Polymer Degradation and Stability 81 (2003) 367–373

www.elsevier.com/locate/polydegstab

The effect of accelerated aging on the surface mechanical

properties of polyethylene
Andreia C. Tavaresa, Joseane V. Gulmineb, Carlos M. Lepienskia, Leni Akcelrudc,*
a
Departamento de Fı´sica, Universidade Federal do Paraná UFPR, Centro Politécnico da UFPR, Caixa Postal 19081, CEP 81531-990, Curitiba/PR, Brazil
b
Instituto de Tecnologia para o Desenvolvimento LACTEC, Caixa Postal 19067, CEP 81531-990 Curitiba/PR, Brazil
c
Departamento de Quı´mica, Universidade Federal do Paraná UFPR, Centro Politécnico da UFPR, Caixa Postal 19081, CEP 81531-990, Curitiba/PR, Brazil

Received 30 October 2002; received in revised form 4 March 2003; accepted 18 March 2003

Abstract
The surface mechanical properties of polyethylene subjected to accelerated aging in Weathering Tester (QUV) and weathe-
rometer (WOM) chambers were determined by a nanoindentation technique. It was observed that the apparent surface hardness
was three times greater than that of the bulk after 1600 h of exposure in the WOM chamber, and a similar increase was also
observed after 800 h of exposure in the QUV chamber. The elastic modulus at tip penetration depths of 500 nm also shows a sig-
nificant increase from 400 to 2000 MPa for samples aged in the QUV chamber during 800 h. For aging times of 200 h and lower the
variations in surface mechanical properties are small and restricted to a thin layer, with thickness lower than 1 mm for samples aged
in both WOM or QUV chambers. The modifications in nanomechanical properties were correlated with the corresponding chemical
processes due to aging, measured through the carbonyl index profile obtained from ATR–FTIR. An interpretation for the surface
cracking of aged LDPE based on differential increases in elastic modulus at different depths is presented.
# 2003 Elsevier Science Ltd. All rights reserved.
Keywords: Nanoindentation; Polyethylene; Mechanical properties; Degradation

1. Introduction such as radiation (mainly UV), temperature variations,

Polymers are sensitive to several types of radiation
such as ultraviolet and gamma rays, electron and neu-
tron beams, ions, and others [1–17]. These external fac-
tors normally affect the surface at higher levels than the
bulk of the polymer, but nevertheless can have a great
influence on the overall performance of these materials,
depending on their application. Aging processes due to
exposure to outdoor conditions is one of the most sig-
nificant processes leading to surface modification, and
has its practical implications.
Polyethylene is widely used as an insulator in high
voltage cables due to its dielectrical properties and low
cost. In a series of applications high voltage cables are
exposed to a combination of aggressive ambient factors
* Corresponding author. Tel.: +55-41-323-8716; fax: +55-41-266-
3582.
E-mail addresses: andreia@fisica.ufpr.br (A.C. Tavares),
joseane@lactec.org.br (J.V. Gulmine), akcel@onda.com.br
(L. Akcelrud).
0141-3910/03/$ - see front matter # 2003 Elsevier Science Ltd. All rights reserved.
doi:10.1016/S0141-3910(03)00108-3
in dry and wet saline environments. The overall result is
the concomitant occurrence of oxidation reactions, cross-
linking and chain degradation, with the development of
internal stresses, and surface cracking. This in turn will
allow water and oxygen from air to penetrate deeper into
the sample and the effect of aging is enhanced [18,19].
A proper evaluation of these aging consequences
would have to assess the variation of the chemical and
morphological modifications with the correspondent
physical properties (mainly mechanical and electrical),
in such a way that a characterization profile from sur-
face to bulk could be outlined. Conventional mechan-
ical tests used in polymers, such as tensile elongation,
describe the combined effects that take place in the
material’s surface and bulk, only very degraded samples
can give significant indication of the extent of degrada-
tion with these methods.
The study of the aging processes in polyethylene is of
great importance for the development of more resistant
materials than available in the past, and to forecast the
limit of their use. Apart from thermal and electrical
368 A.C. Tavares et al. / Polymer Degradation and Stability 81 (2003) 367–373
stresses, polyethylene insulation in high voltage cables is
subjected to mechanical stresses. The presence of surface
cracks in aged polyethylene is a direct evidence of the
development of surface stresses during the aging process.
The effect of the internal mechanical stresses on the elec-
trical performance of polyethylene has been investigated
[20].
Considerable attention has been given to the quanti-
tative evaluation of these aging effects [21–37]. Calori-
metric, spectroscopic and microscopic techniques,
mostly in combination with each other, are methods
usually used to analyze the aging effect on polyethylene.
The most important modification is brought about by
polymer oxidation, and this effect has been recently
addressed by FTIR-spectroscopy to characterize and
quantify the oxidation products, using the attenuated
total reflection (ATR) technique for surface characteriza-
tion [21–28] or transmission spectroscopy for specimens in
the form of microtomized or pressed films [38–45].
Recent work by Suarez [13] has correlated the
mechanical properties of low density polyethylene
(LDPE) with the dose of gamma irradiation. It was
found that the Rockwell hardness increased by a factor
of 40% for lower doses but decreased when the dose
was increased. The results were interpreted in terms of
different contributions of crosslinking at lower doses
and scission of chains at higher doses with concomitant
carbonyl formation.
One way to characterize the degradation profile is to
microtomize the sample and thus access different
depths. Another approach is a non-invasive and non-
destructive method that makes use of the full potential
of the ATR–FTIR technique, using different crystalline
materials and different reflection angles for the infrared
beam at the sample-crystal interface. Such a procedure
allows the analysis of layers up to different depths on a
micrometer scale. In a recent publication [46] we have
described the degradation profile of low density poly-
ethylene taking into account the chemical modifications
occurring from surface to bulk, when submitted to
strictly controlled aging conditions. Two methods were
used: microtomization to the different depths and the
ATR–FTIR method, which provides access to the dif-
ferent depths without physically altering the sample.
In this contribution, we complement the chemical
description of the degradation profile of aged LDPE
with the characterization of the corresponding mechan-
ical properties along the sample’s depth, in particular
the variation of hardness and elastic modulus. With the
nanoindentation technique it is possible to follow the
variation of hardness and elastic modulus with the
depth of indenter penetration in the aged surface. The
results are correlated with the corresponding values of
carbonyl index, in an attempt to provide a full descrip-
tion and interpretation of the aging processes taking
place in LDPE.
The nanoindentation technique uses a hard indenter
tip to penetrate into the sample surface by controlling
the displacement and the applied load during the
indentation process. It is a high resolution method and
has been used to evaluate the mechanical properties of
surfaces, coatings and thin films [47,48]. With this tech-
nique it is possible to determine the mechanical proper-
ties of films and modified layers as thin as 100-nm.
Nanoindentation is the most versatile and powerful
means to measure mechanical properties at nanoscale
resolution [49]. However the technique has not been
widely used to measure the mechanical properties of
polymers as in the case of metals and ceramics, due to
the difficulties associated with to the viscoelastic char-
acter of polymer deformation under stress. In polymers
the values of hardness and elastic modulus measured
from nanoindentation vary depending on the loading
and unloading rates used in the indentation test in a
similar way as in the standard tensile or compressive
tests [50]. Nevertheless, the nanoindentation technique
is a powerful tool for polymer surface characterization,
and can be successfully applied to compare samples of
the same origin submitted to different treatments or
applied stresses [51].
The nanoindentation test consists of applying a load
P to the material and measuring the depth of penetra-
tion h of the diamond tip. The hardness H can thus be
defined as:
P
H¼
A
where A represents the projected contact area of the
indentation which is related with the contact depth by
the expression
A ¼ 24:5h2c
where hc is the contact depth.
The unload curve behaves as a power law function
P ¼  ð h  hf Þ m
where hhf is the elastic displacement and  and m are
constants.
From the unloading curve the contact depth of the tip
into the material during the maximum load is obtained
from
Pmax
hc ¼ hmax  "
S
where hmax is the maximum displacement reached by the
indenter and Pmax is the maximum load applied, and the
peak indentation load respectively, S=dP/dh in the
initial part of unloading curve (contact stifness), and  is
 A.C. Tavares et al. / Polymer Degradation and Stability 81 (2003) 367–373
a parameter which depends only the tip shape and is
equal to 0.75 for triangular based pyramidal (Berko-
vich) tip normally used in nanoindentation.
The reduced elastic modulus Er is related to the mea-
sured values by the relation
pffiffiffi
 1
Er ¼ pffiffiffiffi
2S A
and
    humidity. The exposures were run for 25, 50, 100, 200,
1 1  2 1  2i
¼ þ
Er E Ei
where E and  are the Young’s modulus and the Pois-
son’s ratio for the specimen and Ei and i are the same
parameters for the indenter material.
An important issue that should be pointed out is that
in samples that present a gradient in hardness as a
function of depth, as in the case of surface hardening,
the measured values by nanoindentation do not repre-
sent absolute or quantitative values at a determined
depth. In order to achieve that depth a high load is
necessary to penetrate the harder layers near the sur-
face. Thus the measured hardness will be higher than
that of a homogeneous material with a constant hard-
ness value as a function of depth. The exact value of
hardness at some depth could be obtained if the hard
surface layer could be cut without altering the mechan-
ical properties and surface finishing. However, it is
impossible to cut very thin layers without affecting the
mechanical properties in the region of the new surface.
Cross-section measurements are not an alternative since
it is not possible to measure the hardness for distances
from surface lower than 5 mm.
Furthermore, the earlier equations were derived
assuming homogeneity (modulus and hardness inde-
pendent of location in the material). For the current
materials, aging leads to variations from surface to
bulk, as noted. The figures thus obtained cannot be
regarded as absolute but rather as semi-quantitative
values that describe the effects of accelerated aging and
provide a means for the interpretation of surface crack-
ing.
2. Experimental
Samples of low density polyethylene PB 681/59 were
obtained from OPP Polietilenos do Brasil S.A. Sheets
with dimensions of 11  11  0.1 cm were made from
pellets using aluminum molds and polyester foils in a
Schulz press, model PHS 15. The thermopressing was
made at 130  C at cycles of 5 min with no load, 5 min
with 2 ton and 5 min with 4 ton. The sheets were
allowed to cool at ambient temperature.
369
Two accelerated methods simulating weathering con-
ditions were employed. The first was performed in the
weatherometer (WOM) equipment from ATLAS, model
CI65 following ASTM G26/94 method, procedure 1
[52]. This equipment works with a xenon lamp of 6500
W, with an irradiance of 0.35 0.03 W/m2 (at 340 nm)
simulating the sunlight of Florida at noon, with cycles
of 102 minutes of incident beam at 90 at 63 5 C,
relative humidity of 60 5% and 18 minutes with deio-
nized water spraying at 50 5 C and 80 5% relative
400, 800 and 1600 h. The second aging methodology was
performed in the QUV weathering tester equipment
model QUV spray UV40, following the ASTM G65/96
method [53], using a fluorescent bulb UVB with 0.60 W/
m2 irradiance (at 313 nm), with 8 h cycles of UV irradia-
tion at 60 C and additional 4 h of deionized water spray-
ing at 50 C, with incident beam at 90 . The samples were
submitted to aging for 12, 24, 50, 100, 200, 400 and 800 h.
The hardness and the elastic modulus of LDPE were
obtained using a Nanoindenter XP from MTS Inc.,
using a pyramidal tip (Berkovich type) with triangular
base and apex angle of 65 . The hardness and the elastic
modulus were obtained by using the Oliver and Pharr
method [47], from the unloading curves. In each sample
were performed 16–25 indentations for eight different
loads, from 4 to 200 mN. To minimize the influence of
viscoelastic properties in the measurements the unloading
was performed at the same high rate in all measurements.
The nanoindenter has an optical microscope con-
nected to it, allowing the observer to choose carefully
the region to be indented.
The detailed spectrometric procedure for FTIR
determinations at various depths has been described
elsewhere [46,54]. The spectra were acquired using an
incident IR beam at 450 with a ZnSe crystal, which
reached a depth of 1.17 mm. from surface.
3. Results and discussion
The hardness and the elastic modulus of aged and
non-aged LDPE samples were measured by nanoinden-
tation technique and the Oliver and Pharr method. In
Fig. 1 the variation of elastic modulus is shown as a
function of penetration depth for samples aged at dif-
ferent times in the QUV chamber. For aging times lower
than 400 h the increase in elastic modulus was only
observed for penetrations less than 1 mm deep. With
increasing aging the elastic modulus increased corre-
spondingly, up to depths of 10 mm. This was the max-
imum depth for indentation with the maximum load
used in the tests. The hardness behavior followed the
same trend found for the elastic modulus, as shown in
Fig. 2: the major effects were detected for aging times of
400 h and higher.
370 A.C. Tavares et al. / Polymer Degradation and Stability 81 (2003) 367–373

Fig. 1. Variation of elastic modulus as a function of penetration depth

Fig. 3. Variation of elastic modulus as a function of penetration depth
for samples aged at different times in the QUV chamber.
for samples aged at different times in the WOM chamber.

The results for the elastic modulus and hardness rela-

tive to samples aged in the WOM chamber are shown in
Figs. 3 and 4, respectively. Higher variation in hardness
and elastic modulus are observed for depths lower than
1 mm. It can be observed that the increase in these
properties for different aging times is less pronounced
than those found for QUV aged samples. For aging
times of 400 h however, the values are lower than those
compared with samples aged in QUV chamber.
Figs. 5 and 6 compare the variation of elastic modulus
and hardness for penetrations depths of 500 nm at dif-
ferent aging times in WOM and QUV regimes. It can be
observed that the more pronounced effect of the QUV
over WOM conditions begins to operate above 200 h of
exposure. The curves for aging times of 12, 25 and 50 h
in all graphs are not shown because they are practically Fig. 4. Variation of hardness as a function of penetration depth for
identical to that of the original sample. samples aged at different times in the WOM chamber.

Fig. 2. Variation of hardness as a function of penetration depth for Fig. 5. Variation of elastic modulus for penetrations of 500 nm at
samples aged at different times in the QUV chamber. different aging times in WOM and QUV regimes.
 A.C. Tavares et al. / Polymer Degradation and Stability 81 (2003) 367–373 371

enough to fracture the surface region of the material. It

was verified that the surface of highly aged samples
showed well defined cracks. The QUV aged samples did
not present visible cracks for aging times lower than 400
h. In these samples some cracks with shallow depths
were visible. The cracks are related to fracture induced
in the surface by the aging process associated with
thermal cycling. Despite the fact that hardness also
increased, indicating an increased resistance to plastic
deformation, the stresses were high enough to start
fracturing at some point at the surface. Further thermal
cycling lead to a continuous increase in the elastic
modulus at the surface, promoting slow cracking pro-
pagation as seen in long aged samples.
Furthermore, the simultaneous occurrence of chemi-
Fig. 6. Variation of hardness for penetrations of 500 nm at different cal modifications as oxidation and crosslinking con-
aging times in WOM and QUV regimes. tribute to the development of surface tension and
cracking. Fig. 8A and B taken with the optical micro-
The variation of the carbonyl index obtained from scope illustrate the cracked surface and the nanoinden-
infrared spectroscopy as a function of time is presented
in Fig. 7. It can be observed that the nanomechanical
determinations match very well the chemical degrada-
tion, followed by carbonyl formation, that is LDPE
degradation begins at around 100 h, the effect of the
QUV regime is about twice as aggressive as WOM for
long times of exposure, as 800 h, and the way the
mechanical properties vary with depth is very similar to
the observed chemical degradation profile, determined
by ATR–FTIR [46].
During the thermal cycles the sample increases and
decreases its volume by dilatation. In non-permanent
deformation conditions, the instantaneous deformation
must be nearly the same from bulk to surface. When the
elastic moduli are different at different levels, residual
stresses will build up as the sample is deformed. Our
results showed that the elastic modulus can grow up to
four times from surface to bulk and consequently the
stress build up is very high. Considering a deformation
of 1% in the entire sample, the generated stresses due
different elastic moduli at different levels are high

Fig. 8. The cracked surface of LDPE aged during 1600 h in the WOM
Fig. 7. Variation of the carbonyl index obtained from infrared spec- chamber (A) and 800 h in the QUV chamber (B), showing the marks
troscopy as a function of time. left by the nanoindenter indentation.
372 A.C. Tavares et al. / Polymer Degradation and Stability 81 (2003) 367–373
tation of LDPE samples aged for 1600 h in the WOM
chamber (Fig. 8A), and 800 h in the QUV chamber
(Fig. 8B). The observed effects indicate that a higher
degree of crosslinking must have been achieved in the
QUV chamber, giving raise to higher stresses resulting
in severe cracking in all directions.
The effect of mechanical properties in the electrical
performance of polyethylene has been addressed by
David et al. [20] who reported that LDPE samples
with residual stresses of 8–9 MPa had different per-
formance than those with residual stresses of 5–6
MPa. These results have shown that even when no
cracks were seen at the surface, the increase in resi-
dual stresses induced by aging are an important fac-
tor in the electrical behavior of polyethylene as an
insulator.
4. Conclusions
LDPE was submitted to strictly controlled aging con-
ditions using two different aging regimes: WOM and
QUV. The effect on mechanical properties was mea-
sured by the nanoindentation technique. It was
observed that hardness and elastic modulus values grew
higher from surface to bulk with aging time and that the
QUV regime exerted a more severe degradation effect
than the WOM. The modifications in the elastic mod-
ulus with aging were high enough to promote cracking
at the surface for exposure times of 400 h and above,
and were attributed to a continuous variation of the
modulus with depth.
Mechanical profiles with depth follow the same pat-
tern as found for the carbonyl index. This indicates that
the chemical modifications, brought about by aging, can
be perfectly correlated with the resulting mechanical
properties, and that nanoindentation is a powerful tool
to characterize the degradation profile of aging in poly-
meric materials.
Acknowledgements
J.V.G. thanks LACTEC and L.A., A.C.T and C.M.L
thank CNPq for financial support.
References
[1] Tidjani A, Arnaud R. Polymer 1995;36(14):2841–4.
[2] Zhao T, Bernstorf RA. IEEE—Electrical Insulation Magazine
1998;14(2):26–33.
[3] Gijsman P, Meijers G, Vitarelli G. Polymer Degradation and
Stability 1999;65:433–41.
[4] Barkhudaryan VG. Polymer 2000;41:5787–91.
[5] Tidjani A. Polymer Degradation and Stability 2000;68:465–9.
[6] Carrasco F, Pagès P, Pascual S, Colom X. European Polymer
Journal 2001;37:1457–64.
[7] Guadagno L, Naddeo C, Vittoria V, Camino G, Cagnani C.
Polymer Degradation and Stability 2001;72:175–86.
[8] Gavrila DE. Journal of Applied Polymer Science 1996;59:71–6.
[9] O’Neill P, Birkinshaw C, Leahy JJ, Barklie R. Polymer Degra-
dation and Stability 1999;63:31–9.
[10] Suljovrujic E, K-Popovic Z, Kostoski D, Dojcilovic J. Polymer
Degradation and Stability 2001;71:367–73.
[11] Setnescu R, Silviu J, Setnescu T, Podina C, Osawa Z. Polymer
Degradation and Stability 1998;61:109–17.
[12] Lu S, Buchanan FJ, Orr JF. Polymer Testing 2002.
[13] Suarez JCM, Monteiro EEC, Mano EB. Polymer Degradation
and Stability 2002;75:143–51.
[14] Suarez JCM, Mano EB. Polymer Degradation and Stability 2001;
72:217–21.
[15] Jahan MS, Stovall JC, King MC. Nuclear Instruments and
Methods in Physics Research B 2001;185:323–7.
[16] Sanche L. IEEE Transactions on Electrical Insulation 1993;
28(5):789–819.
[17] Zeid HMA, Ali ZI, Maksoud TMA, Khafagy RM. Journal of
Applied Polymer Science 2000;75(2):179–200.
[18] Montanari GC, Simoni L. IEEE Transactions on Electrical
Insulation 1993;28(5):775–6.
[19] Densley RJ, Bartnikas R, Bernstein B. IEEE Transactions Power
Delivery 1994;9(1):559–71.
[20] David E, Parpal JL, Crine JP. IEEE Transactions on Dielectrics
and Electrical Insulation 1996;3(2):248–57.
[21] Khabbaz F, Albertsson AC, Karlsson S. Polymer Degradrada-
tion and Stability 1998;61:329–42.
[22] Khabbaz F, Albertssson AC, Karlsson S. Polymer Degradation
and Stability 1999;63:127–38.
[23] Lagarón JM, Pastor JM, Kip BJ. Polymer 1999;40:1629–36.
[24] Monney L, Belali R, Vebrel J, Dubois C, Chambaudet A. Poly-
mer Degradadation and Stability 1998;62:353–9.
[25] Sack S, Wagner H, Steger E. Acta Polymer 1985;36:305–10.
[26] Gupta G, Chiang CP. ACS Division of Polymeric Materials
1990:667–71.
[27] Das-Gupta, D.K., Cooper, D.E. Conference Record 1988 IEEE
Internat Symp Electr Insulation, Boston; 5–8 June 1988. p. 335–
338.
[28] Delor F, Barrois-Oudin N, Duteurtre X, Cardinet C, Lemaire
J, Lacoste J. Polymer Degradation and Stability 1998;62:395–
401.
[29] Gorur RS, Bernstein BS. IEEE Transactions Power Delivery
1998;13:316–21.
[30] Gorur RS, Bernstein BS. IEEE Transactions Power Delivery
1998;13:323–9.
[31] Han SO, Lee DW, Han OH. Polymer Degradation and Stability
1999;63:237–43.
[32] Horrocks AR, Mwila J, Miraftab M, Liu M, Chohan SS. Poly-
mer Degradation and Stability 1999;65:25–36.
[33] Severini F, Gallo R, Ipsale S, Nisoli E, Pardi M. Polymer
Degradation and Stability 1999;65:329–35.
[34] Gugumus F. Polymer Degradation and Stability 1999;66:161–72.
[35] Birtwhistle D, Blackmore P, Krivda A, Cash G, George G. IEEE
Transactions on Dielectrics and Electrical Insulation 1999;
6(5):612–9.
[36] Shimizu N, Laurent C. IEEE Transactions on Dielectrics and
Electrical Insulation 1998;5(5):651–9.
[37] Costa L, Jacobson K, Brunella V, Bracco P. Polymer Testing
2001;20:649–59.
[38] Uhniat M, Kudla S. Polymer Degradation and Stability 2001;
71:75–82.
[39] Allen NS, Edge M, Holdsworth D, Rahman A, Catalina
F, Fontan E. Polymer Degradation and Stability 2000;
67:57–67.
[40] Haider N, Karlsson S. Polymer Degradation and Stability 1999;
64:321–8.
 A.C. Tavares et al. / Polymer Degradation and Stability 81 (2003) 367–373
[41] Kurtz SM, Muratoglu OK, Buchanan F, Currier B, Gsell R,
Greer K. Biomaterials 2001;22:1731–7.
[42] Wu Q, Qu B, Xu Y, Wu Q. Polymer Degradation and Stability
2000;68:97–102.
[43] McLaughlin WL, Silverman J, Al-Sheikhly M, Chappas WJ,
Zhan-Jun L, Miller A. Radiation Physics and Chemistry 1999;
56:503–8.
[44] Ciuprina F, Teissèdre G, Filippini JC. Polymer 2001;42:7841–6.
[45] Lee AW, Santerre JP, Boynton E. Biomaterials 2000;21:851–61.
[46] Gulmine JV, Janissek PR, Heise HM, Akcelrud L. Polymer
Degradation and Stability 2003;79:385–97.
[47] Oliver WC, Pharr GM. Journal of Materials Research 1992;
7(6):1564–80.
[48] Doerner MF, Nix WD. Journal of Materials Research 1986;
1:601–9.
373
[49] Baker SP. Encyclopedia of materials: science and technology.
Oxford: Elsevier Science Ltd; 2001.
[50] Flores A, Baltá Calleja FJ. Philosophical Magazine A 1998;
78(6):1283–97.
[51] Beake BD, Leggett GJ. Polymer 2002;43:319–27.
[52] American Standards for Testing and Materials. Standard practice
for operating light-exposure apparatus (Xenon-arc type) with and
without water for exposure of nonmetallic materials, ASTM G26.
Easton; 1994.
[53] American Standards for Testing and Materials. Standard practice
for operating light- and water-exposure apparatus (Fluorescent
UV-condensation type) for exposure of nonmetallic materials,
ASTM G53. Easton; 1996.
[54] Gulmine JV, Janissek PR, Heise HM, Akcelrud L. Polymer
Testing 2002;21(5):557–63.