 

 

 

 

. v D .11.! f‘ 0": .
.‘ldllllw‘ll'j
ll. ‘lu‘.u.ﬁf I
.... ||I|lfv 0 0|! x.‘ .q
. ,v |fnlbux|¥lv . 0 cl.

 

 

 

 

THESIS UNIV RS SITYLBRAR

IIIIIIIIIII||IIIIIIIIIIIIII II IIIII II "II

31293 01691 903

 

 

 

 

 

 

 

 

 

 

 

 

 

 

This is to certify that the

thesis entitled

DESIGN OF INITIATING SYSTEMS FOR ULTRAVIOLET
PHOTOPOLYMERIZATION OF THICK POLYMER
AND COMPOSITE PARTS

presented by

Vijaykumar Narayanan

has been accepted towards fulﬁllment
of the requirements for

M.S. degree in Chem Engineering

 

 

(22“ 2mm”

Major professor

Date OCR é] [377

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

 

 

LIBRARY

Michigan State
Universlty

 

 

 

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

 

MTE DUE MTE DUE DATE DUE

 

'111 4 1 4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

use animus-p.14

 

DESIGN OF INITIATING SYSTEMS FOR ULTRAVIOLET
PHOTOPOLYMERIZATION OF THICK POLYMER AND COMPOSITE PARTS

By

Vijaykumar Narayanan

A THESIS

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

MASTER OF SCIENCE

Department of Chemical Engineering

1997

ABSTRACT

DESIGN OF INITIATING SYSTEMS FOR ULTRAVIOLET
PHOTOPOLYMERIZATION OF THICK POLYMER AND COMPOSITE PARTS

By

Vijaykumar N arayanan

Photopolymerization offers many advantages that could be exploited for polymer
composites processing. In this thesis, several photoinitiator systems were studied to
develop criteria for selecting a suitable initiator for the photocuring of thick polymers and
composites. To aid in this choice, a mathematical model was developed to describe the
photobleaching of photoinitiators dissolved in non-reactive solvents. This model is based
upon the set of coupled partial differential equations describing the spatial and temporal
evolution of the light intensity gradient. The model was veriﬁed experimentally for four

different a—cleavable photoinitiators. A procedure was developed to evaluate the

suitability of several photoinitiators for UV curing of thick polymer and composite parts
based upon their absorption spectra, emission spectrum of the UV light source,
transmittance of the resin and cure time of thick polymer samples. Bis(2,4,6-
trimethylbenzoyl)phenyl phosphine oxide (BAPO) at a concentration of 0.2 wt % was
found to be the most suitable for UV curing of thick vinyl ester samples. Finally, the
mechanical properties of the UV cured composite specimens was found to be comparable

to specimens cured by conventional methods.

ACKNOWLEDGEMENTS

Foremost, I would like to thank my advisor Dr. Scranton for his guidance, support
and encouragement throughout the course of my graduate studies. I will always look to

him for advice in the future too.

I would like to thank Khanh and Rick, for contributing to this work by carrying

out several experiments. Their work is greatly appreciated.

. I would like to express my appreciation to Julie and Shail, for acquainting me

with the lab and for their advice on several matters both technical and non-technical.

I would like to thank Arvind, Bernie, Bharath, Katy and Kiran for making my

experience in the research group interesting and enjoyable.

Finally, I would like to thank my parents for their support and encouragement to

pursue graduate studies.

iii

TABLE OF CONTENTS

CHAPTER 1

INTRODUCTION

1.1 Photopolymerizations ................................................................................................... l
1.2 Photopolymerization: Advantages and Application ..................................................... 1
1.3 Photopolymerization of Composites ............................................................................ 2
1.4 List of References ......................................................................................................... 4
CHAPTER 2

BACKGROUND

2.1 Issues associated with photopolymerization of composites ......................................... 5
2.2 Photocured dental composites ...................................................................................... 6
2.3 Industrial Composites ................................................................................................... 8
2.4 Initiators for Photopolymerization of Composites ..................................................... 13
2.5 List of References ....................................................................................................... 18
CHAPTER 3

RESEARCH OBJECTIVES .......................................................................................... 20
CHAPTER 4

MATHEMATICAL MODEL FOR PHOTOINITIATION OF THICK POLYMERS
PARTS: MONOCHROMATIC PHOTOBLEACHING IN SOLUTION

4.1 Introduction ............................................................................................................... 21
4.2 Description of the System Being Modeled ................................................................ 23
4.3 Materials and Method ................................................................................................ 24
4.4 Development of the Model ....................................................................................... 25
4.5 Results and Discussion .............................................................................................. 27
4.6 Effect of Diffusion ..................................................................................................... 29
4.7 Conclusions ............................................................................................................... 32
4.8 List of References ...................................................................................................... 44
4.9 List of Symbols .......................................................................................................... 45
4.10 Appendix ................................................................................................................... 46

CHAPTER 5
INVESTIGATION OF PHOTOINITIATING SYSTEMS FOR
PHOTOPOLYMERIZATION OF COMPOSITES

5.1 Introduction ................................................................................................................ 55
5.2 Experimental .............................................................................................................. 57
5.3 Results and Discussion ............................................................................................... 60
5.4 Conclusions ................................................................................................................ 68
5.5 List of References ....................................................................................................... 81
CHAPTER 6

SUMMARY AND RECOMMENDATIONS

6.1 Summary .................................................................................................................... 82
6.2 Recommendations for Future Work ........................................................................... 83
6.3 List of References ....................................................................................................... 85

LIST OF TABLES

Chapter 4
Table 4.1 Comparison of theoretical and best fit values of y for different
photoinitiators ............................................................................................... 38
Chapter 5

Table 5.1 Optimum concentration and cure times for three different photoinitiators
for curing 0.9 cm thick sample of Derakane 510A with a light intensity
of 65 mW/cm2 UVA .................................................................................... 67

Table 5.2 Effect of sample thickness on optimum concentration of BAPO ................. 69

vi

LIST OF FIGURES

Chapter 2

Figure 2.1 Molecular structure of Benzoin ethyl ether (BEE), a representative 0t-
cleavage photoinitiator ................................................................................. 16

Figure 2.2 Absorption spectrum of BEE showing the 1t—>1t*and n—>1t* absorption
bands ............................................................................................................. 16

Figure 2.3 Photochemistry of BEE in solution .............................................................. 17
Chapter 4

Figure 4.1 Time evolution of the absorbance spectrum of a bis-acyl phosphine
oxide (BAPO) photoinitiator as it is irradiated with low intensity UV

light. The 2.5 cm thick sample containing 5.1 x 10'4 g ml"l of the
initiator dissolved in acetone was illuminated from above with ﬁltered
light from a mercury lamp. The light was passed through a 10 nm
bandpass ﬁlter centered at 360 nm, and the intensity at the top surface of
the sample was 2 mW cm’2 .......................................................................... 34

Figure 4.2 Elimination of the intensity gradient in a 1cm thick solution of BEE.
The concentration of BEE is 5 x10'4 g/ml and the incident light intensity
is 5 mW/cm2 at 330 nm ................................................................................ 35

Figure 4.3 Elimination of the light intensity gradient in a 1cm thick solution of

TPO. The concentration of TPO is 5 x 10'4 g/ml and the incident light
intensity is 5 mW/cm2 at 360nm .................................................................. 35

Figure 4.4 Change in transmittance as a function of time for a 1cm thick solution of
TPO irradiated with 360nm light of intensity 5 mW/cm2 ............................ 36

Figure 4.5 Change in transmittance as a function of time for a 3cm thick solution of
TPO irradiated with 360nm light of intensity 5 mW .................................... 36

vii

Figure 4.6

Figure 4.7

Figure 4.8
Figure 4.9
Figure 4.10
Figure 4.1 1
Figure 4.12
Figure 4.13

Figure 4.14

Figure 4.15

Figure 4.16

Figure 4.17

Change in transmittance as a function of time for a 5cm thick solution
of TPO irradiated with 360 nm light of intensity 5 mW/ cm2 ...................... 37

Effect of solvent absorbance on the evolution of the transmittance of a
3cm thick solution of TPO in a solvent with an extinction coefficient of
0.5 ml/gm.cm and 0.8 gm/ml density. The concentration of TPO is 5 x
10'4 g/ml and intensity of the incident light is 5 mW/cm2 ............................ 37

Comparison of theory and experiment for photobleaching of TPO ............. 38
Comparison of theory and experiment for the photobleaching of TPO ....... 39
Comparison of theory and experiment for the photobleaching of TPO ....... 39
Comparison of theory and experiment for the photobleaching of BAPO 40
Comparison of theory and experiment for the photobleaching of BAPO 40
Comparison of theory and experiment for the photobleaching of BEE ....... 41

Comparison of theory and experiment for photobleaching of substituted
titanocene ...................................................................................................... 41

Effect of diffusion on the photobleaching rate of BEE. The

concentration is 10'3 g/ml and the light intensity is 2.8 mW/cm2 @
330nm ........................................................................................................... 42

Effect of diffusion on the photobleaching rate of BEE. The

concentration is 10'3 g/ml and the light intensity is 10 mW/cm:2 @
330nm ........................................................................................................... 42

Effect of diffusion on the photobleaching rate of BEE. The

concentration is 3.3 x 10'3 g/ml and the light intensity is 10 mW/cm2 @
330nm ........................................................................................................... 43

viii

Figure 4.18 Effect of diffusion on the photobleaching rate of BEE. The

concentration is 2.79 x 10'4 g/ml and the light intensity is 0.8 mW/cm2
@ 330nm ...................................................................................................... 43

Figure 4.19 Comparison of experimental and model data (with and without
diffusion) for the photobleaching of BEE .................................................... 44

Figure 4.20 Effect of initiator diffusion coefﬁcient on the photobleaching process.

The light intensity is 10 mW/cm2 and the concentration is 3.3 x 10'3
g/ml. All the intensity gradients shown are at time = 900 seconds ............. 44

Figure 4.21 Comparison of the photobleaching rate of BAPO and BEE. Both the
solutions had the same initial absorbance of 1 and solution depth of 2

cm. The light intensity used for bleaching was 2.2 mW/ cm2 ...................... 45
Chapter 5

Figure 5.1 Variation of temperature as a function of time at the bottom of thick
samples which are photopolymerized under adiabatic conditions ............... 74

Figure 5.2 Transmittance spectrum of Derakane 510A in the 300-500 nm range.
The path length used was 1 cm .................................................................... 74

Figure 5.3 Plot of the extinction coefﬁcient of Derakane 510A as a function of
wavelength in the 350-550 nm range. The density of the resin is 1.2
gm/ml. .......................................................................................................... 75

Figure 5.4 Comparison of the transmittance spectra of three different resins in the
300-550 nm range. The path length used was 1 cm in all cases .................. 75
Figure 5.5 Emission Spectrum of the 1000 W Hg vapor lamp in the 280-500 nm

region ............................................................................................................ 76

Figure 5.6 Emission spectrum of 1000 W Hg vapor lamp in the 280-500 nm region
after it is transmitted through 0.9cm thick layer of Derakane 510A resin. .. 76

Figure 5.7

Figure 5.8

Figure 5.9

Figure 5.10

Figure 5.11

Figure 5.12

Figure 5.13

Figure 5.14

Figure 5.15

Figure 5.16

Figure 5.17

Absorption spectrum of Benzoin Ethyl Ether. The concentration is 5.1 x

10'4 g/ml in acetone ..................................................................................... 77
Absorbtion spectrum of TPO. The concentration is 6.4x 10'4 g/ml in 2-
propanol ........................................................................................................ 77

Absorption spectrum of BAPO. The concentration is 5.1 x 10'4 g/ml in
acetone .......................................................................................................... 78

Optimum concentration of TPO for curing 0.9 cm thick sample of

Derakane 510A. The light intensity used was 65 mW/cm2 UVA ............... 78
Optimum concentration of BAPO for curing 0.9 cm thick sample of

Derakane 510A. The light intensity used was 65 mW/cm2 UVA. ............... 79
Optimum concentration of BEE for curing 0.9 cm thick sample of

Derakane 510A. The light intensity used was 65 mW/cm2 UVA. ............... 79
Radicals formed by photolysis of TPO ......................................................... 8O
Radicals formed during complete photolysis of BAPO ............................... 81

Comparison of the absorption spectra of TPO and BAPO in the 320—450
nm spectral region ........................................................................................ 82

Maximum temperature measured by the thermocouple during BAPO
initiated photopolymerizations of 0.9 cm thick samples of Derakane
510A. The light intensity used was 65 mW/cm2 UVA ............................... 82

Maximum temperature measured by the thermocouple during TPO
initiated photopolymerizations of 0.9 cm thick samples of Derakane
510A. The light intensity used was 65 mW/em2 UVA .............................. 83

Figure 5.18 Effect of light intensity on cure time of 0.9 cm thick sample of Derakane
510A ............................................................................................................. 83

Figure 5.19 Comparison of emission spectra of Hg vapor lamp after passing through
0.9 cm thick sample of Derakane 510A resin with and without BAPO
photoinitiator ................................................................................................ 84

Figure 5.20 Absorbance spectrum of Bis [ms—2,4—Cyclopentadien-1-yl] bis[2,6-
difluoro-3-[lH-Pyrrol-l-yl] phenyl titanium. The concentration is 2.4
x10‘3 g/ml in acetone .................................................................................... 84

Figure 5.21 Absorption spectrum of Camphorquinone in the 380-520 nm region.
The concentration is 5.01 x 10’3 g/ml in acetone ......................................... 85

Figure 5.22 Comparison of ﬂexural strength of composites. The bar represents the

average value (of 5 samples) while the error bars represent the high and
low values obtained in the tests .................................................................... 85

xi

Chapter 1

Introduction

1.1 Photopolymerization
Photopolymerization is generally a chain polymerization in which the propagating
active centers (usually radicals or cations) are produced by a photochemical event. A
wide variety of chemical compounds (photoinitiators) are available that produce free
radicals upon absorption of photons of the appropriate energy (typically in the UV or
visible region of the spectrum). The photochemical mechanism of active center
generation has been well characterized, and generally proceeds by (it-cleavage, B-cleavage

or hydrogen abstraction1 ’2.

1.2. Photopolymerization: Advantages and Applications

Light-induced polymerizations offer many advantages that may be exploited in
industrial processes. For example, photopolymerizable formulations are typically
solvent-free, which is important for coating applications because volatile organic
emissions are a major concern. As active centers may be produced rapidly and efﬁciently
using photochemical processes, Photopolymerization offers high production rates and is
very energy efﬁcient compared to thermal systems in which the entire reaction system is
raised to elevated temperatures. Finally, photopolymerizations provide a great deal of
control over the initiation reaction: spatial control, because the light may be directed to

locations of interest in the system; and temporal control, because the light may be readily

shuttered on or off.

The advantages afforded by photopolymerization have led to tremendous growth
in applications of these reactions. Light-induced polymerizations are now widely used for
ﬁlms, inks and coatings on a variety of substrates including paper, metal, plastic and
wood; dental materials for which the initiating light is routed into the mouth using optical
ﬁbers; and a variety of high-tech and electronic applications such as coatings on optical
fibers, replication of optical disks, and fabrication of printed circuit boards. Finally, a
relatively recent application of radiation curing is in the area of stereolithography for the
production of three-dimensional prototypes using data generated by a computer aided

design (CAD) software package.

1.3. Photopolymerization of Composites

While the traditional thin-ﬁlm applications of photopolymerization are
experiencing rapid growth in their own right, the advantages afforded by these reactions
also make them attractive for the development of new processes for producing ﬁber-ﬁlled
polymer composites. Most notably, the spatial and temporal control of initiation offered
by photopolymerizations provides ‘cure on demand’, which could be used to achieve
short cycle times while circumventing many problems and limitations of current
composites production methods. Composites for the durable goods industry are typically
produced by resin transfer molding (RTM) or filament winding processes in which a free
radical polymerization is triggered using thermal initiators. In these thermally initiated

processes, the time required for heat transfer places a lower limit on the cycle time, and

3

attempts to reduce the cycle time by preheating the mold (in RTM) or the resin (in
ﬁlament winding) invariably lead to problems associated with premature cure, or very
complicated ﬂow patterns. Finally, the energy efﬁciency and low capital cost associated
with UV curing may provide additional beneﬁts. Photopolymerization may offer
enhanced energy efﬁciency because it generally is performed at room temperature without
external heating. In addition, the light sources used in photopolymerization may be
considerably less expensive than the large autoclaves used for thermal polymerizations.

As an example of how the cure on demand afforded by photopolymerizations
could lead to improvements in composites manufacturing, it is useful to consider the
RTM process. In this method, a reacting resin is injected into a closed mold containing
the ﬁber preform. Short cycle times necessitate a careful balance between the conflicting
requirements of rapid mold ﬁlling and high polymerization rates. This trade-off leads to a
very complicated flow process in which a reacting liquid is forced through a porous
medium (the preform). As the liquid reacts it becomes more viscous (actually
viscoelastic), leading to several important problems associated with high operating
pressures, poor resin impregnation into the ﬁbers, preform displacement, and gelation of
the resin in the transfer lines. These problems, which ultimately arise from the fact that
the mold-ﬁlling and reaction are coupled to one another, are exacerbated by attempts to
reduce the cycle time by using more reactive resins.

The temporal control offered by photopolymerizations could be exploited to
alleviate many of the problems associated with RTM by: (1) allowing the mold to be

completely ﬁlled with a low-viscosity resin before any reactions takes place, and (2)

4

initiating a fast thermoset reaction after the mold is completely ﬁlled3. This decoupling of
mold ﬁlling and reaction would permit complete control of the time allowed for micro-
flow, and would reduce waste and expensive clean-up by preventing premature reaction
(gelation in the transfer lines). In addition, unlike traditional processes, the photoinitiated
reactions are not limited by the rate of heat transfer from the mold to the resin, and would
consequently offer short cycle time. Finally, since there would be no need for high
pressures or external heating of the die, the proposed process would be very energy
efﬁcient. Other composite processing methods such as ﬁlament winding would beneﬁt in

a similar manner from the cure on demand afforded by photopolymerizations.

1.4 List of References

1. JP. Fouassier, Photoinitiation Photopolymerization and Photocuring Fundamentals
and Applications, Hanser Publishers, 1995

2. JP. Fouassier in Radiation Curing in Polymer Science and Technology (Vol. 2)
(Fouassier, J. P., and Rabek, J.F., eds), pp. 1-61, Elsevier Applied Science, 1993

3. LS. Coons, B. Rangarajan, D. Godshall, A.B. Scranton In Innovative Processing
and Characterization of Composite Materials (Vol. 20) (Gibson, R.T., Chou, T.W.,
Raju, P.K., eds), pp.227-240, ASME, New York, 1995

Chapter 2

Background

2.1. Issues associated with photopolymerization of composites

Photopolymerizations of thick and ﬁber-ﬁlled polymers are more challenging than
polymerizations of thin-ﬁlm systems due to the exponential reduction in light intensity
through the sample resulting from absorption and scattering. For this reason, issues that
are relatively unimportant for thin-ﬁlm photopolymerizations can make the difference
between success and failure for thick and ﬁber-ﬁlled systems. For example, an initiator
concentration that is too high, or even a small resin extinction coefﬁcient at the initiating
wavelength can prevent efﬁcient cure throughout the sample. Therefore, for thick
polymers and composites, proper selection of the initiator formulation and illumination
wavelength are imperative to ensure that the samples cure throughout (including the
portions of the sample furthest from the light).

For photopolymerization of composites for the durable goods industry, light from
the near UV region of the spectrum between 330 and 420 nm is the most convenient
initiating source. This arises as a result of several factors. First, visible light is usually
avoided to prevent premature cure triggered by ambient light. In addition, deep UV light
is inappropriate due to absorbance by the resin and/or ﬁbers. Finally, there are several
commercially available initiators that absorb in this region, and inexpensive mercury
lamps have prominent emissions in this spectral region.

As mentioned previously, careful selection of the illumination source, initiator

6

system (type of initiator, photosensitizer, concentration etc.), resin, ﬁber reinforcement
and composite thickness are imperative for successful photopolymerization of
composites. Thus the cure behavior of the composite is dependent upon several different
variables, which must be carefully controlled. Similar problems have been encountered
in other applications such as photocurable dental restoratives. Hence, it is useful to
brieﬂy examine the literature in this area to obtain some insight into the strategies that

could be used for photopolymerization of composites.

2.2 Photocured dental composites
Photopolymerization has been used in dental applications since 1970, when UV-

]. In

cured ﬁssure sealants were reported using benzoin methyl ether as the photoinitiator
the past few decades, photopolymerization has become prominent for curing dental

materials, and only representative references related primarily to depth of cure issues will

be discussed here. Some applications and issues associated with photopolymerization of

dental materials are described in recent reviews2’3.

Most photocured dental materials
consist of a multifunctional acrylate or methacrylate resin, an initiator and a powdered
glass reinforcement with particle size between 0.7 and 50 um. Today, dental composites

are cured using visible (blue) light due to safety considerations and ease of handling.

The depth of cure is an important aspect of photocured dental composites and
numerous investigators have studied this issue. For example, Kilianz, Cook4, Salako and

Cruickshanks-Boyd5 and Tirtha et (11.6 characterized the hardness of photocured dental

composites as a function of sample depth (several millimeters) and have concluded that

7
the hardness decreases considerably from top to bottom (due to the variation of light
intensity with sample depth). Cook4 conducted a theoretical study of the variation of cure
time with sample thickness and predicted a logarithmic relation between cure depth and
time, in agreement with his experimental results. This model was based upon the
assumption of a static light intensity gradient, and included the effects of an inhibitor that

prevents polymerization below a threshold active center concentration.

Kawaguchi et (11.7 characterized the relationship between cure depth and
transmission coefﬁcient of visible-light-initiated dental composites up to seven
millimeters thick and ﬁtted the results to a linear correlation. These authors also
investigated the effect of particle size on the depth of cure and found that microﬁlled

resins with particle size of 0.04 pm exhibited a lower cure depth than resins ﬁlled with

particles of size ranging from 0.2 to 50 pm. This effect was attributed to enhanced

scattering by the smaller particles. Leung et al.8 studied the post-irradiation cure of

dental composites and reported a continuous increase in hardness that plateaued after

about a day, which they attributed to dark-cure. Similarly, Kolarik et al.9 found a

perceptible increase in glass transition temperature and ﬂexural modulus of the composite

after storage in the dark. In addition, both Kolarik et 01.9 and Taira et (11.10 reported that
elevating the temperature of the illuminated sample resulted in signiﬁcant post-cure (as

evidenced by enhanced thermal and mechanical properties), which they attributed to the

presence of residual living radicals. Finally, Maffezzoli et al.” reported that the degree
of cure of the dental resins under nonisothermal conditions (when the heat of

polymerization is not immediately removed, thereby causing the temperature to increase)

8

was considerably higher than the degree of cure under controlled isothermal conditions.
This review of the dental composites literature provides some important
information that can be exploited for photopolymerization of ﬁber glass composites for
the durable goods industry. Foremost, the work on dental composites has demonstrated
that photopolymerization can be used to cure relatively thick samples (not just thin ﬁlms),
and has provided guidelines for the design of these reaction systems. Indeed, in some
ways the photopolymerization of industrial composites are easier than their dental
counterparts because fewer constraints are placed on the initiation wavelengths, reaction
temperature etc. For example, the industrial composites could be cured with UV rather
than visible light, and a higher temperature may be tolerated (burning the patient is a
concern for dental composites) and can be exploited to obtain higher conversions at

shorter cycle times.

2.3 Industrial composites
While there are literally hundreds of papers covering photopolymerization of
dental composites, there are relatively few articles on photopolymerization of composites
for industrial (durable goods) applications. The discussion of this area will be organized
in two sections. It will begin with a survey of laboratory studies of photopolymerization
in thick and ﬁber-ﬁlled systems to illustrate the trends and operating principles that have
been identiﬁed by several research groups. Next a few commercial applications of

composites photopolymerization that have been reported to date will be discussed.

9

Laboratory studies of composite photopolymerizations

In a series of papers, Ogale and co—workerslz'l5

examined photopolymerization of
acrylate - glass ﬁber systems for three-dimensional stereolithography applications. These
investigators examined a variety of glass reinforcements, including continuous and
chopped glass ﬁbers, as well as 55 um diameter borosilicate glass spheres. Based upon

measurements of the mechanical properties of the resulting composites, these authors

concluded that photolithography of composites is indeed feasible'2’13. These

investigators also used dynamic mechanical analysis to examine the effects of glass,

quartz and carbon ﬁbers on the photopolymerization rate of acrylate composites”. The
authors reported that the glass and quartz ﬁbers had essentially no effect on the cure rate,
while the carbon ﬁbers reduced the rate of photopolymerization due to absorption of

light.

Shi and Ranbyl6 reported photopolymerizations of glass—ﬁber-polyester
composites. These authors used unsaturated polyester resins that were end-capped with
acrylate groups, and included multifunctional acrylic, allylic or acrylether comonomers as
reactive diluents. Chopped e—glass ﬁber mats were used for the composite reinforcement,
and laminates that were approximately 2 mm thick (containing four or ﬁve layers of the
glass ﬁber mats) were prepared. The investigators found that the laminates could be
cured rapidly upon illumination with a high-pressure mercury lamp, and that the resulting
composites exhibited excellent mechanical properties. In addition, they found that the
acrylate or acrylether reactive diluents enhanced the mechanical properties more than the

allylic comonomers. They attributed this trend to fact that the allylic comonomers

10
undergo hydrogen abstraction reactions with the initiator radicals'é. Shi and RanbyI7
also prepared glass-reinforced composites using dendritic methacrylated polyesters as the
matrix. They found that photopolymerization allowed the composites to be cured more
rapidly than thermal polymerization with no deleterious effects on the mechanical
properties. The authors concluded that the dendritic polyesters are promising for

industrial glass-ﬁber-reinforced composites.

Coons et al.‘8 investigated photopolymerizations of thick polymers and glass-
ﬁber-ﬁlled composites using commercially available vinyl ester resins and inexpensive
benzoin ether initiators. These investigators have examined two strategies for efﬁcient
photopolymerization of thick polymers samples and composites using UV light as the
sole (external) initiating source. The ﬁrst method involves careful selection of the
initiating wavelength (such that the resin and ﬁbers are transparent to the initiating
wavelength) and use of photobleaching initiators to ensure penetration of light into the

sample. Photobleaching occurs when the products of photolysis exhibit less absorbance

1,19

than the photoinitiator at the wavelength of interest For example, the efﬁcient

photobleaching of the initiator bis(2,4,6-trimethylbenzoyl) phenyl phosphine oxide is
illustrated in Figure 4.1 (Chapter 4). The ﬁgure illustrates that the broad absorption peak
between about 360 and 400 nm decreases dramatically in intensity as the illumination

time is increased, thereby allowing the light to penetrate more deeply into the sample.

[.18

Coons et a characterized the effects of process variables such as light intensity, cure

time, initiator concentration and ﬁber loading on the mechanical properties of the ﬁnal

20

composite. In agreement with previous results for unﬁlled systems , an optimum

initiator concentration was observed above which or below which the

photopolymerization was less efficient.

The second composite polymerization strategy investigated by Coons et al. '8 was
a dual-cure strategy in which both a photoinitiator and a thermal initiator are included in
the reaction mixture. Here, the heat evolved from an exothermic photopolymerization
elevates the temperature and results in the production of additional active centers by the
thermal initiator. In experiments performed using standard vinyl ester resins,
measurement of temperature proﬁles in situ revealed that the reaction exotherm is high
enough to ensure adequate polymerization and property development without leading to
thermal degradation. These results suggest that the dual cure strategy offers both the
temporal control of the start of the reaction afforded by the photopolymerization, as well

as the enhanced reaction rate and completeness of cure provided by the thermal initiation.

Medenouvo and Fouassier”

investigated photopolymerizations of unsaturated
polyester - styrene resins ﬁlled with glass ﬁbers. In agreement with the studies discussed
above, the authors found that the mechanical properties of the photocured composites
were similar to those produced using conventional methods. One objective of their study
was to identify an initiator system that makes optimum use of the light emitted from a
number of light sources; therefore they experimented with a variety of photoinitiators,
photosensitizers, thermal initiators and additives. Based on their experiments they
reported an optimum formulation containing 0.15 wt% phenonyl oxime ester (the

photoinitiator), 0.02 wt% eosin (a photosensitizer), 0.02 wt% benzanthraquinone (a

photosensitizer), 1 wt% methyl diethanolamine (an electron-donating additive), and 0.25

l2

wt% benzoyl peroxide (a thermal initiator) with the remainder of formulation made up of

the monomer to be polymerized.

Our ﬁnal example of laboratory studies on composite photopolymerizations is

taken from the patent literature. Bellobono25 reported organometallic photoinitiators that,
when used with organic photosensitizers, are able to efﬁciently photoinitiate in the near
UV (300-400 nm) region. Bellobono used his novel initiator systems to photopolymerize
styrenated polyester resins ﬁlled with glass or carbon-ﬁbers to produce composites with
thicknesses of 1-3 mm. In addition he reported that it is possible to prepare composites as

thick as a few centimeters.

Commercial composite photopolymerization systems

As the manufacture of industrial composites by photopolymerization is in its
infancy, few commercial systems have been reported to date. One notable commercial
photopolymerization system for ﬁber-glass composites is a UV-cured vacuum-assisted
RTM process23 developed by Synergistic Composite Systems Inc., using
photopolymerizable vinyl ester resins from Sunrez Corporation. The company claims the
cure on demand afforded by photopolymerization as an important advantage of their
process because mold-ﬁlling problems during infusion can be corrected before the UV

lamps are turned on”. In addition, the resins can withstand temperatures up to 65°C

without curing; therefore the resin may be heated to reduce its viscosity”, thereby
improving mold ﬁlling and fiber impregnation. The process was used to cure relatively

thick composites in a reasonable period of time. For example, 0.5 in (1.2 cm) thick

13

laminates were cured in 7-10 minutes”. The process was tested for manufacture of a

half-scale mid—ship section of a naval combat ship23.

A second composite UV-cure process that has been commercialized is the

AccusetTM ﬁlament winding resin reported by Loctite Corporation”. In the Loctite

process, a photopolymerization stage (after resin impregnation, before winding) is used to
overcome many of the problems and limitations of conventional ﬁlament winding
processes by producing a viscous, gel state that immobilizes the resin and reduces resin
loss. After winding is complete, the cure is completed by a second, thermal stage. The
resin formulations typically contain ~14 wt% of a UV-curable acrylate monomer, with the
remainder of the formulation made up of a standard, thermally-cured, epoxy-based resin.
The company claims that the addition of the UV stage offers many advantages, including
faster winding speeds, shorter cure times, instant resin immobilization (which prevents

resin dripping), and enhanced control of ﬁber placement and ﬁber volume.

2.4. Initiators for Photopolymerization of Composites
A photoinitiator is the chemical species which produces active centers (free
radicals or cations) upon absorption of photons of the appropriate wavelength. In general,
free radical photoinitiators fall into two categories: cleavable and non-cleavable
photoinitiators. The cleavable photoinitiators directly produce free radicals upon
absorption of photons and are further classiﬁed according to the mechanism by which

they undergo photolysis as (it-cleavable, B-cleavable and y-cleavable initiators. The non-

cleavable initiators require a co-initiator such as an amine (electron transfer agent) or an

14

alcohol (a hydrogen donor) in order to produce free radicals.
Most of the commonly used free radical photoinitiators fall under the category of
a-cleavable initiators. Such initiators include several families of compounds such as

benzoin ethers, dialkoxyacetophenones, hydroxy alkyl ketones, benzoyl phosphine

oxides, benzoyl oxime esters, morpholino ketones and amino ketoneszs. Benzoin ethyl

ether (Figure 2.1) is a representative oc—cleavable photoinitiator whose photochemical

mechanism has been well characterized26’27‘28’29.

Figure 2.2 shows the absorption
spectrum of benzoin ethyl ether. It has two distinct absorption bands one in the deep UV

region (below 300 nm) and a second smaller band in the UVA region between 300 and

390 nm. Such a spectrum is characteristic of ketone30. The more intense absorption

. * . . . * .
band in the deep UV corresponds to a 1t—>1t electronic transuion. The 1t—>tt electronic

transition occurs when a it electron of the keto group is promoted to an excited state after

. . , =l<
absorblng a photon. The smaller band in the near UV region corresponds to a n—>1t

electronic transition, which occurs when one of the lone pair of electrons on the oxygen
of the keto group is promoted to an excited state after absorbing a photon of appropriate
wavelength. Figure 2.3 shows the sequence of reactions that occurs when a solution of
BEE in benzene is irradiated with light in the UVA region. The BEE upon absorption of
a photon is promoted to the excited singlet state. It then undergoes efﬁcient intersystem
crossing to the triplet state after which it undergoes photolysis to give rise to two free
radicals (a benzoyl radical and an ethoxybenzyl radical). The above steps occur very

rapidly and the kinetic constants are as shown in Figure 2.3. In the absence of oxygen,

15
the free radicals further react and finally form benzaldehyde, benzil and pinacol ether327.

In the presence of oxygen, however, the radicals may be captured, resulting in the

formation of benzoic acid and esters of benzoic acid”. Finally, if unsaturated

compounds (such as styrene) are present, then the benzoyl free radicals preferentially add

on to the double bonds, thereby initiating a chain polymerization”. The kinetic constant

C-1 25

for oxygen quenching of the free radicals is about 109 1 mol‘1 se and the kinetic

-l 29.

constant for reaction with styrene is about 105 1 mol'1 sec However, in usual

3

solution photopolymerizations, the concentration of dissolved oxygen is about 10' mol/l

and that of the monomer is about 101 mol/l, so the rate of oxygen quenching and
initiation have the same order of magnitude. In bulk polymerizations, however, the
concentration of dissolved oxygen is much lower and the concentration of monomer is
much higher, hence the oxygen quenching reaction does not compete with the initiation

reaction.

oczhg
II I
o H

Figure 2.1. Molecular structure of Benzoin ethyl ether (BEE), a representative 0t-
cleavage photoinitiator

 

1.6 {~

.4
..
14*
I
.t
q

d
I
N
IIJA
I

 

Absorbance

 

 

 

 

0.8T
0.65-
0.45-

l
0.2:—
or.
200 250 300 350 400

Wavelength (nm)

Figure 2.2. Absorption spectrum of BEE showing the 1t—>1t*and n—>1t* absorption
bands

It

Figure 2.3. Photochemistry of BEE in solut

,/ki$c>1011 sec’1

OC2H5 3

I1

ion

10.

ll.

12.

13.

14.

15.

16.

17.

2.5. List of References
M.G. Buonocore, J. Am. Dent. Assn. 80, 324, 1970

RJ. Kilian, in Polym. Sci. and Tech. (Vol. 14), Charles G. Gebelein and Frank K.
Koblitz, eds, pp. 411-417, Plenum Press, 1980

K.S. Anseth, S.M. Newman, C.N. Bowman, in Advances in Polymer Science (Vol.
122), Peppas, N.A., Langer, R.S., eds, pp. 177-217, Springer-Verlag, 1994

W.D Cook, J. Dent. Res. 59, 800, 1979

NO. Salako, D.W. Cruickshanks—Boyd, Brit. Dent. J. 146, 375, 1979

Tirtha, R., Fan, P.L., Dennison, J.B., Powers, J.M., (1982) J. Dent. Res. 61, 1184
M. Kawaguchi, T. Fukushima, K. Miyazaki, J. Dent. Res, 73, 516,1994

R.L. Leung, P.L. Fan, W.M.Jonston, J. Dent. Res, 62, 363, 1983

J. Kolarik, C. Migliaresi, P. Capauana, L. Fambri, Polymers in Medicine, 4, 145,
1991

M. Taira, A.M. Khan, K. Ohmoto, N. Satou, H. Shintani, K. Wakasa, M. Yamaki, J.
Mat. Sci. Lett. 13, 1229, 1994

A.M. Maffezzoli, R. Terzi, L. Nicolais, J. Mat. Sci. :Materials in Medicine 6, 161,
1995

T. Renault, A.A.Ogale, R.L. Dooley, A. Bagchi, C.C. Jala-Almonte, SAMPE
Quaterly 22, 19, 1991

T. Renault, A.A.Ogale, R.L. Dooley, A. Bagchi, C.C. Jala-Almonte, SAMPE
Quaterly 23, 28, 1991

T. Renault, A.A. Ogale, MJ. Drews, Proc. ANTEC ’93, 2352, 1993
T. Renault, A.A. Ogale, Polym. Eng. Sci. 36,551, 1996
Winfang Shi, Bengt Ranby, J. of App. Polym. Sci. 51, 1129, 1994

Wenfang Shi, Bengt Ranby J. Appl. Polym. Sci. 59, 1951, 1996

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

19

L.S. Coons, B. Rangarajan, D. Godshall, and AB. Scranton, ACS Symposium
Series, in press

J. Finter, M. Riediker, O. Rohde, B. Rotzinger, Makromol. Chem., Macromol.
Symp. 24, 177, 1989

J. Guthrie, M.B. Jaganathan, M.S. Otterburn, J. Woods, Polym. Bull. 15, 58, 1986
F. Medenouvo, J .P. Fouassier, Angew. Makromol. Chem. 229, 29, 1995

LR. Bellobono, European Patent 0 421 512 A1, 1991

A. Hudson, Composites Technol., 2, 24, 1996

Y. Okamoto, P. Klemarczyk, S. Levandoski, V. Hanlon, J. Bremmer, SME
Technical Paper EM94-l 1 1, l, 1994

JP Fouassier in Radiation Curing in Polymer Science and Technology (Vol. 2)
(Fouassier, J. P., and Rabek, J.F., eds), pp. 1-61, Elsevier Applied Science, 1993

ED. Lewis, R.T. Lauterbach, H.G. Heine, W. Hartmann, H. Rudolph, J. Am. Chem.
Soc., 97, 1519, 1975

H.G. Heine, H.J. Rosenkranz, H. Rudolph, Angew. Chem, Int. Ed. Engl, 11, 974,
1972

S. P. Pappas, A. Chattopadhyay, J. Am. Chem. Soc., 95, 6484, 1973
R. Kuhlmann, W. Schnabel, Angew. Makromol. Chem, 70, 145, 1978

J .D. Coyle, Introduction to Organic Photochemistry, John Wiley & Sons, 1986

Chapter 3

Research Objectives

From the introduction and background provided, it is clear that photopoly-
merizations offer tremendous advantages over traditional thermal methods, including low
energy requirements, spatial and temporal control of initiation, and high polymerization
rates By careful system design, it is possible to come up with a commercially feasible
process for photopolymerization of composites. Indeed, few commercial processes for
manufacture of composites for the durable goods industry already exist and there is
considerable scope for improvement in technology.

The general objective of this research project is the development of photopoly-
merization processes for high-speed, low-cost production of polymeric composites. To

achieve this general objective, the speciﬁc objectives were:

i. To quantify the photobleaching effect of photoinitiators to aid in making a choice of
photoinitiator for curing of thick polymer and composite parts

ii. To establish a technique for the design of photoinitiating systems for curing of thick
polymer and composite parts

To show that UV curing is a viable technology for manufacture of composites

2O

Chapter 4

Mathematical Model for Photoinitiation of Thick Polymers
Parts: Monochromatic Photobleaching in Solution

Abstract
Many a-Cleavable photoinitiators exhibit a photobleaching effect in which the

absorbance decreases with illumination time when exposed to light of the proper
wavelength. For example, benzoin ethers exhibit photobleaching of the absorbance peak
between 320 and 365nm. This occurs because the products of photolysis do not absorb in
the same region as the original initiator molecule. The net result of the photobleaching is
an increase in transmittance of the photoinitiator system as a function of illumination
time, as long as the other components of the system (e.g. the monomer) do not absorb
signiﬁcantly in the same wavelength range. The phenomenon of photobleaching is
particularly important in the photocuring of thick polymer parts. In this chapter, a
mathematical model is developed for the photobleaching effect of photoinitiators
dissolved in non-reactive solvents. Given the incident light intensity and wavelength and
the initial photoinitiator concentration, the model predicts the variation of the
photoinitiator concentration and light intensity with time and sample depth. The model

was veriﬁed experimentally for four different a—cleavable photoinitiators

21

22

4.1. Introduction
For photopolymerization of thick systems, one of the criteria for choice of
photoinitiator is its ability to undergo photobleaching. In general, when a molecule is
irradiated with light of a wavelength in which it absorbs, and is promoted to the excited
singlet state, it may undergo a number of photophysical processes such as ﬂuorescence,

intersystem crossing to the triplet state, phosphorescence, photolysis, hydrogen

abstraction, singlet or triplet state quenching etc]. However, most of the commonly used

(it—cleavable free radical photoinitiators undergo a fairly simple sequence of

photochemical transitions. They are excited to the singlet state, undergo efﬁcient
intersystem crossing to the triplet state and subsequently undergo photolysis to form free
radicals. The quantum yield of each of these steps is close to unity. Once formed, the
free radicals may further react with one another or with other species in solution (such as
oxygen) to produce other products, and if monomer is present, to initiate polymerization.
If the absorption of the photolysis products is less than that of the photoinitiator
molecule, then the absorbance of the solution decreases with irradiation time (as shown in

Figure 4.1), allowing more light to pass through it. This phenomenon is known as

photobleachingz. For example, the photolysis products of benzoin ethyl ether (BEE)

when irradiated with UV light in a solution of deoxygenated benzene are benzil,

benzaldehyde and picanol ethers3

. These molecules are considerably less absorbing at
365 nm than BEE itself, hence photobleaching is observed upon irradiation of a solution

of BEE at 365 nm. Photobleaching is particularly pronounced in solutions of acyl and

bisacyl phosphine oxides since photolysis results in cleavage of the carbon phosphorus

23

bond, thereby destroying the chromophore which absorbs in the in the near UV-Vis

region (350-410 nm)4. Substituted Titanocenes are another class of photoinitiators which

have excellent bleaching properties in the 440 nm region5 .

Several investigators have emphasized the importance of photobleaching for

curing of thick pigmented coatings and thick polymer and composite parts6‘7. However,
little attention has been paid to quantitatively describing this effect. A quantitative
description of the photobleaching effect would be particularly useful in comparing the
photobleaching effectiveness of different initiators, and thus would provide an important
tool for choosing a photoinitiator for applications such as curing of thick polymer and
composite parts.

In this contribution, a mathematical model was developed for the photobleaching
effect of photoinitiators dissolved in non-reactive solvents. Given the incident light
intensity and wavelength and the initial photoinitiator concentration, the model predicts
the variation of the photoinitiator concentration and light intensity with time and sample
depth. The model equations (a set of coupled partial differential equations) were solved
using software packages (Mathcad and Excel) and the simulation predictions were

compared to experimental results.

4.2. Description of the System Being Modeled
A mathematical model was developed to describe the spatial and temporal
evolution of the light intensity gradient in thick photobleaching systems. The system

being modeled consists of a dilute solution of a photoinitiator (of known depth) which is

24

illuminated with monochromatic light from above. Both the photoinitiator and the solvent
may absorb at the wavelength under consideration, therefore the light intensity changes
with sample depth. However since the photoinitiator undergoes photobleaching, the
intensity gradient in the sample changes with depletion of photoinitiator. In fact at any
point below the sample surface, the intensity increases with time. Thus the light intensity
and photoinitiator concentration in the sample are functions of both time and sample
depth. The model predicts the variation of light intensity and concentration with sample
depth and time, given the incident light intensity, wavelength and the initial photoinitiator
concentration.
4.3. Materials and Method

Materials

The photoinitiators chosen for this study were benzoin ethyl ether (BEE, Aldrich),
2,4,6-trimethyldiphenylbenzoylphosphine oxide (TPO, Aldrich), 2,4,6-
bis(trimethylbenzoyl) phenylphosphine oxide (BAPO, Ciba-Geigy), and bis[n5—2,4—
Cyclopentadien-l-yl] bis[2,6—diﬂuoro-3-[lH-pyrrol-l-yl]phenyl titanium (a substituted
titanocene, Ciba-Giegy). The solvents used were acetone (J .T. Baker), 2—propanol (J .T.
Baker) and methyl ethyl ketone (Mallinckrodt). The substituted titanocene received from
Ciba-Giegy actually contains a blend of the pure photoinitiator and an inert clay. The
mixture was dissolved in methyl ethyl ketone and centrifuged at 2000 rpm for about 10

minutes, in order to precipitate out the inert clay. The other chemicals were used as

received.

25

Experimental Procedure

The photoinitiator solution was exposed to light from a 1000 W mercury-xenon
lamp (Oriel Corp.). A water ﬁlter was used to remove the IR radiation emitted by the
lamp, and. interference ﬁlters were used to select the desired illumination wavelength.
The light intensity at the sample surface was measured using a Scientech 362 energy
meter. A photoinitiator solution of known depth was placed in a cuvette and illuminated
from above with light of the required wavelength for a predetermined period of time. The
sample was then removed from the light and was mixed well before the absorbance
spectrum was collected using a Hewlett Packard 8452A diode array UV-Vis
spectrophotometer. This procedure was repeated for various illumination times until no
further change in the sample absorbance was observed. The spectra collected at various
times were analyzed and used to generate the concentration versus time plots. The

experimental procedure adopted for all the initiators was the same.

4.4. Development of the Model
The spatial and temporal evolution of the light intensity gradient as well as the
concentration gradients of the initiator and the photolysis products are described by the
following set of coupled differential equations (see Appendix, section 1 and 2 for
derivation). Here, I(z,t) represents the light intensity at time ‘t’ and at a depth ‘z’ below
the sample surface. C,(z,t) represents the photoinitiator concentration and CA(z,t), Cg(z,t)

etc. represent the concentrations of the photolysis products ‘A’, ‘B’ etc.

26

 

 

 

d](z.,t)
dz =—[S+k,C,.(z,t)+k,C,,(z.t)+k8CB(z.t)+-.-.lllz.t)
dC,(z,t)_ k, I t C +D 072C,(z,t)
dt -_ NAle (1” (Z’) ’(Z’I) ’ dzz
ac,(z,t)_ k, 1 0C 0+0 32C,(z.t)
8t — NAhv I)" (Z’ ,(2, A dz2
ac,,(z,t) k, 32C,,(z,t)
8t —[NAhV]¢BI(z,t)C,(z,t)+DB dzz
(BIC.
Initial Conditions: C. (2,0) = C,“
CA(z,O)=0
CB(z,0)=0
etc.

Boundary Condition: I(0,t) = I0

The above equations are a set of coupled second order partial differential equations which

(4.1)

(4.2)

(4.3)

(4.4)

are difﬁcult to solve. However, if the effect of diffusion is neglected (as justiﬁed in

section 4.6), then they can be simpliﬁed to the following form.

ﬂ:_a Curl—ﬂCiI
dz
8C
—= "I, III/Ci
l
I(t,0)=l

Ci (0’ Z) : Cm

(4.5)

(4.6)

27

 

 

Where,

Cz:[£1+(t,¢,+k,¢,h J] (47)

in ¢i
ﬂ:[kr¢i—(k/1¢A +ka¢n+m)] (4.8)
(D.
and
_ ki¢i
y_[N,hv) (4.9)

This set of coupled ﬁrst order partial differential equations can be further reduced to an
ordinary differential equation shown below

dc _
dt— = —0Ci0c + ﬂCl-O (e C — l) where c = ln(Ci0/C,) (4.10)

@zchzyhu

The constants 0t and B can be determined from a single photobleaching experiment as

follows (see appendix, section 3 for derivation)

1
a=——ln T - 4.11
Cm, (-) ( >
1 T
=——ln 4° 4.12
ﬂ Curl [70] ( )

where T0 is the initial transmittance and To, is the ﬁnal transmittance (after complete

bleaching) of a thickness ‘1 ’of the sample at the wavelength being considered.

28

Equation (4.10) can be solved numerically to obtain c as a function of 2 at
different times. From this information, it is possible to obtain the initiator concentration
in the sample as well as the light intensity in the sample as a function of time and sample
depth.

The experimental data obtained is a plot of the average photoinitiator
concentration in the sample as a function of time. In order to obtain this information
from the model, the concentration curve at that time must be integrated through the entire

sample depth. This is easily accomplished using numerical integration techniques

4.5. Results and Discussion

Simulation Results

Figures 4.2 and 4.3 show the simulation results for evolution of light intensity
gradient in the photoinitiator solution with time. It can be seen that the severity of the
intensity gradient gradually reduces with time until it becomes constant (no longer
changing with time). The ultimate intensity gradient reached depends upon the
absorbance of the photolysis products. If the products are non absorbing at the
wavelength being considered then the gradient is completely eliminated (Figure 4.2).
However, if the products of photolysis do absorb at the wavelength of interest, then the
gradient is partially eliminated as shown in Figure 4.3. Figures 4.4, 4.5 and 4.6 show the
change the transmittance of the photoinitiator solution with time for different sample
thickness. It is clear that at larger sample depths, the curve has a characteristic sigmoidal

shape, indicating an initial lag or induction period before the intensity begins to increase

29

rapidly. Figure 4.7 shows the effect of solvent absorbance on the gradient elimination
rate. It is clear that even a small extinction coefﬁcient of the solvent at the wavelength of
interest has a considerable effect on the gradient elimination rate because the solvent

constitutes more than 99% of the photoinitiator solution.

Experimental Verification of Model

The method adopted for experimental veriﬁcation of the model is described
below. The photoinitiator solution was irradiated from above for a predetermined period
of time. The solution was then well stirred and its absorbance spectrum is taken. From
the absorbance spectrum, the photoinitiator concentration in the well stirred solution was
determined based upon the extinction coefﬁcient of the initiator and the photolysis
products (using equation 23 from the appendix). This is the experimental data. The
theoretical concentration in the well stirred solution was calculated by integrating the
concentration gradient predicted by the model with respect to 2 then dividing by the total
depth. Thus the model and experiment could be compared.

Figures 4.8 through 4.14 show the comparison of theoretical and experimental
results for the monochromatic photobleaching experiments for four photoinitiators at
different concentrations, light intensities and thickness. It is seen that the model
compares well with experiment in every case. Small deviations could arise be because the
irradiation was not perfectly monochromatic. It was observed that the value of the

constant 7 was different from the value which was theoretically predicted by equation 4.9.,

30

This was true for all the photoinitiators studied. Table 4.1 shows the theoretical and

actual values of y for the different photoinitiators which were investigated. It is seen that
the actual value is about 1.5 to 2.5 times the theoretically predicted value in all cases.
However, for a given photoinitiator and light wavelength, the values of (X, [3 and y are

constant and completely characterize the photobleaching of the initiator at that

wavelength.

4.6. Effect of Diffusion

Model for Photobleaching Including the Effect of Diffusion

When the solution of photoinitiator is irradiated from above, the light intensity
gradient causes different depths of the sample to undergo bleaching at different rates,
thereby resulting in a concentration gradient though the sample depth. The concentration
gradient results in diffusion of the photoinitiator, which in turn affects the rate of
photobleaching. The following is a brief analysis of the effect of diffusion in the
photobleaching process.

Consider a solution of photoinitiator whose photolysis products do not absorb at
the wavelength being considered. Then, the equations describing the photobleaching

effect including the effect of diffusion are:

071(z.t)

 

 

=—k,.C,- (LO/(21.1) (4'13)
dz

dC,(z,t):_[ k.¢,- )C,(z,t)+D.§ﬂf’—t) (4.14)
dt NAhV 3Z-

31

the above equations can be transformed into the following (see appendix, section 5)

 

d u 012“

—-D. ,=a l—e" 4.15

dt ' dz" ( ) ( )
I

where, u = ln[—] (4.16)
0

and, a 2 (W2) (4.17)

NAhV

The initial and boundary conditions are

@zzQu=0
@tzQu2-kCz

I I“

(4. 18)

Equation (4. 15) was solved numerically using the ﬁnite difference method (see appendix,

section 5) and the results were compared to the solution without including the effect of
diffusion (Di=0). In the simulations discussed in this section, a representative value8 of

diffusion coefﬁcient used was 10’4 cmz/sec. If anything, this value of diffusion
coefﬁcient is high for polymerizing systems. Figure 4.15 illustrates the effect of diffusion
on the photobleaching of a dilute solution of benzoin ethyl ether subjected
monochromatic light of moderate intensity. It is clear that at small time scales (about 2
minutes), diffusion has practically no effect. However at larger time scales, there is a
small but observable effect of diffusion on the bleaching process. Figure 4.16 shows the

2

effect of increasing the light intensity. Increasing the light intensity from 2.8 mW/cm to

10 mW/cm2 results in a considerably faster bleaching and effect of diffusion is not
noticeable during the entire bleaching process. Figure 4.17 illustrates the effect of

increasing the photoinitiator concentration. The light intensity used is moderate at 2.8

32

mW/cm2 (the same as Figure 4.15). Again, at small time scales, the effect of diffusion is
not noticeable however, at larger time scales, the effect of diffusion is very clear and there
is a considerable error involved if the diffusion term is neglected. In general, it can be
concluded that at large time scales, the effect of diffusion is more pronounced and factors
such as low light intensity and high photoinitiator concentration which tend to increase
the time scale of photobleaching cause the effect of diffusion to become more important
in the analysis.

Figures 4.18 and 4.19 show the effect of diffusion on the photobleaching of BEE
under the same conditions (initiator concentration and light intensity) used in the
experiment. It is clear that effect of diffusion in this case is quite negligible and hence,
neglecting the diffusion term in the photobleaching model is justiﬁed. The same
conclusion can be drawn for the other photoinitiators based on the time scale of the

experiment and the concentration of photoinitiator used.

Effect of Diffusion Coefficient

Figure 4.20 shows the effect of changing the value of the diffusion coefficient on
the photobleaching process. There are two limiting cases: (1) no diffusion in the sample
(diffusion coefﬁcient = 0) and (2) well mixed sample (effectively inﬁnite diffusion).
Analytical solution to the photobleaching equations can be obtained for both these cases
(see appendix, sections 6 and 7). The light intensity gradient through the sample for cases
involving a ﬁnite diffusion coefficient lie in between the two limiting cases as seen in

5 2

Figure 4.20. As the diffusion coefficient is increased from 10" cm2/s to 10' cm2/s, the

33

light intensity gradient through the sample shifts from one approaching the no diffusion

limiting case towards the well mixed sample limiting case (note that curve for 10'2 cmz/s

case almost coincides with the curve for the well mixed case).

4.7. Conclusions

A mathematical model was developed to characterize the photobleaching of
photoinitiators. The model contains three constants which can easily be determined
experimentally and completely characterize the photobleaching for a given initiator and
wavelength of light. The model was veriﬁed experimentally and good agreement between
theory and experiment was found. The effect of diffusion on the photobleaching was
shown and it was demonstrated that diffusion does not have a signiﬁcant effect under the
conditions studied. The model can be used to compare the photobleaching effectiveness
of different initiators and thereby in making a choice of initiator for curing systems where
the depth of cure is a major concern. For example, Figure 4.21 compares the
photobleaching rate of two initiators namely BEE and BAPO. Obviously, BAPO
bleaches faster than BEE, thereby making it a better choice as a photoinitiator for thick

systems.

34

 

 

   

Intensity = 2mW/cm2 @ 360nm
Thickness = 2.5 cm

Concn. = 5.1 x 10“ g/ml in Acetone

_‘ I
1.4 + , \

1.23. I

 

 

Absorbance

 

 

 

320 340 360 380 400 420 440
Wavelength

Figure 4.1. Time evolution of the absorbance spectrum of a bis—acyl phosphine oxide
(BAPO) photoinitiator as it is irradiated with low intensity UV light. The

2.5 cm thick sample containing 5.1 x 10’4 g ml'1 of the initiator dissolved in
acetone was illuminated from above with ﬁltered light from a mercury lamp.

The light was passed through a 10 nm bandpass ﬁlter centered at 360 nm,

and the intensity at the top surface of the sample was 2 mW cm'z.

0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2

Normalized Intensity

0.1

Figure 4.2.

0.9
0.8
0.7
0.6
0.5
0.4
0.3

Normalized Intensity

0.2
0.1

Figure 4.3.

 

 

1‘11 1

 

iliiiiliviili

    

111'
I

 

—o—0 sec

i;1_ldi'
I

+30 sec
A 90 sec
~ ~~x~~180 sec
+600 sec

llllilLl‘

 

 

 

 

l l J 1 I
1 ﬂ 1 r v r r r I Y t r T 1 Y r r v r r r l

0 0.2 0.4 0.6 0.8 1
Depth (cm)

Elimination of the intensity gradient in a 1cm thick solution of BEE. The
concentration of BEE is 5 x10'4 g/ml and the incident light intensity is 5

mW/cm2 at 330 nm

 

 

I

il-iiilii ILL-ll

'llt

 

 

-.\\
+0 sec I T .....

---<»~ 3 0 s e c “\‘w

~ ' A 90 sec ““’
x 180 sec
I—+—600 sec

illitililiiliiiilit

 

 

 

 

 

1 I 1
“r— 1 T Y r r r 1 T 1 1 T "r 1 1 Y

0 0.2 0.4 0.6 0.8 1
Depth (cm)

 

——<»—

Elimination of the light intensity gradient in a 1cm thick solution of TPO.

The concentration of TPO is 5 x 10'4 g/ml and the incident light intensity is
5 mW/cm2 at 360nm

Normalized Intensity

0.9

 

0.8

.1111141

I

0.7

0.6

ailiiill

0.5

0.4

0.3

0.2

1111111111111

0.1

l l
l

 

.
O I ‘

l

 

 

g

 

 

0 100

l

200

T 7 #7 T T

300
Time (sec)

400

1
I

500

600

Figure 4.4. Change in transmittance as a function of time for a 1cm thick solution of

Normalized Intensity

Figure 4.5. Change in transmittance as a function of time for a 3cm thick solution of TPO

TPO irradiated with 360nm light of intensity 5 rriW/cm2

0.6

 

0.4
0.3
0.2

0.1

 

1

l

 

 

irradiated with 360nm light of intensity 5 mW/cm

0 100

I
I

200

300
Time (sec)

1

400

2

—0—

500

600

0.45

0.4 4’

0.35

O
O)

0.25

O
'0

Normalized Intensity
P
a:

o
L.

0.05 1‘

Figure 4.6.

0.7 ,,

0.6 g

0.4

0.2

Normalized Intensity

0.1

Figure 4.7.

0.5 '

0.3 -

37

 

 

 

 

 

0 200 400 600 800 1000 1200
Tim e (sec)
Change in transmittance as a function of time for a 5cm thick solution of

TPO irradiated with 360 nm light of intensity 5 mW/cm2

 

 

 

T —o—Without Solvent Absorbance '

I ~a—-With Solvent Absorbance

 

 

 

 

 

0 200 400 600 800 1000 1200
Time (sec)

Effect of solvent absorbance on the evolution of the transmittance of a 3cm
thick solution of TPO in a solvent with an extinction coefﬁcient of 0.5

ml/gm.cm and 0.8 gm/ml density. The concentration of TPO is 5 x 10‘4
g/ml and intensity of the incident light is 5 mW/cm2

38

Table 4.1. Comparison of theoretical and best ﬁt values of ”y for different photoinitiators

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Initiator 7 (Theoretical) 7 (Best Fit)
BAPO 2.4x10"4 2.9x10'4
TPO 20x104 5.4x10"4
BEE 0.96 x 10'4 1.7 x 10‘4
Titanocene 4.6 x 10-4 10 x 10-4
(L9 4 \i -
3: \ F. Experimental Data
0-8 j ~v, . 1 Model Prediction
0-7 i \ Thickness = 2.5 cm
0.6 1 ’~ Concn. = 7.6 x 10" g/ml
0' Intensity = 2.2 lecm2 @ 360 nm '

 

 

 

Dimensionless Avg. Concn. (Ci/0.0)
0
0|

 

 

 

0.1 1*» ""~,. _
oils
0 200 400 600 800 1000

Time (sec)

Figure 4.8. Comparison of theory and experiment for photobleaching of TPO

39

 

0.9
0.8

lllllill'lJ

0.6
0.5
0.4
0.3

Dimensionless Avg. Concn. (Cl/0,0)

 

‘1_L

 

0 Experimental Data
. L Model Prediction

 

 

 

Thickness = 3cm
Concentration = 5x10'4 glml
Intensity = 2.2 mW/cm2 @ 360

 

 

 

O

 

l l 1 l
I r r r 1 r r r

 

3'!
a:
i:
q
a:
a.
so

Dimensionless Avg. Concn. (C/Cio)

1 r r I r r

200 400 600 800 1000
Time (sec)

. Comparison of theory and experiment for the photobleaching of TPO

 

0.9
0.8 E
0.7 -
0.6
0.5 f»
0.4
0.3
0.2
0.1 i

 

 

I 0 Experimental Data

I Model Prediction

 

 

 

O. Thickness = 2.5cm
I Concentration = 4.5x10" glml
Intensity = 1.4 lecm2 @ 360 nm

 

 

 

 

l r l 1 O
r r r T I T I

 

T I i I I Y T T I

200 400 600 800 1000
Time (sec)

Figure 4.10. Comparison of the01y and experiment for the photobleaching of TPO

40

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

  

1
’3 ;\l
0 0.9 it- _
3' 3 \ - Experimental Data—I
v o 8 —~ \ .
=- ' g —Theoretical Plot I
g 0.7 f
0. 0.6 i \ Thickness = 2.5 cm
gt 0.5 jg " Concentration = 8.4 x 10" glml
< 3 Intensity = 2.2 mW/cm2 @ 360 nm
3 0.4 T”
at i: .1
E 0.3 EI \
2 0.2 3+—
0 r
g 0.1 :~
0 a ' ' .
O r r r i J, r r . i r' r i . = r
0 1000 2000 3000 4000
Time (sec)
Figure 4.11. Comparison of theory and experiment for the photobleaching of BAPO
1 .
’72 i E—
g 0'9 f ; - Experimental Data I
-f 0-8 :5 I -~Theoretical Plot I
5 0.7 it
5 : I Thickness = 2.5 cm
0 0.6 : I Concentration =5.1 x 10-4 glml
m g_ I Intensity = 2.2 lecm2 @ 360 nm
5 0.5 _ I
3’, 0.4 g,
2 i
.5 °-3 i
2 0.2 3-
o -
.E 0.1 {r \
CI : ‘N _
o i...“‘:~*i~—.rv i.

 

 

 

0 500 1000 1500 2000
Time (sec)

2500

Figure 4.12. Comparison of theory and experiment for the photobleaching of BAPO

4|

 

 

 

 

 

 

 

 

 

 

 

<25 1 ;

3.. 0-9 3” I - Experimental Plot

‘7 0,3 ;_ —Theoretical Plot

C 1

o _

g 0-7 jl Thickness = 2.5cm

0 0,5 g» Concentration = 2.79 x 10" glml

m .. Intensity = 0.8 mW/cm2 @ 330 nm

> 0.5 g '-

< r

3 0.4

2 :

r: 0.3 x

.2 g -

2 0-2 \

o \\

“E- 0.1 -h \\‘:~\

a Z l 1 TR.-._‘__‘
0 . i Y T ~ I—r T T I r Y T '

0 1000 2000 3000

Time (sec)

4000

Figure 4.13. Comparison of theory and experiment for the photobleaching of BEE

 

 

 

 

 

 

 

 

 

 

 

1
’3 .
S- 0.9 is
v 03 g; - Expermental Data
5 1 "Model Prediction
5 0.7 :
o 0.5 3 Thickness = 2.5cm
g3 -f~ Concentration = 2.4 x 10-4 glml
< 0'5 E Intensity = 2.2 mWIcm2 @ 410nm
‘0
m 0.4 i
2 4; "
E 0.3
'2 0.2 - \
2 ' \
0.1 ; \‘\
'5 - \g
0 T Y i I r r r i T . l r i r -~~-1l...._.,__,_,‘ 1 i r r
0 100 200 300 400 500 600
Time (sec)

Figure 4.14. Comparison of theory and experiment for photobleaching of substituted

titanocene

42

 

~‘-—--

0.9 _
0.8 H
0.7
0.6 a

      

0.5 ~-— \ i - ..

100 sec

0.4 - \
0 sec -

0.3 + _ g
0.2 I ‘

r

 

Normalized Intensity (l/Io)

-—-With Diffusion . ' --
0.1 4*" . Without DiffusionIi N‘T‘k
o -— I I I I I
0 0.2 0.4 0.6 0.8 1
Depth (cm)

 

 

 

 

 

Figure 4.15. Effect of diffusion on the photobleaching rate of BEE. The concentration is
10'3 glml and the light intensity is 2.8 mW/cmz @ 330nm

 

 

0.8 +
0.7 ~»

100 sec
0.6 "A" '2“.

0.4 0 sec

I

0.3 -~

Normalized Intensity (Illo)
O
OI

 

0.2 -- ' u

I —With Diffusion I

- WithoutDiffusion I .

o i i . f I

0 0.2 0.4 0.6 0.8 1
Depth (cm)

 

 

 

Figure 4.16. Effect of diffusion on the photobleaching rate of BEE. The concentration is
10'3 g/ml and the light intensity is 10 mW/cm2 @ 330nm

43

 

0.9

0.8

  
    

 

0.7 —— With Diffusion

.- Without Diffusion

 

 

0.6 —~

 

0.5 ~~

0.4 --

0.3 +—

Normalized Intensity (lllo)

0.2 or

0.1 -_

 

 

 

0 0.2 0.4 0.6 0.8 1
Depth (cm)

Figure 4.17. Effect of diffusion on the photobleaching rate of BEE. The concentration is
3.3 x 10'3 glml and the light intensity is 10 mW/cm2 @ 330nm

 

0.98 ~

0.96 -
0.94 __ It = 3600 secI

 

 

 

0.92 —~
+With Diffusion

0.9 +
0.88 ~~

0.86 ~—

 

 

.--x—-WIthout Diffusion

 

Normalized Intensity (Illo)

0.84 +
0.82 —~

0.8 t % t
0 0.5 1 1.5 2 2.5
Depth (cm)

 

 

 

.0—

Figure 4.18. Effect of diffusion on the photobleaching rate of BEE. The concentration is
2.79 x 10’4 glml and the light intensity is 0.8 mW/cm2 @ 330nm

44

 

 

 

 

 

 

 

 

 

  
   
  
  
  
   

 

 

 

 

 

 

 

 

5 100 V. P
a V, 21./7,27”
E 80 + ‘;/,I,~/'/ '-
q 70
N ._
o ,//
ea 60 +
a 50
~I-
E
2 4o .- ——With Diffusion
0
‘5 30 _- -~~ ~Without Diffusion
'3 I - Experimental Data
g 20 ~~
G u ‘
E 10 I»
2
0 - I I I I I I I .
0 500 1000 1500 2000 2500 3000 3500 4000
Time (sec)
Figure 4.19. Comparison of experimental and model data (with and without diffusion)
for the photobleaching of BEE.
1 i
0.9 i I
’5‘ 0.8
a : -o-Dl = 0 cm 2ls
0.7 ——
g 3 +0: = 0.00001 cm2ls
g 0-6 3” .. Dl = 0.0001 cm2/s
_, _
E 0.5 {P ~x—Di = 0.001 cm2ls
E 0.4 3 ~—-Dl = 0.01 cm2 Is
2 0.3 +well mixed
3 :
z 0.2 9*
0.1 :1
0 r , I I I I I I I I I I I I
0 0.2 0.4 0.6 0.8 1
Depth (cm)
Figure 4.20. Effect of initiator diffusion coefﬁcient on the photobleaching process. The

light intensity is 10 mW/cm2 and the concentration is 3.3 x 10'3 glml. All
the intensity gradients shown are at time = 900 seconds

45

 

0.8 -—
0.7

l

0.6 —~

BEE

0'4 BAPO

0.3 ~—
0.2 —~ \~

0.1 w

Dimensionless Concentration (CI/Ch)

 

 

 

o 1 l l
r 1 l r l T T l . l 1 T ‘r l f

0 200 400 600 800
Time (sec)

Figure 4.21. Comparison of the photobleaching rate of BAPO and BEE. Both the
solutions had the same initial absorbance of l and solution depth of 2 cm.

The light intensity used for bleaching was 2.2 mW/cm2

46

4.8. List of References
. J .D. Coyle, Introduction to Organic Photochemistry, John Wiley & Sons, 1986

. J .P. Fouassier, Photoinitiation Photopolymerization and Photocuring F undamentals,
and Applications, Hanser Publishers, 1995

. F.D. Lewis, R.T. Lauterbach, H.G. Heine, W. Hartmann, H. Rudolph, J. Am. Chem.
Soc., 97, 1519, 1975

. K. Dietliker, G. Hug, R. Kaeser, M. Kohler, U. Kolczak, D. Leppard, L. Misev, G.
Rist, W. Rutsch, Novel High Performance Bisacylphosphine Oxide (BAPO)
Photoinitiators, Ciba-Geigy Corporation, 1994

. J. Finter, M. Riediker, O. Rohde, B. Rotzinger, Makromol. Chem, Macromol. Symp.,
24, 177 1989

. K. Dietliker, In Radiation Curing in Polymer Science and Technology, Volume II,
J .P. Fouassier and J .F. Rabek, Eds., Elsevier Applied Science, pp. 155-237,1993

. L.S. Coons, B. Rangarajan, D. Godshall, A.B. Scranton, ACS Symposium Series, in
press

. R.E. Emmert, R.L. Pigford, Chemical Engineers’ Handbook, 4th Ed., R.H. Perry, C.H.
Chilton and S.Y. Kirkpatrick, Eds., McGraw-Hill, pp. 14.22 - 14.23

To
T...
2

47

4.9. List of Symbols

Concentration of photolysis product ‘A’ in solution

Concentration of photolysis product ‘B’ in solution

Concentration of initiator

Initial concentration of initiator solution

Diffusion coefﬁcient of initiator in solvent

Ln(Cio/Ci)

Plancks Constant

Intensity of light at the wavelength being considered

Intensity of light at sample surface

Extinction coefﬁcient of species ‘A’ in solution multiplied by 2.303
Extinction coefﬁcient of species ‘B’ in solution multiplied by 2.303
Extinction coefﬁcient of initiator in solution multiplied by 2.303
Avogadro number

Time variable

Absorbance of the solvent divided by the path length
Transmittance of sample at time 0
Transmittance of sample at time 00

Depth (distance) variable

0L,B,y Constants

‘1’».
4’3
in

V

Quantum yield of species ‘A’
Quantum yield of species ‘B’
Quantum yield of photolysis of initiator
Frequency of light being used

48

4.10. Appendix
I . Assumptions made in the derivation
1. The photoinitiator solution is dilute, hence Beer-Lambert’s law is valid
2. The incident light is monochromatic
3. There are no secondary, thermally driven chemical reactions in which the

photoinitiator participates

N

. Derivation of the Model Equations
The initial concentration of the photointitator solution is CIO and the intensity of
light at the surface of the solution at that wavelength is IW/m2 . Consider a thin layer of

solution of thickness 52 at a depth 2 from the surface. The number of photons striking
this surface in a small time St is given by

i = — (1)
hv

The number of photons absorbed by any species ‘8’ present in the solution is given by

(NACSadzHcSi (2)

The total number of photons absorbed is the sum of the photons absorbed by each of the
species present in the solution which includes the solvent itself. Hence, change in the

number of photons passing through the thickness dz of the solution is given by

49

 

5i = (NAatizixs + kiCl. + kACA + kBCB-i- ..... ) (3)
So, from equations (1), (2) and (3), we have
udlﬁ al5t
=NAa5z— S+kiCi+kACA+kBCB+ ..... ] (4)
hv hv
hence taking limits in equation (4), we get
a"!
_:_[S+kiCi +kACA +kBCB+ ..... ]I (5)
c);

In the small time ﬁt, the number of molecules of product A that will be formed is given by

6A = ¢A[NACia§:.]iki (6)

Due to difference in light intensities at different depths, the concentration of A will also
be different. This difference in concentration will cause diffusion of A which must also

be accounted for in the model. A molar balance for species A in the differential volume

adz give

6A = ¢A[NACia&]ikl. + NAJAaﬁtz — NAJAaéi Z+Az (7)

50

Equation (7) can be simpliﬁed by using Fick’s Law for diffusion and taking the limits to

give

 

 

A
—= ¢Alci +0}, 2 (8)

Similar equations can be written for the species B, C etc., that are formed during the

process as shown

 

 

 

 

2
dCB k B d CB
— = ¢ 8 [Ci + D B 2 (9)
c)! N Ahv 3:.
and so on.
For the initiator, the equation becomes
ac, k i (p [C D 32c,
= — —— - - + . 10
()t N A h V 1 l t a: 2 ( )

Equations (5), (8), (9), (10) together give the final form of the equations describing the

effect.

5i

3. Determination of constants 0t and B

 

 

We have
8 ln 1
:-aCiu-ﬂCi (ll)
8 2.
this equation can be rewritten as
o" In T
= ‘aCm-ﬂct (12)
(9 z
where T = [/10 .
Hence,
T = exp|:-—a sz — ,3 I Cidz] (l3)
0
using the condition that at t=0, C,=CIO
7‘0 = exp[_a Cluz — ﬂ Clﬂz] (14)

using the second condition that at t-)°°, CI=O we get

Tm =exp[—a sz] (15)

Equations (14) and (15) can be solved to obtain 0t and B as shown in equations (1 1) and

(12) in the paper.

52

4. Determination of Initiator Concentration from Absorbance Measurement

After the sample has been irradiated for a ﬁxed period of time and well stirred, its
absorbance is measured at the wavelength of interest in order to determine the
concentration of the remaining (undissociated) photoinitiator. The absorbance of the
sample measured experimentally is actually the sum of the absorbance due to the
photoinitiator and the products of photolysis. The following analysis separates out the
absorbance due to the photolysis products and leads to the equation that is used to
determine the photoinitiator concentration from the absorbance data.

Let Cpf be the ﬁnal concentration (lumped) of all the photolysis products, 8f be the

lumped extinction coefficient of the photolysis products. A... is the ﬁnal absorbance (at
t-)°°) of the photoinitiator solution and ‘b’ is the path length used in the absorbance
measurements. Then

A, = 8f Cmb (16)

The final concentration of the photolysis products is given by

C ‘ =C. (¢AMA+¢BMB+..)
¢IM.

 

(17)

Where MA, MB, M, are the molecular weights of species A, B and I respectively. Hence,
from equations (16) and (17),

Am=e Cb (18)

I) ll!

where,

 

-e ((DAMA +¢BMB+”)

8)
' " cm.

(19)

53

If ‘A’ is the average absorbance of the solution at the wavelength being considered after
it has been irradiated for some time, ‘t’ then we can write

A = e]. Cpl) + epCib (20)
where Cp is the concentration of the photolysis products in the solution.

)(¢,M, +¢BM,+..)

 

CI, = (Cm - C, <1),- M.— (21)
from equations (19), (20) and (21), we get
A = a}, (CM — C, )b + e,C,.b (22)
so the ﬁnal equation for C, (phOtoinitiator concentration) is
g = A — epC,,,b (23)

(8" - 8,, )b

the constant SP is found from equation (18) using the experimentally determined value of

A...

5. Solution of the Simultaneous Diffusion with Photobleaching Equations
The equations describing the photobleaching effect including the effect of

diffusion are (assuming no absorption of the photolysis products at the wavelength of

 

 

interest:
81
—=—k.C.I 24
0.,z , , ( )
ac, _ k,¢,. C +0 32C, 25
8t — NAhv ’ "0"2'.2 ( )

54

Equation (1) can also be written as

_lo71nl

 

 

 

 

C. = — 26
' k, d z ( )
Substituting this in Equation (2), we get the following:
dzlnl _ k0,. 1 am] +0 a-‘lnl (27)
dzdt— NAhv dz "dz3

Integrating this with respect to z and using the fact that 1:10 at 2:0, equation (4) can be

 

 

 

reduced to:
o’lnl 32 in] k.¢.
— D. = ——'———’— l -I 28
dt ' dzz IN,hv]<0 ) ( )
which can be finally transformed into
8 u alga
——D. , =a 1—e" 29
at ' 072'.” ( ) ( )
l
where, u = ln[—J (30)
(l
and, a {Milt—I (31>
NAhv

with the following initial and boundary conditions

55

@zzQuzO
@t=Qu=—kCz

l m

(32)

using the usual Taylor’s series approximations for the derivatives, the ﬁnite difference

solution scheme for Equation (6) can be written as

D, (At

(AzY

 

um,” : (Al)a(l_e“).t )+

(“H-k _2uj.k +uj+l.k) (33)

4

for all the simulations involving Di = 10' cm2/sec, the following step sizes were used

0 A
At = 4sec, Az = 0.05 cm , so that f - 0.16 which is the optimum value of this ratio

2 ..

and gives superior accuracy.

6. Analytical solution for the ‘no diﬂusion’ case

Putting Di = O in equations ( l) and (2) above, we get

 

great (34)
Z

59—3— k"¢" C (35
dt - NAhv ’ )

then, following the same steps as shown previously, equation (6) becomes

56

izaﬂ—e") (36)

This equation can be solved analytically and the solution is:

 

_= . (37)
—€

7. Analytical Solution for the ‘well mixed’ case

Assume that the depth of the photoinitiator solution is L cm and it is always well mixed
in the sense that the concentration of the solution does not vary with depth at any given
time.

The photobleaching equations then are:

 

1 = [Ge—W (38)
£=(1_]0) $1 (39)
dt NAth

where ‘I’ is the intensity at the depth ‘L’ of the sample, and 10 is the intensity of the
initiating light at the top of the sample.
Substituting for I from equation (1), we get a ordinary differential equation for C, which

can easily be solved. The solution is given by

C, = i ln(e“""" — W + I) (40)

l

57

so, the equation for the intensity gradient through the sample is

_l_ 2 (euL-hl _e-h! + i)—%‘ (41)
[0

Chapter 5

Investigation of Photoinitiating Systems for
Photopolymerization of Composites

Abstract

Studies were performed to evaluate the suitability of a variety of photoinitiators
for UV curing of thick polymer and composite parts. A preliminary choice of
photoinitiators was made based upon their absorption spectra, the emission spectrum of
the UV light source and the transmittance of the resin. For each photoinitiator, studies
were performed to determine the effect of light intensity, photoinitiator concentration and
sample thickness on the cure time of the resins. Hybrid photoinitiating systems and dual
photo/thermal cure schemes were investigated in an attempt to further reduce the cure
time. Finally, composite samples were prepared both by photocuring and by thermal

curing and their mechanical properties were evaluated.

5.1. Introduction
The choice of a photoinitiator for curing of thick polymer and composite parts
depends on a number of considerations, including spectrum of the initiating light source,
absorbance of the monomer and other additives, absorbance and scattering properties of
the ﬁller or ﬁber reinforcement and photoinitiator characteristics such as its absorption
spectrum, initiation efﬁciency and its tendency to photobleach. A preliminary choice of
photoinitiator for curing a particular resin can be made based upon the above factors.

However, the final choice of photoinitiator must be based upon actual measurements of

58

59

the time taken to cure the resin. Determination of cure time can be accomplished by
selecting a suitable cure monitoring technique.

Several techniques are available at present for in situ cure monitoring of resins.
These include infrared spectroscopy, photodifferential scanning calorimetry (PDSC),

microwave dielectrometry, viscosity measurements etc. Recent reviews in the area of
cure monitoring techniques for radiation curing have been published by Davies1 and

Rabekz. However, most of these techniques are applicable to thin ﬁlms. Davies1 has
stated that cure monitoring of thick samples is one area in which progress needs to be

made. Hardness measurement is commonly used for cure monitoring in thick dental

3

composites . However, since hardness measurements can be made only during the ﬁnal

stages of cure when the resin is solid, it cannot be used for on-line cure monitoring.

Photo dynamic mechanical analysis was recently used by Renault et al4’5 for in situ
monitoring of cure of neat or ﬁber ﬁlled resins. While this technique shows considerable
promise it also has several limitations. It is based on the bulk tensile modulus of the entire
sample rather than the cure behavior of the sample at a particular depth and requires the
modulus of the resin to be correlated with the extent of cure. For these reasons it is not
very useful for determining cure time of thick polymer samples.

In this study, the polymerization was monitored by measuring the temperature at
the bottom of the sample during photopolymerizations carried out under adiabatic
conditions. This method is based upon the fact that the polymerization process is highly
exothermic, and, under adiabatic conditions, leads to a considerable temperature rise.

Therefore, cure at the bottom of a thick sample can then be monitored by placing a

60

thermocouple on the bottom surface of the container. After the sample is illuminated, the
observed temperature rises due to the heat of polymerization, and ﬁnally plateaus at a
constant value when the polymerization is complete. The time at which the temperature
becomes constant can be regarded as the cure time for that sample. Using this cure
monitoring technique, the effect of several variables such as photoinitiator type,
photoinitiator concentration, light intensity and spectral distribution, sample thickness etc.
were studied and recommendations were made for design of initiating systems for

photopolymerization of composites.

5.2. Experimental

Materials

In these studies, Derakane 510A obtained from Dow Chemicals was used as the
resin. The photoinitiators used included 2,4,6 Trimethyldiphenylbenzoyl phosphine oxide
(TPO, BASF), bis(2,4,6-trimethylbenzoyl) phenylphosphine oxide (BAPO, Ciba-Giegy),
benzoin ethyl ether (BEE, Aldrich) and bis[n5—2,4-Cyclopentadien-1-yl] bis[2,6-difluoro-
3-[lH-Pyrrol-1-yl]phenyl Titanium (a substituted titanocene, Ciba—Giegy). In addition a
two component initiating system consisting of N-methyldiethanolamine (Aldrich) and
camphorquinone (Aldrich) was investigated. Finally, in some systems benzoyl peroxide
(Aldrich) was used as a thermal initiator. The ﬁber reinforcement used in the composites

was an e-glass woven mat. All the chemicals were used as received.

61

Determination of Cure Time

The formulation to be photopolymerized (which consisted of the resin and the
photoinitiating system) was injected into a cylindrical polyethylene vial to the required
thickness (thicknesses ranging from 9 mm to 29 mm were studied). A thermocouple
(Chromium-Aluminium, Omega type K) was introduced through a small hole in the side
of the vial. The thermocouple was positioned at the bottom of the sample, near the wall
of the vial. Care was taken to ensure that the thermocouple was placed at the same
position for every run. The thermocouple was connected to a WS-ABC interface card
(Omega Inc.) which digitized the data and stored it in a PC. The vial was thermally
insulated by a thick layer (several inches) of polyurethane foam in order to ensure that the
polymerization was adiabatic. The sample was irradiated with light from above using a
1000 W Oriel model 66021 medium pressure mercury-xenon arc lamp which was
equipped with a water ﬁlter to eliminate the IR radiation and a dichroic mirror to ﬁlter out
the deep UV radiation (below 280nm) emitted by the lamp. Light intensity was measured
using UVICURE Plus high energy UV integrating radiometer (EIT Inc.). In order to start
the data collection and sample illumination simultaneously, a shutter (Oriel model 71445
electronic safety shutter) was triggered by a digital pulse from the WB-ASC interface
card. The sample was illuminated for a predetermined period of time, and the acquired
temperature-time data was analyzed to determine the cure time for the sample. Figure 5.1
shows the general trend followed by the temperature-time curve during such
photopolymerizations. The time when the curve begins to plateau was taken to be the

cure time for the specimen (200 seconds for the curve in ﬁgure 5.1). Three runs were

3

62

performed for each sample, and the variation in cure time was less than 5%.

Composite Preparation and Testing

The UV cured composite specimens were prepared as follows: woven glass ﬁber
mats were cut to the required dimensions, and were stacked on top of one another before
the UV curable formulation was poured onto the stack. The resin-impregnated ﬁber mats
were placed between two glass plates and pressure was applied to compress the uncured
composite. The system was then exposed to light of 250 mW/cm2 UVA intensity from a
3000 W arcless mercury vapor lamp (Fusion UV Systems, model F450T) for ﬁve
minutes. The cured composite was then cut to the dimensions required for the ﬂexural
test. A similar procedure was followed for the thermally cured composite except that it

was thermally cured in an oven at 75°C for 1 hour and postcured at the same temperature

for 24 hours. The ﬂexural properties of cured composite specimens were determined
using a United STM-20 instrument in accordance with the ASTM D790 method. The
composite specimens prepared for testing were 6 to 7 mm thick, 10 mm wide and 125
mm long. The span length used was 96 mm and the crosshead downdrive rate was 2.6

mm/min, as speciﬁed by the ASTM test procedure.

Absorption and Transmittance Spectra
Transmittance spectra of the resins and absorption spectra of the photoinitiators
used in this study were determined using a Hewlett Packard 8452A diode array UV-Vis

spectrophotometer.

63

Emission Spectrum of Lamp

Emission spectra of the lamp was obtained using a ﬁber optic spectrometer model

SD1000 from Ocean Optics Inc.

5.3. Results and Discussion

Spectral Considerations for the Choice of photoinitiator

A number of commercially available photoinitiators may be used for
photopolymerizations of ﬁlms and coatings. However, for photocuring of thick parts, the
criteria for selection of photoinitiator are more stringent. Foremost, the photoinitiator
obviously must have signiﬁcant absorbance in the region where the lamp emits. In
addition, for thick parts, the photoinitiator must have good absorbance in a spectral region
different from where the other components of the system (resin, additives etc.) absorb.
This is because the resin constitutes more than 99% of the formulation and may
effectively ﬁlter out the light which is required for photoinitiation. Finally, it is best if the
photoinitiator undergoes signiﬁcant bleaching upon illumination This allows light
penetration into the sample upon production of active centers.

Figure 5.2 shows the transmittance spectrum of the Derakane 510A resin in the
300-550 nm range. The ﬁgure illustrates that the resin is opaque to light of wavelength
between 300 and 350 nm and the transmittance increases continuously at higher

wavelengths. Figure 5.3 is a plot of the extinction coefﬁcient of Derakane 510A, which

64

again reveals that the absorbance is very high below 350 nm and decreases rapidly at
higher wavelengths. It is noteworthy that the extinction coefﬁcient of the resin appears to
be low (for example the extinction coefﬁcient of the resin at 365 nm is 1.8 ml/gm cm as
compared almost 2000 ngm cm for BAPO at the same wavelength). However, even
such a small resin extinction coefﬁcient becomes very important because it constitutes
more than 99 % of the photocurable formulation. The extinction coefﬁcient of the resin
being investigated (Derakane 510A) and its dependence on the wavelength in the 300-450
nm region is a very important optical characteristic especially for UV curing thick parts
since it dictates how much of the initiating light will be absorbed by the resin itself (and
hence will not be available for photoinitiation). For example, attempts to photocure the
resin using 330 nm light will fail since the resin is opaque at this wavelength, hence no
photons will be available to the initiator to produce active centers which are required to
initiate the polymerization.

Figure 5.4 compares the transmittance spectra of three different resins: a vinyl
ester (Derakane 510A), a multifunctional acrylate (ditrimethylolpropane tetracrylate,
DTMPTA,) and an epoxy resin (DER 362 from DOW), illustrating that the optical
characteristics of different resins in the 300-500 nm region are signiﬁcantly different.
The DER 362 and the DTMPTA are more transparent in this spectral region and therefore
are easier to photopolymerize using UV light.

Figure 5.5 shows the emission spectrum of a medium pressure mercury lamp
commonly used for UV curing. It has prominent emission lines at 300, 312, 330, 365,

410 and 440 nm. Figure 5.6 shows the emission spectrum of light from the same lamp

65

after it is transmitted through Derakane 510A. It is clear that the resin completely absorbs
the emission lines at 300,312 and 330 nm, however, it transmits the 365, 410 and 440 nm
lines and the transmittance increases at higher wavelengths. These results are it:
concordance with the transmittance spectrum of the resin which was discussed earlier and
serve to reinforce the point that the transmittance spectrum of the resin being cured is
very important for photocuring of thick polymers.

The above results directly affect the choice of photoinitiator or photoinitiating
system which can be used for photopolymerization of Derakane 510A. Obviously, the
photoinitiating system must have signiﬁcant absorbance in the 360-450 nm region in
order to utilize the 365, 410 and 440 nm lines which are emitted by the lamp and
transmitted by the resin. Several classes of commonly used photoinitiators such as
benzoin ethers, dialkoxyacetophenones, hydroxy alkyl ketones, morpholino ketones,
amino ketones etc do not meet this requirement. However, one class of photoinitiators,
namely benzoyl phosphine oxides exhibit considerable absorbance above 360 nm. For
example, Figures 5.7, 5.8 and 5.9 show the absorbance spectra of benzoin ethyl ether,
trimethyldiphenylbenzoyl phosphine oxide (TPO) and bis(2,4,6 trimethylbenzoyl) phenyl
phosphine oxide (BAPO). It is clear that TPO and BAPO have signiﬁcant absorbance at
365 and 410 nm and are therefore good candidates for UV polymerization of Derakane
510A. Another signiﬁcant motivation for selecting the benzoyl phosphine oxides is their

tendency to photobleach efficiently in the UVA spectral region (as shown in Chapter 4).

66

Eﬁect 0f Photoinitiator Concentration on Cure Time

Thermocouple studies performed on neat resin (Derakane 510A) revealed that for
a given sample thickness and light intensity, there exists an optimum concentration of
photoinitiator at which the cure time is a minimum. At initiator concentrations below the
optimum, the cure time decreases with increasing initiator concentration as expected from
simple kinetic considerations. At concentrations higher than the optimum the cure time
actually increases with increasing concentration due to the optical density of the
photoinitiator, which prevents light from penetrating into the sample. Figures 5.10, 5.11
and 5.12 show the effect of photoinitiator concentration on the cure time of neat resin
(Derakane 510A) for TPO, BAPO and BEE respectively. In every case, it is seen that
there exists an optimum initiator concentration at which the cure time is minimum.
However, there is considerable difference in the optimum concentration and optimum
cure times for the different photoinitiators. Table 5.1 summarizes the optimum cure times
and concentrations for these three photoinitiators. Clearly, BEE exhibits a much higher
cure time as compared to the phosphine oxide photoinitiators, a result which is consistent
with the conclusions which were drawn based upon spectral considerations. Table 5.1
also reveals that the BAPO photoinitiator exhibits a lower cure time and a lower optimum
concentration as compared to the TPO. Both these results can be explained based upon
the photochemistry of the two initiators. Figure 5.13 shows that two radicals formed
upon photolysis of a single TPO molecule. In comparison, Figure 5.14 illustrates that
complete photolysis of BAPO results in more than two free radicals. The higher radical

yield translates into high curing efﬁciency and consequently results in lower optimum

67

initiator concentration and lower cure time for the BAPO. A second factor contributing
to this is the higher extinction coefﬁcient of BAPO compared to TPO in the 360-420 nm

region (as shown in Figure 5.15).

Table 5.1. Optimum concentration and cure times for three different
photoinitiators for curing 0.9 cm thick sample of Derakane 510A with a light
intensity of 65 mW/cm2 UVA

 

 

 

 

Photoinitiator Optimum Cure Time Optimum Concentration
(see) (wt.%)
BAPO 160 0.2
TPO 220 0.4
BEE 620 0.3

 

 

 

 

 

Effect of Photoinitiator Concentration on Degree of Conversion

It was observed that the ﬁnal temperature attained during the
photopolymerizations was maximum at the optimum initiator concentration. This is
shown in Figures 5.16 and 5.17 for two different photoinitiators (BAPO and TPO). Since
all the experiments were carried out under adiabatic conditions, a higher temperature can
only mean that more heat was released during the polymerization, implying that more
double bonds were broken. This suggests that not only is the cure time a minimum at the

optimum concentration but also the degree of conversion is maximum. This is in

concordance with the result obtained by Guthrie et al‘3 who concluded that the extent of

polymerization (degree of conversion) of thick ﬁlms (up to 5 mm) of multifunctional

68

acrylate and methacrylate reaches a maximum and then decreases with increasing
concentration of photoinitiator.

These results illustrate that once a choice of photoinitiator has been made based
upon spectral considerations, it is important to carry out further investigations to ascertain
the optimum concentration of the initiator that is suitable for photocuring the sample of
the required thickness. This is because use of the optimum photoinitiator concentration
not only results in minimum cure time but also leads to maximum conversion. From
these results it was concluded that use of 0.2 wt % BAPO was best for UV curing of 9

mm thick samples of Derakane 510A.

Eﬁ‘ect of Intensity on Cure Time

Figure 5.18 illustrates the effect of light intensity on cure time for 9 mm thick
Derakane 510A resin samples cured using the BAPO photoinitiator. Thus it is seen that
increasing the intensity from 10 mW/cm2 to 20 mW/cm2 resulted in a dramatic decrease
in the cure time from 480 to 240 seconds. However, further increase in intensity did not
change the cure time as rapidly. This diminishing effect of increased light intensity is
probably due to saturation of the excited singlet state of the photoinitiator. When the
system is saturated, the rate of production of active centers is governed entirely by the rate

of a—cleavage, rather than by the rate at which the excited states are produced. Thus,

above an intensity of 30 mW/cmz, increasing the light intensity leads to little change in

the rate of production of active centers, and the cure time decreases only slightly.

69

Effect of Sample Thickness on Optimum Concentration

Experiments performed on photocuring of thick polymer parts using BAPO
photoinitiator showed a small decrease in optimum initiator concentration with increase
in sample thickness (Table 5.2). This decrease in optimum initiator concentration enables
more light to penetrate the sample to compensate for the decrease in light intensity that

reaches the bottom of the sample with increased sample thickness.

Table 5.2. Effect of sample thickness on optimum concentration of BAPO

 

 

 

 

Thickness of Cured Resin Optimum Concentration of BAPO
(cm) (wt. %)
0.9 0.20
2.2 0.16
2.9 0.14

 

 

 

 

Hybrid Photoinitiator Systems

As mentioned earlier, one of the considerations for the choice of a photoinitiating
system is its ability to utilize all the light emitted by lamp. The lamp used in this study
emits three lines 365, 410 and 440 nm where the resin transmits. As shown before, the
photoinitiator BAPO absorbs the 365 and 410 nm lines. However, the 440 nm line is not
absorbed (as shown in Figure 5.19) and hence is not used for production of active centers.
Attempts were made to further reduce the cure time by utilizing the 440 nm line by means

of a hybrid photoinitiating system. Essentially, this involves use of a photoinitiating

70

system consisting of BAPO plus a second photoinitiator which absorbs at 440 nm.
Figures 5.20 and 5.21 show the absorption spectra of two photoinitiators (a substituted
titanocene and camphorquinone) which were studied. Both these initiators have good
absorbance at 440 nm. The formulation containing the hybrid photoinitiator system
BAPO+titanocene actually showed a small increase in cure time over the formulation
containing BAPO alone. This is probably because the titanocene has good absorbance at
365 and 410 nm in addition to 440 nm. This results in the light emitted at these
wavelengths also to be absorbed by the titanocene rather than by the BAPO. Since the
titinocene does not have good initiation efﬁciency (due to its bulky structure), it resulted
in an increase rather than in a decrease in cure time. The camphorquinone+amine+BAPO
formulation also did not show any signiﬁcant decrease in cure time. This is probably
because camphorquinone produces active centers by a diffusion controlled reaction with

the amine and this reaction is too slow compared to the rapid (it-cleavage reaction of the

BAPO to have any signiﬁcant effect on the cure time.

Dual Photoﬂ’hermal Cure Systems

Coons et alg, prepared vinyl ester/glass ﬁber composites by photopolymerization
using BEE as the photoinitiator. They showed that addition of a thermal initiator
(benzoyl peroxide) in addition to the photoinitiator resulted in a considerable decrease in
cure time. This was explained based on the fact that the heat released during the
photopolymerization process triggers the thermal initiator thereby resulting in the

production of more active centers and hence faster cure. However, experiments

71

performed with the dual cure systems using BAPO as the photoinitiator and BPO as the

thermal initiator revealed no signiﬁcant reduction in cure time as compared to systems

containing only BAPO. Coons et al8 found that the cure time of 9 mm thick samples
decreased from 900 seconds to about 300 seconds upon addition of 0.2 wt % of benzoyl
peroxide along with 0.1 wt % of benzoin ethyl ether and further increase in the benzoyl
peroxide concentration did not result in further reduction of the cure time. The cure time
of 9 mm thick samples cured using 0.2 wt % BAPO alone was shown to be about 160
seconds, faster than the cure time of dual cured (BEE + BPO) samples. This shows that
the production of active centers by the thermal initiator was not rapid enough to affect the

polymerization rate of samples cured using BAPO alone.

Composite Preparation & Testing Studies

Glass ﬁber reinforced composites, 6 mm thick, containing 50 wt.% e-glass woven
mat as the reinforcement and Derakane 510A as the matrix were prepared and their
mechanical properties were determined. Three types of composites were prepared — (i)
Purely photocured composite which contained 0.2 wt.% BAPO as the initiator (0.2 wt. %
was used since it has earlier been established as the optimum for 9mm thick samples and
the optimum concentration does not change signiﬁcantly with thickness) (ii) Dual
photo/thermal cured composite which contained 0.2 wt.% BAPO and 0.5 wt.% BPO as
the initiating system (iii) Thermally cured composite containing 1.2 wt.% BPO as the
initiator.

Figure 5.22 compares the ﬂexural properties of the composites prepared. It is

72

clearly seen that there is no significant difference in the properties of the composites. It is
important to note that the thermally cured composite was prepared by baking in an oven

for 1 hour at 75°C to achieve complete cure and subsequently postcuring at the same
temperature for 24 hours and it required 1.2 wt.% BPO, whereas the UV cured composite

was cured only for 5 minutes at 250 mW/cm2 UVA intensity using only 0.2 wt.% of the
BAPO photoinitiator. This clearly demonstrates that UV curing leads to considerably
faster cycle times as compared to thermal curing and results in a composite with the same
mechanical properties as the traditional thermally cured composite. Using a thermal
initiator along with the photoinitiator does not have any effect on the mechanical
properties and it was earlier shown that it does not have a signiﬁcant effect on the cure
time either. Hence, there is no advantage in using the dual cure scheme with the BAPO

photoinitiator.

5.4. Conclusions

In this contribution, it has been established that of all the photoinitiators studied,
Bis(2,4,6-trimethylbenzoyl) phenyl phosphine oxide (BAPO) resulted in minimum cure
time for the photopolymerization of thick samples of Derakane 510A using the mercury
vapor lamp as the light source. It was also shown that there exists an optimum
concentration of photoinitiator at which the cure time is a minimum. This optimum
decreased slightly with increase in sample thickness. Studies conducted using hybrid and
dual cure initiating systems with BAPO as one of the initiators revealed that there was no

signiﬁcant advantage of utilizing these schemes. Finally, it was shown that the ﬂexural

73

properties of glass ﬁber reinforced composites prepared by UV curing was comparable to
those prepared by traditional thermal curing. More importantly, the time required for
preparation of the composite by UV curing was considerably lower than the time required
for preparation by thermal curing. This suggests that cycle times for preparation of glass
fiber reinforced composites can be signiﬁcantly reduced by using UV rather than thermal

curing techniques.

74

 

100 _

Temperature (°C)

 

 

 

.1
l 1 I 1
V T T Y 1 y f Y ‘r T 1 r v f 1, v r T T T

0 50 100 150 200
Time (sec)

 

Figure 5.1. Variation of temperature as a function of time at the bottom of thick samples
which are photopolymerized under adiabatic conditions

100
90
80

 

1 I

1111I1111

T

l
T

60
50
40
30

Transmitance %

11111111111111

111

10 3}.

 

O N 1 1 I
T T T ﬁr T T T T T T T T - T T 1 T T T T

300 350 400 450 500 550
Wavelength (nm)

 

 

Figure 5.2. Transmittance spectrum of Derakane 510A in the 300-500 nm range. The
path length used was 1 cm

75

 

 

2.5 . m-

1.5 I

Extinction Coefficient (ml/gm.cm)

 

 

 

l 1 L
o r r r T 1 . r 1 r r r r r r 1 ﬁr 7 r r T T r 1 T

300 350 400 450 500 550
Wavelength (nm)

 

Figure 5.3. Plot of the extinction coefﬁcient of Derakane 510A as a function of
wavelength in the 350-550 nm range. The density of the resin is 1.2 gm/ml.

 

 

   
 

 

 

 

 

 

 

100 -
90 ‘3" ’//// l
. /

80 :e ,.-/
23 70 ; ///
3 60 /
g y / — Derakane 510A
= 50 1— I
E : / ~--~DTMPTA
m :
g 40 ;— ”DER 362 }
b : l
i" 30 i" 1

20 I

10%

300 350 400 450 500 550

Wavelength (nm)

Figure 5.4. Comparison of the transmittance spectra of three different resins in the 300-
550 nm range. The path length used was 1 cm in all cases.

76

4500 ,

 

4000-; H
3500 i— I
3000<§
2500-}

2000-}

 

 

 

T

 

Intensity (Counts)

1500 E

 

1000 }

 

500-;

o 1 1 1
T T T T T I 1 1 I Y I T T T T l I I I I

200 250 300 350 400 450 500
Wavelength (nm)

 

 

—o—

 

—n—
.1
.4

Figure 5.5. Emission Spectrum of the 1000 W Hg vapor lamp in the 280-500 nm region

 

4500
4000-; I

3500 §

3000 }

l
“‘3

2500 5
2000-i

1500 §

 

Intensity (Counts)

 

 

1000-i

 

500-}

 

 

 

o r r r 1 i r 1 1 1 Jr 1 r t I i r 1 ﬂ 1 Jr I 1 1 I 1r r r r r
200 250 300 350 400 450 500
Wavelength (nm)

 

Figure 5.6. Emission spectrum of 1000 W Hg vapor lamp in the 280-500 nm region after
it is transmitted through 0.9cm thick layer of Derakane 510A resin.

77

 

0.9 5-
0.8 T
0.7 «I
0.6
0.5 i
0.4 {—

Absorbance

0.3 :r
0.2 j“

 

 

 

 

.1
I 1 1 1
1 1 1 1 1 1 f 1 1 ‘1 ‘1 1 1 1' 1 r 1 1 1 1 1 1 T

260 300 340 380 420 460
Wavelength (nm)

Figure 5.7. Absorption spectrum of Benzoin Ethyl Ether. The concentration is 5.1 x 10'4
g/ml in acetone

1.2

 

L L

0.6 {I

Absorbance

0.4 {-

 

 

 

0 1 1 1 1
1 1 1 1 1 1 1 . 1 1 1 1 1 1 1 1 1 1 1 T f

260 300 340 380 420 460
Wavelength (nm)

Figure 5.8. Absorbtion spectrum of TPO. The concentration is 6.4x 10'4 g/ml in 2-
propanol

78

(L8

 

03'

11114111

(L5

(L4

1111111111111

Absorbance

(L3

 

(L2

OJ --

 

 

 

 

J ' L 1 L 1
#11111111 rrrﬁf‘1111111111ri1i111ﬁ11

320 340 360 380 400 420 440 460
Wavelength (nm)

Figure 5.9. Absorption spectrum of BAPO. The concentration is 5.1 x 10'4 g/ml in
acetone

900

 

 

800 5-
700 e
600

500 j”

400 j» free/M—
: //’
300.? '//////

200 :

Cure Time (Sec)
I
I
g

100 I

 

 

 

-4
o 1 1 1
T T T T T Y T 1 1 I T T T T T f

0 (L5 1 1.5 2
% TPO by Weight

Figure 5.10. Optimum concentration of TPO for curing 0.9 cm thick sample of Derakane
510A. The light intensity used was 65 mW/cm2 UVA.

400

350 ,

300 j

200

Cure Time (sec)

100 j

50-:

Figure 5.11.

900 .

800

N
O
O

Cure Time (sec.)

400 5

300

250 ~7

150

79

 

l

\—
~-._
\-

 

 

 

1
1 T 1 1 1 1 1 r 1 1 1

0 0.1 0.2 0.3 0.4
% BAPO by wt.

Optimum concentration of BAPO for curing 0.9 cm thick sample of
Derakane 510A. The light intensity used was 65 mW/cm2 UVA.

 

 

 

600

500 f

 

l 1 I - 1
1 1 1 1 1 1 I 1 1

 

I T T ‘. T T T T T T l

 

0.2 0.25 0.3 0.35 0.4 0.45 0.5

Wt. °/o BEE

Figure 5.12. Optimum concentration of BEE for curing 0.9 cm thick sample of Derakane

510A. The light intensity used was 65 mW/cm2 UVA.

80

Figure 5.13. Radicals formed by photolysis of TPO

81

0:0
0

1.5313
{F- + s

Figure 5.14. Radicals formed during complete photolysis of BAPO

82

 

 

 

1.4 “I" IConcentration = 6.4 x 10" glmlI

400

Wavelength (nm)

 

1.2 3—

0.8 3
0.6 9»

0.4 i

Absorbance

 

0.2

lllLlLllJl

 

 

o .IIII
320 360

 

Figure 5.15. Comparison of the absorption spectra of TPO and BAPO in the 320-450 nm
spectral region

 

 

 

 

110
e .1
a 90« .
i
E 803—
13 1
E -
: 70-—«~
.§
x
(B -
E 60f-
50_IIIIIIIIIIIIIIIII.IIIeIIII ..
O 0.05 0.1 0.15 0.2 0.25 0.3

Wt. °/o BAPO

Figure 5.16. Maximum temperature measured by the thermocouple during BAPO
initiated photopolymerizations of 0.9 cm thick samples of Derakane 510A.
The light intensity used was 65 mW/cm2 UVA

100

60

Maximum Temperature ( °C)

50

Figure 5.17.

500
450
400
350
300
250
200

Cure Time (sec)

100
50

90
so ’—

70 7

150 i

83

 

1
-

 

 

 

1 1 1 1
1 ﬂ. 1 1 1 1 1 1 1 1 1 1 1 1 1

0 02 04 05 03 1
“HM%'TPC’

Maximum temperature measured by the thermocouple during TPO initiated
photopolymerizations of 0.9 cm thick samples of Derakane 510A. The
light intensity used was 65 mW/cm2 UVA

 

111

1111L1111111

 

—<1—

1 1 1 1' 1
- T T f 1 T T T T T 7 : T I T T T T T T ‘T T

0 10 20 30 40 50 60 70
Intensity (mWIcm2 UVA)

 

 

Figure 5.18. Effect of light intensity on cure time of 0.9 cm thick sample of Derakane
510A

84

 

 

 

 

 

 

 

 

 

 

 

 

 

4500 ,
4000 3 ——Transmittance of 0.9 cm Thick Sample
j of Pure D510A
3500 T —Transmittance of 0.9 cm thick sample of
’0? : D510A containing 0.2wt% BAPO
E 3000 :—
8
o 2500 E-
B it
7’ 2000 1
c :
:2 1500 3
1000 -} 1
500 3— k
OIIIIIIIIIIIIrIIIIIrIITT'II
250 300 350 400 450 500

Wavelength (nm)

Figure 5.19. Comparison of emission spectra of Hg vapor lamp after passing through 0.9
cm thick sample of Derakane 510A resin with and without BAPO

 

 

 

 

photoinitiator
3
2.5
o 2 i
o -
t: .
to -
E 1.5 -
o
a
n a
< 1f
0.5
o F 1 1 1 1 i 1 1 1 1 i 1 1 1 1 i T T I I
320 370 420 470 520

Wavelength (nm)

Figure 5.20. Absorbance spectrum of Bis ['115 -2,4—Cyclopentadien-l-yl] bis[2,6-diﬂuoro-

3-[1H-Pyrrol-l-yl] phenyl titanium. The concentration is 2.4 x10'3 glml in
acetone

Absorbance

Figure 5.21.

Flexural Strength (MPa)

0.8 _,
0.6 I,
0.4 -

0.2 7

0

 

85

 

 

I 1 1 1
11 11 1 1 Ifrmr—r I YY‘rf—ﬁ 1Irﬁ

380 400 420 440 460 480 500 520

350

300 A

250 E

200

150

100

50

Wavelength (nm)

Absorption spectrum of Camphorquinone in the 380-520 nm region. The
concentration is 5.01 x 10‘3 g/ml in acetone

11111

 

-1111

 

 

 

 

 

 

Photocured Dual Cured Thermally
Cured

Figure 5.22. Comparison of ﬂexural strength of composites. The bar represents the

average value (of 5 samples) while the error bars represent the high and
low values obtained in the tests

86

5.5. List of References

A.K. Davies, Radiation Curing of Polymers [1, DR. Randell ed., The Royal Society

of chemistry, 1991

J.F. Rabek, Radiation Curing in Polymer Science and Technology, Volume I, J.F.
Fouassier and J .F. Rabek eds., Elsevier, 1993

R. Tirtha, P.L. Fan, J .B. Dennison, J .M. Powers, J Dent Res, 61, 1184, 1982
T. Renault and AA. Ogale, Polym Engg. Sci, 36, 551, 1996
T. Renault, A.A. Ogale, M.J. Drews, Proc. ANTEC ’93, 2352, 1993

J. Guthrie, M.B. Jaganathan, M. S. Otterburn, J. Woods, Polymer Bulletin, 15, 51,
1986

K. Dietliker, G. Hug, R. Kaeser, M. Kohler, U. Kolczak, D. Leppard, L. Misev, G.
Rist, W. Rutsch, Novel High Performance Bisacylphosphine Oxide (BAPO)
Photoinitiators, Ciba-Geigy Corporation, 1994

Coons, L.S., Rangarajan, B., Godshall, D., and Scranton, A.B., ACS Symposium
Series, in press

 

Chapter 6

Summary and Recommendations

6.1 Summary

Model for Photobleaching

A fundamental mathematical model was developed to characterize the
photobleaching of initiators when exposed to monochromatic light of the appropriate
wavelength. Given the incident light intensity and wavelength and the initial
photoinitiator concentration, the model predicts the variation of the photoinitiator
concentration and light intensity with time and sample depth. The model was veriﬁed
experimentally for four different Ot—cleavable photoinitiators: benzoin ethyl ether,
trimethyldiphenylbenzoylphosphine oxide, bis(2,4,6—trimethylbenzoyl) phenyl phosphine
oxide and a substituted titanocene. The model proved to be very useful ﬁrstly because it
quantiﬁed the photobleaching effect in terms of three constants which could easily be
determined, and secondly because it helped in comparing the photobleaching
effectiveness of different initiators and thereby in making a choice of photoinitiator for
curing of thick polymer and composite parts. For example it was shown that the phenyl
phosphine oxide initiator (BAPO) undergoes more efﬁcient bleaching than benzoin ethyl

ether.

87

 

88

Investigation of Photoinitiating Systems for Photopolymerization of Composites

Studies were performed to evaluate the suitability of different photoinitiators for
photocuring of thick polymer and composite parts. The benzoyl phosphine oxide
photoinitiators were chosen for further studies based upon their absorption spectra, the
emission spectrum of the UV lamp used and transmitance of the resin. The
photoinitiators chosen were then used to cure thick resin samples and the ﬁnal choice of
initiator was made based upon cure time considerations. Studies were performed to
determine the effect of light intensity, photoinitiator concentration and sample thickness
on the cure time of the resins. It was shown that there exists an optimum photointiator
concentration at which the cure time is minimum and conversion is maximum. The
results indicated that BAPO was the best photoinitiator and the optimum concentration
was 0.2 wt.%. It was observed that increasing the light intensity reduced the cure time
dramatically up to the point when saturation occurred and increasing the intensity beyond
this point did not reduce the cure time as rapidly. Studies performed to evaluate the effect
of sample thickness on cure time indicated that increasing the sample thickness resulted
in a small decrease in the optimum photoinitiator concentration. Hybrid photoinitiating
systems and dual photo/thermal cure schemes were investigated in an attempt to further

reduce the cure time. Finally, composite samples were prepared both by photocuring and

by thermal curing, their mechanical properties were evaluated and found to be equal.

 

89

6.2 Recommendations for Future Work
Choice of Photoinitiating System
BAPO was found to be the best photoinitiator for the system under consideration.
However, it was shown that the 440nm line emitted by the lamp was not utilized and
attempts to utilize this line by adding initiators that absorbed in this region were not
successful. This shows that behavior of hybrid photoinitiating systems is not so simple.
Further studies need to be conducted to understand how hybrid photoinitiator systems

work. Similar studies need to be conducted on dual photo/thermal initiating systems.

Development of a Polychromatic Photobleaching Model

The model developed in this thesis is applicable only if the light source is
monochromatic (or almost monochromatic). However, in several practical applications
involving photobleaching, the light source emits at several different wavelengths. For
example the commonly used mercury vapor lamp has at least 5 different emission lines in
the UVA spectral region. Hence, another area for future work is extension of the

monochromatic photobleaching model to polychromatic photobleaching.

Model for Photopolymerization of Thick Polymer Parts

To arrive at a model for photopolymerization of thick polymer parts, it is
necessary to combine the photobleaching model with a suitable model for the
polymerization kinetics. Most of the models for polymerization kinetics are either too

complicated or make the quasi steady state assumption which is incorrect. One possible

90
kinetic model that can be used is an empirical model proposed by Pusatcioglu et al I. The
model incorporates empirical constants which must be determined from experiment.

However, the model is being widely used to study polymerization kinetics either directly

or with minor modiﬁcationsz’3’4.

Thus, one possible area for further study is the
development of a working model for photopolymerization of thick polymer parts by

combining the photobleaching model with the model for polymerization kinetics.

Development of Cure Monitoring Techinques for Thick Systems

There are very few techniques available for cure monitoring of thick systems.
Development of more efﬁcient and reliable methods for monitoring the cure of thick
systems will undoubtedly lead to a better understanding of the photocuring process and
will result in highly efﬁcient photopolymerization techniques which could eventually

completely replace thermal cure.

6.3. List of References
l. S. Y. Pusatcioglu, A.L. Fricke, J. C. Hassler, J. Appl. Polym. Sci, 24, 937, 1979

2. J. M. Kenny, A. Maffezzoli, L. Nicolais, Composites Science and Technology, 38,
339, 1990

3. M. R. Kamal, S. Sourour, M. Ryan, Polym. Eng. Sci, 13, 59, 1973

P

H. Han, K. W. Lem, J. Appl.Polym. Sci, 28, 3155, 1983