EFFECTS OF PHOTO OXIDATION ON THE MECHANICAL BEHAVIOR OF POLYMERIC

MATERIALS: AN EXPERIMENTAL STUDY

By

Abd-Elrahman Korayem

A THESIS

Submitted to

Michigan State University

in partial fulfillment of the requirements

for the degree of

Civil Engineering—Master of Science

2021

ABSTRACT

EFFECTS OF PHOTO OXIDATION ON THE MECHANICAL BEHAVIOR OF POLYMERIC 

MATERIALS: AN EXPERIMENTAL STUDY

By

Abd-Elrahman Korayem

Here, an experimental study is carried out to understand the degradation of crosslinked poly-

meric materials by exposure to coupled thermo- and photo-oxidation, and their effects on me-

chanical behavior of the material. Our investigation specifically focuses on two cured elastomeric 
adhesives: (i) a silicon-based and (ii) a polyurethane-based adhesive. Samples were prepared and 
subjected to UV radiation at three different temperatures namely 45, 60 and 80◦C, for different ag-
ing periods ranging from 1 to 150 days. Chemical damage is characterized by the FTIR, DSC and 
Swelling tests which help us to monitor the change in the chemical composition of the materials 
along the aging durations and conditions. Mechanical damage is characterized by uniaxial tensile 
failure test, cyclic tensile test and permanent set test. These tests revealed the change of behavior 
of our materials along the aging periods and conditions. The changes in mechanical behavior vary 
depending on the material, aging temperature and duration. We show that behavior of the materi-
als aged in the photo-oxidative environment is in fact a result of a combination of two separable 
modes of damage, Mechanical and Environmental. Accordingly, We are able to further separate 
the environmental damage mode into one created by the effects of temperature and another cre-

ated by the combination of temperature and radiation. This separability has been confirmed with a 
Neo-hookean model, which proved that for a combined state of aging such as photo-oxidation the 
material cannot be characterized by models which consider only single modes of damage. Such 
regimes would require approaches such as the micro-mechanical approach to fully consider the 
effects by all the damage modes found in complex aging phenomena.

iv

v

1

5
7
10

11
13
15
17
17
18
20
20
22
24

28

30

32

33

TABLE OF CONTENTS

LIST OF TABLES

LIST OF FIGURES

1 

INTRODUCTION

2  MATERIALS AND METHODOLOGY
. .

2.1 TESTING . . . . . . . . . .
.
2.2 Mechanical and environmental Damage .

.

.

.

.

.
.

.
.

. .
.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

.
.

3  RESULTS AND DISCUSSION

.
SA . . . . . . . . . . . .

.
3.1 Mechanical Damages in Aged samples .
3.2 Separation of Mechanical & Environmental damage .
.
3.3 Observations . . . . . . . . . .
.

.
.
.
.
.
.
. .
3.3.1
3.3.2  PUA  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .  .
.
.
.
.

Silicon polymer
3.5 DSC test . . . . . . . .
3.6 Swelling test . . . . . . . . .

.
.
. . . .
.
.
.
. .

. . . .
.

3.4 FTIR . . . . . . . . . . . . . .

.
. .
.
.
.
. .
.
.
. .

.
.
.
.
. .
.
.

. .
.
.
.
.
.
.

. .
.
.
.
.
.
.

3.4.1

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.

.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.
.
.
.

.

.
.

.
.
.
.

.

.
.

.
.
.
.

.
.

.
.

.
.

.
.

.
.
.
.

. .

.
.

.
.

.
.
.
.

.
.
.
.

.
.
.
.

.

.
.

.
.
.
.

.
.
.
.

4  MODELLING

5  SIGNIFICANCE IN MODELLING 6 

6  CONCLUSION

REFERENCES

iii

LIST OF TABLES

Table 1:

Properties of study materials .

.

.

.

Table 2:

Aging temperatures and durations .

Table 3:

Cyclic test strain limits .

.

.

.

.

.

.

.

.

.

.

.

.

Table 4:

Table 5:

SA FTIR functional groups detected .

PUA FTIR functional groups detected .

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

.

5

7

8

20

21

iv

LIST OF FIGURES

Figure  1: Full network shown broken down into sub-networks, each susceptible to a 
unique damage mode. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure  2:  a)  Mold  used  for  casting  of  samples  capable  of  producing  20  samples  per 
batch, b) Labelled samples, c) Samples dimensions being taken. . . . . . . . . . . . . . . . . . . . . . .

Figure  3: a) QUV machine used for aging, b) Modified adjustable sample size fixture 
created for our experiments. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 4: a) UTM machine used for testing along with roller grips used for soft material 
gripping, b) Cyclic test template with no repetition of cycle. . . . . . . . . . . . . . . . . . . . . . . . . . 

Figure  5:  Apparatus  used  to  hold  the  samples  at  the  desired  strain  during  the  aging 
period . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 6: a) FTIR scanning machine used, b) DSC Q2000 series machine. . . . . . . . . . . . . . 
Figure 7: a) Failure test of SA aged at 45◦C, b) 60◦C, and c) 80◦C for different aging 
periods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 

Figure  8:  Cyclic  tensile  test  at  constant  and  increasing  amplitudes  for  SA  aged  at  a) 
45◦C, b) 60◦C, and c) 80◦C for different aging periods. . . . . . . . . . . . . . . . . . . . . . . . . .  
Figure  9:  a)  Failure  tensile  test  of  PUA  aged  at  45◦C,  b)  60◦C,  and  c)  80◦C  for 
different aging periods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure  10:  a)  Cyclic  tensile  test  of  PUA  aged  at  45◦C,  b)  60◦C,  and  c)  80◦C  for 
different aging periods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 11: a) Permanent set results for SA strained continuously at 45, 60 and 80  C. for 
different  aging  periods,  b)  Showing  permanent  set  results  for  PUA  strained 
continuously at 45, 60 and 80 ◦C. for different aging periods. . . . . . . . . . . . . . . . . . . . . . .

◦

Figure  12:  a)  Change  in  ultimate  strength  of  SA  along  time  for  the  different  aging 
temperatures of 45, 60 and 80 ◦C. b) Showing the change in ultimate stretch of SA 
along time for the different aging temperatures of 45, 60 and 80 ◦C. . . . . . . . . . . . . . . . . .

Figure  13:  a)  Change  in  ultimate  strength  of  PUA  along  time  for  the  different  aging 
temperatures of 45, 60 and 80 ◦C. b) Showing the change in ultimate stretch of PUA 
along time for the different aging temperatures of 45, 60 and 80 ◦C. . . . . . . . . . . . . . . . . . 

4

6

7

8

9

9

11

11

12

12

13

14

15

Figure 14: Different parameters being calculated from the cyclic test. . . . . . . . . . . . . . . . . .      16

Figure 15: Changes a) in σ* across aging conditions at λ1 in the cyclic test b)in e*
, 
and  c)  in  W*  during  the  first  load  cycle  against  aging  time  at  λ1  in  the  cyclic 
test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17

v

Figure 16: a) Stress softening across different cycles for SA at 45, 60 and 80 ◦C, where 
λn  is  the  stretch  limit  of  cycle  n  in  the  cyclic  test  b)  Showing  relative  residual  strain 
across different cycles for SA at 45, 60 and 80 ◦C, where λn is the stretch limit of cycle n 
in the cyclic test. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

17

18

18

20

Figure 17: a) Change in values of σ* across aging conditions during the first cycle with 
time. b) Showing the change in e∗. c) Showing the change of the W* during the first 
cycle with time. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 
Figure 18: a) Stress softening across different cycles for PUA at 45, 60 and 80 ◦C, where 
λn  is  the  stretch  limit  of  cycle  n  in  the  cyclic  test  b)  Showing  relative  residual  strain 
across different cycles for PUA at 45, 60 and 80 ◦C, where λn is the stretch limit of cycle 
n in the cyclic test. . . . . . . . . . . . . .  . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure  19:  FTIR  scan  of  SA  aged  at  45◦C,  60◦C,  80◦C  for  different  durations 
imposed over an unaged sample results. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure  20:  FTIR  scan  of  PUA  aged  at  45◦C,  60◦C,  80◦C  for  different  durations, 
imposed over an unage sample scan . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Figure 21: Heat flow plot generated from DSC scan of SA aged at 45◦C, 60◦C, 80◦C 
for different durations imposed over an unaged sample results. . . . . . . . . . . . . . . . . . . . . . . 
Figure 22: Heat capacity plot generated from DSC scan of SA aged at 45◦C, 60◦C, 
80◦C for different durations imposed over an unaged sample results. . . . . . . . . . . . . . . . .   24
Figure  23:  Heat  flow  plot  generated  from  DSC  scan  of  PUA  aged  at  45◦C,  60◦C, 
80◦C for different durations imposed over an unaged sample results. . . . . . . . . . . . . . . . . 

21

23

25

Figure 24: Heat capacity plot generated from DSC scan of PUA aged at 45◦C, 60◦C, 
80◦C for different durations imposed over an unaged sample results. . . . . . . . . . . . . . . . .

Figure 25: Change in a) Endothermic melting temperature for SA and b) glassification 
temperature for PUA at 45◦C, 60◦C, 80◦C for different aging durations. . . . . . . . . . . 

Figure 26: Cross-link density of a) SA and b) PUB elastomers at 45◦C, 60◦C, 80◦C 
for different durations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  

Figure  27:  Results  of  the  a)  Projected(P)  results  of  the  failure  test  vs  the 
Experemental(E)  results  for  the  SA  material  at  45◦C,  b)  60◦C,  and  c)  80◦C  for 
different aging periods. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 28: Results of the a) Projected(P) results of the failure test vs the Experimental(E) 
results for the PUA material at 45◦C, b) 60◦C, and c) 80◦C for different aging 
periods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

26

27

27

28

29

vi

1. INTRODUCTION

Due to excellent thermal and chemical resistance of polymers, they are abundantly used in various

applications(Bower, 2002). However, long time exposure to a variety of environmental conditions

along with mechanical damages can cause severe damage to polymer structure and limits its per-

formance over time (Mohammadi et al., 2020). This process is often referred to as aging. Major

types of aging are thermo-oxidative (Mohammadi and Dargazany, 2019), hygrothermal (Nachtane

et al., 2019), hydrolysis (Bahrololoumi et al., 2020) and photo-oxidative aging (Rodriguez et al.,

2020).

Over the lifetime of polymeric materials, photo-oxidative aging becomes a continuous risk to

the performance of equipment and structures that are subjected to UV radiation such as that present

in sunlight(Geuskens and David)(Blaustein and Searle, 2013). With the growing use of polymers

in nearly every application in our daily lives(Chandra and Rustgi), photo-oxidation becomes an

ever more pressing issue for the sustainable use of polymers in their working environments over

extended periods of time (Namazi, 2017). Elastomers are a special class of polymers with a highly

nonlinear constitutive behaviour, which makes their photo-oxidative deterioration much harder to

predict than a linearly behaving material (Rezig et al., 2020). Given the complex molecular struc-

tures that can emerge from aging and the countless effects additives could have on the unaged

and aged material, achieving a robust framework for behavior prediction can be a great challenge.

Additionally, polymeric substances can react differently to photo-oxidation when mixed together

or with other additives that are used as fillers or curing agents(Bigg, 1987)(Khabbaz et al., 1999)

(Liu et al., 2020). This adds a further challenge for achieving a generalized form for characterizing

elastomers. In elastomers, polymer chains consist mainly of a variety of cross-linked hydrocar-

bon chains that are bonded together either through physical bonds forming as hydrogen bonds or

chemical bonds forming as covalent bonds(Flory, 1953)(Hiemenz and Lodge, 2007).

A UV ray carries sufficient energy to cause photolytic degradation of the C-C backbone or

1

cross-links between the carbon chains(Biron, 2015)(Grossman and Gouzman, 2003). This is most

apparent at shorter wavelengths, which are characterized by having higher energy and is sufficient

to cause degradation(Andrew R. George, 2008). This in turn destabilizes the compound and in-

creases it’s reactivity, allowing for rapid oxidation and deterioration of the bonds that hold the

structure of the polymer matrix intact(Tolinski, 2015). The deterioration can cause significant

changes to the mechanical characteristics of polymeric materials, which may be very damag-

ing to the performance of equipment and parts if not considered (Feldman, 2002). Accordingly,

higher energy UV spectra coming from shorter wavelengths found in sunlight, affect polymers the

most(Biron, 2015), so the amount of expected damage to polymeric components due to photo-

oxidation must be considered for products designed for outdoor applications. Another phenomena

that must also be considered is thermo-oxidation, as UV radiation from sunlight is accompanied

with infrared radiation which is experienced by bodies as heat(Bhatia).

Thus for a given application involving UV light and heat, we find that polymers can be affected

through two main mechanisms, i) the damage caused by the heat energy and/or that of UV radia-

tion, which releases free radicals into the matrix, and ii) Oxidation, which occurs due to oxygen

in the environment reacting with the free radicals released (Audouin et al., 1998). The presence

of free-radicals from the heat or UV energy have a destabilizing effect on the matrix, thus each

mode of damage has an amplifying effect on the other, which causes further damage to the mate-

rial. Accordingly, for any full characterization of the actual deterioration that occurs in polymeric

materials, it is crucial that investigative efforts focus on the combination of photo-thermal aging

instead of simply photo induced aging.

Previous studies on the topic usually focus on the general behavior of the aged material, along

with the evolution of certain chemical properties of interest for characterization, such as in the

case of Signor et al(Signor et al., 2003), Pickett(Pickett, 2017), Dupuis et al(Dupuis et al., 2017)

and Wu et al(Wu et al., 2020). Other focuses were directed towards using the Arrhenius functions

which were based on activation energies for predicting shift factors for time-temperature super po-

2

sition and kinetic studies, as in the case of (Cruz-Pinto et al., 1994), (Liu et al., 2018), (Zaghdoudi

et al., 2020) and (Rezig et al., 2020). Most studies on thermo-oxidation or photo-oxidation can

be categorized into two groups i) characterizing the chemical response to correlate the chemical

variations to the mechanical variation, or ii) developing an empirical approach to construct a mas-

ter curve which is extremely difficult in case of coupled thermo-radiative aging due to convoluted

degradation path(Celina, 2013)(Gillen et al., 1997) . Even though Both systems might function

for a certain material or a given aging condition, they cannot be used for generating a generalized

form for all materials in a certain category, such as elastomers. This is due to the variability in

chemical composition and micro-structure between materials, which in turn influences their re-

sponse to different environmental and mechanical stresses(McKeen, 2018)(Rubinstein and Colby,

2003). A further complexity arises from the high variability in materials’ capacity for diffusion

limited oxidation(DLO) due to the differences in their spatial heterogeneity in oxidative degra-

dation. This disparity affects their capacity for oxygen diffusion into the matrix and consumption

during photo-thermo induced oxidative aging, which makes the accurate prediction of the products

of the oxidation from a chemical perspective more complex (Quintana and Celina, 2018). (Gillen

et al., 2000) showed that there exists a synergism between mechanical loading and the amount of

potential environmental damage that can occur in a sample. This shows, that while both of these

modes of damage act on different aspects of the material, they potentially have a complementary

effect that must be isolated in order to properly characterize the materials of interest.

A solution to this issue can be found in micro-mechanical modelling, as instead of trying to fit

a single equation to all materials, a more discrete approach is performed. This method involves

only considering the general effect of each aging factor on a single micro-structural component

of the material and then compounding it up-til the bulk structure (Sadd, 2018). Accordingly, to

improve the effectiveness and accuracy of the formulation created for a certain aging type, it is

necessary to investigate which individual aging factors contribute the most to the behavior of the

material(Mohammadi and Dargazany, 2018)(Bahrololoumi et al., 2019 (in submission).

3

In this study, we investigate the mechanism of damage accumulation for materials exposed to

coupled mechanical and environmental damages modes, and which factors plays a more significant

role on affecting the behavior of two elastomeric materials in each of those modes. Accordingly,

we will study how to superpose mechanical damages such as deformation induced softening with

environmental damage that is induced by couple thermal-radiative damage. We seek to verify our

hypothesis through experimental efforts which investigate the change in mechanical and chemical

properties and behavior of the materials of this study. To investigate this separability the damage

modes are hypothesized to operate on two distinct networks. A Network containing C-C chains,

crosslinked together chemically. The other Network is devoid of chemical crosslinking and com-

prises of C-C chains only. A representation of this can be seen in Figure 1.

Figure 1: Full network shown broken down into sub-networks, each susceptible to a unique damage mode.

4

2. MATERIALS AND METHODOLOGY

The material selection process involved considering different variants of hard and soft polymers.

Our greatest interest was in observing the change in elastic behavior of polymers and accordingly

selecting materials with high visco-elastic properties was our prime interest. Different variations of

silicon adhesive, acrylic, epoxy, polyurethane and other base compounds were investigated, aged

and tested in the expected environmental and mechanical loading conditions that the study would

expand on. Accordingly, we were able to narrow down the material list to four materials initially,

then they were further reduced to two materials which were most representative of the behavior

that would be useful to the study. The study focuses on two variants of elastomeric adhesives,

namely silica-based and polyurethan based adhesives with properties given in Table 1.

Table 1: Properties of study materials

Hardness

Material

Un-Aged

Shore Hardness

Un-Aged
Modulus
Un-Aged

Cured strength

Un-Aged

Elongation at break

Soft

Softer

Polyurethane adhesive

(PUA)

60
(A)

3.2 MPa

7.6 MPa

375 %

Silicone Adhesive

(SA)
32
(A)

0.5 MPa

2.5 MPa

680 %

Sample preparation  was performed according to the ASTM D-412 (Standard, 2006). The sam-
ples were cast separately in a custom mold in contrast to being punched out of a single sheet. This 
is achieved by injecting the adhesive into the mold using a commercial applicator one cavity at a 
time.  Excess material is then removed, and the samples are left to cure over a period of 5 days, 
after which they are removed from the mold using full-body holders to avoid mechanical damage 
prior to the aging process.  To verify the optimal curing and avoid over-curing, the samples were

5

tested at different curing times, to which it was observed that behavior does not change beyond 5

days of curing in room temperature for up to 30 days. Samples were labelled, weighed, measured

and photographed before and after the aging process to track any change that could occur to the

materials due to photo-oxidation.

a)

b)

c)

Figure 2: a) Mold used for casting of samples capable of producing 20 samples per batch, b) Labelled samples, c)
Samples dimensions being taken.

Aging environment: The samples are separately aged in accordance to the ASTM G-151(Standard,

2019) at an irradiation of 1 W/m2/ nm wavelength for three aging durations, 1, 10 and 30 days at
three temperatures, 45, 60 ◦C and 80 ◦C.The reasoning behind the temperature selection is that
we required at least three distinct temperatures to monitor the behavior of the materials with in-

creasing temperatures and investigate whether there is a trend to be witnessed or not. Thte QUV

machine that we use for aging the samples has a minimum temperature setting of 45oC and a max-

imum setting of 80oC. The UV spectrum the samples were subjected to, was one generated by

UVA-340 lamps that cover wavelengths ranging from 295 – 340 nm. These wavelength represent

the greater portion of radiation intensity for sunlight, and accordingly are the most appropriate to

use in the study of the photo-oxidative effects from sunlight(Feldman, 2002). In order to fully un-

derstand the effects pure photo-oxidation have on polymeric materials the samples that were aged
at 45 ◦C were additionally aged to 60 and 150-day periods. This will allow us to properly isolate
the effects of photo-thermal oxidation experienced at the higher temperatures from those of pure

photo-oxidation.

6

a)

b)

Figure 3: a) QUV machine used for aging, b) Modified adjustable sample size fixture created for our experiments.

Table 2: Aging temperatures and durations

Aging Temperature

[◦C]

Aging Duration

[days]

45

60

80

1-10-30-60-150 1-10-30-60-150 1-10-30-60-150

2.1.  TESTING

Uniaxial Tensile tests  were preformed to identify deterioration in material performance accord-
ing aging environment.  To achieve this, specially designed tensile tests were used to analyze the 
behavior of the samples, and were conducted at room temperature.

• Failure test where the samples are continuously stretched until failure, are performed to

derive the constitutive behavior, ultimate strength σ f , ultimate strain 
 f , and toughness T.

• Cyclic test @ increasing amplitude were carried out to capture mechanical damages induced

by stretch history λmax. In essence, the test can capture the evolution of permanent set e1,
energy dissipation D with respect to λmax. Four stretch amplitudes were chosen based on
the initial failure test as shown in Table 3.

All tensile tests were performed on a Test Resources Universal Testing Machine (UTM), Model

311-39 Dual Column Test Machine with a 2,250 lb (10 kN) capacity. The tests were also performed

7

Table 3: Cyclic test strain limits

Material

Strain Limits

Revised Strain Limits

PUA

SA

112.5, 225,
337.5, 405

35, 75,

112.5, 130

87.5, 175,
262.5, 315
45, 87.5,
125, 175

at 50 mm/min extension rate, which was determined experimentally to be the highest rate of loading

to not have an effect on the behavior of the material

a)

b)

Figure  4:  a)  UTM  machine  used  for  testing  along  with  roller  grips  used  for  soft  material  gripping,  b)  Cyclic  test 
template with no repetition of cycle.

Permanent  set  tests  were  designed  and  carried  out  to  capture  the  evolution  of  e1  the  non-
vanishing  residual  strain,  as  a  measure  of  damage  with  respect  to  different  environmental  and 

mechanical loads.  In those tests, virgin samples were initially stretched to 50% strain and held at 
that extension during aging condition as shown in Figure 5.  The stretched samples were placed
in  the  same  aging  conditions  used  for  the  mechanical  testing  for  different  periods.  They  were 
then subsequently removed after the periods had concluded, weighed and measured to inspect the

residual and permanent strain which had developed.

8

Figure 5: Apparatus used to hold the samples at the desired strain during the aging period

FTIR tests were conducted to characterize changes in chemical composition of the matrix in the 
course of photo-oxidative aging using a JASCO FTIR/IR 4000 series machine. Each material was 
scanned with a different refractive crystal and bandwidth due to the difference in transmittance 

properties  between  them.  SA  was  scanned  with  a  Diamond  crystal  under  a  bandwidth  of  400-
4000 cm-1.  PUA on the other hand was scanned with a Germanium crystal under a bandwidth of 
600-4000 cm-1  due to the significant noise it causes in the 400-600 c m-1 range.

a)

b)

Figure 6: a) FTIR scanning machine used, b) DSC Q2000 series machine.

DSC test was performed on both materials to track the change in heat capacity and flow of the 
materials throughout the aging process.  The test was conducted on a Q2000 series Differential 
Scanning Calorimetry machine from TA Instruments, with a temperature scan between 80oC and 
350oC at a heating rate of 10oC/min. Small sections were cut from the aged samples and loaded in

9

the machine where a temperature sweep was performed on them. This test was conducted to give

us insight into the behavior of the material across a wide array of temperatures and enables us to

determine the glassification temperature for PUA along with the Endothermic melting temperature

for SA considering that the glassification temperature for SA is estimated to be near -170oC, which

is too low and cannot be achieved in the DSC test.

Swelling test was performed to characterize the changes in the cross-link density of the matrix

in the course of photo-oxidative aging by submerging the samples into three different solvents

namely, Hexane, Toluene and Pentane until the polymer matrix is saturated with the solvent. The

saturation limit is determined by the complexity and level of entanglement of the polymer matrix

due to cross-linking.

2.2. Mechanical and environmental Damage

Damage induced by environmental and mechanical factors has an accumulating nature. This fea-

ture allows us to identify damage by tracing the incremental changes in the materials with respect

to the aging durations and conditions. A major challenge, however, is to properly isolate the effects

of different damage mechanisms on the constitutive behavior of the materials. Of particular inter-

est is firstly to separate mechanical and environmental effects, and secondly to separate photo-

and thermo-oxidation, which can be achieved through the selected conditions, as the effects of
thermo-oxidation are least apparent at 45 ◦C and most apparent at 80 ◦C. Thermo-oxidative effects
are well studied in literature and accordingly the effects of photo-oxidation can be more easily

deduced. This can only be achieved however if there is a separability between the damage cause

by mechanical stress from our tests and the damage.

10

3. RESULTS AND DISCUSSION

Silicone Adhesive (SA):

a)

b)

c)

Figure 7: a) Failure test of SA aged at 45◦C, b) 60◦C, and c) 80◦C for different aging periods.

a)
Figure 8: Cyclic tensile test at constant and increasing amplitudes for SA aged at a) 45◦C, b) 60◦C, and c) 80◦C for
different aging periods.

b)

c)

Results for the tensile failure tests of SA are presented in Figure 7 and 8, respectively.

The conditions the samples are being subjected to vary in both temperature and duration of

exposure. Samples have a specific behavioral trend across the aging conditions, as samples exhibit
continuous stiffening at 45◦C, but at higher temperatures the material initially exhibits stiffen-
ing and then softens with longer aging periods. This behavior can be understood by considering

aging as a competition between purely photo-oxidative effects and those of photo-thermal degra-

dation.This can be seen through the lens of initial over-curing of the material, which peaks after

a certain amount of energy is absorbed by the material to achieve full-curing, which is then fol-

lowed directly by a phase of chain detachment sub-mechanism, which continues throughout the

11

0150300450Strain [%]00.51.52.5Stress [MPa]Unaged10 days60 days150 daysAging TrendΘ = 45ºC Θ = 60ºC 0150300450Strain [%]00.51.52.5Stress [MPa]Unaged10 days30 days 1 dayInitial HardeningSoftening with time0150300450Strain [%]00.51.52.5Stress [MPa]Unaged10 days30 days 1 dayInitial HardeningSoftening with timeΘ = 80ºC 0150300450Strain [%]00.51.52.5Stress [MPa]Unaged10 days60 days150 daysAging TrendΘ = 45ºC Θ =60ºC 0150300450Strain [%]00.51.52.5Stress [MPa]Unaged10 days30 days 1 dayInitial HardeningSoftening with time0150300450Strain [%]00.51.52.5Stress [MPa]Unaged10 days30 days 1 dayInitial HardeningSoftening with timeΘ = 80ºC remainder of the aging process(Bettina et al., 2017).  This chain detachment,  which is likely to 
be  accompanied  with  a  loss  in  cross-link  density,  decreases  the  ability  of  the  material  to  resist 
mechanical loading and accordingly causes the material to become softer.

Ultimate strength and stain for silicon adhesive decreases regardless of aging duration, with 
the exception of samples aged at 45◦C  for 10 days, suggesting that the material is experiencing 
some form of over-curing up til this condition.

PUA:

Results for the failure and cyclic tests of PUA are presented in Fig.9 and 10, respectively.

a)

a)

Figure 9: a) Failure tensile test of PUA aged at 45◦C, b) 60◦C, and c) 80◦C for different aging periods.

b)

c)

Figure 10: a) Cyclic tensile test of PUA aged at 45◦C, b) 60◦C, and c) 80◦C for different aging periods.

b)

c)

The material behaves similarly across all temperatures and durations with the exception of
80◦C for 30 days where the material loses its intgrity and there is an unexpected softening effect

12

02468Unaged10 Days60 Days150 DaysAging TrendΘ = 45ºC 0100200300400500Strain [%]Stress [MPa]02468Unaged1 Day10 Days30 DaysInitial StiffeningΘ = 60ºC 0100200300400500Strain [%]Stress [MPa]Softeningwith time0100200300400500Strain [%]02468Stress [MPa]Unaged1 Day10 Days30 DaysInitial StiffeningSofteningwith timeΘ = 80ºC Unaged10 days60 days150 days0200400Strain [%]02468Stress [MPa]Θ = 45ºC Unaged1 day10 days30 days0200400Strain [%]02468Stress [MPa]Θ = 60ºC Unaged1 day10 days30 days0200400Strain [%]02468Stress [MPa]Θ = 80ºC dominating  the  behaviour.  Aging  causes  polyurethane  to  stiffen  with  time,  increases  ultimate 
strength, yet decreases the ultimate strain that the material can be extended to.

UV radiation increases the elastic modulus with the longer aging periods. As it can be seen at 
higher temperatures and longer aging durations, increased heat causes significant damage to the 
polymer as it loses it’s mechanical properties.  With the ultimate strength dropping from 6.4 MPa 
to 4.8 MPa, the ultimate strain from 490% to 170% and toughness as well as shown in Figure 12. 
Therefore, we can assume that the dual effect of photo-thermal aging has a softening effect induced 

by cross-link detachment in contrast to the single mode of thermo-oxidative aging as shown in 
(Al-Azhary et al., 2020), which shows that the same material exhibits a continuously hardening 
behavior at higher temperatures. This can be explained by considering that the combination of both 
heat and radiation contains the sufficient amount of energy to act on the cross-links and entangled 

chains causing scission, which limits the entanglements of the longer chains, causing the material 
to lose it’s ability to resist mechanical loading.  This energy or damage mode however, was not 
available in the case of pure photo-oxidation, which is dominant at lower temperatures.

3.1.  Mechanical Damages in Aged samples

Permanent Set results are shown in Figure 11 for samples aged at 1, 10 and 30 days at 45, 60 and 
80◦C.

a)

b)

Figure 11: a) Permanent set results for SA strained continuously at 45, 60 and 80 ◦C. for different aging periods, b)
Showing permanent set results for PUA strained continuously at 45, 60 and 80 ◦C. for different aging periods.

In both silicon and PU compounds, results show increasing permanent set e1 for longer aging

13

Residual Strain [%]Time [hrs]Time[hrs]Residual Strain [%]PUAperiods, and higher temperatures, with almost similar behaviour between the samples aged at 45◦
and 60◦C, and a clear difference with those aged at 80◦C, which can suggest that the optimum
operating temperatures for the materials should be below 60◦C. Similar observations were made
for increasing stretch amplitude λmax and number of cycles nc, as summarized below

e1 ∝ {T, t, λmax, nc}

D ∝

T, t, λmax,

(1)

(cid:40)

(cid:41)

1
nc

The tests also show that the polymer chains and associated cross-links of the silicon adhesive

(SA) have a relatively high level of thermal stability and are less likely to be damaged or sig-

nificantly altered when subjected to thermal and mechanical stresses. On the other hand, PUA

performance at higher temperatures suggests that either the compounds contains volatile compo-

nents that when subjected to heat can dissipate and are replaced with less stable chains/cross-links,
which causes the drop in material behavior observed at 80 ◦C.

Ultimate strength σ f and strain 
 f continuously change in the course of aging which confirms

changing the contribution of cross-linking formation and detachment at different stages of aging.

Failure properties of the SA and PUA samples were shown in Figure 12 and 13, respectively. In

general, one can see that the total toughness of the material T decreases with respect to aging time,
regardless of the behaviour of the σ f and 
 f . At 45◦C however, initial increase of toughness can
be associated to over-curing, the effect of which gradually fades away.

a)
Figure 12: a) Change in ultimate strength of SA along time for the different aging temperatures of 45, 60 and 80 ◦C.
b) Showing the change in ultimate stretch of SA along time for the different aging temperatures of 45, 60 and 80 ◦C.

b)

c)

14

100101102103Time[hrs]00.511.5245 °C60 °C80 °C45 °C60 °C80 °CRelative Ultimate Strength100101102103Time[hrs]00.511.52Relative Ultimate Stretch45 °C60 °C80 °C45 °C60 °C80 °C100101102103Time[hrs]0100300500Toughness[MJ/m3]SAa)
Figure 13: a) Change in ultimate strength of PUA along time for the different aging temperatures of 45, 60 and 80 ◦C.
b) Showing the change in ultimate stretch of PUA along time for the different aging temperatures of 45, 60 and 80 ◦C.

b)

c)

3.2. Separation of Mechanical & Environmental damage

In order to validate the two chain theory’s underlying framework, we need to prove that the change

in the mechanical constitutive behavior of the materials stems from the damage sustained from the

environmental damage, while the ratio could be sustained from mechanical damage is constant

given that the mechanical loading regime remains unchanged. Accordingly, to separate the effects

of mechanical and environmental damage in photo-oxidative aging on the performance of the

polymers, we introduced three dimensionless parameters to describe the effect of environmental
aging on mechanical performance obtained from cyclic and failure tests, namely σ∗, e∗ and W∗.
These parameters represent the ratios of mechanical performance for the same criteria along

the cyclic testing regime. They also eliminate the effect of the environmental damage, since they

are not absolute values, but rather ratios within the same test, and accordingly each parameter is

only compared to another which has sustained the same amount of environmental damage.

Relative softening in stress amplitude, σ*, is derived from cyclic tests @ constant amplitude

with the following equation:

∗
n =

σ

σn
σmax

(2)

where σmax is the maximum stress achieved in the first cycle at a given strain, and σn is the stress

achieved at the nth cycle at constant stretch amplitude. As expected from elastomeric systems
the largest decline is expected to occur in the first cycle, σ∗

n which gradually disappear in

1 > σ∗

15

100101102103Time[hrs]00.511.52Relative Ultimate Strength45 °C60 °C80 °C45 °C60 °C80 °C100101102103Time[hrs]00.511.52Relative ultimate Stretch45 °C60 °C80 °C45 °C60 °C80 °C100101102103Time[hrs]PUA050015002500Toughness[MJ/m3]45 °C60 °C80°C45 °C60 °C80 °CFigure 14: Different parameters being calculated from the cyclic test.

next cycles as can be seen in Figure 15. Hereafter, we will simply use σ∗ := σ∗
1 to represent
the softening that occurs in the material as a result of mechanically damage through the ratio of

retained strength to the original strength at a specific stretch amplitude, here the first stretch limit

in the cyclic test λ1.

The second parameter being calculated is the relative permanent set, e∗, which is calculated by

e∗
n =

en
λ1

(3)

where λ1 is the stretch amplitude of the first set of cyclic tests, and en is the permanent set of cycle
n. Hereafter, we will simply use e∗ := e∗
1, as the strains used to make these calculations are denoted
in Table 3 . This metric shows the evolution of permanent set from the first cycle as a percentage

of that cycle’s stretch amplitude and it indicates how much the material loses its overall elasticity

with time. The third and final metric is the relative hysteresis loss(W*), which is calculated by the

following equation,

(4)

W∗ =

Dn
W

16

Strain [%]0Stress [MPa]ee1maxσ σ 1maxWΔW Where W is the total work done in extending the material in one cycle ∆W, denoted in Eqn.4 as

Dn, is the net work stored by the polymer. This metric allows a further understanding of the loss of

elasticity induced by mechanical damage and how much the mechanical loss changes throughout

the photo-oxidative aging period.

3.3. Observations

3.3.1. SA

a)
Figure 15: Changes a) in σ* across aging conditions at λ1 in the cyclic test b)in e∗, and c) in W* during the first load
cycle against aging time at λ1 in the cyclic test.

b)

c)

a)

b)

Figure 16: a) Stress softening across different cycles for SA at 45, 60 and 80 ◦C, where λn is the stretch limit of cycle
n in the cyclic test b) Showing relative residual strain across different cycles for SA at 45, 60 and 80 ◦C, where λn is
the stretch limit of cycle n in the cyclic test.

As can be seen from Figure 15, all three parameters of stress softening, relative residual strain

and hysteresis loss in SA compound can be considered to be constant and independent of the

temperature and aging time. This means that they are independent from the effects of environmen-

tal aging generally and environmental damage, which includes losses in stiffness and toughness,

17

100101102103Time[hrs]00.511.52σ* 45 °C60 °C80 °C45 °C60 °C80 °CSA100101Time[hrs]-0.500.51e*45 °C60 °C80 °CSA45 °C60 °C80 °C100101102103Time[hrs]00.20.40.60.81W*45 °C60 °C80 °C45 °C60 °C80 °CTime [hrs]SAStress SofteningC λ1 C λ2 C λ1 C λ2 C λ3 C λ4 C λ3 C λ4 Time [hrs]Relative Residual StrainSAspecifically.

In order to ensure that the consistency in behavioral independence from environmental effects 

persists across the entirety of the mater tensile response, we removed the role of stretch amplitude 
by  computing  these  parameters  across  the  different  strain  peaks  present  in  the  cyclic  test  and 

we have shown that then results are consistent and similar at across all of them at high and low 
temperatures alike, as demonstrated in Figure 16.

3.3.2. PUA

a)

b)

c)

Figure 17: a) Change in values of σ* across aging conditions during the first cycle with time. b) Showing the change
in e∗. c) Showing the change of the W* during the first cycle with time.

a)

b)

Figure 18: a) Stress softening across different cycles for PUA at 45, 60 and 80 ◦C, where λn is the stretch limit of
cycle n in the cyclic test b) Showing relative residual strain across different cycles for PUA at 45, 60 and 80 ◦C, where
λn is the stretch limit of cycle n in the cyclic test.

Similar results were obtained for PUA, as it does not exhibit any significant stress softening

along both the aging period and time, as shown in Figure 17a). This performance however is not

maintained in relative residual strain and hysteresis loss, which indicates that the material’s elastic

properties are changing with time and temperature. As can be seen in Figure 17b), the loss of

18

100101102103Time[hrs]00.511.52σ* 45 °C60 °C80 °CPUA80 °C60 °C45°C100101Time[hrs]-0.500.51e*45 °C60 °C80 °C45 °C60 °C80 °C100101102103Time[hrs]00.20.40.60.81W*45 °C60 °C80 °C45 °C60 °C80 °CPUAPUATime [hrs]Stress SofteningTime [hrs]Relative Residual StrainPUAelasticity occurs gradually along the aging period and with increases in temperature similarly. No

specific trend can be deduced however from from Figure 17c), however we can infer that there is

a certain change that occurs for the amount of energy loss of each cycle with respect to time and

temperature.

Both compounds exhibit similar behavior in stress softening across temperatures and aging

duration for the different cycles, as shown in Figure 17 and 18. This indicates that both materi-

als maintain a similar micro-structural assortment of chains, which unwind in the same manner

regardless of the existing damage in the polymer matrix. Results also suggest that the there is no

change in the recoil mechanism responsible for the elastic properties of the materials, but rather

the main change that materials experience during aging is in the mechanism that is in action dur-

ing the loading process. Accordingly, the damage that does occur from environmental factors does

not affect the C-C backbone which gives the materials their elastomeric nature. Thus, the chain

scission that does occur along the aging process cannot be considered must be acting mainly on

chains which do not contribute to the elastic nature of the the materials. The results show that

the mechanical and environmental modes of damage can be considered separable in both

compounds.

The differences in behavior that occur in unloading indicates that the modes of damage are

affecting the components responsible for creating the elastic nature of the materials. The behaviour

is resulted from the competition between the detachment and reformation of c-c links, which is

mostly controlled by the oxygen attack, diffusion rate and consumption rate(Rubinstein and Colby,

2003). This means that the differences in values of both the relative residual strain and hysteresis

loss that occur along the aging durations and temperatures occur due to the different networks of

crosslinks that are broken down and created throughout the aging process and during mechanical

testing. Accordingly, since this damage mode is related to both environmental conditions and the

mechanical loading case, the separability of these modes of damage becomes significantly less for

the recoil/elastic behavior of the materials.

19

3.4. FTIR

Figure 19: FTIR scan of SA aged at 45◦C, 60◦C, 80◦C for different durations imposed over an unaged sample results.

3.4.1. Silicon polymer

Results across the aging temperatures and durations are fairly consistent, with little to no change

in the chemical composition of the SA, but rather a loss in transmittance, signifying a decrease

in molecular mobility relative to that of the unaged material. Another notable change occurring

consistently is that beyond the first period of exposure to UV radiation, little change is occurring

with respect to time in the surface layer of the material.

Table 4: SA FTIR functional groups detected

Peak wavelength cm-1 Functional group Movement type

1079 - 1059

1407

2909

2961

Si-O
B-N
C-H

(Aldehyde)

C-H

(Cyclic Alkanes)

stretching
Stretching

stretching

Stretching

20

Virgin45C 10 days 45C 60 days45C 150 days60C 1 day60C 10 days60C 30 days80C 1 day80C 10 days80C 30 days1000200030004000Wavelength [cm-1]050100150Transmittance [%T]400290914071079 - 10592961SAFigure 20: FTIR scan of PUA aged at 45◦C, 60◦C, 80◦C for different durations,imposed over an unaged sample scan

PUA

The spectral response of the polyurethane samples varies across the aging periods and tempera-

tures, however a consistent observation is the initial drop in transmittance followed by an increase
with respect to time. This phenomena is consistent across all temperatures, but falls short at 60◦C,
as the transmittance level for the sample aged for 30 days drops again.

Table 5: PUA FTIR functional groups detected

Peak wavelength cm-1 Functional group Movement type

853
1087

1140

1485

2316-2370

2855-2968

Si-O
B-N
C-H

(Aldehyde)

C-H

(Cyclic Alkanes)

C-H

(Cyclic Alkanes)

C-H

(Cyclic Alkanes)

stretching
Stretching

stretching

Stretching

Stretching

Stretching

From a general point of observation, both materials same to be undergoing the same phenom-

ena, which is curing, caused by the breakdown and/or creation of new crosslinks which form the

21

Virgin45C 10 days 45C 60 days60C 1 day60C 10 days60C 30 days80C 1 day80C 10 days80C 30 days050100Transmittance [%T]1000200030004000Wavelength [cm-1]2968 - 28551087114014858532370 - 2316PUAaged polymeric matrices. This phenomena however only translates to the mechanical behavior at

lower temperatures, as higher temperatures cause the materials to soften and weaken with time.

This deterioration in performance is not accompanied however with any significant change in func-

tional groups, which further supports the idea that the c-c backbone chains do not experience any

significant deterioration or scission from the environmental stresses they were subjected to. thus it

is logical to conclude that the change that occurs in the material is mostly due to physical bonding

between the chains such as hydrogen bonds (He et al., 2004), as they would not appear with dis-

tinctive peaks on the IR scan(Pretsch et al., 2000). Given that the backbones of the polymers are

indeed intact and no complications are occurring in the chemical structure of the materials, we can

safely assume that even though the damage caused by the environmental and mechanical factors

would operate on the same fronts, which is the current state of crosslinks developed in the material

with little effect on the main chains. This means that with our experiments we can can isolate the

compounding factors from each of these modes of damage in the constitutive behavior and use it

to characterize the constitutive behavior of the materials.

3.5. DSC test

The DSC test was conducted on aged sampled of both materials providing us with an insight on the

change in their heat properties throughout the aging process. Figure 22 and 24 show the variation

in Heat capacity of both materials across different temperatures.

The behavior of SA ad PUA in the DSC test show initial variation in the heat capacity either

increase or decrease. The extent of this variation however drops along the aging period and ap-

proaches that of the unaged material. This shows that regardless of whether either material shows

a change in it’s microstructural configuration causing it to be able to absorb more or less heat, the

final configuration achieved cannot vary significantly from the main configuration seen in the pri-

mary chain and cross-link assortment which gives each material its basic physical characteristics.

considering that the heat flow varies directly with heat capacity according to the following

22

Figure 21: Heat flow plot generated from DSC scan of SA aged at 45◦C, 60◦C, 80◦C for different durations imposed
over an unaged sample results.

equation:

dH
dt

= C p

dT
dt

+ f(T, t)

(5)

Where dH/dt is the heat flow, Cp is the heat capacity and f (T,t) is a function of Heat Flow

due to Kinetic Processes. Considering that the heat flow and capacity mirror each other in both

materials, we are able to deduce that the heat flow due to kinetic processes changes according to

time and temperature in a similar manner for both materials.

The change of the Endothermic Melting Temperature of SA along the aging period is presented

in Figure 25 a) and shows that the temperature fluctuates marginally along the aging process. This

indicates that the physical bods that maintain the solid state of the polymer become marginally

stronger. This is supported by the results of the tensile tests, which show that there is an increase

in stiffness for the material along the aging durations.

The Glassification temperature of PUA along the aging duration is presented in Figure 25

b) and shows that the temperature generally shows an initial rise, however this is followed by

a decline for longer aging durations.

It is apparent from the rest of the data presented in this

23

SAUnaged45C 10 days 45C 60 days45C 150 days60C 1 day60C 10 days60C 30 days80C 1 day80C 10 days80C 30 daysFigure 22: Heat capacity plot generated from DSC scan of SA aged at 45◦C, 60◦C, 80◦C for different durations
imposed over an unaged sample results.

study that the micro-structural evolution at the 30 days of aging is different at 60◦C than the
other two temperatures of the study. The trend of decline in micro-structural integrity due to the

environmental damage sustained from aging is apparent in the presented data, as any enhancement

in micro-structural integrity during the over-curing phase is overridden by the damage accumulated

from aging in the long run.

3.6. Swelling test

The swelling test is performed to describe the changes in the cross-link density throughout the

course of aging by measuring the amount of swelling of each compound in its Hexane, Toluene

and Pentane solvents using the following equation:

(cid:32)
1 − 2Mx
M

(cid:33)

,

(6)

ne
V0

=

ρ
Mx

where ne is the molar cross-linking, V0 is the starting volume of the elastomer, ρ is the density of the

solvent, Mx is the molar volume of the solvent and M is the molar volume of the elastomer(Flory,

24

SAUnaged45C 10 days 45C 60 days45C 150 days60C 1 day60C 10 days60C 30 days80C 1 day80C 10 days80C 30 daysFigure 23: Heat flow plot generated from DSC scan of PUA aged at 45◦C, 60◦C, 80◦C for different durations imposed
over an unaged sample results.

1953, Hiemenz and Lodge, 2007, Rubinstein and Colby, 2003). The response of the materials

varies with respect to different solvents; SA showed a clear swelling response to all three solvents,

with Hexane being the most clear and consistent response across the samples. PUA however, did

not respond well to both Hexane and Pentane, but showed marginal swelling with Toluene, which

is represented in Figure 26

Samples exhibit different behaviors in the test, where the cross-link density of SA marginally

increases and reaches a stabilized plateau after a certain time. Such a plateau in cross-linking can-

not be correlated directly with either the stiffening or softening behavior of the material across the

aging temperatures, given that the level of cross-linking does not rise and fall with the mechanical

behavior. This along with considering that cross-linking is primarily a product of the environ-

mental stress, we can deduce that the mechanical behavior is dependent on other factors than the

products of the environmental damage that occurs in the material.

The behavior of PUA marginally varies across both aging period and time, however at 45oC

the behavior is that of an initial increase in cross-linking until a certain peak level, beyond which

25

Unaged45C 10days 45C 60 days45C 150 days60C 1 day60C 10 days60C 30 days80C 1 day80C 10 days80C 30 daysPUBFigure 24: Heat capacity plot generated from DSC scan of PUA aged at 45◦C, 60◦C, 80◦C for different durations
imposed over an unaged sample results.

a steady decrease is observed. Thermo-oxidation does not seam to affect the development of

cross-links in SA and accordingly it can be assumed that the role of thermo-photo-oxidation on

PUA is comparatively greater from a micro-structural standpoint than on SA. On the other hand,

thermo-photo-oxidation has varying effects on PUA, however across all aging temperatures and

with longer aging durations, cross-link density drops from the peak values that are reached in

the over-curing period of aging. This being the case, it is more challenging to separate the exact

environmental effects of both photo-oxidation and Thermo-photo-oxidation in the case of the PUA.

The results shown in Figure 26 also indicate that there isn’t a proper correlation between the

constitutive behavior of the materials and the density of crosslinks, as both of the results do not

follow each other in their increasing and decreasing values. This leads to the conclusion that even

though the crosslinks have a role to play in the constitutive behavior of the materials, it is indeed

their type, how they are bonded to the polymer chains, the length and amount of chains and the

micro-structure which they create that collaboratively have the main influence on the mechanical

behavior. Thus the contribution of each factor must be considered from a qualitative perspective

26

Unaged45C 10 days 45C 60 days45C 150 days60C 1 day60C 10 days60C 30 days80C 1 day80C 10 days80C 30 daysPUBa)
Figure 25: Change in a) Endothermic melting temperature for SA and b) glassification temperature for PUA at 45◦C,
60◦C, 80◦C for different aging durations

b)

rather than a quantitative one.

a)

b)

Figure 26: Cross-link density of a) SA and b) PUB elastomers at 45◦C, 60◦C, 80◦C for different durations

27

100101102103Time [hrs]-43-42.5-42-41.5-41-40.5-40Temp [ C]45  Co60  Co85  CoSA45  Co60  Co85  Coo100101102103-63-62-61-60-59-58-57-56Temp [ C]45  Co60  Co85  Co45  Co60  Co85  CoPUATime [hrs]o100101102103Time[hrs]-0.1-0.0500.050.1Crosslink Density[mol/cm3]45C60 C80 CSA100101102103Time[hrs]-0.500.5Crosslink Density[mol/cm3]PUA45 C60 C80 C4. MODELLING

For modelling the behavior of the material a simple Neo-Hookean model based on Gaussian theory

was used to represent the response of the material to tensile(James and Guth, 1943)(Rezig et al.,

2020). The formulation of equation.8 is centered around the cross-link density of the material as

the defining parameter which accounts for the effects of aging which include time, temperature

and UV radiation.

Mc =

1
2Vc

σ = ( ρRT
Mc

)((P(1 − 2
f

)(λ − 1
λ

))n)

(7)

(8)

where Mc is the mass density of the rubber, Vc the cross-link density,σ the stress response of

the material measured in MPa, R being the gas constant, T being the aging temperature,P is a

multiplicative factorial, f the functionality of the crosslinks for each specific material, n the stress

factorial and lambda being the strain the material is subjected to. The factorials were adjusted to

fit the unaged behavior of both materials and were maintained at those values for all other plots.

Accordingly the projected compared to the experimental values of the stress were plotted for

strains up 300% as represented in Figure 27 for SA and in Figure 28 for PUA.

a)

b)

c)

Figure 27: Results of the a) Projected(P) results of the failure test vs the Experimental(E) results for the SA material
at 45◦C, b) 60◦C, and c) 80◦C for different aging periods.

Considering that the Neo-hookean model is dependant on the variation in cross-link density as

28

050100150200Strain[%]00.511.522.5Stress[MPa]Unaged-P10 Days-P60 Days-P150 days-PUnaged-E10 Days-E60 Days-E150 days-E050100150200Strain[%]00.511.522.5Stress[MPa]Unaged-P1 Day-P10 Days-P30 days-PUnaged-E1 Day-E10 Days-E30 days-E050100150200Strain[%]00.511.522.5Stress[MPa]Unaged-P1 Day-P10 Days-P30 days-PUnaged-E1 Day-E10 Days-E30 days-Ea)

b)

c)

Figure 28: Results of the a) Projected(P) results of the failure test vs the Experimental(E) results for the PUA material
at 45◦C, b) 60◦C, and c) 80◦C for different aging periods.

the parameter for mechanical behavior representation, it is of little surprise that the experimental

results do not match the projected results of the model. The stress-strain values generated experi-

mentally show a drop in stiffness along the aging process, which is not mirrored by the values of

the cross-link density and accordingly not mirrored by the projected plot. This further illustrates

the separation between the mechanical and environmental damage modes, given that the mechan-

ical mode is clearly not reliant on the most prominent parameter of the environmental damage, i.e

the cross-link density.

29

050100150200Strain[%]0246Stress[MPa]Unaged-P10 Days-P60 Days-P150 days-PUnaged-E10 Days-E60 Days-E150 days-E050100150200Strain[%]0246Stress[MPa]Unaged-P1 Day-P10 Days-P30 days-PUnaged-E1 Day-E10 Days-E30 days-E050100150200Strain[%]0246Stress[MPa]Unaged-P1 Day-P10 Days-P30 days-PUnaged-E1 Day-E10 Days-E30 days-E5. SIGNIFICANCE IN MODELLING

The collected data allowed the study of a relationship between energy dissipation, cyclic stress

and permanent set for both constant- and incremental-amplitude regimes. Results showed clear

possibility of decomposition of mechanical Dmech and environmental damages Denv in most cases

and with respect to most factors considered for characterization; the increase in accumulated en-

ergy happened at later stages of cyclic loading, supporting our earlier findings. The presented

graphs can be extremely useful for modeling of aging in soft materials, by serving as a reference

to describe the error in the multiplicative decomposition of mechanical and environmental dam-

ages which plays a critical factor in simplifying the modeling approach. According to the previous

results, one can safely assume that the constitutive behavior of the samples can be modeled using

classical representation of accumulative damage factor D where

ΨN = (1 − D) Ψ0,

D ≈ DenvDmech

(9)

where ΨN represents the energy of the aged matrix, and Ψ0 represents the energy of virgin

material.

This can serve as a functional alternate route to the shape-function reliant micro-mechanical

approach presented in (Mohammadi et al., 2020), which relies on defining the aged material’s

behavior by interpolating between the behavior of the unaged network and that of a fully damaged

network as presented in the following equation.

ΨN = S 0Ψ0 + S ∞Ψ∞,

(10)

where ΨN represents the energy of the aged matrix, Ψ0 and S0 represent the shape function and

the energy of virgin material. Ψ∞ and S∞ represent the shape function and the energy of fully aged

material.

The damage factor approach can additionally be utilized to complement the shape-function and

30

further define the behavior of the fully damaged network, increasing the accuracy and reliability

of the final prediction.

31

6. CONCLUSION

Considering that both materials used for this study share similar behavior in their cured unaged

condition, it is not unexpected that their behavior with aging on both the mechanical and chem-

ical fronts do not differ significantly. It has been shown in this study that the modes of damage

that the materials exhibit can indeed be separated into environmental and mechanical components,

both of which operate through the polymer chains and their cross-links which holds and fold the

chains together. With the environmental damage occurring mainly in the crosslinked network in

comparison to the mechanical network. Considering that both modes of damage are separable we

can deduce that the chains and links within the matrix of the material would have different config-

urations, which cause them to be affected differently according to the stresses they are subjected

to. We are also able to further breakdown the environmental mode of damage into two distinct

components, that caused solely by photo-oxidation and that caused by the combination of photo

and thermo-oxidation. Upon isolating the exact effects of mechanical and environmental damages

on the individual networks they act on, we would be able to combine these factors into a micro-

mechanical model to properly predict the constitutive behavior of elastomeric materials subjected

to the combined weathering conditions found in applications subjected to solar radiation.

32

REFERENCES

33

REFERENCES

S.  Al-Azhary,  H.  Mohammadi,  and  R.  Dargazany.  Thermo-oxidation  analysis  of  structural  
adhesives:  An  experimen-  tal  study.  In  ASME  2020  International  Mechanical  Engineering 
Congress and Exposition. American Society of Mechanical Engineers, 2020.

the  non-destructive 
A.  B.  S.  Andrew  R.  George.  A  new  spectroscopic  method  for 
characterization  of  weathering  damage  in  plastics.  (English)  [On  weathering  of  polymers]. 
Journal of Advanced Materials, 40(2):41–56, 2008.

L. Audouin, S. Girois, L. Achimsky, and J. Verdu. Effect of temperature on the photooxidation of 
polypropylene films. Polymer degradation and stability, 60(1):137–143, 1998.
A.  Bahrololoumi,  V.  Morovati,  and  R.  Dargazany.  Hydrolytic  aging  in  rubber-likes  materials: 
Micro-mechanical  approach  to  modeling.  International  Journal  of  Plasticity,  1,  2019  (in 
submission).

A.  Bahrololoumi,  V.  Morovati,  E.  A.  Poshtan,  and  R.  Dargazany.  A  multi-physics  constitutive 
model  to  predict  quasi-static  behaviour:  Hydrolytic  aging  in  thin  cross-linked  polymers. 
International Journal of Plasticity, page 102676, 2020.
Z.  Bettina,  N.  Christophe,  S.  Christian,  and  P.  Wulff.  Chemistry,  polymer  dynamics  and 
mechanical properties of a two-part polyurethane elastomer during and after crosslinking. Part I: 
dry conditions. (English) [On crosslinking of polymers]. Polymer, 115(1):77–95, 2017. doi: https://
doi.org/10.1016/j.polymer.2017.03.020.
S. Bhatia. Advanced renewable energy systems,(Part 1 and 2).
D.  Bigg.  Mechanical  properties  of  particulate  filled  polymers.  Polymer  Composites,  8(2):115–
122, 1987.
M.  Biron.  Material  Selection  for  Thermoplastic  Parts:  Practical  and  Advanced  Information. 
William Andrew, 2015.

A.  Blaustein  and  C.  Searle.  Encyclopedia  of  Biodiversity  (Second  Edition),  Ultraviolet 
Radiation.  Academic Press, 2013.
D. Bower. An Introduction to Polymer Physics. Cambridge, 2002.

M. C. Celina.  Review of polymer oxidation and its relationship with materials performance and 
lifetime prediction. Polymer Degradation and Stability, 98(12):2419–2429, 2013.
R. Chandra and R. Rustgi. Biodegradable polymers. 23:1273–1335.

J.  Cruz-Pinto,  M.  Carvalho,  and  J.  Ferreira.  The  kinetics  and  mechanism  of  polyethylene 
photo-oxidation.  Die Angewandte Makromolekulare Chemie: Applied Macromolecular Chemistry 
and Physics, 216(1):113–133, 1994.
A.  Dupuis,  F.-X.  Perrin,  A.  U.  Torres,  J.-P.  Habas,  L.  Belec,  and  J.-F.  Chailan.  Photo-
oxidative degradation  behavior  of  linseed  oil  based  epoxy  resin.  Polymer  Degradation  and 
Stability, 135:73–84, 2017.

34

D. Feldman. Polymer Weathering: Photo-Oxidation. (English) [On photo-oxidation of polymers]. 
Journal  of  Polymers  and  the  Environment,  10(1):163–173,  2002.  doi:  https://doi.org/10.1023/
A:1021148205366.
P. Flory. Principles of polymer chemistry. Cornell University Press, 1953.

G. Geuskens and C. David. The photo-oxidation of polymers. a comparison with low molecular 
weight compounds. In Photochemistry–7, pages 233–240. Elsevier.
K.  T.  Gillen,  M.  Celina,  R.  L.  Clough,  and  J.  Wise.  Extrapolation  of  accelerated  aging  data-
arrhenius or erroneous? Trends in polymer science, 8(5):250–257, 1997.

K. T. Gillen, M. Celina, and M. R. Keenan. Methods for predicting more confident lifetimes of 
seals in air environ-ments. Rubber chemistry and technology, 73(2):265–283, 2000.
Y. He, B. Zhu, and Y. Inoue. Hydrogen bonds in polymer blends. Progress in Polymer Science, 
29(10):1021–1051, 2004.
P. C. Hiemenz and T. R. Lodge. Polymer Chemistry, Second Edition. CRC Press, 2007.
H.  James  and  E.  Guth.  Theory  of  the  elastic  properties  of  rubbers.  Journal  of  Chemical 
Physics, 11:455, 1943.
F.  Khabbaz,  A.-C.  Albertsson,  and  S.  Karlsson.  Chemical  and  morphological  changes  of 
environmentally  degradable  polyethylene 
thermo-oxidation.  Polymer 
Degradation and Stability, 63(1):127–138, 1999.

films  exposed 

to 

Q.  Liu,  H.  Yang,  J.  Zhao,  S.  Liu,  L.  Xia,  P.  Hu,  Y.  Lv,  Y.  Huang,  M.  Kong,  and  G.  Li. 
Acceleratory  and  inhibitory  effects  of  uniaxial  tensile  stress  on  the  photo-oxidation  of 
polyethylene: Dependence of stress, time duration and temperature. Polymer, 148:316–329, 2018.

L. W. McKeen. The effe ct of sterilization on plastics and elastomers. William Andrew, 2018.
H. Mohammadi and R. Dargazany. Thermo-oxidative aging. Tire Technology Review, pages 54–
57, 2018.
H. Mohammadi and R. Dargazany. A micro-mechanical approach to model thermo-oxidative 
aging in elastomers. International Journal of Plasticity, 118:1–16, 2019.
H. Mohammadi, V. Morovati, E. Poshtan, and R. Dargazany. Understanding decay functions and 
their  contribution  in  modeling  of  thermal-induced  aging  of  cross-linked  polymers.  Polymer 
Degradation and Stability, page 109108, 2020.

M.  Nachtane,  M.  Tarfaoui,  S.  Sassi,  A.  El  Moumen,  and  D.  Saifaoui.  An  investigation  of 
hygrothermal aging effe cts on high strain rate behaviour of adhesively bonded composite joints. 
Composites Part B: Engineering, 172:111–120, 2019.
H. Namazi. Polymers in our daily life. (English) [On applications of polymeric materials ]. 
Bioimpacts, 7(2):73–74, 2017. doi: https://bi.tbzmed.ac.ir/FullHtml/bi-17438.

35

J.  E.  Pickett.  An  unusual  photo-oxidation  pathway  for  a  poly  (2,  6-dimethyl-1,  4-phenylene  
oxide) model compound part a polymer chemistry. 2017.   
E. Pretsch, P. B¨uhlmann, C. Affo lter, E. Pretsch, P. Bhuhlmann, and C. Affo lter. Structure 
determination of organic compounds. Springer, 2000.

A.  Quintana  and  M.  C.  Celina.  Overview  of  dlo  modeling  and  approaches  to  predict 
heterogeneous oxidative polymer degradation. Polymer Degradation and Stability, 149:173–191, 
2018.
N. Rezig, T. Bellahcene, A. Aberkane, and A. N. Abdelaziz. Thermo-oxidative ageing Af a Abr 
rubber: effects An mechanical and chemical properties. Journal of Polymer Research, 27(11):1–13, 
2020.

A.Rodriguez, B. Mansoor, G. Ayoub, X. Colin, and A. Benzerga. Effect of uv-aging on the 
mechanical and fracture behavior of low density polyethylene. Polymer Degradation and Stability, 
page 109185, 2020.
M.Rubinstein and R. H. Colby. Polymer physics. OUP Oxford, 2003.
M. H. Sadd. Continuum Mechanics Modeling of Material Behavior. Academic Press, 2018.

A. W. Signor, M. R. VanLandingham, and J. W. Chin. Effe cts of ultraviolet radiation exposure on 
vinyl  ester  resins:  characterization  of  chemical,  physical  and  mechanical  damage.  Polymer 
degradation and stability, 79(2):359–368, 2003.

A. Standard. D412-06 standard test methods for vulcanized rubber and thermoplastic elastomers–
tension. ASTM International, West Conshohoken, PA, USA, 2006.

A.  Standard.  G151-19  standard  practice  for  exposing  nonmetallic  materials  in  accelerated  test 
devices  that  use  labo-ratory  light  sources.  ASTM  International,  West  Conshohoken,  PA,  USA, 
2019.

M. Tolinski. Additives for polyolefins: getting the most out of polypropylene, polyethylene and 
TPO. William Andrew, 2015.
H.  Wu,  Y.  Zhao,  X.  Dong,  L.  Su,  K.  Wang,  and  D.  Wang.  Probing  into  the  microstructural 
evolution  of  isotactic  polypropylene  during  photo-oxidation  degradation.  Polymer  Degradation 
and Stability, page 109434, 2020.

M. Zaghdoudi, A. K¨ommling, M. Jaunich, and D. Wolff. Erroneous or arrhenius: A degradation 
rate-based model for epdm during homogeneous ageing. Polymers, 12(9):2152, 2020.

36