materials

Article
Lifetime Prediction of Polymers: To Bet, or Not to
Bet—Is This the Question?
Ignazio Blanco 1,2 ID

1 Department of Civil Engineering and Architecture, University of Catania, Viale Andrea Doria 6,
95125 Catania, Italy; iblanco@unict.it
2 UdR-Catania Consorzio INSTM, Viale Andrea Doria 6, 95125 Catania, Italy




Received: 18 July 2018; Accepted: 7 August 2018; Published: 8 August 2018


Abstract: Polymers are a great and very important category of organic compounds that have changed
our lifestyle. In the last eighty years, we have used them for the most varied applications, and from
the first structural ones we began to investigate their durability, which can be fatal in the successful
completion of the application for which the material was designed. Over the last thirty years,
the environmental problems related to the disposal of polymers that have completed their lifecycle
have begun to arise, and the need to foresee their end of life has become increasingly urgent. In this
manuscript, the reliability of the lifetime predictions of polymeric materials is faced with comparing
measurements obtained at low temperature with those carried out at high temperatures, in the molten
state. The obtained data were treated by a well-established kinetics model and discrepancies were
observed in the two different conditions (high and low temperatures), which led to a mismatching
between expected and real data. A correction of the data extrapolated from measurements obtained
at high temperatures, by using a novel equation which takes into account the induction period
(IP) of the degradation process, is proposed. Considerations about the useful parameters, namely
initial decomposition temperature (Ti ), activation energy of degradation (Ea ), and glass-transition
temperature (Tg ), to be used for making predictions, are also carried out.

Keywords: lifetime prediction; thermal stability; polymer degradation; thermogravimetric analysis;

induction period

1. Introduction
To realize the great interest on this topic, just put the keyword “lifetime prediction” in the Scopus
search engine. Since data are available, we can observe an exponential growth of scientific works
concerning the lifetime prediction (Figure 1). It is changing from an annual average of fifty manuscripts
in the seventy-eighty, at the present day, to an annual average of eight hundred manuscripts, with an
increase of 1500 percent. Although it is right to remark that in the same period there has been an
overall increase in scientific publications, if we repeat the same operation by entering two other very
common search terms, such as “glass transition” or “calorimetry”, a remarkable growth can obviously
be observed, but that does not reach even half from publications relating to “lifetime prediction”.
Some years ago, the International Confederation for Thermal Analysis and Calorimetry (ICTAC),
realizing this interest, has instituted a special scientific committee in order to set the use of Thermal
Analysis (TA) methods for making lifetime predictions by working closely with Kinetics Committee.
Furthermore, in 2016, for the first time, during the ICTAC Conference, which was held in Orlando,
Florida, a special session on lifetime prediction has been inserted in the congress program that has had
wide and geographically variegated participation [1].

Materials 2018, 11, 1383; doi:10.3390/ma11081383 www.mdpi.com/journal/materials

Materials 2018, 11, 1383 2 of 13
Materials 2018, 11, x FOR PEER REVIEW 2 of 13

Figure 1. Time evolution of manuscripts published on International Journals having "lifetime

Figure 1. Time evolution of manuscripts published on International Journals having "lifetime
prediction" as topic. Source: Scopus (https://www.scopus.com/).
prediction" as topic. Source: Scopus (https://www.scopus.com/).
Many aspects of the lifetime prediction were treated and some interesting conclusions were
Many
drawn. aspects of sixty
First of all, the lifetime
percent of prediction were treated
the presentations coveredand some materials,
polymeric interesting conclusions
highlighting that were
drawn. First of all, sixty
thermo-oxidative aging percent
of polymersof the is presentations
one of the most covered
seriouspolymeric
problems felt materials, highlighting
by the scientific
that thermo-oxidative
community. One of the aging of polymers
addressed issues wasis one
the of
rolethe most and
of aging serious problems felt
the deteriorating by the scientific
of physical and
mechanical
community. Oneproperties, for instance
of the addressed of rubber
issues was the during
role ofusage.
agingIt and
is a very serious problem.
the deteriorating Do you and
of physical
remember
mechanical Challenger?
properties, forThe space of
instance shuttle
rubber exploded
duringshortly
usage.afterIt islaunch
a veryonserious
28 January, 1986 just
problem. Do you
because of the failure of an O-ring. Although in that case the accident
remember Challenger? The space shuttle exploded shortly after launch on 28 January 1986 was attributed to the use of just
rubber with improper glass-transition temperature (Tg), it proves that failure of a polymeric material
because of the failure of an O-ring. Although in that case the accident was attributed to the use
may cause great loss, or even disaster. Thus, accelerated aging in an air-circulating oven at elevated
of rubber with improper glass-transition temperature (Tg ), it proves that failure of a polymeric
temperatures for months and monitoring changes of mechanical properties until failure for
material may cause
extrapolating great loss,
to service or eventodisaster.
temperature Thus,prediction
obtain lifetime accelerated is ofaging in an air-circulating
great importance for aerospace,oven at
elevated
automotive, and petroleum industry applications [2,3]. Understanding the aging process when for
temperatures for months and monitoring changes of mechanical properties until failure
extrapolating
predictingtotheservice
lifetimetemperature
of elastomertocomponents
obtain lifetime predictionthe
and foreseeing is timespan
of great importance
left until they for aerospace,
have to
automotive,
be replacedand petroleum
is essential; industry
generally, applications
we can [2,3]. process
divide the aging Understanding
into physical theandaging process
chemical aging.when
Understanding
predicting the lifetimethe main characteristics
of elastomer of the two aging
components typologies, the
and foreseeing the question
timespan discussed
left untilwasthey
how have
the aging of elastomers can be experimentally investigated. Direct
to be replaced is essential; generally, we can divide the aging process into physical and chemical or indirect method? On
considering that oxidation is the dominant factor in a lot of polymer
aging. Understanding the main characteristics of the two aging typologies, the question discussedapplications, a direct method
like thermogravimetric analysis (TGA), differential scanning calorimetry (DSC), or dynamic
was how the aging of elastomers can be experimentally investigated. Direct or indirect method?
mechanical analysis (DMA) are those most indicated [4–8].
On considering that oxidation is the dominant factor in a lot of polymer applications, a direct method
The other investigated aspect was that of kinetic modeling of the degradation, which has lifetime
like thermogravimetric
prediction as its finalanalysis
goal. The (TGA), differential
different modelingscanning
approaches calorimetry (DSC), or dynamic
of the thermoanalytical mechanical
data obtained
analysis
with(DMA)
the mainare techniques,
those most such indicated
as TGA,[4–8].DSC, and differential thermal analysis (DTA), were
The otherfrom
discussed: investigated aspect
isoconversional (thewas that rate
reaction of kinetic modeling
at constant extent ofof the degradation,
conversion which has
is only a function
lifetime prediction or
of temperature) asmodel-free
its final goal.
methods,Thewhichdifferent modeling
are generally splitapproaches
in two categories:of thedifferential
thermoanalytical
and
integral. The most common differential isoconversional method is that
data obtained with the main techniques, such as TGA, DSC, and differential thermal analysis (DTA), of Friedman [9,10], whilst
wereexamples
discussed: of integral methods are the Ozawa–Flynn–Wall
from isoconversional (the reaction rate [11,12] and Vyazovkin
at constant extent of[13] equations.isOn
conversion only a
the other hand, we have the model-fitting methods (the derivation
function of temperature) or model-free methods, which are generally split in two categories: of kinetic parameters associated
differential
with a particular reaction model that is assumed to represent the conversion dependence of the
and integral. The most common differential isoconversional method is that of Friedman [9,10],
reaction rate). Two of the most used are the Sestak–Berggren [14,15] and the Avrami–Erofeev [16,17]
whilst examples of integral methods are the Ozawa–Flynn–Wall [11,12] and Vyazovkin [13] equations.
equations. In the middle ground is the Kissinger method [18], which yields a reliable estimate of
On the other hand,
activation energy weonly
havewhenthe conversion
model-fitting doesmethods (the derivation
not practically vary with of kinetic
heating parameters
rate. associated
For this reason,
with ait particular reaction model that is assumed
should not generally be called “isoconversional” [19,20].to represent the conversion dependence of the reaction
rate). Two of the most used are the Sestak–Berggren [14,15] and the Avrami–Erofeev [16,17] equations.
In the middle ground is the Kissinger method [18], which yields a reliable estimate of activation
Materials 2018, 11, 1383 3 of 13

energy only when conversion does not practically vary with heating rate. For this reason, it should not
generally be called “isoconversional” [19,20].
Examining the literature of the last thirty–forty years, one comes across an unbridled use of
the initial decomposition temperature (Ti ) and the activation energy of degradation (Ea ) to make
considerations on the materials’ lifetime; at a more careful reading, a question arises spontaneously:
which parameter must be considered to perform correct and reliable lifetime prediction of materials?
The most used parameter to determine the thermal stability of a material (and thus to perform a
prediction on its lifetime) is certainly the Ti [21–23]. However, when we compare compounds with a Ti
sufficiently close with each other, Ea must be considered [24–26], thus giving more importance to the
determination of the kinetic parameters of degradation. Since experiments to obtain kinetic parameters
of degradation are usually performed at temperatures near (or, in some cases, above) melting
temperatures, the values obtained might not be suitable to represent the behavior of polymers in service,
so the lifetime predictions based on these measurements could be questionable. Another question is
added to the previous about the use of the most suitable parameter to make lifetime predictions: could
one make a prediction relying exclusively on data obtained at high temperatures?
Furthermore, thermal degradation in polymeric materials occurs when the energy of vibration
exceeds the primary bonding between atoms (thus, thermal stability is generally ascribed to the
vibration energy of the chemical bonds), whilst the glass-transition phenomenon is related to the
rotation and vibration of molecular chains [27]. According to these considerations and from the
perspective of investigating a possible correlation between Tg and degradation, the temperature of
isothermal experiments was gradually lowered to be close to the glass transition one.
Finally, but not less importantly, the aspects related to the disposal of plastics. The polymer-
packaging industry is actually divided between the need to extend and implement principal packaging
function containment, protection, and preservation, and the need to be easily disposed of or recycled
after use. Thus, for these specific polymers used in food packaging, the study of their thermal properties
is also crucial to improve recyclability or provide a viable alternative. In particular, lifetime predictions
play a key role in facilitating both the design phase and the final disposition. But these predictions,
up to today, have been based on the identification of the critical reaction that limits the life of a material;
therefore, there was a strong need for accelerated lifetime characterization methodologies.
Before answering these questions, we have a certainty: the thermogravimetric (TG) technique
is widely employed to determine kinetic parameters of polymer degradation. It has already been
highlighted, as experiments carried out above the melting temperature of polymers, generally in
dynamic condition, are usually performed to determine kinetic parameters associated with polymer
degradation. Another problem is to give the correct physical meaning to the kinetics parameters used
because the kinetics methods used to evaluate the TG data, obtained in the most different experimental
conditions, are sometimes based on assumptions not always satisfied; thus, great care must be devoted
before making actions towards polymer stabilization based on these results.
With the aim to answer all of these questions, in this manuscript I have tried to correlate the
experimental data obtained in the last 15 years by performing long- and short-term degradation
measurements on several polymeric materials. The novelty of this procedure is represented by
performing degradation experiments at temperatures lower than those usually employed, which were
largely lower than the melting ones and, for a good part of the analyzed polymers, close to their Tg .
Degradation data obtained at these relatively low temperatures were compared to those obtained
at higher temperatures in order to verify the reliability of lifetime predictions of polymers made by
degradation experiments performed at higher temperatures (due to the time problem of the measure).
Particular focus has been devoted to better investigating the role of the induction period (IP) in the
degradation of complex systems, like polymers, and its effect on the prediction.
Materials 2018, 11, 1383 4 of 13

2. Experimental
Materials
Materials 2018, 11, x FOR PEER REVIEW 44 of 13
Materials 2018,
Materials 2018, 11,
11, xx FOR
FOR PEER
PEER REVIEW
REVIEW 44 of
of 13
13
Materials 2018,
2.1. Materials 2018, 11,
11, xx FOR
FOR PEER
PEER REVIEW
REVIEW 4 of
of 13
13
2.
2. Experimental
Experimental
2.
2. Experimental
2. Experimental
All used polymers are reported in Table 1, together with the code used for their identification.
Experimental
2.1.
Polyetherketones
2.1. Materials
(PEK, samples 1–3) were synthesized according to the procedure previously
Materials
2.1.
2.1. Materials
2.1. Materials
Materials
reported [28] and All
All
All
were
used
used used in are
polymers
used polymers
polymers are
powder
are
reported
reported
reported
form.
in
in
in
High-density
Table
Table
Table 1,
1, together
1, together
together with
polyethylene
with the code
with the
the code used(PE,
code used
used for
for
sample
for their
their
their
4), polystyrene
identification.
identification.
identification.
All
All used
used polymers
Polyetherketones polymers
(PEK, are
are reported
reported
samples 1–3)in
in Table 1,
1, together
together with
Table synthesized
were with the
the code
accordingcodetoused
used
the for
for their
their identification.
procedure identification.
previously
(PS, samplePolyetherketones
5), polycarbonate
Polyetherketones
Polyetherketones
(PEK,
(PEK,
(PEK,
(PC, sample
samples
samples
samples
1–3)
1–3)
1–3)
and synthesized
6),were
were
were
poly(methyl
synthesized
synthesized
methacrylate)
according
according
according
to
to the
the (PMMA,previously
procedure
procedure sample 7), were
previously
Polyetherketones
reported [28] and (PEK,used
were samples
in 1–3) form.
powder were High-density according to
synthesized polyethylene the
to (PE, procedure
the sample
procedure 4), previously
previously
polystyrene
standard forreported
Gel
reported [28]
[28] and
Permeation
and were
were used
used in
Chromatography
in powder
powder form.
form. High-density
(GPC)
High-density polyethylene
products purchased
polyethylene (PE,
(PE, sample
from
sample 4), polystyrene
Fluka
4), (Milan, Italy)
polystyrene
reported
reported
(PS, sample [28] and
[28]5),
and were
were used
used in
polycarbonate powder
in (PC,
powder sampleform.
form.6),High-density
High-density
and polyethylene
polyethylene
poly(methyl (PE,
(PE, sample
methacrylate) sample
(PMMA, 4),
4), polystyrene
polystyrene
sample 7),
(PS,
(PS, sample
sample 5),
5), polycarbonate
polycarbonate (PC,
(PC, sample
sample 6),
6), and
and poly(methyl
poly(methyl methacrylate)
methacrylate) (PMMA,
(PMMA, sample
sample 7),
7),
and used without
(PS, sample
(PS, sample
were any5),further
standard 5), purification.
polycarbonate
polycarbonate
for Gel Permeation (PC, samplePE,6),
(PC,Chromatography
sample PS,
6), and
and and PMMA,
poly(methyl
poly(methyl
(GPC) which
methacrylate)
methacrylate)
products were(PMMA,
purchased in powder
(PMMA,
from Flukasample
sample form,
(Milan,7), were
7),
were
were standard
standard for
for Gel
Gel Permeation
Permeation Chromatography
Chromatography (GPC)
(GPC) products
products purchased
purchased from
from Fluka
Fluka (Milan,
(Milan,
were
used as received, standard
while
were standard
Italy) and used for
PC,
for Gel
Gel
without Permeation
which was
Permeation
any further inChromatography
pellet form,
Chromatography
purification. PE, (GPC)
was
(GPC)
PS, andproducts
ground
products
PMMA, in purchased
a mortar
purchased
which were from
before
from
in Fluka
Fluka
powderuse.(Milan,
Polylactide
(Milan,
form,
Italy)
Italy) and
and used
used without
without any
any further
further purification.
purification. PE,
PE, PS,
PS, and
and PMMA,
PMMA, which
which were
were in
in powder
powder form,
form,
Italy)
Italy) and
and used without any further purification. PE, PS, and PMMA, which were in powder form,
(PLA, samplewere
were
were wasused
8) used
used
used
as
as
as
without while
received,
a commercial
received,
received,
any food-packaging
while
while
further
PC,
PC,
PC,
purification.
which
which
which
was
was
was
in
in
in
PE, PS,form,
pellet
product
pellet
pellet
and
and
form,
form,
PMMA,
was
was
was
was
whichinin
ground
used,
ground
ground in
in
were
aa film
a
in form,
mortar
mortar
mortar
powder
before
before
before asform,
use.
purchased
use.
use.
were
were used
used
Polylactide as
as received,
received,
(PLA, sample while
while
8) was PC,
PC, which
which
aa commercial was
was in
in pellet
pellet
food-packagingform,
form, was
was ground
ground
product and in
in
was a
a mortar
mortar
used, inbefore
before
film use.
use.
form,
Polylactide
without anyPolylactide (PLA,
special preliminary
(PLA, sample
sample 8)
8) was
treatments.
was commercial food-packaging
PE terephthalate
a commercial food-packaging product
(PET, and
sample
product and was
9) and
was used,
used, in film form,
polypropylene
in film form, (PP,
Polylactide
Polylactide
as purchased (PLA,
(PLA, sample
sample
without any 8) was
was aa commercial
8)special commercial
preliminary food-packaging
food-packaging
treatments. PE product
product and
and was
terephthalate was
(PET, used,
used, in
in film
sample film9)form,
form,
and
as
as purchased
purchased without
without any
any special
special preliminary
preliminary treatments.
treatments. PE
PE terephthalate
terephthalate (PET,
(PET, sample
sample 9)
9) and
and
sample 10) polypropylene
were
as beverage-
purchased
as purchased withoutwithout
(PP,
and anyfood-packaging
special
any special
sample 10) were preliminary
preliminary
beverage-
products and
treatments.
treatments.
and
were
PE used
terephthalate
PE terephthalate
food-packaging products
after(PET,
(PET,
and
washing
sample
sample
were used
and
9) drying
and
9)after
and to
polypropylene
polypropylene (PP,
(PP, sample
sample 10)
10) were
were beverage-
beverage- and
and food-packaging
food-packaging products
products and
and were
were used
used after
after
polypropylene
remove anywashing
food or
polypropylene
andwater (PP,
(PP,
drying sample
debris.
sample
to remove 10)
All
10) were
polymers
were
any beverage-
beverage-
food or were
water and
and food-packaging
dried under
food-packaging
debris. All polymers products
vacuum
products
were driedatand
and were
room
were
under used
used
vacuum after
temperature
after
at and
washing
washing and
and drying
drying to
to remove
remove any
any food
food or
or water
water debris.
debris. All
All polymers
polymers werewere dried
dried under
under vacuum
vacuum at
at
washing
washing and
and drying
drying to remove
tokept
remove aa any
any food or
foodTheir water
orunder
water debris. All
All polymers
debris. until
polymers were
were dried
dried under
under vacuum
vacuum at
at
kept in a desiccator
room
room
room under
temperature
temperature
temperature vacuum
and
and
and kept
kept
inuntil
in
in use.
desiccator
desiccator
aa desiccator under
under characteristics
vacuum
vacuum
vacuum until
until
use.
use.are
use. reported
Their
Their
Their in
characteristics
characteristics
characteristics Table are
are
are 2 together
reported
reported
reported with
room
room
in Tabletemperature
temperature
2 together and
and
with kept
kept
the in desiccator
in a desiccator
time of the under vacuum
under vacuum
long-term until
until
isothermal use. Their characteristics are
use. Their characteristics are reported
experiment. reported
the time of the
in
in long-term
Table
Table 22 together
together isothermal
with
with the
the time
time experiment.
of
of the
the long-term
long-term
in Table 2 together with the time of the long-term isothermal experiment.
isothermal
isothermal experiment.
experiment.
in Table 2 together with the time of the long-term isothermal experiment.
Table
Table 1.
1. Molecular
Molecular formula
formula of
of the
the tested
tested polymeric
polymeric materials.
materials.
Table 1.Table
Table
Table
1.
1. Molecular
Molecular
1. Molecular
Molecular
formula
formula
formula
formula
of
of the
of the
of the
the
tested
tested
tested polymeric
tested polymeric
polymeric materials.
materials.
polymeric materials.
materials.
Sample
Sample
Sample
Sample Molecular
Molecular Formula
Formula Abbreviation
Abbreviation
Sample
Code
Code Molecular
Molecular Formula
Formula Abbreviation
Abbreviation
Code
Sample Code
Code Molecular Formula
Molecular Formula Abbreviation
Abbreviation
Code

11 PEEK1
PEEK1
11 PEEK1
PEEK1
1 1 PEEK1
PEEK1

22 PEEK2
PEEK2
22 PEEK2
PEEK2
2 2 PEEK2
PEEK2

33 PEEK3
PEEK3
3 33 PEEK3
PEEK3
PEEK3
3 PEEK3

4 44 PE
PE PE
44 PE
PE
4 PE

5 55 PS
PS PS
55 PS
PS
5 PS

6 66 PC
PC PC
66 PC
PC
6 PC
Materials 2018, 11, 1383 5 of 13

Table 1. Cont.
Materials 2018,
Materials 2018, 11,
11, x FOR
FOR PEER
PEER REVIEW
REVIEW 55 of
of 13
13
Materials
Materials 2018, 11, xxx FOR
2018, 11, FOR PEER
PEER REVIEW
REVIEW of
55 of 13
13
Sample Code Molecular Formula Abbreviation

77 PMMA
PMMA
7 77 PMMA
PMMA
PMMA

88 PLA
PLA
8 88 PLA
PLA PLA

9 99 PET PET
PET
99 PET
PET

10 10
10 PP PP
PP
10
10 PP
PP

where:
where:
where: Ar=1,4-substituted
Ar=1,4-substituted
Ar=1,4-substituted phenylene.
phenylene.
phenylene.
where:
where: Ar=1,4-substituted
Ar=1,4-substituted phenylene.
phenylene.

2.2. Long-Term
2.2.
2.2. Long-Term Isothermal
Long-Term Isothermal Degradations
Isothermal Degradations
Degradations
2.2. Long-Term Isothermal Degradations
2.2. Long-Term Isothermal
Long-term
Long-term Degradations
Long-term isothermal
isothermal
isothermal degradations were
degradations
degradations were performed
were performed in
performed in static
in static air
static air atmosphere.
air atmosphere. Alumina
atmosphere. Alumina crucibles
Alumina crucibles
crucibles
Long-term isothermal degradations were performed in static air atmosphere. Alumina crucibles
containing exactly
containing exactly
containing weighed
exactly weighed quantities
weighed quantities
quantities of of the
of the investigated
the investigated polymers
investigated polymers
polymers were were
were put put
put inin
in an an oven
an oven
oven at at different
at different
different
containing
Long-term exactly weighed
isothermal degradationsquantities of the
were investigatedin
performed polymers
static were
air put in
atmosphere. an oven at different
Alumina crucibles
temperatures
temperatures
temperatures and and
and times
times
and times
times as as
as a
a function
function
as aa function
function of of
of the
the
of the polymer
polymer
the polymer tested.
tested.
polymer tested. Samples
Samples
tested. Samples
Samples 1–3 1–3
1–3
1–3 at at
at 270
270
at 270
270 °C °C
°C
°C forfor
for
for 3737
37 months;
months;
37 months;
months;
temperatures
containing samples
exactly
samples 4–7
samples weighed
4–7
4–7 at 150
at
at 150 °C
150 °Cquantities
°C for 37
for
for 37 months;
37 months;
months; of sample
the investigated
sample
sample 88 at
at 150
at 150 °C
150 °C forpolymers
°C for
for 99 months;
months; and
months; were
and
and put in
samples
samples an 10
99 and
and
and oven
10
10 at 150
at 150
150 at°C
°C
°Cdifferent
samples 4–7 at 150 °C for 37 months; sample 88 at 150 °C for 99 months; samples
and samples 99 and at
10 at 150 °C
temperatures for 25
for 25
25
and months.
months.
times as a function of the polymer tested. Samples 1–3 at 270 ◦ C for 37 months;
for months.
for 25 months.
Samples
Samples were weighed
were
◦ C for weighed at
weighed at specific
at specific time
specific time intervals
time intervals
intervals using aaa Mettler
using Mettler AE
Mettler AE 240
AE 240 electronic
240 electronic balance
electronic balance (±1
balance (±1 ×××
(±1
samples 4–7 at Samples
150
Samples were
were 37weighed
months; sampletime
at specific 150 ◦ Cusing
8 atintervals using
for 9a months;
Mettler AE and samples balance
240 electronic 9 and 10 (±1 at× 150 ◦ C
10
10 −5 g). Samples 1–3 were weighed every 30–40 days; samples 4–7 were weighed once a day (first two
−5 g).
10−5 g).
−5 Samples
g). Samples
Samples 1–3 1–3 were
1–3 were weighed
were weighed
weighed every every 30–40
every 30–40 days;
30–40 days; samples
days; samples
samples 4–7 4–7 were
4–7 were weighed
were weighed
weighed once once a
once aa dayday (first
day (first two
(first two
two
10
for 25 months.weeks of
weeks
weeks of measurements),
of measurements), and
measurements), and then
and then three
then three times
three times aaa week
times week (for
week (for 111 month),
(for month), twice
month), twice aaa week
twice week (for
week (for another
(for another
another
weeks of measurements), and then three times a week (for 1 month), twice a week (for another
Samples wereand
month),
month),
month), and
andweighed
successively
successively
successively at specific
once
once time sample
once aaa week;
week;
week; intervals
sample
sample 88 was
8 was
wasusingweighed
weighed
weighed a Mettler
once aaa day
once
once AE (first
day
day 240 electronic
(first
(first 22 months
2 months of
months ofbalance
of
month), and successively once a week; sample 8 was weighed once a day (first 2 months of
− 5 measurements),
(±1 × 10 measurements),
g). Samples 1–3 then three times a week (for 3 months), and successively once a week; samples 9–10
measurements),
measurements), thenwere
then
then
three weighed
times
three times
three times a weekevery
week
aa week
(for 3 30–40 days;
months),
(for 33 months),
(for months), and samples 4–7
successively
and successively
and successively oncewere
once
once
a week;
aa week;
week; weighed
samples
samples 9–10
samples once a day
9–10
9–10
were weighed
were
were weighed once
weighed once aa day
once day (first
day (first 222 months
(first months of
months of measurements),
of measurements), then
measurements), then 33 times
then times aaa week
times week (for
week (for 333 months),
(for months),
months),
(first two weeks
were weighed once aa day
of measurements), (firstand then of
2 months three times a week
measurements), then 33(for
times 1 amonth),
week (for twice a week (for
3 months),
successively once
successively
successively once aaa week
once week (for
week (for 333 months),
(for months), and
months), and finally
and finally once
finally once aaa month.
once month. To
month. To this
To this aim,
this aim, crucibles
aim, crucibles were
crucibles were
were
successively
another month), and once a week (for
successively once3 months),
a week; andsample
finally once8 was a month.
weighed To thisonce aim,a crucibles
day (firstwere2 months
extracted
extracted from
extracted from
from thethe oven,
the oven, cooled
oven, cooled
cooled in in a desiccator
in a desiccator
desiccator for for 1
for 1 h,h, weighed,
h, weighed,
weighed, and and then
and then immediately
then immediately
immediately put put
put in in the
in the
the
extracted from the oven, cooled in aa desiccator for 11 h, weighed, and then immediately put in the
of measurements),
oven again.
oven
oven again.
again.then three times a week (for 3 months), and successively once a week; samples
oven again.
9–10 were weighed To
To avoid
once
avoid
To avoid
avoid problems
problems
problems related2with
a day related
(first
related
related with
months
with the frequent
the
the frequent
frequent entry and
of measurements),
entry
entry and exit
and exit from
exit from
then
from the3oven,
the
the oven,
times
oven, two
two
two sets
a sets
week
sets
sets of the
of
of the
the same
(forsame3 months),
same
same
To problems with the frequent entry and exit from the oven, two of the
samples of
samples of about
of about the
about the same
the same quantities
same quantities were
quantities were put
were put inside
put inside the
inside the oven.
the oven. One
oven. One was
One was used
was used for
used for measurements
for measurements and
measurements and
and
successivelysamples
once aofweek
samples about(for 3 months),
the same quantities andwerefinally
put insideoncethe a month.
oven. OneTo this
was usedaim, for crucibles
measurements were andextracted
the second
the
the second one
second one was
one was left
was left inside
left inside the
inside the oven
the oven throughout
oven throughout the
throughout the experiment.
the experiment. At
experiment. At the
At the end
the end of
end of the
of the experiment,
the experiment, no
experiment, no
no
the
from the oven, second
cooled one in was
a left inside
desiccator thefor oven
1 h,throughout
weighed, the experiment.
and then At the
immediately end of the experiment,
put in the oven no again.
remarkable difference
remarkable
remarkable difference in
difference in mass
in mass loss
mass loss among
loss among the
among the two
the two sets
two sets of
sets of the
of the same
the same samples
same samples were
samples were recorded.
were recorded.
recorded.
remarkable
To avoid problems difference
related in mass
withloss theamong
frequent the two sets and
entry of the samefrom
exit samples thewere oven, recorded.
two sets of the same
2.3. Short-Term
samples of 2.3.
about
2.3.
2.3. Short-Term
the same
Short-Term Isothermal
Isothermal
Isothermal Degradations
quantities were put inside the oven. One was used for measurements and
Degradations
Degradations
Short-Term Isothermal Degradations
the second oneA Awas
A left
Mettler
Mettler inside
TA
TA
Mettler TA 3000
3000
TA 3000 the
TG
TG
3000 TG oven throughout
analyser,
analyser,
TG analyser,
analyser, coupled
coupled theaa experiment.
coupled with
with
with Mettler TC
Mettler TC 10AAt
TC 10A
10A the endas
processor
processor asof
as the experiment,
control
control and
and
A Mettler coupled with aa Mettler
Mettler TC 10A processor
processor control
as control and
and
evaluation
evaluation
no remarkable unit
unit
difference
evaluation unit (Mettler-Toledo,
(Mettler-Toledo,
in mass loss
unit (Mettler-Toledo, Greifensee,
Greifensee,
among the
(Mettler-Toledo, Greifensee, Switzerland),
Switzerland),
two sets of
Greifensee, Switzerland), and
and
the
Switzerland), and a
a Shimadzu
Shimadzu
same DTG-60
DTG-60
samples
and aa Shimadzu
Shimadzu DTG-60weresimultaneous
simultaneous
recorded.
DTG-60 simultaneous
simultaneous
evaluation
DTA-TG apparatus
DTA-TG
DTA-TG apparatus (Shimadzu,
apparatus (Shimadzu, Kyoto,
(Shimadzu, Kyoto, Japan)
Kyoto, Japan) were
Japan) were used
were used for
used for short-term
for short-term degradations.
short-term degradations. The
degradations. The
The
DTA-TG apparatus (Shimadzu, Kyoto, Japan) were used for short-term degradations. The
temperature
temperature calibration
calibration of TGA
of TGA instruments
TGA instruments was
instruments was made
was made according
made according to
according to the
to the procedure
the procedure suggested
procedure suggested by
suggested by the
by the
the
2.3. Short-Term Isothermal
temperature
temperature Degradations
calibration
calibration of
of TGA instruments was made according to the procedure suggested by the

A Mettler TA 3000 TG analyser, coupled with a Mettler TC 10A processor as control and
evaluation unit (Mettler-Toledo, Greifensee, Switzerland), and a Shimadzu DTG-60 simultaneous
DTA-TG apparatus (Shimadzu, Kyoto, Japan) were used for short-term degradations. The temperature
calibration of TGA instruments was made according to the procedure suggested by the two companies
Materials 2018, 11, 1383 6 of 13

and reported in previous works [23,29], based on the change of magnetic properties of 3 metal samples
at their Curie points (isatherm, 142.5 ◦ C; nickel, 357.0 ◦ C; and trafoperm, 749 ◦ C) for Mettler TGA
and using indium (NIST SRM 2232), tin (NIST SRM 2220), and zinc (NIST SRM 2221a) as standard
materials for temperature and a set of exactly weighed samples for mass for Shimadzu DTG. Isothermal
experiments were performed in static air atmosphere in order to exactly reproduce the long-term
degradation conditions.
Samples of about 5 × 10−3 g, held in alumina open crucibles, were heated (20 ◦ C·min−1 ) from
room temperature to the selected one and then maintained at this temperature for 900 min. The weight
of the samples at the start of isothermal heating was considered the initial one. Degradations were
repeated 3 times, and the D average values, where D = (W 0 − W)/W 0 , and W 0 and W, were the weights
at the starting point and during scanning, at various times were in agreement with the experimental
ones within ±3% in every case.

2.4. Kinetics Method

The set of short-term experimental isothermal data was used to evaluate the Ea of our polymers
through the MacCallum [30,31] method, based on the following linear equation:

lnt = a + b/Tiso (1)

where t was the time employed to reach a fixed degree of degradation D, a = ln[F(1 − D)] − lnA, b
= Ea /R, Tiso was the temperature of isothermal degradation, and F(1 − D) a function of degree of
degradation. The apparent activation energy at each degree of degradation was obtained by the slope
of the lnt vs. 1/Tiso straight line.

3. Results and Discussion

The study started with the isothermal degradation of 3 high thermally stable aromatic PEK
(Table 1), samples 1–3. A long-term experiment at a temperature, 270 ◦ C, largely lower than the
melting ones was carried out aiming to correlate data collected in these conditions with those obtained
by short-term experiments performed at temperatures near the melting one. The long-term experiment
was performed in an oven, whilst the short-term ones in a TG balance. An induction period of 1115,
418, and 150 days was observed for samples 1, 2, and 3 respectively, at 270 ◦ C, whilst for the short-term
experiments, no IP was recorded [28].
The different behavior observed for polymer degradation below melting and above it, in the
molten state, had the consequence that, when we tried to insert the value obtained at low temperature
into the kinetic equations obtained at a higher temperature with short-time experiments, a marked
deviation of linearity was observed [28]. This obtained the first result, which means different
degradation apparent activation energy in the two studied conditions suggested caution in the lifetime
prediction performed by using data obtained at high temperatures.
Among the questions that arose spontaneously after these first results were those related to the
typologies of used materials, very thermally stable polymers, and to the temperature chosen for the
long-term experiment, which was, however, only relatively low. Consequently, the next step was to
repeat the comparison between long- and short-term experiments using, this time, four well known
polymers, PE, PS, PC, and PMMA, and a much lower isothermal temperature, 150 ◦ C. It is worth to
note that, for three of the used polymers, the chosen temperature is very close to their Tg .
Whilst for PS (sample 5), PC (sample 6), and PMMA (sample 7), whose Tg is very close to the
isothermal one (Table 2), a long IP was observed (Figure 2), for PE (sample 4), whose Tg is more
lower [32], no IP was observed.
On considering that the IP of PC and PMMA lasted for almost the entire time of the experiment,
the comparison between long- and short-term experiments was possible only for the two polymers
that have significantly lost weight during the long-term experiment, i.e., PE and PS (Figure 3).
Materials 2018, 11, 1383 7 of 13
Materials 2018, 11, x FOR PEER REVIEW 7 of 13

Figure 2. Percentage of undegraded sample (1 − D)%, at 150 °C as a function of heating time (t) for PE
Figure 2. Percentage of undegraded sample (1 − D)%, at 150 ◦ C as a function of heating time (t) for PE
(sample 4 ), PS (sample 5 ), PC (sample 6 ), and PMMA (sample 7 σ) after 200 days of
(sample 4 ), PS (sample 5 4), PC (sample 6 #), and PMMA (sample 7 σ) after 200 days of degradation.
degradation.

Figure 3. Percentage of undegraded sample (1 − D)% at 150 °C as a function of heating time (t) for PE
Figure 3. Percentage of undegraded sample (1 − D)% at 150 ◦ C as a function of heating time (t)
(sample 4 ), PS (sample 5 ), PC (sample 6 ), and PMMA (sample 7 σ) after 1130 days of
for PE (sample 4 ), PS (sample 5 4), PC (sample 6 #), and PMMA (sample 7 σ) after 1130 days
degradation.
of degradation.
No matching among the kinetics data obtained by using the MacCallum equation was observed
No matching
(Figure among
4) for both the kinetics
polymers, but whendata
theobtained bypresent,
IP was not using thePE MacCallum equation
(Figure 4a), the was observed
mismatching was
(Figure 4) different
quite for bothinpolymers, butPS
respect to the when the4b)
(Figure IPthat
wasshowed
not present,
an IP ofPE (Figure
about 4a), The
150 days. the same
mismatching
behavior was
quite was
different in respect
observed to the
by treating thePS (Figure
kinetic data,4b) that showed
obtained an IP of about
in the short-term 150 days.
experiments, The
with same
other behavior
kinetics
methods as
was observed bywell [33,34].
treating the kinetic data, obtained in the short-term experiments, with other kinetics
methods as well [33,34].
Materials 2018, 11, 1383 8 of 13
Materials 2018, 11, x FOR PEER REVIEW 8 of 13

(a) (b)
Figure 4. MacCallum straight lines at various D values: 5% (), 7.5% (), 10% () for (a) PE and (b)
Figure 4. MacCallum straight lines at various D values: 5% (), 7.5% (#), 10% (4) for (a) PE and
PS.
(b) PS.

Table 2. Sample code, abbreviation, molecular weight (Mw), glass-transition temperature (Tg), onset
Table 2. Sample code, abbreviation, molecular weight (Mw ), glass-transition temperature (Tg ), onset
melting temperature (Tonset), and time of long-term experiment of the tested polymeric materials.
melting temperature (Tonset ), and time of long-term experiment of the tested polymeric materials.
Sample Experimental
Sample
Polymer Mw Tg °C Tonset °C Ref. Experimental
Code Polymer Mw Tg C◦ ◦
T onset C Ref. Time/Months
Code Time/Months
1 Polyetherketones 1 n.a. n.d. 417.0 [27] 37
1 Polyetherketones 1 n.a. n.d. 417.0 [27] 37
22 Polyetherketones
Polyetherketones 1 1 n.a.n.a. n.d.
n.d. 379.0
379.0 [27]
[27] 3737
33 Polyetherketones
Polyetherketones 1 1 n.a.n.a. n.d.
n.d. 434.0
434.0 [27]
[27] 3737
44 Polyethylene
Polyethylene (PE)(PE) 60,000
60,000 n.a. n.a. 129.2
129.2 [31]
[31] 3737
5 Polystyrene (PS) 65,000 99.5 n.d. [31] 37
56 Polystyrene
Polycarbonate (PS)
(PC) 65,000 145.9
50,000 99.5 n.d.
n.d. [31]
[31] 3737
67 Polycarbonate
Poly(methyl methacrylate (PC)
(PMMA) 50,000 145.9
65,000 118.1 n.d.
n.d. [31]
[31] 3737
78a Polylactide (PLA) 1
Poly(methyl methacrylate (PMMA) n.a. 65,000 118.160.0 140.4
n.d. [33]
[31] 379
8b Polylactide 2 n.a. 63.5 139.9 [33] 9
8a9a Polylactide
Polyethylene (PLA)
terephthalate 1 1
(PET) n.a.n.a. 60.0
76.9 140.4
229.2 [33]
[34] 925
8b9b Polylactide
Polyethylene 2 2
terephthalate n.a.n.a. 63.5
33.0 139.9
140.8 [33]
[34] 925
10a Polypropylene (PP) 1 n.a. 47.8 151.0 [34] 25
9a
10b
Polyethylene terephthalate (PET) 1 n.a.n.a.
Polypropylene 2
76.9
58.9
229.2
131.6
[34]
[34]
2525
9b Polyethylene terephthalate 2 n.a. 33.0 140.8 [34] 25
10a Polypropylene (PP) 1 n.a. 47.8 151.0 [34] 25
After all these years
10b of long-term2 experiments, n.a.
Polypropylene the first58.9
conclusion131.6was that lifetime predictions
[34] 25
of polymers in service made by degradation experiments at higher temperatures seems not reliable.
The possibility to find
After all these useful
years structure–IP
of long-term correlations
experiments, thebegan to make itswas
first conclusion way.that lifetime predictions
After studying very stable synthetic polymers (samples 1–3)
of polymers in service made by degradation experiments at higher temperatures and model polymers
seems (samples 4–7),
not reliable.
to
Thegopossibility
ahead, with the aim
to find usefulto structure–IP
strength thiscorrelations
structure–IPbegan
correlation,
to maketheits choice
way. fell on two different
commercial PLA (samples 8a and 8b in this manuscript) used
After studying very stable synthetic polymers (samples 1–3) and model for food packaging
polymers [35]. PLA was
(samples 4–
selected for two different reasons:
7), to go ahead, with the aim to strength this structure–IP correlation, the choice fell on two different
-commercial
DespitePLA (samples 8a its
the advertizing, and 8b in this
recycling manuscript)
is not economicallyused for food
viable and, packaging [35].
since it is one ofPLA was
the most
selected for two different reasons:
used polymeric materials in food packaging because of its ability to be degraded, it represents
- one of the
Despite themajor domestic
advertizing, itswaste materials;
recycling is not thus, it could be
economically interesting
viable andituseful
and, since is onetoofpredict,
the mostat
least roughly,
used polymeric itsmaterials
end life ininthe environment.
food packaging because of its ability to be degraded, it represents
- On considering
one of the major the original
domestic purpose
waste of thethus,
materials; workit to verify
could the validityand
be interesting of lifetime
useful topredictions
predict, at
by extrapolating
least the life
roughly, its end kinetics
in thedata obtained at high temperatures, PLA was chosen because
environment.
- of its
On low glassthe
considering transition
original and melting
purpose of thetemperature.
work to verifyInthe
respect to of
validity thelifetime
polymers examined
predictions by
in the past, itthe
extrapolating suggested
kinetics the
datapossibility
obtained of a sizable
at high degradation
temperatures, PLAin was
a reasonable time during
chosen because of its
long-term
low measurements.
glass transition and melting temperature. In respect to the polymers examined in the past,
it suggested the possibility of a sizable degradation in a reasonable time during long-term
measurements.
Materials 2018, 11, 1383 9 of 13

Materials 2018,
Sample 11, x FOR PEER
8a started REVIEW immediately, differently than compound 8b, which showed
to degrade 9 of 13 an

initial IP of about ten days. As expected, the degradation in this case was faster and the two PLA
Sample 8a started to degrade immediately, differently than compound 8b, which showed an
samples, which reached almost complete degradation after just 300 days (Figure 5), a third of the time
initial IP of about ten days. As expected, the degradation in this case was faster and the two PLA
needed by thewhich
samples, polymers
reachedpreviously examined
almost complete (samples
degradation after1–7), to achieve
just 300 a much
days (Figure lower
5), a third ofdegradation
the time
percentage. The degradation data obtained during the long-term experiment were
needed by the polymers previously examined (samples 1–7), to achieve a much lower degradation also in this case
compared with those
percentage. obtained from
The degradation data the short-term
obtained duringone, and a slightly
the long-term mismatching
experiment were alsowasin observed
this case for
compared with
both investigated those obtained
samples [35]. from the short-term one, and a slightly mismatching was observed for
both
In investigated
order to quantifysamplesthe
[35].uncertainty in the lifetime prediction of the studied polymers,
In order to quantify
a comparison among the measured the uncertainty
trend ofinthe
thereal
lifetime predictionand
degradation of the
thatstudied polymers,
extrapolated from a the
comparison among the measured trend of the real degradation and that extrapolated from the
MacCallum equations has been performed, showing about twelve days of difference for sample 8a
MacCallum equations has been performed, showing about twelve days of difference for sample 8a
and about sixteen days for sample 8b [35]. It is worth to note that also in this case, where no IP was
and about sixteen days for sample 8b [35]. It is worth to note that also in this case, where no IP was
present, the real and the expected values were closer than the polymer for which an IP (although the
present, the real and the expected values were closer than the polymer for which an IP (although the
slightest) is present.
slightest) is present.

Figure 5. Percentage of undegraded sample (1 − D)% at 150 °C as a function of heating time (t) for
Figure 5. Percentage of undegraded sample (1 − D)% at 150 ◦ C as a function of heating time (t) for
PLA (sample 8a ) and (sample 8b ) after 300 days of degradation.
PLA (sample 8a #) and (sample 8b ) after 300 days of degradation.
On continuing the research in this direction, two different types of commercial PET (samples 9a
On
andcontinuing the research
9b) and PP (samples in 10b)
10a and this direction, two different
used for beverage and foodtypes of commercial
packaging PETfor
were degraded (samples
more 9a
and 9b)
than two years in isothermal oxidative condition at 150 °C [36]. Long-term degradation showed for
and PP (samples 10a and 10b) used for beverage and food packaging were degraded thatmore
than two years in isothermal oxidative condition at 150 ◦ C [36]. Long-term degradation showed that
sample 9a started to degrade immediately without IP, differently than samples 10a and 10b, which
sampleboth startedatoshort
9ashowed degrade immediately
IP of about without
twenty days. IP, differently
Sample thanappreciable
9b did not show samples 10a andloss
mass 10b, which
and,
thus, IP for
both showed the whole
a short time oftwenty
IP of about the experiment (Figure9b
days. Sample 6).did
Thenot
overall
showmass losses, in mass
appreciable aboutloss
750 days,
and, thus,
IP forwere 1.5, 79, time
the whole 64, and
of 65% for polymers(Figure
the experiment 9a, 9b, 10a, and overall
6). The 10b, respectively. According
mass losses, in aboutto the
750practice
days, were
followed for the other polymers previously investigated, the data acquired from the
1.5, 79, 64, and 65% for polymers 9a, 9b, 10a, and 10b, respectively. According to the practice followed long-term
experiments were compared with thermo-oxidative degradations, carried out on the same materials
for the other polymers previously investigated, the data acquired from the long-term experiments
at higher temperatures. The comparison was obviously performed for the polymers (samples 9a, 10a,
were compared with thermo-oxidative degradations, carried out on the same materials at higher
and 10b) that showed appreciable mass loss during the long-term experiment.
temperatures. The comparison was obviously performed for the polymers (samples 9a, 10a, and 10b)
that showed appreciable mass loss during the long-term experiment.
Materials 2018, 11, 1383 10 of 13
Materials2018,
Materials 2018,11,
11,xxFOR
FORPEER
PEERREVIEW
REVIEW 10of
10 of13
13

Figure6.6.Percentage
Figure Percentage ofundegraded
undegraded sample(1(1 − D)%at
at 150°C
°C asaafunction
function ofheating
heating time(t)
(t) for
Figure 6. Percentage of
of undegraded sample
sample (1 −−D)%
D)% at 150
150 ◦ C as
as a function of
of heating time
time (t) for
for
9a and )and and
(sample 9a 
PET (sample9a
PET and9b9b) andPP
PP(sample 10a
(sample10a and10b
10bσ)
σ)after
after750
750days
daysof
ofdegradation.
degradation.
PET (sample 4 and 9b #) and PP (sample 10a  and 10b σ) after 750 days of degradation.

In the
In
In the case
the case of
case of sample
of sample 9a
sample 9a for
9a for the
for the first
the first time,
first time, differently
time, differently than
differently than what
than what we
what we found
we found for
found for all
for all the
all the other
the other
other
experiments,
experiments, the
the comparison
comparison of
of the
the long-term
long-term experimental
experimental data
data with
with those
those obtained
obtained
experiments, the comparison of the long-term experimental data with those obtained in short-term in
in short-term
short-term
modalityresulted
modality
modality resultedin
resulted in
inaaaperfect
perfect match
perfectmatch (Figure
match(Figure
(Figure 7).7).
7). ForFor
For thethe
the other
other samples,
samples,
other expectations
expectations
samples, havehave
have
expectations alsobeen
also been
also
confirmed.
confirmed.
been confirmed.

Figure 7.MacCallum
Figure MacCallum straightlines
lines atvarious
various Dvalues:
values: 5%(),
(), 7.5%(),
(), 10%(),
(), 12.5%(σ),
(σ), and15%
15%
Figure 7.7. MacCallum straight
straight lines at
at various D
D values: 5%
5% (), 7.5%
7.5% (#), 10%
10% (4), 12.5%
12.5% (σ), and
and 15%
()
() for Sample
for Sample 9a.
Sample 9a.
9a.
(3) for

Forsample
For sample9b,
9b,which
whichpresented
presentedaavery
verylong
longIPIPat
at150
150°C,
°C,no
nomatching
matchingwas waspossible.
possible.Instead,
Instead,for
for
samples10a
samples 10aand
and10b,
10b,whose
whoseIPs
IPswere
werevery
veryshort,
short,aadifference
differenceofofthe
theorder
orderof ofnine
ninedays
dayswaswasrecorded
recorded
in the
in the comparison
comparison with
with real
real data,
data, highlighting
highlighting the
the hypothesis
hypothesis that
that the
the IP
IP greatly
greatly influences
influences the
the
lifetimepredictions
lifetime predictionsachieved
achievedbybyextrapolating
extrapolatingdata
dataobtained
obtainedininthe
themolten
moltenstate.
state.
Materials 2018, 11, 1383 11 of 13

For sample 9b, which presented a very long IP at 150 ◦ C, no matching was possible. Instead, for
samples 10a and 10b, whose IPs were very short, a difference of the order of nine days was recorded
in the comparison with real data, highlighting the hypothesis that the IP greatly influences the lifetime
predictions
Materials 2018, achieved by extrapolating
11, x FOR PEER REVIEW data obtained in the molten state. 11 of 13
Is the discrepancy between real and expected data (RE) dependent on the length of IP? If we
look Is
at the discrepancy
Figure between
8, where this real and
discrepancy expected as
is reported data (RE) dependent
a function of the IPson the length
obtained fromofthe
IP? If we
various
look at Figureperformed
experiments 8, where this discrepancy
with is reportedthe
different polymers, as aexistence
function of the IPs obtainedappears
a relationship from the various
clear and
experiments performed with different polymers, the existence of a relationship appears
probably will help those people who are approaching to realize lifetime predictions by using the IP clear and
probably will helpaccording
length to correct, those people
to thewho areequation:
below approaching to realize lifetime predictions by using the IP
length to correct, according to the below equation:
RE = 13.9 (±4) + 0.134 (±0.01) IP (2)
RE = 13.9 (±4) + 0.134 (±0.01) IP (2)
the extrapolated data from measurements performed in the molten state.

Figure
Figure 8.
8. Relationship
Relationship between
between the
the discrepancy
discrepancy (among
(among the
the real
real and
and expected
expected data)
data) and
and the
the IPs
IPs
observed in the long-term degradation experiments.
observed in the long-term degradation experiments.

4.
4. Conclusions
Conclusions
The
The time
time toto answer
answer the
the questions
questions made
made atat the
the beginning
beginning ofof this
this long
long work
work and
and in
in the
theintroduction
introduction
of
of this
this article
article has
has come:
come: which
which parameter
parameter must
must be
be considered
considered to to perform
perform correct
correct and
and reliable
reliable lifetime
lifetime
prediction of materials? Could one make a prediction relying exclusively
prediction of materials? Could one make a prediction relying exclusively on data obtained on data obtained at high at
temperatures?
high temperatures?
If,
If, on
on the
the one
one hand,
hand, experimental
experimental experience
experience and and the
the literature
literature confirm
confirm the
the importance
importance of of the
the
calculation of the E of degradation to perform a lifetime prediction, the work carried
calculation of the Ea of degradation to perform a lifetime prediction, the work carried out in these years
a out in these
years and, above
and, above all, theall, the correlations
correlations here put here put inbetween
in place place between
the various the data
various data obtained,
obtained, with
with different
different
polymerspolymers and at temperatures,
and at different different temperatures, from long-term
from long-term experiments experiments and the
and the relative relative
IPs, IPs,
prove that
prove that we can not disregard this latter parameter in lifetime predictions.
we can not disregard this latter parameter in lifetime predictions. In particular, a relationship to In particular, a
relationship
help people to in help peoplethe
correcting in data
correcting the by
obtained data obtained
TGA by TGAperformed
experiments experiments performed
at high at high
temperatures
temperatures
is proposed. is proposed.

Author Contributions: I.B. conceived the experimental plan, performed the thermal testing, carried out the
interpretation of the data, and wrote the manuscript.

Funding: Ignazio Blanco is grateful to the MIUR for the grant “Fund for basic research activities”, and to the
Department of Civil Engineering and Architecture of the University of Catania for supporting the project
MATErials LIfe foreCAst (MATELICA).
Materials 2018, 11, 1383 12 of 13

Author Contributions: I.B. conceived the experimental plan, performed the thermal testing, carried out the
interpretation of the data, and wrote the manuscript.
Funding: Ignazio Blanco is grateful to the MIUR for the grant “Fund for basic research activities”, and to
the Department of Civil Engineering and Architecture of the University of Catania for supporting the project
MATErials LIfe foreCAst (MATELICA).
Conflicts of Interest: The author declares no conflict of interest.

References
1. Ozao, R.; Popescu, C. Special Chapter in the recent advancement in North America on the field of Thermal
Analysis and Calorimetry. J. Therm. Anal. Calorim. 2018, 131, 237. [CrossRef]
2. Yang, R. Status and challenge in aging research of polymer composites. Polym. Mater. Sci. Eng. 2015,
31, 181–184.
3. Liu, X.; Zhao, J.; Yang, R.; Iervolino, R.; Barbera, S. A novel in-situ aging evaluation method by FTIR and the
application to thermal oxidized nitrile rubber. Polym. Degrad. Stabil. 2016, 128, 99–106. [CrossRef]
4. Blanco, I.; Bottino, F.A.; Cicala, G.; Latteri, A.; Recca, A. A kinetic study of the thermal and thermal oxidative
degradations of new bridged POSS/PS nanocomposites. Polym. Degrad. Stabil. 2013, 98, 2564–2570.
[CrossRef]
5. Cangialosi, D. Glass Transition and Physical Aging of Confined Polymers Investigated by Calorimetric
Techniques. Handb. Therm. Anal. Calorim. 2018, 6, 301–337. [CrossRef]
6. De Eulate, P.N.G.; Cangialosi, D. The very long-term physical aging of glassy polymers. Phys. Chem.
Chem. Phys. 2018, 20, 12356–12361. [CrossRef] [PubMed]
7. Cicala, G.; Latteri, A.; Mannino, S.; Ognibene, G.; Blanco, I. Influence of Soluble Electrospun
Co-Polyethersulfone Veils on Dynamic Mechanical and Morphological Properties of Epoxy Composites:
Effect of Polymer Molar Mass. Adv. Polym. Technol. 2018, 37, 798–809. [CrossRef]
8. White, C.C.; White, K.M.; Pickett, J.E. Service Life Prediction of Polymers and Plastics Exposed to
Outdoor Weathering; William Andrew Applied Science Publishers: Norwich, UK, 2018; pp. 1–342,
ISBN 978-0-323-49776-3.
9. Friedman, H.L. Kinetics of thermal degradation of char-forming plastics from thermogravimetry. Application
to a phenolic plastic. J. Polym. Sci. Part C Polym. Symp. 1964, 6, 183–195. [CrossRef]
10. Friedman, H.L. New methods for evaluating kinetic parameters from thermal analysis data. J. Polym. Sci.
Part B Polym. Lett. 1969, 7, 41–46. [CrossRef]
11. Flynn, J.H.; Wall, L.A. A quick direct method for the determination of activation energy from
thermogravimetric data. J. Polym. Sci. B Polym. Lett. 1966, 4, 323–328. [CrossRef]
12. Ozawa, T. A new method of analyzing thermogravimetric data. Bull. Chem. Soc. Jpn. 1965, 38, 1881–1886.
[CrossRef]
13. Vyazovkin, S. Modification of the integral isoconversional method to account for variation in the activation
energy. J. Comput. Chem. 2001, 22, 178–183. [CrossRef]
14. Šesták, J.; Berggren, G. Study of the kinetics of the mechanism of solid-state reactions at increasing
temperature. Thermochim. Acta 1971, 3, 1–12. [CrossRef]
15. Šimon, P. Fourty years of the Šesták–Berggren equation. Thermochim. Acta 2011, 520, 156–157. [CrossRef]
16. Fatemi, N.; Whitehead, R.; Price, D.; Dollimore, D. Some comments on the use of Avrami-Erofeev expressions
and solid state decomposition rate constants. Thermochim. Acta 1986, 104, 93–100. [CrossRef]
17. Bruijn, T.J.W.; Jong, W.A.; Van Den Berg, P.J. Kinetic parameters in Avrami–Erofeev type reactions from
isothermal and non-isothermal experiments. Thermochim. Acta 1981, 45, 315–325. [CrossRef]
18. Kissinger, H.E. Reaction kinetics in differential thermal analysis. Anal. Chem. 1957, 29, 1702–1706. [CrossRef]
19. Abate, L.; Blanco, I.; Cicala, G.; Recca, G.; Scamporrino, A. The influence of chain-ends on the thermal and
rheological properties of some 40/60 PES/PEES copolymers. Polym. Eng. Sci. 2009, 49, 1477–1483. [CrossRef]
20. Abate, L.; Asarisi, V.; Blanco, I.; Cicala, G.; Recca, G. The influence of sulfonation degree on the thermal
behaviour of sulfonated poly(arylene ethersulfone)s. Polym. Degrad. Stabil. 2010, 95, 1568–1574. [CrossRef]
21. Dobkowski, Z.; Rudnik, E. Lifetime prediction for polymers via the temperature of initial decomposition.
J. Therm. Anal. 1997, 48, 1393–1400. [CrossRef]
Materials 2018, 11, 1383 13 of 13

22. Dobkowski, Z. Thermal analysis techniques for characterization of polymer materials. Polym. Degrad. Stabil.
2006, 91, 488–493. [CrossRef]
23. Blanco, I.; Oliveri, L.; Cicala, G.; Recca, A. Effects of novel reactive toughening agent on thermal stability of
epoxy resin. J. Therm. Anal. Calorim. 2012, 108, 685–693. [CrossRef]
24. Li, X.-G.; Huang, M.-R.; Bai, H. Kinetics of thermal degradation of liquid-crystalline aromatic polymers.
Angew. Makromol. Chem. 1998, 256, 9–19. [CrossRef]
25. Wang, H.; Tao, X.; Newton, E. Thermal degradation kinetics and lifetime prediction of a luminescent
conducting polymer. Polym. Int. 2004, 53, 20–26. [CrossRef]
26. Abate, L.; Blanco, I.; Cicala, G.; La Spina, R.; Restuccia, C.L. Thermal and rheological behaviour of some
random aromatic polyethersulfone/polyetherethersulfone copolymers. Polym. Degrad. Stabil. 2006,
91, 924–930. [CrossRef]
27. Ebewele, R.O. Polymer Science and Technology; CRC Press: Boca Raton, NY, USA, 2000; ISBN 0-8493-8939-9.
28. Abate, L.; Blanco, I.; Motta, O.; Pollicino, A.; Recca, A. The isothermal degradation of some polyetherketones:
A comparative kinetic study between long-term and short-term experiments. Polym. Degrad. Stabil. 2002,
75, 465–471. [CrossRef]
29. Abate, L.; Blanco, I.; Pappalardo, A.; Pollicino, A. A kinetic study of the thermal and oxidative degradations
of a new poly(arylene)ether copolymer. J. Therm. Anal. Calorim. 2001, 65, 373–380. [CrossRef]
30. MacCallum, J.R. Comprehensive Polymer Science; Allen, G., Ed.; Pergamon Press: Oxford, UK, 1989; Volume 1.
31. Hill, D.J.T.; Dong, L.; O’Donnell, J.H.; George, G.; Pomery, P. Thermal degradation of polymethacrylonitrile.
Polym. Degrad. Stabil. 1993, 40, 143–150. [CrossRef]
32. Fakirov, S.; Krasteva, B. On the Glass Transition Temperature of Polyethylene as Revealed by Microhardness
Measurements. J. Macromol. Sci.-Phys. B 2000, 39, 297–301. [CrossRef]
33. Blanco, I.; Abate, L.; Antonelli, M.L. The regression of isothermal thermogravimetric data to evaluate
degradation Ea values of polymers: A comparison with literature methods and an evaluation of lifetime
prediction reliability. Polym. Degrad. Stabil. 2011, 96, 1947–1954. [CrossRef]
34. Blanco, I.; Abate, L.; Antonelli, M.L.; Bottino, F.A. The regression of isothermal thermogravimetric data to
evaluate degradation Ea values of polymers: A comparison with literature methods and an evaluation of
lifetime predictions reliability. Part II. Polym. Degrad. Stabil. 2013, 98, 2291–2296. [CrossRef]
35. Blanco, I. End-life Prediction of Commercial PLA Used for Food Packaging through Short Term TGA
Experiments: Real Chance or Low Reliability? Chin. J. Polym. Sci. 2014, 32, 681–689. [CrossRef]
36. Blanco, I. Lifetime prediction of food and beverage packaging wastes. J. Therm. Anal. Calorim. 2016,
125, 809–816. [CrossRef]

© 2018 by the author. Licensee MDPI, Basel, Switzerland. This article is an open access
article distributed under the terms and conditions of the Creative Commons Attribution
(CC BY) license (http://creativecommons.org/licenses/by/4.0/).