Infrared Spectroscopic Investigation of the Effects
of Titania Photocatalyst on the Degradation of
Linear Low Density Polyethylene Film for
Commercial Applications.
by
Dylan John Nagle,
B. App. Sci. (App. Chem.), M. App. Chem.
A thesis submitted to the School of Physical and Chemical
Sciences in partial fulfilment of the requirements for the degree
of
Doctor of Philosophy
Queensland University of Technology
October 2009
1
2
Do it once, do it properly; never do it again.
These words of Peter M. Fredericks are probably the greatest lesson I have learnt in my years of study to complete the PhD degree, a lesson that requires continual revisiting. I wish to acknowledge the mentoring of my supervisory team; Peter Fredericks, Llew Rintoul and Graeme George. I am grateful for what I have learned from each one, academically and personally. I also acknowledge the efforts of my family, who have offered their utmost encouragement and support. Likewise my friends and colleagues at QUT. Ultimately it was my wife Mi Jeong who carried me when life was at its most challenging, and celebrated life with me at its most rewarding. I am forever thankful.
3
The work presented in this thesis is, to the best of my knowledge and belief,
original and my own work, except where acknowledged in the text. This material
has not been submitted, either in whole or in part, for a degree at this or any other
university.
Dylan John Nagle
October 2009
4
Abstract ____________________________________________________________8
List of Abbreviations ________________________________________________10
Introduction ______________________________________________________11 1.1 Polymer degradation _____________________________________________ 13
1.1.1 Thermooxidation _____________________________________________________ 14 1.1.2 Heterogeneous vs. homogenous thermooxidation kinetics______________________ 15 1.1.3 Photooxidation _______________________________________________________ 17 1.1.4 Role of hydroperoxides in polyethylene photooxidation _______________________ 20 1.1.5 Stabilisation of commercial polyethylene __________________________________ 23
1.2 Prodegradants___________________________________________________ 27 1.2.1 Titanium dioxide _____________________________________________________ 30 1.2.2 Titania photocatalysis__________________________________________________ 33 1.2.3 Factors affecting titania activity in polymers ________________________________ 35 1.2.4 Surface chemistry of titania _____________________________________________ 36 1.2.5 Surface modification of titania ___________________________________________ 37 1.2.6 Doping _____________________________________________________________ 39 1.2.7 Effect of UVA vs. UVC radiation on polymer – TiO2 systems __________________ 40 1.2.8 Summary of sections 1.1 and 1.2 _________________________________________ 41
1.3 Polymer degradation characterisation techniques _____________________ 42 1.3.1 Characterization of the bulk via physical tests _______________________________ 43 1.3.2 Surface Characterisation________________________________________________ 43 1.3.3 Chemical Characterisation ______________________________________________ 44 1.3.4 Achieving high lateral resolution _________________________________________ 53 1.3.5 Characterisation techniques used in this thesis_______________________________ 57
1.4 Objectives ______________________________________________________ 58 Experimental _____________________________________________________63
2.1 Ciba films investigation ___________________________________________ 63 2.2 Accelerated aging of samples_______________________________________ 65 2.3 Mid-IR spectroscopy _____________________________________________ 67 2.4 Imaging IR Spectroscopy__________________________________________ 68 2.5 Synchrotron experimental _________________________________________ 69 2.6 Scanning electron microscopy ______________________________________ 72
Effect of UV pre-irradiation on the degradation of polyethylene ___________73 3.1 Introduction ____________________________________________________ 73 3.2 Physical characteristics of commercial titanias and general comments ____ 73
3.2.1 Degussa P25 _________________________________________________________ 73 3.2.2 Kronos _____________________________________________________________ 74 3.2.3 Huntsman Tioxide ____________________________________________________ 75 3.2.4 Sachtleben Hombitan __________________________________________________ 76 3.2.5 Section summary _____________________________________________________ 77
3.3 Sample whitening ________________________________________________ 78 3.4 Times to embrittlement for LLDPE film containing titania______________ 80 3.5 IR spectral analysis – control film (undegraded)_______________________ 86
3.5.1 Polyethylene absorption table____________________________________________ 86 3.5.2 Titania absorption in the mid-infrared _____________________________________ 88
5
3.6 Processing agent absorptions _______________________________________ 88 3.7 IR spectral analysis – control film (degraded) _________________________ 90
3.7.1 OH stretc.hing region (3800-3200 cm-1) ____________________________________91 3.7.2 Carbonyl region _______________________________________________________91 3.7.3 Below 1500 cm-1 ______________________________________________________93 3.7.4 Section summary ______________________________________________________93
3.8 Effect of UV irradiation – control film (degraded) _____________________ 94 3.8.1 Control, weatherometer aged samples ______________________________________94 3.8.2 Control, oven aged samples______________________________________________98 3.8.3 Section Summary_____________________________________________________100
3.9 IR spectral analysis – film containing titania (degraded) _______________ 101 3.9.1 Carbonyl region ______________________________________________________101 3.9.2 Fingerprint region ____________________________________________________104 3.9.3 Section summary _____________________________________________________104
3.10 LLDPE containing Degussa P25 (degraded) _________________________ 105 3.10.1 Degussa P25, weatherometer aged samples ______________________________105 3.10.2 Section summary___________________________________________________109 3.10.3 Degussa P25, oven aged samples ______________________________________110 3.10.4 3% Degussa P25 samples ____________________________________________111 3.10.5 Section summary___________________________________________________114
3.11 LLDPE containing Kronos 1002 (degraded) _________________________ 115 3.11.1 1% Kronos 1002, weatherometer aged samples,___________________________115 3.11.2 3% Kronos 1002, weatherometer aged samples,___________________________116 3.11.3 1% Kronos 1002, oven aged samples, __________________________________117 3.11.4 3% Kronos 1002, oven aged samples, __________________________________118 3.11.5 Section Summary __________________________________________________118
3.12 LLDPE containing Huntsman Tioxide (degraded) ____________________ 119 3.12.1 3% Huntsman tioxide A-HR, weatherometer aged_________________________119 3.12.2 3% Huntsman tioxide A-HRF, weatherometer aged________________________121 3.12.3 3% Huntsman tioxide A-HR, oven aged_________________________________122 3.12.4 3% Huntsman tioxide A-HRF, oven aged________________________________123 3.12.5 Section summary___________________________________________________123
3.13 LLDPE containing Sachtleben Hombitan (degraded)__________________ 124 3.13.1 3% Sachtleben Hombitan, weatherometer aged ___________________________124 3.13.2 3% Sachtleben Hombitan, oven aged ___________________________________126 3.13.3 Section summary___________________________________________________127
3.14 Discussion of the effects of titania __________________________________ 128 3.15 Conclusions ____________________________________________________ 131
Multivariate Data Analysis _________________________________________135 4.1 Introduction____________________________________________________ 135 4.2 Data treatment__________________________________________________ 136 4.3 Analysis of samples subjected to oven aging__________________________ 137
4.3.1 Samples without pre-irradiation__________________________________________137 4.3.2 Samples with pre-irradiation ____________________________________________141 4.3.3 UVA vs UVC pre-irradiation: extent of degradation information ________________144 4.3.4 Section Summary_____________________________________________________151
4.4 Weatherometer aging ____________________________________________ 151 4.4.1 Water vapour ________________________________________________________152 4.4.2 UVA vs. UVC pre-irradiation ___________________________________________157 4.4.3 Section summary _____________________________________________________157
4.5 Conclusions ____________________________________________________ 157
6
Obtaining spatial information around titania particles via a model polymer system ____________________________________________________159
5.1 Introduction ___________________________________________________ 159 5.2 Experimental___________________________________________________ 161 5.3 Imaging ATR/FTIR spectroscopy results____________________________ 163
5.3.1 Determination of titania particle location(s)________________________________ 165 5.3.2 Discussion of heterogeneous oxidation ___________________________________ 174
5.4 Conclusions ____________________________________________________ 178 Investigation of degradation in the mid-IR using a synchrotron light
source ____________________________________________________________181 6.1 Introduction ___________________________________________________ 181 6.2 Experimental___________________________________________________ 181 6.3 Synchrotron results and discussion_________________________________ 184 6.4 Conclusions ____________________________________________________ 191
Conclusions ______________________________________________________193
References _______________________________________________________199
7
Abstract There is a need in industry for a commodity polyethylene film with controllable
degradation properties that will degrade in an environmentally neutral way, for
applications such as shopping bags and packaging film. Additives such as starch
have been shown to accelerate the degradation of plastic films, however control
of degradation is required so that the film will retain its mechanical properties
during storage and use, and then degrade when no longer required. By the
addition of a photocatalyst it is hoped that polymer film will breakdown with
exposure to sunlight. Furthermore, it is desired that the polymer film will degrade
in the dark, after a short initial exposure to sunlight.
Research has been undertaken into the photo- and thermo-oxidative degradation
processes of 25 µm thick LLDPE (linear low density polyethylene) film
containing titania from different manufacturers. Films were aged in a suntest or
in an oven at 50 °C, and the oxidation product formation was followed using IR
spectroscopy. Degussa P25, Kronos 1002, and various organic-modified and
doped titanias of the types Satchleben Hombitan and Hunstsman Tioxide
incorporated into LLDPE films were assessed for photoactivity. Degussa P25
was found to be the most photoactive with UVA and UVC exposure. Surface
modification of titania was found to reduce photoactivity. Crystal phase is
thought to be among the most important factors when assessing the photoactivity
of titania as a photocatalyst for degradation. Pre-irradiation with UVA or UVC
for 24 hours of the film containing 3% Degussa P25 titania prior to aging in an
oven resulted in embrittlement in ca. 200 days.
The multivariate data analysis technique PCA (principal component analysis)
was used as an exploratory tool to investigate the IR spectral data. Oxidation
products formed in similar relative concentrations across all samples, confirming
that titania was catalysing the oxidation of the LLDPE film without changing the
oxidation pathway. PCA was also employed to compare rates of degradation in
different films. PCA enabled the discovery of water vapour trapped inside
cavities formed by oxidation by titania particles.
8
Imaging ATR/FTIR spectroscopy with high lateral resolution was used in a novel
experiment to examine the heterogeneous nature of oxidation of a model polymer
compound caused by the presence of titania particles. A model polymer
containing Degussa P25 titania was solvent cast onto the internal reflection
element of the imaging ATR/FTIR and the oxidation under UVC was examined
over time. Sensitisation of 5 µm domains by titania resulted in areas of relatively
high oxidation product concentration.
The suitability of transmission IR with a synchrotron light source to the study of
polymer film oxidation was assessed as the Australian Synchrotron in
Melbourne, Australia. Challenges such as interference fringes and poor signal-to-
noise ratio need to be addressed before this can become a routine technique.
9
List of Abbreviations ATR/FTIR Attenuated total reflectance/FTIR (spectroscopy)
CB Chain breaking
EDAX Energy dispersive X-ray analysis
ETD Everhart-Thornley detector
FTIR Fourier transform infrared (spectroscopy)
FPA Focal plane array detector
HALS Hindered amine light stabiliser
HOMO Highest occupied molecular orbital
HDPE High density polyethylene
IR Infrared
IRE Internal reflection element
LDPE Low density polyethylene
LLDPE Linear low density polyethylene
LUMO Lowest unoccupied molecular orbital
MCT Mercury cadmium telluride
NMR Nuclear magnetic resonance (spectroscopy)
PC Principal component
PCA Principal component analysis
PMMA Polymethyl methacrylate
PVC Polyvinyl chloride
QUT Queensland University of Technology
RI Refractive index
SEM Scanning electron microscopy
S/N Signal-to-noise ratio
SSD Silicon strip detector
UV Ultraviolet
10
Introduction Low-end commodity plastics such as polyethylene are in high demand – 60
million metric tons were produced in 2004 worldwide1. Low Density
Polyethylene (LDPE) and Linear Low Density Polyethylene (LLDPE) are the
two most common forms of polyethylene other than High Density Polyethylene
(HDPE), and are mostly processed as sheets and films for applications in
packaging, shopping bags, agriculture, etc.. Due to our high consumption of
polyethylene, the matter of disposal of used plastic has evolved as a contentious
issue. As a global community we are becoming more successful at recycling
unwanted plastic, but in many situations the added cost of recycling is too heavy
an economic burden, and for industries such as agriculture it is wholly
impractical. Currently the most common method of disposal is burying beneath
soil, which coincidentally prevents the plastic from degrading due to the absence
of sunlight.
In recent times scientists have sought to develop plastics with more controllable
degradation properties to create an environmentally neutral film. An example is
the addition of starch to polyethylene, attempting to make it biodegradable2.
Unfortunately degradable additives such as starch often inhibit mechanical
properties3 and, rather than achieving ‘controllable’ degradation, serve merely to
accelerate the degradation process. This has a clear effect on the properties of the
material in question, such as shelf life, where the polymer is already degrading
before being used.
To combat these issues technology is being developed to more strictly control the
degradation properties of various plastics. A successful approach has been the
addition of a material that will accelerate degradation processes when exposed to
sunlight. Such additives are termed ‘photosensitisers’, and exploit the radical
chemistry occurring during photodegradation. Among other materials, transition
metal salts in particular such as cobalt4, iron5 and nickel6 have been demonstrated
to exhibit photosensitizing effects in polymeric materials.
11
A common photosensitiser is nano-particulate titania, which has been
demonstrated to greatly enhance the degradation properties of various polymers
when exposed to UV radiation7. Titania holds great potential as a photosensitiser
for real world applications as it accelerates the degradation process, hopefully
preventing a buildup of buried undegraded plastic. Once the molecular weight of
a polymer has been sufficiently reduced via photooxidation, microbial or biotic
degradation can proceed8.
While technology such as this is certainly a step in the right direction, the
demand for plastics with a high degree of control over degradation is increasing,
and thus science must look deeper to provide better degradation management.
Beyond simply accelerating the degradation process, it is desirable to pre-
determine the length of time a plastic film will maintain its mechanical properties,
tunable to the situation required. Thus the ultimate objective of this research is to
investigate a method for controlling LLDPE film lifetime, according to the
application.
A method of achieving this goal will be investigated by examining the effects of
pre-irradiation of LLDPE film containing titania with UV before aging in a dark
environment. Titania catalyses oxidation of organic materials by absorbing UV
radiation and creating radical species that are involved in the initiation step of
oxidation processes9. However the concept under investigation is that of pre-
irradiation, which involves the exposure of a polymer containing titania to UV
irradiation in order to create reactive sites throughout the polymer matrix, which
can then proceed to propagate degradation reactions which spread throughout the
material, even in the absence of light, similarly to an infection spreading through
a population10.
By utilising pre-irradiation technology, a measured dose of UV can be applied to
a polymeric material, such as a shopping bag, in order to initiate oxidation
processes. The polymer will then proceed to degrade, within a known time frame
pre-determined by the strength and the time of the UV dosage.
12
In order to achieve an understanding of pre-irradiation and the effect of titania as
a photosensitiser in commercially available LLDPE film, samples containing
several different types of titania from different manufacturers have been exposed
to UV irradiation, and then aged under accelerated conditions while being
periodically monitored by mid-infrared spectroscopy. The mechanism of spread
of oxidation originating at a titania particle has also been examined using
infrared imaging spectroscopy, as well as high lateral resolution spectroscopy
using a synchrotron radiation source.
An understanding of empirical and mechanistic effects of pre-irradiation of nano-
particulate titania with UV on the photodegradation of LLDPE will be developed
by analysis of the data obtained from the experimental methods outlined above.
It is hoped that data will provide a greater understanding of the fundamental
processes involved in titania-catalysed degradation, which can be exploited by
future researchers to assist in developing technology that will allow more
accurate control over the degradation of commodity plastic film.
1.1 Polymer degradation
There are seven processes by which a polymer can degrade11:
1. Thermal: the application of heat
2. Mechanical: the application of force
3. Ultrasonic: the application of sound waves
4. Hydrolytic: attack on certain functional groups along the polymer
chain by water
5. Chemical: attack by corrosive chemicals or gases, such as ozone
6. Biological: attack on certain functional groups by microbes
7. Radiation: absorption of radiation at certain frequencies that induces
reactions
Often, there is not just one process at work in the degradation of a polymer, and
the nature of oxidation processes involved in the breakdown of a particular
plastic will depend on the degradation environment of the plastic. Following the
description of the goals of this project presented in the introduction, it is
13
desirable to develop a plastic film that retains its mechanical properties during its
usable lifetime, and then will disintegrate into particles small enough to allow
microbial action to breakdown the molecular structure of the polymer.
Mechanical degradation is of lesser relevance to this study than other degradation
processes, as the focus is on the breakdown of the film after disposal, by which
time the mechanical properties of the film are no longer relevant. Additionally,
the technology has been designed to oxidise the plastic film without requiring the
application of mechanical degradation processes.
Biotic breakdown of the plastic film is important to ensure that the film is
environmentally neutral; however this will not be discussed further in this thesis
as it does not pertain directly to oxidative degradation. Ultrasonic and hydrolytic
degradation processes are also not relevant to the degradation of waste
polyethylene film for commercial applications. Chemical degradation will be
discussed from the point of view of oxidation, or chemical attack by atmospheric
oxygen. The degradation processes to be investigated in this thesis are termed
thermooxidation (application of heat and attack by oxygen) and photooxidation
(application of radiation and attack by oxygen).
1.1.1 Thermooxidation
There are three principal steps involved in the oxidation of a polyolefin12:
1. Initiation:
By radical generator
I (initiator) 2r
r RH rH R+ + By hydroperoxide
ROOH +R HOO
ROOH RO + HO Scheme 1-1
14
2. Propagation
R O2 ROO
ROO RH ROOH R
2ROOH RO ROO H2O
+
+ +
+ + Scheme 1-2
3. Termination
2ROO ROH R=OO2+ +or ROOR + O2
ROO R+ ROOR
2R R-R Scheme 1-3
H = polymer, R• = polymer macroradical
ese processes
ccurring simultaneously during the degradation of a polyolefin14.
kinetic models are
eveloped to assist in polymer lifetime prediction studies17.
R
Initiation of a polymer chain radical, or macroradical, occurs via the abstraction
of a hydrogen from the carbon backbone by a radical species. Alternatively,
initiation reactions can result from the cleavage of a hydroperoxide, which is
itself an oxidation species. Subsequent attack by O2 on the macroradical results
in the formation of a reactive hydroperoxide radical at the carbon centre.
Oxidation can then spread to other polymer chains13. Radicals are inherently
unstable, and will then terminate by creating hydroxyl groups, carbonyl
functional groups or cross-links. It is not uncommon to see all th
o
1.1.2 Heterogeneous vs. homogenous thermooxidation kinetics
The oxidation reactions presented in Section 1.1.1 are used in combination with
chemical measurements, such as oxygen uptake, to develop models describing
the kinetics of polymer degradation15,16. Ultimately, such
d
15
It has been convention to interpret polymer degradation in terms of the steady
state approximation18. This is at least partially due to the use of oxygen uptake
measurements, which is a bulk measurement. Oxygen uptake curves demonstrate
linearity past the induction period of a polymer19, providing good correlation
with the steady state approximation.
In addition to oxygen uptake measurements, kinetic information has
conventionally been obtained from polymers in solution20. Thus complications
such as radical mobility, polymer chain mobility, morphology, and oxygen
diffusion limitation occurring in solid state systems cannot be correctly
accounted for in a homogenous oxidation model10.
Chemiluminescence data was used by George and Celina13 to propose a
heterogeneous oxidation model for the oxidation of polypropylene. Investigation
of the oxidation of polypropylene powder at 150 °C revealed that a particle
undergoing oxidation could infect a nearby stable particle. It was found that even
after short oxidation times, oxidation products could be observed in localised
zones, which were thought to exist around particles of residual catalyst.
Furthermore, George and Celina postulated that the oxidation of polypropylene
was heterogeneous even within amorphous regions of the films.
This view of localised oxidation zones on polypropylene films was used to
explain the phenomenon of cracking in oxidised polypropylene. Once oxidation
was initiated around a catalyst particle, an oxidation front was formed which
progressed through amorphous regions, resulting in defects on a macroscopic
scale. With further oxidation these linked defects formed cracks in the polymer
surface. This explains why slightly oxidised polypropylene sheets demonstrated
reduced tensile strength, despite only low concentrations of oxidation products.
The effect of reduced tensile strength at low levels of oxidation has been well
demonstrated in the literature21-23.
The concept of heterogeneous oxidation in solid state polymer films, which then
leads to cracking, is fundamental to the chemistry underlying the experiments
carried out in this thesis. The phenomenon of oxidation spreading through a
16
polymer in the solid state proposed by George and Celina can be exploited by the
addition of chromophoric materials to enhance degradation. Similar to the spread
of oxidation from catalyst residues, oxidation spreads from introduced
photocatalysts to enhance degradation.
1.1.3 Photooxidation
A great deal of research has been done on the photooxidation of polyolefins, and
in particular polyethylene, over the last half century24-34. This section describes
some fundamental photooxidation chemistry as described by recognised research
leaders in this field.
Polymers containing only C-C, C-H and C-O single bonds are not expected to
absorb in the UV wavelength range35. For such polymers to degrade
photochemically a chromophore must be present. A chromophore might be an
impurity which is chemically bonded to a polymer chain, either in the middle
section, or at the end of a chain. Alternatively, a chromophore might be an
impurity present as an occlusion and is not chemically bound, but is contained
within the polymer matrix. A typical example is catalyst residues. Finally, a
chromophore might be part of the polymer structure itself, such as double bonds,
etc.. It is via these chromophores that photodegradation reactions are initiated.
The differences and similarities between thermooxidation and photooxidation
have been studied for many decades. In 1954 Rugg et al.25 determined that
thermooxidation of polyethylene resulted in little or no differences in the infrared
absorption intensity of unsaturated moieties. Photooxidation however produced
an overall increase in unsaturation, particularly in terminal vinyl group
concentration, and internal double bonds. Side-chain methylene groups were
found to decrease in concentration.
The degradation pathways favoured by polymers are typically reverse-analysed;
information regarding the structure of oxidation products is obtained using
conventional characterisation methods such as infrared spectroscopy, and from
this the likely degradation pathway is deduced. It is critical, therefore, to
17
understand the relationship between degradation products and the process(es)
that resulted in the products. One of the most indicative degradation products is
the carbonyl group.
Carbonyl groups formed mid-chain, such as ketones, are generally the result of
chain branching reactions, whereas terminal carbonyl groups, such as aldehydes,
are a consequence of β–scission. Ketones can undergo reactions resulting in
cleavage near the carbonyl bond via Norrish type I (resulting in two radical
species) or Norrish type II (yielding a vinyl group and a ketone) reactions.
1. Norrish type I
O O
O
+ CO
+hυ
Scheme 1-4
2. Norrish type II
hυO
H
O
+
O
Scheme 1-5
Polymer conformation, the availability of γ-hydrogens, polymer mobility and
other factors control the probability of Norrish type I and Norrish type II
photoreactions. Below the glass transition (Tg) temperature the rate of formation
18
of Norrish type II depends on the ability to form the cyclic intermediate, and thus
the reaction is limited by the mobility of the polymer chains35. Above Tg the
mobility of the chains is such that the cyclic intermediate is no longer rate
controlling and is kinetically similar to a polymer in solution. Below Tg the lack
of chain mobility prevents separation of the Norrish type I radical species, and
thus does not occur.
Allen and Edge11 describe the importance of carbonyl species in
photodegradation of solid state polymers. Carbonyls are chromophores, and by
absorbing UV radiation, the carbonyl oxygen can be promoted to an excited
triplet state. This may be quenched by ground state molecular oxygen, resulting
in a transfer of energy to the O2 molecule, giving an excited singlet oxygen. This
reacts with unsaturated sites to produce hydroperoxides, according to:
Ohυ
O* O
3O2 1O2+
1O2
O2H
+
Scheme 1-6
The exact significance of singlet oxygen in photooxidation of polymers such as
polyethylene is still disputed. This is due to the fact that much of the
experimental data comes from model system experiments, involving polymers
above the Tg. Experimental evidence suggests that the above mechanism is
inefficient in the absence of ketones, while others theorise various conflicting
mechanisms for the above reaction to proceed. Clearly oxygen, photons,
chromophores and unsaturation combine to result in oxidation; however the exact
mechanism is unknown. It is possible that the many different mechanisms exist
in competition with each other, and the many factors affecting oxidation such as
temperature, incident radiation wavelength, presence and type of chromophores,
19
polymer chain mobility, etc.etera, determine the most likely degradation
pathway.
Other species involved in photooxidation outlined by Allen and Edge are
oxygen-polymer charge transfer complexes. Charge transfer complexes are used
to describe an alternative pathway for the formation of hydroperoxides via attack
by oxygen. Oxygen abstracts an electron from a hydrogen on the polymer
backbone to generate an charge-separated complex. An intermediate of a
polymer radical and hydroperoxide radical is formed, which recombine to give
the final hydroperoxide. However, questions still remain regarding the efficiency
of this reaction, while others argue that once an initial hydroperoxide if formed,
oxygen-polymer charged transfer species are auto-catalysing35. It is likely that in
processed polymers they have little practical significance compared to the effect
of hydroperoxides36.
1.1.4 Role of hydroperoxides in polyethylene photooxidation
In the 1980s Arnaud et al.37,38 produced some important papers regarding the
photooxidation of polyethylene. It was found that most unsaturated groups
formed by Norrish II reactions rapidly disappeared due to subsequent radical
attack. Preferential oxidation sites were carbons in the α-position to the
vinylidene. This was not true for the vinyl groups, and was believed to be due to
low lability of the vinyl hydrogen. After an initial increase in the formation of
vinyl groups upon exposure to UV radiation, the rate of vinyl group formation
was found to parallel that of acid groups, indicating subsequent oxidation. Also,
it was found that vinyl and vinylidene groups were competing for radicals during
photooxidation.
In 1990 Gugumus39 suggested some novel reactions to explain the presence and
relative concentrations of some degradation products of photooxidised LLDPE,
as well as the lack of hydroperoxide accumulation in polyethylene when exposed
to radiation. In contrast to much of the published literature, Gugumus suggested
that the photolytic decomposition of hydroperoxide did not involve a radical
20
species. The proposed mechanisms involved a 6 membered transition state, as
well as the evolution of water as a product.
Gugumus used these 6 membered transition state reactions to propose reactions
that give vinyl, ketone and aldehyde products. Included as an example in Scheme
1-7 is the reaction between a hydroperoxide and polymeric carbon to yield a
ketone.
C
O H
O
H H
HC
hν
C
O H
O
H H
HC
*
C
O
O
H H
HC+
H
Scheme 1-7
The following year Lacoste et al.40 performed a similar study to Gugumus
producing similar results; however Lacoste suggested already established
mechanisms to explain the same degradation products. In order to simplify the
reaction system, LLDPE samples were pre-oxidised by γ–radiation in air slightly
to develop hydroperoxides, and then exposed to UV radiation in the absence of
oxygen so that the degradation products of these hydroperoxides could be
studied. Secondary hydroperoxides were formed and lost during 100 hours of
irradiation. Carbonyl and free alcohol species increased in concentration. End
carboxylic acid groups and esters also increased, along with γ–lactones. Some
vinyl groups were initially lost, although a slight increase in trans-vinylene was
found. Ketones were found to be created by Norrish type I and II cleavage
reactions. In all cases Lacoste et al. used radical chemistry to explain the
formation of oxidation products, give in Scheme 1-8.
21
ROOH RO + OH
RO + OH + RH R + H2O + ROH
R O2+ ROO
ROO + RH ROOH R+
2ROO ROH + R'C(=O)R" + O2
heat orlight
Scheme 1-8
It is apparent that there is not a single, elegant solution to describe the exact
process of polyethylene photooxidation. Different oxidation products, in
differing concentrations, result from different reaction conditions, and even
manufacture of polyethylene41. The deeper one delves into the published
literature, the deeper the divides in the opinion of the polymer degradation
community become apparent. Conjecture and supposition regarding mechanisms
are based on scientific evidence; it is the interpretation of experimental data that
is likely to be debated for some time to come. It is helpful to consider the
mechanism proposed by Tidjani42, proffering a simplified overview of the
polyethylene photooxidation process, stemming from a widely accepted
hydroperoxide intermediate.
22
H
Scheme 1-9 Polyethylene photooxidation pathways proposed by Tidjani42. ©2008 Elsevier Science. By following the various degradation pathways in the Tidjani degradation
scheme, degradation products including esters, alcohols, acids, ketones and vinyl
moieties are expected in photooxidised polyethylene. Although the exact
mechanisms may not be fully agreed upon, it is clear that there is a relationship
between the products, hydroperoxide intermediates and the effects of UV
radiation absorption.
1.1.5 Stabilisation of commercial polyethylene
As we have seen there has been a great deal of research devoted to understanding
and establishing the degradation pathways of polyethylene. However for use in
commercial applications, these degradation processes must be moderated for a
polyolefin film to serve its intended purpose. Thus, antioxidant additives are
included during processing to prolong the lifetime of polyethylene.
C
OOH
hυ
C
H
O
+ OH
C
O
+ H 2O
cage effectC O
C
O
O R C
H
O H
PH
+ P CHC
H2
O
+ H2CE s ter 17 3 5 cm- 1 Al c o hol 3400 cm-1
COOH
O
C H3
+ H 2C CH
Norrish Ior
OH N orri sh II
Acid 1710 cm-1
Keton e 1720 c m - 1 V i n
1yl
0 1c
640 and9 m-1
ra di cal att a c k
C O O H
R
β-scissi o n
O
23
Antioxidants can be classified into two groups36:
• Chain breaking (CB) antioxidants, which trap radicals formed
during the propagation step, and;
• Preventive antioxidants, which stabilise hydroperoxide,
effectively reducing the rate of initiation.
Chain breaking antioxidants are commonly added as stabilisers against
thermooxidation, and are of particular importance for polyolefins due to high
processing temperatures43. These antioxidants trap alkyl radicals, preventing
further propagation reactions:
R CB R-CB+ Scheme 1-10
A common chain breaking type antioxidant is Irganox 1010 pictured in Figure
1-1. Trapping of the alkyl radical occurs at the phenol. Irganox 1010 in particular
has many industrial applications and is used by Ciba, whose films are used in this
thesis. Some hindered amine stabilisers with multiple aromatic groups also
provide UV stability, for example Tinuvin 327 and Chimassorb 81.
OHO
O
4 Figure 1-1 Irganox 1010
24
Some of the most effective preventive antioxidants are nickel dithiolate
complexes, which remove hydroperoxy groups. Studies have shown that polymer
hydroperoxides cannot be detected in polyethylene or polypropylene processed
with nickel dithiolate complexes36. The mechanism for scavenging of
hydroperoxide by the phosphate version of a dithiolate complex published by
Scott in 198344 is included in Scheme 1-11.
Scheme 1-11
Another important class of hydroperoxide decomposing stabiliser is phosphite or
phosphonite stabiliser45. Aryl phosphites, such as pictured in Scheme 1-12,
demonstrate very efficient competition with polymer RH for chain propagating
radicals. Alkyl phosphites are not used as stabilisers as the radical formed in the
reduction of the phosphate radical is alkyl and will create further active radical
species.
25
ArO
P
ArO
ArO
OAr + OOR P
OOR
OAr
OAr
ArO
P
O
OAr
OAr
ArO + RO
P
ArO
ArO
OAr + RO P
OR
OAr
OAr
ArO
P
ArO
ArO
OR +
ArO + OOR Inactive Products Scheme 1-12
Best stabilisation of polyethylene, and many other types of polyolefins for that
matter, is achieved by combining both of these classes of stabiliser45,46.
Typically, this includes high molecular mass or hindered amine light stabilisers
(HALS), in combination with phosphites or phosphonites47. Thus chain breaking
antioxidants are strongest during the early lifetime of the polymer, interrupting
crosslinking reactions and competing with hydroperoxides, while preventive
antioxidants compete with polymer chains for hydroperoxy radicals, helping to
remove them from the system.
Although antioxidants provide a mechanism to prolong the lifetime of a polymer,
especially by preventing degradation reactions during the melt, it is the object of
this thesis to examine methods of accelerating oxidation reactions to produce a
polymer with controllable degradation characteristics, as that is the ultimate goal
of this work.
26
1.2 Prodegradants
The presence of foreign substances, such as metal ions, in a polymer matrix has a
pronounced effect on the degradation of that material6,31,48-50. In 1970 May and
Basharah51 listed the degradation reactions involving metal ions. These reactions
were adapted from an earlier paper produced by Chalk and Smith in 195752, and
are as follows:
RHhυ
R + H
R + O2 ROO
ROO + RH ROOH + R
ROOH Mn++ + +H+ROO M(n-1)+
ROOH + M(n-1)+ RO OH
ROH
+ Mn++
RO RH+ R+ Scheme 1-13
It was found that the catalytic activity of the metals appeared to be related to
their oxidation potential. The order of catalytic activity of the metals is Co > Fe >
Ce = Cu > Mn = Pb > Zn > Ca. This implies that the electromotive force
associated with reduction is related to the catalytic activity of the metals.
Stabilisers such as metal deactivators can be added to the polymer to slow
degradation. Metal deactivators, for example phenylamines, trap the metal ions,
inhibiting their oxidative catalytic effect53.
There are various methods by which metals and/or metallic ions can be included
in a polymeric material. The vast majority of commercially manufactured
polymers contain metal ions as impurities from polymer catalysts11, and these
impurities most often result in accelerated degradation of the material54. Metals
can also be deliberately introduced in the form of ions or complexes as
prodegradants to accelerate oxidation4-6,31,48,55,56. There are also cases where
27
metals can actually interact with the polymer system to work as oxidation
retardants54, and are termed prohibitors.
In 1988 Osawa54 listed five different mechanisms by which a metallic compound
can behave as a prodegradant in a polymer matrix. These reactions are an
extension of the degradation reactions given by May and Bashara involving
metal ions shown in Scheme 1-13.
1. Catalytic decomposition of hydroperoxides
Metal ions can react with hydroperoxides to produce free radicals, according to
the following reactions:
ROOH + Mn+ RO + M(n+1)+ + OH-
ROOH + M(n+1)+ ROO + Mn+ + H+ Scheme 1-14
The reactions are in reverse order to those given by May and Basharah in
Scheme 1-13. However the process of metal catalysis and product formation can
be summarised by:
2ROOHMn+/M(n+1)+
RO + ROO + H2O Scheme 1-15
2. Direct reaction with the substrate
This results in the production of free radicals:
RH + MX2 R + MX + HX
RH + MX R + M + HX Scheme 1-16
28
3. Activation of oxygen
Transition metals may interact with oxygen to produce a charge transfer
complex, which can then create hydroperoxy radicals which react with the
polymer:
Mn+ + O2 M(n+1)+ + O2
O2 + H+ HO2 Scheme 1-17
4. Decomposition of a metallic compound
Energy can initiate the decomposition of a metallic compound to produce a free
radical, which can then go on to react with the polymer:
M + XMXhυ
RH + X R + HX Scheme 1-18
5. Photo-sensitising action
An electron in the metal’s outer shell may be promoted from the ground state to
an excited state by the absorption of radiation. Subsequent transfer of energy to
the polymer molecule upon relaxation induces a radical.
M*Mhυ
M* + RH M + RH*
RH* R + H Scheme 1-19
29
The last reaction type involves semi-conductors, and is often called
semiconductor photocatalysis. A common semiconductor incorporated into
polymeric systems is titanium dioxide.
1.2.1 Titanium dioxide
Titanium dioxide is commonly used in polymer manufacture as a pigment, and
made up about 60% of global pigment production in 200257. Titanium dioxide
(TiO2), or titania, exists in three different crystal lattice structures: rutile, anatase
and brookite. Brookite is not commonly used due to its poor stability, and
therefore the considerable majority of discussion found in the literature regarding
the photoactivity of titania involves either anatase or rutile. Rutile is the most
thermodynamically stable of these forms. Microparticle TiO2 powder is suitable
for use as a white pigment due to its high refractive index and lack of absorption
in the visible range of the spectrum between 380 nm and 700 nm wavelength.
Rutile TiO2 has a refractive index of 2.7, slightly higher than anatase at 2.55.
Anatase and rutile have numerous structural and functional differences.
Commercially available anatase is typically less than 50 nm in size with the
particles possessing a band gap of 3.2 eV, corresponding to a UV wavelength of
387 nm58. The adsorptive affinity of anatase for organic compounds is higher
than that of rutile, and anatase exhibits lower rates of recombination in
comparison to rutile. In contrast, the thermodynamically stable rutile phase
generally contains particles larger than 200 nm with a smaller band-gap of 3.0
eV. The excitation wavelengths extend into the visible spectrum at 410 nm.
Despite this, anatase is generally regarded as the more photochemically active
phase, due to the combined effect of lower rates of recombination and higher
surface adsorptive capacity59.
The different crystal faces of rutile and anatase titania influence the chemistry
occurring at the surface of a titania particle9. The most thermally stable crystal
face of rutile TiO2 is (110), depicted in Figure 1-2a. Anatase has two stable
30
surfaces, (101) and (001), of which (001) is the most common and is given in
Figure 1-3a. The (100) face is less common in nanoparticles. Oxygen
deficiencies in the rutile (110) of titania provide reaction sites for redox
chemistry such as water cleavage and oxygen adsorption.
Figure 1-2 Some crystal faces of rutile titania9. a (110), b(100), c(001) ©2008 Elsevier Science.
31
Figure 1-3 Some crystal faces of anatase titania9. a (101), b(100), c(001) ©2008 Elsevier Science. It is the photoactivity of titania nanoparticles that is a desired property when
using titania as a prodegradant in polymeric materials. The reactions relating to
the photochemistry of titania and its photocatalytic properties have been an
increasing area of interest for some time56,60.
32
1.2.2 Titania photocatalysis
Although titania had been well known as a white pigment in paint due to its
reflective properties, it was not until the first half of the 20th century that research
was first conducted into the phenomenon of paint chalking in sunlight9. Chalking
is the appearance of white powder on the surface of paint, so named for its
similarity with chalk. It was recognised that oxidation and reduction reactions
were occurring simultaneously.
In their review of semiconductor photocatalysis Mills and Le Hunte49 group the
terms ‘photocatalysis’, ‘photoinduced reaction’, ‘photoactivated reaction’ and
‘photosensitisation’, and define them as “a process by which a photochemical
alteration occurs in one chemical species as a result of the initial absorption of
radiation by another chemical species called the photosensitiser”.
If a semiconductor absorbs light of energy greater than the ∆E of the bandgap
(Figure 1-4), an electron (e-) can be promoted from the valence band (HOMO, or
Highest Occupied Molecular Orbital) to the conduction band (LUMO, or Lowest
Unoccupied Molecular Orbital), creating a hole (h+) in the valence band. This
electron-hole pair is termed an ‘exciton’. There are several possible outcomes of
such a reaction, with simple recombination of the e- and h+ being the most
common.
33
Atomic orbitals Molecule Cluster Q-size particle Semiconductor N=1 N=2 N=10 N=2000 N>>2000
LUMO
Figure 1-4 The energy required to excite electrons from the ground state (HOMO) of a
semi-conductor to the excited state (LUMO) decreases with increasing number of units N.
According to Scheme 1-20 and Scheme 1-21, if an electron acceptor or donor
such as oxygen, water, hydrogen peroxide or organic molecule is present, the
electron-hole pair may form a radical species instead of recombining:
Acceptor:
e + O2 O2
e +
+e
H2O2 OH + OH
R + H RH Scheme 1-20
Donor:
+ O2O2
+
+
H2O OH +
R + HRHh
h
h
H
Scheme 1-21
These reduction and oxidation reactions provide the radical species that can
initiate degradation reactions in polymers. Holes are the primary oxidising
Energy
∆E ∆E ∆E ∆E
HOMO
34
species in photocatalytic reactions9, reacting with water to produce hydroxyl
radicals, and with organic molecules to produce carbon-centred radicals. Oxygen
is an important species in these reactions, not only by accepting electrons to
create superoxy radicals, but also by assisting charge separation, resulting in
carbon-centred radical formation61.
1.2.3 Factors affecting titania activity in polymers
Research into the photoactivity of titania in recent decades has uncovered aspects
of titania chemistry that influence its activity in polymeric materials49,54,56,62.
Among other factors, these include particle size63, crystalline structure64, phase
composition65-67, surface area68,69, nature and concentration of lattice defects70,
surface hydroxyl groups71-73, and impurities74,75.
Possibly the single biggest factor affecting the activity of titania is particle size.
This is due to several reasons. The larger a titania particle, the relatively less
surface area is available for reaction with oxygen, reducing its effectiveness.
Also, larger particle size makes it easier to achieve hole – electron recombination.
Revisiting Figure 1-4, it can be seen that the more TiO2 molecules in a particle,
the closer the gap between the HOMO and the LUMO. Thus ideally, to
maximize the photoactive effect of titania in polymer systems for the purpose of
degradation, it is more desirable to have nanoscale particles with good dispersion.
The different effects of micro- and nano-sized titania particles on the degradation
of cumene as a model for polymer degradation was investigated by Allen et al.
and reported in two papers76,77. A clear difference was noted between the activity
of the two, with nano-sized particles considered to be the more active.
Furthermore, nanoparticles were found to influence the degradation of a material
much earlier, starting at the manufacture of the polymer.
Another detrimental effect of large particle size on the photocatalytic properties
of titania is due to whitening; larger particles reflect UV light, which is
demonstrated by titanias’ use as a stabilizer when used with microscale particle
size.
35
In 2001 Cho and Choi78 established the difference between photolytic and
photocatalytic degradation of PVC containing Degussa P25 TiO2. SEM images
revealed that photocatalytic degradation occurred more rapidly in an area
localized around the TiO2 particles, and photolytic degradation occurred at
evenly distributed centers throughout the PVC matrix. Dispersion of the titania
particles was a problem, and micrometer-sized agglomerates were reported to
reduce significantly the photosensitizing effect of TiO2.
Particle size in a polymeric system is determined by the manufacture of the
titania powder, and by agglomeration of the titania once it is mixed with molten
polymer. The propensity of titania to agglomerate in polymers is well recognized,
and researchers have looked to modify the surface of titania particles to achieve
good particle separation, and thus small particle size once the polymer product is
finalised.
1.2.4 Surface chemistry of titania
The ease of molecule adsorption onto the surface of titania catalyst particles has
a significant effect on the activity of titania. The surface of titania particles is
highly heterogeneous, with anatase containing Lewis acid sites and several
different forms of hydroxyl groups79. Anatase also has larger crystal faces than
rutile, and is generally considered more suitable as a catalyst as it demonstrates
higher adsorptivity80. In addition to hydroxyl groups on the surface of titania
particles, defects must be present to trap oxygen to allow catalytic degradation of
organic molecules81.
Degussa P25 titania is one of the most efficient and extensively used commercial
photocatalysts available due to its high surface area, high photoactivity and
minimal impurities81. The high activity of Degussa P25 is thought to be due to a
more positive conduction band potential in rutile compared to anatase. This
allows photogenerated electrons to pass from anatase to rutile, preventing
recombination within the anatase. The removal of electrons by rutile is similar to
the action of dopants, discussed below. Degussa P25 titania is formed as
36
nanoparticles, and thus has much greater surface area than most other forms of
titania. This allows for significantly increased molecule adsorption rates.
The importance of water molecules trapped onto the surface of the titania
particles is still open for debate82. Although it has been established that hydroxyl
radicals form the main oxidising species during titania photocatalysis, there are
several suggested pathways to creating such a species, which may or may not
necessitate water. The evidence in the literature appears incomplete in either
case, and more study is required before the role of water can be properly
determined.
1.2.5 Surface modification of titania
Unmodified nano-titania does not demonstrate good dispersion characteristics in
polymeric systems due to titania-polymer interactions9. To improve the
dispersion of titania nanoparticles researchers have modified the surface of
titania by grafting polymers with better polymer miscibility properties onto the
particle surface83-87.
This effectiveness of this approach is determined by the polarity of the functional
groups present on the polymer chain; the grafting of a short-chain polymer with
polar groups onto the surface of titania nanoparticles allows for improved
particle dispersion in a polymer with polar functional groups87. An example of a
silicon grafting agent ‘WD-70’ is given in Figure 1-5, used by Zan et al.85 in
2004 to improve the dispersion of nano-titania in polystyrene. Dispersion of the
titania was reported to be successful, and the polystyrene with grafted TiO2
degraded at a faster rate than pure polystyrene. The titania particles used in this
case were laboratory-prepared by Zan et al. with a size range of 70-100nm.
37
Figure 1-5 The structure of WD-70 which is grafted onto the surface of TiO2 particles by Zan et al.84 to improve titania dispersity in polystyrene.
Although there is some literature describing the grafting of short-chain polymers
to provide electrostatic interactions with a polymer such as those mentioned,
there is considerably less literature on surface modified titania for application as
a photocatalyst in polyethylene.. The lack of scientific literature is presumably
due to the difficulty of identifying an appropriate polymer for use as a grafting
agent. The absence of functional groups in polyethylene chains disqualifies the
use of grafting agents such as that presented in Figure 1-5.
In 1992 Allen et al.7 conducted a study on low density polyethylene (LDPE)
films containing nine different types of titanium dioxide pigments (coated and
uncoated, rutile and anatase titania). The nature of the coating was not divulged,
except to say that it was organic in nature. It is presumed that the coating was
designed to improve dispersion. Thermooxidative and photooxidative
degradation was compared, and it was found that all the titania pigments acted as
photosensitisers, with uncoated anatase and uncoated fine crystal rutile types
being the most active. The significance of this result lies in that nature of the
titanium dioxide. Even though the titania was pigment grade, manufactured to
behave as a photostabiliser in other applications, it behaved as a prodegradant in
LDPE. Additionally, surface modification to promote better dispersion had a
negative effect on the photocatalytic properties of the titania particles.
It was also found that during thermooxidation, the rate of carbonyl formation as
followed by IR techniques was less dependent on the nature of the particle as the
temperature was increased. It was found that the role of titania particles was
dependant on crystal size and structure, as well as the nature of surface treatment.
The photocatalytic activity of Allen’s titania particles increased with decreasing
particle size.
38
In 2006 Zhao et al.88 reported enhanced photooxidation of polyethylene
containing 0.1 - 1% TiO2. Degussa P25 (a mixture of 75% anatase and 25% rutile
titania) was used. The titania did not appear to be modified, and unfortunately the
size of any agglomerated particles was not reported. SEM images showed
creation of cavities in the polymer following irradiation, which were attributed to
the evolution of small volatiles.
1.2.6 Doping
Doping of titania involves the addition of an element or compound to titania
which can enhance the quantum efficiency of electron-hole separation. Elements
used in doping depend on the application, and are often used to enhance the
photoactivity of titania under visible light89-99.
Vanadium doped titania photocatalysts 98,100-104 are a good example of the
mechanism of doped titania photocatalysis. Martin et al.98 reported on the
mechanism of quantum sized vanadium doped titania nanoparticles in the
oxidation of the dye 4-chlorophenol. The 1-5 nm sized particles synthesized in
the laboratory aggregated to 50 µm size particles; however each crystallite was
reported to be electronically isolated.
V
O O
O O
V
O O
O O
H
V
O OH
O O
Titania particle surface
-
Figure 1-6 Mechanism for charge separation of vanadium doped titania.
Figure 1-6 shows the mechanism given by Martin et al.98 for charge separation
by V(V) (VO2+) doped titania. V(V) attached to the surface of a titania particle
can abstract a hydrogen from an organic molecule. When applied to polymer
degradation, a carbon centered radical formed by hydrogen abstraction can lead
to hydroperoxide formation, according to mechanisms discussed earlier.
39
In this thesis the photocatalytic effect of titania doped with antinomy is reported.
There is apparently no literature discussing such a system, however it is assumed
that antinomy is present to assist in charge-carrier separation similarly to the
example of vanadium given here.
1.2.7 Effect of UVA vs. UVC radiation on polymer – TiO2 systems
Allen has contributed greatly to the knowledge of the activity of titania in
polymeric materials7,11,62,76,77,105-113. However, in 1996 Allen and Katami110 found
an unusual result when they conducted a comparison of aging conditions on the
degradation of linear low density polyethylene films containing titania. Using a
narrow band 365 nm radiation source all types of titania studied, except heavily
coated rutile particles, acted as photosensitisers. However, with narrow band 254
nm irradiation all types of titania acted as UV screeners and stabilised the
polymer. This was explained by the penetration depth of the higher-energy
wavelength light. The 365 nm radiation was said to be absorbed closer to the
surface of the titania particles. However, the 254 nm exhibited "deep crystal
lattice penetration", and the surface functional groups of the titania particles were
not activated. It was concluded that the thermal and photo generation of active
carriers on the surface of pigment particles strongly influences the photoactivity
of titania particles.
Although Allen has reported several times the photostabilising effect of titania
under 254 nm radiation76,77,108,110,114, it appears that there has been limited
supporting experimental evidence. Only two papers77,108 show different
experiments that support the theory of deep crystal lattice penetration. Seeing
that many different factors, such as particle size, temperature, etc., affect the role
that titania plays in polymer degradation, it is reasonable to suggest that more
experimental evidence is necessary to corroborate this theory.
In 2006 Zan et al.115 developed a low density polyethylene film incorporating
titania nanoparticles via a melt-blending technique. Degussa P25 titania was
employed, and an irradiation source of 254 nm was chosen. The result of
40
photocatalytic degradation was investigated via several established analytical
techniques including IR and SEM imaging. The titania was found to act as a
photosensitiser, and the polymer weight loss following degradation was found to
be much higher than for pure polyethylene degradation.
This result can be considered significant, as Allen had found that titania acted as
a mild photostabiliser under 254 nm radiation. This was attributed to the depth of
penetration of incident radiation into the crystal lattice. However, Zan et al.
found that degradation occurred at 254 nm faster than in sunlight, which was
accelerated compared to pure polyethylene. Both researchers used similar
materials under comparable conditions. The difference, however, was the titania
used. Zan et al. used Degussa P25, a nano sized particle, while Allen used a
laboratory prepared experimental grade titania particle.
As stated by Allen et al.105 and Mills and Hunte49, the manufacture history plays
a large a role in the photoactivity of the titania particles. It is possible that the
particle size of Allen’s titania affected the degradation kinetics to a greater extent
than the wavelength of the incident radiation. In all likelihood it may have been a
combination of both factors, as the 254 nm radiation penetrated deeper than 300+
nm light into the crystal lattice structure, and the larger particle size further
restricted the activation of surface groups.
1.2.8 Summary of sections 1.1 and 1.2
The importance of polyethylene as a commodity plastic has resulted in the
thermooxidation and photooxidation degradation pathways being well studied for
many decades. Polyethylene degradation is driven by radical reactions,
eventually giving rise to chain scission, crosslinking of the polymer chains, and
the development of oxygenated functional groups.
It has been revealed that degradation of polymers in the solid state is not a
homogenous process. Oxidation occurs at points in the polymer, such as metal
catalyst residues, and spreads to other areas of the bulk, eventually resulting in
cracks and the loss of mechanical properties. This phenomenon can be exploited
41
by the addition of prodegradant materials to enhance the oxidation process.
Among potential prodegradant materials titania has achieved noteworthy status,
being both highly photoactive and economically viable for commercial
applications.
Producing a polyethylene film with controllable degradation properties is not as
straightforward as merely adding titania powder to a melt. This is due to the
many factors affecting the activity of titania in polymers, including particle
aggregation, particle size, surface properties, crystal phase, etc.. Researchers
have attempted to enhance the activity of titania by surface modification, and
addition of dopants to improve the separation of electrons and holes.
Although the activity of titania under UVA irradiation is well characterised, there
is some contention regarding the effect of UVC on a polymer/titania composite
material, especially where titania particles are between pigment and nano-sized.
Additionally, although these materials have been well studied under conditions
of constant irradiation, it is unknown whether degradation reactions initiated by
titania with UV irradiation will continue to occur rapidly in the dark, such as
might be expected in the lifecycle of waste plastic packaging, or shopping bags.
1.3 Polymer degradation characterisation techniques
There are various approaches employed by polymer researchers to explore,
visualize and quantify physical and chemical aspects of polymer degradation
occurring in the solid state. The applicability of a characterization technique to a
given problem is determined by the nature of the information that is sought. In
order to determine degradation pathways for, say, the design of new antioxidants,
FTIR and NMR provide detailed information regarding molecular structure of
oxidation products. Alternatively tensile strength testing is better suited to
quantifying the physical effects of oxidation on the mechanical properties of a
polymer film.
Areas of interest can be broadly grouped into three categories; physical, or bulk,
effects; surface characteristics; and molecular structural information. A
42
combination of these aspects of polymer degradation provides an overall
description of the entire degradation process.
1.3.1 Characterization of the bulk via physical tests
1.3.1.1 Elongation at break
Although bulk testing is more commonly employed by commercial assessors in
industry116, it is also a powerful tool for researchers in polymer degradation, due
to its ability to detect early signs of degradation117. Elongation at break tests
involve measuring the strain required to break a piece of film of predetermined
dimensions. Madfa et al.118 demonstrated the applicability of elongation at break
testing to LDPE films that had been exposed to natural weathering. Tests showed
that cross-linking induced by radiation absorption became significant after just
one week of weathering. The materials used in this experiment were
commercially produced; however they wholly degraded after just 4 months of
weathering.
Roy et al.55,119-123 employs elongation at break tests among several other
characterization techniques when investigating accelerated polyethylene
degradation. His results demonstrate not only decreased elongation at break at
early stages of oxidation, but also that increasing the concentration of
prodegradant (typically cobalt stearate) have a negative effect on the mechanical
properties.
1.3.2 Surface Characterisation
For solid state polymers exposed to oxidative environments such as natural
weathering or accelerated conditions it is expected that degradation will occur
initially at the surface124. There are numerous methods in existence for
examining the surface of degraded polymers, of which the most common
technique is scanning electron microscopy (SEM).
SEM provides physical information from the surface of degraded polymers, such
the appearance of cracks and voids125,126, and the influence of domains in
43
polymer blends 127. It is particularly useful in characterising the surface of
degraded polymers containing transition metal photocatalysts due to its inherent
image contrast between metals and organic compounds. Thus agglomeration and
distribution of transition metal photocatalysts128 and the immediate environment
surrounding these photocatalysts can be clearly observed.
Zhao et al.88 provided some excellent SEM images demonstrating the appearance
of cavities around titania particles in polyethylene following UV irradiation
(Figure 1-7). In these images the titania particles are difficult to detect, however
as demonstrated later in this thesis backscattered images show the location of
titania particles quite clearly.
Figure 1-7 SEM images obtained by Zhao et al.88 of appearance of cavities in photodegraded polyethylene containing TiO2. © 2007 Elsevier Science.
One of the potential drawbacks to SEM imaging is the loss of sample, as the
sample must be coated with carbon or gold prior to examination. However the
technique does not require a large sample, and as the sample is already degraded,
sample loss does not generally constitute a significant issue when characterizing
the surface of degraded polyolefins with SEM.
1.3.3 Chemical Characterisation
1.3.3.1 Oxygen uptake
As stated earlier, oxygen uptake measurements are used to determine kinetic
information regarding degradation reactions. Zeynalov and Allen112,113,129-131
used oxygen uptake measurements to examine the effects of antioxidants and
44
prodegradants on the degradation kinetics polymers, using the model compound
cumene. Oxygen uptake measurements were used to obtain information such as
the rate of radical scavenging by the antioxidants, nature of the rate dependence
on the concentration of inhibitors and the activity of different phases of titania
nanoparticles.
Although oxygen uptake measurements are useful for investigation of the
kinetics of oxidative degradation, other techniques provide more accurate
chemical information regarding the nature of degradation products, which can
then be used to determine degradation pathways.
1.3.3.2 Nuclear Magnetic Resonance
Nuclear Magnetic Resonance (NMR) allows the polymer chemist to investigate
the properties of a degraded polymer sample on a molecular scale, yielding
information regarding carbon hybridization132, chain scission and cross-linking
phenomena133,134, polymer chain mobility135-137 and morphology138. NMR can be
used at high field frequency to detect specific molecular changes, and at low
power can detect degradation related changes in the bulk sample139.
NMR is not as well represented in the literature as other characterisation tools,
such as FTIR, for polyolefin degradation. There are several possible reasons for
this, including the lack of spatial information, the relative difficulty in obtaining
solid state NMR spectra compared with other characterisation techniques, and
the nature of the information obtained. Additionally, NMR is a destructive
technique, making it difficult to obtain information from bulky samples.
1.3.3.3 Mid-Infrared Spectroscopic Techniques
Vibrational spectroscopic techniques have been used for many decades in the
characterisation of polymers, and the characterisation of polymer degradation
products. The most important of these techniques has been mid-Infrared (IR)
spectroscopy, for its ease of application to many polymeric materials, and the
type and quality of information obtained. Mid-infrared instruments usually
45
examine the wavelength range from 2.5 µm to ~17 µm (4000 cm-1 to ~600 cm-1).
Of IR techniques, transmission has been in use the longest.
Transmission IR
Many journals and books have been written discussing the application of
transmission FTIR for characterisation of polymeric materials, and of
degradation related products. In the 1950s Rugg et al. investigated polyethylene
structure24 and oxidation products25 using IR spectroscopy. Much of the work
performed by Rugg et al. is still used in the literature today, demonstrating the
reliability and reproducibility of transmission IR. Transmission IR can also be
used to quantitatively examine oxidation products of polyolefins140, although in
more recent times emission FTIR has also been demonstrated to be appropriate to
the task141.
Transmission IR can pose some challenges in the study of thin films. If the film
in question is too thick, then over-absorption can occur, distorting the spectrum
by cutting off the top of the CH2 absorption peaks of a polyolefin. At an
absorbance value of 2 or higher less than 1% transmission occurs and detectors
become increasing unreliable. However, when examining oxidation products this
is not usually a problem, as over-absorption can actually be used to enlarge weak
absorptions that might otherwise be difficult to observe clearly. This technique
can be used to examine the carbonyl and fingerprint regions at early stages of
oxidation. If over absorption is unwanted it can often be overcome by cutting off
a thinner section of the polymer for examination.
More challenging than over-absorption of a film is the issue of interference
fringes. An interference fringe is described as a sinusoidal intensity variation due
to interference of radiation that undergoes multiple reflection between two flat
and parallel surfaces142. Figure 1-8 shows an IR transmission spectrum with an
interference fringe (upper plot). It can be seen in Figure 1-9 that light reflected at
the lower n1/n2 interface may reflect again inside the polymer film on the
opposite interface, and upon passing through the sample will interfere
constructively and destructively in a sinusoidal fashion (depending on the
46
wavelength) with the transmitted beam, resulting in the superimposition of a sine
wave on the final spectrum.
Figure 1-8 Avoidance of interference fringes by using polarised light at the Brewster angle143.
Incident IR beam
n2 (air)
n1 (polymer f ilm)Reflection at interface
Sinusoidally interfering waves
n2 (air)
Figure 1-9 Schematic demonstrating the interference fringe phenomenon, which causes the superimposition of a sinusoidal wave over a transmission IR spectrum.
47
For interference fringes to occur in the IR spectra of polymer films the surfaces
must be flat, parallel, and spaced the correct distance apart (film thickness
between 5 µm and 2.5 mm)144. Although the intensity and position of the
polymer absorption bands remain unchanged in the spectrum, interference
fringes can complicate the spectrum by making small peaks difficult to interpret
as they can lie of the shoulder of a sine wave. Additionally, spectral comparison
methods such as overlapping, spectral subtraction and curve fitting become very
difficult to achieve reliably. When investigating early stages of oxidation with IR
transmission methods interference fringes can pose more than a mere nuisance.
There are some methods to alleviate the problem of interference fringes in
transmission spectra. One of the most effective methods is use of the Brewster
angle, suggested by Harrick143 in 1976. This involves orientation of the polymer
film at the Brewster angle with respect to the incident light. A demonstration of
this is included in Figure 1-8, where a transmission spectrum of polyester film
has had interference fringes avoided by use of the Brewster angle and polarised
light.
This method would be expected to be less effective in studying commercial
polyethylene films due to crystal orientation of the polyethylene chains during
film blowing145, which would yield spectral results dependant on film orientation
during sampling146,147. Other methods for reducing fringes include scratching the
surface of the polymer with steel wool to create a rough surface, and clamping
the film between two mid-IR transparent windows144. However neither of these
techniques is applicable to the study of degraded polyolefin films.
The spacing of peaks in interference fringes can be used to determine the
thickness of the film under investigation, if the refractive index of the material is
known144. The thickness is calculated by counting the number of waves over a
wavenumber range, according to:
48
t =∆n
2(ν2−ν1)η Equation 1
Where: t = thickness
∆n = number of waves in spectral range
ν2−ν1 = spectral range (wavenumbers)
η = refractive index of film
ATR/FTIR
Attenuated Total Reflectance FTIR (ATR/FTIR) measures the near surface
layer144, and has demonstrated suitability to the investigation of surface
degradation of polyolefin films148-150. A comparison of the vibrational
spectroscopic techniques transmission IR, emission IR and ATR/FTIR by Delor
et al.151 concluded that ATR/FTIR is a reliable method for the study of the
evolution of degradation products of elastomers by infrared spectroscopy.
ATR/FTIR is an internal reflection technique144 that uses the optical principle of
light passing through a medium of high refractive index internally reflecting
when impinging on a surface of lower refractive index at an angle less than the
‘critical angle’. The critical angle, θc, can be described as the threshold angle
below which light will internally reflect at a boundary between two media, and is
given by:
sin θc =η2η1
Equation 2 Where: η2 = lower refractive index
η1 = higher refractive index
In ATR/FTIR internal reflection is achieved by placing a crystal that is
transparent to mid-IR, and possessing a high refractive index, in optical contact
with the sample (Figure 1-10). When light is internally reflected at the
crystal/sample boundary, evanescent mid-IR waves can be absorbed by the
49
samples molecules. The resulting signal is Fourier transformed to produce an
infrared spectrum. ATR/FTIR crystals can be multi-bounce or single bounce,
depending on the size of sampling area sought, and the strength of the signal
required.
Internal ref lection element (IRE)
Incident mid-IR Exiting mid-IR
Sample Figure 1-10 Schematic of multi-bounce ATR/FTIR.
As ATR/FTIR is a semi-surface technique, it is essential to know the depth of
sample being described by the spectrum obtained. The depth of penetration of
light is dependant the relative difference in refractive index between the IRE and
the sample, and on the wavelength of the radiation. As a mid-infrared spectrum is
obtained over a range of wavelengths, the depth of penetration will vary over the
spectrum, according to the Harrick equation:
dp =λ
2πn1(sin2 θ - n2 )1/221
(12)
Where: dp = depth of penetration
λ = wavelength (nm)
n1 = refractive index of IRE
n21 = ratio of refractive index of sample/objective
θ = angle of incidence
50
This raises some important factors that must be considered when viewing an
ATR/FTIR spectrum. Firstly, higher energy, shorter wavelength light (high-
wavenumber end of the spectrum) penetrates less into the sample, giving a
weaker absorbance than the low wavenumber end of the spectrum. Software used
for the manipulation of spectra usually includes an ATR correction formula that
will balance this disparity in absorption intensity caused by the difference in
penetration depth with wavenumber. Also, the depth of penetration is dependent
on the refractive index of the material and the IRE. Naturally, the refractive
index of the sample cannot be changed, however by using an IRE with a
relatively high refractive index (germanium for example has a refractive index of
4 and is transparent to mid-IR), the depth of penetration can be lowered
significantly151,152.
In this thesis the low wavenumber end of the spectrum (down from
approximately 1800 cm-1) is the region of greatest interest, as it contains
chemical information relating to degradation. ATR correction formulas change
the scaling by decreasing the relative absorption strength of bands in this region,
which is undesirable in this instance as this the region contains the most relevant
information. Especially at low levels of oxidation, ATR correction would cause
already difficult to detect changes to be more difficult to observe. Thus spectra
obtained using ATR/FTIR methods in this thesis have not undergone an ATR
correction, but are analysed in all cases as the raw data obtained when the spectra
were acquired.
1.3.3.4 Polyethylene absorptions in the mid-IR
Different types of polyethylene (high density, low density, branched, etc.)
display different absorption bands in the mid-IR153. Transmission IR spectra of
thick polyethylene specimens will exhibit over-absorption of the main C-H
vibrations, allowing weaker, skeletal vibrations to be more easily identified.
High density polyethylene (HDPE) contains less –CH3 groups than other forms
of polyethylene, and thus has a relatively weaker symmetrical bend C-CH3
absorption at 1378 cm-1. Additionally, HDPE has a higher vinyl unsaturation
51
content, absorbing at 910 and 990 cm-1. In contrast low density polyethylene
(LDPE) and in particular linear low density polyethylene (LLDPE) have a much
higher –CH3 content, and thus a stronger C–CH3 at 1378 cm-1. Due to the higher
numbers of short chain branches in LLDPE, there is a high content of vinylidene
(pendant methylene,
>CH=CH2), absorbing at 888 cm-1.
Crystallinity of the polymer also affects the mid-IR spectrum. A band at 1303
cm-1 increases with increasing amorphous content, while sharp absorptions at
1175 and 1050 cm-1 increasing with higher crystalline content.
1.3.3.5 Depth Profiling by ATR/FTIR
Transmission IR experiments can be performed by coupling an IR microscope to
an FTIR spectrometer154. Similarly, micro-ATR/FTIR experiments can be carried
out by using an ATR objective on the microscope155. An aperture can be
adjusted on a micro-ATR/FTIR spectrometer to measure an area on the sample
smaller than the size of the contact surface. Thus micro-ATR/FTIR allows for
improved lateral resolution, but at the cost of signal/noise ratio, as there is less
reflected light received by the detector.
Do et al.152 performed an experiment to deduce an optimum aperture size for the
measurement of some carbon-filled polymeric materials. It was found that the
minimum aperture setting, which allowed for a best possible compromise
between signal/noise ratio and collection time, was 40 µm. The authors152 then
add that the actual spatial resolution being achieved when using micro-
ATR/FTIR can be determined by dividing the aperture size by the refractive
index of the crystal. In this case it was found to be around 12 µm, which is
approximately the diffraction limit of infrared radiation.
Line-mapping by micro-ATR/FTIR was demonstrated to produce spectra that
could be used to obtain oxidation profiles of a cross-sectioned surface of a
polymer sample156. These profiles evidenced higher levels of degradation
towards the exposed edge of the rubber than compared with the centre. The
52
authors152 concluded that micro-ATR/FTIR was a technique suitable to the study
of polymeric materials, and good quality, low noise spectra could be obtained
using a silicon IRE in just a few minutes.
1.3.4 Achieving high lateral resolution
With respect to the heterogeneous degradation processes occurring in solid
polymeric materials (Section 1.1.2), there is a clear advantage to be able to view
degradation processes with respect to the dimensions of the material in question.
Additionally, the inclusion of natural (crystalline or amorphous regions) and
introduced (prodegradants, inhibitors) heterogeneities within polyolefins
heighten the need for spatial information. Lateral (commonly referred to as
spatial) resolution refers to the smallest distance at which two objects can be
distinguished, and recent developments in IR technology has seen lateral
resolution reduced to below the wavelength of IR light in an air medium.
Imaging ATR/FTIR is a technique that has achieved considerable success in
increasing the spatial resolution available to IR spectroscopists.
1.3.4.1 Imaging ATR/FTIR
There are two primary methods of image acquisition that are of particular interest
to polymer chemists157. When performing projection imaging, an area of interest
is selected and uniformly illuminated by a broad beam. Reflected or transmitted
radiation from the specimen is directed back via a system of optics to an arrayed
set of detectors, known as a focal plane array detector. Scanning imaging
involves moving the sample or detector such that different areas on the specimen
surface are sampled in a raster pattern. In both methods, the signal is received by
the detector and reconstructed to give chemical and spatial information.
Imaging detectors are typically constructed of an m x n array of detectors. An
optical signal impinges on the detector related to a point on the object, and if the
radiation is sufficiently large it will produce a current, I(t), that flows through a
load resistor, RL, and produces a voltage, v(t), according to:
53
v(t) = I(t)RL = [S(t) * h(t)]RRL
Equation 3
Where: S(t) = intensity envelope of the optical signal
h(t) = impulse response function of the detector
R = responsivity (ratio of photosignal to radiation
power incident on detector) specified in units of
amperes per watt
* = denotes convolution
Ideally, detectors should have a high signal-to-noise ratio (S/N), and a high
spectral response. S/N is often limited by the sensitivity of the detector, which
can be described in terms of the current produced per unit of incident radiation.
Signal received by the detector is considered noise if it does not originate at the
conjugate object point. Spectral response refers to the range of wavenumbers
over which the detector will produce useful information. It is often necessary to
sacrifice spectral response in order to increase the S/N of a focal plane array
detector. MCT (HgCdTe) detectors are commonly used in FTIR instruments to
improve the S/N.
MCT detectors offer greater sensitivity when using ATR/FTIR. FPA detectors
consist of a 2-dimensional square array of MCT detectors, usually in the order of
64 x 64, or 128 x 128 detectors. One of the drawbacks of using such a sensitive
photovoltaic detector is its ease of saturation – large amounts of incident
radiation quickly overwhelm the detector. Furthermore, in order to achieve
maximum spectral response with low levels of incident radiation, it must be used
at very low temperatures. This problem is resolved by operating the detector with
liquid nitrogen cooling.
High spatial resolution is important in imaging studies in order to examine as
small an area as possible. As spectroscopic measurements are determined using
photons, spatial coherence of the photons limits the spatial resolution. Spatial
coherence is distance below which the interference of the harmonic signal is
constructive. This is inversely related to the wavelength, according to:
54
K = 2π/λ Equation 4
Where: K = wave vector
λ = wavelength
In imaging spectrometers, pixel resolution plays a role in determining the overall
system resolution. One pixel measures a fixed width, and pixel resolution is
given as the length of an image in one direction divided by the number of pixels
in that direction.
In recent times, the use of µ-ATR/FTIR imaging has introduced some exciting
improvements to the achievable spatial resolution of IR imaging
spectrometers158-161. Chan and Kazarian published new findings in 2003,
reporting the achievement of spatial resolution of 3-4 µm using µ-ATR/FTIR
with a Ge IRE. This is a further improvement on a spatial resolution of 8 µm
achieved by Sommer et al.162 in 2001.
Chan and Kazarian used stringent criteria when determining lateral
resolution159,162. A chemically heterogeneous polymer sample, consisting of
poly(methyl methacrylate) (PMMA) patterned using electron beam lithography
and fixed on a silicon wafer, was examined using µ-ATR/FTIR. Two different
areas (in this case, clean PMMA vs lithographed PMMA) were considered to be
resolved when the spectra showed a 5-95% absorbance profile as a function of
distance, i.e. when the spectra changed from showing 5% of one component to
95%. This was achieved with a dimension limit of 4 µm. The change in
wavelength of light when passing through media of high refractive index is
responsible for such an achievement. The relationship between the refractive
index and wavelength of light is given as:
55
n =λ0
λn Equation 5
Where: n = refractive index
λ0 = wavelength of light in a vacuum
λn = wavelength of light in a medium of refractive index n
The wavelength and velocity of light both decrease with increasing refractive
index of the medium..The shorter wavelength of light when applying µ-
ATR/FTIR techniques with a Ge IRE (RI = 4.0) allows for much greater spatial
resolution than using, say, transmission IR, where the light passes through a
medium of air. The application of this principle allowed Chan and Kazarian159 to
obtain spatial resolution higher than had previously been reported.
1.3.4.2 Synchrotron radiation source
While a conventional FTIR microspectroscopy cannot use an aperture smaller
than approximately 20 µm, due mainly to S/N restrictions, a synchrotron light
source is powerful enough to achieve a reasonable signal with a much smaller
aperture. A synchrotron light source is some 300 times brighter than a light
source on a conventional IR spectrometer163. However, this translates to an
improvement of brightness of 3 orders of magnitude when light is projected
through a pinhole of 10 µm diameter, resulting in greatly improved S/N for MCT
detectors.
Despite the significant attenuation of the light source, spatial resolution of 6 µm
has been reported in the literature164. Many of the applications of synchrotron
FTIR studies are biological in nature, examining organic materials such as
tissues165 and fungi166, where improved spatial resolution has been beneficial,
and imaging ATR/FTIR was not practical.
Utilisation of a synchrotron light source for the investigation of polymer
degradation has not been popular with research scientists. Although little
information can be seen in the literature, an article has been by published by
56
Wetzel and Carter167 who examined the degradation of acrylic polymer
automotive coatings using a synchrotron light source in 1998. The coating was
cross sectioned and spectra obtained in 1 µm steps in transmission mode.
Although this paper demonstrated the applicability of microspectroscopy with a
synchrotron light source to obtain spectral information with high spatial
information it is uncertain whether information was obtained that could not have
been obtained via imaging ATR/FTIR.
Despite the improved spatial resolution of FTIR microspectroscopy using a
synchrotron light source over conventional FTIR microspectroscopy, the spatial
resolution obtainable is still slightly inferior to that of an imaging ATR/FTIR
system. However, as not all samples are suitable to examination by imaging
ATR/FTIR, there is a range of applications for both methods.
1.3.5 Characterisation techniques used in this thesis
There is a broad range of characterisation techniques that have some relevance to
the investigation of the degradation of polyethylene film containing titania.
However, some methods provide more pertinent data than others. In particular,
IR spectroscopy has a long history in this field, and the information regarding
degradation-related functional groups has been well studied. Furthermore IR
spectroscopy allows the researcher to link the products with a degradation
pathway.
Other chemical characterisation techniques are less useful for this study. Raman
spectroscopy provides better lateral resolution, however it is less sensitive to
oxygen containing functional groups compared to IR spectroscopy. Solid-state
NMR requires destruction of the sample, and the spectra are more difficult to
interpret with a broad range of oxidation products in low concentrations. X-ray
photoelectron spectrsocopy (XPS) examines the surface layer of the sample,
whereas when examining the films used in this thesis information was sought
from deeper than the surface of the material. ATR/FTIR penetrates up to 3
microns into the sample, which examines that part of the sample with the highest
concentration of oxidation products. Oxygen uptake measurements provide
57
information regarding the rate of oxidation; however this has been achievable
using IR spectroscopy. SEM and back-scattered SEM images provide
information regarding the dispersion of titania particles and the physical
environment around the particles, and was used for the purpose of
characterisation.
There are also mechanical methods of determining the extent of degradation of
polyethylene film, such as stress-at-break measurements. However, the film was
not required to degrade until it had lost its useful mechanical properties, but was
required to break up into fine particles to allow for microbiotic action to render
the polymer environmentally neutral. Thus all films were degraded until
embrittlement, such that they could no longer be held in their sample holders. In
this situation it was decided that mechanical measurements would not present
information pertinent to the investigation, and so were not conducted.
It was decided therefore that IR spectroscopy in various forms would be the main
characterisation tool in this study principally because of its non-destructive
nature as well as the high level of molecular structural information that it
provides.
1.4 Objectives
There is a strong demand from industry to develop an environmentally neutral
commodity plastic film with controllable degradation qualities for applications
such as shopping bags, packaging, agricultural film, etc.168. LLDPE is a natural
choice for such a material as it has proven applications in these fields, and its
degradation chemistry has been well studied in the literature23,26,27,50,56,120,169.
The cost of the final product is an important consideration, as this research is
being performed with a view to potential commercial applications. With this in
mind LLDPE film blown by Ciba containing commercially available types of
titania from different manufacturers has been chosen as the subject material.
Although there is published research regarding the effect of titania on the
degradation of polyethylene88,107,115,170,171, these studies concern mainly
58
polyethylene and/or titania manufactured in a research science laboratory, and
therefore these results may not be directly transferable to commercial
applications. The research undertaken in this thesis will be directly applicable to
the development of environmentally neutral films.
Titania particle size has been demonstrated to strongly affect the photosensitivity
of polymer-titania composite systems63. Although nanoparticles of titania are
more photoactive then pigment grade titania110, they have a greater tendency to
agglomerate9,63, reducing their effectiveness. The surface of titania particles has
been modified by researchers to encourage better dispersion83,84,86,87, however,
surface modification not only reduces the photoactivity87, but also increases the
cost of a commercial material.
To date there has not been a published study on the effects of UV irradiation or
thermal oxidation of the commercial films used in this thesis. For films used in
agriculture, it is important to know how long these films can be expected to last
when exposed to sunlight. Additionally titania from different manufacturers has
been used in the films, and it would be useful to establish a relative order of
photoactivity of these titanias in polyethylene. Therefore examining the
degradation under simulated solar irradiation will be performed extensively to
determine the activity of titania.
Although the topic is debated among prominent polymer degradation research
scientists39,40, the oxidative degradation of polyethylene is most likely radical
based35, whereby the formation of carbon centred radicals leads to
hydroperoxides, resulting in degradation products. Therefore by controlling the
radical population in a polyolefin one can control, to a greater or lesser extent,
the rate of degradation. This principle has been used in the development of
radical scavenging antioxidants36,43, used to prolong the lifetime of a polyolefin.
When a titania particle absorbs UV radiation an electron/hole pair is created9,49.
If they do not recombine, it is possible for these species to migrate to a nearby
organic molecule, creating a carbon centred radical9,49. In a polyolefin such a
radical species can then react with oxygen to form a hydroperoxide macroradical,
59
which can subsequently infect the polymer10. From such infection sites in a
polymer degradation can be seen to spread throughout the bulk13.
The main objective of this thesis is to exploit this phenomenon to control the
degradation of polyethylene-titania composite material, even in the dark. Radical
species will be initiated in polyethylene-titania film via exposure to UV pre-
irradiation before being subjected to accelerated degradation conditions. It is
anticipated that oxidation will spread from these sites of infection, resulting in
degradation of the bulk polymer film. The concept of pre-irradiation with UV to
control subsequent degradation of polyolefin film in the dark is a novel approach,
not yet seen in scientific literature. In order for the final film to have a viable
commercial application, it is desired that the polymer film should break down
completely in the dark following pre-irradiation after approximately 6 months.
Although a good deal is known about the degradation of LLDPE, there is little
information regarding the chemistry of the environment surrounding a titania
particle during degradation. If more fundamental knowledge regarding the
degradation pathways and chemical species surrounding the particle was
available it is hoped that this could lead to refining the composite material to
create one with strictly controllable properties. State of the art IR techniques
allowing improved lateral resolution will be assessed to determine their
applicability to the study of the degradation around titania particles, especially in
the early stages of oxidation.
The parameters to be studied include pre-irradiation wavelength, exposure time,
and subsequent degradation conditions, which need to be understood in order to
further develop the technology. For example there are reports of titania irradiated
with UVC acting as a stabiliser110, while others have found it has a
photosensitising effect115. Such issues need to be clarified to determine the
optimum conditions for pre-irradiation. Also the difference in activity between
commercially available grades of titania needs to be investigated, along with the
effects of surface modification and doping.
60
This technology would have a clear application for commodities such as
shopping bag film, whereby careful pre-irradiation of a polyethylene-titania
composite film prior to use could result in shopping bags that will degrade in the
dark. Packaging film could be pre-irradiated so that once the expected useable
lifetime of the film has expired, it can be disposed of, and will degrade even
without further exposure to sunlight. It is hoped in the near future we will be
using films that degrade controllably according to the demands of the application
in an environmentally neutral manner.
61
62
Experimental
2.1 Ciba films investigation
10 samples of polyethylene film were obtained from Ciba AG (Postfach
Schwarzwaldallee 215, Basel, Switzerland). The sheets ranged between 22 and
27 µm in thickness and were composed of Dowlex LLDPE. Dowlex is described
on the Dow Plastics company website172:
“Next Generation DOWLEX* NG 5056 E and 5056.01 E are ethylene octene-1
copolymers specifically designed for use in blown film applications requiring the
finished film to show high impact strength and exceptional optical properties, as
well as good retention of properties at low temperatures. Typical applications of
use include lamination, bag-in-box liners and form-fill-seal packaging of frozen
vegetables and liquid foods. DOWLEX NG 5056.01 E is the slip and anti-block
version of the resin.” Dowlex 5056 has a melt index of 1.1 and a density of
0.919.
Dowlex 5056 contains the synergistic stabilisers Irgafos 168 (phosphate), and
hindered phenolic type stabilisers Irganox 10176, Irganox 1010 and Irganox
1330. The stabilisers are present in concentrations of less than 0.5% w/w.
Nine of the polyethylene samples contained titania, and one control without
titania. A list describing the titania in the LLDPE film was supplied by Ciba to
accompany the samples and is in a modified version is provided in Table 1.
63
Table 1 List of types of titania in LLDPE obtained from Ciba.
Titania Titania Description Average Particle Size
(nm)
Loading (%)
Degussa P25 approx 75% anatase, 25% rutile, no surface
modification 25-35 1
Degussa P25 approx 75% anatase, 25% rutile, no surface
modification 25-35 3
Kronos 1002 100% anatase, apparently no surface
modification 20-200 1
Kronos 1002 100% anatase, apparently no surface
modification 20-200 3
Huntsman tioxide A-HR
Organic coating; micronised 100% anatase, water
dispersible 150 3
Huntsman tioxide A-HRF
Organic coating; micronised 100%
anatase, dispersible in organic systems
n/a 3 Sachtleben
Hombitan LW-S-U
Anatase microcrystal with an antimony-doped
crystal lattice 30 3
Sachtleben Hombitan LW-S-
U-HD
Organic coating on anatase microcrystal
30 3
Sachtleben Hombitan LW-S-
12
Organic coating on anatase microcrystal
31 3 Sachtleben
Hombitan LC-S Aluminium and organic
coating on anatase microcrystal
32 3 Control nil n/a 0
64
2.2 Accelerated aging of samples
The samples were subjected to a multilevel factorial design experiment set up to
concurrently investigate several variables. The phenomena under investigation
were the effects of:
• Pre-irradiation of the samples with UVC or UVA light prior to
aging
• Pre-irradiation exposure times of 0 s, 60 s, 3 hrs or 24 hrs
• Subsequent aging in either a weatherometer or an oven
This created 176 unique samples (see Table 2- Table 12, pg
75), which were arranged in random order to minimize
systematic errors. Squares of polyethylene were cut from the
sheets and held inside 35 mm photographic slide mounts to
create individual samples (left). As the number of samples
that could be processed at one time was limited by the
available space in the weatherometer, the samples were
processed in batches of 25. Once the samples aged in the
weatherometer had all achieved embrittlement, the next
batch of samples was processed. Samples that were aged in
the oven were added to the same oven on different shelves.
Figure 2-1 LLDPE film in projector slide sample holder
Before samples were subjected to accelerated aging conditions in some cases
they were pre-irradiated. Pre-irradiation is a key concept in this thesis, and
involved exposing the films to a measured dose of UV irradiation prior to
accelerated aging. UVA pre-irradiation was conducted in a Q-UV aging cabinet,
incorporating a battery of eight 40 W Q-UV-A lamps, at a distance of 5 cm from
the samples (dose rate ~1,200 W/m2). The peak emission was at 340 nm with a
cut-off at 295 nm. All pre-irradiation was conducted at ambient atmospheric
conditions.
UVC pre-irradiation was conducted using 2 x 60 W low-pressure mercury
vapour lamps with single wavelength 254 nm emission, purchased from Heraeus.
65
The system power was approximately 50 W/m at the irradiation platform,
including radiation from a parabolic reflector for collection of stray UV light.
The spectral emission of a low-pressure mercury vapour lamp is a line spectrum
with approximately 90 % of its output at 254 nm.
2
Weatherometer aging was conducted in a Heraeus Suntest CPS+
Weatherometer™ device operating at an irradiation level of approximately
765 W/m2 at the plane of the samples. Temperature ranged between 35 and 45 ºC,
with ambient humidity. Air was drawn from outside the weatherometer by an
internal fan and blown over the samples. The weatherometer was set to a cycle of
72 hours irradiation time, after which the samples were removed from the
weatherometer and ATR/FTIR spectra were obtained. Non-embrittled samples
were returned to the weatherometer subsequent to measurement for another 3 day
cycle. Embrittled samples were removed from the population.
Oven aging was conducted in a Contherm Digital Series Oven™ set to 50 ºC
under atmospheric conditions. The oven was opened weekly, allowing fresh air
inside. Oven aged samples were removed weekly or fortnightly for the first 3
months, and then less frequently for the remainder of the aging time for
ATR/FTIR analysis. The samples were returned to the oven immediately
following analysis.
The temperature of 50 ºC was chosen for accelerated thermooxidation, which is
10 - 20 ºC lower than others reporting in the literature for similar films4,173,174.
The choice of temperature reflected the attempt to best mimic natural conditions
within the time available. Additionally, is has been reported that with increasing
temperature the nature of titania particles (crystal phase, surface modification,
particle size, etc.) has lesser influence on the rate of oxidation7. As the effects of
different forms of titania were under investigation, lower temperatures were
more appropriate.
Samples were considered to have embrittled when they either fractured and
developed tears during the aging process, or when the film was easily punctured
and the material tore easily under the application of light pressure from a
66
relatively blunt object. All weatherometer aged samples were aged to
embrittlement, while all oven aged samples were aged for a minimum of 200
days, and in some cases up to 400 or more.
2.3 Mid-IR spectroscopy
ATR/FTIR was performed on a Nicolet Nexus 870™ spectrometer using a Smart
Endurance macro diamond ATR crystal. 64 scans were co-added at 4 cm-1
resolution. Spectra were manipulated using Grams32 AI software. The spectra
were not ATR corrected.
The carbonyl index of the samples was recorded for each spectrum. The carbonyl
index was obtained by measuring the ratio of the area under the carbonyl peak
(between 1705 cm-1 to 1735 cm-1) to the area under the CH2 deformation peak
(1460 cm-1 to 1475 cm-1).
Multivariate analysis was performed using Solo™, a standalone version of the
PLS-toolbox add-on designed for Matlab by Eigenvector. Outliers were
identified by abnormal Hotelling T2 or Q residuals. In all cases spectra identified
as outliers were examined first to determine that they were true outliers
(abnormal baseline, additive absorptions, etc.). If an abnormality was found, the
spectrum was labeled an outlier and not included in the analysis. If the spectrum
appeared ‘normal’ in all other respects, it was generally included. In general very
few apparent outliers were identified as outliers.
The PCA program in Solo™ was used for all multivariate analysis work. The
spectra were normalized by setting the area to the same value (Area = 1),
followed by mean centering. The PCA model was cross-validated by a ‘leave one
out’ method. Typically 3-5 PCs were used to create a model. Data handling is
described in more detail in Section 4.2.
67
2.4 Imaging IR Spectroscopy
Data were collected using a Varian FT-IR imaging system. The system consists
of a rapid scan Varian 3100 FT-IR spectrometer, a Varian 600 UMA FT-IR
microscope equipped with an ATR objective, and a 32 x 32 liquid nitrogen
cooled mercury cadmium telluride (MCT) focal plane array detector. A
germanium slide-on crystal was used in the ATR objective. Spectra were
processed and images created using Varian Resolution Pro software.
Topas® was used as the polymeric material for solvent casting degradation
experiments covered in Chapter 5. Topas is an ethylene/norbornene copolymer
(Figure 2-2), which is easier to dissolve in solvents than polyethyelene.
xy
Figure 2-2 Molecular structure of Topas®.
Five grams of Topas was dissolved in 50 ml of cyclohexane on a hot plate stirrer
set to 40 ºC and left to dissolve overnight with stirring. Droplets of 2.5 µL of the
polymer/cyclohexane solutions were pipetted onto glass slides in triplicate and
the cyclohexane evaporated off under a fume hood. A glass slide supported
control droplets, not containing titania. A second glass slide supported Topas that
had Degussa P25 titania powder mixed with the Topas solution at approximately
5% loading. The third slide supported droplets onto which Degussa P25 titania
powder had been dusted over the surface before the solvent was evaporated in
the fume hood.
The droplets on the slide were irradiated with UVC described in Section 2.2.
ATR/FTIR spectra were obtained for the droplets on the 4 slides hourly over 4
hours. It was found that the slide supporting Topas with titania deposited on the
surface demonstrated the greatest carbonyl intensity, and it was concluded that
this method would be used for the following step of the imaging ATR/FTIR
experimental.
68
Figure 2-3 ATR Ge crystal slide assembly.
Using the same solution of 5 g Topas in 50 ml cyclohexane, 2.5 µl was deposited
onto the IRE surface of the slide-out germanium ATR/FTIR crystal objective
(Figure 2-3). Degussa P25 titania was deposited onto the wet surface, and the
slide-out crystal assembly with the solvent-cast polymer was left to dry under a
heated vacuum overnight. The assembly was then irradiated with UVC, and
images collected every hour, up till 8 hours of irradiation. Spectra were acquired
at 16 cm-1 resolution with 1064 scans co-added, over a range of 4000–850 cm-1
2.5 Synchrotron experimental
Spectra were acquired at the Australian Synchrotron (Clayton, Victoria) on the
IR beamline175,176 using a Bruker V80v spectrometer with Bruker Hyperion 2000
infrared microscope in micro-transmission mode. An aperture of 10 x 10 µm was
used. Spectra were examined using OPUS 6.5 software.
A Perspex box was built around the microscope stage allow purging with
nitrogen gas (Figure 2-4). The box had a door at the front to access the stage.
Once the doors had been closed they were left shut for 10 minutes to allow for
adequate purging.
69
Perspex purge box
DoorsStage controls
Synchrotron light source
Figure 2-4 Bruker FTIR microscope with Perspex purge box at the Australian Synchrotron.
UVA from an Omnicure® 2000 high pressure 2000 W mercury lamp emitting at
300-500 nm (unfiltered) with a flexible fiber optic cable was used inside the
Perspex box by threading the fiber optic cable through a similarly sized hole in
the Perspex. This was clamped in place 4 cm above the sample, at an
approximately 70 ºC angle from the horizontal, with the power level at 100.
(Previous proof of concept experiments not published in this thesis had
adequately demonstrated that this positioning and light strength provided enough
UVA in atmospheric conditions to degrade a titania containing LLDPE sample
such that a carbonyl absorption could be observed in the IR spectrum after about
10-15 minutes of irradiation).
The material investigated was a polyethylene film blown by members of the
CRC-P project at QUT. It was comprised of LLDPE obtained from DOW
Chemicals, with 1% polyisobutylene (for processability and to improve
tackiness) and 3% Degussa P25 titania. The film was 15 µm thick and clear. A
piece of LLDPE was cut from the sheet, and placed over a metal slide containing
a hole, to allow transmission experiments
70
Tape
Film
Sampled area
Metal plate
Hole in plate
UVA
Fiber optic
Figure 2-5 Schematic representing the experimental set up for synchrotron transmission IR experiments.
The LLDPE film was mapped in micro-transmission mode in a 2 x 3 pattern,
with an aperture of 10 x 10 µm. Two hundred and fifty six scans were co-added
for each spectrum, at a spectral resolution of 4 cm-1. A background was taken via
the hole in the metal plate to the side of the film edge before each spectrum.
Microscope magnif ication objectiveUVA fibre optic
Upper cassegrain
Sample on metalsample holder
Microscope stage
Perspex boxextension forstage movement
Figure 2-6 Photograph taken at the Australian Synchrotron showing the IR microscope and stage. The stage is currently in position for map acquisition, and is moved to bring the sample under the UVA probe when irradiating.
71
After taking a background the stage was moved to a predetermined position
using the microscope controlling software, and the doors to the Perspex box were
opened for 3 minutes to allow air into the box. The film was then exposed to 2
minutes of irradiation, with the doors open. The UVA lamp was switched off, the
doors were closed, and 10 minutes was allowed for the nitrogen purge to remove
most of the air inside the box. A 2x3 map was acquired, which took
approximately 10 minutes. This process was repeated until the film had
undergone a total of 30 minutes of irradiation. Each map took a total of 30
minutes to acquire, which includes the time taken for purging, sample irradiation,
etc. 30 minutes of irradiation resulted in a total experiment time of 8 hours.
2.6 Scanning electron microscopy
SEM images were acquired on a FEI Quanta 200 Environmental SEM equipped
with an Everhart-Thornley detector (ETD), while backscattered electron images
were acquired using a silicon strip detector (SSD). Elemental microanalysis was
conducted using energy dispersive x-ray analysis (EDAX) and all samples were
coated in carbon films.
72
Effect of UV pre-irradiation on the degradation of polyethylene
3.1 Introduction
As stated in Chapter 1, the aim of this project is to utilize UV pre-irradiation to
enhance and control the degradation of polyethylene, such that it will degrade in
the dark. Various types of titania from different manufactures have been
incorporated into polyethylene film, and have been subjected to a range of
degradation conditions. Factors influencing the photoactivity of titania particles
such as size distribution, agglomeration, crystal phase, modification, etc. have
been investigated with a view to understanding the effects of these factors on the
degradation pathway of polyethylene.
This chapter scrutinizes the large amount of data obtained during the
investigation, and examines and compares the photoactivity of different titania
particles
3.2 Physical characteristics of commercial titanias and general comments
3.2.1 Degussa P25
Polyethylene film containing 1% Degussa P25 was translucent, although slightly
hazy. With several layers of thickness the film appeared opaque and glossy
white, with a distinct sheen. If the film was held up to the light numerous small
imperfections were seen included in the sheet. These imperfections were
approximately 200 microns or smaller in diameter and appeared evenly
distributed.
This film lasted only 6 to 12 days in the weatherometer. All samples turned
entirely white very early, and disintegrated into fine particles upon
embrittlement. Rubbing the embrittled material between fingertips resulted in a
73
crumbly, flaky collection of small (sub-millimetre) particles of approximately
even size.
The sample with 3% Degussa P25 loading was also translucent, although much
less so than the 1% sample. Several layers thickness of the film showed the
plastic was white. UV degradation experiments produced similar results to the
1% loading. All samples reached embrittlement in 6 to 12 days, although they
tended to embrittle earlier than the 1% sample. Again the samples whitened and
became opaque very early.
Backscattered SEM images (Figure 3-1) show a particle size in the film of up to,
and in some cases exceeding, 5 µm. The Degussa P25 titania appears
agglomerated and poorly dispersed. This poor dispersion in the polyethylene
film is possibly due to the lack of surface modification of the titania particles.
Figure 3-1 Backscattered SEM images of film containing 3% (left) and 1% (right) Degussa P25. (50 µm size scale).
3.2.2 Kronos
The Kronos 1% loading film was semi-transparent white, with a fine dispersion
of particles. The 3% loading film had a similar, albeit less transparent,
appearance to the 1% film. A size dispersion of particles could be seen similar to
the 1% loading film (Figure 3-2).
74
The Kronos 1% loading film did not undergo a significant colour change with
exposure to UV. Embrittlement typically took between 20 and 30 days. The 3%
films on the other hand whitened considerably and heterogeneously. The 3%
loading material tended to flake when fully embrittled.
Figure 3-2 Backscattered SEM images of film containing 1% (left) and 3% (right) Kronos Titania. (50 µm size scale).
3.2.3 Huntsman Tioxide
The two Huntsman Tioxide samples were very white when observing a single
film and this became opaque with 4 sheets thickness. The plastic film has a very
glossy look and feel. It demonstrated longer embrittlement times than the
Degussa P25 samples. The film containing water dispersible organic-coated
particles (A-HR) had a wide range of embrittlement times, varying from 3 to 21
days. Organic dispersible (A-HRF) titania in LLDPE showed more consistency
in aging times, ranging from 9 to 18 days to embrittlement. The Huntsman
Tioxide samples whitened heterogeneously with exposure to UV.
Backscattered SEM images (Figure 3-3) show good particle dispersion, and
smaller particle size than the Degussa P25 samples, although there is still some
agglomeration, and many particles around 1 µm in size.
75
Figure 3-3 Backscattered SEM images of film containing 3% A-HR (left) and A-HRF (right) Huntsman Tioxide. (20 µm size scale on the left and 50 µm on the right.)
3.2.4 Sachtleben Hombitan
The Sachtleben Hombitan films were all white, and 4 sheets thickness resulted in
opacity. The particles appeared to be evenly dispersed, as no imperfections could
be seen in the film when held up to the light. A tinge of brown could be seen in
the folded antimony-doped and organic coated films, perhaps due to the nature of
the modifiers on the titania. The organic and aluminium coated titania film was
very white and had a very high gloss. No brown tinge could be seen in this film.
The Sachtleben Hombitan films demonstrated the longest times to embrittlement
of all films in the weatherometer, generally between 12 and 30 days, with some
exceptions. The Sachtleben Hombitan LW-S-12 sample exhibited the greatest
degree of whiteness when subjected to UV radiation, whilst the aluminium and
organic coated particles remained relatively clear until embrittlement. The
aluminium and organic coated titania samples tended to tear when embrittled,
rather than become flaky as the whitened samples containing Degussa P25
titania. These samples took up to 45 days to embrittle.
Sachtleben Hombitan titania appeared well distributed in the polyethylene films
(Figure 3-4), with a particle size of under 1 µm. The organic coated particles
demonstrated the best dispersion.
76
Figure 3-4 Backscattered SEM images of film containing 3% LW-S-U (upper left), LW-S-U-HD (upper
right), LW-S-12 (lower left) LC-S (lower right) Sachtleben Hombitan titania. (20 µm size scale).
3.2.5 Section summary
All of the films were translucent, varying slightly in colour between white and
off-white. Higher titania loadings resulted in less transparency. The LLDPE film
containing Degussa P25 titania showed the poorest particle distribution, as well
as significant agglomeration of particles. Titania with organic coatings
demonstrated improved particle dispersal and a narrow particle size distribution.
77
LLDPE films that degraded rapidly in the weatherometer became white, and
disintegrated into small flaky particles when embrittled. LLDPE films that took
longer to degrade in the weatherometer tended to remain translucent, and tore at
embrittlement. The difference in degradation results was due to the photoactivity
of the titania, and the cause of sample whitening is discussed in the following
section.
3.3 Sample whitening
The phenomenon of whitening was examined by SEM. An image taken of a
whitened, degraded polyethylene film containing 1% Degussa P25 titania and the
corresponding backscattered image is shown in Figure 3-5.
Figure 3-5 SEM image of polyethylene containing 1% Degussa P25 titania particles following photodegradation. The image on the right is the backscattered image. Note the appearance of dark areas in the backscattered image around the titania particles, indicating an absence of material at these locations.
The titania particles clearly show up as white dots in the backscattered image. It
can be seen that there are dark areas in close proximity to many of these
particles, often in the shape of a ‘tail’, or a ‘wormhole’, which represents an
absence of material. The effect of titania creating holes in polymeric material
under the application of UV radiation with subsequent whitening has been
demonstrated many times in the literature 77,78,88,177-179.
78
It has been postulated that the whitening of materials containing titania can be
explained by the presence of cavities78. Cavities created by titania-catalysed
photodegradation, such as those observed in Figure 3-5, scatter visible light,
producing a whitening effect. The scattering of light by pores in a material is
utilised by the coatings industry to produce white coatings180, and the pore size
must be one half the wavelength of incident light to achieve maximum
scattering181,182. Thus to scatter visible light (λrange = 400 nm – 750 nm), a pore
diameter of between approximately 0. 2 µm and 0.4 µm is required. A cavity
created by titania particles, such as captured by the SEM images in Figure 3-5, is
therefore capable of scattering visible light such that the LLDPE film appeared
white.
Whitening of the LLDPE film provides an indication of the relative photoactive
strengths of the various types of titania used in this thesis. Samples containing
modified titania, such as the Satchleben Hombitan organic coated titanias, did
not whiten, and took longer to degraded. These films appeared similar to the
control sample at embrittlement, remaining quite clear. Conversely LLDPE film
containing Degussa P25 embrittled quickly and whitened.
The cavities formed by titania affect the physical properties of the film. The loss
of material, as evidenced by the wormholes in Figure 3-5, combined with rapid
and extensive oxidation of the LLDPE chains, causes the film to fall apart with
exposure to UV irradiation. This is evidenced by the homogeneous flaking of the
film at embrittlement, rather than the shrinking and tearing of the control film
and those containing less active titania.
79
3.4 Times to embrittlement for LLDPE film containing titania
The following tables describe the experiments performed for each LLDPE film,
and the time taken for each sample to achieve embrittlement. The first two
columns indicate whether the samples were aged in the oven or suntest, the next
two columns indicated what wavelength of pre-irradiation UV was used, and the
subsequent columns indicated the length of pre-irradiation (a time of 0 sec
indicates that the sample was not subjected to pre-irradiation). The last column
denotes the numbers of days aged until embrittlement was achieved, and a ‘+’
after the number indicates that the sample had still not achieved embrittlement at
the conclusion of the experiment.
Thus, as an example, the 3rd line in Table 2 shows that the LLDPE film
containing 1% Degussa P25 was pre-irradiated with UVA for 3 hours and aged
in the oven, and embrittlement was not achieved in over 330 days of aging.
The experiments were not performed in the order presented in the tables, but
were randomised instead to assist in reducing drift and sample carry over errors
(Section 2.2).
Table 2 Data for the LLDPE film containing 1.00% TiO2 Degussa P25 showing sample aging details and days taken to reach embrittlement. 1.00% TiO2 Degussa P25
Oven Suntest UVA UVC 0 sec 60 sec 3 hr 24 hr Days to emb.
370+ 400+ 330+ 200+ 270+ 400+ 392 24 6 12 6 9 12
80
12 9 3
Table 3 Data for the LLDPE film containing 3.00% TiO2 Degussa P25 showing sample aging details and days taken to reach embrittlement. 3.00% TiO2 Degussa P25
Oven Suntest UVA UVC 0 sec 60 sec 3 hr 24 hr Days to emb.
375 330+ 330+ 181 330+ 381 371 229 12 6 6 9 9 9 6 6
Table 4 Data for the LLDPE film containing 1.00% TiO2 Kronos 1002 showing sample aging details and days taken to reach embrittlement. 1.00% TiO2 Kronos 1002
Oven Suntest UVA UVC 0 sec 60 sec 3 hr 24 hr Days to emb.
400+ 330+ 330+ 370+ 330+ 400+ 330+ 330 27 27 21 27 30 36 24 3
81
Table 5 Data for the LLDPE film containing 3.00% TiO2 Kronos 1002showing sample aging details and days taken to reach embrittlement. 3.00% TiO2 Kronos 1002
Oven Suntest UVA UVC 0 sec 60 sec 3 hr 24 hr Days to emb.
370+ 470+ 370+ 330+ 400+ 370+ 200+ 270+ 24 24 18 15 21 24 15 9
Table 6 Data for the LLDPE film containing 3.00% TiO2 Huntsman Tioxide A-HR showing sample aging details and days taken to reach embrittlement. 3.00% TiO2 Huntsman Tioxide A-HR
Oven Suntest UVA UVC 0 sec 60 sec 3 hr 24 hr Days to emb.
200+ 270+ 415+ 370+ 200+ 470+ 200+ 233 15 15 12 15 15 21 12 3
82
Table 7 Data for the LLDPE film containing 3.00% TiO2 Huntsman Tioxide A-HRF showing sample aging details and days taken to reach embrittlement. 3.00% TiO2 Huntsman Tioxide A-HRF
Oven Suntest UVA UVC 0 sec 60 sec 3 hr 24 hr Days to emb.
200+ 270+ 330+ 270+ 470+ 270+ 270+ 371 15 15 15 18 12 15 9 9
Table 8 Data for the LLDPE film containing 3.00% TiO2 Sachtleben Hombitan LW-S-U showing sample aging details and days taken to reach embrittlement. '+' indicates the sample had not reached embrittlement when the experiment had ended. 3.00% TiO2 Sachtleben Hombitan LW-S-U
Oven Suntest UVA UVC 0 sec 60 sec 3 hr 24 hr Days to emb.
330+ 370+ 470+ 270+ 270+ 415+ 330+ 327 21 21 30 30 24 32 9 12
83
Table 9 Data for the LLDPE film containing 3.00% TiO2 Sachtleben Hombitan LW-S-U-HD showing sample aging details and days taken to reach embrittlement. 3.00% TiO2 Sachtleben Hombitan LW-S-U-HD
Oven Suntest UVA UVC 0 sec 60 sec 3 hr 24 hr Days to emb.
430+ 430+ 415+ 370+ 200+ 430+ 270+ 330 27 24 27 27 30 24 27 12
Table 10 Data for the LLDPE film containing 3.00% TiO2 Sachtleben Hombitan LW-S-12 showing sample aging details and days taken to reach embrittlement. 3.00% TiO2 Sachtleben Hombitan LW-S-12
Oven Suntest UVA UVC 0 sec 60 sec 3 hr 24 hr Days to emb.
200+ 430+ 330+ 200+ 470+ 470+ 200+ 73 18 15 21 18 15 12 12 9
84
Table 11 Data for the LLDPE film containing 3.00% TiO2 Sachtleben Hombitan LC-S showing sample aging details and days taken to reach embrittlement. 3.00% TiO2 Sachtleben Hombitan LC-S
Oven Suntest UVA UVC 0 sec 60 sec 3 hr 24 hr Days to emb.
430+ 470+ 430+ 330+ 200+ 370+ 470+ 392 39 30 27 27 45 33 39 21
Table 12 Data for the control sample showing sample aging details and days taken to reach embrittlement. Control
Oven Suntest UVA UVC 0 sec 60 sec 3 hr 24 hr Days to emb.
266+ 266+ 266+ 266+ 266+ 266+ 266+ 266+ 66 66 60 57 66 66 48 24
These tables show that the samples have been aged in the oven for considerable
periods of time; up to 470 days for some samples. As stated in the objectives
(Section 1.4) it is desirable that the films embrittle completely in the dark after
85
about 6 months of oven aging following pre-irradiation, and thus all samples
have undergone a minimum of 200 days spent aging in a dark environment at 50
°C. All samples tested in the weatherometer were aged to embrittlement.
Two samples have come close to the target of embrittlement in the oven by 6
months; both samples are LLDPE containing 3% Degussa P25 titania (see Table
3), and were pre-irradiated with 24 hours of UVA (181 days) and UVC (229
days) respectively. All other samples exposed to UV pre-irradiation and aged in
the oven took over 200 days to embrittle. It is probable that for a commercial
application, 24 hours of UV pre-irradiation is too long to be practical; however
from the point of view of exploring new technologies, it is important to recognise
that these samples provide evidence of the success of pre-irradiation as a
concept, and a starting platform for the continued research into this new
technology. An in-depth analysis of these, and all samples in the pre-irradiation
experiment, is provided in the following sections.
3.5 IR spectral analysis – control film (undegraded)
3.5.1 Polyethylene absorption table
The infrared characterisation of polyethylene is well known, and the early work
published by Rugg et al. in the 1950’s 24,25 is still referred to 169,183,184. A
complete description of straight chain alkane vibrations from C3H8 through to n-
C19H40 was provided by Snyder and Schachtschneider 185,186. The differences in
absorption in the mid-IR of different types of polyethylene, and the effects of
crystalline and amorphous regions were discussed in Section 1.3.3.4. An
ATR/FTIR spectrum of the control film used in this thesis is provided in Figure
3-6, and a table assigning the absorptions is provided in Table 13.
86
4000 3500 3000 2500 2000 1500 1000
0.0
0.1
0.2
0.3
0.4
0.5
Abso
rban
ce
Wavenumbers (cm-1)
Figure 3-6 ATR/FTIR spectrum of the LLDPE control film. This film does not contain
titania.
Table 13 Mid-infrared absorption table for Dowlex 5056 G polyethylene.
Absorption (cm-
1)
Appearance Assignment 24,25,186,187
~2953 Weak, shoulder -CH3 antisymmetric stretc.h
2914 Very strong -CH2 antisymmetric stretc.h
2847 Very strong -CH2 symmetric stretc.h
2690-2630 Weak, broad collection
of small bands
Various CH2C and CH3C
structural features 188, possible
overtones
1471, 1463 Strong, doublet -CH2 deformation
~1445 weak, shoulder -CH3 antisymmetric bend
1212, 1195, Weak, sharp bands Methyl bending vibrations
729, 718 Strong, sharp CH2 rocking modes
87
3.5.2 Titania absorption in the mid-infrared
4000 3500 3000 2500 2000 1500 1000
0.0
0.1
0.2
0.3
0.4
0.5
Subraction result
Abs
orpt
ion
Wavenumbers (cm-1)
Kronos film Control film
Figure 3-7 Spectral subtraction result (Black) of control film (Magenta) subtracted from
1.00% Kronos film (Blue). Spectra are to scale. The spectra have been offset.
Anatase and Rutile titania absorb mid infrared light below ~850 cm-1 189. The
strength of this absorption increases with increasing titania concentration. The
presence of titania does not appear to otherwise significantly affect the infrared
spectrum of polyethylene, which can be seen in Figure 3-7 where the subtraction
result demonstrates the additional titania absorption in the Kronos film, while the
CH2 absorption peaks are largely similar aside from some minor broadening in
the CH stretch region.
3.6 Processing agent absorptions
Absorptions occur in the spectra of the Ciba films that cannot be assigned to
polyethylene, and are representative of an organic material that has been added to
88
the LLDPE film. Spectral subtraction was used to extract a spectrum of this
material from the polyethylene (Figure 3-8), and possible band assignments are
provided in Table 14.
1750 1500 1250 1000 750
0.000
0.005
0.010
0.015
0.020
0.025
0.030
Abso
rban
ce
Wavenumbers (cm-1)
Figure 3-8 Subtraction spectrum showing additive absorption peaks.
Table 14 Mid-infrared absorption table for processing agent present on the surface of polyethylene Absorption (cm-1) Appearance Assignment 24,25,186,187
1490, 1460 Strong, doublet -CH2 deformation
1398, 1361 Medium, doublet t-butyl –(CH3)3
1212, 1195 Medium, doublet t-butyl –(CH3)3
1081 Medium, sharp isopropyl –(CH3)2 (?)
906 Weak, sharp Vinyl (?)
854 Weak, sharp Vinyl (?)
776 Weak, sharp t-butyl symmetric skeletal stretc.h
719 Strong -CH2 rock
89
There are no absorptions in the spectrum above 1500 cm-1, which excludes the
possibility of aromatic or carbonyl functional groups on the compound.
Additionally, this spectrum only appears on one side of the polymer film, and is
on the surface. It is likely that it is some kind of aliphatic processing agent,
possibly a film blowing, anti-static or anti-slip agent. Similar compounds are
known to be used in some film blowing processes190. The spectrum of this
additive looks similar to polyisoprene or polybutylene, although some
absorptions are not so closely matched that a positive identification can be made.
There is a great deal of processing agents used by different manufacturers for
different purposes, making positive identification extremely difficult. Soxhlet
extraction techniques were performed using a variety of solvents in attempts to
isolate this additive, without success.
The focus of this thesis is to obtain results that will be directly transferable into
‘real world’ applications. The materials subjected to study are commercially
available ones, and contain additives and processing agents such as these. In the
context of shopping bags and other applications, it reasonable to expect that the
side subjected to sunlight cannot be controlled, and the effect of pre-irradiation
and aging processes must be investigated in a manner that variables such as
additive concentrations, processing agents on the surface, machine direction of
the blown film, etc., are randomised. Thus the side of the film was not
deliberately taken into consideration when aging the film. The effect of the
processing agent on degradation is investigated in Chapter 4.
3.7 IR spectral analysis – control film (degraded)
Oxidative degradation of the polyethylene film resulted in significant changes in
the infrared absorption spectrum (Figure 3-9). In all cases an increase in
absorption occurred in the carbonyl region (1850-1650 cm-1), indicating the
presence of degradation products including some kind of oxygenated functional
group. A broad increase in absorption was noticed below 1300 cm-1, with some
specific peaks that can be related to esters and unsaturation. The OH stretc.hing
region displays evidence of carboxylic acid type OH formation.
90
4000 3500 3000 2500 2000 1500 1000
0.0
0.1
0.2
0.3
Abso
rban
ce
Wavenumbers (cm-1)
Control Control 66 days weatherometer Subtraction result
Figure 3-9 ATR/FTIR spectral subtraction result (Blue) of unaged control sample (Black)
subtracted from 66 days weatherometer aged control sample (Red). Spectra have been
offset.
3.7.1 OH stretc.hing region (3800-3200 cm-1)
Absorption at the high wavenumber end of the spectrum indicating the presence
of alcohol oxidation products is expectedly weak using ATR/FTIR
spectroscopy144. For reasons discussed in Section 1.3.3.3, the spectra have not
undergone ATR correction, and thus the high wavenumber end of the spectrum
does not display full strength absorptions. However subtraction spectra such as
that shown in Figure 3-9 indicate that some alcohol functional groups are present
in small quantities.
3.7.2 Carbonyl region
Examination of the carbonyl region of the spectrum of degraded polyethylene
shows multiple, overlapping bands. Curve fitting of this region has been
91
performed to extract peak positions. It can be seen in Table 15 that more
assignments have been made than there have been bands fitted. The carbonyl
region of these degraded polymers is in most cases fairly broad and undefined,
and addition of further bands to the curve-fitting calculations does not result in a
more accurate fit. As was seen in Section 1.1.3 the carbonyl region of degraded
polymers, such as polyethylene, shows the absorptions of many different kinds of
carbonyl-containing degradation products. In a curve fitting exercise it is useful
to assign the major bands, whilst being aware that multiple absorptions are likely
to be hidden under the same peak. The results are shown in Figure 3-10, and the
peak assignments are given in Table 15.
1800 1700 1600
Wavenumbers (cm-1)
Figure 3-10 Carbonyl region of Figure 3-9.
Table 15 Curve fitting results of the carbonyl region for the control film
Absorption (cm-1) Assignment 184,187
1785 Lactones, anhydrides, peracids
1763 Peresters, anhydrides
1733 Esters and aldehydes
1710 Ketones and carboxylic acid
1641 (sharp) RCH=CH2
1639 Carboxylates
92
3.7.3 Below 1500 cm-1
Several changes occur in the absorption spectrum of polyethylene below 1500
cm-1 with oxidation (Figure 3-11). A feature that is common to all polyethylene
films studied was the broad absorption increase from approximately 1400 cm-1 to
600 cm-1. This absorption increase is likely to be a complex combination of
signatures arising from absorptions such as C-O stretc.hes of esters, anhydrides,
carboxylates, etc.. An increase in absorption intensity is in this region is common
to all degraded samples, both those containing titania and those without. The
strong absorption at 1180 cm-1 is assigned to an ester C-O stretc.hing
vibration187.
1250 1000 750
-0.005
0.000
0.005
0.010
0.015
0.020
0.025
0.030
Abso
rban
ce
Wavenumbers (cm-1)
Figure 3-11 Expansion of subtraction result in Figure 3-9 below 1500 cm-1.
3.7.4 Section summary
The LLDPE used in this study oxidises to produce absorption peaks in the
infrared spectrum that agree with exhaustive studies published in the literature
over decades. A weak, broad absorption above 3200 cm-1 indicates the presence
93
of alcohols. Investigation of the carbonyl region between 1900 and 1650 cm-1
shows that esters and acids form a large proportion of degradation products, with
some anhydride and lactone absorptions at higher wavenumbers. The area below
1400 cm-1 has a very broad increase in absorption, indicating the presence of
various oxygenated degradation products, with a strong absorption at 1180 cm-1
assigned to an ester C-O stretc.h. An aliphatic processing agent is present on one
side of the film, which appears to contain a high concentration of tertiary methyl
groups.
3.8 Effect of UV irradiation – control film (degraded)
Degradation experiments in the weatherometer and in the oven were performed
according to the details provide in Section 2.2. Section 3.8 covers the control
film, and will examine the development of the carbonyl index, drawing
comparisons between the films. The carbonyl index was acquired according to
the method described in Section 2.3.
3.8.1 Control, weatherometer aged samples
This section will examine the effect that different times of exposure to UV
irradiation had on the control sample aged in the weatherometer. UVC radiation
will be investigated first, followed by UVA radiation.
94
3.8.1.1 Effect of pre-irradiation – UVC
0 10 20 30 40 50 60 700.00
0.05
0.10
0.15
0.20
0.25
0.30C
arbo
nyl i
ndex
Days aged in weatherometer
0 secs 60 secs 3 hrs 24 hrs
Figure 3-12 Carbonyl index plots of the control film pre-irradiated with UVC for 0 secs, 60
secs, 3 hours and 24 hours and subsequent exposure in the weatherometer. Note the earlier
time to embrittlement combined with lower carbonyl index at embrittlement of the 24 hour
exposed sample. Second order polynomial trend lines have been fitted to depict a trend, and
do not imply reaction mechanism or theory.
The carbonyl index plots in Figure 3-12 show the effects of pre-irradiating
LLDPE with UVC, and subsequent weatherometer aging. Two key points can be
recognised from observation of the plots in Figure 3-12.
Firstly, significant UVC exposure shortens the lifetime of the polymer. The
carbonyl index plot of the sample exposed to 60 secs of UVC is virtually
indistinguishable from the untreated sample, while the samples exposed to 3
hours and 24 hours of UVC demonstrate accelerated degradation. In this case, 3
hours of pre-irradiation causes a reduction of the time to embrittlement of the
polymer by one third, and 24 hours pre-irradiation reduced this by a further one
third. Embrittlement is defined in Section 2.2.
95
Secondly, the slopes of the carbonyl index plots in Figure 3-12 imply that
oxidation is occurring more rapidly in the samples with longer UVC exposure
times. If the slopes had been similar across all samples, with only an offset on the
y-axis to demonstrate an increased concentration of oxidation products formed
during UVC treatment, it could be concluded that the pre-irradiation had merely
started degradation earlier. However, the steeper slopes indicate a faster rate of
reaction, which is attributed to the polymer samples with longer UVC exposure
times being more susceptible to further photooxidation in the weatherometer than
those samples with lesser UVC pre-treatment.
The carbonyl index plots clearly demonstrate that UV irradiation increases the
rate of degradation. Furthermore, larger doses of UVC irradiation appear to have
had a significant effect on the rate of oxidation product formation. The spectra of
the 3 hour and 24 hour UVC pre-irradiated samples prior to weatherometer aging
show an absorption at 1641 cm-1, which is assigned to a vinylic absorption (Table
15). In addition to the creation of unsaturated polymer chains, the films have
shrunk in the sample holder, indicating extensive crosslinking35. Crosslinked
polymers have a higher concentration of hydrogens tertiary to backbone carbons,
which are more susceptible to hydroperoxide attack11.
Section 1.1.2 covered the effect of oxidation occurring initially in defective sites
in polymer, and spreading from there to the polymer bulk. High doses of UV
radiation, and in this case particularly UVC, appear to have created reactive sites
via double bonds and crosslinks, which has resulted in a faster rate of
degradation of the bulk polymer. This is evidenced by the steep slope of the
carbonyl plots of the UVC pre-irradiated samples. Thus ‘weakening’ of the
polymer via UV irradiation has resulted in shorter embrittlement times.
96
3.8.1.2 Effect of pre-irradiation – UVA
0 10 20 30 40 50 60 700.00
0.05
0.10
0.15
0.20
0.25
0.30C
arbo
nyl i
ndex
Days aged in weatherometer
0 secs 60 secs 3 hrs 24 hrs
Figure 3-13 Carbonyl index plots of the control film pre-irradiated for 0 secs, 60 secs, 3 hrs
and 24 hours with UVA radiation. Second order polynomial trend lines have been added.
In contrast to their UVC pre-irradiated counterparts, the carbonyl index plots of
the UVA pre-treated samples in Figure 3-13 do not demonstrate significant
differences. For all UVA pre-irradiated samples, there is little change in the
carbonyl region following pre-irradiation. Close examination of the carbonyl
absorption region of the 24 hour pre-irradiated sample prior to weatherometer
aging does show some absorption; however this is barely above the signal noise.
Despite the similar carbonyl index plots, UVA pre-irradiation affected the
outcome of film degradation. Increasing the length of pre-irradiation resulted in
shorter embrittlement times, as samples pre-irradiated for 0 secs, 60 secs, 3 hours
and 24 hours embrittled in 66, 66, 60 and 57 days in the weatherometer
respectively (Table 12).
97
3.8.1.3 Section summary
Pre-irradiation of LLDPE film under UVC or UVA before aging in a
weatherometer decreases the time taken to achieve embrittlement. The film must
be irradiated for a significant period of time to produce an effect. Irradiation
results in reactive sites in the polymer matrix that are susceptible to
hydroperoxide formation, increasing the rate of degradation. Although oxidation
occurs more quickly, the heavily irradiated samples show less oxidation product
formation at embrittlement, suggesting that other degradation mechanisms are
occurring.
Pre-irradiation with UVC has a more significant impact than with UVA on the
time taken to embrittlement of the control sample. This is most likely due to the
higher energy of UVC, which results in the formation of more reactive sites. The
degrading effect of UVC is reflected by the faster embrittlement times, and lower
carbonyl index at embrittlement of the UVC pre-irradiated samples.
3.8.2 Control, oven aged samples
This section will examine the effect that different times of exposure to UV
irradiation have on the control sample aged in the oven at 50 °C. UVC pre-
radiated samples will be investigated first, followed by UVA pre-radiated
samples.
98
3.8.2.1 Effect of pre-irradiation – UVC
0 50 100 150 200 250 3000.0
0.1
0.2
0.3
0.4
0.5
0.6
Car
bony
l ind
ex
Days aged in oven
0 secs 60 secs 3 hrs 24 hrs
Figure 3-14 Carbonyl index plots for UVC pre-irradiated control samples aged in the oven.
Second order polynomial trend lines have been added. Error bars showing standard
deviation have been added to this figure and Figure 3-15; however they have been omitted
from the other figures for clarity
Pre-irradiation of the control film with UVC has strongly affected the
degradation of the polymer. Figure 3-14 shows little difference in the carbonyl
index of the non-irradiated and the 60 sec irradiated films.
99
3.8.2.2 Effect of pre-irradiation – UVA
0 50 100 150 200 250 3000.00
0.02
0.04
0.06
Car
bony
l ind
ex
Days aged in oven
0 secs 60 secs 3 hrs 24 hrs
Figure 3-15 Carbonyl Index plots for UVA pre-irradiated control samples aged in the oven.
Second order polynomial trend lines have been added. An outlier has been removed from 3
hr sample.
Similarly to those results seen for the weatherometer aged control sample, UVA
pre-irradiation did not have as large an impact on the rate of oxidation as UVC
pre-treatment (Figure 3-15). Even the sample with 24 hours exposure to UVA
did not degrade significantly until after 100 days in the oven. Note that the
carbonyl index scale on the y-axis shows that the LLDPE control film samples
are still in the early stages of oxidation.
3.8.3 Section Summary
Oven aged control samples (LLDPE film without titania) take much longer to
degrade than weatherometer aged control samples. None of the samples have
achieved embrittlement in the oven, despite aging for over 266 days. Pre-
100
irradiation with UVC results in more significant carbonyl product formation than
UVA with similar irradiation times.
3.9 IR spectral analysis – film containing titania (degraded)
Films containing titania degraded in the weatherometer and in the oven
considerably faster than the control. Not all films containing titania behaved the
same, with some films achieving embrittlement much faster than others. The
following sections will examine the effect of titania on the IR spectra, and
compare the degradation of LLDPE film containing titania from different
manufacturers.
3.9.1 Carbonyl region
An interesting feature of the degradation of this film is the carbonyl absorption at
embrittlement. Figure 3-16 shows spectra of 1% Degussa P25 film aged in the
oven and weatherometer compared with the control aged in the weatherometer.
101
1800 1750 1700 1650 1600
0.01
0.02 1% Degussa P25 weatherometer
Abs
orba
nce
Wavenumbers (cm-1)
Control in weatherometer 1% Degussa P25 oven
Figure 3-16 Comparison of control film without pre-irradiation embrittled in the
weatherometer, 1% Degussa P25 without pre-irradiation after 330 days in the oven, and
1% Degussa P25 without pre-irradiation embrittled after 12 days in the weatherometer.
Both the 1% Degussa P25 containing samples had embrittled, while the control sample had
not. Spectra are to scale.
It is immediately evident that the carbonyl region of the weatherometer aged film
containing 1% Degussa P25 titania at embrittlement has not developed to the
extent of the control film at the end of its lifetime in the weatherometer. This
implies that there are less oxygenated degradation products present in the
material at the point of failure in the 1% Degussa P25 film weatherometer aged,
than in the control film, and therefore degradation processes other than those
resulting in the formation of products containing carbonyl species are combining
to result in embrittlement. The ‘wormhole’-like cavities shown in the SEM
images in Figure 3-5 demonstrate the loss of material due to irradiation of titania
particles. It was shown that this effect was strongest in the materials containing
Degussa P25, implying that these particles are most active in degrading the
polymer matrix.
102
The material that has been destroyed by the titania particles has in all likelihood
been converted directly to volatiles including CO2 and H2O191. The likelihood of
this is substantiated by multivariate statistical analysis presented in Section 4.4.1.
This at least partially explains both the clear absence of material evident in the
SEM images, and the lack of carbonyl absorption strength in the infrared spectra.
The titania is contributing to the faster embrittlement times when the LLDPE is
subjected to extensive UV irradiation, by forming cavities which could weaken
the polymer matrix.
The magenta spectrum in Figure 3-16 is of the LLDPE film containing 1%
Degussa P25 and aged in the oven for 330 days (it had not embrittled after this
time). The film had not undergone UV pre-irradiation, and does not show a
strong absorption at 1732 cm-1, an absorption present in both the weatherometer
aged control and 1% Degussa P25 films. This supports the degradation scheme
proposed by Tidjani, given in Scheme 1-9. According to this scheme, oxygen
centred radicals formed by the absorption of radiation by hydroperoxides give
rise to ester oxidation products (among other products). Thus in samples exposed
to significant amounts of UV (such as the weatherometer aged samples) can be
expected to have ester products in higher concentrations. Alternatively,
hydroperoxides that decompose by heat produce a carbon centred radical, which
is more likely to result in acid oxidation products. This explains the relatively
lower ester concentration in the samples aged in the oven compared to those aged
in the weatherometer.
Further to the differences in absorption of mid-IR caused by the different aging
conditions (oven vs. weatherometer), it is worth noting the similarity of the
carbonyl region of the control sample and LLDPE film containing 1% Degussa
P25. Despite a difference in overall absorption intensity, as discussed in earlier
paragraphs, the shape of the absorption peaks occurring in this region are very
similar. This indicates similar relative concentrations of degradation products,
with different absolute concentrations. It is apparent from these observations that
the titania is speeding up the process of embrittlement, however the oxidation
pathway of the material is not changing.
103
3.9.2 Fingerprint region
Corresponding with the growth of an absorption in the carbonyl region with
increasing oxidation, a peak formed in the infrared spectra at 1178 cm-1. This is
most likely to be the C-O stretc.h of an ester peak. Also, there is a broad increase
in absorption in this region (Figure 3-17).
1400 1350 1300 1250 1200 1150 1100 1050 1000 950 9000.00
0.01
0 days
Abs
orpt
ion
Wavenumbers (cm-1)
3 days 6 days
Figure 3-17 IR Spectra of 1% Degussa P25 without pre-irradiation aged for 0 days, 3 days,
and 6 days in the weatherometer. Note the increase appearance of an ester C-O absorption
at 1180 cm-1, combined with genearally higher absorption in this region. The absorption at
1080 cm-1 is the processing agent.
3.9.3 Section summary
The presence of photoactive titania caused the LLDPE film containing 1%
Degussa P25 to embrittle after just a few days in the weatherometer. Degradation
could be followed in the IR spectrum by an increase in carbonyl absorption and a
broad increase in absorption of the fingerprint region.
104
The carbonyl region of the IR spectra of samples aged in the weatherometer
containing 1% Degussa P25 and the control show very similar relative absorption
intensities, implying that there are similar relative amounts of oxidation products.
However the control sample had a much stronger carbonyl absorption overall,
indicating a greater extent of oxidation at embrittlement. The lack of oxidation
product build-up in the weatherometer aged LLDPE film containing titania is
probably due to the effects of other degradation processes, such as weakening of
the film due to cavities.
Oven aged samples showed very little ester absorption intensity in the IR
spectrum, confirming the applicability of the Tidjani degradation scheme to these
LLDPE film samples.
3.10 LLDPE containing Degussa P25 (degraded)
3.10.1 Degussa P25, weatherometer aged samples
All samples containing Degussa P25 titania degraded very quickly in the
weatherometer. This made it difficult to notice any significant effects of pre-
irradiation on the embrittlement times of these samples.
Figure 3-18 shows the effect of different exposure times of UVC on the carbonyl
absorption of the polymer containing 1% Degussa P25. Sixty seconds exposure
did not appear to result in significant changes to this region. Three hours
exposure shows a gain in absorption of carbonyl containing species such as
ketones, esters and acids. Twenty-four hours exposure resulted in a substantial
carbonyl absorption, as well as an unsaturation peak at 1641 cm-1. The formation
of unsaturation to varying degrees occurred in samples exposed to all forms of
radiation, including pre irradiation with UVA and UVC, and the weatherometer.
However, the absorption band was sharpest and relatively most intense in
samples exposed to large doses of UVC.
105
3.10.1.1 Effect of pre-irradiation – UVC
1850 1800 1750 1700 1650 1600
-0.001
0.000
0.001
0.002
0.003
0.004
0.005
0.006 0 secs
Abs
orpt
ion
Wavenumbers (cm-1)
60 secs
24 hrs 3 hrs
Figure 3-18 Comparison of the carbonyl region of LLDPE film containing 1% Degussa P25
after pre-irradiation with UVC for 24 hours (Blue), 3 hours (Green), 60 seconds (Red) and 0
seconds (Black).
The 1% Degussa P25 film showed some differences to the control film when
aged in the weatherometer. Figure 3-19 gives the carbonyl index plots for the
weatherometer aged samples.
106
0 2 4 6 8 10 12 140.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
0.20
0.22
0.24 0 secs 60 secs 3 hrs 24 hrs
Car
bony
l ind
ex
Days aged in weatherometer
Figure 3-19 Carbonyl index plots for UVC pre-irradiated LLDPE film containing 1% Degussa P25 titania and aged in the weatherometer. Order polynomial trend lines have been added.
The effect of titania can be seen in these carbonyl plots when compared with
those of the control sample given in Figure 3-12. The samples have achieved
embrittlement much earlier than the control sample, and pre-irradiation with
UVC has reduced that time even further. Additionally, the carbonyl index at
embrittlement is much lower than the control samples. These observations imply
quicker degradation rates, with less oxygenated functional groups at
embrittlement.
The lower concentration of oxygenated degradation products at embrittlement
suggests that other degradation processes are having a significant effect. IR
spectra of these materials exhibit a more intense unsaturation absorption band
than the control sample. Furthermore, as discussed in Section 3.9.1 the polymer
appears to have been ‘burned away’ by titania, and the cavities left have
weakened the material, hastening the embrittlement.
107
3.10.1.2 Effect of pre-irradiation – UVA
The samples pre-treated with UVA radiation showed similar trends to those
treated with UVC. Figure 3-20 shows the carbonyl plots for these samples.
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 150.00
0.01
0.02
0.03
0.04
0.05
0.06
0.07
0.08
0.09
0.10
Car
bony
l ind
ex
Days aged in weatherometer
0 secs 60 secs 3 hrs 24 hrs
Figure 3-20 Carbonyl index plots for UVA pre-irradiated LLDPE film containing 1%
Degussa P25 titania and aged in the weatherometer. Second order polynomial trend lines
have been added.
Samples pre-irradiated with UVA embrittled by 12 days, which is comparable to
the UVC irradiated samples. The carbonyl index of the samples at embrittlement
is similar for all samples irradiated with UVA, with the exception of the sample
irradiated for 24 hours. However, it was noted that the sample actually appeared
embrittled after 6 days, and by 9 days the polyethylene film had been nearly
completely destroyed. The most significant difference between the samples
irradiated with UVA or UVC and aged in the weatherometer is the carbonyl
index of the samples at embrittlement of the 24 hour pre-irradiated samples. The
108
higher carbonyl index at embrittlement of the 24 hour UVC irradiated sample
indicates a higher concentration of oxidation products in this sample.
Considering that the sample entered the weatherometer at a higher starting
carbonyl index, it is clear that significant oxidation had taken place during the 24
hours of UVC irradiation. In fact, the starting carbonyl index of 0.9 is
comparable to the carbonyl index of samples at embrittlement after several days
of aging in the weatherometer. This indicates that although the sample had
undergone significant oxidation under UVC irradiation the sample had not
embrittled, that is to say chain scission reactions had not proceeded to the extent
that the film began to fall apart. The sample irradiated with 3 hours of UVC
shows a similar trend.
3.10.2 Section summary
Degussa P25 titania greatly reduces the time taken to embrittlement in the
weatherometer compared to the control film, resulting in some cases in a 10-fold
increase in the degradation rate. Pre-irradiation with UV did not have a great
impact on embrittlement times, with the exception of the sample irradiated for 24
hours with UVC. This sample demonstrated higher carbonyl index at the start of
aging and at embrittlement, indicating a higher concentration of oxidation
products, but not a greater extent of chain scission reactions.
109
3.10.3 Degussa P25, oven aged samples
3.10.3.1 Effect of pre-irradiation – UVC
0 50 100 150 200 250 300 350 400 4500.00
0.05
0.10
0.15
0.20
0.25
0.30 0 secs 60 secs 3 hrs 24 hrs
Car
bony
l ind
ex
Days aged in oven
Figure 3-21 Carbonyl index plots for UVC pre-irradiated LLDPE film containing 1%
Degussa P25 titania and aged in the oven. Second order polynomial trend lines have been
added.
The samples pre-irradiated for 24 hours and 3 hours reached embrittlement in the
oven at 24 days and 392 days respectively. It appears as though the 24 hour
irradiated sample was already close to embrittlement before oven aging. This is
confirmed by comparison with the carbonyl index plot of the weatherometer
aged sample shown in Figure 3-19, where the 24 hour UVC treated polymer
embrittled after just 3 days.
110
3.10.3.2 Effect of pre-irradiation – UVA
0 50 100 150 200 250 300 350 4000.00
0.02
0.04
0.06
0.08
0.10 0 secs 60 secs 3 hrs 24 hrs
Car
bony
l ind
ex
Days aged in oven
Figure 3-22 Carbonyl index plots for UVA pre-irradiated LLDPE film containing 1% Degussa P25 titania and aged in the oven. Second order polynomial trend lines have been added.
Figure 3-22 shows that pre-irradiation with UVA has been much less effective
than UVC in accelerating degradation in the oven. Pre-irradiation for 60 secs
with UVA has almost no effect on the rate of degradation of the polymer. Even
those samples exposed to higher doses of UVA are not strongly deviating from
the curve for the non-irradiated sample, and after nearly 400 days of oven aging
these samples only show a relatively moderate carbonyl index.
3.10.4 3% Degussa P25 samples
Increasing the titania loading from 1% to 3% affected the degradation of the
polyethylene film to an extent. The weatherometer aged samples generally
embrittled earlier, taking around 6-9 days. Furthermore, in some cases the
carbonyl index actually starts to drop away, indicating that some of the
111
oxygenated functional products have degraded even further to form volatiles
such as CO2, H2O and small organic molecules.
-1 0 1 2 3 4 5 6 7 8 9 10 11 12 13-0.04-0.020.000.020.040.060.080.100.120.140.160.180.200.220.240.260.280.300.320.34
Car
bony
l ind
ex
Days aged in weatherometer
0 secs UVC 60 secs UVC 3 hrs UVC 24 hrs UVC 0 secs UVA 60 secs UVA 3 hrs UVA 24 hrs UVA
Figure 3-23 Carbonyl index plots for UVC and UVA pre-irradiated LLDPE film containing 3% Degussa P25 titania and aged in the weatherometer. Second order polynomial trend lines have been added. The data for samples aged for 60 secs UVC and 24 hours UVC contain outliers.
With the exception of the 24 hr pre-irradiated sample, the samples pre-treated
with UVC and aged in the oven did not demonstrate significantly different
lifetimes (Figure 3-23). Figure 3-24 shows that the carbonyl index of these
samples is not greatly different at embrittlement. The lifetime of the 24 hr pre-
irradiated sample is shortened by about 150 days; however the slope of the
carbonyl index plot is quite similar to the lesser pre-irradiated samples.
112
0 50 100 150 200 250 300 350 4000.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
C6 B C7 C14 C11 C12 C9 C13
Car
bony
l ind
ex
Days aged in oven
Figure 3-24 Carbonyl index plots for UVC and UVA pre-irradiated LLDPE film containing 3% Degussa P25 titania and aged in the oven. Linear or second order polynomial trend lines have been added.
The samples pre-irradiated under UVA and oven aged produced some different
results. The sample exposed to 24 hours of UVA actually embrittled after just
180 days of aging, and at a very high carbonyl index of over 0.35. Figure 3-24
shows that these samples have a much higher carbonyl index then their 1%
Degussa P25 counterparts after a similar length of time spent in the oven.
Exposure to 3 hrs or less UVA did not seem to greatly affect the degradation of
this material.
The higher carbonyl index of the oven aged samples at embrittlement, or after
long periods of time in the oven, indicate that there is more oxidation occurring
in these samples compared with the weatherometer aged samples. The creation of
reactive sites in the polymer films due to pre-irradiation by UV proceed to be
oxidised further in the oven, and can eventually result in embrittlement. By
contrast, weatherometer aged samples embrittle earlier and at lower carbonyl
113
indexes, demonstrating the greater effect of non-oxidation related processes, and
the burning of the material by titania.
In addition, the higher carbonyl index of the 24 hour UVA and UVC pre-
irradiated samples was much more pronounced in the 3% Degussa P25 samples
then the 1% Degussa P25 samples. This phenomenon was not found in the
weatherometer aged samples. It suggests that at higher concentration of Degussa
P25, the wavelength of pre-irradiation becomes less relevant, as there is enough
titania to form a sufficient quantity of reactive sites that can induce more rapid
polymer aging.
The reactions provided in Scheme 1-21 describe how titania can produce carbon
centered radicals. Macroradicals can then proceed to crosslink the material
(Section 1.1.1). The information contained in the carbonyl plots seen thus far
indicate that this is occurring in the LLDPE, specifically that the titania is
absorbing UV irradiation to form a charge separated species, which is giving rise
to macroradicals, resulting in degradation processes such as crosslinking.
3.10.5 Section summary
Pre-irradiation of polymer film containing 1% Degussa P25 titania with UV light
results in faster degradation than non-pre-irradiated samples. UVC has a much
more significant effect on this material than UVA. Despite the faster degradation,
the polymers are still taking a long time to embrittle in the oven – a film exposed
to 3 hours of UVC took over 1 year to embrittle.
The 3% Degussa P25 film behaves differently to the 1% film, with the samples
in the weatherometer tending to embrittle earlier. Three hours or less dosage with
UVC or UVA did not have a significant effect on the degradation of these films,
although 24 hours of pre-irradiation greatly shortened the lifetime in both cases.
Titania is having a significant effect on the polymer by introducing reactive sites
in the polymer chains, which proceed to react in a dark environment. With
increased concentrations of Degussa P25, the wavelength of pre-irradiation light
114
becomes less relevant, provided that the polymer has been irradiated for a
sufficiently long period of time.
3.11 LLDPE containing Kronos 1002 (degraded)
The films containing Kronos 1002 titania degraded much more slowly than the
Degussa films. Overall the titania appeared to be much less active. Only heavy
doses of irradiation affected the degradation outcomes.
3.11.1 1% Kronos 1002, weatherometer aged samples,
0 5 10 15 20 25 30 35 400.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
Car
bony
l ind
ex
Days aged in weatherometer
0 secs UVC 60 secs UVC 3 hrs UVC 24 hrs UVC 0 secs UVA 60 secs UVA 3 hrs UVA 24 hrs UVA
Figure 3-25 Carbonyl index plots for UVC and UVA pre-irradiated LLDPE film containing
1% Kronos titania and aged in the weatherometer. Polynomial trend lines have been added.
Figure 3-25 shows the carbonyl index plots for all LLDPE films containing 1%
Kronos titania and aged in the weatherometer, UVA and UVC pre-irradiated
combined. The films generally degraded in about half of the time of the control
sample, showing that titania has some effect on the degradation.
115
3.11.2 3% Kronos 1002, weatherometer aged samples,
0 5 10 15 20 25 300.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
Car
bony
l ind
ex
Days aged in weatherometer
0 secs UVC 60 secs UVC 3 hrs UVC 24 hrs UVC 0 secs UVA 60 secs UVA 3 hrs UVA 24 hrs UVA
Figure 3-26 Carbonyl index plots for UVC and UVA pre-irradiated LLDPE film containing
3% Kronos titania and aged in the weatherometer. Second order polynomial trend lines
have been added.
Increasing the concentration of Kronos 1002 titania from 1% to 3% resulted in
shorter embrittlement times in the weatherometer. Overall, the samples degraded
between 15 and 25 days, and the carbonyl index at embrittlement was higher than
the 1% samples. Pre-irradiation had less of an effect on the 3% loading film.
116
3.11.3 1% Kronos 1002, oven aged samples,
0 100 200 300 4000.0
0.1
0.2
0.3
0.4
0.5
Car
bony
l ind
ex
Days aged in oven
0 secs UVC 60 secs UVC 3 hrs UVC 24 hrs UVC 0 secs UVA 60 secs UVA 3 hrs UVA 24 hrs UVA
Figure 3-27 Carbonyl index plots for UVC and UVA pre-irradiated LLDPE film containing
1% Kronos titania and aged in the oven. Second order polynomial trend lines have been
added.
The lack of impact that pre-irradiation had on these samples is show in Figure
3-27. As has been the trend with all samples, the 24 hr UVC pre-irradiated
sample showed a significantly different carbonyl index plot, although even this
sample took 330 days to embrittle in the oven. None of the other samples had
achieved embrittlement at the conclusion of the experiment.
117
3.11.4 3% Kronos 1002, oven aged samples,
0 100 200 300 4000.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
Car
bony
l ind
ex
Days aged in oven
0 secs UVC 60 secs UVC 3 hrs UVC 24 hrs UVC 0 secs UVA 60 secs UVA 3 hrs UVA 24 hrs UVA
Figure 3-28 Carbonyl index plots for UVC and UVA pre-irradiated LLDPE film containing 3% Kronos titania and aged in the oven. Second order polynomial trend lines have been added.
The carbonyl index plots of the oven aged 3% loading Kronos 1002 film look
very similar to those of the 1% film show in Figure 3-27. Apart from the 24 hr
UVC pre-irradiated sample, none of the films had achieved embrittlement.
3.11.5 Section Summary
The Kronos 1002 titania appears to be relatively inactive in polyethylene film
with respect to UV treatment. Pre-irradiation has not had a significant impact,
except for the sample treated for 24 hrs with UVC. The strong photosensitising
effect of titania when irradiated with large doses of UVC it’s a common trend in
all samples studied. The Kronos samples took a considerable time to age in the
weatherometer, and the oven aged samples also degraded slowly. The carbonyl
index plots of the oven aged samples are similar to those of the control,
indicating that pre-irradiation did not have a strong effect on the outcome of
118
oxidation. Increasing the loading from 1% to 3% resulted in somewhat shortened
lifetimes in the weatherometer, although there was little effect on the oven aged
samples.
3.12 LLDPE containing Huntsman Tioxide (degraded)
The films containing Huntsman Tioxide titania were reasonably sensitive to UV,
degrading in about 15 days in the weatherometer. UV pre-irradiation also
affected the rate of degradation of film aged in the oven.
3.12.1 3% Huntsman tioxide A-HR, weatherometer aged
This Huntsman tioxide film contained titania that was 100% anatase, and water
dispersible. It was more active than the 100% anatase Kronos films. Figure 3-29
shows the weatherometer aging results. Films tended to break down after about
15 days. Pre-irradiation did not have a significant impact on carbonyl index, with
the exception of the 24 hour UVC pre-irradiated sample.
119
0 2 4 6 8 10 12 14 16 18 20 220.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16C
arbo
nyl i
ndex
Days aged in weatherometer
0 secs UVC 60 secs UVC 3 hrs UVC 24 hrs UVC 0 secs UVA 60 secs UVA 3 hrs UVA 24 hrs UVA
Figure 3-29 Carbonyl index plots for UVC and UVA pre-irradiated LLDPE film containing 3% Huntsman Tioxide A-HR and aged in the weatherometer. Second order polynomial trend lines have been added. The sample exposed to 60 secs of UVC gave an apparently anomalous result,
taking significantly longer to embrittle. There is no obvious reason for this, and it
is postulated that this inconsistency was due to a factor not controlled in this
experiment, such as perhaps variable film thickness, higher antioxidant
concentration, heterogeneous titania dispersion or human error. When working
with real-world samples, such as these films are, anomalous or unusual results
can appear quite regularly.
120
3.12.2 3% Huntsman tioxide A-HRF, weatherometer aged
0 2 4 6 8 10 12 14 16 18 200.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
Car
bony
l ind
ex
Days aged in weatherometer
0 secs UVC 60 secs UVC 3 hrs UVC 24 hrs UVC 0 secs UVA 60 secs UVA 3 hrs UVA 24 hrs UVA
Figure 3-30 Carbonyl index plots for UVC and UVA pre-irradiated LLDPE film containing 3% Huntsman Tioxide A-HRF and aged in the weatherometer. Second order polynomial trend lines have been added.
The Huntsman tioxide A-HRF samples behaved similarly to the A-HR equivalent
in the weatherometer. The samples pre-irradiated with UVA did not degrade
faster than the untreated samples, and the 24 hour UVA treated sample took the
longest time to achieve embrittlement, with lower carbonyl index measurements.
UVC pre-irradiation did not lead to significantly different carbonyl plots to UVA
pre-irradiation, a difference some other films have shown.
121
3.12.3 3% Huntsman tioxide A-HR, oven aged
0 100 200 300 4000.000.020.040.060.080.100.120.140.160.180.200.220.240.260.280.30
Car
bony
l ind
ex
Days aged in oven
0 secs UVC 60 secs UVC 3 hrs UVC 24 hrs UVC 0 secs UVA 60 secs UVA 3 hrs UVA 24 hrs UVA
Figure 3-31 Carbonyl index plots for UVC and UVA pre-irradiated LLDPE film containing 3% Huntsman Tioxide A-HR and aged in the oven. Second order polynomial trend lines have been added. Pre-irradiation of LLDPE containing 3% Huntsman tioxide A-HR with UVA or
UVC did not have a strong effect on the outcome of oven aging, with the
exception of the sample pre-irradiated for 24 hours of UVC. Again, it can be seen
that this treatment causes earlier embrittlement times, and a higher carbonyl
index at embrittlement, however the sample still took 250 days to embrittle.
122
3.12.4 3% Huntsman tioxide A-HRF, oven aged
0 100 200 300 4000.000.020.040.060.080.100.120.140.160.180.200.220.240.260.280.30
Car
bony
l ind
ex
Days aged in oven
0 secs UVC 60 secs UVC 3 hrs UVC 24 hrs UVC 0 secs UVA 60 secs UVA 3 hrs UVA 24 hrs UVA
Figure 3-32 Carbonyl index plots for UVC and UVA pre-irradiated LLDPE film containing 3% Huntsman Tioxide A-HRF and aged in the oven. Second order polynomial trend lines have been added.
As with the weatherometer aged samples, pre-irradiation does not appear to have
a significant impact on the degradation of the polyethylene film containing 3%
Huntsman tioxide A-HRF and aged in the oven. The exception to this is the 24 hr
UVC pre-irradiated sample.
3.12.5 Section summary
The polyethylene films containing Huntsman tioxide titania demonstrated
increased sensitivity to UV radiation than the control sample, although not as
much as the Degussa P25 films. The samples degraded in about 15 days in the
weatherometer, and did not demonstrate significant differences according to UV
pre-treatment.
Pre-irradiation had little effect on the oven aging of these films. The A-HR
material showed increased carbonyl index measurements during oven aging with
pre-irradiation, but this effect could not be observed in the A-HRF material. The
123
material did not always behave consistently, possibly due to manufacture
disparities such as titania distribution and film thickness.
3.13 LLDPE containing Sachtleben Hombitan (degraded)
Generally, the Sachtleben Hombitan films were among the least responsive to
UV treatment of all the films containing titania studied. Weatherometer aged
samples often took over 20 days to embrittle, and up to 45 days. Only those
samples pre-irradiated with 24 hrs of UVC had achieved embrittlement in the
oven. Of the 4 films containing Sachtleben Hombitan titania, the LLDPE film
containing LW-S-12 (organic coating on anatase microcrystal) was the most
active.
3.13.1 3% Sachtleben Hombitan, weatherometer aged
Pre-irradiation had very little effect on samples aged in the weatherometer.
Figure 3-33 shows the carbonyl index plots for the film containing antinomy
doped titania particles. Higher doses of UV pre-treatment resulted in slightly
shortened lifetimes in the weatherometer.
124
0 5 10 15 20 25 30 350.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
Car
bony
l ind
ex
Days aged in weatherometer
0 secs UVC 60 secs UVC 3 hrs UVC 24 hrs UVC 0 secs UVA 60 secs UVA 3 hrs UVA 24 hrs UVA
Figure 3-33 Carbonyl index plots for UVC and UVA pre-irradiated LLDPE film containing 3% Sachtleben Hombitan LW-S-U titania and aged in the weatherometer. Second order polynomial trend lines have been added.
The LW-S-U film (Figure 3-33) demonstrated longer embrittlement times than
the LW-S-12 film (Figure 3-34). However neither samples showed a significant
response to pre-irradiation.
125
0 2 4 6 8 10 12 14 16 18 20 220.00
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
0.18
Car
bony
l ind
ex
Days aged in weatherometer
0 secs UVC 60 secs UVC 3 hrs UVC 24 hrs UVC 0 secs UVA 60 secs UVA 3 hrs UVA 24 hrs UVA
Figure 3-34 Carbonyl index plots for UVC and UVA pre-irradiated LLDPE film containing 3% Sachtleben Hombitan LW-S-12 titania and aged in the weatherometer. Second order polynomial trend lines have been added.
3.13.2 3% Sachtleben Hombitan, oven aged
The oven aged samples proved to be similar to the results seen so far. Pre-
irradiation with 24 hours of UVC shortened the embrittlement time, however
even these samples generally took over 300 days to achieve embrittlement.
126
0 50 100 150 200 250 300 350 400 4500.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
Car
bony
l ind
ex
Days aged in oven
0 secs UVC 60 secs UVC 3 hrs UVC 24 hrs UVC 0 secs UVA 60 secs UVA 3 hrs UVA 24 hrs UVA
Figure 3-35 Carbonyl index plots for UVC and UVA pre-irradiated LLDPE film containing 3% Sachtleben Hombitan LC-S titania and aged in the weatherometer. Second order polynomial trend lines have been added.
Figure 3-35 shows the carbonyl index plots of oven aged samples. Samples
which received minimal pre-irradiation appear to be degrading very slowly, and
extrapolation from the plots indicates they would take years to embrittle. Heavy
pre-irradiation still resulted in an embrittlement time of just less than 400 days.
The LW-S-12 film gave improved results, with the 24 hr UVC pre-treated
sample achieving embrittlement in 75 days. The remainder of the samples of this
film appear to be following the general pattern. For example, the sample pre-
irradiated for 60 secs with UVC does not appear close to embrittlement after over
450 days in the oven.
3.13.3 Section summary
The Sachtleben Hombitan films proved resistant to UV pre-irradiation effects.
They took longer than other films to embrittle in the weatherometer, and after
over 400 days spent in the oven many films still appear intact, with low carbonyl
127
indexes. The most active of these films was the LW-S-12 film; however this film
was considerably less responsive to UV treatment than the polyethylene film
containing Degussa P25 titania.
3.14 Discussion of the effects of titania
The results provided in this chapter have demonstrated the effect of UV pre-
irradiation of LLDPE film, with and without titania, on the relative rates of
degradation. UV pre-irradiation decreased the time taken to embrittle, for both
weatherometer and oven aged samples. SEM images showed that the Degussa
P25 titania exhibited poor dispersion, and a tendency to agglomerate into large
particles. Other types of titania were much more evenly distributed, and had a
narrower size distribution also.
All LLDPE samples containing titania degraded faster than the control sample,
although oven aged samples were difficult to gauge precisely as mostly they had
not degraded after over 200 days spent aging. Thus all modified and unmodified
titanias used in this study exhibit prodegradant qualities. Of these types of titania,
Degussa P25 demonstrated the most reactivity, and samples containing Degussa
P25 degraded in ca. 200 days in the dark.
The effect of rutile acting as a dopant to assist in electron-hole separation was
discussed in Section 1.2.4. The results from this large study demonstrate the
relative strength of Degussa P25 as a photocatalyst, and therefore there is little
doubt as to the importance of electron-hole separation when identifying a strong
photocatalyst. Additionally, the Degussa P25 titania was not surface modified,
and thus there exists a large amount of surface area for oxygen adsorption. The
factors affecting titania photoactivity were listed in Section 1.2.3, namely particle
size, crystalline structure, phase composition, surface area, nature and
concentration of lattice defects, surface hydroxyl groups, and impurities. Degussa
P25 exhibits small particle size, high surface area (and thus high availability of
surface defects to oxygen), surface hydroxyl groups and low impurities.
128
Kronos titania does not contain a mixed crystal phase, being 100% anatase.
Kronos exhibited the lowest photoactivity, and it is thought that this was due to
the higher chance of electron-hole recombination. The combination of electron-
hole recombination and the lower surface area, leading to reduced availability of
surface defects, resulted in Kronos displaying the lowest photoactivity of the
titanias investigated in this study.
The Huntsman Tioxide samples were coated with an organic modifier to improve
dispersibility. However these coatings appear to have lowered the available
surface area, reducing the photoactive potential; likewise for the Satchleben
Hombitan coated samples, which also showed poor photoactivity. The antimony-
doped titania performed poorly; however this could be due the doping effect of
antimony, which has not been reported elsewhere in the literature. Overall, these
results indicate that the available surface of titania, particle size (regardless of
agglomeration when distributed in an organic phase) and the crystal phase are the
dominant factors when considering the potential of titania as a photocatalyst.
Increasing the titania loading from 1% to 3% had little significant affect on the
relative rates of degradation of the LLDPE films. It is likely that this has arisen
due to agglomeration of the particles, reducing the available surface area. This is
supported by the SEM images presented in Section 3.2, which show larger titania
particle sizes in the 3% loading samples compared to the 1% loadings for
Degussa P25 and Kronos films.
The data presented in this chapter demonstrate the effectiveness of pre-irradiation
to accelerate degradation in LLDPE film, even in the dark. By exposing the
sample containing 3% Degussa P25 to 24 hours of UVA or UVC, embrittlement
was achieved in approximately 200 days. This result indicates that a significant
concentration of highly photoactive titania, pre-irradiation with a large dose of
UV radiation and higher temperatures are required to greatly accelerate the
oxidation of LLDPE film. The continuing challenge for research into this
technology will be likely to be focussed on methods to reduce the loading of
titania, and improving particle dispersion in addition to utilising a smaller
particle size with a narrower distribution. Additionally, it would be desirable to
129
reduce the intensity of UV pre-irradiation by increasing the photosensitivity of
the film.
It can be seen from the carbonyl index plots of the control sample aged in the
weatherometer that UVC pre-irradiation not only reduces the time taken to
embrittlement, but the carbonyl absorption is not as intense also. This effect was
also found in LLDPE containing titania. However, this was reversed for the oven
aged samples – samples that had undergone 24 hours of UVC irradiation
degraded with a much higher carbonyl index.
This phenomenon is explained when considering the effect of UV irradiation on
polyethylene. As was written in Section 1.1.3, UV irradiation induces
crosslinking and unsaturation. Titania photoreactions also produce macroradicals
that can result in similar products (Scheme 1-20, Scheme 1-21). Oxidation
reactions are initiated at these reactive sites that have now become sensitive to
oxidation, and proceed to spread throughout the bulk (Section 1.1.2).
The more active the titania is, the more reactive sites are produced. In the case of
weatherometer aging, degradation processes not showing an oxygenated
functional group signature (e.g. crosslinking and cavity formation) dominate the
degradation process, reducing the carbonyl index at embrittlement. However for
oven aged samples, the reactive sites are attacked by oxygen to produce
oxidation products containing a carbonyl functional group. This is in agreement
with the literature discussed in Section 1.1.4.
The similarity of the relative carbonyl absorption intensities in all samples
strongly indicates that the reaction pathway is not affected by titania. Hence
titania is considered to be catalysing degradation of the film by increasing the
number of radicals available for reaction. Photosensitisation products such as
crosslinks and unsaturation are in higher concentration, resulting in a more rapid
rate of oxidation.
Allen had described a stabilising effect of pigment-grade titania particles on
organic material when irradiated with UVC (Section 1.2.7). This effect was not
130
found in the course of these experiments, including samples containing pigment
grade titania, modified titania or unmodified nano-titania. It is possible that this
effect was unique to the sample set that was investigated in that experiment;
however those results should not be extrapolated into other systems. It is
concluded from the experiments conducted in this thesis that titania has a
sensitising, not stabilising, effect on polyolefins irradiated with UVC.
3.15 Conclusions
The activity of the different titanias as a prodegradant in polyethylene is
summarised in Table 16.
131
Table 16 Summary of titania activity.
Titania Oven
lifetime
Weatherometer
lifetime (max)
Effect of pre-
irradiation Degussa P25 ~200 days 12 days Greatly reduced
lifetime in
weatherometer and
oven with large
doses
Huntsman
Tioxide A-
HR/F
>300 days 15 days Oven samples
showed higher deg.
rate
Sachtleben
Hombitan
LW-S-12
>300 days 18 days Oven samples
showed higher deg.
rate
Sachtleben
Hombitan
LW-SU/HD,
LC-S
>300 days 30 days Oven samples
showed higher deg.
rate
Kronos 1002 >300 days 30 days Reduced effect
Most active
Least Active
The carbonyl index plots and comparisons of times taken to embrittlement have
revealed that Degussa P25 is the most photoactive titania in LLDPE.
Furthermore, the most important titania characteristics that determine the
photoactivity of a titania particle are available surface area for oxygen adsorption
and crystal phase. Titania particles are thought to create photosensitised regions,
which result in faster rates of degradation. Pre-irradiation with 24 hours of UV
can result in a plastic film that will degrade in around 200 days in the dark at
132
50 °C. This result demonstrates potential for technology involving pre-
irradiation.
In order to develop the technology required to create a LLDPE film with more
accurately controllable degradation properties, more information regarding the
chemistry of the degradation processes occurring is required. During the course
of these experiments a great deal of spectra data have been acquired. These data
have been subjected to statistical analysis in order to glean as much information
as possible from this unique data set.
133
134
Multivariate Data Analysis
4.1 Introduction
In Chapter 3 it was seen that a large body of data had been collected comparing
the degradation of LLDPE containing nano-titania particles from different
manufacturers, under different degradation environments, and with different
kinds of pre-treatment. While techniques such as spectral subtraction and
carbonyl index plots can elucidate some information regarding degradation
processes, standard ‘data mining’ methods such as these struggle to cope with the
size and complexity of a data set like the one in this study.
Multivariate data analysis (also known as Chemometrics when specifically
applied to chemical data) allows the user to highlight aspects of data that are
varying in relation to each other. It is especially applicable to large data sets,
such as a series of measurements taken over time, and has the potential to extract
information that is difficult to see with the unaided eye.
This is not to suggest that multivariate data analysis techniques can see
something that is not there. Rather, they allow data to be presented in such as
way that only relevant information is examined. In the analogy of ‘mining’ data
for information, chemometrics is an excavator and sorter combined.
It is hoped that the spectral data collected on the photooxidation of LLDPE
containing titania holds information that may be exploited to produce a polymer
with controllable degradation properties. A chemometric investigation is the
most likely method of data mining to discover this information, and expand the
knowledge base regarding the degradation of polyethylene photosensitised with
titania.
135
4.2 Data treatment
Principal Component Analysis (PCA) is a multivariate data analysis technique
that decomposes data into one or more components. These components describe
the variations that are happening in the data set. The first component, Principal
Component 1 (PC1) describes the most significant variance, PC2 describes the
second most significant variance, and so forth. Eventually, a PC and all
subsequent PCs will describe mostly noise. PCs from this point are not
considered significant.
The data must be pretreated in order to remove artifacts such as signal strength
variations, baseline differences, or other variations that can affect the outcome of
a PCA model. Firstly, the data are normalised, by giving the total area of each
spectrum the same value (arbitrarily set to 1). This removes differences in signal
strength due to artifacts such as poor contact with the internal reflection element
when performing ATR/FTIR, possible light scattering or slight variations in
refractive index, etc.. Secondly, the data are mean centered, which creates an
average spectrum, and the variables (absorption at a given wavenumber) are
described in terms of negative or positive variance from the mean. If the data are
not mean-centered, the first PC does not describe variance, but actually shows an
average spectrum, and describes how far other spectra are from this average. As
our interest is not the mean, but variations around the mean, the data are mean-
centered first.
Data that do not contain information relevant to the investigation were removed.
The spectral ranges under investigation varied between models, however in all
cases data above 1900 cm-1 were removed prior to any statistical analysis, as at
low levels of oxidation relevant information is difficult to distinguish from signal
noise.
A PCA calculation on a data set is termed a model, and the model needs to be
validated with external data to determine its reliability. Cross-validation methods
were used, in which one spectrum (a data subset) is removed from the data set
(test set) to leave the data set minus the subset (model building set), and the
136
model was recalculated. The spectrum removed was then predicted by the model.
Reiterations were performed until every spectrum had been left out. Two cross-
validation methods were used according to the number of spectra; if there were
less than 20 spectra, then the leave-one-out method was used. In the case of 20 or
more spectra, a “venetian blind” model was used, in which regularly spaced
spectra (eg 3rd, 4th, 5th etc) from the test set was used in the model building set
while the remaining spectra were used as a validation set. Venetian blind cross
validation uses less computing power and less time to calculate the model, as
leave-one-out cross validation is inappropriate for large data sets due to the
complexity of the calculation.
4.3 Analysis of samples subjected to oven aging
4.3.1 Samples without pre-irradiation
It is necessary to investigate any effects of titania upon degradation processes in
polyethylene in the absence of irradiation in order to distinguish these
phenomena from photoactivated effects. The sample containing 3% Degussa P25
demonstrated much higher carbonyl intensity than the control after similar times
of aging in the oven at 50 ºC without either sample being subjected to any
irradiation (Section 3.10.4). This implied that titania is increasing the rate of
oxidation, even in the absence of light.
PCA was used to make a comparison between the control sample, and the sample
containing 3% Degussa P25 titania. Both samples were oven aged and did not
undergo any UV pre-irradiation.
137
-0.015 -0.01 -0.005 0 0.005 0.01 0.015 0.02-5
-4
-3
-2
-1
0
1
2
3
4
5x 10-3
Scores on PC 1 (75.91%)
Sco
res
on P
C 3
(5.9
1%)
00
07
17
24
35
54
80
115
150
180
217
244
286 327
375
18 33
61
94
136 177
225
266
Figure 4-1 Scores plot for control film (green star) and film containing 3% Degussa P25 (red triangle). Both films were aged in the oven and were not pre-irradiated. The numbers next to each point represent the days spent aging in the oven.
900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
Wavenumber
Load
ings
on
PC
1 (7
5.91
%),
Load
ings
on
PC
3 (5
.91%
)
Figure 4-2 Loadings plots for PC1 (blue) and PC3 (green) from Figure 4-1.
138
Figure 4-1 shows a plot of the scores for PC1 vs. PC3. PC1 describes the general
trend of carbonyl absorption development with time, and the inverted peaks of
the processing agent spectrum demonstrate that the processing agent is actually
disappearing from the spectrum. (An absorption assignment table for the
processing agent is provided in Section 3.6). Not only is the processing agent
disappearing with more time spent aging in the oven, but the PCA model gives
some indication as to whether the processing agent is being oxidised, or simply
being lost as a volatile compound.
Comparing the relative scores of the control sample after around 200 days of
aging with that of the sample containing 3% Degussa P25 titania after a similar
length of aging, it is seen that they have very different scores on PC1. The
loadings plot of PC1 is describing variances occurring mostly to the processing
agent spectrum. The processing agent absorptions at 1080 cm-1 and around 1375
cm-1 are both being lost as the carbonyl absorption increases in intensity. After
200 days spent aging in the oven, the control samples still have a negative score
on PC1, and therefore still contain some amount of processing agent. The heavily
oxidised LLDPE film containing 3% Degussa P25 titania, however loses its
absorptions due to the processing agent after a similar period of time spent in the
oven. Thus we can see that the processing agent is being lost from the more
heavily oxidised sample, and it is inferred that this represents an oxidation of the
processing agent.
PC3 appears to be describing the difference in degradation processes between
that of the processing agent and the polymer. Consider that PC1 tells us that as
the carbonyl absorption increases, the absorption peaks of the processing agent
decrease. However, PC3 shows us that the methyl absorptions at 1390 cm-1 and
1361 cm-1 of the processing agent are actually increasing with carbonyl
absorption. In addition, there is a new absorption at 1100 cm-1 that decreases
with increasing carbonyl, and an absorption band at 1280 cm-1 which is also
decreasing with increasing carbonyl. Both of these bands are indicative of a low
molecular weight ester C-O stretc.h, such as formate or acetate ester.
139
Due to the difficulties in determining the exact nature of the processing agent,
interpretation of these data is not straightforward. However, it appears that the
processing agent has degraded to form a low molecular weight oxidation product,
which may have then further oxidised to form volatile compounds which do not
appear in the spectra. The processing agent has disappeared from the spectra of
the heavily oxidised sample containing titania faster than from the control sample.
The data do not only yield information about the processing agent, however, but
also about the LLDPE.
The most interesting information to come from this investigation is the lack of
separation between the control and the sample containing titania. Processing
agent chemistry aside, the two materials appear to be degrading by the same
process. This has been established by the calculation of many PCA models of
different systems comparing non pre-irradiated, oven aged LLDPE samples.
Rather than present pages of examples showing a lack of separation according to
degradation chemistry between the control and samples containing titania, an
example of the degradation chemistry of the processing agent has been provided
here to demonstrate the information obtained by application of PCA.
When examining scores plots for separation, one is looking for grouping of the
data. If groups of data can be identified, then the loadings plots will reveal on
what basis they are separated. In this instance PC1 shows significant separation
of the control and titania-containing sample, however PC1 is describing the
general extent of degradation, and shows that the titania-containing samples are
simply more degraded. By comparing the value of the PC1 score and the
degradation time, it is clear that the effect of the titania is to speed up the
degradation process considerably, however the lack of separation on the PC3 axis
tends to indicate that the degradation processes are similar.
PC2 contained information not relevant to degradation chemistry, but showed
fluctuations in the strength of the additive absorption and did not separate the
data, while PC3 did not separate the control sample or titania containing sample
either. It is apparent then that in the absence of irradiation, the control sample is
140
degrading in a similar manner to the titania-containing sample, albeit more
slowly.
4.3.2 Samples with pre-irradiation
Multivariate statistical analysis of the spectral data obtained from the oven aging
of the UV pre-irradiated samples containing 3% Degussa P25 revealed
interesting chemistry regarding the effect of UV irradiation. A comparison was
made using the carbonyl region of the spectra of the samples that were non-
irradiated, 3 hours of UVC pre-irradiated, and 24 hours of UVA irradiated. (24
hours of UVC pre-irradiation was found to be too harsh to make accurate
comparisons. 3 hours of UVC pre-irradiation has a more similar effect to 24
hours of UVA pre-irradiation).
5 10 15 20 25 30-0.06
-0.04
-0.02
0
0.02
0.04
0.06
Sample
Sco
res
on P
C 1
(91.
43%
)
00 07
17 24
35
54 80 12
24 61
95
32
70 70 98
115
150
180
217
244
286 327 375
123
162
199
241
282
330
371
140
181
Figure 4-3 Scores plot for PC1 comparing 3% Degussa P25 containing film non-irradiated (red triangle), 24 hours UVA pre- irradiated (green star) and 3 hours UVC pre- irradiated (black circle). Samples were aged in an oven. The label ‘Sample’ on the x-axis refers to the spectrum number in the series of spectra.
141
1680 1700 1720 1740 1760 1780 1800 1820 1840 1860 1880-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
0.35
Wavenumber
Load
ings
on
PC
1 (9
1.43
%)
Figure 4-4 PC1 for Figure 4-3. This PC describes the general variances seen in the spectra during the process of degradation. .
PC1 shows the trend of increased carbonyl absorption with time spent in the
oven. This PC describes over 90% of the total variance in the carbonyl region.
PC2 and PC3 describe much less, and are discussed below.
142
-0.015 -0.01 -0.005 0 0.005 0.01 0.015-12
-10
-8
-6
-4
-2
0
2
4
6
8x 10-3
Scores on PC 2 (4.97%)
Sco
res
on P
C 3
(1.8
4%)
00
07
17
24
35
54
80
12
24 61
95
70 98
150
180
217 244
327
375
123
162 199
241 282 330 371
140 181
Figure 4-5 Scores plot for PC2 and PC3 comparing 3% Degussa P25 containing film non-irradiated (red triangle), 24 hours UVA pre-irradiated (green star) and 3 hours UVC pre-irradiated (black circle).
1680 1700 1720 1740 1760 1780 1800 1820 1840 1860 1880-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
Wavenumber
Load
ings
on
PC
2 (5
.39%
), Lo
adin
gs o
n P
C 3
(1.8
7%)
Figure 4-6 PC2 (blue) and PC3 (green) for Figure 4-5.
143
The loadings plots show that PC2 describes the absence of the carbonyl
absorption at 1713 cm-1 and the presence of a carbonyl absorption at 1732 cm-1.
Referring to Section 3.7.2, these are assigned to ketones/carboxylic acids, and
esters/ketones respectively. PC3 shows another ester absorption at higher
wavenumbers, around 1740-45 cm-1, and also some absorptions around 1860 cm-
1. These latter absorptions are probably due to less common carbonyl containing
degradation products involving cyclic structures, anhydrides, lactones and
peracids
Figure 4-5 shows that the samples subjected to pre-irradiation score more highly
on PC3 than do the non-irradiated samples. Thus we know that these samples
contain more of the lesser absorbing oxygenated functions such as anhydrides,
lactones, etc. which have absorptions above 1800 cm-1. The samples pre-
irradiated with UVA score more highly on PC2, suggesting less carbonyl at 1713
cm-1. The other samples seem to be spread over this PC.
The trends suggested by this analysis are that pre-irradiation of samples
containing titania results in slightly higher ester, lactone and anhydride
concentrations. This confirms Tidjani’s degradation pathway schematic in
Section 1.1.4. In this diagram we see that degradation involving UV results in
various products containing carbonyl functionality which can absorb at higher
wavenumbers, while the thermal aging route to the left side of the diagram gives
rise to acids, which typically absorb around
1713 cm-1, as seen in this multivariate statistical analysis.
4.3.3 UVA vs UVC pre-irradiation: extent of degradation
PCA information can be used to obtain plots describing the extent of degradation
similar to the carbonyl plots in Chapter 3. An extent of degradation plot using
multivariate statistical information has distinct advantages over a conventional
plot obtained by plotting the area under the carbonyl absorption. When using a
chemometric approach, a much wider spectral range can be used, and selected
absorptions that do not contain relevant information can be omitted. This is
144
compared to the carbonyl index plots obtained by measuring the area under the
carbonyl peak, such as those presented in Chapter 3.
Additionally, if there is more than one influence leading to variances in a
spectrum over time, then the PC that describes the variances occurring only from
degradation can be separated from the rest of the spectral information and used to
obtain an extent of degradation plot. This is relevant when studying real world
materials, where we have already seen the effect of processing agent on the
spectrum of the materials being investigated. Finally, a PCA calculation requires
only a few minutes to set up, and seconds to calculate. Compare this with
conventional carbonyl plots, where depending on the number of spectra, it takes
hours of measuring and calculation to obtain the area under the carbonyl peak
and plot the results.
Spectral data obtained from the 3% Schatleben Hombitan LW-S-U-HD (anatase
crystal with organic coating) titania in LLDPE film aged in an oven at 50 ºC was
subjected to a PCA investigation. Two samples were used: one was pre-irradiated
with 24 hours of UVC, and the other pre-irradiated with 24 hours of UVA. The
PCA results are presented below.
145
2 4 6 8 10 12 14 16 18 20 22-0.02
-0.015
-0.01
-0.005
0
0.005
0.01
0.015
Sample
Sco
res
on P
C 1
(79.
93%
)
12 24
61
00 12
24 61
95
123 162 199
241 282 330
371
123
162 199 241
282
C096330
Figure 4-7 PC1 scores for 3% Schatleben Hombitan LW-S-U-HD titania in LLDPE film aged in an oven at 50 ºC. 24 hours UVA pre-irradiated (red triangle) vs 24 hours UVC pre-irradiated (black circle).
900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
Wavenumber
Load
ings
on
PC
1 (7
9.93
%)
Figure 4-8 PC1 loadings plot from Figure 4-7.
146
PC1 describes variances occurring with oxidation: the growth of the carbonyl
peak, and loss of methyl deformation modes. The data are separated due to the
size of the initial carbonyl absorption following pre-irradiation: the UVC pre-
irradiated sample entered the oven with a much stronger absorption
Taking the values for the scores values and plotting them against days aged in the
oven provides a curve describing the extent of degradation (Figure 4-9).
-0.004
-0.002
0
0.002
0.004
0.006
0.008
0.01
0.012
0.014
0.016
0 50 100 150 200 250 300 350 400
Days aged in oven
PC1
scor
e
Figure 4-9 PC1 scores against time taken from Figure 4-7. The sample is 3% Schatleben Hombitan LW-S-U-HD titania in LLDPE film, and the black points represent 24 hours UVC pre-irradiated, and the red points 24 hours UVA pre-irradiated. Polynomial trendlines have been fitted. The scores have been offset to allow a better comparison.
It can be seen from the extent of degradation plot in Figure 4-9 that the two
trendlines are nearly parallel. This is a clear indication that the rates of
degradation in the oven of these samples are the same, regardless of the type of
irradiation that they were subjected to initially. A carbonyl index plot taken by
measuring the area under the carbonyl absorption and plotted against time is
presented in Figure 4-10.
147
-0.1
0
0.1
0.2
0.3
0.4
0.5
0 50 100 150 200 250 300 350
Days aged in oven
Carb
onyl
inde
x
Figure 4-10 Carbonyl index plot for 3% Schatleben Hombitan LW-S-U-HD titania in LLDPE film aged in an oven at 50 ºC. 24 hours UVA pre-irradiated (red) vs 24 hours UVC pre-irradiated (black). Polynomial trendlines have been fitted.
It is evident that the plots presented in Figure 4-9 and Figure 4-10 are
significantly different. As the plots derived from PCA contain a great deal more
spectral data, and the loadings plots demonstrate that these data are directly
related to degradation, it is likely that Figure 4-9 is the more accurate one. The
UVC pre-irradiated sample had embrittled after 370 days of aging, whereas the
UVA pre-irradiated sample was still intact by this time. It is probable that
plotting only the carbonyl area does not reveal all of the degradation related
variances such as unsaturation loss or gain and crosslinking, and as the sample
reaches embrittlement the carbonyl intensity will start to drop off due to the
evolution of small volatile degradation products (Section 1.1.4).
A comparison between PCA derived data and carbonyl index information to
graphically represent the extent of degradation is provided in the following plots.
For this example, PCA has been performed on the same range in the carbonyl
region (ca. 1705 – 1735 cm-1) as the range used for the carbonyl index
measurements. The PE film contained 1% Kronos titania and was aged in the
oven. Trendlines have been fitted to the data, and the r2 values are reported in the
plots.
148
0 100 200 300 400
-0.04
-0.02
0.00
0.02
0.04
0.06
0.08
0.10
0.12
Scor
e on
PC
Days aged in oven
0 secs (r2=0.28) 24 hrs UVC (r2=0.86) 60 secs UVA (r2=0.04)
Figure 4-11 Extent of degradation plot for selected samples containing 1% Kronos titania and aged in the oven. The carbonyl region selected for PCA analysis corresponded to that used for carbonyl index calculations.
149
0 100 200 300 400
0.0
0.1
0.2
0.3
0.4
0.5C
arbo
nyl i
ndex
Days aged in oven
0 secs (r2=0.91) 24 hrs UVC (r2=0.95) 60 secs UVA (r2=0.95)
Figure 4-12 Carbonyl index plot for selected samples containing 1% Kronos titania and aged in the oven. Considering that a similar region of the IR spectrum of the degraded materials
was examined for both plots, it would be expected that the plots should look
similar. Indeed, the plots are very comparable, indicating that the two methods
are presenting the same information in different ways. This result was found to
be repeatable in other series of data, and demonstrates that data obtained via PCA
for extent of degradation plots are reliable and can be used to show the
progression of oxidation in a series of data.
The most significant difference between the two types of plots is the closeness of
fit. The r2 values strongly suggest that the carbonyl index data are more robust,
especially at low levels of oxidation. This is reflected in the plots shown in
Figure 4-9 and Figure 4-10, where the carbonyl index data also show better fit.
It appears that despite the advantage of being able to examine a broader range of
spectral data when employing PCA to determine the extent of degradation,
carbonyl index derived plots provide better accuracy. Therefore one should
150
consider the region(s) of spectra that ought to be considered, and the accuracy of
the data required, when deciding which plot is the most suitable for a given task.
4.3.4 Section Summary
PCA is a useful tool for the exploration of large amounts of spectral data. It can
sometimes highlight very subtle differences, and the ability to choose certain
areas of the spectra for investigation is a considerable advantage over
conventional techniques. Scores and loadings plots provide an easily
interpretable method of data representation.
UV pre-irradiation of LLDPE film containing photoactive titania with
subsequent oven aging has produced some differences in the resulting
degradation products. UV irradiation has formed more complex degradation
structures, including anhydrides, esters and possibly some cyclic oxygenated
functions. These are found only in very small concentrations, and most of the
aging products give absorptions typical of acid and ketone carbonyls. This is in
close agreement with Tidjani’s degradation pathway schematic.
Information regarding the extent of degradation can be obtained from PCA data
plotted against time. There are several advantages in this method, including a
broader range of spectral data that can be analysed, and the speed of calculation.
However, a comparison of carbonyl index plots and PCA derived plots over the
same spectral range reveals that the carbonyl index data are more robust with
regards to closeness of fit.
4.4 Weatherometer aging
In Section 4.3 it was seen that PCA can be used to extract data regarding the
distribution of degradation products, and to obtain information relating to the rate
of degradation similar to carbonyl index plots. This section will apply PCA as an
exploratory tool to mine IR spectral data obtained from samples aged in the
weatherometer.
151
4.4.1 Water vapour
Investigation of the data obtained from the weatherometer aging of LLDPE film
containing Degussa P25 and Huntsman titanias (these are listed as the most
active titanias in Table 16, Section 3.14) revealed the presence of water vapour in
the spectra.
-0.04 -0.02 0 0.02 0.04 0.06 0.08-6
-4
-2
0
2
4
6x 10-3
Scores on PC 1 (91.16%)
Sco
res
on P
C 4
(0.8
6%)
03
06 09
12
15
00
03
03
06 12 15
18
21
24
27
30
33
36
36 39
48 51
54 57
60 63
66
Figure 4-13 Scores for PC1 and PC4 for the LLDPE film containing 3% Huntsman A-HR titania (100% anatase, water dispersible; red triangle) and control sample (black circle). Both samples were aged in the weatherometer and were not pre-irradiated
152
1300 1400 1500 1600 1700 1800
-0.06
-0.04
-0.02
0
0.02
0.04
0.06
0.08
0.1
0.12
Wavenumber
Load
ings
on
PC
4 (0
.86%
)
Figure 4-14 Loadings plot for PC4 from Figure 4-13.
The distribution of samples on PC4 (Figure 4-13) shows a grouping of the
samples containing titania. Examination of the loadings plot of PC4 in Figure
4-14 reveals a ‘spectrum’ that contains signature absorptions (small, sharp
absorptions sitting above the noise) of a water vapour spectrum. The titania
containing samples scored highly on this PC, indicating that there is relatively
more of this spectrum contained in their spectra. The presence of water vapour is
not unique to this sample: it can be found in most of the spectra of degraded
LLDPE containing active titania types. It should be noted that PC4 concerns less
than 1% of the total variance in the spectra between 1900 cm-1 and 1200 cm-1,
and demonstrates the effectiveness of PCA in detecting small phenomena in
complex spectra, and its ability to highlight these differences and present them in
a visually comprehensible manner.
To the best of the author’s knowledge, water vapour has not been found trapped
in plastic film containing photoactive prodegradants before, and no literature
discussing such a phenomenon could be found. Water vapour was not found in
any of the oven aged samples, nor was it found in the weatherometer aged
153
samples containing titania types that were relatively inactive, such as those
manufactured by Kronos or Satchleben Hombitan.
Section 1.1.1 listed the reactions involved in oxidation of a polyolefin. The
initiation and propagation steps are reviewed here:
Initiation:
By radical generator
I (initiator) 2r
r RH rH R+ + By hydroperoxide
ROOH +R HOO
ROOH RO + HO Scheme 4-1
Propagation
R O2 ROO
ROO RH ROOH R
2ROOH RO ROO H2O
+
+ +
+ + Scheme 4-2
Comparing these with the reactions involving titania;
Acceptor:
e + O2 O2
e +
+e
H2O2 OH + OH
R + H RH Scheme 4-3
154
Donor:
+ O2O2
+
+
H2O OH +
R + HRHh
h
h
H
Scheme 4-4
it can be seen that the water vapour likely plays an important role in the
degradation process.
It was mentioned in Section 1.2.2 that exciton holes (h+) play an important role in
titania-catalysed degradation. Thus absorption of UV light by the titania particle
produces a hole, which gives rise to a macro radical and a hydrogen radical. The
macro radical reacts with oxygen in the propagation step to produce a
hydroperoxide, which among other possible reactions can combine with another
hydroperoxide to produce a water molecule (3rd step of Scheme 4.4-2). It is
conceivable that this water molecule can then react further with exciton holes to
produce hydroxyl radicals, which may then act as initiators for new degradation
reactions. These reactions are summarised in Scheme 4-5 below.
155
+ R + HRHh
R O2 ROO
ROO RH ROOH R
2ROOH RO ROO H2O
+
+ +
+ +
+ H2O OH +h H
I (initiator) 2r
r RH rH R+ +
Scheme 4-5
According to this reaction pathway, to produce a water molecule two
hydroperoxide molecules must combine. Therefore hydroperoxides would need
to be in a relatively high concentration for such recombination to occur. It is
expected that the more active a titania is in producing radicals, the more
hydroperoxides are formed, and the higher the likelihood of two hydroperoxide
molecules combining to produce a water molecule.
The cavities produced by titania vigorously oxidising the polymeric material
(Section 3.2) also play a role. The water vapour was noticed in the most oxidised
materials, and these materials also whitened, which was attributed to the
presence of cavities (Section 3.3) and confirmed by SEM. These cavities contain
water vapour in high enough concentrations to be detectable in the infrared
spectra via PCA. As mentioned earlier it is possible that the water re-enters the
reaction cycle to produce more degradation initiating molecules, however as the
water vapour is apparently collecting inside the cavities, it is more probable that
the exciton holes preferentially react with carbon chains rather than with water.
156
4.4.2 UVA vs. UVC pre-irradiation
PCA analysis of samples pre-irradiated with UVA or with UVC and aged in the
weatherometer showed no separation on principal components. Pre-irradiation
did not significantly enhance degradation of the samples, and the degradation-
related information in the IR spectra was found to closely represent the
information found in the carbonyl plots. In many cases, and especially in the
more heavily pre-irradiated samples and those samples with the most photoactive
titania types, there were insufficient data points, (i.e. insufficient number of
spectra) to form a credible PCA calculation. This is due to the short times to
embrittlement, and therefore in some cases only 2 or 3 spectra were acquired.
4.4.3 Section summary
LLDPE film containing photoactive titania and aged in the weatherometer
formed cavities caused by the degradation of material around the titania particles.
Water vapour collected in these cavities, and is detectable in the infrared spectra.
It is likely that this water is created by the combination of two hydroperoxide
moieties, and it is also possible that it re-enters the reaction pathway to create
more oxidation initiating species. Pre-irradiation with UVA or UVC has little
effect of the degradation outcome of LLDPE film containing titania and aged in
the weatherometer, with the samples unable to be distinguished by PCA.
4.5 Conclusions
Multivariate data mining techniques have been used to explore large amounts of
spectral data, providing an advantage over conventional techniques not just in
time saved to perform an analysis, but representation of the data in visual ways
that enhance relevant areas of variance. Principal component analysis has been
used to effectively explore the IR spectral data collected from LLDPE film
containing titania and subjected to different forms of pre-treatment and aging
conditions.
Pre-irradiation with UV was found to promote degradation, decreasing the time
taken for the LLDPE to degrade in the oven. Some differences, such as higher
157
ester concentrations, were found in the spectra of those samples that underwent
more significant periods of pre-irradiation and contained active titania types. In
general, however, the titania served to enhance the degradation process without
greatly changing the types of degradation products formed. The experimental
evidence here supported Tidjani’s degradation pathway schematic presented in
1.1.3.
PCA information was also used to create plots showing relative rates of
degradation. Using PCA scores of a degradation-related PC has some clear
advantages over using conventional carbonyl index for such plots, including the
ability to select only particular or relevant areas of the spectrum to be analysed,
and the speed of calculation. However, a comparison of carbonyl index plots and
PCA derived plots over the same spectral range revealed that the carbonyl index
data have a closer agreement, and fitted trendlines possessed a superior closeness
of fit.
A potentially important discovery found through application of PCA was that of
water vapour, which was established to reside in the cavities in the LLDPE film
produced by titania when subjected to aging in the weatherometer. The water
molecules are likely to be produced by the combination of two hydroperoxides. It
is possible that water molecules may then re-enter the reaction pathway by
reaction with titania-generated exciton holes to produce more oxidation initiation
species. The relevance of water is titania photoreactions is discussed in Section
1.2.4.
Furthermore it has been established that titania does not change the degradation
pathway, however, by sensitising regions of the polymer when subjected to UV
light it behaves as a photocatalyst, increasing the overall rate of degradation. This
is in agreement with the observations discussed in Chapter 3. It is therefore
relevant to examine the changes occurring around titania particles in order to
detect reactive regions, and examine any regions of faster rates of degradation.
Recent changes in IR technology have enabled scientists to reach beyond the
traditional limits of spatial resolution, allowing the investigation of
heterogeneous oxidation via mid-IR spectroscopy.
158
Obtaining spatial information around titania particles via a model polymer system
5.1 Introduction
To this point the focus of the thesis has involved the investigation of data
acquired by taking single-point mid-IR spectra of the degraded LLDPE films
containing various titania types from different manufacturers. By characterising
the oxidation products and comparing the relative rates of degradation, it has
been possible to determine the effects of pre-irradiation on the degradation
processes of bulk LLDPE containing a prodegradant.
The purpose of pre-irradiation was to initiate oxidation reactions, which would
then propagate further oxidation throughout the bulk (see Section 1.1.2).
Considering the success of pre-irradiation in shortening the lifetimes of LLDPE
film aged in an oven, it is likely that the film degraded heterogeneously, as
shown by the SEM images which depicted the heterogeneous distribution of
titania particles. Higher concentrations of oxidation products are expected in the
close vicinity of the photoactive titania nanoparticle centres.
There would be obvious advantages to obtain chemical information regarding the
heterogeneous degradation processes occurring around the titania particles and
propagation of oxidation from particles into the bulk, to assist in the development
of a polymer film that will degrade quickly and controllably. Information such as
the rate of spreading, optimum distance between particles, optimum particle size,
etc. is obtainable if one can monitor the spatial progress of degradation
throughout the bulk.
Until now it has not been possible to obtain chemical data following
heterogeneous oxidation around prodegradant particles in the infrared. Although
chemiluminescence has been used to demonstrate heterogeneous oxidation10,13,192,
this technique does not provide the wealth of information regarding the
development of oxidation products, and thus the ability to trace the oxidation
159
pathway, that is available in the mid-IR. Previously oxidation data has been
acquired from single points (such as the data presented so far in this thesis), and
in order to obtain spatial information line mapping with ATR/FTIR has been
employed.152
Difficulties arise when seeking to monitor the development of oxidation around a
selected prodegradant particle using ATR/FTIR (transmission-mode IR
spectroscopy is generally unsuitable due to sample thickness, interference fringes
in films and poor spatial resolution as discussed in Section 1.3.3). ATR/FTIR
requires the sample to be in optical contact with IRE, which is most commonly
achieved by the application of controlled pressure. However the IRE cannot
remain in constant contact with the sample during oxidation, as the sample must
be subjected to accelerated oxidation conditions between spectra acquisition.
This introduces two hindrances to the study of oxidation. Firstly, the identical
spot on the sample surface must be re-located between accelerated oxidation to
study the environment surrounding the same titania particle, which can be
difficult for some samples. Secondly, ATR measurements on an identical
location require repeated application of pressure to obtain spectra. Mechanical
stresses have been demonstrated to affect photochemical degradation rates of
polymers 193, and it is therefore reasonable to suggest that repeated ATR contact
will affect degradation processes, resulting in an incorrect model of the
degradation pathway.
The imaging ATR/FTIR study of the oxidation of a model aliphatic polymer
presented in this chapter addresses both of these problems, as discussed in further
paragraphs. Additionally the spatial information obtained via imaging ATR/FTIR
(Section 1.3.3.3) is ideally suited to the study of heterogeneous oxidation,
conditional to the size of the heterogeneous domains under investigation.
The size of heterogeneous domains largely determines the suitability of a
particular technique to obtain spatially resolved chemical information. Referring
to the SEM images of Degussa P25 titania in LLDPE (Section 3.2.1, Figure 3-1),
some titania particles are up to 5 µm in diameter, and it is therefore reasonable to
160
suggest that a minimum lateral resolution of 5 µm is required in order to observe
chemical changes occurring around a titania particle for this study.
Imaging ATR/FTIR with an IRE of high refractive index can provide lateral
resolution of up to 3 – 4 µm159. Theoretical aspects of these methods have been
discussed in the introduction (see Section 1.3.4). Furthermore, the ability to
obtain hundreds of spectra in one image (Section 1.3.4.1), and the relative
accessibility of the instrumentation (compared to, for example, a synchrotron
light source) promises potential for the investigation of heterogeneous oxidation
around titania centres.
The novel concept presented in this thesis involves the solvent casting of an
aliphatic model polymer directly onto the IRE surface. The method, described in
detail in the following section, circumvents the need to re-locate the identical
location on the polymer surface, and ATR pressure is not applied as the material
is in good optical contact from the start. Contingent upon a prodegradant particle
being within the imaging area on the IRE surface, heterogeneous oxidation in a
sensitised region around a particle can be monitored in real time, without the
need for removal of the sample from the IRE surface.
5.2 Experimental
The investigation of degradation around a titania particle was performed using an
imaging ATR/FTIR spectrometer at Queensland University of Technology
(QUT). Imaging ATR/FTIR is a mid-IR spectroscopic technique that collects
spectral information in a spatial context, to a lateral resolution of up to 4 µm159.
Factors affecting the lateral resolution capability, and other aspects of imaging
ATR/FTIR, have been discussed in Section 1.3.4.1. The imaging ATR/FTIR
spectrometer used for the research presented in this thesis employed a 32 x 32
Focal Plane Array (FPA) detector, resulting in images containing 32 x 32 pixels.
Each pixel represents the spectrum of a specific part of the sample imaged onto a
particular MCT detector element in the FPA. Images are created in a number of
different ways including band area or intensity, ratios of band areas or intensities,
and more complex methods such as principal component analysis. In this case
161
images were constructed by ratioing the area under the carbonyl absorption to the
CH2 deformation absorption, and the pixels were assigned a colour on an
arbitrary colour scale according to the numerical value of the ratio result, with
red showing high values and blue showing low values.
LLDPE is unsuitable for these oxidation studies as it is largely insoluble at room
temperature, and the likelihood exists of oxidation reactions occurring at the high
temperatures required to dissolve polyethylene. The experiment is designed to
investigate the heterogeneous oxidation of polymeric materials, and is not
restricted to polyethylene. The suitability of several model aliphatic polymers
(for example polyisobutylene, polypropylene) was investigated, and it was
experimentally determined that Topas® (an aliphatic polymer containing a
norbornene moiety (Figure 5-1)) was most appropriate.
xy
Figure 5-1 Molecular structure of Topas®.
Topas was dissolved in cyclohexane and solvent cast onto the IRE surface shown
in Figure 5-2. Degussa P25 titania was deposited onto the surface and the whole
assembly was exposed to a total of 8 hours of UVC according the experimental
procedure recorded in Section 2.4. Images were recorded hourly, and they are
presented in Figure 5-3.
162
Ge InternalRef lection Element
Figure 5-2 ATR/FTIR objective assembly.
5.3 Imaging ATR/FTIR spectroscopy results
The images presented in Figure 5-3 immediately begin to show an increase in the
amount of light blue (indicating an increase in carbonyl absorption) after 1 hour
of irradiation, and after 5 hours there appears to be a significant concentration of
oxidation products that register 0.05 on the carbonyl index scale. By 6 hours the
degradation begins to occur more rapidly, and after 8 hours of irradiation the
oxidation is quite extensive.
163
Figure 5-3 Images taken of Topas containing TiO2 and irradiated with UVC. Each figure is labelled with the cumulative irradiation time. The numbers on the x and y axes represent the number of pixels. Each pixel is 1.2 µm in width. The pixels have been smoothed using Varian software.
164
The oxidation process occurring is clearly heterogeneous, as Figure 5-3-I
contains pixels ranging in colour from green to red, covering 0.05 to 0.15 on the
carbonyl index. Importantly, the heterogeneity is occurring with a domain size of
around 5-10 pixels, or 6-12 µm, across (see Figure 5-4 below). These domains of
more rapid oxidation are thought to be the photosensitised regions, caused by
titania particles as discussed in Section (3.14).
These images demonstrate the applicability of imaging ATR/FTIR to the
heterogeneous investigation of the oxidation of polymers. For imaging
ATR/FTIR to be suitable, the domain size of the heterogeneity would need to
cover at least 3 pixels, translating to approximately 4 µm in diameter.
Comparison with the size of some of the larger titania agglomerates shown in the
SEM images of LLDPE containing Degussa P25 (Figure 3-1), this lateral
resolution would be adequate for the detection of titania particles and oxidation
in the surrounding polymer.
Domains of higher oxidation product concentration
Domain of lower oxidation product concentration
Figure 5-4 Image taken from Figure 5-3I, illustrating heterogeneous domains.
5.3.1 Determination of titania particle location(s)
The current FPA technology prevents direct detection of titania particles on the
surface of the IRE, as the spectral range does not reach below 900 cm-1, which is
higher than the Ti-O absorption in the mid-IR. An ATR/FTIR spectrum of the
Degussa P25 titania used in this experiment is included in Figure 5-5, showing
165
the strong Ti-O absorption below 800 cm-1. The exclusion of this region from the
FPA spectral range forces reliance on the OH stretc.hing absorption above 3000
cm-1 and bending vibration at 1635 cm-1 in order to detect titania directly.
3500 3000 2500 2000 1500 1000
0.0
0.1
0.2
Abso
rban
ce
Wavenumbers (cm-1)
Figure 5-5 ATR/FTIR spectrum of Degussa P25 powder. The O-H stretc.hing absorption (3600 – 3000 cm-1) and bending absorption (1635 cm-1) are due to hydroxyl groups on the surface of the titania particles.
Other factors in this experiment affect the ability to detect titania. Importantly,
the likely location of titania particles must be considered with respect to the
depth of penetration of the IR light into the polymer film. Furthermore the
surface of the film being measured is not the surface in direct contact with
oxygen or under direct UV irradiation, which may also affect the spectra.
166
190 nm
IRE
Topas f ilm
Titania particleImaged area
Figure 5-6 Schematic showing Topas film with titania cast onto IRE surface. The depth of penetration of ATR/FTIR is shown by the dashed black line. The titania particle locations are hypothetical.
Figure 5-6 is a schematic representing the Topas film containing titania particles
cast onto the IRE surface. The Harrick equation presented in Section 1.3.3 can be
used to determine the depth of penetration (dp) of the IR radiation into the Topas
film, which at 3500 cm-1 is;
2857 nm
2π x 4(sin2 45 - (1.5/4)2 )1/2dp =
= 190 nm
with an ATR angle of incidence of 45 °, and a refractive index of 1.5 for
Topas194. A penetration depth of 190 nm is not a well defined cut-off point, as
the strength of the signal from the evanescent waves weakens exponentially as
they move through the sample. However it provides an approximate depth to
which IR analysis is performed, and hence in order to observe hydroxyl
stretc.hing absorptions from a titania particle the particle ought to be within
approximately 190 nm of the IRE surface. This is represented by the schematic in
Figure 5-6, with a titania particle slightly impinging on the edge of the imaged
area. .
The images in Figure 5-3 show the earliest signs of oxidation in the upper left
corner. Any correlation between sensitisation by titania and the relatively rapid
oxidation occurring at this location can be substantiated by a relatively higher
167
OH stretc.hing absorption strength in this corner of the image acquired before
oxidation UV irradiation had commenced.
The poor signal-to-noise ratio of these spectra largely prohibited the use of
conventional spectral exploratory tools such as spectral subtraction and area
comparison, and so PCA was used to analyse the spectra obtained in the image
shown in Figure 5-3-A, i.e. Topas containing Degussa P25 titania prior to UV
irradiation. It was hoped that due to OH absorptions on the surface of the titania
which are visible in the mid-IR spectra, any separation in the data according to
an absorption intensity difference in this region might indicate the presence of
titania.
-0.3 -0.2 -0.1 0 0.1 0.2 0.3-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
Scores on PC 1 (74.80%)
Sco
res
on P
C 2
(10.
33%
)
0hr69 0hr97 0hr100
0hr227
0hr244 0hr254
0hr267
0hr268
0hr309
0hr312
0hr314
0hr318 0hr362
0hr365
0hr370 0hr386 0hr387
0hr393
0hr400
0hr414
0hr419 0hr427
0hr430
0hr442
0hr452
0hr454
0hr457
0hr463
0hr470
0hr486
0hr489
0hr503 0hr504
0hr520
0hr524 0hr543
0hr545
0hr561
0hr590 0hr616
0hr674
0hr706
0hr707
0hr708
0hr732
0hr744
0hr745
0hr763
0hr773
0hr775 0hr777
0hr796
0hr799
0hr805
0hr807
0hr812
0hr830
0hr831 0hr838
0hr843
0hr873
0hr874
0hr901
0hr905
0hr922
0hr923 0hr927
0hr929
0hr931 0hr932
0hr933 0hr937 0hr954 0hr965
0hr974 0hr986
0hr990
0hr994
0hr1009
0hr1015 0hr1018 0hr1021
0hr1022
0hr1023
Figure 5-7 Scores plot for PCs 1 and 2 based on the OH stretc.h region of the spectra obtained in the image from Figure 5-3-A. The spectra are numbered from 1 to 1024, starting from the lower left corner of the image, and increasing sequentially from left to right. Every 32nd spectrum begins a new row above the previous row. The <number>hr prefix refers to the time of UV irradiation. All spectra have been obtained from the image of the sample prior to irradiation, hence the 0hr prefix before all sample labels.
168
Figure 5-7 shows the scores plot for PCs 1 and 2 based on the OH stretc.hing
region of the spectra. Figure 5-8 and Figure 5-9 are the corresponding loadings
plots, accounting for 75% and 10% of the data variation respectively. Following
the description of the spectra labelling method provided in the caption to Figure
5-8, the spectra in the upper left corner of the image are labelled higher than 800
in the series, and belong to the first 5 in every row of 32 spectra. Thus some
separation of these spectra from the other spectra is the image is sought, which
might indicate the possible location of surface OH groups on a titania particle.
Figure 5-7 shows that there is separation of the data on PC1, corresponding to the
location of the spectra in the image. Thus it would appear that PC1 is describing
some systematic artefact present in the spectra due the imaging technique, and
not the presence of titania. Thus this PC was discounted from the investigation,
and PC2 was explored, which was shaped more in accordance with a likely OH
absorption.
3100 3150 3200 3250 3300 3350 3400 3450 3500 35500.116
0.118
0.12
0.122
0.124
0.126
0.128
0.13
Wavenumber
Load
ings
on
PC
1 (7
4.80
%)
Figure 5-8 Loadings plot for PC1 from Figure 5-7.
169
3100 3150 3200 3250 3300 3350 3400 3450 3500 3550-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
Wavenumber
Load
ings
on
PC
2 (1
0.33
%)
Figure 5-9 Loadings plot for PC2 from Figure 5-7.
The scores plot of PC2 in Figure 5-7 contains 1024 spectra, resulting in a
complicated plot from which it is difficult to observe any clear discrimination of
the data. It is expected that if PC2 is describing an OH stretc.hing absorption
signature to a titania particle, and assuming photosensitisation of the upper left
corner of the imaged area by titania, then the cumulative scores in this area of the
image might be higher than another area of the image were oxidation was slower
to occur.
Two 5 x 5 pixel matrices were selected from the image of the Topas film prior to
irradiation, the first from the upper left corner, and the second from an area
demonstrating a slower oxidation rate. These areas are pictured in Figure 5-10.
170
1
2
Figure 5-10 Image of Topas on IRE prior to irradiation. The boxed area labelled 1
represents a region that oxidised quickly, compared to the boxed area labelled 2, which oxidised more slowly.
The PC2 scores for each pixel was determined and summed to determine the
‘total score’ of the 5 x 5 pixel area. Figure 5-11 and Figure 5-12 show the spectra
number corresponding to these pixels, and the score of each pixel on PC2, with a
summed total.
1
171
993 994 995 996 997
961 962 963 964 965
929 930 931 932 933
897 898 899 900 901
865 866 867 868 869
32
31
30
29
28
1 2 3 4 5
27
6Row
Column
-0.024 0.095 0.076 0.014 -0.03
0.075 0.015 0.004 -0.005 0.040
0.089 0.041 0.129 0.110 0.078
0.047 0.087 0.078 0.059 0.043
0.001 -0.01 0.01 0.002 -0.004
32
31
30
29
28
1 2 3 4 5
27
6
Figure 5-11 5 x 5 pixel selection from Box 1 in Figure 5-10 showing the spectrum number (left) and the scores on PC2 for each pixel (right).
Sum of scores in each pixel for PC2 in Box 1 = 1.03
2
172
304 305 306 307 308
272 273 274 275 276
240 241 242 243 244
208 209 210 211 212
176 177 178 179 180
10
9
8
7
6
16 17 18 19 20
5
21Row
Column
0.006 0.007 0.073 0.049 0.037
0.012 -0.040 -0.017 -0.045 0.083
0.020 0.025 -0.042 -0.034 -0.041
-0.027 0.006 0 -0.012 -0.019
0 0.003 -0.010 -0.007 0.013
10
9
8
7
6
16 17 18 19 20
5
21
Figure 5-12 5 x 5 pixel selection from Box 2 in Figure 5-10 showing the spectrum number (left) and the scores on PC2 for each pixel (right).
Sum of scores for each pixel for PC2 in Box 2 = 0.013
It can be seen that area represented by Box 1 in Figure 5-10 has a much high
cumulative score on PC2 than the area in Box 2. PC2 appears representative of
OH stretc.hing absorption, and it is concluded that there is a strong likelihood
that the area of the film corresponding to the upper left corner of the image
contains a titania particle(s), determined by detection of the OH stretc.hing
absorption, which has resulted in sensitisation to oxidation of the Topas film.
While this result demonstrates the probability that the images presented in Figure
5-3 are showing the heterogeneous degradation of polymer film sensitised to
173
oxidation around titania particles, relative rates of oxidation can be examined by
comparing the carbonyl index values of the spectra acquired over the course of
UV irradiation.
5.3.2 Discussion of heterogeneous oxidation
In order to examine the images to determine the spread of oxidation, carbonyl
index values were calculated for the spectra that fell across a line map drawn
through the 4th pixel of each row, as shown in Figure 5-13. These values have
been plotted against their corresponding row number, and displayed in Figure
5-14. Some line maps have been omitted for clarity.
Figure 5-13 Cross section of image to plot carbonyl index.
174
Figure 5-14 Carbonyl index line maps for the 4th spectra in each row. Only selected line maps are shown for clarity.
The images presented in Figure 5-3 indicate that where the line map has been
acquired the carbonyl index is quite variable, and this is represented by the
significant point-to-point variations in the line maps of Figure 5-14. Overall there
is a trend towards higher values at the right side of the plot, correlating with the
upper left corner of the images. This corresponds with the observation of the
domain of higher carbonyl intensity found in the images, thought to be around
titania-sensitised regions.
The relative rates of oxidation can be obtained by plotting the carbonyl index for
each pixel against time. Points corresponding to regions of high (row 29) and low
(row 7) carbonyl index have been plotted, along with the average of all points for
comparisons sake. Figure 5-15.
175
Figure 5-15 Rates of degradation for every 5th pixel and average of all pixels.
Figure 5-15 reveals the photosensitising effect of titania on Topas. There is a
clear induction period during which little or no carbonyl moiety containing
degradation products are formed, however the length of this induction period is 1
hour longer at point corresponding to row 7. At row 29, where it was previously
shown that there is a high probability of a titania particle in the immediate
environment, the induction period only lasts 2 hours. Once the induction period
has ended the polymer tends to degrade at a rate apparently independent of
titania photosensitisation effects.
The adduced evidence strongly indicates that titania is present in the imaged
location of the solvent-cast Topas polymer, and that the titania has had a
photosensitising effect, resulting in heterogeneous degradation and more rapid
oxidation of the polymer surrounding the titania. This experiment is considered a
significant step forward for the study of heterogeneously oxidising polymers
using infrared spectroscopy, for the following reasons:
• It provides an advantage over other spatially-resolved techniques such as
chemiluminescence imaging by supplying chemical information
contained within the mid-IR spectra. This can used to trace the
degradation pathway of materials, examine different oxidation products
176
and identify domains of varying sensitivity to oxidation that might
correspond to different concentrations of components in a polymer blend.
• Polymer oxidation can be studied in real time, and as data are collected
from the same area of polymer, the effect of impurities such as catalyst
residues can be monitored.
• Samples are not subjected to pressure from ATR/FTIR techniques,
removing the influence of mechanical stress on the oxidation process.
Additionally, the same area of the polymer does not need to be re-located.
• Imaging ATR/FTIR with an IRE of high refractive index allows for
spatial resolution of around 4 µm, which is a large improvement on
physical limitations of transmission spectra through a medium of air. This
also allows for the collection of a large amount of data on a small
sampling area, dependent on the number of MCT elements in the FPA.
There are some technical issues regarding polymeric materials that need to be
addressed before this technique can be used for routine analysis of
heterogeneously degrading polymers. These include:
• Polymers need to be dissolved in solvents at room temperature to avoid
initiation of oxidation reactions at elevated temperatures. These solvents
are also required to be volatile and non-aggressive to the bond holding the
IRE into the ATR assembly. For certain polymers, such as polyethylene,
finding the appropriate solvent could prove to be quite difficult.
• Certain polymers have a tendency to shrink due to crosslinking when
oxidising, and lift off the IRE surface. Additionally, polymers tend to lift
off the IRE surface during solvent volatilisation.
• It is difficult to determine the imaging location on the IRE surface. This
might affect heterogeneous polymers with a low concentration of a
177
secondary component, such as catalyst particles, as there is a reduced
likelihood that imaged area will contain a particle.
• There is no clear method for determining the thickness of the polymer
film once it has been solvent cast.
Imaging ATR/FTIR is still a new field of infrared spectroscopy, and the
instrumentation is under continual research and development. Some areas of
instrumentation that require improvement include:
• Poor signal-to-noise ratio of the FPA
• FPA spectral range cut-ff at 900 cm-1, which prevents the identification of
signature absorptions of some materials and oxidation products.
• Varying baselines, anomalous dispersion and bad pixels can result in two
spectra of the same material appearing dissimilar, with dissimilar
absorption intensities.
• Attenuation of signal, resulting in the need for low spectral resolution and
a high number of scans.
5.4 Conclusions
A novel experiment in which a model polymer system containing titania was
photooxidised and imaged in real time to demonstrate the heterogeneous
development of oxidised domains. It is thought that these domains of higher
carbonyl product concentration are likely due to the photosensitisation effect of
titania particles in the immediate vicinity. The existence of such domains implies
that the titania is not homogeneously distributed throughout the polymer.
PCA was used to demonstrate the existence of a region of greater OH absorption
in the upper left corner of the images, corresponding to where the carbonyl
178
concentration was most intense. The Degussa P25 titania mid-IR spectrum
contains OH signature intensities, present as functional groups on the surface of
titania. It is considered likely that titania in this region contributed to the OH
absorption in the images, and the titania has sensitised regions of the Topas to
photooxidation. This was supported by the greater overall gain of carbonyl
absorption intensity in the spectra acquired from the region thought to contain
titania.
An induction period was found to prelude more rapid acceleration of oxidation
rate, which demonstrated the photosensitising effect of titania. Areas that were
likely to contain titania oxidised rapidly after 2 hour of UV exposure, compared
to 3 hours of regions further from titania particles.
The novel technique presented represents a significant advance in imaging
ATR/FTIR spectroscopy, and has acquired previously unobtainable data in real-
time. In particular the ability to acquire spatially-resolved chemical data without
the need to force ATR/FTIR contact or re-locate a position on the sample surface
is advantageous. While it holds great potential for the study of heterogeneous
systems, there are a number of technical and instrumental issues that need to be
addressed before this can become a routine technique.
179
180
Investigation of degradation in the mid-IR using a synchrotron light source
6.1 Introduction
The advantages of high lateral resolution in imaging or mapping spectroscopy
were highlighted in the previous chapter (Section 5.35.1), where the evolution of
oxidation product formation was observed in a domain 5 µm in diameter. IR
spectroscopy with a synchrotron light source is another technique that has shown
potential to achieve high lateral resolution195, and hence is expected to have some
usefulness in the investigation of the spread of oxidation from titania catalyst
particles at the very early stages of degradation.
This study has examined the suitability of IR spectroscopy with a synchrotron
light source to investigate the titania-photocatalysed degradation of polyethylene
film, with a view to the acquisition of data with a lateral component at the
earliest stage of oxidation.
6.2 Experimental
The experimental procedure was described in Section 2.5; however some further
explanation of the sample choice is required. A LLDPE film blown by members
of the project at QUT was used instead of the films produced by Ciba discussed
in Chapters 3 and 4. Among other advantages, this was because the purpose of
the investigation in this case was not primarily to assess the suitability of the film
for commercial applications, but to assess the suitability of transmission IR with
a synchrotron light source to examine the early stages of polymer degradation.
The film produced by QUT was clear, and the titania was better dispersed than
the Ciba produced films. Additionally, at 15 µm thickness the QUT produced
film was 10 µm thinner than the Ciba produced film, allowing better oxygen
permeability.
The advantage of using a synchrotron light source over a conventional source for
this type of investigation is the capacity to achieve diffraction limited lateral
resolution. On laboratory bench top systems with an internal glowbar source this
181
would result in too great a reduction in signal strength; however with the high
brightness and high degree of focus of synchrotron sourced light, even the small
fraction of light that manages to pass through the aperture is sufficient to obtain
quality spectra. Thus, theoretically, an aperture can be set to provide a beam size
of 3 µm x 3 µm at the sample surface, although lateral resolution will be larger
than this because of diffraction effects26.
The Bruker Hyperion 2000 microscope at the Australian Synchrotron had two
single-point MCT detector options. One detector provided a greater spectral
range
(3800 cm-1 to 550 cm-1) at the cost of signal-to-noise ratio. This detector was
chosen over the second option, which provided improved signal-to-noise ratio at
the cost of spectral range, and was effective only to 750 cm-1. It was hoped that
titania would be detectable in the mid-IR, which would allow degradation
information to be related to the location of titania particles. As titania absorption
starts at 750 cm-1 and continues to lower wavenumbers, the detector providing a
more suitable spectral range was selected.
During the course of performing experiments it was found that the signal-to-
noise ratio was too poor to allow an extremely small aperture size. Eventually it
was established that a 10 µm x 10 µm aperture with a spectral resolution of 4 cm-
1 and 256 scans provided the best compromise between lateral resolution, noise,
spectral resolution, titania absorption and acquisition time.
IR maps were obtained during the process of sample irradiation. These maps
were 2 contiguous steps down by 3 contiguous steps across to provide a total of 6
pixels (Figure 6-1). All stage movements were automated, and the stage was
taken back to the same place for each measurement. When the sample was to be
irradiated with UVA for a 2 minute exposure, the stage was brought across to
position the sample under the UV probe (See Section 2.5, Figure 2-6).
Subsequent to irradiation the stage was moved back to the sampling position and
a map was acquired. Maps were obtained from the same place to achieve two
182
goals: to examine any progress of heterogeneous oxidation, and to minimize any
variation in the interference fringe for PCA analysis.
10 µm
M1 M2 M3 10 µm LLDPE film
M6 M5 M4
Figure 6-1 Schematic showing map pattern for LLDPE acquired in micro-transmission mid-IR mode at the Australian Synchrotron. The pixels in the map were acquired sequentially, from map point 1 (M1) to point 6 (M6). Pixels were contiguous 10 µm x 10 µm squares.
Interference fringes, discussed in Section 1.3.3.3, proved to be difficult to
eradicate. Various methods were employed, such as having the film on an angle
during data acquisition, and cutting the film at an angle to sample the cut surface.
The method that met with the most success was to place a small piece of film in
optical contact with a KBR slide; however the film did not remain in contact for
longer than one or two minutes. None of these techniques were successful in
reliably removing the interference fringe, and ultimately it was decided to
continue with data acquisition. By revisiting the same point on the polymer film
surface for mapping, it was hoped that as long as the thickness of the film did not
change during irradiation, the interference fringe pattern at each point in the map
should be identical. From the point of view of assessing the suitability of these
techniques for future studies, it was of interest to investigate whether a
chemometric analysis of the data would be able to overcome changes in the
spectra due to variations in the interference fringe pattern.
183
6.3 Synchrotron results and discussion
Figure 6-2 shows a typical spectrum acquired at the Australian Synchrotron of
LLDPE containing 3% Degussa P25 titania, obtained using the experimental
procedure described in Section 2.5. The sinusoidal baseline is characteristic of an
interference fringe. Titania does not absorb strongly in this spectrum, and the low
wavenumber end of the spectrum suffers from poor signal-to-noise ratio, making
titania detection quite difficult.
4000 3500 3000 2500 2000 1500 1000 500
0.0
0.5
1.0
1.5
2.0
2.5
3.0
3.5
Abs
orba
nce
Wavenumbers (cm-1)
Figure 6-2 A mid-IR spectrum of LLDPE containing 3% Degussa P25 titania obtained using a Bruker Hyperion 2000 microscope with an MCT detector with a synchrotron source, a 10 µm x 10 µm aperture, 4 cm-1 spectral resolution and 256 scans. Some data points are missing, presumbably due to conversion from a Bruker format to a Grams32 AI compatible. As spectra in the Bruker format cannot be read or manipulated by other software, it was necessary to convert to a more suitable format.
PCA proved to be a valuable tool to address challenges created by poor signal-to-
noise ratio and interference fringes when analysing the carbonyl and fingerprint
regions. As mentioned earlier, the detector with a broader detection range at the
184
cost of signal-to-noise ratio was employed for data acquisition. Subjection of the
data to PCA resulted in a greatly improved signal-to-noise ratio of the spectra,
which is key to an investigation of degradation around a titania particle as the
absorption of oxidation products in early stages of degradation is likely to be
weak, making it difficult to distinguish from noise.
PCA was particularly effective in assisting with interference fringes. Firstly, it
should be pointed out that the interference fringes remain in the loadings plots
after PCA analysis, which suggests that there are some changes occurring to the
sinusoidal pattern. This is probably due to changes in the thickness of the
LLDPE film as it begins to crosslink during UV irradiation. Despite the lingering
presence of the fringes, improvement in the signal-to-noise ratio and the
highlighting of absorptions that are changing in the series of spectra allows
absorptions to stand out clearly above the interference fringe. This makes any
small changes occurring in the data much more accessible to investigation.
The benefit of using PCA to data mine mid-IR spectra of this nature was clearly
demonstrated when analysing the data obtained in the experiment described
Section 6.2. The PCA result of the region below 1900 cm-1 from the first map
point is provided in the following figures. PC1 appears to describe systematic
changes as it steadily decreases with increasing irradiation time, while
subsequent PCs describe noise.
To help interpretation of the figures, the reader is reminded of the schematic
presented in Figure 6-1 which describes the order of pixels in the map, starting at
M1. In the following figures the number of minutes the sample had undergone
irradiation at the time the map was acquired is provided by a number following
an underscore. Thus as an example M1_18 represents the spectrum acquired
from the first pixel in the map, by the time the film had undergone 18 minutes of
UV exposure. The sample number in the following scores plots refers to the
sequence in the series of spectra collected during 30 minutes of irradiation.
185
2 4 6 8 10 12 14 16-12
-10
-8
-6
-4
-2
0
2
4
6
8x 10
-3
Sample Number
Sco
res
on P
C 1
(87.
98%
)
M1_0
M1_2
M1_4
M1_6
M1_8
M1_10
M1_12
M1_14
M1_16
M1_18
M1_20
M1_22
M1_24
M1_26 M1_28 M1_30
Figure 6-3 Scores plot for PC1 of the first map point. Each point represents a spectrum acquired during the course of the experiment. The sample number on the x-axis represents the sequential number of the spectrum.
600 800 1000 1200 1400 1600 1800-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
Wavenumber
Load
ings
on
PC
1 (8
7.98
%)
Figure 6-4 Loadings plot for PC1 from Figure 6-3.
186
It is not immediately clear what PC1 is describing in this case. Certainly the
interference fringe is visible, and the scores plot could be describing some
systematic change. It does not appear to be degradation related, as the PC scores
drop steadily for 10 minutes of irradiation, then begin to increase again until 30
minutes of irradiation. It is possible that the PC is describing changes related to
interference fringes, CH absorption variations, or some other artefact.
Due to the unsuitability of PCA over this broad spectral region, only the carbonyl
region has been investigated for the 6 pixels in the map, and the results provided
below. The data have been over fitted to show PCs 2 and 3, which appear quite
noisy. PC1 is blue, PC2 is green, and PC3 is red.
M1
2 4 6 8 10 12 14 16-6
-4
-2
0
2
4
6
8x 10-3
Sample Number
Sco
res
on P
C 1
(89.
15%
), P
C 2
(6.0
1%),
PC
3 (1
.05%
)
M1_0
M1_2
M1_4
M1_6
M1_8 M1_10 M1_12
M1_14
M1_16 M1_18 M1_20
M1_22
M1_24
M1_26 M1_28
M1_30
1700 1720 1740 1760 1780 1800
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
Wavenumber
Load
ings
on
PC
1 (8
9.15
%)
M2
2 4 6 8 10 12 14 16-8
-6
-4
-2
0
2
4
6
8
10x 10-3
Sample Number
Sco
res
on P
C 1
(87.
68%
), P
C 2
(6.9
1%),
PC
3 (1
.75%
)
M2_0
M2_2
M2_4
M2_6
M2_8 M2_10
M2_12
M2_14 M2_16 M2_18 M2_20
M2_22
M2_24 M2_26 M2_28 M2_30
1710 1720 1730 1740 1750 1760 1770 1780 1790 1800
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
Wavenumber
Load
ings
on
PC
1 (8
7.68
%)
187
M3
2 4 6 8 10 12 14 16-6
-4
-2
0
2
4
6
8x 10-3
Sample Number
Sco
res
on P
C 1
(77.
17%
), P
C 2
(14.
34%
), P
C 3
(2.8
3%)
M3_0
M3_2
M3_4 M3_6
M3_8
M3_10
M3_12 M3_14 M3_16 M3_18 M3_20
M3_22
M3_24 M3_26 M3_28
M3_30
1710 1720 1730 1740 1750 1760 1770 1780 1790 1800
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
Variable
Load
ings
on
PC
1 (7
7.17
%)
M4
2 4 6 8 10 12 14 16-4
-3
-2
-1
0
1
2
3
4x 10-3
Sample Number
Sco
res
on P
C 1
(53.
31%
), P
C 2
(34.
12%
)
M4_0 M4_2
M4_4
M4_6
M4_8
M4_10 M4_12
M4_14
M4_16 M4_18
M4_20
M4_22
M4_24
M4_26
M4_28
M4_30
1710 1720 1730 1740 1750 1760 1770 1780 1790
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
0.3
Wavenumber
Load
ings
on
PC
1 (5
3.31
%),
PC
2 (3
4.12
%)
M5
2 4 6 8 10 12 14-8
-6
-4
-2
0
2
4
6
8
10
12x 10-3
Sample Number
Sco
res
on P
C 1
(79.
37%
), P
C 2
(16.
44%
), P
C 3
(1.8
6%)
M5_2
M5_4
M5_6
M5_8
M5_10
M5_12
M5_14 M5_16 M5_18 M5_20 M5_22 M5_24
M5_26 M5_28 M5_30
1700 1710 1720 1730 1740 1750 1760 1770 1780 1790
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
Wavenumber
Load
ings
on
PC
1 (7
9.37
%)
M6
188
2 4 6 8 10 12 14-6
-4
-2
0
2
4
6
8
10
12x 10-3
Sample Number
Sco
res
on P
C 1
(80.
21%
), P
C 2
(14.
00%
), P
C 3
(2.6
9%)
M6_2
M6_4
M6_6
M6_8
M6_10
M6_12 M6_14
M6_16 M6_18 M6_20 M6_22 M6_24 M6_26
M6_28
M6_30
1710 1720 1730 1740 1750 1760 1770 1780 1790 1800
-0.2
-0.15
-0.1
-0.05
0
0.05
0.1
0.15
0.2
0.25
Wavenumber
Load
ings
on
PC
1 (8
0.21
%)
Figure 6-5 Scores plots (left) and loadings plots (right) for each pixel in a 3 x 2 map (see Figure 6-1) of LLDPE containing 3% Degussa P25 titania irradiated with UVA. Spectra were recorded every 2 minutes for 30 minutes total irradiation time. Each point in the scores plots represents a spectrum, and the label describes the pixel number and length of UV exposure of the film in minutes. The sample number on the x-axis represents the sequential number of the spectrum.
There is a clear trend in the scores and loadings plots presented in Figure 6-5.
PC1 seems to be the only PC describing a systematic change in the data with
time, starting at a high score which diminishes with time. PCs 2 and 3 mostly
describe noise, with the possible exception of the 4th map point. And each PC1
loadings plot shows similarly positioned absorption peaks.
It is unlikely that the PCA investigation in this instance is detecting any
degradation related changes in the spectra. Inspection of the loadings plots shows
upward pointing absorptions at 1718 cm-1, 1735 cm-1, 1750 cm-1 and 1773 cm-1.
The high starting scores continuing to low finishing scores indicate that these
absorptions are actually disappearing with time. Closer inspection of these
loadings plots reveals that these absorptions are describing water vapour, which
in this case is in the surrounding air.
The likelihood of water vapour absorptions present in the spectra was confirmed
by the regularly high value of the score of the data in all map pixels collected
after the 10th minute of irradiation. It had been noted while conducting the
experiment that the 3 minute nitrogen purge before data collection was
accidentally omitted, and therefore more water vapour is seen as this point. There
189
are several possible reasons why water vapour was seen to be decreasing over the
course of the experiment.
The experiment was conducted between the hours of 11pm and 7am. Until 11pm
there had been 2 persons around the instrument, however following the
commencement of the experiment there was only 1, reducing the source of
atmospheric water. Additionally, after initially monitoring the experiment in the
early stages, the person conducting the experiment would leave the room while
waiting for the purge and data collection.
The late hour at which the experiment was performed could also have had an
influence, as there were less people in the vicinity, and less doors being opened
to the outside. The Australian Synchrotron has an air purging system which helps
to keep the air at low humidity, and while there was less human traffic this may
have help it to work more efficiently, contributing to lower water vapour.
Regardless of the water vapour content, there does not appear to be any signs of
oxidation in the spectra. This was not expected, as proof of concept experiments
conducted prior to this one had found the appearance of a carbonyl after 10-15
minutes of irradiation with the same source. However in the optimisation
experiments the spectra were measured using diamond ATR/FTIR, which is a
surface sensitive technique. In the experiment discussed here, spectra were
collected in transmission and hence measured much more the bulk of the
material, decreasing the relative concentration of any degradation products.
Additionally, despite the best efforts to allow oxygen back into the purge box
described in Section 2.5, there would have been a constant pressure of purge gas
away from the sampling area, preventing atmospheric oxygen from circulating
back into the box. Any lack of oxygen would naturally result in a decrease in
oxidation product formation.
190
6.4 Conclusions
Micro-transmission mid-IR spectroscopy with a synchrotron light source allows
for improved spatial resolution at the sample surface, down to a possible 3 µm.
With the view to correlate oxidation-related absorptions to distance from a titania
particle, an MCT detector providing a spectral range broad enough to encompass
titania absorption was selected. This incurred some penalties in data collection,
however. Primarily, the necessary 5 µm or better lateral resolution could not be
achieved due to a poor signal-to-noise ratio; the uppermost lateral resolution that
could be attained was 10 µm. Furthermore the signal-to-noise ratio required 256
scans at 4 cm-1 spectral resolution, which becomes time consuming when
acquiring spectra in a 6 pixel map every two minutes. Notwithstanding the long
acquisition time, the spectra retained a significant level of noise, particularly at
the low wavenumber end of the spectrum.
There were other challenges faced when analysing a film in the mid-IR at the
Australian Synchrotron. The instrument possessed a Perspex purge box with
doors opening at the front (see Figure 2-4). This made measurements acquired
over time very cumbersome, as the environment inside the purge box requires up
to 10 minutes or more to remove absorptions from external water and carbon
dioxide when taking sensitive measurements. Additionally, the purge source is
nitrogen gas, which must be replaced with dry air in order to conduct oxidation
experiments.
Interference fringes could not be removed from the spectra, despite using
different methods such as having the film at an angle, and placing the film in
optical contact with a medium transparent in the mid-IR. Despite the difficulties
caused by interference fringes when performing data manipulation such as
spectral subtraction, the sensitivity of PCA analysis surpassed the problem by
highlighting absorptions that are changing with time. PCA analysis also
demonstrated an ability to largely remove noise considerations, and was able to
clearly detect the variation of water vapour in the air during the course of the
experiment, despite allowing for 10 minutes of purging time inside the Perspex
box.
191
Future studies of the oxidation of LLDPE film could be successful if the purge
gas can be replaced with dry air or oxygen. However due to limitations of the
detector it is difficult to detect titania particles, which remains a challenge if
information regarding degradation changes around a titania particle is desired.
192
Conclusions The primary aim of the work reported in this thesis has been to exploit the nature
of titania photocatalysis in order to advance the technology that will lead to the
development of a commercial plastic film with controllable degradation
properties, even in the dark. To this end the concept of pre-irradiation has been
thoroughly examined using commercially available titania from different
manufacturers, blown in a LLDPE film by a well known and respected chemical
production company, Ciba. Pre-irradiation is a novel concept, and involves the
exposure of the film containing titania photosensitiser to UV irradiation in order
to initiate oxidation reactions prior to aging in the dark.
Nine different samples of 25 µm thick LLDPE film, containing 1-3% loadings of
titania including Degussa P25, Hunstman Tioxide, Satchleben Hombitan and
Kronos were subject to investigation. The degradation of the samples was
followed by mid-IR spectroscopy to determine the effects of pre-irradiation
wavelength (UVA vs. UVC), length of pre-irradiation time, and aging conditions
(accelerated aging in the oven at 50 °C, and suntest aging).
SEM images showed that the Degussa P25 titania tended to agglomerate into
particles up to several micron across, while modified titanias exhibited much
better size and particle distribution. When exposed to UV irradiation some of the
samples turned white – this occurred only in the samples containing photoactive
titania, and the whiteness was found to be the result of light scattering caused by
the titania particles completely destroying the surrounding polymer to form a
cavity the shape of a ‘wormhole’.
ATR/FTIR spectra of degraded polymer film demonstrated that samples exposed
to UV irradiation developed a higher concentration of oxidation products
containing ester moieties, along with various other products that absorb at higher
wavenumbers such as lactones and anhydrides. Oven aged samples however
tended to form acids, confirming the degradation pathways proposed by
Tidjani42. Importantly, it was found that although titania accelerates the
193
degradation of LLDPE film, the degradation products, and relative
concentrations of these products, is not affected by titania. It is concluded that
titania behaves as a catalyst by providing radicals for degradation to occur,
without changing the degradation pathway.
The LLDPE film containing 3% Degussa P25 titania and pre-irradiated for 24
hours with UVA embrittled in approximately 200 days in the oven. This is a
significant reduction in the lifetime of the polymer compared to the control
sample pre-irradiated for 24 hours with UVA, whose carbonyl index plot did not
show signs of significant oxidation occurring after 260 days in the oven. The
reduction in the lifetime of the LLDPE film containing titania after pre-
irradiation, in conjunction with the photocatalytic nature of titania that
accelerates degradation without changing the degradation pathway, has been
accepted as strong evidence that pre-irradiation is a successful concept.
Some dissimilarity was observed between pre-irradiation with UVA and UVC,
whereby samples exposed to UVC often degraded more rapidly than those
exposed to UVA. This was attributed solely to the higher energy of UVC, which
resulted in more aggressive degradation of the LLDPE prior to aging. The
degradation products and hence pathways were found to be similar, however, and
in all situations UVA and UVC irradiation was found to accelerate aging.
This has an impact on the research performed by Allen 7 and co-workers. Allen
found that pigment grade titania has a photostabilising effect when subjected to
UVC irradiation. SEM images of the titania used in this study illustrated that it
had agglomerated into pigment grade sized particles, however at no point was the
LLDPE film stabilised by titania when exposed to UVC. It is concluded that the
results found by Allen et al. were applicable only to that unique data set, and
titania acts as a photosensitiser when subjected to UVC irradiation.
Titania surface modification was found to play a more important role in reducing
the photoactive potential of the titania than particle size and aggregation,
probably due to the lack of available sites for oxygen trapping. Degussa P25
titania was demonstrated to be significantly more photoactive than the other
194
forms of titania. This was attributed to the crystal phase, whereby the minority
rutile fraction was considered to be acting as a dopant, assisting in electron/hole
separation. The high photoactivity of Degussa P25 titania despite significant
agglomeration and lack of surface modification implies that titania activity is
dependent on the efficiency of electron/hole separation.
Increasing the loading from 1% to 3% had a moderate effect of increasing the
rate of degradation. Low doses of UVA and UVC pre-irradiation did not greatly
affect the rate of degradation, and several hours at least of pre-irradiation was
required in order to achieve embrittlement significantly faster. It is likely that
higher rates of degradation when increasing the loading was not seen due to the
tendency of the particles to aggregate, reducing the available surface area. This
was supported by SEM evidence.
The order of photoactivity of titania in the LLDPE films was determined to be
Degussa P25 >> Huntsman Tioxide > Satchleben Hombitan > Kronos.
Over the course of the pre-irradiation experiment thousands of mid-IR spectra
had been collected to form a broad, comprehensive data set. This data was
subjected to the multivariate data analysis technique PCA, which was found to
not only handle the large amount of data quickly and efficiently, but was also
easily tailored to investigate various aspects of the degradation.
PCA confirmed that titania acted as a photocatalyst by increasing the rate of
reaction, without altering the degradation pathway. It was used to provide an
alternative measure of oxidation by plotting the scores of a series of spectra on a
principal component describing degradation against time, to result in a plot
describing the rate of degradation. This has the advantage over conventional
carbonyl index plots of including much more spectral data in the analysis,
providing a plot describing the rate of change in the whole spectrum rather than
just the carbonyl region. The data did not show the closeness of fit of carbonyl
index derived data.
195
The ability of PCA to detect very slight changes in a spectrum was demonstrated
by the detection of water vapour in the ‘wormholes’ caused by titania completely
oxidising the LLDPE. The existence of water vapour in these cavities was
previously unknown, and it is possible that it plays some secondary role in the
degradation processes of the LLDPE.
Heterogeneous oxidation was investigated in a novel experiment by imaging
ATR/FTIR, and domains of more rapid carbonyl product formation were
discovered. For the first time spectral information with a lateral component was
collected in real time, and revealed the existence of localised domains of
increased oxidation rate. It was thought that these regions corresponded to the
location of titania particles. Although titania could be directly observed in the IR
spectrum obtained using and FPA, PCA was used to demonstrate that regions of
high OH concentration corresponded to sensitised domains. The OH signature
was attributed to functional groups present on the surface of the titania particles.
Despite the potential of imaging ATR/FTIR for the study of heterogeneously
degrading polymers, some challenges need to be addressed before it can become
a more routine technique.
IR transmission spectroscopy with a synchrotron light source did not prove as
useful as hoped for the investigation of the early stages of oxidation production
formation in a lateral context. Despite various issues complicating the acquisition
of data, it was found that interference fringe issues and poor signal-to-noise
could be somewhat compensated for by subjection of the data to PCA. Once
again PCA demonstrated its usefulness in the analysis of mid-IR spectral data
sets by overcoming interference fringe issues, and by graphical representation of
small changes in the spectra, such as the drop in ambient humidity overnight. For
future studies it is likely however that imaging ATR/FTIR techniques will supply
more useful information.
This thesis has demonstrated that UV pre-irradiation of an LLDPE film with
photoactive titania particles can result degradation of the film in the dark in a
greatly accelerated time frame. It follows that the concept of pre-irradiation holds
potential for the development of a plastic film with controllable degradation
196
properties for commercial outcomes. The way forward for this technology is to
increase the photosensitivity of the LLDPE film, and to continue to expand the
knowledge of the degradation chemistry occurring around titania nano-particles,
in order to optimise the tunablity of the product for various applications.
197
198
References
1. www.plasticsinfo.co.za. 2004. 2. N/A. (Personal Products Co., USA). Application: BE: 1975, p 11 pp. 3. Willett, J. L.; Westhoff, R. P., Annual Technical Conference - Society of
Plastics Engineers 52nd, 1572 1994. 4. Roy, P. K.; Surekha, P.; Rajagopal, C.; Chatterjee, S. N.; Choudhary, V.,
Polymer Degradation and Stability 91, 1791 2006. 5. Sekine, Y.; Fujimoto, K., Journal of Material Cycles and Waste
Management 5, 107 2003. 6. Kaczmarek, H.; Sionkowska, A.; Kaminska, A.; Kowalonek, J.; Swiatek,
M.; Szalla, A., Polymer Degradation and Stability 73, 437 2001. 7. Allen, N. S.; Khatami, H.; Thompson, F., European Polymer Journal 28,
817 1992. 8. Arnaud, R.; Dabin, P.; Lemaire, J.; Al-Malaika, S.; Chohan, S.; Coker,
M.; Scott, G.; Fauve, A.; Maaroufi, A., Polymer Degradation and Stability 46, 211 1994.
9. Fujishima, A.; Zhang, X.; Tryk, D. A., Surface Science Reports 63, 515 2008.
10. George, G. A.; Celina, M., Environmental Science and Pollution Control Series 21, 277 2000.
11. Allen, N. S.; Edge, M., Fundamentals of Polymer Degradation and Stabilization; Springer, 1992.
12. Denisov, E. T. In Environmental Science and Pollution Control Series; Amin, M. B.; Hamid, S. H.; Maadhah, A. G., Eds.; Marcel Dekker: NY, 2000.
13. Celina, M.; George, G. A., Polymer Degradation and Stability 40, 323 1993.
14. Bandyopadhyay, P. K.; Shaw, M. T.; Weiss, R. A., :Polymer-Plastics Technology and Engineering 24, 187 1985.
15. James C. W. Chien, E. J. V. H. J.: 1968, p 381. 16. Gijsman, P.; Hennekens, J.; Vincent, J., Polymer Degradation and
Stability 39, 271 1993. 17. Colin, X.; Fayolle, B.; Audouin, L.; Verdu, J., International Journal of
Chemical Kinetics 38, 666 2006. 18. Gugumus, F., Polymer Degradation and Stability 52, 145 1996. 19. Scheirs, J.; Bigger, S. W.; Billingham, N. C., Polymer Testing 14, 211
1995. 20. Lawrence, J. B.; Weir, N. A., Journal of Applied Polymer Science 18,
1821 1974. 21. Gupta, A. P.; Saroop, U. K.; Gupta, V., Journal of Applied Polymer
Science 106, 917 2007. 22. Tidjani, A.; Dasilva, A. O.; Fanton, E.; Arnaud, R., Journal of
Macromolecular Science, Pure and Applied Chemistry A36, 633 1999. 23. Rymarz, G.; Klecan, T.; Sobocinska, A., Polimery (Warsaw, Poland) 38,
588 1993.
199
24. Rugg, F. M.; Smith, J. J.; Wartman, L. H., Journal of Polymer Science 11, 1 1953.
25. Rugg, F. M.; Smith, J. J.; Bacon, R. C., Journal of Polymer Science 13, 535 1954.
26. Beachell, H. C.; Tarbet, G. W., Journal of Polymer Science 45, 451 1960. 27. Beachell, H. C.; Nemphos, S. P., Journal of Polymer Science 21, 113
1956. 28. Lacoste, J.; Singh, R. P.; Arnaud, R.; Lemaire, J., Recent Trends Chem.
React. Eng., [Proc. Int. Chem. React. Eng. Conf.] 1, 312 1987. 29. Cruz-Pinto, J. J. C.; Carvalho, M. E. S.; Ferreira, J. F. A., Angewandte
Makromolekulare Chemie 216, 113 1994. 30. Tyler, D. R., Journal of Macromolecular Science, Polymer Reviews C44,
351 2004. 31. David, C.; Trojan, M.; Daro, A.; Demarteau, W., Polymer Degradation
and Stability 37, 233 1992. 32. George, G. A., Mater. Forum FIELD Full Journal Title:Materials Forum
19, 145 1995. 33. Fanton, E.; Pellereau, B.; Arnaud, R.; Lemaire, Polymer Photochemistry
6, 437 1985. 34. Peeling, J.; Clark, D. T., Journal of Polymer Science, Polymer Chemistry
Edition 21, 2047 1983. 35. Rabek, J. F., Polymer Photodegradation: Mechanisms and experimental
methods Springer, 1996. 36. Al-Malaika, S.; Scott, G., Degradation and Stability of Polyolefins, 283
1983. 37. Arnaud, R.; Moisan, J. Y.; Lemaire, J., Macromolecules 17, 332 1984. 38. Fanton, E.; Pellereau, B.; Arnaud, R.; Lemaire, J., Polymer
Photochemistry 6, 437 1985. 39. Gugumus, F., Polymer Degradation and Stability 27, 19 1990. 40. Lacoste, J.; Carlsson, D. J.; Falicki, S.; Wiles, D. M., Polymer
Degradation and Stability 34, 309 1991. 41. Gijsman, P.; Sampers, J., Polymer Degradation and Stability 58, 55 1997. 42. Tidjani, A., Polymer Degradation and Stability 68, 465 2000. 43. Al-Malaika, S.; Scott, G., Degradation and Stability of Polyolefins, 247
1983. 44. Scott, G., Pure and Applied Chemistry 55, 1615 1983. 45. Habicher, W. D.; Bauer, I., Environmental Science and Pollution Control
Series 21, 81 2000. 46. Gugumus, F., Polymer Degradation and Stability 50, 101 1995. 47. Gugumus, F. L., Environmental Science and Pollution Control Series 21,
1 2000. 48. Freitas, A. R.; Vidotti, G. J.; Rubira, A. F.; Muniz, E. C., Polymer
Degradation and Stability 87, 425 2005. 49. Mills, A.; Le Hunte, S., Journal of Photochemistry and Photobiology A:
Chemistry 108, 1 1997. 50. Barr-Kumarakulasinghe, S. A., Polymer 35, 995 1994. 51. May, W. R.; Bsharah, L., Industrial & Engineering Chemistry Product
Research and Development 9, 73 1970. 52. Chalk, A. J.; Smith, J. F., Transactions of the Faraday Society 53, 1214
1957.
200
53. Dabald, A. R., Rubber Age (New York) 71, 621 1952. 54. Osawa, Z., Polymer Degradation and Stability 20, 203 1988. 55. Roy, P. K.; Surekha, P.; Rajagopal, C.; Chatterjee, S. N.; Choudhary, V.,
Polymer Degradation and Stability 90, 577 2005. 56. Sipinen, A. J.; Rutherford, D. R., Journal of Environmental Polymer
Degradation 1, 193 1993. 57. Winkler, J., Titanium Dioxide; Vincentz Network: Hannover, 2003. 58. Mills; Andrew; Hill; George; Crow; Matthew; Hodgen; Stephanie, Thick
titania films for semiconductor photocatalysis; Springer: Heidelberg, 2005.
59. Hurum, D. C.; Agrios, A. G.; Gray, K. A.; Rajh, T.; Thurnauer, M. C., Journal of Physical Chemistry B 107, 4545 2003.
60. Ivanova, S. R.; Minsker, K. S.; Zaikov, G. E., Oxidation Communications 25, 325 2002.
61. Thompson, T. L.; Yates, J. T., Jr., Topics in Catalysis 35, 197 2005. 62. Allen, N. S., Polymer Degradation and Stability 2, 155 1980. 63. Zhang, Z.; Wang, C.-C.; Zakaria, R.; Ying, J. Y., Journal of Physical
Chemistry B 102, 10871 1998. 64. Bickley, R. I.; Gonzalez-Carreno, T.; Lees, J. S.; Palmisano, L.; Tilley, R.
J. D., Journal of Solid State Chemistry 92, 178 1991. 65. Linsebigler, A. L.; Lu, G.; Yates, J. T., Jr., Chemical Reviews
(Washington, D. C.) 95, 735 1995. 66. Watson, S. S.; Beydoun, D.; Scott, J. A.; Amal, R., Chemical Engineering
Journal (Amsterdam, Netherlands) 95, 213 2003. 67. Lee, S.-K.; Robertson, P. K. J.; Mills, A.; McStay, D.; Elliott, N.;
McPhail, D., Applied Catalysis, B: Environmental 44, 173 2003. 68. Ohtani, B.; Zhang, S.; Handa, J.; Kajiwara, H.; Nishimoto, S.; Kagiya, T.,
Journal of Photochemistry and Photobiology, A: Chemistry 64, 223 1992. 69. Dagan, G.; Tomkiewicz, M., Journal of Physical Chemistry 97, 12651
1993. 70. Heller, A.; Degani, Y.; Johnson, D. W., Jr.; Gallagher, P. K., Proceedings
- Electrochemical Society 88-14, 23 1988. 71. Yang, Q.; Xie, C.; Xu, Z.; Gao, Z.; Li, Z.; Wang, D.; Du, Y., Journal of
Molecular Catalysis A: Chemical 239, 144 2005. 72. Beydoun, D.; Amal, R.; Low, G. K. C.; McEvoy, S., Journal of Physical
Chemistry B 104, 4387 2000. 73. Sato, S., Langmuir 4, 1156 1988. 74. Thompson, T. L.; Yates, J. T., Jr., Chemical Reviews (Washington, DC,
United States) 106, 4428 2006. 75. Luo, Z.; Gao, Q., Journal of Photochemistry and Photobiology, A:
Chemistry 63, 367 1992. 76. Allen, N. S.; Edge, M.; Ortega, A.; Sandoval, G.; Liauw, C. M.; Verran,
J.; Stratton, J.; McIntyre, R. B., Polymer Degradation and Stability 85, 927 2004.
77. Allen, N. S.; Edge, M.; Sandoval, G.; Ortega, A.; Liauw, C. M.; Stratton, J.; McIntyre, R. B., Polymer Degradation and Stability 76, 305 2002.
78. Cho, S.; Choi, W., Journal of Photochemistry and Photobiology A: Chemistry 143, 221 2001.
79. Hadjiivanov, K. I.; Klissurski, D. G., Chemical Society Reviews 25, 61 1996.
201
80. Kontos, A. I.; Arabatzis, I. M.; Tsoukleris, D. S.; Kontos, A. G.; Bernard, M. C.; Petrakis, D. E.; Falaras, P., Catalysis Today 101, 275 2005.
81. Piscopo, A.; Robert, D.; Weber, J. V., Journal of Photochemistry and Photobiology, A: Chemistry 139, 253 2001.
82. Fox, M. A., Solar Energy Materials and Solar Cells 38, 381 1995. 83. Tang, Q.; Lin, J.; Wu, Z.; Wu, J.; Huang, M.; Yang, Y., European
Polymer Journal 43, 2214 2007. 84. Zan, L.; Wang, S.; Fa, W.; Hu, Y.; Tian, L.; Deng, K., Polymer 47, 8155
2006. 85. Zan, L.; Tian, L.; Liu, Z.; Peng, Z., Applied Catalysis A: General 264,
237 2004. 86. Farrokhpay, S.; Morris, G. E.; Fornasiero, D.; Self, P., Colloids and
Surfaces A: Physicochemical and Engineering Aspects 253, 183 2005. 87. Ju-Nam, Y.; Lead, J. R., Science of The Total Environment 400, 396
2008. 88. Zhao, X. u.; Li, Z.; Chen, Y.; Shi, L.; Zhu, Y., Journal of Molecular
Catalysis A: Chemical 268, 101 2007. 89. Cheng, P.; Li, W.; Zhou, T.; Jin, Y.; Gu, M., Journal of Photochemistry
and Photobiology, A: Chemistry 168, 97 2004. 90. Wu, J. C. S.; Chen, C.-H., Journal of Photochemistry and Photobiology,
A: Chemistry 163, 509 2004. 91. Iketani, K.; Sun, R.-D.; Toki, M.; Hirota, K.; Yamaguchi, O., Materials
Science & Engineering, B: Solid-State Materials for Advanced Technology B108, 187 2004.
92. Sonawane, R. S.; Kale, B. B.; Dongare, M. K., Materials Chemistry and Physics 85, 52 2004.
93. Gole, J. L.; Stout, J. D.; Burda, C.; Lou, Y.; Chen, X., Journal of Physical Chemistry B 108, 1230 2004.
94. Klosek, S.; Raftery, D., Journal of Physical Chemistry B 105, 2815 2001. 95. Jeon, M. S.; Yoon, W. S.; Joo, H.; Lee, T. K.; Lee, H., Applied Surface
Science 165, 209 2000. 96. Iwasaki, M.; Hara, M.; Kawada, H.; Tada, H.; Ito, S., Journal of Colloid
and Interface Science 224, 202 2000. 97. Subrahmanyam, M.; Rao, K. V. S.; Babu, M. R.; Srinivas, B., Indian
Journal of Environmental Protection 18, 266 1998. 98. Martin, S. T.; Morrison, C. L.; Hoffmann, M. R., Journal of Physical
Chemistry 98, 13695 1994. 99. Bickley, R. I.; Gonzalez-Carreno, T.; Gonzalez-Elipe, A. R.; Munuera,
G.; Palmisano, L., Journal of the Chemical Society, Faraday Transactions 90, 2257 1994.
100. Zhao, Z.; Liu, Q., Catalysis Letters 124, 111 2008. 101. Yang, X.; Cao, C.; Hohn, K.; Erickson, L.; Maghirang, R.; Hamal, D.;
Klabunde, K., Journal of Catalysis 252, 296 2007. 102. Kubacka, A.; Fuerte, A.; Martinez-Arias, A.; Fernandez-Garcia, M.,
Applied Catalysis, B: Environmental 74, 26 2007. 103. Mohamed, M. M.; Al-Esaimi, M. M., Journal of Molecular Catalysis A:
Chemical 255, 53 2006. 104. Serpone, N.; Lawless, D.; Disdier, J.; Herrmann, J.-M., Langmuir 10, 643
1994.
202
105. Allen, N. S., Degradation and Stabilisation of Polyolefins; Applied Science Publishers: New York, 1983.
106. Allen, N. S.; Chirinis-Padron, A.; Henman, T. J., Polymer Degradation and Stability 13, 31 1985.
107. Allen, N. S.; Edge, M.; Corrales, T.; Catalina, F., Polymer Degradation and Stability 61, 139 1998.
108. Allen, N. S.; Edge, M.; Corrales, T.; Childs, A.; Liauw, C. M.; Catalina, F.; Peinado, C.; Minihan, A.; Aldcroft, D., Polymer Degradation and Stability 61, 183 1998.
109. Allen, N. S.; Edge, M.; Holdsworth, D.; Rahman, A.; Catalina, F.; Fontan, E.; Escalona, A. M.; Sibon, F. F., Polymer Degradation and Stability 67, 57 2000.
110. Allen, N. S.; Katami, H., Polymer Degradation and Stability 52, 311 1996.
111. Corrales, T.; Peinado, C.; Allen, N. S.; Edge, M.; Sandoval, G.; Catalina, F., Journal of Photochemistry and Photobiology A: Chemistry 156, 151 2003.
112. Zeynalov, E. B.; Allen, N. S., Polymer Degradation and Stability 86, 115 2004.
113. Zeynalov, E. B.; Allen, N. S., Polymer Degradation and Stability 91, 931 2006.
114. Allen, N. S.; Barcelona, A.; Edge, M.; Wilkinson, A.; Merchan, C. G.; Ruiz Santa Quiteria, V., Polymer Degradation and Stability 86, 11 2004.
115. Zan, L.; Fa, W.; Wang, S., Environmental Science and Technology 40, 1681 2006.
116. Pospísil, J.; Pilar, J.; Billingham, N. C.; Marek, A.; Horák, Z.; Nespurek, S., Polymer Degradation and Stability 91, 417 2006.
117. Parris, D. R., Proceedings of International Wire and Cable Symposium 38th, 105 1989.
118. Al-Madfa, H.; Mohamed, Z.; Kassem, M. E., Polymer Degradation and Stability 62, 105 1998.
119. Roy, P. K.; Titus, S.; Surekha, P.; Tulsi, E.; Deshmukh, C.; Rajagopal, C., Polymer Degradation and Stability 93, 1917 2008.
120. Roy, P. K.; Surekha, P.; Rajagopal, C.; Choudhary, V., Journal of Applied Polymer Science 108, 2726 2008.
121. Roy, P. K.; Surekha, P.; Rajagopal, C.; Choudhary, V., eXPRESS Polymer Letters 1, 208 2007.
122. Roy, P. K.; Surekha, P.; Rajagopal, C.; Chatterjee, S. N.; Choudhary, V., Polymer Degradation and Stability 92, 1151 2007.
123. Roy, P. K.; Surekha, P.; Rajagopal, C.; Choudhary, V., Journal of Applied Polymer Science 103, 3758 2007.
124. Nagle, D. J.; Celina, M.; Rintoul, L.; Fredericks, P. M., Polymer Degradation and Stability 92, 1446 2007.
125. Suarez, J. C. M.; Mano, E. B., Polymer Degradation and Stability 72, 217 2001.
126. Wang, S.; Yu, J.; Yu, J., Journal of Polymers and the Environment 14, 65 2006.
127. Davis, G., Materials Characterization 51, 147 2003. 128. Zhao, G.; Tian, Q.; Liu, Q.; Han, G., Surface and Coatings Technology
198, 55 2005.
203
129. Zeynalov, E. B.; Allen, N. S.; Calvet, N. L.; Stratton, J., Dyes and Pigments 75, 315 2007.
130. Zeynalov, E. B.; Allen, N. S., Polymer Degradation and Stability 91, 3390 2006.
131. Zeynalov, E. B.; Allen, N. S., Polymer Degradation and Stability 85, 847 2004.
132. Dick, C. M.; Liggat, J. J.; Snape, C. E., Polymer Degradation and Stability 74, 397 2001.
133. Maxwell, R. S.; Chambers, D.; Balazs, B.; Cohenour, R.; Sung, W., Polymer Degradation and Stability 82, 193 2003.
134. Palmas, P.; Colsenet, R.; Lemarié, L.; Sebban, M., Polymer 44, 4889 2003.
135. Assink, R. A.; Celina, M.; Minier, L. M., Journal of Applied Polymer Science 86, 3636 2002.
136. Assink, R. A.; Gillen, K. T.; Sanderson, B., Polymer 43, 1349 2002. 137. Gillen, K. T.; Assink, R. A.; Bernstein, R., Polymer Degradation and
Stability 84, 419 2004. 138. Assink, R. A.; Celina, M.; Gillen, K. T.; Clough, R. L.; Alam, T. M.,
Polymer Degradation and Stability 73, 355 2001. 139. Somers, A. E.; Bastow, T. J.; Burgar, M. I.; Forsyth, M.; Hill, A. J.,
Polymer Degradation and Stability 70, 31 2000. 140. Stivala, S. S., Journal of Polymer Science, Part A: General Papers 3, 4299
1965. 141. George, G. A.; Celina, M.; Vassallo, A. M.; Cole-Clarke, P. A., Polymer
Degradation and Stability 48, 199 1995. 142. Bertie, J. E. In Handbook of Vibrational Spectroscopy; Chalmers, J. M.;
Griffiths, P. R., Eds.; John Wiley and Sons Ltd: Chichester, 2002. 143. Harrick, N. J., Applied Spectroscopy 31, 548 1977. 144. Hannah, R. W. In Handbook of Vibrational Spectroscopy; Chalmers, J.
M.; Griffiths, P. R., Eds.; John Wiley and Sons Ltd: Chichester, 2002. 145. Chambon, F.; Srinivas, S., Journal of Plastic Film & Sheeting 15, 297
1999. 146. Ajji, A.; Zhang, X.; Elkoun, S., :Annual Technical Conference - Society
of Plastics Engineers 62nd, 2045 2004. 147. Noda, I.; Dowrey, A. E.; Marcott, C., Modern Polymer Spectroscopy, 1
1999. 148. Choi, H. S.; Shikova, T. G.; Titov, V. A.; Rybkin, V. V., Journal of
Colloid and Interface Science 300, 640 2006. 149. Kaczmarek, H.; Oldak, D.; Malanowski, P.; Chaberska, H., Polymer
Degradation and Stability 88, 189 2005. 150. Stark, N. M.; Matuana, L. M., Polymer Degradation and Stability 86, 1
2004. 151. Delor, F.; Barrois-Oudin, N.; Duteurtre, X.; Cardinet, C.; Lemaire, J.;
Lacoste, J., Polymer Degradation and Stability 62, 395 1998. 152. Do, T.-T.; Celina, M.; Fredericks, P. M., Polymer Degradation and
Stability 77, 417 2002. 153. Chalmers, J. M.; Hannah, R. W.; Mayo, D. W. In Handbook of
Vibrational Spectroscopy; Chalmers, J. M.; Griffiths, P. R., Eds.; John Wiley and Sons Ltd: Chichester, 2002.
204
154. Sommer, A. J. In Handbook of Vibrational Spectroscopy; Chalmers, J. M.; Griffiths, P. R., Eds.; John Wiley and Sons Ltd: Chichester, 2002.
155. Fitzpatrick, J.; Reffner, J. A. In Handbook of Vibrational Spectroscopy; Chalmers, J. M.; Griffiths, P. R., Eds.; John Wiley and Sons Ltd: Chichester, 2002.
156. Fredericks, P. M. In Handbook of Vibrational Spectroscopy; Chalmers, J. M.; Griffiths, P. R., Eds.; John Wiley and Sons Ltd: Chichester, 2002.
157. Koenig, J. L., Microspectroscopic Imaging of Polymers American Chemical Society: New York, 1998.
158. Chan, K. L. A.; Hammond, S. V.; Kazarian, S. G., Analytical Chemistry 75, 2140 2003.
159. Chan, K. L. A.; Kazarian, S. G., Applied Spectroscopy 57, 381 2003. 160. Chan, K. L. A.; Kazarian, S. G.; Mavraki, A.; Williams, D. R., Applied
Spectroscopy 59, 149 2005. 161. van der Weerd, J.; Kazarian, S. G., Applied Spectroscopy 58, 1413 2004. 162. Sommer, A. J.; Tisinger, L. G.; Marcott, C.; Story, G. M., Applied
Spectroscopy 55, 252 2001. 163. Williams, G. P. In Handbook of Vibrational Spectroscopy; Chalmers, J.
M.; Griffiths, P. R., Eds.; John Wiley and Sons Ltd: Chichester, 2002. 164. Jilkine, K.; Gough, K. M.; Julian, R.; Kaminskyj, S. G. W., Journal of
Inorganic Biochemistry 102, 540 2008. 165. Miller, L. M.; Dumas, P., Biochimica et Biophysica Acta (BBA) -
Biomembranes 1758, 846 2006. 166. Kaminskyj, S.; Jilkine, K.; Szeghalmi, A.; Gough, K., FEMS
Microbiology Letters 284, 1 2008. 167. Wetzel, D. L.; Carter, R. O., III, AIP Conference Proceedings 430, 567
1998. 168. Feuilloley, P.; Cesar, G.; Benguigui, L.; Grohens, Y.; Pillin, I.; Bewa, H.;
Lefaux, S.; Jamal, M., Journal of Polymers and the Environment 13, 349 2005.
169. Prasad, A.; Mowery, D., Annual Technical Conference - Society of Plastics Engineers 55th, 2310 1997.
170. Jin, C.; Christensen, P. A.; Egerton, T. A.; Lawson, E. J.; White, J. R., Polymer Degradation and Stability 91, 1086 2006.
171. Zhiyong, Y.; Mielczarski, E.; Mielczarski, J.; Laub, D.; Buffat, P.; Klehm, U.; Albers, P.; Lee, K.; Kulik, A.; Kiwi-Minsker, L.; Renken, A.; Kiwi, J., Water Research 41, 862 2007.
172. http://www.resinex.be/Datasheets/Dowlex/DOWL_NG5056E.pdf. 2007. 173. Sharma, N.; Chang, L. P.; Chu, Y. L.; Ismail, H.; Ishiaku, U. S.; Mohd
Ishak, Z. A., Polymer Degradation and Stability 71, 381 2001. 174. Tarantili, P. A.; Kiose, V., Journal of Applied Polymer Science 109, 674
2008. 175. Creagh, D.; Tobin, M.; Broadbent, A.; McKinlay, J., AIP Conference
Proceedings 879, 615 2007. 176. Creagh, D.; McKinlay, J.; Dumas, P., Vibrational Spectroscopy 41, 213
2006. 177. Eren, T.; Okte, A. N., Journal of Applied Polymer Science 105, 1426
2007. 178. Nakayama, N.; Hayashi, T., Composites Part A: Applied Science and
Manufacturing 38, 1996 2007.
205
179. Fa, W.; Zan, L.; Gong, C.; Zhong, J.; Deng, K., Applied Catalysis B: Environmental 79, 216 2008.
180. Rennel, C.; Rigdahl, M., Colloid and Polymer Science 272, 1111 1994. 181. Belyi, V. N.; Malevich, V. L.; Shraiber, Y.; Khilo, N. A., Journal of
Applied Spectroscopy 75, 199 2008. 182. Chambon, E.; Florentin, E.; Torchynska, T.; Gonzalez-Hernandez, J.;
Vorobiev, Y., Microelectronics Journal 36, 514 2005. 183. Cran, M. J.; Bigger, S. W., Applied Spectroscopy 57, 928 2003. 184. Salvalaggio, M.; Bagatin, R.; Fornaroli, M.; Fanutti, S.; Palmery, S.;
Battistel, E., Polymer Degradation and Stability 91, 2775 2006. 185. Schachtschneider, J. H.; Snyder, R. G., Spectrochimica Acta 19, 117
1963. 186. Snyder, R. G.; Schachtschneider, J. H., Spectrochimica Acta 19, 85 1963. 187. Lin-Vien, D.; Colthup, B. N.; Fateley, W. G.; Grasselli, J. G., Infrared
and Raman Characteristic Frequencies of Organic Molecules; Academic Press: London, 1991.
188. Lawson, E. E.; Edwards, H. G. M.; Johnson, A. F., Spectrochimica Acta Part A: Molecular and Biomolecular Spectroscopy 51, 2057 1995.
189. Farmer, V. C.; Editor, Mineralogical Society Monograph 4: The Infrared Spectra of Minerals, 1974.
190. Hummel, D. O., Atlas of Plastics Additives: Analysis by Spectrometric Methods, by Dietrich O. Hummel, 2003.
191. Yang, R.; Liu, Y.; Yu, J.; Wang, K., Polymer Degradation and Stability 91, 1651 2006.
192. Blakey, I.; Billingham, N.; George, G. A., Polymer Degradation and Stability 92, 2102 2007.
193. Tyler, D. R., Journal of Macromolecular Science, Polymer Reviews C44, 351 2004.
194. Yokoyama, S.; (Tomoegawa Paper Co., Ltd., Japan). JP: 2009, p 12pp. 195. Levenson, E.; Lerch, P.; Martin, M. C., Journal of Synchrotron Radiation
15, 323 2008.
206