Engineering Failure Analysis 92 (2018) 1–11

Contents lists available at ScienceDirect

Engineering Failure Analysis

journal homepage: www.elsevier.com/locate/engfailanal

Influence of hydraulic oil on degradation behavior of nitrile rubber

T
O-rings at elevated temperature
⁎
Weitao Lou, Weifang Zhang, Hongxun Wang , Tingzhu Jin, Xuerong Liu
School of Reliability and Systems engineering, Beihang University, Beijing 100191, China

A R T IC LE I N F O ABS TRA CT

Keywords: The degradation of nitrile rubber O-rings exposed to hydraulic oil and air were investigated by
Nitrile rubber O-rings adopting accelerated aging test. The chemical structures and mechanical properties of the aged
Degradation samples were evaluated by measuring attenuated total reflection Fourier transform infrared
Hydraulic oil (ATR-FTIR) spectroscopy, crosslinking density, weight loss, mechanical properties and fracture
Crosslinking
morphology. The ATR-FTIR results indicate that chemical structure changed significantly due to
Chain scission
loss of additives and oxidation reactions. Hydraulic oil and temperature played a critical role in
surface chemical changes. The crosslinking density changes were attributed to the competition
between crosslinking and chain scission. The mechanical test results show that there existed large
differences in changes of mechanical properties for samples under different conditions, which
correlated to loss of fillers, the changes in crosslinking density and the formation of defects.
Additionally, the fracture morphologies demonstrate that the formation of hardened brittle outer
layer, voids and agglomerates also promoted the decrease in mechanical properties.

1. Introduction

Nitrile rubber is well known for its good wear resistance, oil resistance and impermeability, and has been widely used as seals and
gaskets in the sealing fields of oil and gas especially in hydraulic system such as piston seals, sealing of piston rods and axial sealing.
However, due to the existence of unsaturated bonds and volatile components in rubber materials, the severe environmental factors
such as oxygen, radiation, high temperature, aggressive media and mechanical stress can accelerate the aging process of seals, which
leads to the degradation in mechanical properties and network structure [1–8]. Eventually, the seals lose the ability of deformation
recovery, resulting in failure of sealing system due to leakages of gas and/or liquid [9,10]. Additionally, natural aging of rubber seals
is a relatively slow process under actual working conditions, thus it needs a long time to gain the performance degradation data.
Accordingly, the artificial accelerated degradation test conducted by simulating actual service conditions can be applied to in-
vestigate the degradation behaviors and mechanisms of rubber seals for shortening test duration.
When the seals are used in the hydraulic system of electro-hydraulic servo valve, they unavoidably are immersed into the hy-
draulic oil. During the immersion process, hydraulic oil gradually diffuses into the interior of seals, which led to swelling of rubber
network structure. Furthermore, the coupled effect between hydraulic oil and temperature has a negative impact on thermal stability
of rubber materials. Additionally, due to the complexity of rubber material compounds containing rubber matrix, various fillers and
additives, the degradation of nitrile rubber materials immersed in hydraulic oil is an extremely complex process accompanied by a
series of physical and/or chemical reactions [11,12]. The degradation of nitrile rubber materials in oil often occurs through the follow
processes: the extraction of soluble components (e.g., plasticizers and antioxidants); the reactions of unsaturated bonds (e.g.,

⁎
Corresponding author.
E-mail address: by1514127@buaa.edu.cn (H. Wang).

https://doi.org/10.1016/j.engfailanal.2018.05.006
Received 7 November 2017; Received in revised form 25 January 2018; Accepted 9 May 2018
Available online 16 May 2018
1350-6307/ © 2018 Elsevier Ltd. All rights reserved.
W. Lou et al. Engineering Failure Analysis 92 (2018) 1–11

carbon‑carbon double bonds and nitrile groups); thermal oxidative aging reactions (crosslinking and chain scission). Additionally, the
loss of additives weakens the antioxidant ability of rubber matrix structure, and causes the worse compatibility between rubber
matrix and additives [13]. The crosslinking and chain scission reactions make the network structure denser and denser, and restrict
the movement of molecular chains [14]. Subsequently, these changes can affect seriously physical and mechanical properties of the
seals. Besides, the oil type has a direct influence on oil resistance of the seals. The biodiesel has more excellent dissolving capacity
than diesel fuel according to the principle of “like dissolves like” [13–15].
The degradation behaviors of nitrile rubber materials (mostly the unfilled rubber and rubber sheet) in oil have been studied
extensively by many researchers. Pan et al. [11]. and Akhlaghi et al. [16]. reported the effects of crude oil and lubricant oil on the
properties of hydrogenated nitrile rubber elastomer and radial lip seal. The deterioration of nitrile rubber [13,17], hydrogenated
nitrile rubber [18] and the fuel hose [19] in different types of biodiesel such as palm biodiesel, rapeseed biodiesel and sweet oil were
investigated. Buckley et al. [20]. and Pazur et al. [21]. showed the effect of liquid media on lifetime predictions of nitrile rubber O-
rings. Additionally, the compatibility of nitrile rubber with diesel fuel [22], biodiesel [15,23] or their mixtures [13,24] were also
studied. In spite of these significant efforts, most studies are limited to the performance degradation of nitrile rubber in oil. There are
relatively less literatures focused on the degradation behaviors and mechanisms of the nitrile rubber O–rings in hydraulic oil.
The objective of this research is to investigate the degradation behaviors and mechanisms of nitrile rubber O-rings immersed in
hydraulic oil at elevated temperatures (70 °C, 90 °C and 110 °C) by the systematic accelerated thermal aging tests. Meanwhile, the
contrast tests of nitrile rubber O-rings exposed to air under the same conditions (temperature and aging time) were conducted to
further reveal the influence of hydraulic oil on the degradation process. Subsequently, the changes in chemical structures of the aged
samples were evaluated by attenuated total reflection-Fourier transform infrared spectroscopy (ATR-FTIR) and solvent swelling test.
The physical and mechanical properties of the samples before and after aging were assessed by weight loss, mechanical properties
and fracture morphologies. By comparing and analyzing the changes in physical properties, mechanical properties and chemical
structural of the samples under different aged conditions, we concluded how the hydraulic oil and elevated temperature affected the
degradation behaviors and mechanisms of the nitrile rubber seals during the thermal aging process.

2. Experimental

2.1. Material and hydraulic oil

Vulcanized nitrile rubber O-rings with 35% acrylonitrile were provided by Changsha 5712 Aircraft Industry Corporation, Ltd,
China. The material was compounded according to the formulation shown in Table 1, using a two-roll mill at a constant temperature
of 151 °C and a pressure of 60 to 120 kg/cm2 for 40 min. Then, the materials were processed into O-rings, which measured
Φ17 mm × 6 mm (inner diameter x cross section diameter, Fig. 1).
The hydraulic oil used in immersion test is 46# aviation hydraulic oil, which was supplied by Great Wall Lube Oil, China. Table 2
shows the main components of hydraulic oil for immersion test. The density of the hydraulic oil is 0.862 kg/L at 25 °C. The kinematic
viscosity is 38.12 mm2/s at 50 °C. The flash temperature is 165 °C, and the acid value is 0.78 mg KOH/g.

2.2. Aging methods

To investigate the effect of hydraulic oil on the degradation behaviors and mechanisms of nitrile rubber O-rings, the accelerated
aging tests (immersion in oil and exposure to air) were performed in air-circulating ovens. The immersion tests were conducted to
simulate the hydraulic oil environment of O-rings in electro-hydraulic servo. The samples were placed in covered jars containing the
hydraulic oil, which occupied 60% space of container. In addition, for comparison, the exposure tests in air were also conducted
under the same aging conditions (temperature and aging time) with the immersion test. Then, the containers for immersion and
exposure tests were placed in air-circulating ovens, and the test temperatures were selected at 70 °C, 90 °C and 110 °C. The aging time
was 2 days, 4 days, 8 days, 16 days, 32 days and 64 days. To obtain the reliable results, each group of experiment set up three parallel
samples for the following measurements.

Table 1
Constituents of the NBR samples.
Constituents phr

Nitrile rubber 100

Zinc oxide 5
Sulfur 2.5
Carbon black 50
Plasticizer 18
Stearic acid 1.5
Antioxidant 2

2
W. Lou et al. Engineering Failure Analysis 92 (2018) 1–11

Fig. 1. The diagram of O-ring seal.

Table 2
The main components of hydraulic oil.
Components Mass fraction/%

Mineral oils 97.0–98.0

Antioxidant 264 0.2–0.4
Tricresyl phosphate 0.2–0.5
Lanolin 0.4–0.6
Dibutyl Phosphite 0.1–0.2
N,N-Din-butyl-diethylamino methylene-benzotriazole 0.01–0.05

2.3. Characterization methods

After the oil immersion test, the aged samples were removed from the vessels, and the excess oil on the sample surface was cleared
with filter paper. Due to the swelling effect, the oil diffused into the samples and extracted the smaller molecules and additives. To
reduce the effects of oil on subsequent characterization methods and to acquire the reliable test results that could be compared with
the samples aged in air, the samples aged in hydraulic oil were dried in a vacuum oven at 110 °C to remove the oil (via a desorption
process) up to a constant weight, and the samples aged in air were also treated under the same conditions to improve accuracy.

2.3.1. ATR-FTIR analysis

Attenuated total reflection Fourier transform infrared (ATR-FTIR, Thermo Nicolet 6700) spectroscopy was applied to study
changes in the chemical structure of the surfaces of the seals immersed in hydraulic oil or exposed to air. The range of spectra is
400–4000 cm−1 with a resolution of 4 cm−1. The ATR-FTIR spectrometer with a diamond crystal was used to collect the data. The
scans at three different locations on one of the aged samples (Fig. 1) were averaged to more accurately identify the peaks of the aged
sample.

2.3.2. Crosslinking density

To evaluate changes of the rubber network structure, the equilibrium swelling behaviors of the aged samples were studied by
conducting swelling tests. The samples were cut from the aged samples. The length, width and thickness of the samples are 20 mm,
10 mm and 2 mm. The tests were conducted in acetone at 23 °C for 96 h. Four parallel samples were measured over time for each
exposure environment, and the average values are shown in this paper. The Florye-Rehner equation was applied to calculate the
crosslinking density of the three-dimensional network:

1 ⎡ ln(1 − Vr ) + Vr + μVr2 ⎤
Ve = −
V⎢⎣ Vr1/3 − Vr /2 ⎥
⎦ (1)

where Ve was the crosslinking density, μ was polymer-solvent interaction parameter, V was the molecular volume of acetone, and Vr
was the volume fraction of nitrile rubber after immersion in acetone [5].

2.3.3. Weight loss

Weight loss of the aged samples was measured by adopting electronic balance which has a resolution of 0.001 g. Three parallel
samples were measured over time for each exposure environment, and the average values are shown in this paper. The percent weight
loss (WL) was calculated by applying the following equation:

3
W. Lou et al. Engineering Failure Analysis 92 (2018) 1–11

M1 − M2
%WL = × 100%
M1 (2)
Where M1 was the initial weight of the sample and M2 was the final weight of the sample. For oil immersion test, M2 was the
weight of the sample after vacuum drying.

2.3.4. Scanning electron microscopy (SEM) analysis

Changes in the tensile fracture morphology of the samples before and after aging in different conditions were analyzed using
scanning electron microscopy (SEM, JSM-6480). After tensile tests, all the samples were immediately placed into a sealed bag to
prevent the fracture surface from being polluted.

2.3.5. Tensile properties

Mechanical properties of the aged samples were estimated by tensile tests. The tensile samples were conducted on each group
according to GB/T 5720:1993 (China) using a CMT5504 electronic tensile testing machine at the strain rate of 50 mm/min. The
Young's modulus (secant modulus at 5% strain), tensile strength and elongation at break of three samples were determined for each
experimental group.

3. Results and discussion

3.1. ATR-FTIR analysis

In order to determine the chemical structure changes on the surface of the samples before and after aging, the ATR-FTIR spectra of
the unaged and aged samples under different conditions were conducted. Fig. 2 shows the ATR-FTIR spectra of the samples immersed
in hydraulic oil at 110 °C for different aging days. The peak at 2232 cm−1 belonged to the –CN stretching vibration. Peaks at
2918 cm−1 and 2845 cm−1 were assigned to the stretching vibration of the CH2 group. Peaks at 1733 cm−1, 1644 cm−1 and
1532 cm−1 corresponded to C]O stretching vibration, C]C stretching vibration and additives such as curing package [1,18], re-
spectively. The peak at 1430 cm−1 was assigned to the –CH2– deformation vibration. The peak at 962 cm−1 was attributed to the
CeH deformation vibration, which was characteristic of the trans-1, 4-structure of the butadiene unit. Additionally, the new peak at
3394 cm−1 was mostly attributed to the NeH stretching vibration and hydroxyl groups. The appearance of the new peak at
3394 cm−1 indicated the oxidation reactions occurred in the matrix. Meanwhile, the appearances of new peaks at 3185 cm−1,
1651 cm−1, 1627 cm−1 and 1418 cm−1 were assigned to NeH stretching vibration, C]O stretching vibration, NeH deformation
vibration and the CeN stretching vibration, respectively, which implied the formation of amide groups (R-CO-NH2) [3,15,16]. The
weak absorption peak at 1644 cm−1 implied the residual unsaturated carbon-carbon double bond (C]C) still remained in rubber
molecular chains. However, the peak at 1644 cm−1 disappeared after 2 days of aging, which may be due to further curing and/or
oxidation throughout aging [3]. Furthermore, at elevated temperature, the residual carbon-carbon double bonds are prone to oxi-
dation compared to the further curing.
Additionally, Fig. 2 also shows the FTIR spectra of hydraulic oil before and after exposure for 32 days at 110 °C. The peak at
2921 cm−1 and 2854 cm−1 belonged to symmetric and asymmetric stretching vibration of the CH2 groups, respectively. The peak
around 1469 cm−1 was assigned to the bending vibrations of the CH2 and CH3 groups. Peaks at 1383 cm−1 was attributed to the

Fig. 2. ATR-FTIR spectra of the samples immersed in hydraulic oil at 110 °C up to 32 days.

4
W. Lou et al. Engineering Failure Analysis 92 (2018) 1–11

Fig. 3. ATR-FTIR spectra of the samples aged in oil and air for different aging days at 70 °C and 110 °C.

bending vibrations of CH2 groups. The results showed the typical characteristic absorption peaks of hydraulic oil consisted of hy-
drocarbons with different length of carbon chain. The peak around 1751 cm−1 was attributed to the carbonyl ester functional groups,
which indicated the presence of acyl esters in unaged and aged oil. The weak absorption peak at 1751 cm−1 originated from the
additives. After aging, the slight increase of the absorption peak was mainly due to the migration of additives and/or oxidation
products of the aged seals, which was in agreement with the decrease of the peak at 1733 cm−1 for the FTIR spectrum of the aged
seals.
To further observe how hydraulic oil and elevated temperature influence the changes of chemical groups of the aged samples
during degradation process, the typical infrared spectra of the samples immersed in hydraulic oil and exposed to air at 70 °C and
110 °C are shown in Fig. 3. Additionally, the absorbance intensities of different IR bands were normalized to the 1036 cm−1 peak
(unaged) to characterize the changes of IR band intensities in quantities, shown in Fig. 4. It can be seen from Fig. 3 and Fig. 4. After
immersion test with different days at 110 °C, the peaks at 2918 cm−1, 2845 cm−1, 1733 cm−1, 1532 cm−1, 1430 cm−1 and 962 cm−1
decreased obviously in intensity as aging time increased. The decreases of the band at 2918 cm−1 and 2845 cm−1 in intensity was
mainly due to the loss of paraffin [1,12,16]. The changes at 1733 cm−1 were probably because of the loss of additives or the
consumption of carbonyl groups [12,17]. Additionally, the appearance of R-CO-NH2 happened after 64 days of immersion in oil at
70 °C, but not in air. The hydroxyl groups were more prone to be formed in oil compared to that of the samples exposed to air.
However, at 110 °C, not only the reactions of the nitrile group immersed in oil for 8 days appeared earlier than that of the samples
exposed to air for 32 days, but also the peak intensity of the hydroxyl group increased more rapidly in oil than that of the samples

Fig. 4. Highlighted area of the ATR-FTIR spectra of nitrile rubber O-rings aged under the listed conditions. Spectra were normalized to the
1036 cm−1 peak: (a) 3500 cm−1-2150 cm−1, (b) 1800 cm−1-900 cm−1.

5
W. Lou et al. Engineering Failure Analysis 92 (2018) 1–11

Fig. 5. Changes in crosslinking density of the samples before and after aging at different temperatures: (a) immersion in oil; (b) exposure to air.

exposed to air. Therefore, with increasing temperature, the oxidation of the sample was intensified, and the effects of hydraulic oil on
the degradation process were more serious than that from air under the same temperature. Especially, the synergistic effect between
higher temperature (at 110 °C) and hydraulic oil had a significant impact on the degradation of the rubber seals. This was principally
due to the fact that the more rapid migration of additives in oil at higher temperature led to the poorer resistance to oxidation of the
matrix. On the other hand, the hydraulic oil would promote the internal dissolution and swelling of the network structure, resulting
in the stretching and disentanglement of molecular chains, the larger space of molecular chains and enhancing the activity of
molecular chains. In addition, the reactions of the nitrile group might made the polarity change and the oil resistance of the nitrile
rubber decline, which resulted in the decrease of chain flexibility and weakened the interaction among molecular chains. Finally, the
thermal oxidative aging of the molecular chains was further accelerated under the synthetic effects of multiple factors.

3.2. The changes in crosslinking density

The changes in crosslinking density of the samples aged in oil and air at different temperatures are shown in Fig. 5. At 70 °C, the
crosslinking density of the samples immersed in oil increased slightly with increasing aging time. Furthermore, there existed a similar
trend for the samples exposed to air at 70 °C. At 90 °C, with increasing aging time, both the crosslinking density of the samples aged in
oil and air increased continuously. However, this increase in air was faster than that in oil. Additionally, at 110 °C, the crosslinking
density of the samples immersed in oil and exposed to air increased significantly at an early stage, and then decreased slightly and
kept stable after 16 days, respectively. Furthermore, this increase in oil was relatively slower than that in air during the initial 8 days
of aging, but faster from 8 to 16 days.
Previous studies have proved that crosslinking and chain scission reactions often occurred simultaneously during the thermal
aging process [1,14]. The changes of crosslinking density were regarded as the result of the competition between crosslinking and
chain scission. If the influence of crosslinking reaction predominated, the crosslinking density would increase, otherwise decrease,
which could well explain the changes in crosslinking density of samples aged in oil and air at different temperatures in this study.
Besides, due to the presence of the residual unsaturated carbon-carbon double bonds in the matrix, the crosslinking consisted of the
further crosslinking and the oxidation-induced crosslinking. For the crosslinking, the absorption peak of the unsaturated double
bonds disappeared after 2 days of aging according to the FTIR analysis, which illustrated that the further crosslinking mainly occur in
the early stage, accompanied by the oxidation-induced crosslinking. Then, the oxidation-induced crosslinking held the dominant
position. The analysis results demonstrated that the crosslinking play a mainly role at lower temperatures (70 °C and 90 °C), resulting
in increase of crosslinking density, whereas at higher temperature (110 °C), the crosslinking predominated at the early stage leading
to the rapider increase of crosslinking density, and then the chain scission took over gradually after 16 days of aging.
The samples exposed to air showed a faster increase in crosslinking density than that immersed in oil, which implied the more
serious degradation for the samples aged in air. However, thermal stability of the samples aged in oil should have become weaker due
to the faster loss of additives, such as antioxidants and plasticizers, than that of the samples aged in air, leading to more prone to
oxidation reactions. Furthermore, for the immersion test, the samples were completely immersed in hydraulic oil, which directly cut
off oxygen from the samples, whereas a small amount of oxygen still penetrated into the seals through slow diffusion in the oil. These
phenomenon caused the lower oxygen concentration at the surface of the seals for oil-immersed samples compared to air-exposed
samples, and slowed further oxidation of the samples immersed in oil [2,25,26]. At 70 °C, the oxygen consumption required for the
oxidation reactions was below the oxygen content that diffused into the samples through the oil. Thus, there is no obvious difference
between the two curves in different mediums. Nevertheless, at 90 °C and 110 °C, the supply of the oxygen could not meet the
requirements of the oxidation reactions leading to slow the oxidation process. These results indicated that the hydraulic oil and the
temperature played a main role in the crosslinking density during the aging process, and the latter was more direct.

6
W. Lou et al. Engineering Failure Analysis 92 (2018) 1–11

Fig. 6. Weight loss of the samples before and after aging at different temperatures: (a) immersion in oil; (b) exposure to air.

3.3. Mass changes

Fig. 6 shows the weight loss of the samples immersed in oil and exposed to air at different temperatures. The changes in weight of
the samples immersed in oil showed similar trends at three temperatures. The weight of the samples immersed in oil decreased
sharply during the first 4 days, and then increased slowly. Moreover, for the samples in oil after 4 days of exposure, the higher the
temperature, the less the weight loss. However, for the samples exposed to air, the similar behavior was also observed at 90 °C and
110 °C, an initial weight loss increased with increasing aging time, followed by a slightly decrease thereafter, but the weight loss was
not obvious at 70 °C compared to that of the samples at the higher temperature.
Previous studies have demonstrated that the swelling, solvent extraction and degradation would occur when the elastomers were
immersed in oil at high temperature [13,17]. Moreover, the swelling and solvent extraction had greatly effects on the degradation
process of rubber materials [20]. The swelling caused the expansion of the network structure and weakened the intermolecular forces
between the molecular chains and additives, which led to the poor compatibility between the additives and polymer matrix. Ad-
ditionally, the soluble components (such as plasticizers and stabilizers) of the elastomers would dissolve in oil. Thus, solvent ex-
traction caused an initially rapid reduction in the weight of the samples after immersion. However, the extraction of additives led to
the decrease in the thermal stability and antioxidant performance, which accelerated the process of oxidation reactions. The oxi-
dation reactions made the network structure denser and denser. The denser network structure restricted the migration of the ad-
ditives and small molecule solvent, which illustrated the difference of the changes in weight between high temperature and low
temperature. On the other hand, accompanied by the reaction products bonding to the polymer matrix, the weight of the aged
samples would slightly increase at the later stage [20]. For the samples exposed to air, the initial rapid decrease in weight was due to
the volatilization of small molecules which included the volatile additives and the oxidation induced small molecules. The higher the
temperature, the faster the volatilization rate. By comparison, we found that the effect of temperature on the changes of weight in air
was more noticeable than that from oil, and the oil had a more significant impact on weight changes vis-a-vis temperature.

3.4. Fracture morphology (SEM)

In order to further elevate what caused the failure of rubber seals, the typical fracture morphologies of the samples before and

Fig. 7. SEM of the tensile fracture surface of the unaged seals.

7
W. Lou et al. Engineering Failure Analysis 92 (2018) 1–11

Fig. 8. SEM of the tensile fracture surface of the seals exposed to air at 110 °C after 64 days.

after aging in air and oil was evaluated by SEM as shown in Figs. 7-9. By analyzing the morphology characteristic of fracture surface,
including the fracture source, crack growth region and final rupture region, we could identify that the sample was ductile fracture or
brittle fracture.
Fig. 7 shows the fracture morphologies of the unaged sample. The crack growth region was relatively rough (Fig. 7 (a)) and the
final rupture region was smooth (Fig. 7 (b)), which relatively conformed to the ductileness fracture feature. At a magnification of
1000× (Fig. 7 (c)), we observed that the additives uniformly distributed in the matrix and no obvious defects occurred.
Fig. 8 and Fig. 9 show the fracture morphologies of the sample aged in air and oil for 64 days at 110 °C, respectively. The fracture
source appeared in the specimen interior and showed the phenomenon that the part of rubber matrix peeled off from the matrix
surface (Fig. 8 (a)). However, two fracture sources occurred in the fracture surface of the sample immersed in oil (Fig. 9 (a)), which
indicated the more defects were formed and led to appearance of stress concentrations. Additionally, the linear stripes were observed
around the fracture source, and the crack growth region was relatively smooth and the final rupture region was rough (Fig. 8 (c)),
which matched well with brittleness fracture feature. However, the fracture surface of the samples immersed in oil turned rougher,
and showed the wider linear stripes and more obvious defects (such as, voids, secondary cracks and particle agglomerates) than air
(Fig. 9 (c) and Fig. 9 (d)). These changes were mainly because the migration of additives, swelling of network structure, crosslinking

Fig. 9. SEM of the tensile fracture surface of the seals immersed in oil at 110 °C after 64 days.

8
W. Lou et al. Engineering Failure Analysis 92 (2018) 1–11

Fig. 10. Changes in elongation at break of the samples before and after aging at different temperatures: (a) immersion in oil; (b) exposure to air.

and chain scission occurred, which resulted in the deformation of the network structure and the rearrangement of molecular chains
[15,19]. Furthermore, the reactions of the polar group (nitrile groups) could alter the distribution of internal energy and the dis-
solution of nonpolar molecular chains. The physical and chemical changes resulted in the separation of additives from matrix and the
non-uniform distribution of density [5,19]. When the aged samples were subjected to force, stress concentration could be generated
around voids and aggregates, and then gradually formed the secondary crack with stress loading. Meanwhile, at the edge of the
fracture surface, we observed a hardened brittle outer layer, and the layer had the expansion trend from the edge to the core of the
samples, implying that the aging of the samples gradually conducted from the surface to the internal structure. Moreover, the layer of
the samples immersed in oil was thicker than that of the samples exposed to air, whereas the latter was denser due to the sharp
increase of the crosslinking density. An explanation for the phenomenon was the oil and diffusion-limited oxidation (DLO) effects
which limited the diffusion of oxygen into the interior [26]. All these results demonstrated that the oil had a significant impact on the
aging of the seals.

3.5. Changes in mechanical properties

The mechanical properties of the nitrile rubber are closely related to the network structure, thus the changes in the matrix
structure including the physicochemical and chemical changes would lead to the degradation of mechanical properties. The changes
in elongation at break, tensile strength and Young's modulus of the samples immersed in oil and exposed to air at different tem-
peratures are shown in Figs. 10-12, respectively.
With increasing aging time, it was obvious from Fig. 10 that the elongation at break of the samples decreased gradually after oil
immersion and air exposure at 70 °C and 90 °C, but showed slower drop after 8 days. At 110 °C, the elongation at break showed a
rapid decline trend during the first days and then kept steady at the later stages. Especially, the elongation at break almost lost

Fig. 11. Changes in tensile strength of the samples before and after aging at different temperatures: (a) immersion in oil; (b) exposure to air.

9
W. Lou et al. Engineering Failure Analysis 92 (2018) 1–11

Fig. 12. Changes in Young's modulus of the samples before and after aging at different temperatures: (a) immersion in oil; (b) exposure to air.

elasticity after 32 days of aging, and decreased 99% over the aging process. Additionally, the decrease in the elongation at break was
more pronounced for higher temperatures and air exposure. This could be attributed to the increase of crosslinking density and the
loss of fillers forming the denser network structure which restricted the movement of segments and led to the decrease in the
elongation at break.
It was shown from Fig. 11 that the tensile strength of the samples decreased slightly with increasing aging time at 70 °C, probably
due to the loss or migration of fillers and/or the destruction of network structure caused by swelling. At 90 °C, the tensile strength
showed a slight increase and then decreased continually after reaching a maximum value, due to the significant increase of the
crosslinking density. The moderation crosslinking of rubber restricted the relative sliding of molecule chains, which contributed to
the transference and dispersal of the force, leading to the increase of the tensile strength. However, the higher crosslinking density
caused the decrease of the number of effective network chains on orienting and the activity limitations of the segments, which were
not beneficial for the dispersion of stress, and then resulted in the decrease of the tensile strength [5,27,28]. Meanwhile, with
increasing aging time, the tensile strength decreased significantly at 110 °C, and then remained constant after 32 days, due to loss of
fillers and the significant increase of crosslinking density. Based on the changes in tensile strength, we found that the higher tem-
perature would lead to the more significant decrease in tensile strength. Furthermore, the degradation in tensile strength of the
samples exposed to air was more serious than that of the samples immersed in oil.
Fig. 12 shows the changes in Young's modulus of the samples immersed in oil and exposed to air at different temperatures. The
Young's modulus of the samples increased with increasing aging time and temperature. Additionally, all the Young's modulus of the
samples immersed in oil at three temperatures were lower than that of the samples exposed to air, indicating the more serious
degradation for the samples exposed to air. The increase in Young's modulus of the samples aged at 70 °C and 90 °C was attributed to
the increase in crosslinking density, indicating that the oxidative crosslinking dominated over chain scission in the sample and the
stiffening of the rubber. Nevertheless, the substantial increases in Young's modulus of the samples aged at 110 °C objected to the data
that crosslinking density decreased in the middle and later stages of aging. These contradictory phenomena between the crosslinking
density and Young's modulus was probably due to the generation of internal cavitation caused by extensive oil-driven bound [29,30].
The formation of cavitation was observed by tensile fracture morphologies of the aged samples after 32 days of immersion in oil at
110 °C.

4. Conclusions

The degradation behaviors and mechanisms of nitrile rubber O-rings immersed in hydraulic oil and exposed to air were in-
vestigated by the systematic accelerated thermal aging tests, the following conclusions can be made:

1. The ATR-FTIR results show that the amide groups and the hydroxyl groups were formed during the aging process. The effects of
hydraulic oil on the changes in the surface chemical structure were more significantly than air. Besides, the higher temperature
had a greater influence on the chemical structure.
2. The crosslinking density results reveal that the crosslinking density increased with increasing aging time, whereas the crosslinking
density decreased slightly and kept stable at 110 °C due to the domination of the chain scission after immersion in oil and exposure
to air for 16 days, respectively. The results indicate that the hydraulic oil and the temperature affected the markedly crosslinking
density, especially the temperature.
3. In general, the weight of the samples decreased in the early stage and increased slightly in the later stage. The effect of tem-
perature on the changes of weight in air was more markedly than that in oil, and the oil had a more significant effect on weight
changes compared to temperature.
4. The fracture morphology results show that the morphology characteristic of fracture surface changed significantly for the samples

10
W. Lou et al. Engineering Failure Analysis 92 (2018) 1–11

before and after immersion in oil and exposure to air. Moreover, compared to exposure to air, the thicker oxide layer, more voids,
agglomerates and cracks were formed on the fracture surface of samples immersed in oil.
5. From the mechanical test study, the elongation at break decreased and the Young's modulus increased with increasing aging time
and temperature. Furthermore, the changes in the tensile strength of the samples had great differences at different temperatures.
These phenomena correlated to the loss of fillers, the changes in crosslinking density and the formation of defects. These results
imply that the influence of temperature on mechanical properties was greater than the hydraulic oil.

Acknowledgements

This work was supported by the Technical Foundation program (grant JSZL32013B001) from the Ministry of Industry and
Information Technology of China JSZ132013B001.

Conflicts of interest

The authors declare no conflict of interest.

References

[1] J. Zhao, R. Yang, R. Iervolino, B.V.D. Vorst, S. Barbera, The effect of thermo-oxidation on the continuous stress relaxation behavior of nitrile rubber, Polym. Degrad. Stab.
115 (2015) 32–37, http://dx.doi.org/10.1016/j.polymdegradstab.2015.02.013.
[2] J. Wise, K.T. Gillen, R.L. Clough, Quantitative model for the time development of diffusion-limited oxidation profiles, Polymer 38 (1997) 1929–1944, http://dx.doi.org/10.
1016/S0032-3861(96)00716-1.
[3] S. Perraud, M.F. Vallat, M.O. David, J. Kuczynski, Network characteristics of hydrogenated nitrile butadiene rubber networks obtained by radiation crosslinking by electron
beam, Polym. Degrad. Stab. 95 (2010) 1495–1501, http://dx.doi.org/10.1016/j.polymdegradstab.2010.06.016.
[4] Q. Zhao, X. Li, J. Gao, Aging of ethylene–propylene–diene monomer (EPDM) in artificial weathering environment, Polym. Degrad. Stab. 92 (2007) 1841–1846, http://dx.
doi.org/10.1016/j.polymdegradstab.2007.07.001.
[5] J. Liu, X. Li, L. Xu, P. Zhang, Investigation of aging behavior and mechanism of nitrile-butadiene rubber (NBR) in the accelerated thermal aging environment, Polym. Test.
54 (2016) 59–66, http://dx.doi.org/10.1016/j.polymertesting.2016.06.010.
[6] S. Bhattacharjee, A.K. Bhowmick, B.N. Avasthi, Degradation of hydrogenated nitrile rubber, Polym. Degrad. Stab. 31 (1991) 71–87, http://dx.doi.org/10.1016/0141-
3910(91)90097-B.
[7] F. Delor-Jestin, N. Barrois-Oudin, C. Cardinet, J. Lacoste, J. Lemaire, Thermal ageing of acrylonitrile-butadiene copolymer, Polym. Degrad. Stab. 70 (2000) 1–4, http://dx.
doi.org/10.1016/S0141-3910(00)00035-5.
[8] F.S. Ahmed, M. Shafy, A.A.A. El-Megeed, E.M. Hegazi, The effect of γ-irradiation on acrylonitrile–butadiene rubber NBR seal materials with different antioxidants, Mater.
Des. 36 (2012) 823–828, http://dx.doi.org/10.1016/j.matdes.2011.02.066.
[9] S. Akhlaghi, U.W. Gedde, M.S. Hedenqvist, M.T.C. Braña, M. Bellander, Deterioration of automotive rubbers in liquid biofuels: a review, Renew. Sust. Energ. Rev. 43 (2015)
1238–1248, http://dx.doi.org/10.1016/j.rser.2014.11.096.
[10] D. Zeng, Q. He, T. Li, J. Hu, T. Shi, Z. Zhang, Z. Yu, R. Liu, Corrosion mechanism of hydrogenated nitrile butadiene rubber O-ring under simulated wellbore conditions,
Corros. Sci. 107 (2016) 145–154, http://dx.doi.org/10.1016/j.corsci.2016.02.028.
[11] L. Pan, J. Tan, X. Han, P. Li, W. Zhang, Effects of elevated temperature and crude oil on the properties of a hydrogenated nitrile butadiene rubber elastomer, J. Appl. Polym.
Sci. 133 (2016), http://dx.doi.org/10.1002/APP.44012.
[12] J. Zhao, R. Yang, R. Iervolino, S. Barbera, Changes of chemical structure and mechanical property levels during thermo-oxidative aging of NBR, Rubber Chem. Technol. 86
(2013) 591–603, http://dx.doi.org/10.5254/rct.13.87969.
[13] A.S.M.A. Haseeb, H.H. Masjuki, C.T. Siang, M.A. Fazal, Compatibility of elastomers in palm biodiesel, Renew. Energy 35 (2010) 2356–2361, http://dx.doi.org/10.1016/j.
renene.2010.03.011.
[14] J. Zhao, R. Yang, R. Iervolino, S. Barbera, Investigation of crosslinking in the thermooxidative aging of nitrile–butadiene rubber, J. Appl. Polym. Sci. 132 (2014), http://dx.
doi.org/10.1002/APP.41319.
[15] S.M. Alves, V.S. Mello, J.S. Medeiros, Palm and soybean biodiesel compatibility with fuel system elastomers, Tribol. Int. 65 (2013) 74–80, http://dx.doi.org/10.1016/j.
triboint.2013.03.026.
[16] S. Akhlaghi, M.S. Hedenqvist, M.T.C. Braña, M. Bellander, U.W. Gedde, Deterioration of acrylonitrile butadiene rubber in rapeseed biodiesel, Polym. Degrad. Stab. 111
(2015) 211–222, http://dx.doi.org/10.1016/j.polymdegradstab.2014.11.012.
[17] F. Guo, X. Jia, M. Lv, L. Wang, R.F. Salant, Y. Wang, The effect of aging in oil on the performance of a radial lip seal, Tribol. Int. 78 (2014) 187, http://dx.doi.org/10.1016/j.
triboint.2014.05.017.
[18] B. Alcock, J.K. Jørgensen, The mechanical properties of a model hydrogenated nitrile butadiene rubber (HNBR) following simulated sweet oil exposure at elevated
temperature and pressure, Polym. Test. 46 (2015) 50–58, http://dx.doi.org/10.1016/j.polymertesting.2015.06.010.
[19] M. Coronado, G. Montero, B. Valdez, M. Stoytcheva, A. Eliezer, C. García, H. Campbell, A. Pérez, Degradation of nitrile rubber fuel hose by biodiesel use, Energy 68 (2014)
364–369, http://dx.doi.org/10.1016/j.energy.2014.02.087.
[20] G.S. Buckley, C.M. Roland, Influence of liquid media on lifetime predictions of nitrile rubber, J. Appl. Polym. Sci. 131 (2014) 2928–2935, http://dx.doi.org/10.1002/app.
40296.
[21] R.J. Pazur, J.G. Cormier, K. Korhan-Taymaz, Service life determination of nitrile O-rings in hydraulic fluid, Rubber Chem. Technol. 87 (2014) 239–249, http://dx.doi.org/
10.5254/rct.13.87916.
[22] W. Trakarnpruk, S. Porntangjitlikit, Palm oil biodiesel synthesized with potassium loaded calcined hydrotalcite and effect of biodiesel blend on elastomer properties,
Renew. Energy 33 (2008) 1558–1563, http://dx.doi.org/10.1016/j.renene.2007.08.003.
[23] F.N. Linhares, H.L. Corrêa, C.N. Khalil, M.C.A.M. Leite, C.R.G. Furtado, Study of the compatibility of nitrile rubber with Brazilian biodiesel, Energy 49 (2013) 102–106,
http://dx.doi.org/10.1016/j.energy.2012.10.040.
[24] L. Zhu, C.S. Cheung, W.G. Zhang, Z. Huang, Compatibility of different biodiesel composition with acrylonitrile butadiene rubber (NBR), Fuel 158 (2015) 288–292, http://
dx.doi.org/10.1016/j.fuel.2015.05.054.
[25] K.T. Gillen, R.L. Clough, Rigorous experimental confirmation of a theoretical model for diffusion-limited oxidation, Polymer 33 (1992) 4358–4365, http://dx.doi.org/10.
1016/0032-3861(92)90280-A.
[26] A. Kömmling, M. Jaunich, D. Wolff, Effects of heterogeneous aging in compressed HNBR and EPDM O-ring seals, Polym. Degrad. Stab. 126 (2016) 39–46, http://dx.doi.org/
10.1016/j.polymdegradstab.2016.01.012.
[27] V. Vijayabaskar, V.K. Tikku, A.K. Bhowmick, Electron beam modification and crosslinking: influence of nitrile and carboxyl contents and level of unsaturation on structure
and properties of nitrile rubber, Radiat. Phys. Chem. 75 (2006) 779–792, http://dx.doi.org/10.1016/j.radphyschem.2005.12.030.
[28] Q. Wang, F. Wang, K. Cheng, Effect of crosslink density on some properties of electron beam-irradiated styrene–butadiene rubber, Radiat. Phys. Chem. 78 (2009)
1001–1005, http://dx.doi.org/10.1002/app.40296.
[29] S. Akhlaghi, A.M. Pourrahimi, M.S. Hedenqvist, C. Sjöstedt, M. Bellander, U.W. Gedde, Degradation of carbon-black-filled acrylonitrile butadiene rubber in alternative fuels:
transesterified and hydrotreated vegetable oils, Polym. Degrad. Stab. 123 (2016) 69–79, http://dx.doi.org/10.1016/j.polymdegradstab.2015.11.019.
[30] S. Akhlaghi, A.M. Pourrahimi, C. Sjöstedt, M. Bellander, M.S. Hedenqvist, U.W. Gedde, Effects of ageing conditions on degradation of acrylonitrile butadiene rubber filled
with heat-treated zno star-shaped particles in rapeseed biodiesel, Polym. Degrad. Stab. 138 (2017) 27–39, http://dx.doi.org/10.1016/j.polymdegradstab.2017.02.011.

11