
devices were identified some time ago as promising sources of tunable infrared radiation.
However, until recently they proved to be less than ideal due to difficulties such as low
optical damage thresholds, spectral bandwidth limitations, and complicated wavelength
control methods.“ Lately, significant improvements in the production of nonlinear
optical materials and the availability of stable injection-seeded single-mode or ultrashort-
pulse high-peak-power solid-state laser pump sources has resulted in new developments
in optical parametric devices.” These devices require a coherent diffraction-limited pump
source, such as an injection—seeded Nd:YAG laser, to deliver the necessary beam quality,
peak power, and pulse energy, as well as high pulse-repetition rate, and narrow

bandwidths. The NszAG lasers operate in the nanosecond and picosecond regimes. For

 

37
ultrafast spectroscopic applications, femtosecond-pulse lasers such as TizAle, /dye lasers

have proven useful.” Parametric devices using these pump sources and newly developed
nonlinear optical materials such as AgGaS2 and AgGaSe2 can now generate infrared
radiation throughout the mid-1R (fingerprint) region.” Previously identified materials
such as LiNbO3 and LiIO3 have transmission characteristics that limit them to

wavelengths shorter than 4 pm and 5 pm, respectively. Now, essentially any wavelength

from the UV to the mid-1R can be produced by selecting different crystals (nonlinear
optical materials) and pump sources.” Use of this type of light source is currently limited
to high end systems because of its relatively high cost compared to the Glower or Globar
sources.

Infrared detectors consist of two general types, thermal and photon detectors.
Thermal detectors measure a change in a physical property of a material as it is heated;
whereas, photon detectors use changes in the electrical properties of a semiconductor
caused by incoming photons, to detect IR radiation.” Thermal detectors tend to be
nonselective detectors, in that they have a response directly proportional to the incident
energy and largely independent of the wavelength.” Photon detectors, on the other hand,
are much more selective detectors since they have a response curve that is strongly
dependent on the wavelength of the incident radiation. Traditionally nonselective thermal
detectors have been the better choice for spectroscopy, because of their relative
wavelength independence” and versatility.” However, these types of detectors tend to
have slow response times and low relative sensitivity.” With improvements in
semiconductor technology and the increasing need for fast response times and better

sensitivity, photon detectors are gaining favor. Thermal detectors include thermocouples,

38
therrnopiles, thermistors, bolometers, and pyroelectric detectors. Some examples of

photon detectors include photoconductive, and photovoltaic cells.”

Thermocouples and bolometers are two commonly used detectors. Bolometers
have the advantages over thermocouples of faster response times and greater sensitivities.
A bolometer uses the temperature dependence of a material’s electrical resistance to
meaure the intensity of incident radiation. Bolometer’s are often constructed of
semiconductors. Although, superconducting materials have been used for improved
sensitivities and faster response times.”

The fast response requirements of the rapid-scanning interferometers in FTIR
instruments, and the convience of room temperature operation is often meet by the use of
a pyroelectric detector.” This detector’s response is based on the temperature sensitivity
of the residual electrical polarization that can be induced in pyroelectric crystals, such as
triglycine sulfate.” When this type of material is polarized by an electrical field a
residual polarization is retained by the material after the field is removed. The voltage
across this crystal is then temperature dependent and is measured by electrodes placed on
the crystal faces. This device is essentially independent of wavelength from the near
infrared through the far infrared.”

Photoconductive and photovoltaic detectors have the advantages of rapid response
times and high sensitivity compared to thermal detectors. They have the disadvantage of
being limited in their detection range, providing poor preformance in the far-IR region.”
In addition, they often require liquid nitrogen or helium cooling. Both of these detectors
depend on the quantum interaction between photons from the incident light and the

semiconductor material in the detector.”

39
Overall for the slower dispersive 1R instruments, the cheaper thermal detectors are

generally the best choice. For high speed applications and especially rapid FTIR work, a
photon detector is essential.” More detailed information on the various types of detectors
available for IR spectroscopy is presented by Colthup et. al in reference” and Cuirczak in

reference.”
II. 4.4. Sampling Techniques for Infrared Spectroscopy

The physical state (i.e. solid, optically dense) of most polymers makes careful
sample preparation imperative since it is generally necessary for the IR beam to be
transmitted through the sample. Therefore, at least some sample preparation is necessary
before a spectrum can be obtained. In addition, the quality of the spectrum is often
dependent on the care taken in preparing the sample. Poorly prepared samples that are
too thick, nonuniform; or contain impurities, can result in poor or misinterpreted data.
There are a number of methods for polymer sample preparation available depending on
the physical state of the sample. For solid polymers, 1R samples may be prepared by
dissolution, dispersion into KBr pellets or mulls, thin film casting, microtoming.

microsampling, and attenuated total reflectance (ATR).”'”'”

For liquid polymer samples,
liquid films can be examined by placing several drops between two salt plates.”
The method of dissolution requires the identification of a solvent that dissolves

the polymer under investigation but does not absorb in the IR region of interest. For this

reason this method is seldom used despite the advantage of excellent reproducibility.” A

40
number of authors, such as D. O. Hummel and J. E. Stanfield have summarized the IR

absorption spectra of commonly used solvents.

For dispersive 1R systems, the spectrum of the solvent is removed from the
polymer spectrum by placing the solvent in the reference beam. It is generally
advantageous to prepare solutions with as high a concentration as possible. For polymer
systems, this method is often most valuable for cases where only a narrow spectral region
is of interest for which a suitable solvent exists.” With FTIR systems there is no problem
removing the solvent absorption bands, unless water is the solvent. The strength and
breadth of the IR absorption spectrum of water in aqueous systems, such as hydrogels,
can easily surpass the dynamic range limitation of the instrument. In order to subtract the
water spectrum, the FTIR spectrum must be examined so that the dynamic range of the
detector is not exceeded by the presence of excessive amounts of water.”

One of the more commonly used sample preparation techniques for the
investigation of polymer systems is to make film samples. This method has the
advantage of providing a means for studying polymers in their solid state. The
dimensional requirements of the film are such that it is large enough to cover the entire
cross section of the light beam, thick enough to provide sufficiently intense bands for
groups that are present in low concentrations, but thin enough to allow beam penetration.
The film thickness may also need to be reduced to prevent saturation of the detector in IR
regions that absorb more strongly than others. Typical dimensions are on the order of 15
mm x 5 mm with thicknesses in the range of 0.001 to 1 mm.” Some of the techniques for
forming polymer films include solvent casting, melt casting, and hot pressing. Solvent

casting is useful for polymers that can be readily dissolved in a volatile solvent. While

 

41
melt casting can be used for therrnoplastics, some samples require hot pressing in which a

hydraulic press is used to provide thinner sample films.”

Another IR sample preparation method often used for polymers is the pressed disk
technique. This technique involves grinding the polymer sample into a fine powder, and
then mixing a small amount with an IR transparent matrix, such as powdered potassium
bromide. This mixture is then molded at high pressures in a special die. Often the die is
evacuated” and pressures range up to 1,000 N/mmz.” A disadvantage of this technique is
that many of the matrix materials (salts) absorb water and exhibit water absorbance peaks
near 3450 cm’l and 1635 cm".” A similar technique which is more useful for water-
sensitive samples, is the mull method. In this method the polymer is ground into a
powder and dispersed into a liquid phase, commonly mineral oil (Nujol),”
hexachlorobutadiene, or perfluorocarbon.” This paste is squeezed between two 1R
transparent plates. The major disadvantage of this method is the limitation of spectral
regions that can be observed due to absorbance bands of the dispersion material.” A table
of transmission ranges and refractive indices for a number of common matrix materials is
presented by Siesler in reference.”

Two final sample preparation methods are microtoming and microsampling.
Microtoming involves the use of a sharp glass or diamond knife to cut extremely thin
samples with thicknesses down to the nanometer scale. The technique is rather tedious
and requires considerable practice to master, but may be useful for highly crosslinked
polymers for which a solvent can’t be found. In order be microtomed, a sample it must

be suitably hard so as not to deform greatly under the knife pressure. Softer samples

often must be microtomed under cyrogenic conditions. Microsampling is a technique that

42

is used to obtain IR spectra from very small specimens, such as a microtomed sample.”
This technique requires the attachment of an optical device that reduces the image of the
monochromator slit at the sample position.” This type of system can introduce several
sources of error that need to be considered. As an example, diffraction phenomena or
imperfections in the optical system can cause false radiation that will result in smaller
intensities of the absorption bands.’2

Finally, for polymer samples for which satisfactory transmission spectra can not
be obtained by any of the above techniques, it may be possible to use attenuated total

67.68

refection (ATR) to acquire an IR spectra. ATR is based on the phenomenon of total
internal reflectance between two materials of different refractive indices. Experimentally,
the sample under investigation with a refractive index of n2 is placed in direct contact
with the reflecting surface of a prism. The prism has a refractive index of n, that is larger
than n,. Total internal reflectance occurs at an angle above the critical angle of incidence,

52.59

01., of the light beam. This angle is defined by the relation,
sin one = nz/nI (nl > n,) Eq. (7)
The penetration depth of the light beam into the sample is generally on the order of
several micrometers. This provides the absorption path without requiring transmission of
the radiation through the entire thickness of the sample. The spectrum obtained is very
similar to the transmission spectrum. The only difference is that the absorption bands
observed with ATR at longer wavelengths have higher intensities compared to
transmission spectra.”
A second type of reflection spectroscopy is reflection-absorption (RA)

spectroscopy. This technique makes use of external reflections. A thin film of the

 

 

ix
.
h
‘4

43
sample is placed on a reflecting metal surface and irradiated with a light beam having an

angle of incidence usually between 70“ and 89". This technique has been used for the
detection of thin polymer films on metal surfaces, and investigations of the adhesion

between polymers and metals.”

11.4.5. IR Application to Polymers

Infrared spectroscopy, especially FTIR, is very useful in both the qualitative and
quantitative characterization of the functional groups present in polymers. Additionally,
infrared spectroscopy can be used for the detection and identification of additives, the
study of oxidation and degradation reactions, and the investigation of molecular
interactions, configurations, and hydrogen bonding.” Hydrogen bonding is of particular
importance for hydrogels such as poly(vinyl alcohol) (PVA) and poly(2-hydroxyethyl
methacrylate) (PI-IEMA) which contain hydroxyl groups. In general, the presence of
hydrogen bonding results in a frequency shift and intensity increase in the 1R peak
corresponding to the proton donating group.” IR spectra have also found use in the study
of polymer deformation, stress relaxation, fracture, and fatigue.” Finally, IR techniques
can be used to investigate kinetics,” network formation and identification,7| and
diffusion.””

One of the major reasons for the continuing interest in hydrogels is their potential
application as biocompatible materials in medical implants. Such applications include

membranes, structural materials, and matrices for the immobilization or diffusion

controlled release of bioactive agents.” FTIR is useful for studying the interactions

44

between polymers and blood due to the development of digitized spectra which allow
manipulations (such as spectral subtraction) to be readily performed. This feature is of
great importance in the study of many biological and hydrogel systems since the large
water absorption bands, which can obscure the spectra of interest may be subtracted

52.70.73
out 0

For example, Gendreau et. al have used FTIR/ATR to study the adsorption of
blood plasma proteins onto various surfaces. The sensitivity and absorbance subtraction
ability of FTIR spectroscopy was crucial in allowing them to construct a time-resolved
picture of the proteins’ rapid adsorbance including changes that occurred in a small
percentage of the proteins. These investigators further used their FTIR/ATR technique in
kinetic studies to determine total as well as relative protein adsorbance by plotting
individual 1R band intensities versus time.”

Infrared spectroscopy has been successfully used in the analysis of the
polymerization process. With the use of high speed FT IR spectrometers, kinetic studies
can often be performed in situ to determine reaction order and activation energy. IR
spectra can also be used to identify the structural changes and the extent of cross—linking
that occur in a polymerization reaction.” The extent of cross-linking can often be
followed from the decrease in the double bond stretching mode at 1640 cm’l if a suitable
internal intensity standard can be found.” Hasirci used IR spectra to identify the major
source of cross-links in hydrogels synthesized from 2-vinylpyridine and divinyl-
benzene/ethylbenzene monomers initiated with gamma radiation. He found that the
major source of cross-links was the bifunctional divinylbenzene monomer." Earhart et.

al used FTIR spectroscopy in their investigation of the effect of poly(vinyl alcohol) on

the mechanism and kinetics of the copolymerization of vinyl acetate and butyl acrylate.

45
These investigators used the FTIR spectra to monitor the increase in the concentration of

acetate moieties on the polymer chains resulting from the grafting of vinyl acetate onto
poly(vinyl alcohol).74 Finally, Gulari et. al used FTIR spectroscopy to follow, in real
time, the conversion of the bulk polymerization of both styrene and methyl methacrylate.
By monitoring the integrated peak area for the IR absorbance of the carbon-carbon double
bond (1643 cm"), they were able measure conversion with less error than methods
involving the ring breathing mode of the aromatic ring at 1000 cm"| as an internal
reference.”

IR spectra often play an important role in the synthesis of hydrogels by providing
a convenient means of characterizing the final products. For example, Allcock and Kwon
used IR spectroscopy to help characterize the structures of hydrogels formed from
glyceryl phosphazenes. They were specifically interested in the cross-linking and binding
ability of the hydroxyl units present on the glyceryl side groups. IR spectroscopy was
used to help verify the synthesis of isopropylideneglyycerol, and to aid in the
identification of the hydrolytic degradation products of glyceryl polyphosphazenes by
measuring the 1R absorbance of the O-H (3500 - 3200 cm") and P=N (1250 cm")
groups.72 Edwards et. al used IR and Raman spectroscopy to characterize copolymers of
methyl methacrylate with butadiene. They compiled a complete table of the
wavenumbers and vibrational assignments for the 50:50 methyl methacrylate-1,3-
butadiene copolymer.”‘”

Hydrogen bonding can play a significant role in both the reactivity of the

monomer and the structure of the resulting polymer. Several investigators have looked at

the effect of hydrogen bonding between different solvents and acrylic and methacrylic

46

” For example, Saini et al reported that they were able to use IR spectra of

acids.”
copolymers of methacrylarnide and methyl methacrylate polymerized in different

solvents to show that the copolymer composition was solvent dependent.” Other

-8 811.111

researchers found similar effects for the polymerizaton of acrylic,” ' methacrylic, and
itaconic” acids. Bajoras and Makuska used infrared spectra to suggest that hydrogen
bonding between a mixed dimethyl sulphoxide (DMSO) - water solvent and acrylic acid
resulted in a marked decrease in the polymerization activity of the acid. They concluded

that this decrease was the result of a decreased probability for the formation of

tt-complexes caused by an increase in the activity of the acid molecules in the hydogen

bond complexes when in the presence of water.m

Furthermore, the effect of hydrogen bonding on the molecular structure of poly
(vinylpyrrolidone) (PVP) and PI-IEMA, which are hydrogels commonly used for soft
contact lens, was studied using both FTIR and Raman spectroscopy by Monti and
Simoni."2 The investigators used these two techniques to determine how the adsorbed
water interacted with the hydrogels by comparing the spectra of dried and hydrated soft
contact lenses. They found that the water interacts with the hydrophilic groups of the
hydrogel, such as carbonyl groups, by hydrogen bonding. Furthermore, they determined
that the strength of this interaction is dependent on which hydrophilic group is involved,
and that for PVP lenses the water alters the conformation of the hydrophobic groups.”

Changes in the structure of the crystalline regions of poly (tetramethylene
terephthalate) due to an applied stress were studied by Ward and Wilding using FTIR.”
They found that differences in the spectra between the stressed and unstressed state could

be attributed to changes in the trans/gauche isomerism of the polymer chains. Under

 

47
stress the conformation changed from a gauche-trans-gauche form (identified by OH

rocking bands at 917 cm'I and 752 cm") to an all trans conformation (identified by CH
rocking bands at 962 cm'l and 840 cm").” An additional use has been made of infrared
spectroscopy in the study of the stereoregulatity of poly(methyl methacrylate)
(PMMA).”'”'” Dybal et al used the infrared spectra of PMMA to find that in a
stereocomplex of isotactic and syndiotactic PMMA the extended chain form of the

syndiotactic PMMA is preferred.”

11.4.6. Physical Basis and Background for Raman Spectroscopy

As discussed earlier, infrared spectroscopy involves vibrational transitions due to
adsorption of photons between the actual energy levels in the infrared region of the
electromagnetic spectrum. In contrast, Raman spectroscopy involves the inelastic
scattering of photons rather than their direct absorption or emission. When radiation is
incident on a molecule several possible interactions may take place and the incident
radiation may be absorbed, transmitted, refracted, diffracted, or reflected. A fraction of
the light is elastically scattered and, therefore, unchanged in energy (Rayleigh
scattering)” A smaller fraction of the radiation will be scattered inelastically giving rise
to the Raman effect. Both the Rayleigh and Raman scattering are very weak with

intensities of roughly IRayleigh = 10‘150urcc’ and IRaman == 10'8 Isomc,” resulting in the

need for high power, monochromatic laser sources. The intensity of the molecular
scattering is dependent on the frequency of the excitation source to the forth power, as

well as the polarizability of the scattering molecules as shown in Equation 8.”

 

48

a

(a. >,.,,. ‘

Here the subscripts m and 11 represent a transition from the m vibrational level to the n

27n5 4
Imn = FIDO/0 +vmn) E

Eq. (8)

 

 

level. Alpha, (1, is the po component of the polarizability tensor, where p6 assumes nine

possible designations ( i.e. xx, xy, xz, yz, etc.). The net result is a strong dependence of
the Raman intensity on the excitation frequency with higher frequency excitation leading
to a larger Raman scattering cross-section and higher intensity Raman signals.

The inelastic scattering of radiation results in two possible processes. Molecules

in the ground state give Raman scattering with energy h(v0 — v1). These transitions result

in a reduction of the energy and is known as Stokes scattering. The second possibility
results when vibrationally excited state molecules inelastically scatter back to the ground

state, giving an energy h(vo + v1). These transitions result in an increase in energy and

are known as anti-Stokes scattering. The change in energy of the molecule in both Stokes
and anti-Stokes scattering corresponds to the vibrational energy levels of the molecule.
This gives rise to the Raman spectrum.

The ratio of the Stokes and anti-Stokes intensity is governed by the sample

temperature. The ratio can be determined from the Boltzman distribution,

anti - Stokes intensity _ (v + v,,,)4 _(,,,.m,m
Stokes intensity (v - V,,,)4

 

Eq. (9)

when the system is at thermal equilibrium.” At ordinary temperatures most of the
molecules exist in the ground state and therefore Stokes lines have greater intensities than
anti-Stokes lines. It is possible to use the ratio of Stokes and anti-Stokes to determine the

temperature of a sample.

49
While infrared and Raman spectra both involve vibrational and rotational energy

levels, they provide complementary information. This is because the intensity of the
spectral band depends on how effectively the photon energy is transferred to the
molecule. Furthermore, the mechanism for the photon/energy transfer differs in the two
techniques.” In order for a molecular vibration to be Raman active, the vibration must be
accompanied by a change in the polarizability of the molecule. In Raman spectroscopy,

an electromagnetic wave of frequency v interacts with the polarizable electron cloud,

causing it to oscillate. The oscillating electron cloud coupled to the molecular vibrations
gives inelastic scattering in all directions. The vibrational analysis of molecules through
the use of group theory, allows the calculation of the number and activity of the Raman
bands expected in the spectrum of the molecule.

In general, the Raman scattered light from an unpolarized random sample will be
polarized to some extent.“I For a specific scattering geometry, the polarization is
dependent upon the symmetry of the molecular vibration giving rise to the line. In solid
samples the polarization is more complicated. However, the directionality in the crystal
can be used to obtain Raman data. Polarization measurements in conjunction with
dichroic measurements from IR spectroscopy can be a powerful tool for the structural
determination of polymers.”

Near 1R Fourier transform (FT) Raman spectroscopy is designed to eliminate the
fluorescence problem encountered in conventional Raman spectroscopy.” If a Raman
sample is even slightly fluorescent, the fluorescence will normally be much more intense
and obscure the Raman signal. The use of an excitation frequency below the threshold

for the fluorescence process, such as 1064 nm, will remove many problems associated

50
with fluorescence interference. The Raman scattering from 1064 run will occur in the

near-IR region. Conventional FT IR instrumentation with added filters for rejection of the
Rayleigh line can be used with little modification. FT Raman spectroscopy holds a
significant promise for bulk chemical characterization of biomaterials and drug delivery
systems.” Conventional Raman spectroscopy is not widely used due to problems
associated with background fluorescence and photodegradation of molecules sensitive to
visible light. Further advantages of PT Raman spectroscopy include the requirement for
little sample preparation and the rapid analysis time for most materials.”

The advantages of F1 Raman are counterbalanced by the difficulties in the
distinction between inherently weak Raman and strong Rayleigh scattering. 1n the
Fourier transform process, the noise in the intense Rayleigh line is redistributed across
the entire spectrum.” Therefore, FT Raman spectroscopy must have an extremely
effective means of removing stray light at the laser wavelength. Furthermore, since lower
frequency 1R radiation is used as the excitation source, the Raman scattering cross-
section is reduced, and the IR radiation can cause excessive local heating of the sample.

When the laser wavelength used to excite the Raman effect lies under an intense
electronic absorption band of a chromophore, a considerable resonance enhancement of
the Raman signal by a factor of 103 to 10" may occur."7 The enhanced vibrational modes
are generally totally symmetric and are associated only with the electronic chromophore
being excited. Accordingly, the enhancement can increase the Raman signal in certain
cases. However, nonlinear variations in intensities that depend in a complex way on the

proximity of the laser excitation wavelength to the electronic maximum of the sample can

 

 

Eb

51
be seen in certain samples.” Therefore care must be exercised in quantitative analysis to

account for this nonlinearity in peak intensity.

”.4. 7. Hardware Requirements for Raman Spectroscopy

The monochromatic and intense light provided by lasers offers clear advantages
for Raman Spectroscopy. Therefore with the advent of lasers in 1960, lasers quickly
replaced all other sources becoming the standard.” Today there are many different types
of lasers available which range greatly in frequency and power output. Not only can high
power Ar’ and Kr’ lasers be used directly, but they can also be used for pumping dye
lasers and tunable solid state lasers in order to increase the range of the exciting
frequency. Frequency doubling of NszAG lasers at high conversion rates and intense
diode lasers have also increased the availability of Raman sources. This ability to change
excitation wavelength can help to eliminate the problems caused by fluorescence of
samples and low detector sensitivity in regions of the spectrum.

Gratings used in early monochromators for Raman spectroscopy were
mechanically ruled. Modern holographically ruled gratings are now commonly available
providing dramatic improvements in signal-to-noise ratios. Furthermore, the introduction
of concave holographic gratings in place of the conventional plane gratings, has aided in
increasing the throughput since concave gratings require no additional optical elements.
Since stray light rejection and throughput are of critical importance, improvements in

double and triple monochromator design have also increased the signal-to—noise ratios.

52
Finally, higher quality materials and improvements in the manufacture of optical

components have increased the light collection of Raman spectrometers.

The recent advances in detector technology have again changed the future of
many Raman techniques. Photomultiplier tubes and photodiode arrays are now being
replaced in Raman spectrometers by charge couple devices (CCD). CCD detectors are a
class of extremely low noise devices that often consist of millions of pixels.” CCD
detectors can be cooled to liquid nitrogen temperatures giving remarkably low dark
current levels, thereby allowing for long exposure time experiments. Although CCD
technology has had more than 20 years to mature, advances are continually being made in
CCD manufacturing technology and design techniques.” New CCD devices can provide
nearly photo-counting performance in each pixel and simultaneously achieve dynamic
ranges between 10‘ to 1 counts and 106 to 1 counts.90 CCD detectors also offer the
possibility of 2-dimensional imaging which may be used to determine the homogeneity of
samples. CCD detectors can be coated to have high sensitivity in the UV, VIS, x-ray and
IR regions. The high sensitivity of CCDs in the 700-l,000 nm region has reduced the
problems caused by intense visible fluorescence of some samples.

Fiber-based devices for Raman spectroscopy are becoming more important in a
variety of applications.” Fiber optic technology allows the excitation and emitted light to
be routed to the point of interest using a small sampling probe, thereby offering the
advantages of flexibility and remote sampling. Small portable, fiber optic Raman
systems can now be purchased commercially from several sources.

The last major component of a Raman system is the data acquisition system. The

improvements in speed and power of compact computers allows complete analysis

 

53
systems to run on microcomputers. Collection of large numbers of data points, data

manipulation, background removal, spectrum plotting and analysis can now be accom-
plished with smaller and faster computers. For analysis of complex multicomponent
data, new computer systems also include spectral identification programs based on
growing spectral libraries.

Technological advances, such as those described above, allow high quality Raman
data to be obtained for systems and samples that would have been impossible to
investigate only 15 years ago. Limitations of old techniques have been removed and new
applications are being continually developed. Furthermore, as technology increases,
Raman spectroscopy systems are becoming less expensive, thereby opening a wider

variety of potential applications.

11.4.8. Sampling Techniques for Raman Spectroscopy

As previously mentioned, sampling techniques for Raman require little
preparation in many cases. This is an advantage over 1R techniques which require careful
sample preparation. In solution sampling for Raman spectroscopy, the polymer solution
may simply be contained in a glass or quartz capillary tube and the scattered light viewed
at a right angle to the excitation radiation. The laser beam excitation source is narrow,
collimated, and unidirectional, so it can be manipulated in a variety of ways to access
many sample configurations.” As mentioned earlier, water has an extremely weak
Raman spectrum and is an excellent solvent for Raman spectroscopy. Thus, solution

studies on water soluble or swellable polymers are increasing in importance.“2

 

 

e»
n-

p
-
__l~' e
4‘“ I:

54
Solid sampling techniques for Raman spectroscopy depend on the nature of the

sample. Highly scattering and turbid samples are most often studied in a front surface
geometry, while for clear samples right-angle scattering can be used. Powdered samples
can be packed into glass tubes or cavities for a front surface or a transverse sampling
geometry. Fiber samples can be examined by lumination of the fiber at 90° to the fiber
axis and the radiation is collected at 90° to both the fiber axis and the excitation source.
Deformation caused by high temperatures or swelling processes in solid polymeric
materials can be studied using Raman polarization measurements and varying the sample
orientation.

Microsampling in Raman can easily be obtained because the coherent visible light
source can be tightly focused. In micro-Raman, microscopes are coupled to the Raman

spectrometer allowing samples from 2 to 10 pm in size to be examined.” The ability to

examine small spot sizes allows micro-Raman to detennine heterogeneities in solid
polymer samples. Micro-Raman can also be used to study small volume gas and liquid
samples down to the nanoliter range.”

Surface-enhanced Raman scattering (SERS) spectroscopy has been established as
a powerful method for elucidation of the structure of molecules adsorbed on specially
prepared metal surfaces.” The enhancement effect can increase Raman scattering by a
factor of 103 - 10".” Furthermore, adsorption of molecules on the SERS active metal
surface causes fluorescence quenching in some highly fluorescent compounds, thereby
reducing the fluorescence interference. However for this sampling technique, the overall

theoretical approach to successfully explain these phenomena is not yet clear.”

55

[1.4. 9. Application to Polymers

Raman spectroscopy is extremely valuable in many different areas of polymer
characterization.” In addition to the identification of the specific type of polymer, it can
be used to determine functional groups, end-groups, structure, conformation, kinetics of
polymerization and crosslinking, orientation of chains; and to follow changes in structural
parameters as the polymers are exposed to environmental or mechanical stresses.

Raman spectroscopy can be used for quantitative work as well as qualitative
studies. Both added internal standard and reference bands are used for quantitative work.
Ratioing techniques have been used to determine the concentration of unreacted
monomer in polymer systems. Additionally, the band ratio technique has been
successfully employed to determine the composition of copolymers.” The intensity ratios
of the characteristic bands of two homopolymers, which have structural units in the chain
similar to those expected in the copolymer, were used to determine the composition of
each monomer in the copolymer.””

Quantitative micro-Raman spectroscopy has been used to determine impurities in
polymer systems. The ability to study small samples will allow the analysis of impurities
which are contaminating polymer materials.” In situ solid state analysis of excipents or
additive molecules within a polymeric biomaterial have been performed with Raman

spectroscopy.” Davies et. al have studied the role of additive-polymer interaction with

drugs by monitoring shifts in Raman bands indicative of chemical changes at the

56
molecular level.” A major feature of this study was the ability to detect quantitatively the

molecular structures of a drug and polymer by Raman spectroscopy.

Micro-Raman and F1 Raman have both been used to monitor the appearance or
disappearance of specific functional groups within a polymer after exposure to
degradation processes such as hydrolysis or irradiation.” In biomedical polymers, a
significant degree of interest has been focused on the analysis of polymer degradation for
drug release. Raman spectroscopy allows the degradation process in aqueous biological
systems to be studied in situ.” The use of micro-Raman coupled with fiber optics may
allow Raman spectroscopy to be used for in vivo analysis and diagnostics.”

Raman spectroscopy of thin films less than 5 pm thick is extremely difficult

owing to the small scattering volume of the sample.”5

A technique know as waveguide
Raman spectroscopy (WSR) has been developed to overcome this limitation. The
radiation is coupled into the polymer sample using a prism and propagates by total
internal reflection. This technique has been used to study organic films, laminates, and
orientation studies of isotactic poly(methyl methacrylate).76 Furthermore, the partial
orientation of small guest molecules in poly(vinyl alcohol) matrices have been analyzed
by the WSR technique.”

Recently, near-infrared FT Raman spectroscopy applications of waveguide
Raman spectroscopy (WRS) have been demonstrated.” In the study, waveguide Raman
spectroscopy measurements using near-infrared excitation and a Michelson interfero-

meter were combined with fiber optics for examination of submicron cellulose acetate

and poly(vinyl alcohol) films. FT Raman spectra of films containing small molecule

 

 

 

57

chromophores imbedded in polymer matrices were obtained. This technique seems to
have potential for advancement in many areas.

Solution studies of polymers relate the conformation of the polymer chain to other
solution properties. The structure of the polymer chain changes upon dissolution or
swelling and undergoes transformation with changes in the pH, ionic strength, or salt
content of the solution.98 All of these properties can have large effects on hydrogels, and
since water is an excellent Raman solvent, Raman spectroscopy is attractive for studying
changes in polymer chain structure. For example, Koenig et. al have investigated the
conformation of syndiotactic polymethacrylic acid (PMAA) at various degrees of
neutralization in aqueous solutions.” They determined progressive changes in C-COOH
motions in a neutralization range from 0.1 to 0.4‘ fractions.

Studies of polymers in the solid state can yield information on the structure and
conformation of polymer backbones. For polymers such as polymethacrylic acid
(PMAA) in aqueous solutions, the local conformation does not seem to change drastically
from solid state to solution.” Koenig et. al have shown that the Raman spectra of PMAA
in the solid state and in aqueous solution are strikingly similar.” The only frequency
shifts seen were attributed to the difference in the state of hydrogen bonding in the solid
and aqueous solution. The similarity in Raman spectra suggest that little, if any,
detectable conformational changes in the PMAA structure occurs with dissolution.
Koenig has also studied poly(ethylene glycols) (PEG) in aqueous solutions.“ The
Raman spectra suggest that the PEG molecules in aqueous solution have changed from a

helical structure to a new, less ordered but not completely disordered structure.

58
Raman scattering from the solid state can be used to obtain information about the

"” The molecular orientations in

distribution of molecular orientations in a solid polymer.
solid polymers have been shown to be capable of providing quantitative information
about the second and fourth moments of the orientation distribution functions in non-

I(H

crystalline polymers. Raman spectroscopy also offers the possibility of obtaining
information about non-crystalline regions of crystalline polymers. The orientation of
non-crystalline regions of polymers may improve the understanding of properties such as
elastic modulus and shrinkage.

Raman spectra can be used to obtain information on the stereoregularity of
polymers from both the solution and solid state, while preserving the sample integrity.
The stereoregulatity of a polymer greatly effects the morphological and physical

l ()2

properties of the polymer. Tudor et. al have shown that polymer morphology is a
crucial factor in the hydrolysis rate of polymers used for drug release. In this work, FT
Raman spectroscopy was used to indicate the morphology of poly(anhydride) samples for
use in drug release polymers?5 Furthermore, the ordered structure of poly(methyl
methacrylate) has been extensively studied by Raman techniques in order to determine

the stereoregularityf“m4

These studies show the promise of Raman techniques for non-
destructive orientation and stereoregularity studies of polymers in solution and solid state.

Raman spectroscopy can be used to follow the polymerization process and
provide information on the extent of polymerization and structural information on the
end-product.76 The intensity of the C = C stretching vibration of the monomer decreases

as the polymerization proceeds. Ratioing of peak intensities to initial intensities can give

direct information on conversion as a function of time. Gulari et. al have used Raman

 

59
spectrosc0py to monitor the polymerization rate of bulk methyl methacrylate (MMA)75

By following the reduction in the scattering intensity ‘of the carbon double bond, the
percent conversion of MMA as a function of time was monitored with Raman
spectroscopy. Perhaps the most interesting finding in this research was the detection of a
new peak at 1545 cm" present only during rapid polymerization period. It was concluded
that the peak is due to a free radical intermediate formed by the addition of AIBN
(azobis-(isobutyronitrile)) to MMA. This free radical is probably stabilized through
conjugation with either the cyan group or the ester group. The fact that a strong signal
was obtained from a reaction intermediate which is in low concentration and short lived
suggests that there is significant enhancement of the Raman band by resonance.75

The recent advances in Raman techniques have made it possible to explore the
kinetic behavior of faster polymerizations.76 The polymerization reaction can be followed
to completion with Raman techniques. Many techniques lose their sensitivity after the
gel point and are therefore limited in following high conversion reactions. Raman has
been shown to work well for bulk polymerization of fast reacting monomers.

Large scale process monitoring seems to be possible using Raman techniques to
measure a change in a reactant or the concentration of a product. To overcome the
problems of sample geometry, non-invasive fiber optic systems have been built to carry

- 105.")6
OUt remote SCDSlIlg.

The near-infrared Fourier transform Raman technique can be
significantly enhanced by use of fiber optic assemblies.”7 As mentioned previously, near-
infrared FT Raman is limited by sample alignment problems and the generation of

excessive temperatures. However, both problems can be reduced with the easier

positioning of the fiber probe and the lower laser power densities required for meaningful

 

 

60

signal to noise ratios.”7 Due to this lower power requirement near-infrared FT Raman
spectra of systems which contain highly absorbing or fluorescing molecules can now be

easily obtained without large temperature increases.

11.4.10. Raman Cure Monitoring

Raman spectroscopy is attractive for monitoring the curing of polymeric systems
because it can provide real-time measurement of the vibrational structure of the sample.
The consumption of double bonds during polymerization affords the possibility of
monitoring the vinyl stretching frequency using raman scattering as a means of tracking
the polymerization reaction kinetics.

Raman spectroscopy was used to monitor the cure in situ during the cationic
photopolymerization of a divinyl ether photosensitized with anthracene. Since Raman
spectroscopy arises from molecular vibrations, the consumption of double bonds can be
directly measured during polymerization reactions. By monitoring the double bond
concentration, Raman spectroscopy gives a direct method of obtaining the propagation
rate during the reaction. Furthermore, the short intrinsic time scale makes Raman
spectroscopy attractive as a time-resolved technique to monitor cure in these high speed
reactions.

A series of Raman experiments were performed on cationic photopolymerizations
of a divinyl ether initiated with a diaryliodonium salt of hexafluoroantimonate
photosensitized by anthracene. The photosensitizer allowed the initiating wavelength to

be shifted from the deep UV region of the spectrum to the near UV region around 350

61

run.”9 Isothermal Raman experiments were performed for a series of reaction temp-
eratures, and used to determine the overall activation energy of the polymerization
reaction. Since the Raman profiles give a direct measure of conversion versus time, the
rate of reaction was determined during the photopolymerizations. Finally, the Raman
profiles were used in conjunction with previously determined photosensitization rate
constants'"9 to determine the kinetic constant for the propagation reaction. The resulting
profiles for the propagation rate constants as functions of conversion provide important
insight into the nature of cationic photopolymerizations.

As mentioned earlier, the fluorescence monitoring technique developed to follow
cure gave a direct measure of the photosensitization reaction rate and the production of

109 The combination of the fluorescence and Raman techniques

active cationic centers.
allows both the photosensitization rate and the propagation rate to be determined in situ

for these high speed cationic polymerizations. The Raman technique was compared with

the earlier fluorescence method in order to verify both reaction monitoring techniques.

11.5. Cure Monitoring Using Photo-Differential Scanning Calrimetery

In general carbocationic polymerization kinetics are complex and vary from
system to system. Therefore a simple, coherent, kinetic expression of general validity is
not available for cationic polymerizations.” There are several difficulties that contribute
to this deficiency. First, the reactivity of the carbocationic center depends strongly on the
proximity of the counterion and a number of propagating species may be identified,

including ion pairs, solvated ions, or aggregates. The reactivities of the separated or

62
solvated cations are greatly increased over ion pairs and have different chemical

characteristics.l '8 Second, the pseudo steady-state active center concentration assumption
is invalid for cationic polymerizations. In contrast to free radical polymerizations,
combination of active centers does not occur and termination is generally suppressed in
cationic polymerizations. Therefore, in mostwcases the rates of initiation and termination
are not equal, and non-steady state calculations are required.”8 In addition, transfer
reactions are typically important in cationic polymerizations and the polymerization rate
constants are affected by the nature of the counter-ion and solvent. Kinetic analysis of
cationic polymerizations is not impossible, but requires careful individual consideration
for each system under investigation.

Photo-differential scanning calorimetry (PDSC) offers a simple method to
characterize the kinetics of photopolymerization reactions. Since the polymerizations are
highly exothermic, the reaction rate may be measured by monitoring the rate at which
heat is released from the polymerizing sample. Therefore the profiles of reaction heat
versus time provided by PDSC may be used to characterize the reaction kinetics and

9

evaluate polymerization rate constants.ll Several authors have used PDSC to char-

acterize free radical photopolymerizations?“”"22

however, application of the technique to
cationic photopolymerizations of divinyl ethers is more problematic. First, the
polymerizations of divinyl ethers proceed very rapidly, and care must be taken to ensure
that the reaction rate does not exceed the time resolution of the instrument. Since the

DSC response time is on the order of 2-3 seconds,”"23

relatively low light intensities must
be used to increase the reaction time for the divinyl ether photopolymerizations.I6

Second, photopolymerizations of divinyl ethers are highly exothermic, and to maintain

63
isothermal conditions during the reactions, very small sample sizes must be used (0.5 -

1.5 mg). If larger samples sizes are used, the heat evolved during the reaction generally
cannot be removed on the time scale of the reaction.

In this this research, PDSC was used to study the photosensitized cationic
homopolymerization of a model divinyl ether monomer. Divinyl ethers have received
increased attention in recent years because they generally exhibit very low vapor

pressures, relatively low viscosities and low toxicity,'24

but polymerize very rapidly to
form films that exhibit excellent clarity, adhesion, and solvent resistance. There are now
a number of commercially available divinyl ether monomers, however relatively few
detailed kinetic studies have been reported. Understanding the kinetics of these cationic
photopolymerizations is important due to the increasing number of applications for rapid,
solvent-free curing of polymer films. This recent surge in applications of UV—initiated
photopolymerizations has been motivated by at least two factors: environmental concerns
about the production of volatile organic emissions, and the need for high-speed reactions
to enhance production rates (see refs 4,27,124,125).

A series of PDSC experiments were performed on cationic photopolymerizations
of a divinyl ether initiated with a diaryliodonium salt of hexafluoroantimonate
photosensitized by anthracene. The photosensitizer allowed the initiating wavelength to
be shifted from the deep UV region of the spectrum to the near UV region around 350
nm.‘26 Isothermal PDSC experiments were performed for a series of reaction
temperatures and light intensities. The profiles were used in conjunction with previously

determined photosensitization rate constants'26 to determine the kinetic constant for

propagation. The kinetic constant for termination was determined from a series of

 

(,4
unsteady-state “dark-cure” reactions at different temperatures and light exposure times.

The resulting profiles for the propagation and termination rate constants as functions of

conversion provide important insight into the nature of the cationic photopolymerizations.

11.6. Luminescence Temperature Monitoring

11.6.1. Need for Temperature Monitoring

Most of the early work on cationic photopolymerizations focused on the
identification and synthesis of appropriate monomers and initiators, (for reviews see
references 4 and 5) and only recently have detailed studies of the reactions been

""9 For example,109 a novel in situ laser-induced fluorescence technique for

reported.I08
monitoring the initiation reaction in these high-speed polymerizations was reported. Due
to its extremely short intrinsic time scale, the fluorescence technique provided a means to
characterize the kinetics of these polymerizations which proceed too rapidly to be
monitored by traditional methods such as differential scanning calorimetry (DSC). These
studies indicate that although the reactions are extremely rapid (the systems typically
react to completion in a few seconds), most of the conversion occurs in the last 100
milliseconds due to a dramatic increase in the reaction rate. This characteristic reaction
profile was attributed to thermal runaway resulting from the large amount of heat released
by the rapidly polymerizing system.

Luminescence techniques was developed to monitor the sample’s internal tem-

perature in real-time during polymerization. The temperature measurement may be

65
accomplished in a straightforward manner by monitoring emission energy, lifetime, or

intensity. The most simplistic method is to monitor the emission intensity from a
fluorophore, which is typically a strong function of the local temperature. The
temperature sensitive lanthanide probes can be purchased commercially or produced in a
one step synthesis. Several lanthanide candidates hace been evaluated, most having the
general form of Eu(btfa)3 and Eu(hfa)3. Successful completion of these experiments will
provide a detailed kinetic picture of the polymerization reaction, including the effect of
temperature.
Initial temperature profiles have been obtained using thin wire thermocouples embedded
into films (Figure 11.1.). An Omega (WB-FAI B8 [BM 8-Channel) High Speed data
acquisition card installed in an IBM PC was used to acquire temperature data as function
of time. The Omega system is capable of recording up to 2500 temperature measure-
ments per second. Fully formulated systems (monomer and initiator) were coated on
glass substrates at an average of 0.067 mm thickness. A thermocouple wire (0.02 mm
diameter) was placed in the film before the exposure to the UV lamp. The lamp and
temperature acquisition are started at the same time, recording the temperature rise during
the course of the reaction. The Omega system was used to obtain initial temperature
profiles to establish the parameters for the temperature measurement.

The importance of thermal runaway will be determined by monitoring the
temperature in situ during cationic photopolymerizations of DVE-3. Since these studies
are complicated by the rapid rate at which the reaction takes place as well as the thin-film

geometry and the need for an optically-transparent, noninvasive probe, traditional

66

DVE-3 + Ar21+SbF6

70 ~~

I "all. "u." H

Temperature (C)

30 -~ "

 

20 *-

 

0 IO 20 30 40 50 60 '10 80 90 1m 1 lo I 20

Time (sec)

Figure 11.1. Temperature profile for a thin film reaction of DVE-3 on a glass substrate.

67
methods of temperature monitoring are inadequate. For example, thermocouples are

invasive, opaque, and even in their smallest form, have a size-scale comparable to the
film thickness. Raman spectroscopy could theoretically be used to determine temperature
by comparison of the magnitudes of the Stokes and the anti-Stokes lines. However,
Raman is characterized by an inherent signal weakness and is often obscured by sample
fluorescence. Luminescence techniques for temperature monitoring have also been
reported in the literature.”0 For example, temperature-sensitive excimer formation,'” and

10
have been used to

a second technique based upon temperature-induced spectral shiftsl
characterize temperature in fuel sprays, and organic solvents. These techniques are not
applicable to polymerizing systems because they are sensitive to changes in viscosity.

In this research, a luminescence scheme for in situ monitoring of temperature
during cationic photopolymerizations of DVE-3 photosensitized by anthracene was
developed. Due to its short intrinsic time scale, luminescence techniques are attractive
for monitoring the local temperature of these reactions which may proceed to completion
in a few seconds. Furthermore, luminescence techniques provide a non-intrusive method

for monitoring the temperature since trace amounts of the luminescence probe may be

readily incorporated into the reaction mixture. Tris(B-diketone) chelates of europium

fulfill the requirements of a temperature-dependent luminescence probe for our studies.
These probes exhibit a linear response for a wide temperature range, and are insensitive to
changes in viscosity. In addition, the excitation wavelength of these probes is ideally

suited for our studies.

68

11.6.2. Probe Requirements

In addition to exhibiting a temperature-dependent response, the luminescence
temperature probe for our studies must meet specific spectral requirements. Anthracene,
which absorbs between 320 and 390 nm, was used as a photosensitizer to expand the
initiating window to the near-visible region of the spectrum. The anthracene
photosensitizer makes it possible to initiate polymerization in the near-UV or visible
wavelengths as a result of a direct interaction between an excited state of a
photosensitizer and the initiator.109 In our experiment, 351 nm or 364 nm ultraviolet (UV)
light was used to initiate the photopolymerization. Therefore, these wavelengths are the
most convenient for excitation of the temperature probe, since they would eliminate the
need for two laser beams. In addition, the emission wavelengths of the temperature probe
must not overlap with the fluorescence of the anthracene photosensitizer in the 425 nm
region.

A class of compounds which satisfies both spectral requirements is the lanthanide

substituted tris(B-diketone) chelates, which absorb from 280 to 380 nm”2 and emit in the

620 nm range.”3

These compounds exhibit an interesting “antenna” effect which is
responsible for the temperature dependence of the luminescence. The light absorption by
the lanthanide itself is very weak; however, the ligand plays the role of an antenna by
absorbing light and efficiently transferring the excitation energy to the metal ion.”‘ This

non-radiative intramolecular transfer from the excited state of the ligand to the highly

luminescent europium metal center results in a number of interesting luminescent

69

““3"” have characterized the chemical and

properties. A number of authors
photophysical properties of these chelates. These authors found that the luminescence
efficiency varies greatly with the identity of the ligand, degree of coordination,

temperature, and solvent.”2"”

Sager et. al reported that the underlying photophysical
process proceeds by the following four-step mechanism: (1) ground singlet absorption to
the excited singlet state; (2) radiationless intersystem crossing from the excited singlet to
the lowest lying triplet state; (3) transfer of energy to the chelated ion; and (4)

characteristic ionic fluorescent emission.”2

11.7. Modeling of the Formation of Highly Crosslinked Polymers

Very few models for the formation of highly crosslinked polymer networks by
cationic photopolymerizations have been reported in the literature. However, free-radical
crosslinking reactions have received considerable attention, and many of the approaches
used for free-radical reactions could also be appropriate for cationic polymerizations.
Models which have been developed to describe free radical crosslinking reactions may be
divided into three classes'27 1) statistical models, 2) kinetic descriptions, and 3) kinetic
gelation models based upon computer simulated random walks. The statistical models
consider an ensemble of polymer structures built from monomer units according to
probabilistic rules for bond formation. The recursive nature of the polymer structure is
exploited to obtain expressions for statistical averages of the ensemble. While this
approach is useful for lightly-crosslinked polymer networks formed in reactions

containing small concentrations of a multifunctional crosslinking agent, it is of limited

70

value for highly crosslinked homopolymerizations of multifunctional monomers.
Reaction nonidealities such as extensive intramolecular cyclization, heterogeneity, and
topological constraints are difficult to treat using this approach.

The kinetic approach based upon the solution of a set of differential equations
describing the disappearance of reactants is also of limited value for describing highly
crosslinked cationic photopolymerizations. Even the basic step of identifying the reactive
species to consider in the kinetic model is a nontrivial task for cationic polymerizations
because the reactivity of a propagating cationic center depends upon the identity and the
location of its counter ion. For kinetic modeling purposes it is common to assume several
types of propagating centers ranging from completely free (unpaired) ions to covalent
species; however, the choice of the number of such species to include is somewhat
arbitrary. The analysis is further complicated by the difficulty in evaluating the kinetic
constants, and the fact that the quasi-steady-state assumption is not valid for the reactive

centers.”a The largest shortcoming of the kinetic approach is the inability to provide

structural information about the crosslinked polymer network.128
The final type of model for free radical copolymerization/crosslinking reactions
are kinetic gelation models based upon random walk on cubic lattices. The first kinetic

9

gelation model was reported by Manneville and de Seze.” These authors distributed
bifunctional and tetrafunctional monomer units randomly on a cubic lattice, then started
the simulation by allowing a fraction of the sites to simultaneously become active centers.
At each step during the simulation, an active center was chosen at random and allowed to

"choose" one of its first or second nearest neighbors. If the chosen site had an unreacted

functionality, a bond was formed and the active center was transferred, while if the

71
chosen site had an active center, the centers annihilated one another and termination was

achieved. Manneville and de Seze recognized that their model incorporated two drastic
simplificationsz'29 constraining the system to a cubic lattice, and ignoring the possibility
of molecular motion.

The kinetic gelation approach has been refined and modified by several authors.

. 3
In a series of papers,””" 3

Herrrnann and collaborators attempted to include inert solvents,
and monomer mobility. Mobility was included by allowing monomer units and
neighboring solvent molecules to periodically exchange positions. Even with monomer
mobility included in this manner, the authors found that an unreasonably large fraction of
the radicals became trapped by becoming surrounded by inert species. This trapping
resulted in a maximum attainable conversion which was typically less that 60%. Boots

134-136

and collaborators relaxed the assumptions of instantaneous and simultaneous

initiation by running simulations in which initiation occurred at a much slower rate than

termination. Finally, Simon et a1'37

allowed an exponentially decaying rate of initiation.
Although kinetic gelation models can account for reaction heterogeneity, and to
some extent topological constraints, they are currently unable to accurately describe free
radical reactions. While many of the simulations predict correct trends such as increasing
gel point conversion with decreasing crosslinker concentration, they are presently unable
to account for many of the molecular phenomena which can significantly affect the
reaction characteristics. As recognized by Manneville and de Seze in their original
paper,‘29 most of the limitations of the models arise from the choice of a lattice, and the

inability to properly account for molecular motions. For example, constraining the

system to a cubic lattice limits the number of units an active center may react with to six,

72
and significantly distorts the bond angles in the polymer backbone chain. The backbone

bond angle is important because the polymer conformation will determine the
probabilities of cyclization and of trapping. The inability to account for monomer and
polymer mobility precludes the possibility of accurately including diffusion-controlled
reactions. Furthermore, Boots and DotsonI38 illustrated that failure to account for polymer

mobility results in an overestimation of the heterogeneity in the reaction.

11.8. List of References

1. G. Odian, Principles of Polymerization, 2nd Ed., Wiley, New York, NY, 1981. 22.

2. RH. Plesch, “Propagation Rate ConStants in the Cationic Polymerization of
Alkenes,” Cationic Polymerization and Related Processes, E.J. Goethals ed.,
IUPAC, Oxford, UK, 1 (1984).

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

4. J.V. Crivello, “Cationic Polymerization-Iodonium and Sulfonium Salt Photo-
initiators,” Adv. Polym. Sci., 62, 1, (1984).

5. J.V. Crivello and LL. Lee, “Alkoxy-Substituted Diaryliodonium Salt Cationic
Photoinitiators,” J. Polym. Sci., Polym. Chem, 27, 3951 (1989).

6. A. Ledwith, S. Al-Kass, A. Hulme-Lowe, “Mechanisms for Photoinitiation of
Cationic Polymerization,” Cationic Polymerization and Related Processes, E.J.
Goethals, Ed., IUPAC, Oxford, UK, 275, 1984.

7. J.V. Crivello, “Recent Progress in the Design of New Photoinitiators for Cationic
Polymerization,” Cationic Polymerization and Related Processes, EJ. Goethals,

Ed., IUPAC, Oxford, UK, 289, 1984.

8. A. Reiser, Photoreactive Polymers, Wiley, New York, NY, 1989.

10.

11.

12.

l3.

14.

15.

16.

17.

18.

19.

20.

73

S.P.Lapin, “Radition-Induced Cationic Curing of Vinyl-Ether-Functionalized
Urethane Oligomers,” Radiation Curing of Polymeric Materials, C.E. Hoyle and
J.F. Kinstle, Eds., ACS Symposium Series 417, ACS, Washington DC, 363
(1989).

J .V. Crivello and J.H.W. Lam, The Photoinitiated Cationic Polymerization of
Epoxy Resins,” Epoxy Resin Chemistry, R.S. Bauer, Ed., ACS Symposium Series,
114, ACS, Washington DC, 1 ( 1979).

SP. Pappas, “Photoinitiation of Cationic and Concurrent Radical-Cationic
Polymerization. Part V.,” Prog. Org. Coat, 13, 35 (1985).

J .V. Crivello, “Photoinitiated Cationic Polymerization of Coatings,” Organic
Coatings, Science and Technology, Vol. 5, GD. Parfitt and A.V. Patsis, Eds.,
Marcel Dekker, Inc., New York, NY, 35, 1983.

W.R. Watt, “Photosensitized Epoxies as Basis for Light-Curable Coatings,” Epoxy
Resin Chemistry, R.S. Bauer, Ed., ACS Symposium Series 114, ACS, Washington
DC. , 17, 1979.

F. Loshe and H. Zweifel, “Photocrosslinking of Epoxy Resins,” Adv. Polym. Sci.,
78, 61 (1986).

J .V. Crivello and DA. Conlon, “Aromatic Bisvinyl Ethers: A New Class of Highly
Reactive Therrnosetting Monomers,” J. Polym. Sci., Polym. Chem, 21, 1785
(1983).

J.V. Crivello, J .L. Lee, D.A. Conlon, “Photoinitiated Cationic Polymerization with
Multifunctional Vinyl Ether Monomers,” J. Rad. Cur., 10, 6 (1983).

J.V. Crivello and J .L. Lee, “UV Cure of Epoxy-Silicone Monomers,” Radiation
Curing of Polymeric Materials, C.E. Hoyle and J.F. Kinstle, Eds., ACS Symposium
Series 417, ACS, Washington DC, 398 (1989).

J.V. Crivello and IL Lee, “The Synthesis, Charaterization, and Photoinitiated
Cationic Polymerization of Silicon-Containing Epoxy Resins,” J. Polym. Sci.,
Polym. Chem, 28, 479 (1990).

J. Pelgrims, “Cationic UV Polymerization of Cyracure Cycloaliphatic Epoxies,”
Pigm. Resin Technol., 16, 4 (1987).

J.V. Crivello , D.A. Conlon, D.R. Olson, and K.K. Webb, “The Effects of Polyols
as Chain Transfer Agents and Flexibilizers in Photoinitiated Cationic
Polymerization,” J. Rad. Cur., 13, 3 (1986).

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

74

J.V. Crvello, D.A. Conlon, D.R. Olson, and K.K. Webb, “Accelerators in UV
Caioinc Polymerization,” Polym. Paint Colour 1., 178, 696 ( 1988).

A. Ledwith, , Macromol. Chem. Suppl., 3, 348 (1979).

RF. Eckberg and K.D. Riding, “UV Cure of Epoxysiloxanes and Epoxysilicones,”
Radiation Curing of Polymeric Materials, C.E. Hoyle and J .F. Kinstle, Eds., ACS
Symposium Series 417, ACS, Washington DC, 382 (1989).

V.S. Anderson, R.G.W. Norrish, “The Polymerization of Styrene Sensitized by
Molecules in the Triplet State,” Proc. Roy. Soc. (London), 251, 1, (1959).

G. Manivannan, J.P. Fouassier, “Primary Processes in the Photosensitized Polym—
erization of Cationic Monomers,” J. Polym. Sci., Polym. Chem, 29, 1113 (1991).

E.W. Nelson, T.P. Carter, A.B. Scranton, “Photosensitization of Cationic Photo-
polymerizations by Anthracene and its Derivatives,” Proc. of ACS, Poly. Mat. Sci.
Eng., ACS Fall Meeting, Chicago, 69, 363, (1993).

J .G. Kloosterboer, “Network Formation by Chain Crosslinking Photopoly-
merization and its Applications in Electronics,” Adv. Polym. Sci., 84, 1 (1988).

Crivello, J .V. “The Synthesis and Cationic Polymerization of Novel Epoxide
Monomers,” Polymer Eng. and Sci. , 32, 1462 (1992).

Decker, C.; Moussa, K. “Real-Time Monitoring at Ultrafast Curing By UV-
Radiation and Laser Beams,” J. Coatings Tech., 62, 55 (1990).

J.G. Kloosterboer and G.F.C.M. Lijten, “Chain Cross-linking Photopolymerization
of Tetraethyleneglycol Diacrylate,” Crosslinked Polymers, RA. Dickie, S.S.
Labana, and RS. Bauer, Eds., ACS Symposium Series 367, ACS, Washington
DC, 409 (1988).

C. Decker, K. Moussa, “Real-Time Kinetic Study of Laser-Induced Poly-
merization,” Macromolecules, 22, 4455 (1989).

C. E. Hoyle, J. M. Torkleson, Photophysics of Polymers, ACS Symposium Series
358, ACS, Washington DC, (1987), and references therein.

H. Moravrety, Photophysical and Photochemical Tools in Polymer Science, M. A.
Winnick, ed., Reidel, Dordrecht, 1 (1986), and references therein.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

45.

46.

47.

75
F.W. Wang, W.J. Pummer, B.M. Fanconi, and En-Shinn Wu, “Fluorescence
Monitoring of Viscosity and Chemical Changes During Polymerization,” in

Photophysics of Polymers, ACS Symposium Series, Vol. 417, American Chemical
Society, Washington DC, (1987).

F.W. Wang, R.E. Lowry, and W.H. Grant, "Novel Excimer Fluorescence Method
for Monitoring Polymerization: l. Polymerization of Methyl Methacrylate,”
Polymer, 25, 690 (1984).

A. Stroeks, M. Shorhun, A.M. Jamieson, and R. Shimha, Polymer, 29, 467 (1988).

F.W. Wang, R.E. Lowry, B.M. Fanconi, “Novel Fluorescence Method for Cure
Monitoring of Epoxy Resins,” Polymer, 27, 1529 (1986).

SF. Scarlata and J.A. Ors, “Fluorescence Polarization: A Method to Monitor
Polymer Cure,” Polymer Comm, 27, 41 (1986).

C.S.P. Sung, I-J. Chin, W-C.Yu, “A Novel Fluorescence Technique for Monitoring
Cure Reactions in Epoxy Networks,” Macromolecules , 18, 1512 (1985).

C.S.P. Sung, E. Pyun, H-L. Sun, “Characterization of EpoxyCure by UV-Visible
and Fluorescence Spectroscopy: Azochromophoric Labeling Approach,”
Macromolecules, 19, 2922 (1986).

J.C. Song, C.S.P. Sung, “Intrinsic Phosphorescence of Curing Agent for Cure
Characterization of Epoxy Composites,” ACS PMSE Proceedings, 67, 497, (1992).

J .C. Song, C.S.P. Sung, ACS Polymer Preprints, 32-3, 362 (1991).

P. Sundell, S. Jonsson, A. Hult, “Photo-Redox Induced Cationic Polymerization of
Divinyl Ethers,” J. Polym. Sci., Polym Chem, 29, 1525 (1985).

L. R. Gatechair and SP. Pappas, “Cationic Photoinitiation Efficiency,” Proc. Org.
Ctg. and App. Polym. Sci. Division, 46, 707, (1982).

J. L. Dektar, N. P. Hacker, “Photochemistry of Diaryliodonium,” J. Org. Chem, 55,
639 (1990).

R. J. DeVoe, M. R. V. Sahyun, E. Schmidt, N. Serpone, D. K. Sharma, Can. J.
Chem, 66, 319 (1988).

Y. Yagci, I. Lukac, W. Schnabel, “Photosensitized Cationic Polymerization Using
N-ethoxy-2-methylpyridinium hexafluorophosphate,” Polymer, 34, 1130 (1993).

48.

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

76

N. Johnen, W. Schanbel, S. Kobayashi, J.-P. Fouassier, J. Chem. Soc. Faraday
T rans., 88, 1385 (1992).

P.C. Painter, M.M. Coleman, J .L. Koenig, The Theory of Vibrational Spectroscopy
and its Applications to Polymeric Materials, John Wiley & Sons: New York, NY,
1982.

H. Ishida, (Ed.); Fourier Transform Infrared Characterization of Polymers,
Polymer Science and Technology; Plenum Press: New York, NY,Vol. 36, 1987.

“Instrumentation ’94,” Chemical and Engineering News, 72:11, 32, (1994).

H.W. Siesler, K.Holland-Moritz, Infrared and Raman Spectroscopy of Polymers,
Practical Spectroscopy Series, Marcel Dekker, Inc.: New York, NY, Vol. 4, 1980.

“PittCon showcases real-world Raman spectroscopy, NEWSBREAKS,” Laser
Focus World, 30:4, 9 (1994).

F. W. Billmeyer, Textbook of Polymer Science, 3rd ed., John Wiley & Sons, Inc.:
New York, NY, 1984.

A.M. Tudor, C.D. Melia, M.C. Davies, P.J. Hendra, S. Church, A.J.Domb, R.

Langer, “The Application of Fourier Transform Raman Spectroscopy to the
Analysis of Poly(anhydride) Homo- and Co-polymers,” Spectrochim. Acta., 47A,
1335-1343 (1991).

J. G. Grasselli, M. K.Snavely, B. J. Bulkin, Chemical Applications of Raman
Spectroscopy, John Wiley & Sons, Inc.; New York, NY, 1981.

L.H. Sperling, Introduction to Physical Polymer Science, 2nd ed., John Wiley &
Sons, Inc.: New York, NY, 1992.

J.F. Rabek, Experimental Methods in Polymer Chemistry, Phyical Principles and
Applications, John Wiley & Sons, Inc.: New York, NY, 1980.

NB. Colthup, L.H. Daly, S.E. Wiberley, Introduction to Infrared and Raman
Spectroscopy, 2nd ed., Academic Press: New York, NY, 1975.

P. J. Hendra, “Raman Spectroscopy,” In Polymer Spectroscopy; D.O. Hummel, Ed.,
Verlag Chemie GmbH: Weinheim/Bergstr., 112-185, 1974.

J. L. Koenig, Spectroscopy of Polymers; ACS Professional Reference Book; ACS:
Washington, DC, 1992.

62.

63

65.

66.

67

68

69

70

71.

72.

73.

74.

77

J. D. Ingle, S. R. Crouch, Spectrochemical Analysis, Prentice Hall: Englewood
Cliffs, New Jersey, 1988.

BM. Fanconi, “Fourier Transform Infrared Spectroscopy of Polymers - Theory and
Application,” J. of Test. Eval., 12, 33-39 (1984).

U. Simon, F. K. Tittel, “Optical Parametric Device Serve as Sources for IR
Spectroscopy,” Laser Focus World, 30:5, 99 (1994).

E. W. Ciurczak, “Detectors in Infrared Spectroscopy,” Spectroscopy, 8:8, 12
(1993).

S. C. Brown, A. B. Harvey, “Polymers,” In Infrared and Raman Spectroscopy, Part
C; E. G. Brarne, and J. G. Grasselli, Eds., Practical Spectroscopy Series; Marcel
Dekker, Inc.: New York, NY, Vol 1, 873-932, 1977.

. S. E. Polchlopek, “Attenuated Total Reflectance,” In Applied Infrared Spectroscopy;

D. N. Kendall, Eds; Reinhold, New York, NY, 462, 1966.

. R. L. Harris, G. R. Svoboda, “Determination of Alkyd and Monomer-Modified

Alkyd Resins by Attenuated Total Reflectance Infrared Spectrometry,” Anal.
Chem, 34:12, 1655-1657 (1962).

. H. G. Tompkins, “Infrared Reflection - Absorption Spectroscopy,” In Methods of

Surface Analysis; A. W. Czandema, Eds., Elsevier, Amsterdam 447, 1974.

. R.M. Gendreau, S. Winters, R.I. Leininger, D. Fink, GR. Hassler, R.J. Jakobsen,

“Fourier Transform Infrared Spectroscopy of Protein Adsorption from Whole
Blood: Ex Vivo Dog Studies,” Appl. Spectrosc., 35, 353-357 (1981).

V.N. Hasirci, “PVNO-DVB Hydrogels: Synthesis and Characterization,” J. Appl.
Polym. Sci., 27, 33-41 (1982).

HR. Allcock, S. Kwon, “Glyceryl Polyphosphazenes: Synthesis, Properties and
Hydrolysis,” Macromolecules, 21, 1980—1985 (1988).

R.M. Gendreau, R.J. Jakobsen, “Fourier Transform Infrared Techniques for
Studying Complex Biological Systems,” Appl. Spectrosc., 32, 326-328 (1978).

NJ. Earhart, V.L. Dimonie, M.S. El-Aasser, J .W. Vanderhoff, “Infrared Studies on
the Grafting Reactions of Poly(vinyl alcohol),” In Polymer Characterization:
Physical Property Spectroscopic and Chromatographic Methods; C.D. Craver, T.
Provder, Eds., American Chemical Society, Vol. 227, 333-341 (1990).

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

78

E. Gulari, K. McKeigue, K.Y.S. Ng, “Raman and FTIR Spectroscopy of
Polymerization: Bulk Polymerization of Methyl Methacrylate and Styrene,”
Macromolecules, 17, 1822- 1825 ( 1984).

H.G.M. Edwards, A.F. Johnson, I.R. Lewis, “Applications of Raman Spectroscopy
to the Study of Polymers and Polymerization Processes,” J. Raman Spectrosc., 24,
475-483 (1993).

Edwards, H. G. M.; Johnson, A. F.; Lewis I. R.; Wheelwright, S. J.; “Raman and
FTIR Spectroscopic Sudies of Copolymers of Methylmethacrylate with Butadiene,”
Spectrochimica Acta, 1993, 49A, 457- 464.

G. Sain, A. Leoni, F. Simone, “Solvent Effects in Radical Copolymerization III.
Methacrylamide,” Die Makromolekulare Chemie, 147, 213-218 (1971).

A.Chapiro, J. Dulieu, “Influence of Solvents on the Molecular Associations and on
the Radication Initiated Polymerization of Acryalic Acid,” European Polymer
Journal, 13, 563-577 (1977).

R. Y. Makushka, G. I. Bayoras, Y. K.Shulskus, A. B. Bolotin, Z. A. Roganova, A.
L. Smolyanskii, “Effect of Complex Formation on Reactivity of Acrylic and
Methacrylic Acids in Radical Polymerizations,” Polymer Science U.S.S.R., 27:3,
634-641 (1985).

G. Bajoras, R.Makuska, “Peculiarities of Radical Homo- and Copolymerizations of
Acrylic, Methacrylic, and Itaconic Acids in Complexing Solutions,” Polymer
Journal, 18:12, 955-965 (1986).

P. Monti, R. Simoni, “The Role of Water in the Molecular Structure and Properties
of Soft Contact Lenses and Surface Interactions,” Journal of Molecular Structure,
269, 243-255 (1992).

I.M. Ward, M.A.Wilding, “Infra-red and Raman Spectra of Poly (m-Methylene
Terephthalate) Polymers,” Polymer, 18, 327-335 ( 1977).

J. Dybal, J. Spevacek, B. Schneider, “Ordered Structures of Syndiotactic
Poly(methyl Methacrylates) Studied by a Combination of Infrared, Raman, and
NMR Spectroscopy,” J. Polym. Sci. : Part B: Polym. Phys, 24, 657-674 (1986).

J. Dybal, J .Stokr, B. Schneider, “Vibrational Spectra and Structure of Stereoregular
Poly(Methyl Methacrylates) and of the Stereocomplex,” Polymer, 24, 971-980
(1983).

 

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

98.

79

J. Dybal, S. Krimm, “Normal-Mode Analysis of Infrard and Raman Spectra of
Crystalline Isotactic Poly(Methyl Methacrylate),” Macromolecules, 23, 1301-1308
(1990).

T.Hirschfeld, B. Chase, “FF-Raman Spectroscopy: Development and Justification,”
Appl. Spectrosc., 40, 133 (1986).

M. C. Davies, J. S. Binns, C. D. Mellia, D. Bourgois, “Fourier Transform Raman
Spectroscopy of Polymeric Biomaterials and Drug Delivery Systems,” Spectro-
chimica Acta, 46A, 277-283 (1990).

S. A. Asher, C. R. Johnson, “Raman Spectroscopy of Coal Liquid shows that
Fluorescence Interference is Minimized with Ultraviolet Excitation,” Science, 225,
311 (1984).

P. P. Suni, “Custom Photodetector arrays Meet Design Challenges,” Laser Focus
World, 30:4, 131 (1994).

E. D. Jungbluth, “Fiber based Devices Light Up CLEO,” Laser Focus World, 30:5,
88 (1994).

W. L. Peticolas, “Applications of Raman Spectroscopy to Macromolecules”
Biochimie, 57, 417 (1975).

J .S. Suh, K.H. Michaelian, “Surface-Enhanced Raman Spectroscopy of Acrylamide
and Polyacrylamide Adsorbed on Silver Colloid Surfaces: Polymerization of
Acrylamide on Silver,” J. of Raman Spectrosc., 18, 409-414 (1987).

Surface Enhanced Raman Scattering, R. Chang, F. Furtak, Eds., Plenum Press:
New York, NY, 1982.

M. L. Howe, K. L. Watters, R. G Greenler, “Investigation of Thin Surface Films
and Adsorbed Molesules Using Laser Raman Spectroscopy,” J. Phys. Chem, 80,
382 (1976).

N. E. Schlotter, J. F. Rabolt, “Measurrnents of Optical Anisotropy of Trapped
Molecules in Oriented Polymer Films by Waveguide Raman Spectroscopy (WRS),”
Applied Spectroscopy, 38:2, 208 (1984).

C. G. Zimba, V. M. Hallmark, S. Turrell, J. D.Swalen, J. F. Rabolt, “Applications
of Fourier Transform Raman Spectroscopy to Studies of Thin Polymer Films,” J.
Phys. Chem, 94, 939-943 (1990).

J. L. Koenig, “Raman Scattering of Synthetic Polymer - A Review,” Applied
Spectroscopy Reviews, 4:2, 233-306 (1971).

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

110.

111.

80

J. L. Koenig, A. C. Angood, J. Semen, J. B. Lando, “Laser-Excited Raman Studies
of the Conformational Transition of Syndiotactic Polymethacrylic Acid in Water,”
J. Am. Chem. Soc., 91, 7250 (1969).

J.L. Koenig, A.C. Angood, “Raman Specta of Poly (Ethylene Glycols) in Solution,”
J. Polym. Sci., 8, 1787-1796 (1970).

D.I. Bower, “Polarised Fluorescence and Raman Spectroscopy,” Structure and
Properties of Oriented Polymers, Polymers Applied Sciences: London, 1975.

G. Odian, Principles of Polymerization; John Wiley & Sons, Inc.: New York, NY,
619, 1991.

A. Neppel, I. S. Bulter, “Raman Spectra of Fully Deuteriated Syndiotactic and
Isotactic Poly(methyl methacrylate),” J. Raman Spect., 15:4, 257-263 (1984).

J. Purvis, D. I. Bower, “A Study of Molecular Orientation in Poly(Methyl
Methacrylate) by means of Laser-Raman Spectroscopy,” Polymer, I, 645-654
(1974). .

K. P. J. Williams, “Remote Sampling Using a Fiber-Optic Probe in Fourier
Transform Raman Spectroscopy,” J. Raman Spectrosc., 21, 147 (1990).

P. J. Hendra, G. Ellis, D. J. Cutler; “Use of Optical Fibers in Raman Spectroscopy,”
J. Raman Spectrosc., 19, 413 (1988).

E. N. Lewis, V. F. Kalasinsky, I. W. Levin, “Near-Infrared Fourier Transform
Raman Spectroscopy Using Fiber-Optic Assemblies,” Anal. Chem, 60, 2658-2661
(1988).

C. Decker, D. Decker, RadTech 94 North American Proceedings, RadTech
International North America, 602, 1994.

E. W. Nelson, T. P. Carter, A. B. Scranton, “Fluorescence Monitoring of Cationic
Photopolymerizations of Divinyl Ethers Photosensitized by Anthracene
Derivatives,” Macromolecules, 27, 1013-1019 (1994).

K. F. Schrum, A. M. Williams, S. A. Haerther, D. Ben-Amotz, “Molecular
Fluorescence Thermometry,” Anal. Chem, 66, 2788-2790 (1994).

A. M. Murray, L. A. Melton, Appl. Opt, 24, 2783-2787 ( 1985).

81

112. W. F. Sager, N. Filipescu, F. A. Serafin, “Substituent Effects on Intramolecular
Energy Transfer. 1. Absorption and Phosphorescence Spectra 0 Rare Earth [3—
Diketone Chelates,” J. of Phys. Chem, 69(4), 1092—1100(1965).

113. F. Halverson, J. S. Brinen, and J. R. Leto, “Luminescence of Europium Hexafluoro-
acetylactetonate,” J. of Chem. Phys., 40 (10), 2790-2792 (1964).

114. B. Alpha, R. Ballardini, V. Balzani, J-M. Lehn, S. Perathoner, N. Sabbatini,
“Antenna Effect in Luminescent Lanthanide Cryptates: A Photophysical Study,”
Photochemistry and Photobiology, 52 (2), 299-306 (1990).

115. M. F. Richardson, W. F. Wagner, D. E. Sands, “Rare-Earth Trishexafluoroacetyl-
acetonates and Related Compounds,”J. Inorg. Nucl. Chem, 30, 1275-1289 ( 1968).

116. S. Sato, M. Wada, “Relations between Intromolecular Energy Transfer Efficiencies
and Triplet State Engergies in Rare Earth B-Diketone Chelates,” Bulletin of the
Chemical Society of Japan, 43, 1955-62 (1970).

117. E. J. Schimitschek, J. A. Trias, “IR and Raman Spectra of Europium(III) B-
Diketoneates,” J. Inorgan. Nucl. Chem, 32, 811-831 (1970).

118. J. P. Kennedy, and E. Marechal, Carbocationic Polymerization, John Wiley & Son,
New York, 1982.

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

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

121. K. S. Anseth, C. M. Wang, and C. N. Bowman, “Kinetic Evidence of Reaction
Diffusion during the Polymerization of Multi(meth)acrylate Monomer,”
Macromolecules, 27, 650 (1994).

122. K. S. Anseth, C. M. Wang, and C. N. Bowman, “Reaction Behavior and Kinetic
Constants for Photopolymerizations of Multi(meth)acrylates,” Polymer, 35, 3243
(1994).

123. J. G. Kloosterboer, G. F. C. M. Lijten, in Cross-Linked Polymers, R.A. Dickie, S .S.
Labana, and R. S. Bauer, Eds., ACS Symposium Series 367; ACS: Washington DC,
1987.

124. C. G. Roffey, Photopolymerization of Surface Coatings, Wiley, New York, 1981.

125. S. P. Pappas, UV Curing, Science and Technology, Vol. 2, Technology Marketing
Corporation, Norwalk, CT, 1985.

126

127.

128.

129.

130.

131.

132.

133.

134.

135.

136.

137.

138.

82

E. W. Nelson, T. P. Carter, A. B. Scranton, “The Role of the Triplet State in the
Photosensitization of Cationic Polymerizations by Anthracene," J. of Poly. Sci.,
Poly. Chem, 33, 247-256 (1995).

AB. Scranton and NA. Peppas, “A Statistical Model of Free-Radical Copolym-
erization/Crosslinking Reactions,” J. Polym. Sci. Polym. Chem, 28, 39 (1990).

K. Dusek, “Formation-structure Relationships in Polymer Networks,” Brit. Polym.
J., 17, 185 (1985).

P. Manneville and L. de Sez, in "Numerical Methods in the Study of Critical
Phenomena," edited by I. della Dora, J. Demongeot and B. Lacolle, Springer, Berlin
(1981).

H.J. Herrmann, D.P. Landau and D. Stauffer, “New Universality Class for Kinetic
Gelation,” Phys. Rev. Lett.,49, 412 (1982).

H.J. Herrmann, D. Stauffer and DP. Landau, “Computer Simulation of a Model
Irreversible Gelation,” J. Phys. Math. Gen., 16, 1221 (1983).

D. Matthews-Morgan, D.P. Landau and H.J. Herrmann, “Effects of Solvent in a
Kinetic Gelation Model,” Phys. Rev.,Bll, 6328 (1984).

R. Bansil, H.J. Herrmann and D. Stauffer, “Computer Simulation of Kinetics of
Gelation by Addition Polymerization in a Solvent,” Macromolecules, 17, 998
(1984).

J.G. Kloosterboer, G.M.M. van de Hei and H.M.J. Boots, “Inhomogeneity During
the Photopolymerization of Diacrylates: DSC Experiments and Percolation
Theory,” Polym. Commun., 25, 354 (1984).

H.M.J. Boots and RB. Pandey, “Kinetic Gelation Model, Qualitative Percolation
Study of Free Radical Crosslinking Polymerizations,” Polym. Bull, 11, 415 (1984).

H.M.J. Boots, J.G. Kloosterboer and G.M.M. van de Hei, “Inhomogeneity During
Bulk Polymerization of Divinyl Compounds,” Brit. Polym. J.,17, 219 (1985).

GP. Simon, P.E.M. Allen, D]. Bennett, D.R.G. Williams, E.H. Williams, “Nature
of Residual Unsaturation During Cure of Dimethacrylates Examined by CPPEMAS
”C NMR and Simulation Using a Kinetic Gelation Model,” Macromolecules, 22,
3555 (1989).

H.M.J. Boots and NA. Dotson, “The Simulation of Free-radical Crosslinking
Polymerization: The Effect of Diffusion,” Polym. Comm, 29, 346 (1988).

CHAPTER III. OBJECTIVES

Based upon the introduction and background chapters, it is clear that UV-initiated
cationic photopolymerizations of bisvinyl ethers exhibit considerable potential for the
development of high—speed, low-cost, pollution-free coatings and inks. However,
because appropriate monomers and initiators have been developed only recently, the
characteristics of these reactions including the structure and properties of the resulting
highly crosslinked polymer films have received very limited attention. Fundamental
characterization of the photopolymerization reactions and the resulting polymer structures
is imperative if the potential of these reactions is to be realized, and optimum
formulations and curing procedures are to be established.

The broad objective of this research effort is to provide fundamental experimental
and theoretical investigations of UV-initiated cationic photopolymerizations of bisvinyl
ethers. In order to characterize these high speed reactions, new novel techniques for cure
and temperature monitoring will be developed and used to determine reaction rate
profiles and temperature profiles during the cationic photopolymerization. These new
methods and techniques should be general in their ability to monitor cure and temperature
and will be widely applicable in high speed polymerization reactions. Furthermore the
photosensitization reaction between anthracene and diaryliodonium salts will be

investigated to provide further understanding of this process. These studies will

83

 

84

complement the more macroscopic studies currently in the literature, and will provide
answers to lingering questions about the nature of these reactions.

Specific objectives of this project are:

i. to obtain complete, time-resolved profiles of the monomer conversion as a
function of time under a variety of reaction conditions;

ii. to obtain detailed, time-resolved profiles of the temperature of the system as a
function of time during the course of the reactions;

iii. to investigate the effect of the wavelength and intensity of the incident UV-light
on the characteristics of the reactions;

iv. to provide further understanding of the mechanism of photosensitization of the
cationic photopolymerization by anthracene and similar compounds;

v. to provide new methods and techniques for monitoring cure in high speed
reactions, and for obtaining in-situ temperature profiles during reaction.

CHAPTER IV. IN SIT U FLUORESCENCE SPECTROSCOPY:

CURE MONITORING TECHNIQUE

IV .1. Introduction

Cationic photopolymerizations of vinyl ethers were investigated using in situ,
time resolved fluorescence spectroscopy. Due to its short intrinsic timescale,
fluorescence spectroscopy provides a means to characterize these polymerizations which
are too rapid to be monitored by traditional methods. The fluorescence intensity of the
photosensitizer (anthracene or its derivatives) was monitored as a function of time with
spectra collected in intervals as short as 2 milliseconds. An observed reduction in
fluorescence intensity was attributed to consumption of the photosensitizer, providing a
means to monitor the production of active cationic centers. The reaction rate was found
to increased as the initiator or photosensitizer concentrations were increased. Also,
reactions photosensitized by anthracene and 9,10-dimethylanthracene produced the fastest
rates, followed by 9-vinylanthracene, while the polymerization rate for 9,10—diphenyl-
anthracene was considerably slower due to steric hindrance or resonance effects. These
results illustrate the tremendous potential of in situ, time-resolved fluorescence spec-

troscopy for monitoring polymerizations.

85

86

In this chapter, the kinetics of cationic photopolymerizations of vinyl ethers were
investigated using in situ laser-induced fluorescence spectroscopy. The fluorescence
intensity of the photosensitizer was monitored during the reaction to obtain a direct
measure of the rate of initiation, and to provide indirect information on the rate of
propagation. Fluorescence spectroscopy is particularly attractive as an in situ, time-
resolved technique for monitoring these high speed polymerizations because it has an
extremely short intrinsic timescale (typically 10'9 seconds). The response time for the
monitoring technique is determined by the speed of the detection system. In these
studies, a high-speed diode array detection system was employed, and fluorescence
spectra were collected in intervals as short as 2 milliseconds. Therefore, the fluorescence
monitoring method has a considerably faster response time than DSC and real-time IR
methods, and provides ample time-resolution to characterize the cationic photo-
polymerizations. In addition, since the fluorescence method allows a relatively wide
window of wavelengths to be monitored (as opposed to a single wavelength), spectral
changes arising from more than one functionality can be monitored simultaneously.

The fluorescence monitoring scheme reported here is considerably different than
the viscosity-sensitive methods reported in the literature. First, because cationic photo-
polymerizations proceed to completion in a matter of seconds, these studies require
considerably faster sampling methods than the epoxy resin studies discussed in the
background chapter. However, a more fundamental difference between our fluorescence
monitoring scheme and the viscosity-sensitive probe approach is that the fluorescence of
one of the reactants (the photosensitizer), and not that of an inert probe is monitored.

Therefore, this technique provides a more direct measurement of the reaction itself, and

87
does not lose sensitivity at the gel point. In the following chapter, a description of the

fluorescence monitoring technique will be given, along with a discussion of its benefits

and limitations.

IV.2. Experimental

IV.2.]. Materials

In these studies, 3,6,9,12-tetraoxatetradeca-1,13-diene (DVE-3) (GAF Chemicals
Corp.) was used as the monomer. 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)iodoniumhexafluoroantimonate. Initiator
concentrations specified in the remainder of this chapter correspond to the total UV9310C
concentration. The anthracene, 9-vinylanthracene, 9,10-diphenylanthracene and 9,10-
dimethylanthracene photosensitizers were all purchased from Aldrich Chemical Company

and were used as received.

IV.2.2. Absorption and Fluorescence Measurements

Absorption spectra of monomer, initiator and photosensitizers were obtained
using a Hewlett Packard UV-VIS 8452A diode array spectrophotometer. The laser-
induced fluorescence studies were performed in the LASER Laboratory at Michigan State
University. The fluorescence and photosensitization were both excited with a Coherent

Innova 200 Argon ion laser operating at 363.8 nm. A Newport 845HP-01 Digital Shutter

88
system was opened with an electric pulse from the detector controller, ensuring that the

initiation and fluorescence acquisition were started simultaneously. Fifteen rnilliwatts of
unfocused laser radiation (measured with a Scientech 362 Power/Energy meter) in a ~3
mm dia. beam was directed to the quartz capillary tube containing the sample, whose
longitudinal axis was perpendicular to the laser beam direction. The quartz capillary was
approximately 1 inch in length with an inner diameter of 1 mm and an outer diameter of 2
mm. A typical reaction system contained ~1 wt % initiator (UV9310C) and ~10'2 wt % of
photosensitizer in monomer (DVE-3). The fluorescent light was collected at an angle of
90° from the incident beam and 90° from the long axis of the quartz tube (see Figure
IV.1.). The fluorescence signal was collected using a Spex 1877 Triplemate spectrometer
with a subtractive dispersion filter stage and a Spectrograph stage. An EG&G Princeton
Applied Research Model 1421 intensified diode array, cooled to -20 "C to minimize dark
charge levels, was used to detect the signal. The data was analyzed with an EG&G
Princeton Applied Research Model OMA III detector controller. To simulate typical
process conditions for coating, ink, adhesive and electronic applications, all reactions
were initiated at room temperature and were performed without external heating or
cooling.

A fluorescence spectrum spanning 300 nm centered around 450 nm was collected
with 150 groove/mm gratings in both the filter and spectrograph stages. The OMA III
detection system has the capability of collecting this entire spectrum in 16.67 ms.

However, if smaller regions of the spectrum are collected, less time is required per

89

 

 

 

 

 

 

Intensified
diode arra
Z , y
Spectrometer
Electronic
shutter -
I
I ' Sample
Photoexcitation
(laser beam)

Figure IV.1. Experimental Setup for Fluorescence Cure Monitoring.

90
acquisition. Spectral regions from 50 to 100 nm were collected in intervals as short as 2

milliseconds per spectrum.

IV.3.1. Results and Discussion

IV. 3. I . Absorption Experiments

Absorption spectra for the monomer, initiator and photosensitizer are shown in
Figure IV .2. The figure illustrates that the initiator exhibits a maximum at 210 nm, a
second local maximum at 240 nm and rapidly decreasing absorption up to 290 nm. The
monomer (DVE-3) absorption lies completely below 240 nm, leaving a clean spectral
region above 300 nm. Anthracene and its derivatives exhibit strong absorption in the
300—400 nm region. Due to this strong absorption, the initiating wavelength for the
photopolymerization may be shifted to this region by including anthracene or its

derivatives as photosensitizers.

IV. 3.2. Fluorescence Experiments

Figure IV.3. contains a representative series of fluorescence spectra obtained
during the cationic photopolymerization of DVE-3. In this figure, each spectrum
represents the total fluorescence intensity collected during a 17 ms exposure time; other

experiments were performed with intervals as short as 2 ms per scan. This time

91

 

 

4 is

8 3+ %DVE'B’ UV9310C

t:

35 2 ~~ ,.

g 1 ._ Anlhrooenei
o " W .
190 240 290 340 390

Wavelength (nm)

Figure IV.2. Absorption Spectrum of Monomer (DVE-3), Initiator (UV9310C) and
Photosensitizer (Anthracene).

92

 

 

 

 

 

mtensnty
“‘52..
<’

 

 

\ a - 900 ms

gs® ; “s

W“
v“
—.—‘
fi“

 

 

 

v—V—g

52.—ii? . T , 1' r; '5; ',--—-5.:1200ms
350 400 450 500 550
wavelength (nm)

 

 

 

Figure IV.3. Polymerization of DVE-3 monomer (0.5 wt % initiator) produced by
exciting anthracene (2.8 x10” wt %), at 363.8 nm, and monitored by measuring the
intensity of the anthracene fluorescence at time intervals of I 7 ms.

93
dependent plot (with time as the depth axis) consists of twenty spectra for the anthracene

fluorescence excited at 363.8 nm (15 mW laser power). As illustrated in Figure IV.3.,
under the experimental conditions shown, the anthracene fluorescence was found to
decrease slowly for the first second, and then drop dramatically over the next 100 ms.
Finally, after about 1.5 seconds, the fluorescence intensity levels off and maintains a
relatively low steady-state value. Under other experimental conditions, the time required
to attain the final (relatively low) steady-state fluorescence intensity was as short as 0.4
seconds.

The decrease in anthracene fluorescence intensity illustrated in Figure IV.3. may
be attributed to the consumption of anthracene during the photosensitization reaction.
The photosensitization likely proceeds by an electron-transfer mechanism illustrated in
Figure 1V.4.l The mechanism in Figure IV .4. illustrates the important photochemical
and electron transfer steps in the reaction. Some possible proton and hydride transfer
steps which may occur after the cation-radical is formed are of no consequence for the
monitoring scheme, and are omitted from the figure. A more detailed reaction
mechanism is reported in reference 2. According to this mechanism, when the
anthracene is exposed to 363.8 nm light, it absorbs photons and populates the first
excited singlet state. From the first excited singlet state, a fraction of the anthracene
undergoes inter-system crossing to populate the lowest triplet state (which may be
represented as having unpaired electrons in the 9 and 10 positions). As shown in step 1
of Figure 1V.4., the triplet state anthracene may then form an excited state complex with

the initiator. In the exciplex, the anthracene may undergo electron transfer to the

94

1) 00%”)1‘» M) "i“? :1“... M331
2) [:HMWI—‘r‘i + "2"

3) Nzl‘e—>M+Ar-

4) 000+ ”meg‘

sew—4

Figure 1V.4. Photosensitization mechanism based on electron transfer from
triplet state anthracene to initiator.

 

95
initiator forming a radical-cation which is capable of initiating cationic polymerization

(step 2 of Figure 1V.4.). As a result of this set of reactions, the anthracene becomes
attached to the polymer as a chain end, and loses the aromatic nature of the center ringI
(steps 4 and 5 of Figure 1V.4.). This mechanism (including the role of the triplet state)
was recently corroborated by pulsed laser lifetime studies of the singlet and triplet states.2

The above mechanism for photosensitization suggests that the reduction in the
fluorescence intensity illustrated in Figure IV.3. arises from the loss of the aromaticity of
the center anthracene ring during the photosensitization reaction. Therefore, as long as no
other anthracene-consuming reactions are taking place, the anthracene fluorescence
provides an indication of the number of cationic active centers formed in the reaction
system. Each time an anthracene molecule is consumed in the photosensitization
reaction, one active cationic center is formed, therefore the magnitude of the fluorescence
decrease at a given time (the difference between the initial steady-state fluorescence and
the value at the time of interest) provides a measurement of the number of propagating
active centers in the reaction system. In this manner, real-time monitoring of the
anthracene fluorescence provides valuable kinetic information. The reduction in
fluorescence intensity provides a direct measure of the rate in the initiation step of the
reaction as well as an indication of the number of propagating active centers.

A series of experiments were performed to ensure that the fluorescence decrease
could be attributed entirely to the photosensitization reaction, and that no other
anthracene-consuming side reactions were present. Anthracene may participate in a few
photo-induced side reactions such as dimerization and oxidation, therefore studies were

performed to determine the extent to which these reactions contributed to the

96
fluorescence decrease. Experiments were performed for two distinct systems: one

containing only monomer and photosensitizer, and another containing monomer,
photosensitizer, and initiator. The system containing no initiator showed very little
decrease in fluorescence intensity even after a few minutes of irradiation. In contrast, the
system containing all three components exhibited a rapid decrease in fluorescence
intensity in a few seconds (much like that shown in Figure IV.3.). These results indicate
that, at least for the polymerization time-scale, the photo-dimerization and oxidation
reactions do not lead to a decrease in fluorescence intensity and our observed

fluorescence decrease may be attributed entirely to the photosensitization reaction.

IV.3.3. Initiation and Polymerization Kinetics

Based upon the above discussion, it was concluded that fluorescence spectroscopy
could be used for kinetic studies of cationic photopolymerizations by monitoring the
fluorescence intensity of the photosensitizer (anthracene or its derivatives). A series of
fluorescence spectra such as the ones shown in Figure IV.3. were analyzed by measuring
the fluorescence intensity at a selected wavelength (typically 425 nm) as a function of
reaction time. Figures IV .5. - IV.9. contain plots of data collected in this manner. A
series of experiments were performed to investigate the effects of a variety of variables
on the polymerization kinetics, including the effect of the initiator concentration, the
photosensitizer concentration, and the functionalization of anthracene.

Figure IV.5. contains plots of the fluorescence intensity versus time for reaction

systems containing anthracene (~10’2 wt %) and initiator (0.5 wt % or 1 wt %) in DVE-3

97

monomer. The general shape of the curves in Figure IV.5. seems to be characteristic of
these cationic photopolymerizations. Immediately after the shutter is opened, the
fluorescence intensity increases until it attains a relatively constant, quasi-steady—state
value. Subsequently, the fluorescence intensity slowly decreases as the anthracene is
consumed and active centers are formed. At some threshold value of time (~ 750 ms for
the 1 wt % initiator case, and ~ 1250 ms for the 0.5 wt % initiator case) the fluorescence
intensity falls dramatically (losing ~70% of its value) in less than 100 milliseconds. This
seems to indicate that the rate of the photosensitization reaction increases dramatically at
this point, possibly due to elevated temperatures arising from the heat of polymerization.
A calculation of the adiabatic polymerization temperature illustrates that this temperature
increase may be substantial. Based upon estimates of the heat capacity of the monomer
(Cpm ~ 400 J/mol K by the empirical method of Stemling and Brown’) and the heat of

polymerization for vinyl ethers (AHp ~ 60 kJ/mol‘) the adiabatic temperature rise was

calculated to be ~150 °C.

If the observed sudden decrease in fluorescence intensity arises from thermal run-
away of the initiation and/or propagation reactions, then the rate constants must increase
substantially in the adiabatic temperature range. Therefore, a consideration of the
activation energies of the photosensitization and propagation reactions is important. The
quenching of photo-excited molecules is diffusion controlled with an activation energy of
~15 kJ/mol’ for small molecules in organic solutions, while the activation energy of
propagation for vinyl ethers is approximately 30 kJ/mol.6 Using an Arrhenius equation
for a 125 degree temperature rise (a conservative estimate of the temperature increase),

the rate constants of initiation and propagation are found to increase by factors of 6.0 and

98
35, respectively. The overall rate of polymerization, which depends upon both initiation

and propagation, is characterized by an activation energy ER = E1 + Ep, where ER, E1 and
Ep are the activation energies for overall polymerization, initiation, and propagation,
respectively. Based upon this equation, the activation energy for polymerization is 45
kJ/mol, and the polymerization rate increases by a factor of approximately 210 for a
125 °C temperature change.

In the above discussion of the activation energy for overall polymerization,
termination was assumed to be negligible. In a cationic polymerization active centers do
not terminate with one another, and chain transfer normally leads to an active center
which is capable of propagation. Therefore, the only possible termination arises from
combination with the counter ion or reaction with contaminants which give unreactive
cations. The possibility of combination with the hexafluoroantimonate ion is low because
of the size of the anion, justifying the assumption of zero termination. The inclusion of
termination would decrease the value of ER since the activation energy for termination
would be subtracted.

Based upon the above discussion of the reaction energetics, the sudden decrease
in fluorescence intensity exhibited in Figure IV.5. is attributed to a large increase in
temperature which causes a large increase in reaction rate. To investigate whether the
intensity decrease arises in part from temperature-dependent fluorescence, a series of
studies were performed in a thermostated cell. In the adiabatic temperature range, the
fluorescence intensity of the photosensitizer was found to remain essentially constant or
increase slightly. Therefore the fluorescence decrease exhibited in Figure IV.5. cannot

arise from temperature-dependent fluorescence of the photosensitizer, and is attributed to

99
consumption of the anthracene during photosensitization. Therefore, based upon these

fluorescence studies, the number of active centers, and therefore the rate of polymeriza-
tion, may be characterized both by the overall rate of fluorescence decrease, and by the
time for the onset of the rapid drop in fluorescence intensity. In order to verify the actual
temperature rise, in situ monitoring of temperature during these cationic photo-

polymerizations is currently under investigation by laser-induced fluorescence.

IV.3.4. Effect of Initiator Concentration

Figure IV.5. illustrates just how rapidly the reaction occurs when the sample is
irradiated with laser light. A cursory observation would suggest that in both cases, the
liquid monomer sample becomes a solid polymer almost immediately after the shutter is
Opened. The detailed time-resolution provided by the laser experiment illustrates that, as
expected by simple kinetic considerations, the polymerization rate is noticeably higher for
the sample containing the higher concentration of initiator (as indicated by the time for
the rapid decrease in fluorescence). In either case, the reaction is very rapid with most of

the fluorescence decrease occurring in less than 1.25 seconds.

IV. 3. 5. Effect of Photosensitizer Concentration

Figure IV.6. contains plots of fluorescence intensity versus time for systems

containing 1 wt % initiator, and various concentrations of anthracene in monomer. The

100

1m_ -——--—-—0 . 03 8 wt% Anthracene
1 wt% UV9310C

 
 
  
  

-------- <~—----0 . 02 8 wt% Anthracene
0 . 5 wt:% UV9310C

 

 

7‘50 1000 12'50 1500

tirre (ms)

0 250 300

Figure IV.5. Fluorescence intensity at 425 nm for polymerization of DVE-3 monomer

with varying concentrations of initiator (UV9310C) and 10'2 wt % of anthracene as the
photosensitizer.

101

0.200 wt% Anthracene
—6— 0.038 wt% Anthracene
—-— 0.004 wt% Anthracene

 

0»- -

 

 

  

E. m‘ A r A
w .1 l l
c 1
9
E 40D<fl

.1-

K.

:

 

 

Figure IV.6. Fluorescence intensity for polymerization of DVE-3 monomer with varying
concentrations of anthracene photosensitizer and 1 wt % initiator ( UV9310C).

102
figure illustrates that as the anthracene concentration is increased from 0.004 wt % to

0.038 wt %, the time for the onset of the rapid fluorescence decrease changes nearly
threefold from ~2000 msec to ~750 msec. Therefore, the reaction rate is very sensitive to
the anthracene concentration, and increases significantly as the anthracene concentration
is increased an order of magnitude from 0.004 wt % to 0.038 wt %. If the photosensitizer
concentration is increased another order of magnitude (to ~02 wt %) the time required
for the rapid drop in fluorescence decreases further (~400 msec), but at a smaller
incremental rate (by less than a factor of two). Therefore, the reaction rate becomes less
sensitive to the photosensitizer as the anthracene concentration is increased. This trend
may be explained by the fact that the high anthracene concentrations lead to a high

optical density (O.D. ~2 for 100 um sample thickness in the 0.2 wt % anthracene case)

which limits the penetration of the light into the sample. Therefore, many of the
anthracene molecules near the back of the sample are not excited, and cannot produce
active cationic centers. This hypothesis is supported by the fact that in the first few
seconds of the experiment, the polymerization was limited to approximately the front half

of the sample.

I v.3. 6. Effect of Anthracene F unctionalization

Experiments were performed using a variety of anthracene derivatives as
photosensitizers. The purpose of these experiments was to determine the effect of
anthracene functionalization on the rate of the cationic photopolymerizations. In

Figure IV.7., the fluorescence intensity for reaction systems containing 0.004 wt %

103

 

 

9000- 3.3
b "3"
any; s ~33? 0.004 wt% 9-VA
-: @333 1 wt:% UV9310C
. ‘F's’m.
KID-3% ”6&3
; «3,3 . —-— 0.004 wt% Anth.
any, {33 1 wt:% UV9310C
‘1'. ”a,
B‘mfi cO‘i‘égiafi
o— q icfiéil’;
E . . r
o— (I;
3000- 3?.
2111)-
1cu)- RN
0 . r ' r r 1 1 I ' l ' 1 ' j
0 5(1) 1(XX) 15(1) 201) 25(1) (XXI) 3500
t1me(ms)

Figure I V. 7. Fluorescence intensity for polymerization of DVE-3 monomer with
UV9310C and 10'3 wt % of 9-vinylanthracene and anthracene as the photosensitizers.

104
anthracene or 9-vinylanthracene as the photosensitizer is plotted versus time. The figure

illustrates that the reaction is somewhat more rapid for anthracene than for 9-vinylanth-
racene since the rapid decrease in fluorescence occurs in about 2 seconds for anthracene,
and in about 2.3 seconds for 9-viny1anthracene. It is interesting to note that the curve for
9-vinylanthracene exhibits a steeper slope and a larger fluorescence decrease in the early
stages of polymerization. This may arise from its ability to participate in the propagation
of the active center through its vinyl functionality.

Fluorescence studies were also performed using two other anthracene derivatives
as photosensitizers: 9,10—diphenylanthracene and 9,10-dimethylanthracene. The
fluorescence intensity of 9,10-diphenylanthracene has been shown to be relatively
insensitive to changes in viscosity,7 therefore these studies were performed in part to
investigate the possibility that the decrease in fluorescence intensity observed in Figures
IV.5. and IV.7. are due to viscosity effects. In Figure IV.8., the fluorescence intensity for
systems containing 10'2 wt % 9,10—diphenylanthracene or 9,10—dimethylanthracene as
photosensitizers are plotted versus time. The figure illustrates that, as in the previous
experiments with anthracene, the fluorescence intensity of these probes initially exhibits a
gradual decrease with time, then shows a dramatic drop as the reaction proceeds. The
fact that the general shapes of the fluorescence profiles are the same as those for
anthracene and 9-vinylanthracene indicate that viscosity effects are not responsible for the
decrease in fluorescence.

Comparison for Figures IV.5. and IV.8. illustrate that for the 9,10-diphenyl-

anthracene system, the decay time was 2.5 times longer than that of the 10'2 wt %

105

--I— 0.022 wt% 9,10-DPAnth

"3,3,; 1 wt% UV9310C
Fara,
111m ”5&3
. ,3“ ———o-— 0.020 wt% 9,10-DMAnth
15 7 ‘\.

 

i”
-— '1
(D .L
c 7000 ‘1‘
.02. 1
C “i
-- 6000 1,
i
511!) I:
C.

Figure IV.8. Polymerization of DVE-3 monomer with UV9310C and 10" wt % of

9,10-diphenylanthracene and 9, IO-dimethylanthracene as the photosensitizers, monitored
by photosensitizer fluorescence intensity.

106
anthracene system and nearly 2 times longer than that of the 9-vinylanthracene system.

In contrast, the 9,10-dimethylanthraccne decay time was the same as the decay time for
anthracene. These results indicate that the 9,10-diphenylanthracene photosensitizer
reaction is hindered due to steric effects of the phenyl rings or formation of a stable
cation-radical due to electronic stabilization. The methyl groups of 9,10—dimethyl-
anthracene seem to have little effect on the reaction when used as the photosensitizer.
These results are consistent with the electron transfer mechanism for photosensitization

shown in Figure IV .4.

IV.3. 7. Fluorescence Experiments Without Monomer

The photosensitizer fluorescence intensity was also monitored for a system
containing photosensitizer and initiator but no monomer. The reaction system without
the divinyl ether monomer will not include any temperature effects arising from the heat
of polymerization. Since these polymerizations are highly exothermic and very rapid, the
reaction may take place under nearly adiabatic conditions with a significant increase in
temperature. This increase in temperature could in turn increase the reaction rate further.
Figure IV.9. contains a plot of anthracene fluorescence intensity versus time for 0.02 wt
% photosensitizer and 1.2 wt % of initiator in methanol. Comparison of Figure IV.9.
with Figures IV.5.- IV .8. reveals a dramatically different fluorescence profile for the
system without monomer. The system containing only photosensitizer and initiator

exhibits a fluorescence intensity which decays exponentially over a time period of twelve

intensity
i

107

 

 

Figure I V.9.
10" wt % of

O
m-

#—
m-
—L
O
—L
N

Fluorescence intensity at 425 nm for photosensitization reaction of 2.2 x
anthracene as the photosensitizer and 1.2 wt % of diaryliodonium hexa-

fluoroantimonate ( UV9310C) in methanol.

108
seconds, with no sharp decrease at an intermediate time as was observed for the systems

which also contained DVE-3. This suggests that the rapid decrease in fluorescence seen
in Figures IV.5. - IV.8. is due to a temperature increase arising from the heat of
polymerization. This hypothesis is supported by the fact that the time scale of the
photosensitization reaction in methanol is seconds as compared to milliseconds for the
divinyl ether reaction systems. The time constant for the best-fit exponential decay was

found to be 4.7 seconds for the curve in Figure IV.9.
I V. 3. 8. Initial Increase in Fluorescence Intensity

During the initial stage of the fluorescence monitoring experiments, an increase in
the signal was observed for all reactions. However, the photosensitization reaction
mechanism should lead to a monotonic decrease in the anthracene concentration as active
centers are generated. Therefore, a set of experiments were performed to determine the
cause of this initial increase in the fluorescence intensity.

Figure IV.10. shows a typical fluorescence monitoring curve for a cationic
photopolymerization of DVE-3. As can be seen in the figure, the signal initially increases
dramatically during the first 500 ms. The figure also shows a fluorescence profile where
the sample in moved to new unreacted section of the sample. If the build-up was an
inherent aspect of the photosensitization reaction, the increase should repeat for the new
location. However, the new location does not show a similar increase, indicating that the

increase in not caused by the photosensitization reaction. Since the new location started to

fluorescence intensity

2500

2000

1500

1000

500

 

 

109

l 1 l n l 1 I 1 l 1 J

 

0.0

0.5 1.0 1.5 2.0 2.5 3.0
time (s)

Figure IV.10. Fluorescence monitoring of DVE-3 with 2.4 x [0'2 wt % anthracene and
1 wt % initiator. The sample was moved during the reaction to a new unreacted section.

110
react only after the sample was moved, the third curve in Figure IV. 10. is a repositioning

of the new location profile to time zero. The repositioned curve and the typical curve
illustrate very well the difference in the build-up at the start of the two runs.

Further evidence that the build-up is not an inherent feature of the reaction is
shown in Figure IV.11. In this experiment, the photosensitization reaction was started,
but the fluorescence emission was block between the sample and the detector. If the
initial build-up was caused by the reaction, the increase should be reduced since the
reaction had been underway for approximately 250 ms when the detector was started.
However, the figure shows that the increase was unchanged even through the reaction
was started well before the detector was started. This indicates that the increase is related
to the detector and not the reaction system.

Figure IV.12. shows three different curves in which the initiation shutter and the
detector were started at different times relative to each other. First the initiation shutter
was delayed 1 second relative to the start of the detector acquisition. If the detector was
causing the build-up, starting the detector before the reaction may allow the detector to
“warm-up” and therefore no build-up would be seen at the start of the reaction. However,
Figure IV .12. show that the build-up is still seen for this curve. This indicates that a
“warm-up” period is not required by the detector. Second, the initiation reaction was
started and the fluorescence emission was allowed to reach the detector. However, the
detector acquisition was not started for one second after the initiation. In this case, no
initial build-up is seen, indicating that the build-up in the detector is a light related
phenomena. The third curve in Figure IV.12. is typical experiment where the initiation

shutter was opened simultaneously with the start of the detector acquisition.

111

 

 

 

2500 -
" Path Blocked after Sample

2000 -
>
:1:
a) .
C
(D
g 1500 -
a) l
o .1
c
8 1000 ‘
(I)
(D
h
o I
2
”" 500 4

4.
I ' 1 . m ' I ' r ' I ' f
0.0 0.5 1.0 1 .5 2.0 2.5 3.0

time (s)

Figure I V.11. Initiation Started at time equal zero, but the fluorescence emission was
not allowed to reach the detector for the first 250 ms of the reaction.

112

  

2500 - Delay on Shutter
Delay on Detector A mfiawA‘AA‘fi
' 3 _. Cf? Afifiéwbfi x
if " 3133, 93333394018}, A”
2000 _ “Xxx with» - 1 XX

l

E

fluorescence intensity

"5?

 

 

 

time (s)

Figure I V. 12. Different time delays of the initiation shutter and the detector to determine
the cause of the initial build-up.

113
As detailed in the section above, this build-up in the anthracene fluorescence

intensity was determined to be an instrument artifact and not inherent to the photo-
sensitization reaction. The initial detector response seems to be related to the light
reaching the detector surface. One possible explanation for this build-up is an increase in
the intensifier efficiency during the start of the reaction. The intensifier operates in a
similar fashion to a photomultiplier tube. When a photon strikes the intensifier surface a
number of the photons are released which then cascades toward the detector surface. This
process requires light to initiate and may take a finite length of time to reach its full value
of the enhancement. However, the important information determined using this
technique is not affected by the initial increase. Finally, this build-up was not seen with

the new CCD detector that was purchased and implemented.

1V.4. Conclusions

This chapter demonstrates the utility of in—situ laser-induced fluorescence
spectroscopy for monitoring cationic photopolymerizations of vinyl ethers photo-
sensitized by anthracene and its derivatives. Due to its extremely short intrinsic
timescale, fluorescence spectroscopy provides a means to characterize the kinetics of
these polymerizations which are too rapid to be monitored by traditional methods such as
DSC. In these studies, the fluorescence intensity of the photosensitizer was monitored as
a function of time with spectra collected in intervals as short as 2 milliseconds. The
profile of the fluorescence intensity versus reaction time invariably exhibited the same

general shape with an initial increase to a quasi-steady-state value, followed by a gradual

114

decrease over several hundred milliseconds, followed in turn by a dramatic reduction in a
period of less than 100 milliseconds to a final, relatively constant value.

The observed reduction in fluorescence intensity is attributed to the consumption
of the anthracene in the photosensitization process, with the associated production of
active cationic centers. Experiments were performed to demonstrate that anthracene-
consuming side reactions such as photo-induced dimerization did not contribute to the
reduction in fluorescence. Therefore, it was concluded that the magnitude of the
fluorescence decrease provides a direct measure of the rate of the initiation step, as well
as an indication of the number of propagating active centers. The sharp decrease in
fluorescence was attributed to large temperature increases arising from the heat of
polymerization in these nearly adiabatic reaCtions (experiments were performed without
external heating or cooling to simulate typical process conditions in coating, adhesive,
and electronic applications). This conclusion was corroborated by experiments
performed with initiator and photosensitizer but in the absence of monomer. In these
experiments, which would have no effects from the heat of polymerization, the
fluorescence intensity exhibited no sharp reduction in intensity, but rather decreased
exponentially over a period of twelve seconds.

The rate of the cationic photopolymerizations were characterized in terms of the
overall rate of the fluorescence decrease, and the time for the onset of the rapid reduction
in fluorescence intensity. Experiments were performed to determine the effects of the
initiator concentration, the photosensitizer concentration, and anthracene functional-
ization on the polymerization rate. As expected, the reaction rate was found to increase

as the initiator concentration or the photosensitizer concentration was increased. A

115
practical concentration limit for the photosensitizer of approximately 0.2 wt % for 500

um films was established due to high optical densities which limit the penetration of light

into the sample. Reactions photosensitized by anthracene and 9,10-dimethylanthracene
produced the fastest rates, while the slowest rates were observed for 9,10—diphenyl-
anthracene, perhaps due to steric hindrance or electronic interference caused by the
phenyl rings. The 9,10-diphenylanthracene studies also illustrated that the fluorescence
intensity decrease was insensitive of viscosity changes.

The fluorescence monitoring technique described here offers a non-intrusive, time
resolved method for monitoring cationic photopolymerizations. Based upon this
technique, our studies have provided previously inaccessible information about the rate of
the cationic photopolymerizations and the effects of various reaction conditions.
However, the technique has limitations. For example, although these studies based upon
the magnitude of the photosensitizer fluorescence provide direct information about the
rate of generation of active centers, they provide little information about the rate at which
the active centers propagate once they are formed. Also, although the response time is
considerably faster than the current alternatives, it is limited by the speed of the detection
system (2 ms in our case). Finally, although the studies reported here illustrate the
tremendous potential of in situ, laser-induced fluorescence spectroscopy for monitoring
polymerizations, they provide only one example of a much more general technique.
Many other fluorescence monitoring schemes which provide complementary information
could be developed. For example, the use of laser-induced fluorescence for the in situ
monitoring of temperature during these cationic photopolymerizations will be used to

investigate this polymerization reactions, and strategies for using reactive probes to

116

monitor the propagation reaction are under development. Clearly, there is considerable
potential for novel applications of time-resolved fluorescence spectroscopy for

monitoring polymerizations.

IV.5. List of References

1. E.W. Nelson, T.P. Carter, and AB. Scranton, “Photosensitization of Cationic
Photopolymerizations by Anthracene and its Derivatives,” Proceedings of American
Chemical Society, Poly. Mat. Sci. Eng., ACS Fall Meeting, Chicago, 69, 363 (1993).

2. E.W. Nelson, T.P. Carter, A.B. Scranton, “ The Role of the Triplet State in the
Photosensitization of Cationic Polymerization by Anthracene,” J. of Polym. Sci. Part
A: Polym. Chem. 33, 247-256 (1995).

3. A. Bondi, “Estimation of The Heat Capacity of Liquids,” I & E C Fundamentals, 5,
442 (1966).

4. S. Jonsson, P-E. Sundell, A. Hult, “Photoredox Induced Cationic Polymerization of
Divinyl Ethers,” Radtech 90-North America, Radiation Curing Conference and
Exposition Proceedings, 417 (March 25-29,1990).

5. REM. Allen, C.R. Patrick, Kinetics and Mechanisms of Polymerization Reactions,
John Wiley and Sons, New York, 79, 1974.

6. Encyclopedia of Polymer Science and Engineering, Second Edition, 1.1. Kroschwitz,
Ed., 17, 453, 1989.

7. F.W. Wang, W.J. Pummer, B.M. Fanconi, En-Shinn Wu, Photophysics of Polymers,
C.E. Hoyle, J.M. Torkelson, Eds., American Chemical Society, Washington DC,
(1987).

117

CHAPTER V. PHOTOSENSITIZATION MECHANISM:

EXPERIMENTS AND MODELS

V.1. Introduction

In this chapter, the photosensitization mechanism for cationic polymerizations
initiated by diaryliodonium salts photosensitized by anthracene were investigated using
fluorescence and phosphorescence spectroscopy. In situ photosensitizer fluorescence
measurements confirmed that the photosensitization reaction proceeds by an electron
transfer process. Transient phosphorescence studies demonstrated that electron transfer
occurred from the triplet excited state of anthracene to the initiator, with an intrinsic
kinetic rate constant of 2 x108l/ mol - s. Further evidence for the role of the triplet state
was provided by an observed 7-fold decrease in the polymerization rate upon addition of
a triplet state quencher. Finally, numerical solution of the photophysical kinetic
equations indicated that the triplet state concentration was approximately three orders of
magnitude higher than that of the singlet state, and that 94-96 % of the active cationic
centers are produced by reaction of the initiator with the triplet state. These results
indicate that the electron transfer occurs primarily from the triplet state of anthracene,
with the singlet state providing only a minor contribution to the photosensitization

reaction.

118
In this chapter, a series of studies were performed to determine which anthracene

excited state (singlet or triplet) is involved in photosensitization of diaryliodonium salts.
Anthracene phosphorescence and fluorescence lifetime studies were performed to
determine the effect of the initiator on the lifetimes of the triplet and singlet. The role of
the triplet state was investigated further from studies performed using an anthracene

triplet quencher and from computer simulations of the photophysical-state populations.

V.2. Experimental

V.2.1. Materials

Anthracene, 9,10-dihydroanthracene, 9,10-diphenylanthracene and 9-phenyl-
anthracene were purchased 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-1,3-hexanediol, and ~50 wt % bis(4-dodecyl-
phenyl)iodoniumhexafluoroantimonate. In this initiator, various dodecyl isomers have
been attached to the phenyl rings of the diphenyliodonium salt in order to impart
solubility in the monomer. These alkane substituents have little effect on the photo-
sensitization reaction with anthracene.I Initiator concentrations specified in the remainder

of this chapter correspond to the total UV9310C concentration.

119

V.2.2. Laser Fluorescence Measurements

The laser-induced fluorescence studies were performed in the LASER Laboratory
at Michigan State University. The fluorescence and photosensitization were both excited
with a Coherent Innova 200 argon ion laser operating at 363.8 nm. A Newport 845HP-01
digital shutter system was opened with an electronic pulse from the detector controller,
ensuring that the illumination and fluorescence acquisition were started simultaneously.
Fifteen milliwatts of unfocused laser radiation (measured with a Scientech 362
power/energy meter) in a ~3—mm-diameter beam was directed upon the quartz capillary
tube (1 mm inner diameter, 2 mm outer diMeter) placed perpendicular to the beam
direction. The fluorescence was collected at an angle of 90° from the incident beam and
90° from the longitudinal axis of the quartz tube. The fluorescence signal was analyzed
using a Spex 1877 Triplemate spectrometer with a subtractive dispersion filter stage and a
spectrograph stage. An EG&G Princeton Applied Research Model 1421 intensified diode
array, cooled to -20 0C to minimize dark charge levels, was used to detect the signal, and
was interfaced to an EG&G Princeton Applied Research Model 1463 OMA IH detector

controller.

V.2.3. Phosphorescence Lifetime Measurements

Phosphorescence lifetime studies were performed using a nanosecond pulsed

Nd:YAG laser (Quanta-Ray DCR-l) for the excitation source. The fundamental output

120
of the NszAG was frequency tripled giving the third harmonic (355 nm) with a pulse

width of 8 ns and a 10 Hz pulse rate. Eight mW (Scientech 365 power/energy meter) of
laser light was directed through a photodiode trigger (Motorola MRD-500) and into a
quartz cuvette containing the sample. The phosphorescence was dispersed with a Spex
1680A monochromator, detected by a Hamamatsu R928 photomultiplier, and the signal
was analyzed with a Tektronix DSA602A storage oscilloscope. The phosphorescence
was monitored at 670 nm with a 600 nm long pass filter placed at the monochromator
entrance slit. In order to remove oxygen, the methanol/anthracene samples were
subjected to freeze-pump-thaw cycles until the residual pressure when frozen was below

5.0 x 10-5 torr.

V.2.4. Fluorescence Lifetime Measurements

The picosecond-lifetimes were obtained using a single-photon timing (SPT)
instrument constructed in-house. SPT (also known as time-correlated single-photon
counting) measures ultra-fast luminescence lifetimes from the time-delay between the
arrival of a laser excitation pulse at the sample and the detection of the first collected
emission photon. This sub-nanosecond timing is accomplished using a time-to-amplitude
converter (TAC; Tennelec Model TC 862) which outputs a voltage proportional to the
measured delay time. The TAC timing is started and stopped by pulses from a pair of
constant-fraction discriminators (CFDs; Tennelec Model TC 455). The CFDs produce
output pulses at the same relative height of the input trigger pulse, even if the input pulse

height is fluctuating shot-to-shot (which can be a problem for laser pulses and single-

121
photon pulses). The “start timing” input is ordinarily generated from the laser pulse and

the “stop timing” input generated by the first detected emission photon; however, the
instrument may also be run in “reverse—timing” mode where these assignments are
reversed. The luminescence from the sample is collected, collimated and focused into a
SPEX 1681 monochromator. The detector is a thermoelectrically cooled Hamamatsu
Model R1564 microchannel plate, which provides an electrical pulse for each detected

single-photon event. The TAC output voltage for each At is converted to a digital value

by a multi-channel analyzer (Nucleus, Model 8000AT) and placed into the appropriate
“bin” of a histogram of number of events vs. delay time. This histogram looks precisely
like a classically produced luminescence decay curve, and provides the same lifetime
information about the sample. The laser system is comprised of a Coherent Antares
mode-locked Nd:YAG laser which is frequency doubled, and the 2nd harmonic used to
pump a sync-pumped, cavity-dumped Coherent 702 dye laser running rhodamine 6G dye.
This laser was operated at 610 nm and subsequently frequency-doubled to 305 nm and
used as the excitation source for the anthracene. The (uncorrected) autocorrelation width

of the laser pulses is ~ 6 ps. The total SPT instrument response time is less than 50 ps.

V.2.5. Absorption and Fluorescence Measurements

Absorption spectra of initiator and photosensitizers were obtained using a Hewlett
Packard UV-VIS 8452A diode array spectrophotometer. An Aminco-Bowman Series 2
Luminescence Spectrometer was used to measure the fluorescence of the various

compounds. The temperature of the fluorescence samples was controlled by a

122

Polyscience 9100 refrigerated constant temperature circulator connected to a jacketed

cuvette holder in the spectrometer.

V.3. Results and Discussion

V.3.1. Photosensitization Mechanism

These results indicate that photosensitization of diaryliodonium salts by
anthracene proceeds by electron transfer primarily form the anthracene triplet excited
state. The important photochemical steps of the proposed mechanism are shown in
Figure V.1. In the first step of this mechaniSm, the photosensitizer absorbs light and
populates the first excited singlet state. From the excited singlet state, the anthracene
may return to the ground state by fluorescence or radiationless deactivation, or undergo
intersystem crossing to the triplet state. In the second step of the mechanism, an exciplex
is formed from the association of a triplet state anthracene and a ground state initiator
molecule. This exciplex formation is the only diffusion controlled step in the mechanism
and therefore is the rate limiting step in the reaction. In the exciplex, an electron may be
transferred from the anthracene to the initiator (step 3 of the mechanism), resulting in the
formation of a radical-cation in the 9 and 10 positions of the anthracene (this location is
favored due to resonance stabilization). This radical-cation may undergo several possible

reactions, however the important result for our purposes is the production of cationic

123

1. Photoexcitation

.. é(000)—11» ——»(ooo*)

2. Exciplex Formation
3 * e e 3 e e *
(~) + Aral SbF5 T—_——-[ N2. prs]

3. Electron Transfer

[3~A'2'$SbFse] *——> ©O© SbFse + Arzl -

Figure VJ. Proposed photosensitization mechanism based on electron transfer from
triplet state anthracene to initiator.

124
active centers capable of initiating polymerization. The possible photolysis products of

the photosensitized diaryliodonium salts have been examined in considerable detail by
Dektar and Hacker.2

Not shown in Figure V.1. is the possibility of the excited singlet state anthracene
forming an exciplex with a ground-state initiator molecule. Once formed, the singlet-
state exciplex could undergo a redox reaction to produce cationic active centers in a
manner analogous to steps 2 and 3 of Figure V.1. As discussed later in this chapter, a
series of studies to determine the relative importance of reaction from the singlet state to
reaction from the triplet state were performed, and the fraction of active centers produced
by each path were evaluated. Previous fluorescence studies3 corroborate the proposed
photosensitization mechanism shown in Figure V.1. First, the observed decrease in
fluorescence is consistent with the electron-transfer scheme since anthracene is consumed
when the central ring loses an electron. In contrast, during the energy transfer mechanism
the anthracene would be regenerated and not consumed. The order of reactivity of the
anthracene derivatives is also consistent with the electron transfer scheme.

Our results illustrated that anthracene and 9,10-dimethylanthracene resulted in the
most rapid reactions, followed by 9-vinylanthracene and 9,10-diphenylanthracene, re-
spectively.3 The relatively low reactivity of 9,10—diphenylanthracene may be attributed to
steric effects associated with the bulky aromatic substituents in the 9 and 10 positions,
and to the enhanced resonance stabilization afforded the radical-cation by these
substituents. Similarly, the vinyl substituents of 9-vinylanthracene produces some steric
hindrance and resonance stabilization of a 9,10 radical-cation compared with anthracene

or 9,10-dimethylanthracene.

125

Further evidence for the electron transfer mechanism for photosensitization may

be obtained from thermochemical considerations.4 The electron transfer scheme is
thermodynamically feasible if the excited state energy of the photosensitizer (E ‘) is

greater than the difference between the energy ( E "X ) required to oxidize the photo-

39".?

sensitizer and the energy (E,;‘,‘,’ ) released by reduction of the initiator.5 Therefore, the
magnitude of the free energy (AG) of photosensitization should be negative. As
illustrated for clarity in Figure V.2., the free energy change of photosensitization (AG)
can be readily determined using the Rehm-Weller Equation:6
AG = (E33333 - 15.2.3? ) — E‘ Eq. 1

Figure V.2. illustrates the division of the photosensitization process into two hypothetical
steps. During the process energy is released in the transition from excited state to ground
state (A61 is negative) however energy is consumed in the redox reaction (A62 is

positive). Some authors have mistakenly used the excitation energy in place of the donor

excited state energy when calculating AG.”

The energy required to oxidize anthracene has been reported to be 106 kJ/mol.7
Similarly, the reduction energy for the diaryliodonium initiator (E322?) has been
determined to fall in the range of -19 to -48 kJ/mol."a The excited-state energy of the
photosensitizer may be calculated from the frequency of the 0-0 fluorescence (singlet
state) or phosphorescence (triplet state) transition. As shown in Figure V.3., the values of
E* are 315 and 176 kJ/mol for the singlet and triplet states, respectively.9 Based upon

these values, AG for electron transfer from the singlet and triplet states of anthracene to

the diaryliodonium salt were determined to be -175 kJ/mol and -36 kJ/mol respectively.

126

Free Energy of Photosensitization Reaction

 

A"+I AG > on+1red
A*+I—é§1—> A+I—A—C}z-> on+1red
AG-E“ AG=EOX _Ered

1" 2 sens init

A* - Excited State Anthracene
(singlet or triplet)

A - Ground State Anthracene
on - Oxidized Anthracene

1 - Initiator

I red - Reduced Initiator

Transition Energy Released from Excited State to Ground State (E*)

S, -——-> So+hv hv=315kJ/mo1

T1 —-> So + hv hv = 176 kJ/mol

Figure V.2. Hypothetical pathway and energetics of the photosensitization reaction.

127

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2 Internal Conversion
I g;
S, 1 / InterSystem Crossing
A :E
a 1i 0 ' T1 8
.9 1 8 1. 5
in i 8 a: 8
8 i {3 -315kJ/mol ii -176 kJ/mol 95’
:3 is 8 ‘8.
2 1713:ii E ”if 8
1 1' 1 :i in
so 0 9-1
Ground State Anthracene

Figure V.3. Electronic energy level diagram for anthracene fluorescence and
phosphorescence.

128

The 0—0 transition of the fluorescence and phosphorescence were used in the
calculation since they represent the transition to the ground state of the photosensitizer.
The 0-0 transition is the highest energy transition (Figure V3.) and gives the most

negative value for AG. However the wavelength which gives AG = 0 can be calculated to

be 855 nm, which is well past both fluorescence and phosphorescence of anthracene.
Therefore, both fluorescence and phosphorescence of anthracene have favorable free
energy values making the electron transfer reaction therrnochemically feasible.

The above analysis illustrates that photosensitization is thermodynamically
feasible from both the singlet and triplet state of anthracene. However, in order to
determine the relative amounts of reaction from each state the kinetics of the photo-
sensitization must be considered. Figure V.3. shows the relative energy levels for the
singlet and triplet states of anthracene. The distinction can be important because the two
states have considerably different lifetimes, and may interact differently with molecules
that may be present in the reaction system. For example, some compounds may act as
triplet state quenchers, but may have little or no effect on the singlet excited state.

Lifetime considerations would seem to favor electron transfer from the triplet
state rather than the singlet state. The triplet state lifetime is three orders of magnitude
longer than that of the singlet state, therefore it is much more likely that a triplet state
molecule will encounter an initiator molecule during its lifetime. Also, the triplet state
contains two unpaired electrons and can be represented as a diradical in the 9 and 10
positions. This facilitates donation of an electron from the triplet state anthracene. For
example, Andersen et al.In demonstrated that the radical polymerization of styrene could

be photosensitized by anthracene and concluded based upon lifetime considerations that it

129
was triplet-state anthracene that interacted with styrene. Furthermore, it has been

demonstrated that standard radical initiators may serve as photosensitizers for onium salts
through a similar electron transfer process.H However, both the reactivity and relative
populations of the singlet and triplet states need to be investigated to determine the

relative number of active centers generated from each state.

V.3.2. Fluorescence and Phosphorescence Lifetime Studies

Transient fluorescence and phosphorescence studies were performed to determine
the anthracene singlet and triplet state lifetimes both in the presence and the absence of
the initiator. Oxygen is a known triplet quencher, and was removed using several freeze-
pump-thaw cycles to increase the signal to noise ratio. Transient phosphorescence
profiles for anthracene in methanol and anthracene plus initiator in methanol are shown in
Figure V.4. In this experiment, the phosphorescence was collected at 670 nm with a 600
nm long pass filter located at the monochromator inlet. The figure illustrates that the
phosphorescence decay rate increased nearly 26-fold when 2.0 wt % initiator was
included in the sample. This dramatic reduction in the triplet lifetime upon the addition
of the initiator provides direct evidence that electron transfer occurs from the triplet state
of anthracene during the photosensitization process. The phosphorescence decay profiles
in Figure V.4. were fit with single exponential decay functions yielding a time constant of

9.55 x 10” s in the absence of initiator, and 3.85 x 10'7 s in the presence of 2.0 wt %

initiator. Since the initiator is present in great excess its concentration may be assumed to

130

 

 

 

 

5~ 1
i, -——-———— 6.6x10‘3 Anthracene in MeOH
4. '13 ——~— 6.6x10‘3 Anthracene and
3 i l 2 wt% Initiator in MeOH
> "1
-*-'-' - 1 1.
8 3 l
.93 . t
C ‘1
2e ”1,,
d i. a; 3 .
1_ p333}
owfi ...... ~
0 10 20 30 4o 50 60
time (usec)

Figure V.4. Transient phosphorescence profiles of 6.6 x 10" wt % anthracene in
methanol with and without 2.0 wt % initiator.

131

remain constant during the reaction. Therefore, an apparent rate constant for
photosensitization can be calculated assuming that the relaxation of the triplet state is first
order, as shown in Equations 2 and 3. The solution to these rate expressions are the
exponential decay function shown in Equation 4 and 5.

d3A
[dt1=—<k.,..,.+k.,,.,..>[-‘A.] Eq. 2

d3A
—[dt—] = —(kphns

 

+ km... + kiwi-‘11.] Eq. 3

3
[ A] -(kphos +kmmracl )'

[3A0]
[3A] 3

 

.
+1: )I

.renphns E q . 5

+1:

phos nnnmd

 

Where [3A] represents the triplet state anthracene concentration at time t, [3A0] represents
the initial triplet concentration, kphos is the rate constant for phosphorescence, knonmd is
the rate constant for all non-radiative processes, and k‘senphos is the pseudo rate constant
for photosensitization. The rate constants may be evaluated based upon the exponential

time constants as shown below:

1

k + k =
( phas normal) 955 x 10-6 seC

 

Eq. 6

. l
“”h”) _ 3.85 x 10'7 sec

 

a +k +k Eq7

phm nonradit

Combination of Eq. 6 with Eq. 7 yields a value of 2.50 x 10“ s’I for the pseudo

kinetic rate constant, k' (k‘ = kmpho_3[1]). Substitution of the diaryliodonium

senphos ’ senphos

concentration into this expression ([1] ~ 0.014 i 0.004 mol/l) yields a value of 2 x 10"

132
l/mol-s for the intrinsic kinetic constant for photosensitization from the triplet state,

ksenphos'

Transient fluorescence studies were also performed to determine the effect of the
initiator on the singlet-state lifetime. As shown in Figure V.5., a small reduction in the
fluorescent lifetime from 4.2 us to 3.2 ns was observed. These values are consistent with
literature values for the fluorescent lifetime of anthracene which range from 5.03 ns to
2.64 ns.m3

Reaction rate constants from fluorescence lifetime experiments were calculated
assuming all reduction in lifetime results in production of active centers. The transient
fluorescence profiles in Figure V.5. were fit with single exponential decay functions. The
time constant for anthracene excited at 305 nm and monitored at 425 nm was found to be

4.2 x 10'9 s in the absence of initiator, and 3.2 x 10'9 s in the presence of initiator. Since

the initiator is in great excess, it was again assumed to remain constant during the
reaction. In an analysis analogous to that outlined earlier for the triplet state, a value of

5.7 x 109 l/mol~s was found for the intrinsic kinetic constant for initiation from the singlet

state, ksenfluor

Comparison of the rate constants for photosensitization reveals that the singlet
state rate constant is over an order of magnitude higher than the triplet state rate constant.
However, to determine the relative importance of the singlet and triplet states for the
production of cationic active centers, the relative populations of the states must be

considered. Since the triplet state has a much longer lifetime, its population may exceed

133

 

 

 

 

 

1200-
1000- ~ 6.6x10'3 wt°/o Anthracene and
- 2 wt% Initiator in MeOH
300- Exponential Fit
g '1 ‘3‘
g 600. .. = 6.6x10‘3 wt°/o Anthracene in MeOH
g . '~ ' Exponential Fit
400-
200-
0.. ....
01% 1'0 1'5 2'0 2'5 30
time (ns)

Figure V.5. Transient fluorescence profiles for 6.6 x 10" wt % anthracene in methanol
with and without 2 wt % initiator.

134
the population of the singlet state by orders of magnitude. The relative population of the

triplet and singlet state depends primarily on the intersystem crossing rate constant (km).
Experimental measurement of this constant is difficult and literature values range from

7.38 x 107 to 1.12 x 108 s‘I for anthracene in solution.”9

The underlying photophysics of the sensitization reaction can be described by
equations 8-15. These equations represent the anthracene electronic state transitions as
well as the reaction of excited-state anthracene with initiator. Other possible

photochemical reactions of anthracene have been excluded due to their relative

 

 

 

 

unimportance.
hv + Ae—kahi—)'A Eq. 8
'A lffluo'” >A+hv Eq.9
‘A ——kmmi—> A Eq. 10
lA—-]—cl;~£C—>3A Eq. 11
3A kph“ 3 A + hvp Eq. 12
3A —-lfmmd—) A Eq. 13
3A + 1 k‘e’W’ws > Rt Eq. 14
'A + 1 k‘flfl‘w’ > Rs Eq. 15

Here 1 represents the diaryliodonium salt and A represents ground state anthracene, while
1A and 3A represent anthracene singlet and triplet excited states, respectively. In addition,
kabs is the rate constant for photon absorption, and kfluo, is the rate constant for
fluorescence. Finally, Rs represents active centers created by reaction of the initiator with

singlet—state anthracene while Rt represents active centers created from reaction with the

135
triplet state. The remaining symbols used in the equations have been defined earlier in

this chapter.
Based upon the above reaction pathway for photosensitization, balances on the

various reaction species yield the following set of coupled, first and second order kinetic

 

 

equations.
% _—. kaMll/‘l — [A]) + (kflm + km,“ )[‘A] + (kph... + k,,,,,,,,d )[3A] Eq. 16
”—2? = k,,,.([A] —1'A1) — (1,3,, + k......)1'A1- k...1‘A1 - k...,......t'A]111 Eq.17
£531]- = 1311.1] — (k... + 13,33,131] - k....,....1-‘A1111 Eq. 18
51% = -k...,....1'A1111— k........13A1111 Eq- 19
”if” = k.........1'A1111 Eq. 20
”if” = k....,.....13A1111 Eq. 21

Solution of this set of equations requires the evaluation of the various kinetic constants.
The sum of the rate constants kfluor and knonmd were determined from the fluorescence
lifetime measurements described previously, yielding (kfluo, + knonmd) = 1.38 x 10" 5".
Based upon the known values of the laser power, the beam cross-sectional area, and the
photon flux, the value of kabs was calculated to be 2.79 s'I for our experiments. Finally,
as discussed above, the value for the intersystem crossing constant (km) has been
reported to range from 7.38 x 107 to 1.12 x 10” s'I for anthracene solutions.” Based upon
these values of the kinetic constants, the set of differential equations were solved

numerically using the Rosenbrock method for stiff systems.‘4

136

/ ' —-— Ground State Anthracene
\ ,A' + Singlet State Anthracene
0-3 ‘ -—+—— Triplet State Anthracene
\f —v— Active Centers from Singlet

 

~ Active Cbnters from Triplet

concentration (mM)

 

 

time (s)

Figure V.6. Simulation results for the concentration profiles of the reaction species. The
initial [A] was taken to be 0.5 mM with [I] = 14.5 mM.

137

Figure V.6. contains simulation results for the concentration profiles of various

species in the reaction system. For the range of km used (7.38 x 107 to 1.12 x 10" s"). the

fraction of active centers produced from the triplet state was found to range from 94 to 96
%. Therefore, assuming that all of the reaction with the singlet state results in the
formation of active centers, only 4-6% of the active centers arise from the singlet state.
As shown in Figure V.6., the simulation results indicate that the anthracene concentration
decays exponentially with time, while the number of active centers produced increases
monotonically. The singlet and triplet state concentrations each pass through a maximum
then decay towards zero. Except for reaction times less than a few milliseconds, the
triplet state concentration is always approximately three orders of magnitude greater than
that of the singlet state. Finally, based upon the simulation results, the time constant for
the exponential decay of the anthracene concentration ranges from 3.0 to 3.7 seconds.

In a final experiment, the steady state anthracene fluorescence intensity was
monitored as a function of time for reaction systems containing photosensitizer and

initiator in methanol. For a reaction system containing 1.0 x 10'2 wt % anthracene and 1

wt % initiator, the photosensitizer fluorescence intensity was found to decay
exponentially with a time constant of 4.7 seconds.3 This fluorescence decrease was
attributed to consumption of the anthracene by electron transfer during the
photosensitization reaction. In Figure V.7. the experimental photosensitizer fluorescence
intensity profile is plotted along with the simulation results for the anthracene

concentration as a function of time. The simulation results are based upon the

k

phm + mmrud ’ senllunr ’ k.\‘¢'nphn.\‘

independently measured rate constants (k + k

uar nunrud ’

) and the

138

 

 

 

 

' 20000

. :1
A Anthracene fluorescence at 425 nm S
5 + Simulation Results _ 16000 a
E, . 8
c ‘3
.g .3. F 12000 8
g r; J I .3 . 5-
: 3‘, ., 3.1..
8 3 ‘\ - 8000 g
g ‘\ ‘4-
0 0.0001 - ' '2

r 4000

0.0000 . . . . 7'"
o 2 4 12

 

time (s)

Figure V. 7. Comparison of computer simulation and experimental photosensitizer
fluorescence intensity for reaction of anthracene (0.5mM - 1.0 x 10'2 wt %) and diary-
liodonium hexafluoroantimonate (14.5 mM - 1 wt %) in methanol.

139

literature value of k1... , with no adjustable due to an instrumental artifact, the fluorescence

intensity and simulation results show good agreement.

V. 3. 3. Triplet Quencher Studies

If the photosensitization reaction takes place primarily from the triplet state of the
photosensitizer, the introduction of a phosphorescence quencher should dramatically
lower the rate of active center production. For example, dihydroanthracene (DHA) will
react with triplet anthracene” by transferring a hydrogen atom to the triplet state
anthracene yielding two monohydro-anthracyl radicals which subsequently dimerize.”
The resulting anthracene dimer does not photosensitize the diaryliodonium salt.
Therefore, DHA effectively decreases the triplet anthracene concentration but has
essentially no effect on the singlet state anthracene concentration. Transient fluorescence
and phosphorescence studies were performed to determines the effect of DHA on the
lifetimes of the singlet and triplet states of anthracene. Systems containing 5 x 10'3 wt %
anthracene in methanol exhibited a two-fold reduction in the triplet state lifetime but
essentially no change in the singlet state lifetime upon addition of 5 x 10'2 wt % DHA.
Based upon these results, the rate constant for reaction of DHA with triplet state
anthracene was calculated to be 1.8 x 107 l/mol-s. These studies verify that the triplet
state was quenched without significantly effecting the singlet-state lifetime. The DHA

will therefore act as a retarder or inhibitor towards triplet-state photosensitization by

effectively competing with the initiator for the triplet anthracene.

140
The above discussion illustrates that DHA reacts only with the triplet state of

anthracene without directly effecting the singlet state concentration. Therefore DHA
provides a means of distinguishing the anthracene excited state (singlet or triplet) from
which the photosensitization occurs. If the electron transfer to the diaryliodonium salt
occurs primarily from the triplet state of anthracene, the addition of DHA to a cationic
photopolymerization should reduce the polymerization rate by inhibiting the production
of active centers. If, however, the photosensitization occurs primarily form the singlet
state, the addition of DHA to the reaction system should have relatively little effect on the
production of active centers. To determine the effect of DHA on the cationic
photopolymerization of a divinyl ether (3,6,9,12,-tetraoxatetradeca-1,13-diene) the
polymerization was monitored using a previoUsly described fluorescence technique.3 A

reaction system containing 2.6 x 10'2 wt % anthracene and 2 wt % initiator exhibited a 7-

fold reduction in polymerization rate upon addition of 0.21 wt % of DHA. As shown,
Figure V.8. exhibits the sudden decrease at 5.5 seconds indicating more than a 7-fold
reduction in the rate of photosensitization upon addition of DHA, compared to reactions
with similar anthracene and initiator concentrations. This considerable reduction in
polymerization rate upon addition of DHA provides further evidence that the

photosensitization reaction occurs predominantly from the triplet state of anthracene.

V3.4. Competing Reactions

Other photo-induced reactions of anthracene which could interfere or compete

with photosensitization were also investigated. Dimerization and oxidation are two

141

10000-

-+-— 2.61102 wt°/e Anthracene

9
3’ W 0.21 wt% 9,1013—1A
aooo- . 2wt% initiator

1.
l

l l
.‘m
N-“
__ 3...... .... ~0-~

 

 

b
'7, 6000 -
C ' .
.93
.E L
4000- l, 3
3? M‘ $81.16“, M‘wu‘p
q \ehww
2“” r T f I I I l
o 2 4 6 8 1o
tim e (s)

Figure V.8. Polymerization of DVE-3 monomer with 2.1 wt % initiator and 10'2 wt %

anthracene as the photosensitizer in the presence of 0.21 wt % of the triplet quencher
9, I 0-dihydroanthracene.

142
possible reactions which remove anthracene and form non-reactive compounds. For

example, a stable photodimer is formed when anthracene is excited with 350 nm light by
the following reaction:
A+‘A —) A2 Eq. 22
Once formed the dimer is not capable of photosensitizing the reaction. Experiments were
performed to determine the rate of dimer formation. The dimer reaction can be monitored
by measuring its UV-VIS absorption spectrum at 280 nm. The half-life for the dimer
reaction was found to be ~ 3 hours, therefore it was determined that the dimerization
reaction is not largely interfering with photosensitization.
The second possible interfering reaction is the oxidation of anthracene. Oxygen is

known to react with anthracene by the following paths:

o,+'A -> 0,3,4 Eq. 23
02+3A —> 0,311 Eq. 24
0,34 —-> 02 + A Eq. 25
A+Oz3A—)AOZ+A Eq.26

In these reactions no effort has been made to remove or lower the oxygen concentration,
and the photosensitization reaction reaches completion in under 2 seconds for all cases.
The oxygen may be acting as a competing reaction of photosensitization, but the oxygen
is at a much lower concentration than the initiator. Oxygen concentration calculated from
comparison of similar organic solvents" was found to be ~1 x 10" M. At this
concentration the oxygen is not appreciably interfering with the photosensitization

reaction.

143

V. 3.5. Fluorescence of the Products from Photosensitizing

Diaryliodonium Salts with Anthracene

Anthracene has been shown to photosensitize diaryliodonium salt photoinitiators

l-3.l7.2(l

by electron transfer leading to the production of cationic active centers. In this
process. an excited state anthracene molecule transfers an electron to the iodonium salt.
which results in the decomposition of the diaryliodonium salt. The photochemistry of
diaryliodonium salts including photosensitization by several compounds has been

7

examined to great extent.“ However, there has been some question of the anthracene

3. 7
' Furthermore, there are some

photophysical state that is transferring the electron.
remaining questions as to the finial result of the anthracene after the photosensitization
reaction.

In the photosensitization reaction, an anthracene cation radical is produced when
an electron is transferred to the diaryliodonium salts. Dektar and Hacker have reported
the production of phenylanthracenes by reaction of anthracene cation radicals with aryl
radicals formed by fragmentation of diaryliodonium salts, followed by proton loss.2
However, they could only account for ~30 % of the anthracene consumed by the
photosensitization reaction. In this research, anthracene fluorescence intensity has been
used to monitor the photosensitization reaction in situ.3 The anthracene fluorescence
intensity decreases during the reaction as the anthracene is consumed. However, if

phenylanthracene is produced from this reaction, phenylanthracene fluorescence should

be seen to grow in as the reaction proceeds. Since no phenylanthracene fluorescence is

144

seen in the experiments, we assume that the anthracene was converted to a non—
fluorescent species. If anthracene losses the aromatic nature of its central ring by reaction
with an aryl radical followed by a second reaction to neutralize the positive charge. its
fluorescence will be greatly reduced.

In order to determine if phenylanthracene is a major product produced during the
photosensitization reaction, experiments were conducted using UV-VIS absorption and
fluorescence measurements. Phenylanthracene and 9,10-diphenylanthracene were
compared to anthracene and solutions of anthracene and diaryliodonium that were
photochemically reacted. The absorption and fluorescence of 9,10-dihydroanthracene
was also compared as a model for anthracene products which have lost the aromatic
nature of the center ring.

Absorption of all the anthracene derivatives were measured in methanol solutions
at similar molar concentrations. Methanol was used as the solvent since it has excellent
optical properties and also contains an extractable proton. However. the experiments
were also done in the divinyl ether monomer. Figure V.9. shows the relative absorption
of the anthracene compounds in the 300 to 450 nm region. At 365 nm, the molar

absorptivity of anthracene was determined to be ~3000 l/mol-cm, while the absorptivity

of 9-phenylanthracene and 9,10-diphenylanthracene were found to be ~6000 and ~4500

l/mol-cm respectively. Therefore, if either of the compounds are produced, they would

absorb more strongly at 365 nm than anthracene. In constant, the absorption of 9,10-

dihydroanthracene is essentially zero at 365 nm.

145

  
   

3- r\
/ r
l — — 4.6 x 10" M 9—Phenylanthraoene
l — - - 5.2 x10‘4 M 9,10-Dihyroanthraoene
23 / -- - -~ ~62 x 10" M Anthracene

—— 5.5 x 10" M 9,10-Dbhenylanthraoene

absorption

 

 

wavelength (nm)

Figure V.9. Relative absorption of anthracene and its derivatives.

146

 

 

 

”-1 ._
700.. .' 1.3’“ ‘. -- ~ - 5.5x10’4M9,10-Dphenylanthraoene
. i \ - - 4.6 x10“4 M 9-Phenylanthraoene
soo- '. - - - - 6.2x 10'4MAnthtacene
a; - / i j“ ‘1 —5.2x10'4M 9,10-Dhyroanthraoene
g 500- , \' \ ‘
g " I l\ j \ \1
E ' I r 1 i
o 300_ l
m \
e . I ,' 1.. ‘
é 300‘ I \L
. I . \ .
100.1 1 .i \ \ I 3
‘ [7 I , 3 x \ '“‘ .
Al/A " -- -‘—‘--..E_EE_L\~"‘-‘-‘:L'__~_;;-:—_+A
o I ' I T I ' I ' I ' I ' I 1 I I I
375 400 45 450 475 500 525 550 575
wavelength (nm)

Figure V.10. Relative fluorescence intensity of anthracene and its derivatives.

147
The fluorescence of the same solutions was measured with excitation at 365 nm.

The fluorescence emission was measured from 370 to 600 nm. Figure V.10. show the
relative intensity and spectral shape of the anthracene derivatives. As expected, 9,10-
dihydroanthracene shows no fluorescence when excited with 365 nm light. The relative
fluorescence intensity increases from anthracene, 9-phenylanthracene to 9,10-diphenyl-
anthracene. The 9-phenylanthracene intensity is 8 time that of the anthracene
fluorescence, and 9,10-diphenylanthracene is 12 time more intense. These studies were
performed to highlight the intensity of fluorescence and differences in spectral shape of
the anthracene compounds.

As shown in previous work3, when anthracene fluorescence intensity is followed
during photosensitization of diaryliodonium Salts in methanol, no change in the position
or shape of the peaks is seen. As shown above, if 9-phenylanthracene or 9,10-
diphenylanthracene is being produced in appreciable amounts, changes in the
fluorescence spectra should be observed. Figure V.11. show the fluorescence of an
anthracene and diaryliodonium solution before and after photosensitization. The
anthracene fluorescence has decreased considerably, however the spectral shape has
changed only slightly. The small change in the spectrum can be seen when the relative
intensity of the first and third peaks of the anthracene are compared. In the reacted
solution the third peak has grown in intensity slightly compared to the first peak. This
increase occurs around 425 nm where 9-phenylanthracene does have a spectral peak.
However, the amount of change is small compared to the overall reduction in the

fluorescence intensity.

148

 

10 -
9 ..
_ . . - - - - Noexposure
8 - ' '3 After exposure to light >330 nm

. '. Anthracene 2.0 x 10'3 wt %
' ' Initiator 3.0 x 10'2 wt % in Methanol

fluorescence intensity
0|
I

 

 

 

wavelength (nm)

Figure V.11. Fluorescence intensity of anthracene before and after photosensitization of
diaryliodonium salts.

149
The mechanism for the formation of phenylanthracenes by reaction of aryl

radicals with the anthracene cation radical pr0posed by Dektar and Hacker2 involves
proton loss from the anthracene cation radical to remain aromatic in nature. However, we
have shown that diaryliodonium salts can also be photosensitized by 9,10-
dimethylanthracene and 9,10-diphenylanthracene3. which do not contain available
hydrogen atoms at the 9-10 positions. The 9,10-dimethylanthracene was shown to have a
similar rate to anthracene, however 9,10-diphenylanthracene photosensitization rate was
determined to be much slower, indicating an inference caused by the phenyl substituents.3

9-Phenylanthracene has also been found to photosensitize diaryliodonium salts.
The fluorescence intensity was followed both in methanol and during a cationic
polymerization of DVE-3. Figure V.12. shows the fluorescence profile for photo-
sensitization of diaryliodonium by 9-phenylanthracene. The fluorescence profile was
obtained by methods that have been described elsewhere.3 The fluorescence profile
decreases as the reaction proceeds again with no shift in the position of the peaks and no
new fluorescence peaks from reaction products.

In this research, the exact products from the anthracene photosensitization
reaction have not been determined. However, the final product produced in these
reactions is non-fluorescent in the 400 - 500 nm region when excited at 365 nm. Our
fluorescence studies indicate that large amounts of 9-phenylanthracene are not produced
by the photosensitization of diaryliodonium salts by anthracene. The important processes
in the reaction are the production of cationic active centers and the loss of anthracene

fluorescence as the photosensitization reaction occurs.

150

 

 

 

 

16000
14000
12000
.2: .
E 10000 \\\.\‘\‘ ‘t a
3 ~ \‘\\,.\\\~ » .
_ ‘V‘\\§\\‘r\3§\‘ §\‘_\_:b;c ‘\
8000 ~ ;‘ “NM \
. 3 ‘~\ -. ‘ \\\. 33“.‘\,\‘ 3\.j.\1\
\‘I‘t ‘\\\\\\\E\T‘\‘Q §‘\\\\\.\\\{\T 3.; -
El 0000 .agsaeg§§§§:s- “ i ”
- e s Q ‘\
4000
2000
0 . I I
375 400 425 450 475 500 525
vuanlengfli(nn1

Figure V.12. Fluorescence profile during polymerization of DVE-3 monomer (1 wt %
initiator) produced by exciting 9-phenylanthracene (1.1 X 10'2 wt %), at 363.8 nm, and
monitoring the fluorescence intensity at time intervals of 70 ms.

151

VA. Conclusions

In conclusion, our studies indicate that photosensitization of diaryliodonium salts
by anthracene occurs by electron transfer primarily from the excited triplet state of the
photosensitizer to the initiator (see Figure V.1.). The observed 26-fold increase in the
rate of anthracene phosphorescence decay upon addition of initiator (see Figure V.4.)
provides direct evidence for reaction from the triplet state. Based upon these studies, the

intrinsic rate constant for photosensitization from the triplet state was found to be

k

mph“ = 2 x108 l/mol-s. Further evidence for the importance of the triplet state is
provided by the 7-fold decrease in the polymerization rate for a reaction system
containing an anthracene triplet-state quencher (DHA) compared to corresponding system
with no quencher. Finally, numerical solution of the photophysical kinetic equations
indicate that the triplet state concentration is approximately three orders of magnitude

higher than that of the singlet state, and that 94-96 % of the active cationic centers are

produced by reaction of the initiator with the triplet state.

v.5. List of References

l. J. V. Crivello, “Cationic Polymerization-Iodonium and Sulfonium Salt
Photoinitiators”, Adv. Polym. Sci., 62, 1 (1984).

2. J. L. Dektar, N. P. Hacker, “Photochemistry of Diaryliodonium”, J. Org. Chem, 55,
639 (1990).

10.

11.

12.

13.

14.

15.

16.

17.

18.

152

E. W. Nelson, T. P. Carter. A. B. Scranton, “Fluorescence Monitoring of Cationic
Photopolymerizations: Divinyl Ether Polymerizations Photosensitized by Anthracene
Derivatives,” Macromolecules, 27,1013 ( 1994).

S. P. Pappas, J. H. Jilek, Photogr. Sci. and Eng, 23, 141 (1979).

. G. J. Kavarnos, 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. F. Vogtle, Ed.,
Topics in Current Chemistry 156, Springer-Verlag, New York, 27, 1990.

D. Rehm, A. Weller, Isr. J. Chem, 8, 259 (1970).

P. Sundell, S. Jonsson, A. Hult, “Photo-Redox Induced Cationic Polymerization of
Divinyl Ethers,” J. Polym. Sci., Part A: Polym Chem, 29, 1525 (1991).

Y. Yagci, I. Lukac, W. Schnabel, “Photosensitized Cationic Polymerization Using N -
ethoxy-Z-methylpyridinium hexafluorophosphate”, Polymer, 34, 1130 (1993).

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

V. S. Andersen, R .G. W. Norrish, “ The Polymerization of Styrene Sensitized by
Molecules in the Triplet State”, Proc. Roy. Soc. (London), 251, 1 (1959).

S. Jonsson, P.-E. Sundell, A. Hult, Radtech 90-North America, Radiation Curing
Conference and Exposition Proceeding, March 25-29, 417, 1990.

R. F. O'Dowd, A. O'Hare, J. Cooke, 1.1K. Taaffe, “Laser System and Data Analysis
Techniques for Absorbing and Non-Absorbing Excited Transients,” J. Phys. E: Sci.
Instrum., 15, 736 (1982).

I. B. Berlman, Handbook of Fluorescence of Aromatic Molecules, 2nd ed., Academic
Press, New York, 356, 1971.

R. L. Borrelli, C. 8. Coleman, Difl'erential Equations - A Modeling Approach,
Prentice - Hall Inc., 1987.

F. C. Bevington, Photochemistry and Reaction Kinetics, Polymer Chemistry, P. G.
Ashmore, F. S. Dainton, T. M. Sugden, Ed., University Press: Cambridge, 202, 1967.

R. G. W. Norrish, J. P. Simons, “The Polymerization of Styrene Sensitized by
Molecules in the Triplet State,” Proc. Roy. Soc. (London), 251, 4 (1959).

R. J. DeVoe, M. R. V. Sahyun, E. Schmidt, N. Serpone, D. K. Shanna, “Electron
Transfer Sensitized Photolysis of ‘Onium Salts,” Can. J. Chem. 66, 319 (1988).

Guerry, D. Jr. Vanderbilt University Ph.D. Thesis, 1944.

153

19. E. W. Nelson, T. P. Carter, A. B. Scranton, “The Role of the Triplet State in the
Photosensitization of Cationic Polymerizations by Anthracene,” J. Polym. Sci., Part A
Polym. Chem.,33, 247, (1995). I

20. S. P. Pappas, “Photoinitiation of Cationic and Concurrent Radical-Cationic
Polymerization. Part V,” Prog. Org. Coat., 13, 35 (1985).

21. J. V. Crivello, Organic Coatings, Science and Technology, G. D. Parfitt, A. V. Patsis.
Ed., Marcel Dekker, Inc., New York, Vol. 5, 35, 1983.

CHAPTER VI. IN SI TU RAMAN SPECTROSCOPY:

CURE MONITORING AND KINETICS

V1.1. Introduction

In this chapter, Raman spectroscopy was used to monitor the cure in situ during
the cationic photopolymerization of a divinyl ether photosensitized with anthracene.
Since Raman spectroscopy arises from molecular vibrations, the consumption of double
bonds can be directly measured during polymerization reactions. By monitoring the
double bond concentration, Raman spectroscopy gives a direct method of obtaining the
propagation rate during the reaction. Furthermore, the short intrinsic time scale makes
Raman spectroscopy attractive as a time-resolved technique to monitor cure in these high
speed reactions.

A series of Raman experiments were performed on cationic photopolymerizations
of a divinyl ether initiated with a diaryliodonium salt of hexafluoroantimonate
photosensitized by anthracene. The photosensitizer allowed the initiating wavelength to
be shifted from the deep UV region of the spectrum to the near UV region around 350
nm.‘ Isothermal Raman experiments were performed for a series of reaction temp-

eratures, and used to determine the overall activation energy of the polymerization

reaction. Since the Raman profiles give a direct measure of conversion versus time, the

154

155
rate of reaction was determined during the photopolymerizations. Finally, the Raman

profiles were used in conjunction with previously determined photosensitization rate
constants1 to determine the kinetic constant for the propagation reaction. The resulting
profiles for the propagation rate constants as functions of conversion provide important

insight into the nature of the cationic photopolymerizations.

VI.2. Experimental

V1.2. 1. Materials

In these studies, 3,6,9,lZ-tetraoxatetradeca-l,l3-diene (DVE-3) (GAF Chemicals
Corp.) was used as the monomer. The monomer was dried over molecular sieves to
remove any traces of water. 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)iodoniumhexafluoroantimonate. Initiator concentrations
specified in the remainder of this chapter correspond to the total UV9310C concentration.
The anthracene photosensitizer (Aldrich Chemical Company) was used as purchased.
The spectral region over which cationic initiators are effective may be expanded by the
addition of a variety of photosensitizers.2 These compounds make it possible to initiate
the photopolymerization using near-UV or even visible wavelengths of light?“4
Therefore, a typical reaction system would contain ~l wt % initiator (UV9310C) and

~10'2 wt % of photosensitizer in monomer (DVE-3).

156

V1.2.2. Ramon Experiments

The photopolymerization kinetics were investigated by monitoring the Raman
spectrum as a function of time. The Raman samples were excited with 250 mW of 496.5
nm light from a Coherent Innova 200 Argon ion laser. The photosensitization reaction
was initiated with 363.8 nm light from a Coherent Innova 7O Argon ion laser. The
initiation laser output of fifteen milliwatts (measured with a Scientech 362 Power/Energy
meter) was attenuated with different optical density filters. Since the reaction is highly
sensitive to the initiation intensity, the optical density filters were used to control the rate
of the polymerization reaction. If the initiation power is two high, the reaction will
thermally runaway causing the polymer to char and lose optical clarity. When the laser
power is set too low, the reaction rate is much slower than typical industrial process
conditions.

A Newport 845HP-Ol Digital Shutter system was opened with an electric pulse
from the detector controller, ensuring that the initiation beam and the acquisition were
started simultaneously. Shuttering only the initiation beam allowed the Raman spectra to
be collected before and after initiation of the polymerization reaction. This permitted
dark cure rates to be obtained by terminating the initiation beam and following the
reaction with the second (non-initiating) beam.

Both laser beams were focused on a quartz capillary tube placed perpendicular to
the beam direction. Since the beams were focused to a small spot size, the alignment of

the two beams was imperative to insure that both beams were sampling the same reaction

157

 

/

 

 

 

 

 

 

 

 

Sample Spectrometer
Electronic /
Shutter Ll uld Ndltro en
| 00! ed

 

 

 

 

363.8 nm

 

496.5 nm Raman Beam

Figure V1.1. Experimental setup for Roman cure monitoring.

158
volume. The quartz capillary was approximately 1 inch in length with an inner diameter

of 1 mm and an outer diameter of 2 mm. The Raman scatter was collected at an angle of
90° from the incident beam and 90° from the long axis of the quartz tube (see Figure
V1.1.). The sample holder was jacketed with cooling water and the temperature was
controlled with a circulating water bath.

The Raman signal was collected using a Spex 1877 Triplemate monochromator
with a subtractive dispersion filter stage and a second spectrograph stage. The scattered
light was passed though a depolarizer, placed before the monochromator, to remove any
polarization effects on the signal intensity. A gated EG&G Princeton Applied Research
Model 1530 C/CUV CCD, cooled to -120°C, to minimize dark current levels, was used to
detect the signal. The data was analyzed ‘with a OMA 4000 detector controller and
software. A 1000 cm"1 wide Raman spectrum, centered at 535 nm, was collected with

600 gr/mm gratings in the filter stage and 1200 gr/mm in the spectrograph stage.

V1.3. Results and Discussion

V1.3.]. Raman Cure Monitoring

As mentioned in the background chapter, the major problem associated with
Raman spectroscopy is the interference from molecule fluorescence. For our system,
which contains a highly fluorescent probe (anthracene), the Raman spectrum must be
outside the range of the anthracene photosensitizer fluorescence. Anthracene

fluorescence is centered at ~ 425 nm and is essentially zero by 500 nm. In addition, the

159
Raman cross-section is related to the 4th power of the incident light energy.5 Therefore.

the lowest wavelength (highest energy) outside the anthracene fluorescence (496.5 nm)
was chosen as the scattering beam. Furthermore, our instrumental setup (gratings and
detector) is most sensitive for wavelengths in the 500 nm range.

To monitor the reaction kinetics, the peaks in the Raman spectrum which
correspond to the- reactive double bonds were identified. In addition, a peak whose
intensity is independent of cure was selected as an internal standard. The three major

peaks in Figure V1.2. were identified by comparison with known compounds of similar

6,7 l

composition. The sharp peak at 1322 cm’ corresponds to the CH bending of the
hydrogens attached to a vinyl carbon. The two peaks at 1458 cm'1 and 1469 cm"
represent the wagging and bending of the ethyl-ether hydrogens (O-CHz-CHz-O),
respectively. Both of these peaks should remain relatively unchanged during the
polymerization reaction, since they represent unreactive sections of the monomer
molecule. Finally, the doublet at 1622 cm'l and 1636 cm'l represents the C=C stretching
of the vinyl ether bonds. The 1622 cm'1 peak results from the trans configuration of the
vinyl ether bond as shown in Figure V1.3. The 1636 cm" peak represents the cis isomer
of the vinyl ether bond. The cis isomer has slightly more steric hindrance than the trans
conformation, resulting in the lower Raman intensity.

When the initiation beam is blocked and a Raman spectrum is collected no
reaction is seen. However, if the sample is exposed to the initiation beam the reaction
proceeds to completion in a few seconds depending upon the reaction conditions. This

indicates that no thermal initiation is occurring. If the power of the Raman scattering

beam is increased to 500 mW, thermal initiation will occur when the laser beam is

160

 

 

 

 

 

 

4000 " h
3500 - DVE-3 containing 1 wt % initiator
J and 10-2 wt % anthracene
3000 _ monomer
.. -------- polymer
2500 -
. ,n.
2‘ i i
71') 2000 - i "‘1
c i
0) . l i l
E 1 i
°’ 1500 - l '1 1
l 1
. 1 I x
l J ‘.
I 1.
x
1
k
\\

 

 

Raman shift (cm-1 from 496.5 nm)

Figure V1.2. Raman spectra of DVE-3 before and after cationic photopolymerization.

161

=C Stretching Bands Of Vinyl Ethers

 

H

H H \ /H

\ _ / C——C

C— / \ -.
/ \ . H 0..
H 0.9. C /
c . _1

Trans 1622 cm'1 \ Cis 1636 cm

Figure V1.3. Cis and Trans isomers which result in the ~1630 cm" doublet in the

' monomer Raman spectrum.

162
allowed to illuminate the sample for approximately 2-3 minutes. Figure V1.2. compares

the DVE-3 Raman spectra of monomer and polymer, after isothermal cationic
photopolymerization at 25°C. The relative intensities of the 1630 cm'I doublet versus the
1458 cm'1 peak indicates the conversion of double bonds. Furthermore, increases in the
peaks at 1133 cm’l and 1282 cm'1 correspond to the production of polymer. Since the
Raman intensity is proportional to the double bond concentration, monitoring either the
1322 or 1622 cm'1 peaks versus the 1458 cm'1 peak provides a direct measure of
conversion.

Figure V1.4. shows 60 Raman spectra collected at intervals of 0.792 seconds for a
duration of 50 seconds in situ during the cationic photopolymerization of DVE-3 with 1
wt % initiator and 10'2 wt % anthracene ph0tosensitizer. The temperature of the reaction
was controlled at 20°C using the jacketed sample holder. As expected, the ~1630 cm'l
doublet and the 1322 cm'1 peak decrease as the reaction proceeds. Moreover, the 1458
cm'I band remains relatively unchanged during the reaction, further indicating the
conversion of monomer to polymer. As demonstrated in Figure V1.4., the Raman
monitoring technique allows a complete spectrum (1000 cm") to be monitored in situ.
This provided a much more detailed reaction profile than can be achieved by monitoring a
single wavelength during the reaction.

Using the initial 1322 cm’1 peak intensity, the double bond conversion can be
calculated for each spectra as the reaction proceeds. Figure V1.5. is a plot of double bond
conversion versus time for the same polymerization reaction shown in Figure V1.4. In
this experiment, Figure V1.5. indicates that the conversion starts initially slow and

increases as the reaction proceeds, reaching a maximum conversion of ~ 70 percent. The

163

1500

S

intensity

500

 

 

 

Raman shift (cm-1 from 496.5 nm)

Figure V1.4. Raman spectra during a cationic photopolymerization of DVE-3 with 1 wt
% initiator and 10’2 wt % anthracene.

164

  

 

 

‘ I
0.6 -
C
.9
2
g) 0.4 -
c
o
O
‘2 .
g I DVE-3 containing 1 wt "/0 initiator and
3. 102 wt % anthracene at 20°C
C
-; 0.2 -
0 O - t I 1 I I I r 1 I j u
0 10 20 30 40 50
time (s)

Figure V1.5. Double bond conversion during a cationic photopolymerization of DVE-3
calculated from the 1322 cm" Raman band.

165
Raman monitoring studies indicated that the final conversion value for the cationic

photopolymerization of DVE-3 is dependent on the rate of the reaction. The final
conversions were found to range from 60 to 85 percent when the temperature of the
reaction is varied from 20 to 50 °C.

As described in Chapter IV., a fluorescence monitoring technique was developed
that followed the anthracene photosensitizer fluorescence as the reaction proceeds.1
Since the fluorescence technique monitors the production of active centers directly and
the polymerization reaction indirectly, the Raman technique should verify the ability of
the fluorescence technique to monitor cure profiles. Figure V1.6. shows the relationship
between the fluorescence and Raman monitoring techniques. The two curves represent
the same sample mixture run at the same experiment conditions (initiation intensity,
temperature, integration time, sample rate, etc.). The decrease in photosensitizer
fluorescence intensity corresponds to the production of active centers and therefore an
increase in the rate of the polymerization reaction.

The Raman curve shown in Figure V1.6. follows the intensity of the 1322 cm'I
peak during the photopolymerization reaction. As can be seen from Figure V1.6. , the two
techniques show good agreement in the rate of reaction as indicated by the decrease in
signals at the same reaction time. This comparison helps to validated the fluorescence
monitoring method for determining cure profiles. Moreover by using both of the
techniques, the rate of propagation and initiation can be measured directly during these

high speed reactions and used to determine the rate constants for the reaction.

166

  
  
  

 

 

 

400 —o— Anthracene fluorescence
" intensity at 425 nm
. ~---- Raman intensity at 1322 cm-1
DVE-3 with 102 wt °/0 anthracene - 2000
2‘ 300 - and 1 wt % initiator
17>
c
g .
.E
c F
‘0 200 - 1600
E
at
m d
100 I - 1200
O I I I ' U I I j I 1 W I Y I Y '
0.00 0.25 0.50 0.75 1 .00 1 .25 1 .50 1 .75 2.00

time (s)

Figure V1.6. Comparison of fluorescence and Ramon monitoring techniques.

Aigsueiug eoueosaionu

167

V1.3.2. Efiect of Light Intensity

As mentioned, the Raman monitoring studies have indicated that the final
conversion value for the cationic photopolymerization of DVE-3 is dependent on the rate
of the reaction. Experiments were run at many different excitation intensities to
determine the effect of light intensity on the reaction profile. In order to increase the
stability of the laser output, the 363.8 nm laser beam was attenuated with a neutral density
filter instead of adjusting the laser power. The excitation intensity was measured with the
power meter directly before the lens used to focus the beam in to the sample. A range of
optical densities form 0.1 to 3.0 were used to adjust the light intensity.

The light intensity studies indicated that the rate of reaction for the cationic photo-
polymerization of DVE-3 photosensitized by anthracene is extremely sensitive to the
excitation intensity. Reactions performed at different intensities ranged from under 1 sec
to several minutes to reach the final conversion. Optical densities under 0.5 displayed
reaction rates that resulted in extreme temperatures and degradation of the polymer. In
this case, the polymer loses optical clarity and the Raman spectra changes drastically.
The final conversion can not be determined since all the Raman peaks are lost including
the internal standard. If optical densities greater than 3.0 were used to attenuate the laser
beam, the reaction rate was too slow to compare with reaction rates needed for most
applications. Therefore, neutral density filters between 1.0 and 3.0 were used to produce
reaction rates which were able to be maintained isothermal and easily monitored by the

Raman system.

168

V1.3.3. Efiect of Temperature

Isothermal reactions were performed at a series of temperatures between 20 and
50 °C to determine an overall activation energy for the polymerization reaction. A
maximum temperature of 50 °C was used to ensure that no thermal initiation occurred
during the photopolymerization. An average initial reaction rate was determined for each
temperature by measuring the time required to reach a specific conversion. A
conversions of 20 percent were chosen, and the activation energy was determined from an
Arrhenius fit to the data.

Figure V1.7. shows Raman conversion‘profiles for the cationic polymerization of
DVE-3 photosensitized with 1.0 x 10’2 wt % anthracene and 1 wt % initiator controlled at
temperatures form 20 to 50°C. As expected the reaction rate and the total conversion
increased with increasing temperature. Based upon an Arrhenius fit, the overall

activation energy (ER) for the cationic polymerization was determined to be 25.1 i 6.1

kJ/mol at 20 percent conversion. Since the Raman method measures the overall reaction

rate, these values represent an overall activation energy which includes initiation,
propagation, and termination.

ER=E,,+E,-ET Eq.l.

For equation 1 to be valid, the active center production must continue throughout

the reaction. Based upon previous measurements of the kinetic constants for

photosensitization,8 the photosensitizer is only approximately 40 percent consumed at 20

percent conversion for each temperature. Therefore, active centers should be produced

169

   

 

 

v ' v
_ ------- v‘ "V'm-.-
0.8 ’vv6;'1v “
776' ‘ ‘A“““‘ m‘
I
. y" n A . '93? -
C " ‘
.9 0.6 .. v A“
‘3 I'- “
2
8 . i
U I I
E 0.4 - Temperature
8 ' 20 0C
§ . . 30 °C
.5 J ‘ 40 CC
0.2 ' 50 CC
0.0 - A1 1 1 I T I I I I I
O 10 20 30 4o 50

time (S)

Figure VI. 7. Roman conversion profiles for isothermal reactions ranging from 20 to
50 "C.

170
throughout the photopolymerization reaction for all temperatures. The activation energ ,,

determined from Raman spectroscopy also agreed well with the activation energy

determined from Photo-Differential Scanning Calorimetry (PDSC) studies (ER = 26.5 i

3.2kJ/mol).9

V1.3.4 Initiation Rate Constant

In these photopolymerizations, active cationic centers are produced through a
photosensitization reaction in which an electron is transferred from an excited state
anthracene molecule to the initiator.8 Although the reaction proceeds by a multi-step
photochemical mechanism,° the rate of generation of active centers may be described by

the following simple expression:

Rate of active center generation = k,[A][l] Eq. 2.

where [A] is the anthracene concentration, [I] is the initiator concentration, and k,- is an
initiation rate constant which accounts for a number of photophysical steps including
excitation, intersystem crossing, exciplex formation, and electron transfer.° According to
this convention, k, is dependent on the incident light intensity, and must be adjusted for
each initiation intensity. In our reactions, the initiator (diaryliodonium salt) concentration
greatly exceeds the anthracene concentration and can be assumed to be constant during
the reaction.8 Therefore a pseudo-first-order kinetic constant, It," = k,[l], may be used to
describe the rate of generation of active centers. The above analysis is based upon the

assumptions that one active center is produced per anthracene and initiator molecules'o‘”

171
and that all reactive centers are capable of propagating the reaction, as indicated by

previous fluorescence measurements.l

The overall rate of change of the active center concentration may be derived by
combining the rate of generation by photosensitization with the rate of consumption by
termination:

d[M*]
dt

 

= kj[A]—k,[M+] Eq. 3.

This equation may be integrated (with the initial condition [Mo+] = 0 and the

exponentially decreasing anthracene concentration) to yield the following equation:

[M1] = [4012—57- (e-kh _ e-kr') Eq- 4.

In this equation, the termination rate constant, k, was determined in earlier work.9 The
initiation rate constant, k.- was determined from solving the photophysical rate
expressions, using the initiation intensity and the anthracene absorption coefficient as
described elsewhere.8 For the Raman experiments, the incident light intensity used in this
calculation was determined by measuring the power of the initiation laser beam (363.8
nm). Using the incident light intensity value and the cross sectional area of the laser
beam, k,-t was determined to be 6.21 x 10'2 s'1 at 25 C by solving the corresponding set of
rate expressions.8

The activation energy for the photosensitization reaction was determined from
fluorescence measurements of the photosensitization reaction between anthracene and the
diaryliodonium salt. As previously described, the rate constant, k,‘, was determined for

several different temperatures and fit to an Arrhenius equation to yield an average

172

activation energy, E], of 28.5 i 1.6 lemol. Using this activation energy, the temperature-

dependence of k,’ was described. Based upon the values of k,, k,‘, E, and A0, equation 4
was used to obtain profiles of the active center ([M]) concentration as a function of time

for each of the Raman reactions.
V1.3.5 Propagation Rate Constant

As previously mentioned, the instantaneous conversion can be determined with
the Raman technique, since the vinyl bond conversion as a function of time is measured
directly. The conversion profile was obtained by setting the initial Raman spectra equal
to zero conversion and calculating conversion. for each of the following time slices. The
rate of polymerization (RP) can be determined from the derivative of the conversion
profile. As shown in Figure V1.7., the conversion data was fit with a triple exponential
expression. Since noise in the actual data can cause large variations in the derivative, the
exponential fit of the conversion profile was used to calculate the rate of polymerization.

According to kinetic analysis, the rate of propagation is given by a propagation
rate constant multiplied by the active center concentration and the monomer

concentration as shown below:

kk,‘ _., _,
RP: d =kp[MIM+]=[AO]k'”_ k: (e *I —e *v )[M] Eq. 5.

 

 

If a number of propagation centers with different reactivities exists, this equation may be
used to fit the polymerization data to yield an apparent propagation constant that

comprises all types of propagating centers (e.g. ion pairs, separated ions, etc.).9 In this

173
chapter, this approach will be used to interpret changes in the value of the apparent

propagation constant, k,,.

The instantaneous vinyl bond concentration ([M]) can be calculated from the
Raman conversion profile and the initial vinyl bond concentration using the following

equation:
[M ] = [M0] X (1 - fractional conversion) Eq. 6.

Equation 6 gives the vinyl bond concentration in mol/l of vinyl bond, since [M] represents
mols of double bonds per liter. Figure V1.8. verifies the importance of temperature
during the cationic photopolymerization of DVE-3, as characterized by the increase in
conversion and rate of the reaction with temperature.

Profiles of the rate of polymerization (propagation) plotted versus vinyl bond
conversion obtained by using equations 6 are shown in Figure V1.9. The figure illustrates
that as the temperature is increased the maximum value of R,, increase accordingly. As
the reaction takes place, the value of RI) increases to a maximum and then decreases with
further vinyl bond conversion. Figure V1.9. also demonstrated that as the rate of reaction
increases with increasing temperature, the limiting conversion for the cationic
photopolymerization also increases. This indicates that the final conversion of the
reaction is a function of the rate of the reaction. Finally, the values for R,, determined
from Raman spectroscopy agrees very well the values determined by the PDSC

technique.9

0.08 -

0.07 - j

0.06 J

0.05 -

Rp (mol/l‘s)
O
2
l

 

 

174

Temperature
20 0C

 

--- 30°C
. ..... 40°C
,‘. --—--50°C

Figure V1.8. Rate of reaction for several diflerent temperatures versus time for a
cationic photopolymerization of DVE-3 with 1 wt % initiator and 10"? wt % anthracene.

175

 

0.08— __
‘ .’.’-’ .~ ‘°\
/ °\.
0'07 ' / O \ ~ \ Temperature
'1 l.’ '\ 2000
0.06 - .,- ‘-\ - - - 30°C
‘ I '\. ---- 40°C

 

 

 

0.0 0.2 0.4 0.6 0.8
vinyl bond conversion

Figure V1.9. Overall rate of propagation versus conversion for DVE-3 photosensitized
with 10'2 wt % anthracene and 1 wt % initiator.

176

For each Raman experiment, a profile of the quantity k,,[M+] may be obtained
using the empirical profiles of RI, and [M]. Based only of the Raman data, this product

may be determined using the following version of equation 5.

R
kp[M+] :[—ML] Eq. 7.

Profiles of k,,[M"] obtained in this manner are shown in Figure V1.10. As shown in the
figure, kp[M] initially increases sharply, and then levels off at a plateau value between
0.001 and 0.01 5'1 depending upon the reaction temperature (as temperature is increased a
higher value of k,,[M*] is observed). Finally, the value of k,,[M+] decreases as a limiting
conversion is reached. The shape of the profiles in Figure [V1.10. indicates that the
apparent propagation rate constant k,, changes as the reaction proceeds. If k,, remained
constant throughout the reaction, the change in k,,[M+]would be governed by the active
center concentration. Therefore, the k,,[M*] value would increase monotonically until it
reached a maximum when all of the photosensitizer is consumed.

More information about the behavior of k,, may be obtained by coupling the
Raman data with the k; and k, values discussed previously. Profiles of the apparent
propagation rate constant, k,, were obtained by substituting the values of [A0], ks, and k,
along with the empirically determined profiles of [M] and R,, into equation 5. Figure
VI.11. shows kp as a function of reaction conversion. The curves exhibit a similar
characteristic shape to the k,,[M*] profiles of Figure V1.10., although the plateau value is
much higher due to the small active center concentration. The similarity in shape

indicates that relative changes in kp are much larger than those for [M]. The final

177

 

 

 

 

0.01 1
5: 15-3 .4 4'" Temperature 3;
3i :1 I —-— 20 oc 4;
i it"? —-— 30 cc 3
3“ j :7 ——----~—~ 40°C 1
l i. —V— 50 °C ‘
JO .5
7 .
1E-4 - I A
. I ,
7 f I ' I ' I ' I
0.0 0.2 0.4 0.6 0.8

vinyl bond conversion

Figure V1.10. k,,[MV versus conversion for the cationic photopolymerization of DVE-3
at temperatures ranging from 20 to 50 "C.

178

 

 

 

Temperature ‘

1 A. —-— 20 0C, a

: _.._ 30 00 t

: ~~-+-40°C :

1 —v— 50 00 I

. 1

1
0.1 . , . l f f . r
0.0 0.2 0.4 0.6 0.8

vinyl bond conversion

Figure V1.11.

k,, versus conversion of the photopolymerization of DVE-3 at various
temperatures.

179
magnitude obtained for kp values were between 10 and 40 l/mol*s, which also compare

very satisfactorily with values from the PDSC data ( 5-30 l/mol"‘s).9

As described in a previous contribution,9 a possible explanation for the initial
large increase in the apparent rate constant for propagation is a change in the reactivity of
the carbocation species due to the proximity of the counterion. In this stage of the
reaction, polymerization leads to a large local increase in viscosity which reduces
diffusional mobility. However, the active carbocation retains considerable mobility by
propagating with double bonds of different monomer units. This type of mobility of the
active center has been termed “reaction diffusion.” In contrast, since the counterion
cannot undergo similar diffusion, the large hexafluoroantimonate counterion experiences
a decrease in mobility as the viscosity increases. During this process, the counterion and
carbocation may become separated and since separated ions are orders of magnitude more
reactive than ion pairs,12 the apparent value of kp increases dramatically during the early
stages of the reaction. This type of “reaction diffusion” is well known for free radical

systemsl3"4

and may aid in explaining the shape of the curves in Figures V1.10. and
V1.11. The final decrease in kp arises from active center trapping during these highly
crosslinked polymerizations. A limiting conversion at which the reaction ceases is
observed despite the presence of active centers. Similar profiles were obtained with the
PDSC technique giving an independent verification of the two techniques. As previously
mentioned, the above analysis is based upon the assumptions that one active center is

10.”

produced per anthracene and initiator molecules and that all reactive centers are

capable of propagating the reaction.1

180

V1.4. Conclusions

In this contribution Rarnan spectroscopy was used to monitor cationic photo-
polymerizations of a divinyl ether. Using the Raman technique a complete spectra of
approximately 1000 cm'l can be monitoring in situ during the cationic photo-
polymerization. Since molecular vibrations associated with the vinyl bonds are
measured, the Raman technique offers a direct method for monitoring vinyl bond
conversion during a polymerization reaction. The conversion profiles provided by the
Raman technique were used to characterize the reaction kinetics and evaluate
polymerization rate constants.

The Raman experiments were used to determine kinetic constants for a series of
unsteady-state divinyl ether polymerizations at different temperatures and light exposure
intensities. The kinetic constant for propagation was obtained from the Raman
experiments in conjunction with previously determined photosensitization rate constants.
As expected, an increase in the initiating light intensity or an increase in the reaction
temperature resulted in an increased reaction rate. This increase in reaction rate was
found to result in higher final conversions for the cationic photopolymerization.

An apparent propagation rate constant that comprises all types of propagating
centers (e.g. ion pairs, separated ions, etc.) with different reactivities was determined
from the Raman reaction profiles. The apparent propagation constant, k , initially
increases sharply, and then levels off at a plateau value (as temperature is increased a

higher value of k,, is observed). Finally, the value of k,, decreases as a limiting conversion

181
is reached. This profile shape indicates that the apparent propagation rate constant k,,

changes greatly as the reaction proceeds. The initial large increase in the apparent rate
constant for propagation may be explained by a change in the reactivity of the carbocation
species due to the proximity of the counterion. During the polymerization, the counterion
and active carbocation become separated due to a decrease in mobility of the large
hexafluoroantimonate counterion as the system viscosity increases. Since separated ions
are orders of magnitude more reactive than ion pairs,l the apparent value of k,, increases
dramatically during the early stages of the reaction. The final decrease in kp arises from
active center trapping during these highly crosslinked polymerizations. Therefore a
limiting conversion at which the reaction ceases is observed despite the presence of active
centers.

The Raman conversion profiles for propagation were compared with Photo—
Differential Scanning Calorimetry profiles from earlier experiments. The two techniques
displayed good agreement for the overall activation energy and the propagation rate
constant, k , verifying the data obtained by each techniques. The profiles of propagation
as functions of conversion provide important insight into the nature of these high-speed
cationic photopolymerizations. Understanding the kinetics of these cationic photo-
polymerizations is important due to the increasing number of applications for rapid,

solvent-free curing of polymer films.

10.

11.

12.

182

V1.5. List of References

. E. W. Nelson, T. P. Carter, A. B. Scranton, “Fluorescence Monitoring of Cationic

Photopolymerizations: Divinyl Ethers Photosensitized by Anthracene Derivatives,”
Macromolecules, 27,1013 ( 1994).

J.V. Crivello, “Cationic Polymerization -- Iodonium and Sulfonium Salt
Photoinitiators,” Adv. Polym. Sci., 62, 1 (1984).

. W.R. Watt, “Photosensitized Epoxides as a Basis for Light-Curable Coatings," Epoxy

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

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

P.C. Painter, M.M. Coleman, and J.L. Koeing, The Theory of Vibrational

Spectroscopy and its Applications to Polymeric Materials, John Wiley & Sons, New
York, 1982.

H.W. Siesler, K. Holland-Moritz, Infrared and Raman Spectroscopy of Polymers,
Practical Spectroscopy Series, Marcel Dekker, Inc.: New York, NY, Vol. 4, 1980.

R.M. Silverstein, G.C. Bassler, and T.C. Morrill, Spectrometric Identification of
Organic Compounds, 5T“ ed., John Wiley & Sons Inc., NY, 113, 1991.

E.W. Nelson, T.P. Carter and AB. Scranton, “The Role of the Triplet State in the
Photosensitization of Cationic Polymerizations by Anthracene,” J. Polym. Sci; Part
A: Polym. Chem, 33, 247-256 (1995).

E.W. Nelson, J .1.. Jacobs, A.B. Scranton, K.S. Anseth, and C.N. Bowman, “Photo-
differential Scanning Calorimetry Studies of Cationic Polymerizations of Divinyl
Ethers,” Polymer, in press.

J. L. Dektar, and NP. Hacker, “Photochemistry of Diaryliodonium,” J. Org. Chem,
55, 639-647, (1990).

R. J. DeVoe, M. R. Sahyun, E. Schmidt, N. Serpone, and D. K. Sharma, “Electron
Transfer Sensitized Photolysis of ‘onium Salts,” Can. J. Chem. 66, 319 (1988).

J. P. Kennedy and E. Marechal Carbocationic Polymerization, John Wiley & Son,
New York, 1982.

183

13. K. S. Anseth, C. M. Wang and C. N. Bowman, “Kinetic Evidence of Reaction
Diffusion during the Polymerization of Multi(meth)acrylate Monomers,” Macro-
molecules, 27, 650 (1994). A

14. K. S. Anseth and C. N. Bowman, “Reaction Diffusion Enhanced Termination in
Polymerizations of Multifunctional Monomers,” Polymer Reaction Engineering, 1(4),
499-520, (1992-93).

CHAPTER VII. IN SI T U PHOTO-DIFFERENTIAL SCANNING

CALORIMETRY: CURE MONITORING AND KINETICS

VII.1. Introduction

Photo-differential scanning calorimetry (PDSC) experiments were used to
determine effective kinetic constants for propagation and termination for a series of
unsteady-state divinyl ether polymerizations at different temperatures and light
intensities. For these cationic photopolymerizations the reaction rate and limiting
conversion were both found to increase as the reaction temperature was increased. At all
temperatures the profile for the propagation rate constant, k, exhibited a dramatic
increase at the start of the reaction, plateaued at a value between 5 and 30 1/mol*s
(depending upon temperature), and then decreased as the reaction reached a limiting
conversion due to trapping of the active centers. The effective termination rate constant,
k,, was very low, with active center lifetimes approaching twenty minutes. Finally, the
overall activation energy for polymerization was found to be 26.5 i 3.2 lemol.

Photo-differential scanning calorimetry (PDSC) offers a simple method to
characterize the kinetics of photopolymerization reactions. Because the polymerizations
are highly exothermic, the reaction rate may be measured by monitoring the rate at which

heat is released from the polymerizing sample. Therefore, the profiles of reaction heat

184

185
versus time provided by PDSC may be used to characterize the reaction kinetics and

evaluate polymerization rate constants.' Several authors have used PDSC to characterize
free radical photopolymerizations"5 however application of the technique to cationic
photopolymerizations of divinyl ethers is more problematic. First, the polymerizations of
divinyl ethers proceed very rapidly, and care must be taken to ensure that the reaction rate
does not exceed the time resolution of the instrument. Since the DSC response time is on
the order of 2-3 seconds,°‘7 relatively low light intensities must be used to increase the
reaction time for the divinyl ether photopolymerizations.8 Second, photopolymerizations
of divinyl ethers are highly exothermic, therefore to maintain isothermal conditions
during the reactions very small sample sizes must be used (0.5 - 1.5 mg). If larger sample
sizes are used, the heat evolved during the reaction generally cannot be removed on the
time scale of the reaction.

A series of PDSC experiments were performed on cationic photopolymerizations
of a divinyl ether initiated with a diaryliodonium hexafluoroantimonate salt photo-
sensitized by anthracene. Isothermal PDSC experiments were performed for a series of
reaction temperatures and light intensities. The profiles were used in conjunction with
previously determined photosensitization rate constants9 to determine the kinetic constant
for propagation. The kinetic constant for termination was determined from a series of
unsteady-state “dark-cure” reactions at different temperatures and light exposure times.
The resulting profiles for the propagation and termination rate constants as functions of

conversion provide important insight into the nature of the cationic photopolymerizations.

186

VII.2. Experimental

VII.2.I. Materials

In these studies, 3,6,9,12-tetraoxatetradeca-l,l3-diene (DVE-3) (ISP) was used as
the monomer. The monomer was dried over molecular sieves to remove any traces of
water. As in previous spectroscopic studies,9"" a commercially available initiator
(UV9310C - GE Silicones) was used. The initiator formulation includes 5-10 wt % linear
alkylate dodecylbenzene, ~50 wt % 2-ethyl-1,3-hexanediol, and ~50 wt % bis-
(4-dodecylphenyl) iodoniumhexafluoroantimOnate. Initiator concentrations specified in
the remainder of this chapter correspond to the total UV9310C concentration. The
photosensitizer, anthracene, was purchased from Aldrich Chemical Company and was
used as received. A representative reaction sample contained 10'2 wt % anthracene and 1

wt % initiator in DVE—3.

VII.2.2. PDSC Experiments

The PDSC experiments were conducted using a differential scanning calorimeter
equipped with a photocalorimetric accessory (Perkin Elmer, DSC-DPA 7). The
photocalorimeter accessory included transfer optics to produce full beam ultraviolet light
of varying intensity and a monochromator to produce light of a given wavelength. The

initiation light source was a 100 watt mercury arc lamp used in conjunction with neutral

187
density filters (Melles Griot) to control the intensity of the incident light. The strong

mercury line at 365 nm was used to excite the anthracene photosensitizer. The
polymerization reactions were run isothermally by means of a refrigerated recirculating
chiller (NESLAB, CFT-25) maintained at 5 °C. Small sample sizes (0.5 - 1.5 mg) were
required in order to limit the total heat released during the polymerization reaction so that
isothermal conditions could be maintained. The samples were placed in uncovered
aluminum DSC pans and cured with UV light intensities varying from 0.4-1.6 mW/cm2 at

various reaction temperatures.

VII.3. Results and Discussion

VII. 3. 1 . Effect of Light Intensity

Experiments were performed at several different excitation intensities in order to
determine the effect of light intensity on the reaction profiles. The light intensity at 365
nm was adjusted by placing a neutral density filter after the excitation monochromator. A
range of optical densities from 0.5 to 2.0 were required to adjust the light intensity. At
optical densities less than 0.5, the reaction proceeded very rapidly which prevented
isothermal conditions. Optical densities greater than 2.0 resulted in very slow reactions
limiting the number of reactions that could be completed in a reasonable time.

Figure VII.l. shows the relationship of reaction rate and light intensity. As

expected, an increase in the initiating light intensity resulted in an increased reaction rate

188

 
 
  
 
 
  

1.1 x 10'2 wt %Anthraoene, 1.0 wt % Initiatorin wee
x = 365 nm, temperature = 30°C
Each curve is the average euwe for a certain light intensity

— Light Intensity = 0.56 chm2
-------- Light Intensity=0.19chm2

3181396 energy (W/g)
N

 

 

01“?“
1 m I 1 1 x T f I T I u I
0 2 4 6 a 10 12 14
time(min)

Figure VIIJ. PDSC exotherms for the cationic polymerization of DVE—3 at diflerent
light intensity values.

189
as indicated by the exotherm reaching its maximum more rapidly. For example, at 0.56

mW/cm2 the heat flux reached the maximum at 2.5 i 0.7 minutes with an average peak
area of 660 i 59 J/g, while for initiating intensities of 0.19 mW/cm2 the maximum was
reached at 5.4 i 0.7 minutes and had an average area of 553 :t 14 J/g. Each of the curves

in Figure VIII. is an average of several profiles obtained at each light intensity, therefore
the average curves are slightly shorter and wider than the original data. However, the
average curves represent the general shape and provide the correct average time at the

peak maximum and the average area.

VII.3.2. Effect of Temperature

Reactions were performed at a series of temperatures between 20 and 50 "C in
order to determine an overall activation energy for the polymerization reaction. A
maximum temperature of 50 °C was used to ensure that no thermal initiation occurred.
An average initial reaction rate was determined for each temperature by measuring the
time required to reach 20 % conversion. This conversion was chosen since it represents
the maximum conversion for the lowest temperature experiment. The activation energy
was determined from an Arrhenius fit to the data.

Figure VII.2. shows PDSC exotherms for the cationic polymerization of DVE-3
photosensitized with 1.0 x 10'2 wt % anthracene and 1 wt % initiator controlled at
temperatures from 20 to 50 °C. As expected, the reaction rate and total conversion

increased with increasing temperature as illustrated by the fact that the exotherms reach

190

 

0.06 -
u i’o\'\.
0.05 - 1" ‘1
j 1, Temperature
1 I .
o 04 1’ t 20 cc
- ii ‘1‘ ----- 30 00
o ‘ .
'S 0 03 J ,' i. ------- 50 °C
0 .
é

 

 

 

time (min)

Figure VII.2. PDSC exotherms for the cationic polymerization of DVE-3 at temp-
eratures ranging from 20 to 50 °C.

191
the maximum in a shorter time and exhibit larger integrated heat as the temperature is

increased. Based upon an Arrhenius fit, the overall activation energy for the cationic

polymerization was determined to be 26.5 i 3.2 kJ/mol. Since the PDSC measures the

total heat of reaction, this value represents an overall activation energy including
initiation, pr0pagation, and termination.

ER=EP+El-ET Eq. 1.
For equation 1 to be valid, the active center production must continue throughout the
reaction. Based upon previous measurements of the kinetic constants for photosensitiza-
tion,9 the photosensitizer is not completely consumed until well after the time for the peak
maximum, therefore the active center concentration continues to increase at the peak

maximum (the point where the reaction rate was measured). Therefore, equation 1 can be

used to represent the overall activation energy for the photopolymerization reaction.
VII.3.3. Termination Rate Constant

To determine the termination rate constant, k,, for the polymerization of DVE-3,
dark cure experiments were performed. In these studies the polymerization samples were
exposed to the initiating light for a short period of time, after which the light source was
removed. The remaining cure was monitored with no exposure to the light source, and
the reaction rate was found to decay exponentially. Samples that were exposed for 2
minutes or less had total double bond conversions of less than 5 percent. Since the

change in the vinyl bond concentration during the dark cure was small, it was assumed

192
that the reduction in the polymerization rate was caused by the decline in the number of

active centers due to termination and not due to depletion of monomer.
The PDSC exotherms were fit with an exponential decay curve from the point at
which the excitation source was removed with an average exponential time constant of

17.3 i 0.4 minutes. Therefore, the termination rate constant, k,, was calculated to be
0.0010 1' 0.0002 5". These results illustrate that the cationic centers exhibit a relatively

long lifetime and relatively small termination rate constants compared to free radical
systems. This arises from the fact that, in contrast to free radicals, cationic centers do not
terminate with one another. Termination in cationic systems typically occurs due to
combination with the counterion or other nucleophilic species. As will be shown later,
the termination rate constant for the polymerization reaction was determined to be

approximately 5 times lower than the initiation rate constant.

V11.3.4. Initiation Rate Constants

In these photopolymerizations active cationic centers are produced through a
photosensitization reaction in which an electron is transferred from an excited state
anthracene molecule to the initiator.9 Although the reaction proceeds by a multi-step
photochemical mechanism,° the rate of generation of active centers may be described by

the following simple expression:

Rate of active center generation = k,[A][I] Eq. 2

where [A] is the anthracene concentration, [I] is the initiator concentration, and k, is an

initiation rate constant which accounts for a number of photophysical steps including

 

193
excitation, intersystem crossing, exciplex formation, and electron transfer.9 According to

this convention, k, is dependent on the incident light intensity. In these reactions the
initiator (diaryliodonium salt) concentration greatly exceeds the anthracene concentration
and can be assumed to be constant.9 Therefore a pseudo-first-order kinetic constant, k,’ =
ki[l], may be used to describe the rate of generation of active centers. The above analysis
is based upon the assumptions that one active center is produced per photosensitizer and
initiator molecules and that all reactive centers are capable of propagating, as indicated by
previous fluorescence measurements.'0

The overall rate of change of the active center concentration, [MI], may be derived
by combining the rate of generation by photosensitization with the rate of consumption
by termination:

d[M*:
dt

 

 

=k:[A]-k,[M+] Eq. 3.

This equation may be integrated (with the initial condition [M5] = 0 and the

exponentially decreasing anthracene concentration) to yield the following equation:
+ k: -k-°t -kr
[M ]=[Ao]k—_—F(e ' '8 ') Eq.4.
t i

In this equation, the termination rate constant, k, was evaluated above while the initiation
rate constant, k, may be determined from solving the photophysical rate expressions as
described in earlier work.” For the PDSC experiments the incident light intensity used in
this calculation was determined by measuring the energy absorbed by graphite discs

placed in the sample and reference cells. Using the incident light intensity value, k: was

194
determined to be 4.03 x 10'3 s'| at 25 °C by solving the corresponding set of rate

expressions.9

The activation energy for the photosensitization reaction was determined from
fluorescence measurements of the photosensitization reaction between anthracene and the
diaryliodonium salt. The rate constant, k, was determined for several different
temperatures and fit to an Arrhenius equation to yield an activation energy, E,, of 28.5

$1.6 kJ/mol. Using this activation energy, the temperature-dependence of k; was

described. Based upon the values of k,, k:, E,, and A0, equation 4 was used to obtain
profiles of the active center ([M]) concentration as a function of time for each of the

PDSC reactions.
VII.3.5. Propagation Rate Constant

As previously mentioned, the rate of polymerization (R) is directly proportional
to the rate at which heat is released from the polymerizing sample and is therefore
proportional to the height of the DSC exotherm measured in W/g. For a polymerization

rate in units of mol(double bonds)/(L sec), the proportionality constant is the density (p -
the density of the reaction mixture) divided by the heat of polymerization, AHP, and the

instantaneous rate of polymerization may be calculated from the following relationship:

d[ ] height of exotherm (W / g) * p

R =_=

P dt AH Eq. 5.

P

 

195
According to kinetic analysis, the rate of propagation is given by a propagation rate

constant multiplied by the active center concentratiOn, [M'], and the vinyl bond

concentration, [M], as shown below:

M + k kt. -k‘t -kt
R,,=£[;1-t-l=k,,[M][M1=[Ao]-k-i’_—E(e ' —e )[M] Eq.6.

If a number of propagating centers with different reactivities exist, this equation may be
used to fit the polymerization data to yield an apparent propagation constant that
comprises all types of propagating centers (e.g. ion pairs, separated ions, etc.). In this
paper we will use this approach to interpret changes in the value of the apparent
propagation constant, kp.

The vinyl bond concentration ([M])‘and the conversion may be calculated from
the integral of the reaction rate profile (equation 5). Therefore, the instantaneous double

bond concentration may be calculated form the PDSC profile using the following

 

 

equations:
exotherm area (J / )*
[M1= [M.]-[ M g P] 5.7.
P
Conversion = exotherm area (I l g)* 0 Eq. 8.
AH ”[MO]

Equation 7 gives the vinyl bond concentration in mol/l when p is the density of the

reaction mixture. The heat of reaction for the vinyl bond was determined to be 75 kJ/mol
by reacting a monovinyl ether to high conversion. A complete profile of the double bond
concentration or conversion may be obtained by determining the area under the DSC

curve as a function of time and applying equation 7 or 8.

196
For each PDSC exotherm, a profile of the quantity kp[M'] may be obtained using

the empirical profiles of Rp and [M]. Based only on the PDSC data, this product may be

determined using the following version of equation 6.

 

kP[M+]= R” Eq. 9.
[M]

Profiles of the rate of polymerization (propagation) obtained by applying
equations 5 and 8 to a set of PDSC exotherms are shown in Figure VII.3. The figure
illustrates that as the temperature is increased the maximum value of RP increases
accordingly. Furthermore, the conversion at the peak maximum is shifted to higher
values as the temperature increases. Therefore Figure VII.3. verifies the importance of
temperature during the cationic photopolymerization of DVE-3, as characterized by the
increase in conversion and rate of the reaction with temperature.

Profiles of kle‘] obtained in this manner are shown in Figure VII.4. As shown in
the figure, kP[M’] initially increases sharply and then levels off at a plateau value between
0.001 and 0.01 s'l depending upon the reaction temperature (as temperature is increased a
higher value of kp[M’] is observed). Finally, the value of kp[M‘] decreases as a limiting
conversion is reached. The shape of the profiles in Figure VII.4. indicates that the
apparent propagation rate constant k,» changes as the reaction proceeds. If k, remained
constant throughout the reaction, this change in kp[M’] would be governed by the active
center concentration. Therefore, the kP[M*] value would increase monotonically until it

reached a maximum when all of the photosensitizer is consumed.

0.06 -

0.05 -

 

 

197

 

I' ~
1' \.
I' ‘.
1' ‘.
D’.‘. \I
I ‘.
Temperature
1' \
." '\
,’ ‘\ 20 0C
‘I '\
.\
\. -------- 30 cc
\
‘\.
x -------- 40 cc
\0
\l
\
------—- 50 cc
‘1
\I
\-
\O
‘0
\Q
\l
‘0
‘0
\-
\
.\
\I
‘l
‘0
\
\ ‘\
'\
K
'\
‘\
".
\-
‘-
‘0
\-
"~‘
.\

percent conversion

Figure VII.3. Overall rate of propagation versus conversion for DVE-3 photosensitized
with 1.0 x 10'2 wt % anthracene and 1 wt % initiator.

198

I All.

 

 

 

 

 

IE-3 E 1'
: .0
: 1fl "r"
A - :' :
'7 15-4 1 ‘ l
3 : Temperature '
g : 20°C
3:: : ----- 30°C
15-5 5 . """"" 40 °C
1 1a ------- 50 0C
1E-6 ;

 

o 10 20 30 40 so 60 70 30 90
percent conversion

Figure VII.4. k, [M] versus conversion for the cationic photopolymerization of DVE-3
at temperatures ranging from 20 to 50 "C.

199
More information about the behavior of kp may be obtained by coupling the PDSC

data with the k... and k, values discussed previously. Profiles of the apparent propagation
rate constant, kp were obtained by substituting the values of [A0], k,’, and k, along with the
empirically determined profiles of [M] and RP into equation 6. Figure V115. shows kp as a
function of vinyl bond conversion. The curves exhibit a similar characteristic shape as
the kp[M*] profiles of Figure VII.4., although the plateau value is much higher due to the
small active center concentration. The similarity in shape indicates that relative changes
in k, are much larger than those for [M]. The final magnitude obtained for kp values were
between 5 and 30 l/mol*s, which compare satisfactorily with values from the literature.”
One possible explanation for the initial large increase in the apparent rate constant
for propagation is a change in the reactivity of the carbocation species due to the
proximity of the counterion. In the early stage of the reaction, polymerization leads to a
large local increase in viscosity which reduces diffusional mobility. The active
carbocation retains considerable mobility through propagation since the cation effectively
moves through the reaction mixture by reacting with new monomers. This type of
mobility has been termed “reaction diffusion” and may become the dominant mechanism
for mobility of the active center in highly crosslinked polymerizations."l2 However, the
large hexafluoroantimonate counterion experiences a decrease in mobility as the viscosity
increases since it cannot diffuse through the reaction mixture in a similar manor. During
this process, the counterion and carbocation become separated, and since separated ions
are orders of magnitude more reactive than ion pairs,13 this would lead to a large increase

in the apparent propagation rate constant, kp. This type of reaction drffusron’ rs well

200

‘v'own-o..-‘-..
.OO’" .-'.'---.~.
'.' -V-.-
.r-'- ‘fi.-.
5 §
’0 '\
c ....................
'n .....-o....... . .. ‘
-.” ..---' -. .‘
l" .\
......
oo-.. ‘
----.- ---------_----- .‘.~
..........
§
~
~
-
~
‘
\
s

 

 

 

 

 

10 ‘
Ii

1 1 f g.
”a? I E E
*— 1 :
E q 5 E
3 101 Temperature
xa 5 20 oC
i -------- 30 00 i
10‘2 E . ........... 40 °C g
' ----------- 50 0C 5

103 ' I I U I U I ' I I I Tr I I I v I U I t I

010 20 30 40 so 60 70 so 90
percent conversion

Figure VII.5. kp versus conversion for the photopolymerization of DVE-3 at various
temperatures.

201
known for highly crosslinked free radical systems“J2 and may aid in explaining the shape

of the curves in Figures VII.4. and VII.5. The final decrease in kp arises from active
center trapping during these highly crosslinked polymerizations. Therefore a limiting
conversion at which the reaction ceases is observed despite the presence of active centers.
As previously mentioned, the above analysis is based upon the assumptions that one
active center is produced for each photosensitizer molecule consumed and that all

reactive centers are capable of propagating '0
VII.4. Conclusions

In this contribution photo-differential scanning calorimetry (PDSC) was used to
monitor cationic photopolymerizations of a divinyl ether. The PDSC method offers a
direct method for evaluating the heat generated during a polymerization reaction. The
heat of reaction profiles provided by PDSC were used to characterize the reaction kinetics
and evaluate polymerization rate constants. However, the reaction system was extremely
rapid and highly exothermic, making DSC measurements particularly challenging. To
maintain isothermal reaction conditions, low light intensities and very small sample sizes
(0.5 - 1.5 mg) were used to increase the reaction time for the divinyl ether
photopolymerizations.

The PDSC experiments were used to determine kinetic constants for a series of
unsteady-state divinyl ether polymerizations at different temperatures and light exposure
times. Kinetic constants for propagation and termination were obtained from the PDSC

experiments in conjunction with previously determined photosensitization rate constants.

202
As expected, an increase in the initiating light intensity or an increase in the reaction

temperature resulted in an increased reaction rate. This increase in reaction rate was

found to result in higher final conversions as determined from the total heat of reaction.
To determine the termination rate constant, k,, PDSC dark cure exotherms were fit

with an exponential decay from the point at which the excitation source was removed (the

exponential time constant was 17.3 :t 0.4 minutes). These results illustrate that the

cationic centers exhibit a relatively long lifetime and relatively small termination rate
constants compared to free radical systems, arising from the fact that cationic centers do
not terminate by combination with themselves.

An apparent propagation rate constant that comprises all types of propagating
centers (e.g. ion pairs, separated ions, etc.) with different reactivities was determined
from the PDSC reaction profiles. The apparent propagation constant, kp, initially
increases sharply and then levels off at a plateau value (as temperature is increased a
higher value of kp is observed). Finally, the value of k, decreases as a limiting conversion
is reached. This profile shape indicates that the apparent propagation rate constant kp
changes greatly as the reaction proceeds. The initial large increase in the apparent rate
constant for propagation may be explained by a change in the reactivity of the
carbocation species due to the proximity of the counterion. During the polymerization,
the counterion experiences a decrease in the mobility of the large hexafluoroantimonate
counterion as the viscosity of the system increases. However, the active carbocation
retains diffusional mobility by reacting with vinyl bonds and may lead to separation of
the two species. Since separated ions are orders of magnitude more reactive than ion

pairs,'3 the apparent value of k, increases dramatically during the early stages of the

203
reaction. The final decrease in k, arises from active center trapping during these highly

crosslinked polymerizations. Therefore, a limiting conversion at which the reaction
ceases is observed despite the presence of active centers.
The PDSC profiles for the propagation and termination rate constants as functions

of conversion provide important insight into the nature of these high-speed cationic

 

photopolymerizations. Understanding the kinetics of these cationic photopolymerizations
is important due to the increasing number of applications for rapid, solvent-free curing of

polymer films.

 

VII.5. List of References

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

2. J. G. Kloosterboer “Network Formation by Chain Crosslinking Photopolymerizations
and Its Application in Electronics,” Adv. Polym. Sci. , 84, l (1988).

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

4. K. S. Anseth, C. M. Wang and C. N. Bowman “Kinetic Evidence of Reaction
Diffusion during the Polymerization of Multi(meth)acrylate Monomers,” Macro-
molecules, 27, 650 (1994).

5. K. S. Anseth, C. M. Wang and C. N. Bowman, “Reaction Behavior and Kinetic
Constants for Photopolymerizations of Multi(meth)acrylate Monomers,” Polymer, 35,
3243 (1994).

6. C. Decker and K. J. Moussa “Real-Time Monitoring at Ultrafast Curing By UV-
Radiation and Laser Beams,” Coatings Tech. , 62, 55 ( 1990).

7. J. G. Kloosterboer, G. F. C. M. Lijten, in Cross-Linked Polymers, R. A. Dickie, S. S.

Labana, and R. S. Bauer, Eds., ACS Symposium Series 367; ACS: Washington DC,
(1987).

8. J. V. Crivello, J. L. Lee and D. Conlon “Photoinitiated Cationic Polymerization with
Multifunctional Vinyl Ethers,” J. Rad. Curing 10, 6 ( 1983).

204

9. E. W. Nelson, T. P. Carter, A. B. Scranton “The Role of the Triplet State in the
Photosensitization of Cationic Polymerizations by Anthracene,” J. of Poly. Sci., Poly.
Chem, 33, 247 (1995).

10. E. W. Nelson, T. P. Carter, A. B. Scranton “Fluorescence Monitoring of Cationic
Photopolymerizations: Divinyl Ethers Photosensitized by Anthracene Derivatives,”
Macromolecules, 27,1013 (1994).

 

11. G. Odian, Principles of Polymerization, 2nd Ed., Wiley, NY, 22, 1981.

12. K. S. Anseth and C. N. Bowman “Reaction Diffusion Enhanced Termination in
Polymerizations of Multifunctional Monomers,” Polymer Reaction Engineering, 1(4),
499-520 (1992-93).

13. J. P. Kennedy, and E. Marechal Carbocationic Polymerization, John Wiley & Son,
New York, 1982.

 

CHAPTER VIII. IN SI T U LUMINESCENCE TEMPERATURE PROFILES

VIII.1. Introduction

In this chapter, the temperature-dependent luminescence of tris([3-diketone)

chelates of europium was used for in situ temperature measurements during cationic
photopolymerizations of vinyl ethers. These molecular-level luminescent probes
provided a real-time, noninvasive method for monitoring temperature during these high-
speed polymerizations. Two specific probes, tris(benzoyl-1,1,1-tn'fluoroacetone)
europium and tris(1,1,1,5,5,5-hexafluoroacetylacetone) europium met several stringent
spectral and performance requirements for application in our system. The luminescence
from these probes exhibits a reproducible temperature dependence over a wide
temperature range and is not sensitive to changes in viscosity. In situ temperature profiles
obtained using this novel technique verified the importance of thermal effects during
these highly exothermic photopolymerizations. These studies have demonstrated the
utility of the of the tris(B-diketone) europium chelates for characterizing the temperature
during high-speed photopolymerizations that cannot be monitored using conventional
techniques.

Most of the early work on cationic photopolymerizations focused on the

identification and synthesis of appropriate monomers and initiators (for reviews see

205

206
references 1 and 2). Not until recently have the detailed studies of the reaction been

reported?"4 For example, in a previous contribution“ Nelson et al. reported a novel in situ
laser-induced fluorescence technique for monitoring the initiation reaction in these high—
speed polymerizations. Due to its extremely short intrinsic time scale, the fluorescence
technique provided a means to characterize the kinetics of these polymerizations which
proceed too rapidly to be monitored by traditional methods such as differential scanning
calorimetry (DSC). These studies indicate that although the reactions are extremely rapid
(the systems typically react to completion in a few seconds), most of the conversion
occurs in the last 100 milliseconds due to a dramatic increase in the reaction rate. Nelson
et al. attributed this characteristic reaction profile to thermal runaway resulting from the
large amount of heat released by the rapidly polymerizing system.

Due to its short intrinsic time scale, luminescence techniques are attractive for
monitoring the local temperature of these reactions which may proceed to completion in a
few seconds. Furthermore, luminescence techniques provide a non-intrusive method for
monitoring the temperature since trace amounts of the luminescence probe may be readily

incorporated into the reaction mixture. Tris(B-diketone) chelates of europium fulfill the

requirements of a temperature-dependent luminescence probe for our studies. As
described in the remainder of this chapter, these probes exhibit a linear response for a
wide temperature range, and are insensitive to changes in viscosity. In addition, the

excitation wavelength of these probes is ideally suited for our studies.

207

VIII.2. Probe Requirements

In addition to exhibiting a temperature-dependent response, the luminescence
temperature probe for these studies must meet specific spectral requirements.
Anthracene, which absorbs between 320 and 390 nm, was used as a photosensitizer to
expand the initiating window to the near-visible region of the spectrum. The anthracene
photosensitizer makes it possible to initiate polymerization in the near-UV or visible
wavelengths as a result of a direct interaction between an excited state of a
photosensitizer and the initiator.4 In this experiment, 351 nm or 364 nm ultraviolet (UV)
light was used to initiate the photopolymerization. Therefore, these wavelengths are the
most convenient for excitation of the temperature probe, since they would eliminate the
need for two laser beams. In addition, the emission wavelengths of the temperature probe
must not overlap with the fluorescence of the anthracene photosensitizer in the 425 nm
region.

A class of compounds which satisfies both spectral requirements is the lanthanide

substituted tris(B-diketone) chelates, which absorb from 280 to 380 nm5 and emit in the

620 nm range.°

These compounds exhibit an interesting “antenna” effect which is
responsible for the temperature dependence of the luminescence. The light absorption by
the lanthanide itself is very weak, however the ligand plays the role of an antenna by
absorbing light and efficiently transferring the excitation energy to the metal ion.7 This

non—radiative intramolecular transfer from the excited state of the ligand to the highly

luminescent europium metal center results in a number of interesting luminescent

208
properties. A number of authors‘°'ll have characterized the chemical and photophysical

properties of these chelates. These authors found that the luminescence efficiency varies
greatly with the identity of the ligand, degree of coordination, temperature, and solvent.°'7
Sager et. al reported that the underlying photophysical process proceeds by the following
four-step mechanism: (1) ground singlet absorption to the excited singlet state; (2)
radiationless intersystem crossing from the excited singlet to the lowest lying triplet state;
(3) transfer of energy to the chelated ion; and (4) characteristic ionic fluorescent

emission.5

VIII.3. Experimental

VIII.3. I . Materials

The monomer, 3,6,9,12-tetraoxatetradeca-1,13-diene (DVE-3) and the
photosensitizer, anthracene, were purchased from Aldrich Chemical Co. The monomer
was dried over molecular sieves to remove any traces of water and the photosensitizer
was used as received. The cationic 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 initiator concentrations
specified in the remainder of this chapter correspond to the total UV9310C concentration.

A free radical polymerization system was used to identify any viscosity
dependencies of the temperature probes. The monomer, 2-hydroxethyl acrylate (HEA),

was purchased from Aldrich Chemical Co. This system was polymerized in solution with

209
N ,N-Dimethyl-formamide (DMF) in order to maintain optical clarity. The initiator for

this polymerization was 2,2-azobisisobutyronitrile (AIBN), purchased from Polyscience

Inc.
VIII.3.2. Synthesis of Temperature Probes

The temperature probes under primary investigation were tris(benzoyl-1,l,l-
trifluoroacetone) europium (Eu(btfa)3), and tris(1,l,1,5,5,5-hexafluoroacetylacetone)
europium (Eu(hfa)3). Both of these probes were synthesized as described below. The
ligands 1,l,1,5,5,S-hexafluoroacetylacetone (hfa) and benzoyl-l,l,l-trifluoroacetone
(btfa) were purchased from Lancaster Synthesis Inc., and the europium (HI) chloride
hexahydrate was purchased from Aldrich Chemical. The europium compound was added
to approximately 50 ml of distilled water. The pH was increased to 8.5 by the addition of
ammonium hydroxide. As the europium solution was stirred, a stoichiometric amount of
the ligand was added. After allowing the solution to react over night with stirring, the
precipitate was filtered and dried for 24 hours at 70°C. This procedure was based on

modifications of those reported in references 5,6,9-11.
VIII.3.3. Absorption and Luminescence Measurements
Absorption spectra of monomers, initiators, photosensitizers, and luminescent

temperature probes were obtained using a Hewlett Packard UV-vis 8452A diode-array

spectrophotometer. For the temperature studies, the luminescence was monitored using

210
an Aminco-Bowman Series 2 Luminescence Spectrometer. The excitation frequencies

were varied from 340 to 364 nm. Based upon the absorption spectrum of each lanthanide
chelate, the optimum concentration for measuring the luminescence was determined. A
concentration that absorbed too strongly would result in an artificially low luminescence
reading, due to a shorter path length resulting from the low penetration depth into the
sample. The temperature was controlled with a Polyscience 9100 refrigerated constant-
temperature circulator connected to a jacketed cuvette holder inside the luminescence
spectrometer. A mixture of ethylene glycol and water was used as the heating fluid.

The luminescence monitoring of the temperature probes in situ during the cationic
photopolymerization was performed in the MSU LASER laboratory. The luminescent
temperature probe and cationic photopolymerization reaction were both excited with a
Coherent Innova 200 Argon ion laser tuned at either 363.8 or 351.1 nm. A Newport
845HP-01 digital shutter system was opened with an electric pulse from the multichannel
analyzer, ensuring that the illumination and the acquisition were started simultaneously.
Fifteen milliwatts of unfocused laser radiation (measured with a Scientech 362
Power/Energy meter) in an ~3 mm dia. beam was directed on to the quartz capillary tube
containing the sample. The sample geometry and instrument set-up have been described
in further detail in an earlier publication.4 The quartz capillary was approximately 2.5 cm
in length with an inner diameter of 1 mm and an outer diameter of 2 mm. The fluorescent
light was collected at an angle of 90° from the incident beam and 90° from the
longitudinal axis of the quartz tube. The temperature of the sample was controlled by a

jacketed tube holder coupled to a thermostated water bath.

 

I

 

211

The luminescence signal was collected using a Spex 1877 Triplemate
spectrometer with a subtractive dispersion filter stage and a spectrograph stage. An
EG&G Princeton Applied Research Model 1530-C/CUV CCD detector, cooled to -120
°C to minimize dark charge levels, was used to detect the signal. The data was analyzed
with an EG&G Princeton Applied Research Model OMA IV detector controller.
Luminescence spectra spanning 50 to 100 nm centered at 620 nm were collected with
150 groove/mm gratings in both the filter and spectrograph stages. These spectra were

typically collected in intervals of 50 ms.

VIII.4. Results and Discussion

VIII.4. I . Probe Selection

The two most promising temperature sensitive luminescence probes, Eu(hfa)3 and
Eu(btfa)3, were selected based on their ability to meet the system requirements. The

structures of these lanthanide B-diketone chelates are shown in Figure VIII.0. In these

chelates six of the nine coordination sites are occupied by organic ligands and the
remaining three sites are occupied by water molecules. The choice of the organic ligand
is important since it determines the absorption wavelength and helps to impart solubility
in the monomer. These two chelates meet the solubility and spectral requirements for
these studies.

The absorbance spectra of Eu(btfa)3 and Eu(hfa)3 are shown in Figure VIII.2. The

figure illustrates that the extinction coefficient of Eu(btfa)3 is an order of magnitude

 

 

 

212

F U P B-dlketones RC ,1:
r--\IC Cf—F ‘1:
F F

 

o o
Hexafluoroacetylacetone Benzoyltrifluoroacetone
HFA BTFA
Substituted B—diketone Tris Chelates

H
R
\cfiQ/
I 4', %fi/R
\ > Oi/o M = Europium
1"-..““l. N;
H ‘\___,"7 lf§0 R : ’ CF3
c—O 1 ‘\\\
l / O\\\ )C—R
R I 5:5
/ ‘n
R

Figure VIIIJ. The structure for the two lanthanide B-diketone chelates, Eu(btfa)3 and
Eu(hfa)3.

absorbance

   

 

213

__.__ DVE-3 .M.

:10 """°— Ell(btfa)3 0.001 wt%f AA 0633300
[—4— Eu(hfa)3 0.01 wt% 11 X o0
\ l l

o l

l
l

 

 

wavelength (nm)

Figure VIII.2. Absorption spectra of Eu(btfa)3 (0.001 wt %), Eu(hfa)3 (0.01 wt %), and
DVE-3 in methanol.

214
larger than that of Eu(hfa)3. In addition, the Eu(btfa)3 spectrum is shifted toward the red

compared to the Eu(hfa)3 spectrum, and absorbs much mere strongly then Eu(hfa)3 at 364
nm. While it may appear that the 364 nm Eu(hfa)3 absorbance is too low to produce
strong luminescence, Eu(btfa)3 at 0.001 wt % and Eu(hfa)3 at 0.01 wt % were found to
have similar luminescence intensities (Figure VIII.3.). Due to its stronger absorption at
351 nm, the Eu(hfa)3 emits a more intense signal when excited at this wavelength.
Therefore the characterization of the cationic photopolymerizations was carried out using
351 nm as the excitation wavelength.

For the DVE-3 reaction system, it is important that the luminescence from the
temperature probe does not overlap with the fluorescence signal from the anthracene
photosensitizer (380-500 nm). For Eu(btfa)3 and Eu(hfa)3 the temperature sensitive
emission peak occurs at around 620 nm (Figure VIII.3.), and therefore does not overlap
with the anthracene peak. The fluorescence in the 375 nm range in Figure VIII.3. was
attributed to the organic ligands and does not change with temperature. As described
earlier, the energy transfer from the organic ligands to the metal center becomes less
effective, and the luminescence of the metal center decreases, as the temperature of the

system increases.

VIII.4.2. Temperature Studies

The temperature-dependent luminescence of Eu(btfa)3 in the monomer is shown

in Figure VIII.4. The figure illustrates that, as expected, an increase in the temperature

leads to a decrease in the luminescence intensity. A plot of the luminescence intensity at

 

intensity

140-

12)-

  

100-

 

13

 
 

215

-—°-- DVE-3
—°— Eu(btfa)3
—°—-— Eu(hfa)3

 

 

emission (nm)

t:—

‘
V .
Q—.'. k-

L—
v
C
.

 

 

Figure VIII.3. Luminescence spectra of Eu( btfa )3 (0.001 wt %) and Eu(hfa)3 (0.01 wt %)

in DVE-3 and of DVE-3.

216

‘23

A

35‘

’5‘
U0
UI
0°

l
“‘-.‘
\
5.“
I'I’
/

‘3
1

1
is
U:
(3’

 

luminescence intensity

 

 

 

r t
m- l 0
. j 1, —— 55 C
l
- i \ o
' t . x
1 \\
4" ,- --—— 75 °C
204 /: ">\
0 1 - J i F I I
a!) 610 M0 6!) 640 650

emission (nm)

Figure VIII.4. Luminescence spectra at varying temperatures for Eu( btfa )3 at 0.001 wt
% in DVE-3, excited at 364 nm.

217

16)-

 

140-
120-

1(1)-

    

luminescence intensity

 

 

403 1617 = 231.24 — 2.53 (T)
20 ' ab 1 4'0 ' s'o ' 63 “lo So

temperature (°C)

 

Figure VIII.5. The luminescence intensity of Eu( btfa) 3 at an emission of 617 nm plotted
at different temperatures, resulting in a linear calibration curve.

218

618 nm (peak maximum) versus temperature, is shown in Figure VIIIS. This intensity
varies linearly with temperature (the correlation coefficient is 0.9976) and may therefore
be used to estimate the temperature of a system if the measurement is performed using the
same experimental conditions (concentrations and instrument settings).

The temperature-dependent luminescence of Eu(hfa)3 is shown in Figure VIII.6.
Again, the luminescence intensity decreases with increasing temperature, however in
contrast to the Eu(btfa)3 profiles discussed above, the Eu(hfa)3 luminescence exhibits two
distinctive peaks centered at 618 and 622 nm. The intensities of these peaks undergo an
inversion in relative intensity with the 622 nm peak exhibiting a larger intensity at low
temperatures and the 618 nm peak exhibiting a higher relative intensity at high
temperatures. Equations 1 and 2 illustrate the linear dependence of each peak intensity on
temperature (the correlation coefficients are 0.99849 and 0.99788 for the 618 and 622 nm
peaks respectively).

16,8 =150—1.6(T) Eq. 1.

1622 = 200 - 2.4(T) Eq. 2.
To take advantage of the inversion in the relative intensities of the two peaks, a detector-
insensitive calibration was developed based upon an intensity ratio defined as the
intensity of the 618 nm peak divided by the modified intensity of the 622 nm peak (see
Figure VIII.7.). As shown in the figure, a plot of the resulting intensity ratio versus
reciprocal temperature is linear with a correlation coefficient of 0.99997). This
calibration curve is independent of operating conditions such as instrument settings since

the peaks act as internal standards.

219

$55"?

5

l

luminescence intensity

 

 

 

 

04... r . . . 1
an no 3:1

emission (nm)

Figure VIII.6. Luminescence spectra at varying temperatures for Eu(hfa )3 at 0.01 wt %
in DVE-3 excited at 340 nm.

220

   

' Data

 
 

 

 

 

 

1.5- Calibration Curve
3?:
VJ
1::
B
.S 1.0-
Q)
.2
E
8 I 6590

0.5“ 6‘8 = -0.70+——°—

I622 — 200 T
0.0 r T I ' T ' ' ‘ ' "
30 40 50 60 70 80

temperature (’C, reciprocal scale)

Figure VIII. 7. The relative intensity, referring to the intensity of the 618 peak divided by
the modified intensity of the 622 peak, of Eu(hfa)3 in DVE-3 (0.01 wt %) plotted versus

reciprocal temperature.

221

To investigate the effect of elevated temperature on the luminescence of the
lanthanide chelates, a series of studies were performed for temperatures ranging from 30
to 130 °C. The resulting luminescence/temperature calibrations for Eu(hfa)3 and
Eu(btfa)3 are shown in Figure V1118. As the figure illustrates, the Eu(hfa)3 curve was
best fit by a single straight line throughout the entire temperature range, while the
Eu(btfa)3 curve was best fit by an exponential decay. Therefore the Eu(hfa)3 probe is
better suited for the high temperature studies since it does not lose sensitivity as the
temperature is increased. Theoretically this probe should be able to monitor much higher
temperatures approaching its degradation temperature. However, since the luminescence
intensity decreases as the temperature rises, the maximum measurable temperature is
limited by the detector’s low-level sensitivity. If the initial temperature is set to the
maximum detector reading, the dynamic range of the detector limits the size of the
temperature range. Therefore, the maximum effective temperature range is effectively

limited by the dynamic range of the instrument.

VIII.4.3. Viscosity Independence

Many probes which exhibit temperature-dependent luminescence are also
sensitive to changes in viscosity.12 However, a viscosity-dependent response is
unacceptable for a luminescent probe used for monitoring temperature during
polymerization. These systems undergo a large change in viscosity, and any viscosity-
dependent response will likely dominate the desirable temperature-dependence.

Therefore, a series of studies were performed to determine whether the luminescence of

222

 

 

250
‘3»

2m- I
>5 1622:270-140)
93 150' "
.5 ° ° 030
8 ‘ E (It? °
5 100- o )3 °°
8 - E.I(btfa)3 _
Q)
.E J
E
'2 50-1

1622=-1.7+220EXP(-('l‘-35)/55)
0 I I I I I 1 I
40 a) m 1(1) 120

temperature (°C)

Figure VIII.8. High temperature calibration curves for Eu(btfa )3 and Eu( hfa )3 excited at
364 nm and monitored at an emission of 620 nm.

 

223
Eu(hfa)3 and Eu(btfa)3 were sensitive to viscosity. A calibration curve was produced for

a liquid monomer system ( ~2 cps) and compared to a calibration for a solid polymerized
system. Therefore, if the calibrations are the same for the two extremes of liquid and
solid, the luminescence response is insensitive to viscosity.

A solution polymerization was chosen for preparing the solid samples to ensure
optical clarity and permit easy removal of the polymer from the quartz cuvettes. DMF
was chosen as the solvent due to its relatively high boiling point (153 °C) and because the
resulting polymer was optically clear. A suitable ratio of monomer to solvent was found
to be 40/60. The polymerizations were carried out at a temperature of 65 °C with an
initiator concentration of 0.5 wt % and a B-diketone tris chelate concentration of 10'2
wt %.

Profiles of luminescence intensity versus temperature for the HEA/DMF system
are shown in Figure VIII.9. When corrected for the presence of the AIBN initiator, the
luminescence intensities show good agreement in both the liquid and solid state,
indicating that the luminescence-temperature probe Eu(hfa)3 is not sensitive to changes in
viscosity. The viscosity independence of the europium probes is not surprising since the
temperature dependent energy transfer is internal and should not be effected by the
translational or rotational motion of the molecule. Equivalent viscosity experiments were

performed with Eu(btfa)3 yielding similar results.

 

 

224

 

 

10_ quuidsample40’6OI-IEA/DMF
o Solidpolymersample(40’601-lEA/D1V1F)
. reactedwith05wt%AlBN
Bothsamplescmtain0.01 wt %Ena-IFA)3
3. 8- i
'E'o"
: .
8 a
.E‘ o
a) 6‘
8
a) - l
o
8 4-
g 3
2. ‘ 2
o
2" a
O f r I ' I ' l ' I
3) 40 50 60 70 80

temperature (°C)

Figure VIII.9. Temperature calibration curves for 0.01 wt % Eu(hfa)3 excited at 351 nm
in liquid and solid systems.

 

 

225

VIII.4.4. Inherent System Limitations

The above discussion illustrates the utility of the lanthanide chelates for in situ
temperature monitoring. These molecular-level probes exhibit a linear temperature
dependence over a wide temperature range and, in the case of Eu(hfa)3, may provide a
detector-insensitive calibration. However, these systems have several inherent
limitations. Perhaps the largest limitation arises from the sensitivity of the luminescent

response to solvent. Because water effectively competes with the B-diketones for the

coordination sites on the lanthanides, the luminescence is diminished in the presence of
water. Therefore these probes cannot be used With any aqueous systems. In DVE-3, the
luminescence intensity decreases over a period of days possibly due to the slow exchange
of the ligand and the solvent. In addition, the shape of the luminescence spectra of the
chelates is dependent on the solvent. For example, as shown in Figure 10, the two
distinctive Eu(hfa)3 peaks observed in DVE-3 are not present for experiments performed
in methanol and methylene chloride. This dependence on solvent necessitates a

calibration to be established for each system.
VIII.4.5. Temperature Profiles for the Cationic Photopolymerizations
Since the luminescence temperature probes were found to be sensitive to changes

in temperature but insensitive to viscosity, they were investigated for use in monitoring

the reaction temperature during the cationic photopolymerizations of DVE-3. A series of

 

 

 

226

51‘
1
\
-/

   

—--- Eu(hfa)3 in MeCl2
--°-— Eu(hfa)3 in MeOH
—e— Eu(hfa)3 in DVE-3

3.11.?
\
7R

luminescence intensity

 

 

 

emission (nm)

Figure VIII.10. The luminescence of Eu(hfa)3 (0.01 wt %) in methanol, methylene
chloride, and DVE-3.

 

227
studies were performed to ensure that the lanthanide probes did not affect the

polymerization kinetics and that the presence of the initiator and photosensitizer did not

disrupt the luminescence response of the temperature probes. The B-diketone tris chelates

are electronically neutral, therefore they will not undergo electrostatic interactions with
the propagating active centers. However it was anticipated that the lanthanide probes
could possibly affect the reaction by interacting with the initiator or the anthracene
photosensitizer.

A non-reacting system (a sample containing DVE-3, anthracene, and an europium
probe) was tested to determine the time stability of the temperature-sensitive probe. No
initiator was added to ensure that the system was non-reactive and remained isothermal
when exposed to the beam from the argon ion laser. Essentially no change in
luminescence was observed over the time scale of the experiment (30 sec) for both
Eu(hfa)3 and Eu(btfa)3. Next, to investigate the effect of the initiator, the temperature-
sensitive luminescence of a sample containing hexanol, anthracene, initiator, and an
europium probe was monitored. Hexanol was chosen to approximate DVE-3 in this non-
reactive system. The hexanol system was then exposed to the initiating wavelength
causing the photosensitization to begin. For both probes (Eu(hfa)3 and Eu(btfa)3), there
was no change in the temperature sensitive probe luminescence over a substantial period
of time. This indicates that the temperature probes were not being consumed during the
photosensitization reaction. By monitoring the anthracene fluorescence intensity as a
function of time, the photosensitization reaction rate was found to be unaffected by the
addition of the europium probes, indicating that the temperature probes are not adversely

affecting the photosensitization reaction.

 

 

 

228
The cationic photopolymerizations were carried out in the MSU LASER

laboratory, therefore calibration curves were generated using the Triplemate spectrometer
and CCD detector. To produce the calibration curve, a non-reacting sample containing
anthracene and a europium probe in DVE-3 was used. The sample tube was placed in a
jacketed cell holder (as described in the experimental section), then the temperature was
increased while the luminescence at 620 nm was measured (Figure VIII.11.).

In preliminary reaction-monitoring experiments, the europium luminescence at
620 nm was measured at 50 ms intervals, and the calibration curve was used to generate
the temperature profile. The temperature reading at the start of the reaction was adjusted
to match room temperature (25 °C). The luminescence from the europium probe
indicated that the temperature increased gradually from ~25 to ~55 °C at which point the
reaction temperature increased rapidly (Figure VIII.12.). Therefore these studies illustrate
that, as expected, a marked increase in temperature is associated with the sudden increase
in the reaction rate. The reaction temperature continues to increase; however, due to the
instrumental dynamic range limitations mentioned above, higher temperatures were not
able to be monitored in this system. In addition, at elevated temperatures the sample
loses optical clarity. However, the important temperature at which the thermal runaway
occurs (~55 °C for these specific reaction conditions) can be readily obtained from the

temperature profile.

 

 

229

7000-

 

30004

luminescence intensity
5

 

2000-

 

 

V r V l I I r 1 I I I I
30 4O 50 60 70 80 90

temperature (°C)

 

Figure VIII.11. Temperature calibration for 1.6 x 10’2 wt % Eu(hfa)3 in DVE-3 with 1.4
x 10’2 wt % anthracene excited with 15 mW at 351 nm using an argon ion laser.

230

8.“ S
l 1 l
D

0

temperature (°C)

L»

O
1

0

4

0000

H
1

l

 

20 U l I 1 1

time (s)

Figure VIII.12. Temperature profile of the cationic polymerization of DVE-3 1.6 x 10'2
wt % anthracene and 1.0 wt % initiator monitored with 1.0 x 10'2 wt % Eu(hfa)3.

231

VIII.4. 6. Further Applications

Due to their temperature dependence and viscosity insensitivity, the B-diketone

tris chelates of europium may be attractive for temperature monitoring in a variety of
systems. The technique could find application in any system in which an in situ, non-
intrusive measurement of temperature is desired. Moreover, due to their viscosity
insensitivity, the probes are appropriate for systems in which a change in phase occurs.

To illustrate the broader applicability of this technique in other reaction systems,
the in situ luminescence during a free radical polymerization of 2-hydroxyethyl acrylate
(HEA) was investigated for both probes. The reaction was a solution polymerization of
40/60 HEA/DMF in order to insure optical clarity during the reaction. The
polymerizations were performed in 3.5 ml cuvettes heated to 65°C by a jacketed cell
holder. AIBN was used as the initiator, and the lanthanide probes were present in
concentrations of 10'2 wt %. Temperature excursions may occur due to the heat of
polymerization, and the luminescence technique provides a temperature measurement
which is averaged over the width of the cuvette. Luminescence measurements were
easily performed in a standard luminescence spectrometer since the reaction takes
minutes to complete as opposed to seconds in the case of cationic polymerizations.

Figure VIII.13. shows the temperature profile produced from the luminescence
intensity at 620 nm (the peak maximum) during a free radical polymerization of HEA. A
calibration curve was prepared using the procedures described previously for the cationic

system. As the reaction proceeds the exothermic heat of polymerization will cause the

 

 

232

temperature (°C)
8
l

 

   

time (min)

Figure VIII.13. The temperature profile generated from the luminescence intensity of

Eu(hfa)3 in a free radical polymerization of HEA/DMF (40/60) with an initiator
concentration of 0.2 wt %.

233
system temperature to increase above the jacket temperature, if heat evolution rate is

higher than the heat transfer rate. Therefore, a dramatic increase in the temperature was
observed at ~15 minutes, corresponding to an increase in the reaction rate due to
autoacceleration. After reaching a peak value of around 100 °C, the reaction is essentially
complete and the temperature of the solution returns to that of the jacket (65°C).
Therefore, the luminescence ultimately returns to approximately the same intensity as
when the sample was a liquid (confirming that the luminescence is independent of
viscosity). The profile shown in Figure VIII.13. has the well-characterized shape
indicative of free radical polymerizations undergoing the gel effect,l3 thereby illustrating

the broader applicability of these temperature probes.

VIII.5. Conclusions

In conclusion, luminescence can provide a novel in situ, non-destructive method
for monitoring the temperature in high speed reactions. Molecular-level luminescent
probes provide extremely rapid response without requiring a macroscopic probe to be
imbedded in the sample. For temperature monitoring during photopolymerizations of

vinyl ethers, two B-diketone tris chelates of europium [Eu(hfa)3 and Eu(btfa)3] were

found to meet several stringent requirements. These probes exhibit a reproducible
temperature dependence over a wide temperature range and, in the case of Eu(hfa)3, may
provide a detector-insensitive calibration. Calibration curves of luminescence intensity
versus temperature were generated for the temperature range of interest (between 20 and

90 °C). In addition, these probes meet specific spectral requirements, such as excitation

 

234
in the 350 nm range and emission above 500 nm. Finally, the luminescence response of

these probes was found to be independent of viscosity.

These temperature profiles for cationic photopolymerizations verified the
importance of thermal effects in these high-speed polymerizations. The luminescence
from the europium probe indicated that the temperature increased gradually from ~25 to
~55 °C, then increased dramatically in a period of a few hundred milliseconds. These
results confirm that a thermal runaway effect is responsible for the large increase in
polymerization rate, as previously postulated by Nelson et al.4 While the temperature-
sensitive probes appear to accurately monitor the temperature prior to thermal runaway,
they are unable to determine the maximum reaction temperature due to detector dynamic
range limitations and problems associated with the sample losing optical clarity.
Furthermore, the luminescence of the probes was found to be sensitive to sample
components (solvent, initiators, and water) and seemed to degrade at long times.
However, this can be overcome with the use of system specific calibration curves and the
use of fresh samples.

These studies have demonstrated the utility of the temperature-dependent

luminescence of tris(|3—diketone) europium chelates for characterizing the temperature

during high-speed photopolymerizations that cannot be monitored using conventional
techniques. However the applicability of these probes is not limited to this reaction
system. These probes could find application in any system in which an in situ, non-
intrusive measurement of temperature is desired. For example, due to their temperature
dependence and viscosity insensitivity, the probes are appropriate for systems in which a

change in phase occurs. The broader applicability of these probes was illustrated by our

 

235
studies of the HEA free radical polymerizations. Finally, a host of other lanthanide

compounds could possibly be used as temperature sensitive fluorescent probes. For
example, commercially available probes worthy of further study include tris(6,6,7,7,8,8,8-
heptafluoro-2,2-dimethyl-3,5-octanedionato) europium and tris(2,2,6,6-tetramethyl-3,5-

heptanedionato) europium.

VIII.6. List of References

1. J. V. Crivello, “Cationic PolymerizatiOn-Iodoium and Sulfonium Salt Photo-
initiators”, Adv. Polym. Sci., 62, 1 (1984).

2. J. V. Crivello, J. L. Lee, “Alkoxy—Substituted Diaryliodonium salt Cationic Photo-
initiators”, J. Polym. Sci., PartA Polym. Chem. 27, 3951 (1989).

3. C. Decker, D. Decker, “Kinetic and Mechanistic Study of the UV-Curing 0 Vinyl
Ether Based Systems,” RadTech 94 North American Proceedings, RadTech
International North America, 602, (1994).

4. E. W. Nelson, T. P. Carter, A. B. Scranton, “Fluorescence Monitoring of Cationic
Photopolymerization: Divinyl Ether Polymerizations Photosensitized by Anthracene
Derivatives,” Macromolecules, 27, 1013-1019 (1994).

5. W. F. Sager, N. Filipescu, F. A. Serafin, “Substituent Effects on Intramolecular
Energy Transfer. 1. Absorption and Phosphorescence Spectra of Rare Earth [3-
Diketone Chelates,” J. of Phys. Chem, 69(4), 1092-1100 (1965).

6. F. Halverson, J. S. Brinen, and J. R. Leto, “Luminescence of Europium and
Hexafluoroacetylacetonate,” J. of Chem. Phys., 40 (10), 2790-2792 (1964).

7. B. Alpha, R. Ballardini, V. Balzani, J -M. Lehn, S. Perathoner, N. Sabbatini, “Antenna
Effect in Luminescent Lanthanide Cryptates: A Photophysical Study,” Photo-
chemistry and Photobiology, 52 (2), 299-306 (1990).

8. A. M. Murray, L. A. Melton,Appl. Opt., 24, 2783-2787 (1985).

 

236

9. M. F. Richardson, W. F. Wagner, D. E. Sands, “Rare-Earth Trishexafluoro-
acetylacetonates and Related Compounds,” J. Inorg. Nucl. Chem, 30, 1275-1289
(1968).

10. S. Sato, M. Wada, “Relations between Intramolecular Energy Transfer Efficiencies
and Triplet State Energies in Rare Earth B-Diketone Chelates,” Bulletin of the
Chemical Society of Japan, 43, 1955-62 (1970).

11. E. J. Schimitschek, J. A. Trias, “IR and Raman Spectra of Europium(I[I) B-Di-
ketonates,” J. Inorgan. Nucl. Chem, 32, 811-831 (1970).

12. K. F. Schrum, A. M. Williams, S. A. Haerther, D. Ben-Amotz, “Molecular
Fluorescence Thermometry,”Anal. Chem, 66, 2788-2790 ( 1994).

13. G. Odian, Principles of Polymerization, John Wiley & Sons, Inc.: New York, NY,
1991 p. 286.

 

 

CHAPTER IX. CONCLUSIONS AND RECOMMENDATIONS

The work reported in this thesis has contributed to the understanding of the
kinetics and mechanisms of cationic photopolymerizations of divinyl ethers. The
fundamental experimental and theoretical investigation of these UV-initiated cationic
polymerizations presented in this thesis complement the more macroscopic studies
currently in the literature, and are imperative if the potential of these reactions is to be
realized. Furthermore, novel cure and temperature monitoring methods have been
developed to characterize these high speed polymerization reactions. These new methods
and techniques are general in their ability to monitor cure and temperature and will be
widely applicable in high speed polymerization reactions. In this chapter, some general

conclusions and suggestions for further work will be presented.

IX.1. Summary of Results

This research has provided the first detailed picture of the kinetics, structure, and
properties of model cationic photopolymerization systems, and has resulted in the
development of a host of analytical techniques which may now be used to characterize a
broader set of cationic photopolymerizations. Several distinct aspects of the photo-

polymerizations were investigated: 1) A series of fluorescence cure monitoring studies

237

 

238
based upon a novel experimental technique were used to investigate the effects of several

reaction variables; 2) A series of studies, including fluorescence and phosphorescence
lifetime measurements, were performed to characterize the photosensitization mechanism
by anthracene and its derivatives; 3) A series of in situ, isothermal reaction studies were
performed using Raman spectroscopy to directly monitor propagation; 4) A series of
isothermal and dark-cure photo-differential scanning calorimetry studies were performed
to characterize the reaction kinetics; and 5) A novel time-resolved, in situ temperature

monitoring scheme was developed based upon B-diketone trischelates of lanthanide salts.

Results of each of these studies are briefly described below.

IX.I.1. Fluorescence Cure Monitoring

Cationic photopolymerizations of vinyl ethers were investigated using in situ, time
resolved fluorescence spectroscopy. Due to its short intrinsic timescale, fluorescence
spectroscopy provides a means to characterize these polymerizations which proceed too
rapidly to be monitored by traditional methods. The fluorescence intensity of the
photosensitizer (anthracene or its derivatives) was monitored as a function of time with
spectra collected in intervals as short as 2 milliseconds. An observed reduction in
fluorescence intensity was attributed to consumption of the photosensitizer, providing a
means to monitor the production of active cationic centers. The reaction rate increased as
the initiator or photosensitizer concentrations were increased. Also, reactions photo-
sensitized by anthracene and 9,10-dimethylanthracene produced the fastest rates, followed

by 9-vinylanthracene. The polymerization rate for 9,10-diphenylanthracene was

 

 

239

considerably slower due to steric hindrance or resonance effects. These results illustrated
the tremendous potential of in situ, time-resolved fluorescence spectroscopy for

monitoring polymerizations.

IX] .2. Characterization of the Photosensitization Mechanism

The photosensitization mechanism for cationic polymerizations initiated by
diaryliodonium salts photosensitized by anthracene was investigated using fluorescence
and phosphorescence spectroscopy. In situ photosensitizer fluorescence measurements
confirmed that the photosensitization reaction proceeds by an electron transfer process.
Transient phosphorescence studies demonstrated that electron transfer occurred from the
triplet excited state of anthracene to the initiator. Further evidence for the role of the
triplet state was provided by an observed 7—fold decrease in the polymerization rate upon
addition of a triplet state quencher. Finally, numerical solution of the photophysical
kinetic equations indicated that the triplet state concentration was approximately three
orders of magnitude higher than that of the singlet state, and that 94-96 % of the active
cationic centers are produced by reaction of the initiator with the triplet state. These
results indicate that the electron transfer occurs primarily from the triplet state of
anthracene, with the singlet state providing only a minor contribution to the

photosensitization reaction.

 

 

240

IX.I.3. In Situ Ramon Studies

Raman spectroscopy was used to monitor cationic photopolymerizations of a
model divinyl ether. Using this technique complete spectra of approximately 1000 cm'1
were collected in situ during the cationic photopolymerization. Since molecular
vibrations associated with the vinyl bonds are measured, the Raman technique offers a
direct method for monitoring vinyl bond conversion, thereby allowing the reaction
kinetics to be characterized at different temperatures and light intensities. Increases in the
initiating light intensity or the reaction temperature resulted in an increased reaction rate
and an enhanced limiting conversion. The apparent propagation constant exhibited an
initial sharp increase and leveled off at a plateau value then decreased as a limiting
conversion was reached. The propagation rate constant and activation energies obtained
from the Raman studies were in good agreement with the results based upon photo-

differential scanning calorimetry.

IX] .5. Photo-Differential Scanning Calorimetry Studies

Photo-differential scanning calorimetry (PDSC) experiments were used to study a
series of divinyl ether polymerizations at different temperatures. The apparent
propagation rate constant, kp, initially increased sharply, then maintained a plateau value
until finally decreasing. The initial large increase in kp was explained by a change in the

reactivity of the carbocation species due to the decreased mobility of the large

 

241

hexafluoroantimonate counterion as the system viscosity increases. The separated ions
are orders of magnitude more reactive than ion pairs dramatically increasing the apparent
value of kp. The final decrease in kp arises from active center trapping which results in a

limiting conversion for these highly crosslinked polymerizations.

IX.I.4. Temperature-Sensitive Luminescence Studies

The temperature-dependent luminescence of tris(B-diketone) chelates of europium

was used for in situ temperature measurements during cationic photopolymerizations.
These molecular-level luminescent probes provided a real-time, noninvasive method for
monitoring temperature during these high-speed polymerizations. Two specific probes,
tris(benzoyl-l,1,l-trifluoroacetone) europium and tris(l,1,l,5,5,5-hexafluoroacetyl-
acetone) europium met several stringent spectral and performance requirements for
application in our system. The luminescence from these probes exhibits a reproducible
temperature dependence over a wide temperature range and is not sensitive to changes in
viscosity. In situ temperature profiles obtained using this novel technique verified the
importance of thermal effects during these highly exothermic photopolymerizations.
These studies have demonstrated the utility of the of the europium chelates for
characterizing the temperature during high-speed photopolymerizations that cannot be

monitored using conventional techniques.

 

242

IX.2. Suggestions for Future Work

The experimental and analytical techniques developed for a model vinyl ether
polymerization can now be used to characterize a broader range of systems which could
find application as industrially relevant solvent-free coatings. Since the methods
described in this thesis allow a large amount of information to be collected in a short
period of time, a wide range of monomers, initiators, and photosensitizers can now
efficiently and systematically be studied. To date a majority of the research on cationic
photopolymerizations has focused on the development of new monomers and initiators,
and there are now a wide variety of both commercially available monomers, thermally
stable initiators, and photosensitizers. However, many of these cationic polymerization
systems have received only cursory kinetic investigations.

To address both the scientific and industrial objectives of cationic photo-
polymerizations for high speed coating systems, two criteria can be used for the selection
of monomers: i) to provide fundamental information about the underlying relationships
between the monomer structure and the characteristics of the polymerizations, sets of
monomers whose molecule structures differ in a systematic manner were identified and
ii) monomers which were deemed most promising for industrial applications. As
described in detail below, future research performed in the following areas would aid in
understanding cationic photopolymerizations: i) characterization of the initiation and
photosensitization mechanisms for new initiator/photosensitizer combinations; ii)

determination of the relationships between the monomer structure and the reaction

 

243

kinetics; iii) studies of the relationships between the monomer structure and the
properties of the resulting polymers; and iv) investigations of the propagation reactions

based upon linear polymerizations of monofunctional monomers.

IX.2. 1. Characterization of the Photosensitization Mechanism

Several commercial cationic initiators are now available and should be
investigated. These included iodonium and sulfonium salts, iron-arene complexes, and
pyridinium complexes. These initiators are available commercially from a variety of
sources including G.E. Silicones, Union Carbide, and Ciba Geigy. The choice of the
counter ions may have a dramatic influence on the rate of the propagation reaction and
the properties of the polymer film. Typical counter ions used for these systems include
Sbe, AsF6, PF6, and BF4, and all four should be included in these studies. The size and
acidity of the counter effects the actual propagating species during the cationic
polymerization. The counter ion that produces the fastest reaction rates of is the
hexafluoroantimonate, however there are potential environmental concerns associated
with the heavy metal counterions (moreover, in microelectronics applications the
presence of charged metal ions can have a detrimental effect on the performance of the
device). Therefore, alternative counter-ions which due not contain heavy metals will be
emphasized in our research. Available metal-free counter ions include trifluoro-
methanesulfonate and terakis(pentafluorophenyl) borate.

Using the experimental methods described previously, the effects of the identities

of the initiator and the counter ion on the observed polymerization kinetics can be

244

examined. In many cases differences in the observed polymerization rate are actually
caused by differences in the rate of active center generation, and the efficiency of the
initiator (at the wavelength of interest) is an important parameter. Since these cationic
photoinitiators absorb light primarily in the deep-UV region of the spectrum,
photosensitization is commonly used to shift the initiating wavelength to the near-UV or
visible region. Photosensitization may occur by one of many possible mechanisms,
including energy transfer, or electron transfer form the singlet or triplet excited states. An
understanding of the operative photosensitization mechanism is important for both
scientific and practical purposes since these reactions determine the rate of active center
generation, and thereby play a pivotal role in the polymerization process.

In Chapter V, a complete set of ‘ photophysical studies demonstrated that
photosensitization of diaryliodonium salts by anthracene occurs by electron transfer from
the triplet excited state of the photosensitizer to the initiator. In future research, these
studies could be expanded to other important initiation systems. For example,
fluorescence and phosphorescence lifetime studies combined with photophysical rate
calculations can now be used to characterize the photosensitization of sulfonium salts by
anthracene and its derivatives. Due to the less favorable reduction potential of the
sulfonium salts, the results for these systems should differ substantially form those for the
iodonium salts. Measurement of the anthracene fluorescence and phosphorescence
lifetimes in the presence and the absence of the initiator allow the intrinsic rates constants
from these states to be determined. These rate constants, when combined with the
photophysical rate expressions, allow the overall and individual rates of active center

production to be determined. This analysis, which involves no adjustable parameters (all

 

 

245
kinetic constants are independently measured) exhibited remarkable agreement with

experimental profiles of the anthracene concentration. Other photosensitizers such as
pyrene and pereylene may be used to extend the initiation wavelength further into the red.
Although these photosensitizers have received less attention than anthracene, they could
provide important flexibility in the design of the initiating system (this flexibility may be
important in systems containing dyes or pigments). Therefore, the mechanism and
effectiveness of various photosensitizers can now be examined using the techniques

described in previous chapters.

IX.2.2. Characterization of the Photopolymerization Reactions

Examination of the relationship between the monomer structure and the polymer
characteristics by performing a series of integral studies will provide a comprehensive
understanding of these complex reactions. Experiments performed in each of the
following areas will provide detailed information about cationic photopolymerizations: i)
the effect of the type of the reactive functionality by comparing studies performed with
vinyl ethers, propenyl ethers, silicone-containing monomers, and vinylphenoxy alkanes;
ii) for multifunctional monomers, the role of the chain length between reactive groups can
be characterized by considering a homologous series of monomers; iii) to provide
information about the propagation reaction as well as potential competing reactions,
linear polymerizations of monovinyl ethers can now be performed so that the resulting
polymers may be analyzed in detail using techniques that require dissolution. For each of

these experiments the polymerizations kinetics will be characterized using in situ

246
fluorescence, Raman and Photo-Differential Scanning Calorimetry (PDSC) methods

describe previously in this thesis, and the polymer properties can be characterized using

standard techniques.

IX.2.3. Relationships between the Monomer Structure and the Reaction Kinetics

Three of most exciting classes of monomers which undergo cationic
photopolymerizations are vinyl ethers, propenyl ethers, and silicone-containing
monomers (epoxides and vinyl ethers). Studies with each of these monomers to
determine the precise role of the reactive group on the polymerization kinetics can now be
performed. While it is clear that each of these groups polymerize very rapidly, detailed
kinetic studies are currently unavailable for most of these monomers. Fundamental
studies of highly purified systems can eliminate the effects of impurities (if present these
impurities could obscure the comparisons between monomers). In addition, for
industrially-relevant systems studies can be performed without additional purification to
characterize the reaction kinetics under typical industrial conditions. Together these
studies will allow the kinetic effects of any of residual impurities to be determined.

Directly monitoring the consumption of the reactive (unsaturated) functionalities
using in situ Raman spectroscopy, is ideally suited for these studies not only because it is
a direct method, but also because it offers appropriate time resolution and allows the
sample temperature to be controlled. In addition, because the technique provides a
spectrum rather than the intensity at a single wavelength, information may be obtained

from more than one spectral feature. For each monomer under consideration, Raman cure

247

profiles can now be obtained at a series of temperatures as described in Chapter VI.
These Raman experiments provide a direct measure of conversion as a function of time
and may be combined with the photosensitizer studies (which provide active center
concentration as a function of time) to evaluate the effective propagation rate constant
throughout the reaction. These studies provide fundamental kinetic information which
may be applied in a variety of reaction conditions, including industrial processes. Using
the thermostated sample cell, studies can be performed at a variety of temperatures, and
the activation energies for overall polymerization, propagation, and photosensitization
determined. These investigations would provide important information on the influence
of the monomer structure on the activation energies. This is particularly important due to

the highly exothermic nature of these reactions.

IX.2.4. Effect of Multifunctional Monomer Structure on Limiting Conversions

1n polymerizations of multifunctional monomers a limiting conversion
significantly less than unity is typically observed due to the restricted mobility of the
pendent functionalities and the associated trapping of active centers. This limiting
conversion can have a substantial effect on the polymer pr0perties, therefore it is
important to understand and control this effect. In future work, two different approaches
can be used to characterize and control the limiting conversion. First, for homo-
polymerizations of multifunctional monomers, the role of the chain length between

reactive groups may be characterized by considering several homologous series of

 

248
monomers. In these studies three distinct systems would be characterized: a series of

oligomeric ethylene glycol divinyl ethers, a series of alkylvinyl ethers, and a series of
bis(vinylphenoxy)alkanes. In a second set of studies copolymerizations of multi-
functional monomers with monofunctional monomers (containing one polymerizable
moiety) will be investigated. The addition of the monofunctional monomers will increase
the maximum attainable conversion due to the enhanced mobility in these systems.

In this thesis, the homopolymerizations of the monomer 3,6,9,l2-tetra-
oxatetradeca-1,13-diene (DVE-3) has been extensively characterized. This monomer
contains two polymerizable vinyl ether functionalities separated by a chain of three
ethylene glycol groups. A series of DVE monomers containing two to six ethylene glycol
units between the reactive functionalities is now commercially available. This
homologous series of monomers would allow the role of the chain length between
reactive groups to be characterized. It is anticipated that the maximum attainable
conversion will increase as the separation between reactive functionalities is increased
(due to the enhanced mobility of the pendent functionality), and may influence the
reaction kinetics as well. These studies will be performed using the Raman scattering
technique described previously, and the results obtained from the spectroscopic
techniques will be compared with results obtained using photo-differential scanning
calorimetry (PDSC) in order to corroborate the results. In addition, similar studies can be
performed for two other homologous series of cationically polymerizable monomers:
bis(vinylphenoxy)alkanes (alkoxy-styrene units connected by alkane units of varying
lengths”); and alkylvinyl ethers (divinyl ethers in which the reactive functionalities are

separated by alkyl chains).

249

In the second set of studies, copolymerizations of multifunctional monomers with
monofunctional monomers may be investigated using the Raman scattering technique.
The maximum attainable conversion of mixtures of mono and di-functional monomers
should increase due to the more flexible network structure, and may create a substantial
effect on the overall polymerization kinetics. Monomers such as cyclohexanedimethanol
monovinyl ether (CVE) and 1,4-bis[vinyloxy methyl]-cyclohexane (CHVE) should have
very similar intrinsic reaction rates and should therefore provide information on limiting
conversion due to changes in monomer structure. This strategy could provide an

industrially feasible method for controlling the limiting conversion.

IX.2.5. Relationships between the Monomer Structure and Polymers Properties

In addition to the kinetic studies described above, experiments can be performed
to establish the relationship between the monomer structure and the properties of the
resulting polymer films. A complete characterization of the thermal properties of the
polymer films may be obtained using the differential scanning calorimeter (DSC), thermal
gravimetric analysis (T GA), and thermal mechanical analysis (TMA) instruments
available at the Composite Materials and Structures Center (CMSC) at Michigan State
University. These experiments will allow the glass transition temperature (if any) the
degradation temperature, and the expansion coefficient of each film to be established.
The mechanical properties of the films can be thoroughly investigated using a
Rheometrics RMS 800 Dynamic Mechanical Analyzer (again available at the CMSC).

Using standard procedures, this technique may be used to measure the loss and storage

250
moduli of the films for a spectrum of frequencies. The results of these experiments will

be correlated with the kinetic studies to establish the effeCts of polymerization conditions,
and polymer characteristics such as residual vinyl bond concentration, on the thermal and

mechanical properties of the polymers.

IX.2.6. Investigations of the Propagation Reactions

Based upon Linear Polymerizations

Further insight into the characteristics of these cationic photopolymerizations can
be provided by a series of experiments performed with linear polymerizations of
monofunctional monomers. Crosslinked polymers cannot be characterized using many
traditional experiment techniques which require dissolution of the sample. However the
linear polymerizations of the monofunctional monomers produce polymers which may be
analyzed using traditional techniques such as gel permeation chromatography (GPC) and
solution-state NMR. These techniques may allow mechanistic questions such as the
extent of chain transfer, and the occurrence of possible side reactions to be answered. For
example, the GPC analysis of the resulting linear polymers will allow the primary chain
length to be determined and will provide information about the extent of chain transfer
(which results in shorter chain lengths). One possible side reaction is acetal formation in
cationic polymerizations of vinyl ethers. Solution-state NMR can be used to determine
the extent of acetal formation in these systems. Several monofunctional vinyl ether

monomers have recently become commercially available and can now be used for these

251
studies. These monomers include propenyl ether of propylene carbonate (PEPC),

cyclohexanedimethanol monovinyl ether (CVE) and hydroxybutyl vinyl ether (HBVE).
IX.2. 7. Efiect of Light Source and Intensity

The rate of cationic photopolymerization of divinyl ethers has been shown to be
very sensitive to the intensity of the initiation source. However, more detailed systematic
studies of this effect need to be performed. Since the number of active centers produced
per unit time may influence the network structure, the intensity of initiation will effect the
overall performance of the coating material. Therefore, the kinetic monitoring techniques
described in this work can be used to investigate the effect of intensity on different

monomer, initiator and photosensitizer systems.
IX.2.8. Swelling Studies

The swelling properties and chemical resistance of each film may be thoroughly
characterized using a variety of solvents, including water, methanol and methyl ethyl
ketone. The equilibrium weight fraction of each solvent absorbed by the films can be
determined. Any change in the properties of the films upon exposure to the solvents may
then be thoroughly characterized. Furthermore, the extractable weight fraction of each
sample should be determined. These experiments involve swelling the polymers in
solvent for several days or even weeks to allow any extractable fraction to diffuse out of

the polymer network, then drying the samples to constant weight. Upon removal of the

 

I'IIIIII‘II .Iln’l tlIllll lll'llltlll.tl|lll III IIIIIIIII'IIJIIIIIII Illa... Ii II

|.I.I. IIIII

III III II Ill III 'Ill

252
extractables, the thermal and mechanical properties of the films can again be characterized

to monitor any changes.

   

HICHIGRN STRTE

 

 

 

1| Nil

3‘12

 

 

 

   
 

   

UNIV. LIBRARIE
“Will” llllllllllll"lllllllmllll
93 14210680

1