Polymer Degradation and Stability 79 (2003) 385397

www.elsevier.com/locate/polydegstab

Degradation prole of polyethylene after articial accelerated weathering

J.V. Gulminea, P.R. Janissekb, H.M. Heisec, L. Akcelrudd,*
a
Instituto de Tecnologia para o Desenvolvimento LACTEC, Caixa Postal 19067, CEP 81531-990 Curitiba/PR, Brazil
Curso de Farmacia e Bioqumica, Universidade Tuiuti do Parana, Rua Marcelino, Champagnat 505, CEP 80710-250 Curitiba/PR, Brazil
c
Institute of Spectrochemistry and Applied Spectroscopy at the University of Dortmund, Bunsen-Kirchho-Str. 11, D-44139 Dortmund, Germany
d
Universidade Federal do Parana UFPR, Departamento de Qumica, Centro Politecnico da UFPR, Caixa Postal 19081, CEP 81531-990, Curitiba/PR, Brazil
b

Received 30 April 2002; received in revised form 14 October 2002; accepted 21 October 2002

Abstract
Elucidation of the chemical changes that take place during the aging of polyethylene sheared electrical cables, and the correlation
of these changes with physical properties need experiments with accelerated aging assays. These were carried out using the most
representative polyethylene grades found in todays market: low density (LDPE), linear low density (LLDPE) and high density
(HDPE). The samples were exposed to UV- and xenon arc radiation using a varying exposure time and dierent temperature cycles.
The changes in the material structure and properties were analyzed by dierential scanning calorimetry (DSC), scanning electron
microscopy (SEM), FTIR-spectroscopy, density and hardness measurements. During degradation polar groups in the polyethylene
are generated, as well as chain scission and cross-linking. As net eects of these processes respective increases in crystallinity, density and hardness and surface cracking at late stages of degradation are observed. By varying the angle of incidence in attenuated
total reection (ATR) experiments using FTIR-spectroscopy it was possible to detect variations in composition of distinct layers
from the surface down to a depth of 1.2 mm. Deeper layers were reached by microtoming the samples in layers down to 60 mm. The
main chemical modications were carbonyl formation of various kinds which were identied in the FTIRATR spectra of degraded
polyethylene samples.
# 2002 Elsevier Science Ltd. All rights reserved.
Keywords: Polyethylene; Degradation; FTIRATR

1. Introduction
Polyethylene has been widely employed as an insulating material due to its dielectric properties combined
with long durability, low cost, easy manufacture, and
good chemical resistance [1,2]. Due to its chemical
inertness, a long term resistance to the most adverse
conditions could in principle be anticipated. Eventually,
however, it became evident that, when submitted to
prolonged exposure to weathering conditions [37],
extruded polyethylene used in insulation of energy distribution cables was susceptible to photo [820], thermal
[2137] and mechanical [38] degradation. The important
kinetic aspects of polyethylene degradation have been
* Corresponding author. Tel.: +55-41-336-7507; fax: +55-41-2663582.
E-mail addresses: [email protected] (J.V. Gulmine), janissek@
zaz.com.br (P.R. Janissek), [email protected] (H.M. Heise),
[email protected] (L. Akcelrud).

addressed by Gugumus [3843]. The failures observed

result from the chemical modications that take place
within the material and lead synergetically to performance changes. The study of the aging of polyethylene
through weathering is of great importance for the
development of more resistant materials than available
in the past, and of additives that will extend the durability of such insulating polymers.
Much work has been carried out aiming at the analysis of the chemical modications generated by aging and
its consequences on the dielectric properties. Frequently
used methods rely on calorimetric, spectroscopic and
microscopic techniques, mostly in combination with
each other [5,9,11,21,22,4454]. Among recent investigations aspects of enhanced degradability of polyethylene using various additives have also been studied
[9,44]. The most important modication is through
polymer oxidation, and this eect has been recently
addressed by FTIR-spectroscopy to characterize and
quantify the oxidation products, using the attenuated

0141-3910/03/$ - see front matter # 2002 Elsevier Science Ltd. All rights reserved.
PII: S0141-3910(02)00338-5

386

J.V. Gulmine et al. / Polymer Degradation and Stability 79 (2003) 385397

total reection (ATR) technique for surface characterization [9,21,4449] or transmission spectroscopy for
specimens in the form of microtomed or pressed lms
[26,5561].
Of particular relevance is information regarding the
qualitative and quantitative aspects of the degradation
below the surface, i.e., the depth prole of degradation
products within a given sample, after having been submitted to specic aging conditions. One way to determine the degradation prole is to microtome the sample
and, thus, to have access to dierent depths. However,
in order to get the prole of very ne layers, with
thickness less than 10 mm for example, the use of
microtomy becomes complicated. It is very dicult to
cut smooth slices of that order. A non-invasive and nondestructive method to approach this problem is to make
use of the full potential of the FTIRATR technique,
using dierent crystal materials and dierent reection
angles for the infrared beam at the samplecrystal interface. Such a procedure allows the analysis of layers up to
dierent depths below 2 mm. Furthermore, with the
appropriate mathematical treatment it is also possible to
isolate distinct layers for analysis, disregarding the inuence of the preceding layers. A review on rened experimental techniques will be given, further on.
In this paper, our contribution is dedicated to the
development of a methodology for analyzing and evaluating several polyethylene based materials after accelerated aging. It will be useful in the future to characterize
alterations and failures in aged materials based on the
diagnostic techniques chosen by us, with a particular
emphasis on the use of infrared ATR spectroscopy.
Results that were obtained from microtomed samples at
dierent depths and those obtained by FTIRATR spectroscopy are combined. The following, in todays market
most important commercial polymer grades were studied:
low density polyethylene (LDPE), linear low density
(LLDPE) and high density polyethylene (HDPE).

for LDPE, LLDPE and HDPE respectively, at cycles

with no load for 5 min, with 1.62
108 Pa for 5 min, and
with 3.25
108 Pa for another 5 min. The sheets were
allowed to cool down to ambient temperature.
Thin sample layers were cut with a microtome, model
RM2165, from LEICA (Wetzlar, Germany). The
microtoming was performed at 20  C with the samples
embedded in an inclusion gel with a metallic support
attached to the cryo-microtome.
2.2. Sample aging
Two accelerated methods that simulate dierent
weathering conditions were employed. For the rst
method, the Weather-Ometer (WOM) equipment from
ATLAS (Chicago, USA), model CI65, was used following the ASTM G26/95 method, (method A) [62].
This instrument works with a xenon lamp of 6500 W,
providing an irradiance of 0.35  0.03 W/m2 (at 340 nm)
to simulate the sunlight of Florida at noon, with cycles
lasting 102 min with an incident radiation beam at 90
and a temperature of 63  5  C, a relative humidity of
60  5% and additional 18 min with a spray using deionized water, a temperature of 50 5  C and a relative
humidity of 80  5%. The individual tests lasted for 25,
50, 100, 200, 400, 800 and 1600 h, respectively. The
second accelerated aging method was performed using
the Weathering Tester (QUV) equipment, model QUV
spray UV40, from QPanel Lab Products (Cleveland,
USA) following the ASTM G53-96 method, using a
uorescent bulb UVB with 0.60 W/m2 irradiance (at 313
nm), with cycles of UV irradiationwith an incident beam
at 90 and a temperature at 60  C lasting for 8 h and an
additional period with a spray of de-ionized water and a
temperature at 50  C lasting for 4 h. The samples were
submitted to the aging process for a duration of 12, 24, 50,
100, 200, 400 and 800 h, respectively [63].
2.3. Analytical methods

2. Experimental
2.1. Sample preparation
Three dierent grades of polyethylene samples were
obtained as pellets from OPP Polietilenos S.A. (Camacari, Brazil): LDPE (code PB 681/59), LLDPE (code
LFH 0726/20 AF), and HDPE (code HF 150), all stabilized for processing with Inganox 1076 (octadecyl 1-3(3,5-di-tert.butyl-4-hydroxyphenyl)-propionate)
from
Ciba Geigy and Weston 399 (Tris(nonyl phenyl)phosphite) from General Electric. Polymer sheets were
made from pellets with dimensions of (11
11
0.1 cm),
using aluminum molds and polyester foils in a
SCHULZ press, model PHS 15 (Joinville, Brazil). The
thermo-pressing was carried out at 130, 140 and 150  C,

2.3.1. Determination of apparent density

The density was measured following ASTM D792-91
[64] using a SARTORIUS analytical balance (Gottingen, Germany), model BP210S, and the specic accessory (model YDK01) for density measurements.
Isopropanol from NUCLEAR PA. (Sao Paulo, Brazil)
with a purity of 99.5% and a density of 0.78336 g/cm3
at 22  C was used as a liquid medium. Four measurements were carried out for each sample of the new and
aged materials, and the results were averaged.
2.3.2. Hardness
The new and aged samples were tested in a MICROTEST equipment (Sao Paulo, Brazil), model 720b-SB,
with a Shore D Durometer (providing a weight of 5 kg
during 30 s) and a CHRONOS cronometer following

J.V. Gulmine et al. / Polymer Degradation and Stability 79 (2003) 385397

387

ASTM procedure No. D2240-91 [65]. The results from

ve measurements for each sample were averaged.
2.3.3. Dierential scanning calorimetry (DSC)
The equipment from NETZSCH (Selb/Bavaria, Germany), model STA 40S, was used to test the new and
aged polymer specimens with scans from 20 to 300  C at
a rate of 5  C/min, under a nitrogen atmosphere during
the rst step and with the introduction of oxygen in a
second step (both at a ow rate of 55 ml/min).
2.3.4. Scanning electron microscopy (SEM)
The sample surfaces were coated with a layer of gold
with a thickness of 810 nm by sputtering using a Baltec
(Liechtenstein, Switzerland) SCD 005 equipment, and
then analysed in a Philips (Hillsboro, U.S.A.) XL 30
electron microscope using a voltage of 10 kV. Photomicrographs were taken with a 500-fold magnication.
2.3.5. FTIR-spectroscopy
The equipment used was a FTIR-spectrometer from
BOMEM (Quebec, Canada), model DA8. Spectra were
recorded with a resolution of 2 cm1 using Bartlet (triangular) apodization; 128 interferogram scans were
averaged to give spectra from 400 to 5000 cm1 when
recorded in transmission and from 650 to 5000 cm1
when using the ATR technique. For detection a thermal
DTGS-detector (deuterated triglycine sulfate doped
with alanine in CsI windows) was used. The ATR
accessory was from Spectra-Tech (Shelton, CT, USA),
model 300 allowing a variable angle of incidence for the
reection measurements between 30 and 60 ; crystals of
ZnSe (refractive index nZnSe=2.4 at 1000 cm1, useful
range 20,000700 cm1, mean refractive index 2.42) and
of Germanium (refractive index nGe=4.0 at 1000 cm1,
useful range 5000900 cm1, mean refractive index 4.0).
2.3.6. EDSenergy dispersive spectroscopy
Chemical analysis (microanalysis) in the scanning
electron microscope (SEM) was performed by measuring the wavelength and intensity distribution of X-ray
signal generated by a focused electron beam on the
specimen, with the attachment of the energy dispersive
spectrometer (EDS) to the electron microscope. The
sample preparation was the same used for SEM analyses. The voltage was 20 kV, and the magnication was
200
. The scanned area was 350
470 mm. The composition was determined by using internal standards.

3. Results and discussion

3.1. Apparent density measurements
As illustrated by Fig. 1, the apparent density
increased with the time of weathering exposure. These

Fig. 1. Variation in density of dierent polyethylene grades versus

accelerated aging duration under WOM- (a) and QUV-weathering
conditions (b), respectively.

results can be attributed to increases in crystallinity and

at the same time to cross-linking reactions, since in both
cases the material becomes denser due to a tighter
packing, and also due to the incorporation of oxygen,
rendering the polymer heavier.
The aging within the QUV-equipment had a more
pronounced eect on density than that within the
WOM-unit, based on the same exposure time. This
result was expected, since the irradiance (radiant incident power per unit area upon a surface, usually
expressed in W/m2 or J/m2) in QUV is practically twofold the corresponding to WOM. The magnitude of the
irradiance takes into account the wavelength of the
incident light. In QUV the wavelength is 313 nm, of
greater energy than the 340 nm wavelength in WOM,
both located in the most deleterious radiation range for
organic compounds: 295385 nm [66,67]. The largest
variation in density was observed in polyethylene with
low initial density and crystallinity (LDPE): from
0.9224 to 0.9427 g/cm3 for WOM and to 0.9551 g/cm3
for the QUV weathering tester, comparing non-aged
samples with those submitted to the longest exposure
times (800 h), in each chamber. It noteworthy that these
density values are comparable to those corresponding to
HDPE. In LDPE the amorphous content is higher than
LLDPE and HDPE, rendering the material more susceptible to oxidative degradation and cross-linking [9].
The two low density sample types studied showed a
similar behavior when working with the WOM-equipment,

388

J.V. Gulmine et al. / Polymer Degradation and Stability 79 (2003) 385397

as can be seen from the coincidental closeness of

the experimental data points in Fig. 1a. For QUVweathering conditions, the LDPE-material behaved similarly, but showing a considerable increase in density for
the last stage of exposure (3.8% of its initial value, after
800 h). This renders the sample to be an outlier, because
through the severe cracks in the surface, as documented
by SEM (see also Fig. 5c), degradation will take place
even at larger depths. On the other hand, the maximum
increase in density observed for HDPE was around 1%
under both aging conditions, as depicted in Fig. 1.
3.2. Hardness measurements
In Fig. 2 the results from hardness measurements are
shown. The aged samples had higher values of hardness
than found for the untreated material, and classication
is always unambiguous through the relation of
measurement results: HDPE > LLDPE > LDPE. These
changes in hardness suggest that the aging process leads
to a closer packing of the material in the macroscopic
scale, resulting in harder and denser products. In the
last stages of aging investigated here, this packing can
be attributed to modications in the amorphous regions
which are more labile to oxidation with subsequent
chain scission. The result of this are shorter molecules
with an increased mobility so that secondary crystallization processes can take place. The polar groups
introduced by oxidation can also interact through

Fig. 2. Variation in hardness of dierent polyethylene grades versus

accelerated aging duration under WOM- (a) and QUV-weathering
conditions (b), respectively.

intermolecular dipolar forces or hydrogen bonds, and

thus can undergo further crystallization processes as
well [4,9,44,68,69].
Apart from crystallization, cross-linking reactions
also lead to increases in hardness. Altogether, these
processes result in a nearly linear increase with time,
except for QUV aged LDPE during 800 h. The hardness
was then signicantly decreased, and this could be
explained in terms of micro-ssures created simultaneously. They would lead to a brittle, mechanically
weaker material. The processes involved in cracking are
now under investigation using nanomechanical testing,
and the results will be published soon [70].
3.3. DSC measurements
The original materials showed melting endothermic
peaks at 113, 126 and 132  C for LDPE, LLDPE and
HDPE, respectively. In Fig. 3 the calorimetric curves for
the melting region of LDPE are presented for new and
WOM-aged materials. New thermal energy absorption
peaks (shoulders) and a broadening of the endotherm
graph were observed due to degradation. These results
can be attributed to changes in crystallite sizes, molecular weight dierences, that are brought about by
chain breaking and secondary recrystallization . These
shoulders have been noticed before for aged polyethylenes, such as LDPE [4,9,24,44], MDPE (medium
density polyethylene) [69] and HDPE [27,68].
Apart from dierences in crystallinity, DSC results
allowed the evaluation of the degradation temperature
of the materials, when taken as the rst derivative of the
heat ux curves in the region of temperatures higher than
160  C. The point of inection in this region corresponds
to the temperature where degradation begins. It is
important to point out that these tests dier from the well
known OIT measurements, which is performed isothermally. In our case the oxidation temperature was
determined through dynamic scans of temperature. Also
in this case each sample behaved dierently: their thermal

Fig. 3. DSC thermograms for non-degraded and dierently aged

LDPE material.

J.V. Gulmine et al. / Polymer Degradation and Stability 79 (2003) 385397

stability was found in the order: HDPE >

LLDPE > LDPE. Tertiary carbon radicals are more
stable than secondary and primary ones, thus requiring
less activation energy to be formed. It is expected then
that a larger number or ramication sites would lead to
a larger number of radicals, resulting in less stable
materials. The formation of tertiary radicals is stimulated during the DSC heating scan at a larger rate in
LDPE, resulting in lower thermal resistance when compared with LLDPE and HDPE. Fig. 4a and b show that
the oxidation temperature decreases rapidly at the rst
stage of degradation (from 0 to 100 h under WOM- and
from 0 to 24 h under QUV-conditions). For longer
exposure times, the oxidation temperature remains
almost unchanged, and the three polyethylene grades
show the same behavior in practice. After the initial
stages, the samples suered from aging eects such as
chain scissions and cross-linking, with similar results as
observed in the DSC scans. As a consequence of this, it
can be stated that the oxidation temperature as determined in DSC analysis is not a good parameter to
characterize the long-term aging in polyethylene samples, and other methods ought to be explored [24,27,69].

389

In Fig. 5b the surface features of a LDPE sample

submitted to 1600 h WOM exposure are presented.
Cracks are seen all over the surface, showing a preferential propagation of micro-cracks with further
ramications in other directions. This eect can be
related to the specic aging conditions, but also to the
processing conditions such as the thermo-pressing during sample preparation. Sample cutting for microscopyunder the application of mechanical stressled
to the formation of larger cracks, perpendicular to those
due to aging. The QUV exposure, when applied for a
period of 800 h, had a more dramatic eect as seen in
Fig. 5c. The material is highly degraded, the cracks form

3.4. Results from SEM

The photomicrograph of Fig. 5a shows that the surface of new and non-degraded material is smooth,
without cracks, holes and free of any kind of defects.

Fig. 4. Variation of the temperature of oxidation under WOM conditions (a) and under QUV conditions (b) (data were taken from the rst
derivative of the DSC curves from above 160  C).

Fig. 5. SEM micrograph of a non-degraded LDPE sample surface (a),

for a LDPE sample surface aged under WOM conditions for 1600 h
(b), and for a LDPE sample surface aged under QUV conditions for
800 h (c).

390

J.V. Gulmine et al. / Polymer Degradation and Stability 79 (2003) 385397

a kind of mosaic pattern with no preferential propagation tendency. The importance of the wavelength of the
irradiation is again evident.
3.5. FTIR-spectroscopy
3.5.1. Attenuated total reectance (ATR) spectroscopy
ATR-spectroscopy is a useful technique to study of
samples showing large absorptivities in the mid-infrared
spectral range. Due to the low penetration depth of the
radiation experienced with this technique mainly surface
degradation can be studied. Another and more unique
application of ATR-spectroscopy is optical depth proling by which a component concentration or gradient
in structural changes can be probed near the bulk
material surface. Only a few spectroscopic techniques
allow the molecular structure to be probed as a function
of thickness on the order of micrometers. There have
been several approaches to obtain information on the
depth prole. A review on several approaches used in
the past has been given by Urban [71].
The rst investigations on surface depth proling
were performed using the well-established reection
theory for weakly absorbing homogeneous samples.
Penetration depth is varied by changing the reection
angle of incidence, polarization of the infrared radiation
and/or the refractive index by employing a dierent
crystal material. For this, the eective thickness
approach had been proposed, making use of the ATRspectra as if they were recorded in transmittance. However, for non-homogeneous materials the quantitative
relations must be modied.
By taking the inhomogeneity of the sample composition into account, more rened approaches were published. One example is the analysis of reectance spectra
recorded from polypropylene surfaces oxidized by O2
corona discharge, reported by Carlson and Wiles [72],
who used the eective thickness equation for thin lms
[73]. Hirschfeld proposed that the depth prole could be
obtained by applying a Laplace transformation to the
absorptance spectrum [74]. Several scientists advanced
the latter algorithm [7577]. However, as Fina pointed
out, the latter approaches were valid only for vibrational bands with small absorptivities [78]. Huang and
Marek [71,79] used exact optical theory in combination
with an iterative calculation. However, for all cases a
guessed prole is necessary for starting the evaluation.
Recently, two strategies were presented by Ishida and
coworkers. One method needs various ATR-spectra
recorded at dierent reection angles [80], whereas in
the other the intensities at various frequencies in a given
spectrum recorded with one single angle are taken to
calculate the depth prole in three calculation steps [81].
The sample is considered to be a anisotropic medium
consisting of isotropic stratied layers, and the calculating steps include estimation of the volume-fraction

prole by linear equations of absorbance, linear leastsquares tting of absorbance and non-linear tting of
reectance. The rst method (multiple angle approach)
utilizes an extinction coecent prole as a tting parameter [80], whereas the second (multiple frequency
approach) utilizes the volume-fraction prole as a tting
parameter [81].
In this context, a recent paper was published on the
accurate estimation of the mean square electric eld in
infrared ATR-spectroscopy also under strongly absorbing conditions, which is a necessary requisite for quantitative analysis without the a priori knowledge of the
optical constants of the material studied [82]. All these
approaches need rather sophisticated mathematics, so
that the simple strategy as outlined by Sack et al. [21]
was applied in this investigation.
3.5.2. Evolution of degradation with time
In Fig. 6 the main dierences observed between the
spectra of the original and aged LDPE materials are
shown for samples treated under WOM and QUV conditions. The dierences among the infrared spectra of
the various commercial polyethylene grades was
recently published by us allowing the distinction of each
type of polymer [83]. In the ATR spectrum of the nondegraded LDPE sample (Fig. 6a), were observed bands
in the interval between 3400 and 3200 cm1 and above
1600 cm1, which were attributed to the absorption of
the stabilizers: the bands at 3400 and 3200 cm1 corresponded to the phenolic OH group of Irganox and the
band near 1600 cm1 to the CC to the aromatic ring
of both additives (Irganox and Weston 399).With aging
the bands in the 34003200 cm1 region become broader
and coalesce, and the band at 1600 cm1 gradually disappears. One reason for the increased concentration of
antioxidant at the surface is its migration during the
thermo-pressing sample preparation. Transmission
spectra of thin PE foils prepared from the same pellet
material cannot provide these spectral features, because
the technique is not surface-sensitive enough. Traces of
the stabilizer compound are still observable in the spectra
of the other two samples aged under dierent conditions
(Fig. 6b and c). After extended weathering exposure
these absorptions are no longer noticeable.
Another aspect that deserves to be commented on is
related to the bands in the region of 11001300 cm1,
that are more pronounced for the samples aged in the
WOM chamber. They result from deposition of salts on
the samples surface, originating from the sprayed water
simulating rain inside of the chamber. This assumption
was conrmed by EDS analyses, discussed below.
The absorption band around 1714 cm1, which can be
assigned to the CO stretching vibration of a ketone
group, grows in intensity with extended aging, and at
the same time new bands begin to appear. These
absorption bands indicate that more than one oxidation

J.V. Gulmine et al. / Polymer Degradation and Stability 79 (2003) 385397

391

Fig. 6. FTIRATR spectrum of non-degraded LDPE material (a), of aged LDPE material under QUV conditions for 200 h (b), and under WOM
conditions for 400 h (c) (measurement conditions: ZnSe crystal, reection angle of incidence: 45 ; the latter two spectra were oset for clarity).

product is formed, as detailed in Fig. 7. These new

bands show up after 100 h of exposure in the WOMequipment and after 24 h under QUV-weathering conditions. The irradiance dierences of the two chambers
are reected in faster aging in QUV conditions, with
carbonyl splitting beginning with only 24 h exposition.
The carbonyl bands were assigned to CO stretching
vibrations in aldehydes and/or esters (1733 cm1) carboxylic acid groups (1700 cm1) and g-lactones (1780

cm1) [9,23,44,68].The latter are seen after 200 h exposure in both aging regimes.
The carbonyl index, dened as the ratio between the
integrated band absorbance of the carbonyl around
1714 cm1 and that of the PE-polymer bands (1470
cm1), characterizes the degree of oxidation for each
polyethylene sample. The results for the samples after
dierent weathering exposure are shown in Fig. 8. These
curves were calculated from ATR spectra, using a ZnSe

Fig. 7. FTIRATR spectrum of LDPE aged during 400 h under WOM conditions showing the formation of various bands that can be assigned to
stretching vibrations of various carbonyl groups formed by the polymer oxidation processes.

392

J.V. Gulmine et al. / Polymer Degradation and Stability 79 (2003) 385397

that have already been mentioned, other relevant ones

are: the wavelength of the radiation, the eciency of the
contact between sample and crystal, and the sample
area in contact with the crystal (the sample can either
fully or only partially cover the crystal) [84]. These
parameters are correlated through Eqs. (1) to (4):
dp

l
2n1 sin2 eff  n221

where dp
l
n1
n21=n2/n1
e

Fig. 8. Oxidation rates for the three polyethylene grades investigated

and characterized by the variation of the carbonyl index (for its denition, see text) versus accelerated aging duration under WOM conditions (a) and under QUV conditions (b); ATR-measurement
conditions: ZnSe crystal, reection angle of 45 .

3.5.3. ATR-measurements for the analysis of the

degradation depth prole
By changing the angle of incidence and the ATR
crystal, the depth of the infrared radiation reaching into
the sample can be varied. Apart from the parameters

penetration depth, in mm
wavelength, in mm
refractive index of the ATR crystal
ratio of the refractive indexes of the
sample (n2) and of the crystal (n1)
eective angle of reection incidence,
in degrees

The eective angle of incidence is calculated by using

Eq. (2):


1 sinIRA  IRE
eff IRA  sin
2
n1
where

IRA
IRE

crystal with a reection angle of 45 to provide the same

depth probing. The three polymers behaved dierently:
the LDPE material was most sensitive to oxidative
aging compared to the other two polymer grades. The
stability against oxidation was in the following order:
HDPE > LLDPE > LDPE. The oxidation rates at short
exposure to aging conditions within the WOM-equipment are generally larger compared to the eects
observed for half exposure times under QUV-conditions. It is noteworthy that the low density polyethylene
grade (LLDPE) and the HDPE-samples submitted to
WOM exposure did not show such an accentuated
increase in the carbonyl index at the last stages of aging,
specially when compared to those observed for samples
under QUV weathering conditions. This result could be
related to the competition between oxidation and crosslinking processes, and it seems to indicate that under
WOM exposure cross-linking is favored over oxidation.
These considerations are merely speculative, since no
crosslinking density measurements were carried out.

1=2

angle measured in the accessorys scale,

in degrees
angle of the crystals face, in degrees

The eective path length of the beam (EPL) is inuenced by the refractive indexes of the sample and of the
ATR crystal, the wavelength and the number of reections (N) inside the crystal. The relationships among
these variables are:
N l=tcoteff

and
EPL Ndp

where l and t are the length and thickness of the ATR

crystal, respectively. The crystal length in particular is
dened as the distance between the centers of the
entrance and exit apertures.
Using Eqs. (1)(4), the parameters listed in Table 1
were calculated. The values of the penetration depth
(dp) and of the eective angle (e) were needed to analyze the spectra of the aged polyethylene samples for
dierent depths. By using the carbonyl index (i.e., normalized absorbance values) instead of the absorbance
band area, the inuence of the number of internal
reections, changes in the contact between sample and
crystal, and the magnitude of the electrical eld of the
infrared radiation at the interface sample/ATR plate
can be eliminated.

J.V. Gulmine et al. / Polymer Degradation and Stability 79 (2003) 385397

Table 1
Parameters used for the calculation of the degradation proles in aged
polyethylene (data are given for the wavenumber of carbonyl stretching vibrational band at 1714 cm1.
l
IRA e
Crystal ~
material (cm1) (cm)

n1

Ge

4.0 1.5 0.48

0.39
0.33

ZnSe

1714

5.83

n2

dp
N
(mm)

35
45
55

37.5
45.0
52.5

45
50
55
60

45.0 2.4 1.5 1.17

47.9
0.97
50.85
0.84
53.8
0.76

EPL
(mm)

21.7 10.5
16.7 6.5
12.8 4.25
16.7 19.5
15.0 14.6
13.6 11.4
12.2 9.2

IRA is the angle measured in the ATR accessory, e the eective

angle of incidence, n1 and n2 are the refractive indices of crystal and
sample, respectively; dp is the penetration depth of the evanescent eld
of the electromagnetic infrared radiation; N is the number of reections inside the ATR crystal under the given experimental conditions
and EPL is the eective pathlength of the beam inside the sample
medium)

By varying the angle of incidence, the depth reached by

the radiation is also varied, as already pointed out. But
when a larger depth is reached in an appropriate ATR
experiment, all the contributions from the preceding
layers are summed up, and to separate the contribution
of a chosen single layer, it is necessary to take into
account the absorptions from the layers above. By using
a relatively simple mathematical treatment, Sack et al.
[21] developed an equation that made it possible to
normalize the carbonyl absorption through the use of
the carbonyl index, and also to obtain the concentration
related to only one particular layer. The schematics of
this process is illustrated in Fig. 9. Several assumptions
were made that the refractive index n2 of the sample is
constant for all samples investigated and the PE concentration can also be approximated to be constant,
that is independent of sample depth and state of oxidation. Furthermore, an exponential function is used
for the description of the concentration prole of the
oxidation products.
The average carbonyl group concentration relative to
each layer can be obtained from Eq. (5):


 


Ai1c
dpi1
Aic
dpi

Ai

5
Aip dpi  dpi1
Ai1p dpi  dpi1
for i=2,3,... and where Aic and Aip stands for the
absorbance of the carbonyl groups and the polymer
vibrational bands at depth di, respectively. The index
(i1) refers to the preceding adjacent layer. For i=1,
the formula breaks down to A 1 =A1c/A1p.
The calculations as described resulted in the data
which are plotted in Fig. 10, showing the degradation
prole of the samples within the rst micrometer layer

393

(the mean depths within each individual layer, as accessed

by dierence ATR spectroscopy, were taken into account
for calculating the abscissa values). The dierence in the
oxidation rate, obtained by the two weathering methods,
becomes evident from the two diagrams that show the
evolution of the degradation from the surface to the
bulk of the materials with time. At the beginning, the
carbonyl content at layers close to the surface is evidently higher, but with time the carbonyl content in
deeper layers increases steadily, diminishing also the
dierence in oxidation with depth. In samples that were
severely deteriorated, like the ones submitted to 1600 h
under WOM conditions and to 800 h under QUV
weathering (not shown in Fig. 10, because out of scale)
a dierence in the oxidation product concentrations is
no longer noticeable. A plateau is reached within the
possibilities of the ATR technique as applied in this
investigation, that is for a depth down to 1.2 mm.
To avoid this limitation and to get access to deeper
layers, microtomy was used. Films were cut consecutively at a depth up to 10 mm to analyze the degradation at deeper layers of a sample that was submitted
to WOM conditions for 400 h. The microlms were
analyzed by transmission measurements, and the results
were compared to those obtained by the ATR technique. The results are shown in Fig. 11. A satisfactory
agreement was found between the two methods, since
an exponential tting of the points obtained by ATR
and microtomy could be done. This demonstrates that
FTIRATR spectroscopy is a specially suitable tool for
depth prole surface analysis, with the additional
advantage of not requiring any previous treatment of
the sample, and that it can be used in conjunction with
microtomy to reach a broad range of depths. The
exponential behavior of the oxidation can be explained
by the dependence of oxygen diusion inside the material [8588].
Another technique with the mapping of depth proles
of oxidation products within degraded PE samples along
a cross-section cut using simple ATR-microscopic techniques, based on silverhalide ber probes, will be reported elsewhere [89]. Such a technique is also applicable for
surface micro-domain analysis of uneven samples.
3.5.4. EDSenergy dispersive spectroscopy
The EDS surface mapping of one representative sample (WOM, 1600 h) is presented in Fig. 12. The presence
of sodium chloride (NaCl), silicon oxide (SiO2) and
aluminum oxide (AlO3) was detected. Silicon and aluminum oxides absorptions are seen in FTIRATR
spectra in the region of 11001300 cm1. This is due to
the incomplete purication of the sprayed water by the
ion exchange resin of the equipment, leaving a certain
amount of salts in the water. The peak seen around 2.1
keV is originated by the gold utilized to cover the surface of the sample.

394

J.V. Gulmine et al. / Polymer Degradation and Stability 79 (2003) 385397

Fig. 9. Schematic representation of the dierent depths reached by the IR-radiation using dierent ATR measurement conditions by changing, e.g.,
the reection angle: dp1, dp2 and dp3 are the calculated depths dening the thickness of the layers 1, 2, and 3 from the surface. The values of A1, A2
and A3 are the absorbance values corresponding to each penetration depth. The values of A 1 , A 2 and A 3 are the average carbonyl indexes for the
respective layers, as obtained by subtracting the contribution of the preceding layers.

Fig. 10. Depth proles obtained for the oxidation products within
LDPE under WOM conditions (a) and QUV weathering (b), which
were based on the spectral evaluation of the normalized band area
around 1714 cm1 using dierent angles of reection and the ATR
crystal material of ZnSe and Ge (see also Table 1).

Fig. 11. Depth prole obtained for the oxidation products within
LDPE under WOM conditions for 400 h, as analyzed by ATR spectroscopy and microtomy based on the evaluation of the normalized
band area around 1714 cm1.

J.V. Gulmine et al. / Polymer Degradation and Stability 79 (2003) 385397

395

Fig. 12. EDS surface spectrum of aged LDPE (WOM, 1600 h). Magnication of 200
. The peak seen around 2.1 keV is originated by the gold
utilized to cover the surface of the sample.

4. Conclusions
The accelerated aging within the three most important, commercially available polyethylene grades was
systematically studied using articial weathering conditions as made possible with WOM and QUV equipment. QUV conditions were more severe than those of
the WOM unit, since the irradiance in the latter is twofold more intense than in the rst mentioned. The stability against aging observed for the dierent
polyethylene materials was found the following order:
HDPE > LLDPE > LDPE.
Increases in density and hardness with aging indicated
an increase in the polymers crystallinity and/or the
occurrence of cross-linking reactions. DSC scans
showed a broadening of the melting endotherm peak
and the appearance of new peaks for PE-samples
degraded by accelerated aging, which were attributed to
changes in crystallite sizes, molecular weight dierences,
that are brought about by chain breaking, and secondary recrystallization.
At late stages of aging the polymer material
became brittle and fragile. The surface examination
by SEM revealed that for samples under WOM
exposure the cracks seemed to have a preferential
propagation, whereas for QUV the damage was
much more severe with cracks propagating in all
directions.

The observed morphological changes, induced by the

dierent accelerated aging methods, were synergetically
related to the chemical modications observed by the
FTIRATR technique. The vibrations from mainly
occurring carbonyl groups were assigned in the spectra.
The variation of the ATR reection angle allowed the
characterization of isolated layers from the surface
down to 1.2 mm, and the results agreed well with those
obtained by the spectroscopic analysis of microtomed
samples, as observed in the exponential tting of the
combined results of both techniques.

Acknowledgements
We wish to thank COPEL (Companhia Paranaense de
Energia ) for nancial support, Dr. Gabriel Pinto de
Souza and M.Sc. Marilda Munaro for the DSC-measurements, Dr. Maurcio Pereira Cantao for the SEM- and
Dr. Carlos Mario Garcia for the XRD-measurements.
JVG thanks LACTEC for a doctoral scholarship and
LA thanks CNPq for a research grant. HMH acknowledges the support given by a CNPq travel grant. The
author from ISAS (HMH) also acknowledges gratefully
the nancial support given by the Ministerium fur
Schule, Wissenschaft und Forschung des Landes
Nordrhein-Westfalen and by the Bundesministerium fur
Bildung und Forschung, Germany.

396

J.V. Gulmine et al. / Polymer Degradation and Stability 79 (2003) 385397

References
[1] Nalwa HS. Handbook of low and high dielectric constant materials and their applicationsmaterials and processing, vol. 1. San
Diego: Academic Press; 1999.
[2] Bartnikas R, Eichhorn RM. Engineering dielectrics, vol. IIA.
Philadelphia: ASTM Publications; 1983.
[3] Carrasco F, Pages P, Saurina J, Colom X. Anidad LII 1995;
458:2317.
[4] Hamid SH, Amin MB. J Appl Polym Sci 1995;55:138594.
[5] Severini F, Gallo R, Ipsale S, Nisoli E, Pardi M. Polym Degrad
Stab 1999;65:32935.
[6] Gijsman P, Hennekens J, Janssen K. Am Chem Soc 1996:622
35.
[7] Scoponi M, Cimmino S, Kaci M. Polymer 2000;41:796980.
[8] Tidjani A, Arnaud R. Polymer 1995;36(14):28414.
[9] Khabbaz F, Albertsson AC, Karlsson S. Polym Degrad Stab
1998;61:229342.
[10] Pickett JE. Macromol Symp 1997;115:12741.
[11] Horrocks AR, Mwila J, Miraftab M, Liu M, Chohan SS. Polym
Degrad Stab 1999;65:2536.
[12] Karlsson S, Hakkarainen M, Albertsson AC. Macromolecules
1997;30:77218.
[13] Bauer DR. Polym Degrad Stab 1995;48:25967.
[14] Philippart JL, Posada F, Gardette JL. Polym Degrad Stab 1995;
49:28590.
[15] Gijsman P, Dozeman A. Polym Degrad Stab 1996;53:4550.
[16] Gijsman P, Sampers J. Polym Degrad Stab 1997;58:559.
[17] Andrei C, Andrei G, Moraru M, Moise G. Polym Degrad Stab
1999;64:16571.
[18] Tidjani A. Polym Degrad Stab 2000;68:4659.
[19] Lau WWY, Pan JQ. Polym Degrad Stab 1998;60:45964.
[20] Wolter KD, Tanaka J, Johnson JF. IEEE Trans Electr Insul
1982;17(3):24852.
[21] Sack S, Wagner H, Steger E. Acta Polym 1985;36:30510.
[22] Han SO, Lee DW, Han OH. Polym Degrad Stab 1999;63:23743.
[23] Setnescu R, Jipa S, Osawa Z. Polym Degrad Stab 1998;60:377
83.
[24] Budrugeac P, Segal E. J Therm Anal 1998;53:8018.
[25] Liauw CM, Childs A, Allen NS, Edge M, Franklin KR, Collopy
DG. Polym Degrad Stab 1999;65:20715.
[26] Allen NS, Edge M, Holdsworth D, Rahman A, Catalina F,
Fontan E, Escalona AM, Sibon FF. Polym Degrad Stab 2000;
67:5767.
[27] Dorner G, Lang RW. Polym Degrad Stab 1998;62:43140.
[28] Liu M, Horrocks AR, Hall ME. Polym Degrad Stab 1995;
49:15161.
[29] Audouin L, Langlois V, Verdu J. Die Angew Makromol Chem
1995;232:112.
[30] Celina M, Gillen KT, Clough RL. Polym Degrad Stab 1998;
61:23144.
[31] Ortiz RA, Salas EAR, Allen NS. Polym Degrad Stab 1998;
60:195204.
[32] Boubakeur A, Mecheri Y, Boumerzou M. Ann Chim Sci Mater
2000;25:45770.
[33] Corrales T, Catalina F, Peindado C, Allen NS, Fontan E. J
Photochem Photobiol A: Chem 2001;5319:112.
[34] Allen NS, Katami H. Polym Degrad Stab 1996;52:31120.
[35] Sharma N, Chang LP, Chu YL, Ismail H, Ishiaku US, Ishak
ZAM. Polym Degrad Stab 2001;71:38193.
[36] Goldberg VM, Kolesnikova NN, Paverman NG, Kavun SM,
Stott PE, Gelbin ME. Polym Degrad Stab 2001;74:37185.
[37] Costa L, Luda MP, Trossarelli L. Polym Degrad Stab 1997;
58:4154.
[38] Gugumus F. Polym Degrad Stab 2000;68:33752.
[39] Gugumus F. Polym Degrad Stab 2002;76(2):32940.
[40] Gugumus F. Polym Degrad Stab 2002;76(3):95110.

[41]
[42]
[43]
[44]
[45]
[46]
[47]
[48]

[49]
[50]
[51]
[52]
[53]
[54]
[55]
[56]
[57]

[58]
[59]

[60]
[61]
[62]

[63]

[64]

[65]

[66]
[67]
[68]
[69]
[70]
[71]

[72]
[73]
[74]
[75]

Gugumus F. Polym Degrad Stab 2002;76(3):38191.

Gugumus F. Polym Degrad Stab 2002;75:5571.
Gugumus F. Polym Degrad Stab 2002;75:13142.
Khabbaz F, Albertssson AC, Karlsson S. Polym Degrad Stab
1999;63:12738.
Lagaron JM, Pastor JM, Kip BJ. Polymer 1999;40:162936.
Monney L, Belali R, Vebrel J, Dubois C, Chambaudet A. Polym
Degrad Stab 1998;62:3539.
Gupta G, Chiang CP. ACS Div Polym Mater 1990:66771.
Das-Gupta DK, Cooper DE. Conference Record 1988 IEEE
Int Symp Electr Insulation, Boston, 58 June 1988. p. 335
338.
Delor F, Barrois-Oudin N, Duteurtre X, Cardinet C, Lemaire J,
Lacoste J. Polym Degrad Stab 1998;62:395401.
Gorur RS, Bernstein BS. IEEE Trans Power Deliv 1998;13:316
21.
Gorur RS, Bernstein BS. IEEE Trans Power Deliv 1998;13:323
9.
Birtwhistle D, Blackmore P, Krivda A, Cash G, George G. IEEE
Trans Dielectr Electr Insul 1999;6:6129.
Shimizu N, Laurent C. IEEE Trans Dielectr Electr Insul 1998;
5:6519.
Costa L, Jacobson K, Brunella V, Bracco P. Polym Testing 2001;
20:64959.
Uhniat M, Kudla S. Polym Degrad Stab 2001;71:7582.
Haider N, Karlsson S. Polym Degrad Stab 1999;64:3218.
Kurtz SM, Muratoglu OK, Buchanan F, Currier B, Gsell R,
Greer K, Gualtieri G, Johnson R, Schaner S, Sevo K, Spiegelberg S, Shen FW, Yau SS. Biomaterials 2001;22:17317.
Wu Q, Qu B, Xu Y, Wu Q. Polym Degrad Stab 2000;68:97102.
McLaughlin WL, Silverman J, Al-Sheikhly M, Chappas WJ,
Zhan-Jun L, Miller A, Batsberg-Pedersen W. Radiat Phys Chem
1999;56:5038.
Ciuprina F, Teisse`dre G, Filippini JC. Polymer 2001;42:78416.
Lee AW, Santerre JP, Boynton E. Biomaterials 2000;21:85161.
Standard practice for operating light-exposure apparatus (xenonarc type) with and without water for exposure of nonmetallic
materials, ASTM G26-95. Annual book of ASTM standards.
Philadelphia (PA): American Society for Testing and Materials;
1995.
Standard practice for operating light- and water-exposure apparatus (uorescent UV-condensation type) for exposure of nonmetallic materials, ASTM G53-96. Annual book of ASTM
standards. Philadelphia (PA): American Society for Testing and
Materials; 1996.
Standard test methods for density and specic gravity (relative
density) of plastics by displacement, ASTM D792. Annual book
of ASTM standards. Philadelphia (PA): American Society for
Testing and Materials; 1991.
Standard test method for rubber propertydurometer hardness, ASTM D2240. Annual book of ASTM standards. Philadelphia (PA): American Society for Testing and Materials;
1991.
Kockott D. Polym Degrad Stab 1989;25:181208.
Andrady AL. Adv Polym Sci 1997;128:4794.
Valadez-Gonzalez A, Cervantes-Uc JM, Veleva L. Polym Degrad
Stab 1999;63:25360.
Dorner G, Lang RW. Polym Degrad Stab 1998;62:42130.
Tavares AC, Gulmine JV, Leni A, Lepienski CM. Polym Degrad
Stab (submitted for publication).
Urban MW. Attenuated total reectance spectroscopy of polymerstheory and practice. Washington: American Chemical
Society; 1996 [Chapter 7].
Carlsson DJ, Wiles DM. Can J Chem 1970;48:2397406.
Harrick NJ, Du Pre FK. Appl Optics 1966;5:173943.
Hirschfeld T. Appl Spectrosc 1977;31:28992.
Fina LJ, Chen G. Vibr Spectrosc 1991;1:35361.

J.V. Gulmine et al. / Polymer Degradation and Stability 79 (2003) 385397

[76] Shick R, Koenig JL, Ishida H. Appl Spectrosc 1993;47:1237
44.
[77] Shick R, Koenig JL, Ishida H. Appl Spectrosc 1996;50:10828.
[78] Fina LJ. Appl Spectrosc Rev 1994;29:30965.
[79] Huang J, Urban MW. Appl Spectrosc 1993;47:97381.
[80] Ekgasit S, Ishida H. Appl Spectrosc 1997;51:4615.
[81] Ekgasit S, Ishida H. Appl Spectrosc 1997;51:148895.
[82] Ekgasit S. Appl Spectrosc 1998;52:7736.
[83] Gulmine JV, Janissek PR, Heise HM, Akcelrud L. Polym Testing
2001;21:55763.

397

[84] Standard practices for internal reection spectroscopy, ASTM E

573-96. Annual book of ASTM standards. West Conshohocken
(PA): American Society for Testing and Materials, 1996.
[85] Gugumus F. Polym Degrad Stab 1996;52:13144.
[86] Gugumus F. Polym Degrad Stab 1996;52:14557.
[87] Gijsman P, Meijers G, Vitarelli G. Polym Degrad Stab 1999;
65:43341.
[88] Gugumus F. Polym Degrad Stab 1997;55:2143.
[89] Gulmine JV, Janissek PR, Kupper L, Heise HM. Appl Spectrosc
[in preparation].