Influence of crosslinker and water on cyclic properties of carboxylated nitrile butadiene rubber (XNBR)

Polymer Testing 67 (2018) 309–321

Contents lists available at ScienceDirect

Polymer Testing journal homepage: www.elsevier.com/locate/polytest

Material Properties

Influence of crosslinker and water on cyclic properties of carboxylated nitrile butadiene rubber (XNBR)

T

Doris Darya Fleischmanna,∗, Florian Arbeiterb, Raimund Schallerc, Armin Holznerc, Wolfgang Kerna,d, Sandra Schlögla a

Polymer Competence Center Leoben GmbH, Roseggerstraße 12, 8700 Leoben, Austria Materials Science and Testing of Polymers, Department of Polymer Engineering and Science, Montanuniversitaet Leoben, Otto Glöckel-Straße 2, 8700 Leoben, Austria c Semperit Technische Produkte GmbH, Triester Bundesstraße 26, Wimpassing 2632, Austria d Chemistry of Polymeric Materials, Department of Polymer Engineering and Science, Montanuniversitaet Leoben, Otto Glöckel-Straße 2, 8700 Leoben, Austria b

A R T I C LE I N FO

A B S T R A C T

Keywords: Carboxylated nitrile butadiene rubber XNBR Network structure Water exposure Cyclic testing Hysteresis

This study evaluates the effect of network structure (ionic versus covalent crosslinks) on the mechanical behavior of carboxylated nitrile butadiene rubber (XNBR) under cyclic loading. Additionally, the influence of water on the mechanical properties of the XNBR networks is assessed. Reversible and dissipated energies of cyclic tests (with constant maximum strain in each cycle) and step cycle tests (with increasing maximum strain in each cycle) of non-swollen and swollen samples are determined as a function of the used crosslinking system. Energies obtained from step cycle tests are compared to total absorbed energies obtained from monotonic tensile tests. The results show a good correlation between network structure and cyclic behavior. Ionic rearrangement processes are found to be more pronounced in samples with higher ionic crosslinker content. The introduction of covalent crosslinks imparts spatial restriction to ionic rearrangement and improves the cyclic performance of XNBR films upon water exposure.

1. Introduction The introduction of ionized (or ionizable) groups into rubbers allows for tailoring the physical and mechanical properties in a wide range. So-called ionomers exhibit thermoplastic behavior and rehealing properties upon temperature increase [1–3]. One example of an ionomer is carboxylated nitrile butadiene rubber (XNBR), which possesses ionizable carboxylic acid groups in addition to acrylonitrile units and carbon-carbon double bonds. Due to its high polarity, it is typically used in oil-resistant applications such as seals and hoses, or it is applied as memory shape material. In latex form, typical applications include gloves, binders and coatings [4–8]. The acrylonitrile content directly correlates with the glass transition temperature and polarity of the rubber and thus allows for the tuning of physical and mechanical properties. Along with the co-monomer content and the molecular weight distribution of the rubber, the final material behavior is also governed by the network structure, i.e. type and number of crosslinks [7]. The structure of XNBR enables the formation of both covalent and ionic crosslinks. In particular, covalent crosslinks are generated across the carbon-carbon double bonds of the butadiene units in a thermal

∗

Corresponding author. E-mail address: darya.fl[email protected] (D.D. Fleischmann).

https://doi.org/10.1016/j.polymertesting.2018.03.021 Received 12 January 2018; Received in revised form 5 March 2018; Accepted 11 March 2018 Available online 12 March 2018 0142-9418/ © 2018 Elsevier Ltd. All rights reserved.

curing process by using peroxides or sulfur (in combination with accelerators), or by photo-polymerization with multifunctional thiols as crosslinker. Furthermore, covalent links can be introduced across the carboxylic acid groups by using polyamine salts, polyisocyanates, polyepoxides or polyols [9,10]. Regarding the formation of ionic bonds, typical curing agents are ZnO and MgO, which not only form divalent crosslinks between two carboxylate groups but also aggregate to socalled multiplets (or clusters) consisting of several ion-counterion units. These ionic clusters are large crosslinks and reach diameters of up to 10 nm [11–16]. It was shown that not only divalent but also monovalent ions undergo cluster formation and can therefore be used as crosslinker [16]. Clusters are able to undergo ionic rearrangement processes and thus significantly influence material properties [1–3]. In preceding work, the effect of network structure on the static performance of XNBR was investigated [17]. In engineering applications, failure of rubber goods usually does not occur within one single loading but rather results from crack growth under cyclic loading [18]. Thus, in this study, focus was put on the effect of network structure on the cyclic behavior of XNBR. Elastically stored and dissipated energies of XNBR samples were determined in dependence on the network structure and compared with corresponding energies obtained from

Polymer Testing 67 (2018) 309–321

D.D. Fleischmann et al.

monotonic tensile tests (which can be seen as single-cycle tests). Elastomers exhibit a characteristic stress-softening effect upon cyclic loading, which has been first investigated in detail by Holt for filled elastomers in the 1930s [19] and was analyzed thoroughly by Mullins in the 1940s [20,21]. Harwood [22] found in the 1960s that both unfilled and filled elastomers exhibit such a hysteretic behavior. This characteristic behavior of elastomers is of high importance for material properties such as abrasion, tear resistance and heat build-up. It is caused by several mechanisms involving (i) breakage of filler agglomerates, (ii) bond rupture (intensified by the addition of filler, which causes local stress concentrations), (iii) orientation of polymer chains upon stretching followed by closer package and/or crystallization, (iv) internal friction, (v) molecule slipping, disentanglements, network rearrangements (irreversible processes) as well as (vi) viscoelastic behavior of elastomers (time-dependent deformation) [18,20–26]. The exposure of elastomers to liquids causes enormous changes in the network structure and material behavior due to the diffusion of molecules of the liquid into the elastomeric bulk, which leads to swelling and forces the polymer chains apart [17,27–29]. The investigation of the hysteretic nature of swollen elastomers and hydrogels is described in several articles. Webber et al. performed cyclic tests on double-network hydrogels which showed a large hysteresis loop during the first cycle followed by smaller loops in subsequent cycles. This effect was attributed to the existence of dissipative processes and was published for the first time in polymer gels [30]. Another study focused on cyclic testing of silica hydrogel hybrids [31]. Chai et al. [32–34] developed a compression device to subject swollen elastomers to cyclic testing. The investigation of the swelling behavior of nitrile butadiene rubber (NBR) and polychloroprene rubber (CR) in biodiesel revealed a reduction of the strength of elastomers upon swelling. In further investigations [35], a model was developed to describe the Mullins effect of swollen elastomers and experimentally tested with NBR, showing a good agreement of the proposed model with experimental results. Cantournet at el. designed a 3D model to display dissipative processes that occur during cyclic testing, which was experimentally proven by testing natural rubber (NR) and styrene butadiene rubber (SBR) [23]. As the use of XNBR as glove material, binder or coating requires resistance to water [36–38], the effect of water exposure on the cyclic performance of XNBR comprising different crosslinker was assessed. The results show a strong dependence of the cyclic properties of XNBR films on network structure and water exposure, thus evidencing the importance of network design on the material performance of final products.

Table 1 Composition of XNBR samples. Sample name

XNBR (phr)

KOH (phr)

ZnO (phr)

TMPMP (phr)

Sulfur (phr)

XNBR-KOH XNBR-0.5ZnO XNBR-ZnO-S XNBR-ZnOthiol XNBR-2ZnO

100 100 100 100

1.8 1.8 1.8 1.8

– 0.5 0.5 0.5

– – – 1

– – 1 –

100

1.8

2

–

–

Fig. 1. Cyclic loading to a constant strain for ten times (inset); illustration of energy components of one hysteresis (the sum of dissipated and elastically stored energy is the total absorbed energy).

2. Experimental 2.1. Materials and chemicals XNBR latex (Nipol LX556) with a pH value of 8.5, an acrylonitrile content of 31% and a dry rubber content of 45 wt.-% was supplied by ZEON (Japan). Aqueous sulfur and zinc oxide dispersions were provided by Semperit Technische Produkte (Austria). The photoinitiator ethyl-2,4,6-trimethyl-benzoyl-phenylphosphinate (Irgacure TPO-L) was obtained from BASF (Germany), the trifunctional thiol trimethylolpropane tris-3-mercaptopropionate (TMPMP) was purchased from Bruno Bock Thiochemicals (Germany) and the antioxidant Ralox LC was supplied by Solvadis (Germany). The emulsifier Tween®20 and potassium hydroxide (≥85%) were purchased from Sigma Aldrich (US). All chemicals were used without further purification. The compositions of the XNBR samples that were used in this work are given in Table 1. 2.2. Preparation of cured XNBR films Fig. 2. Comparison between stress-strain curve (monotonic loading) and step cycle tests; (a) the unloading curve obtained by step cycle testing is linearly interpolated up to the point of intersection with the stress-strain curve; (b) energy components.

2.2.1. Thermal pre-curing of XNBR latex The pH value of the XNBR latex was adjusted to 10.2 by adding a KOH solution (1 wt.-%) under stirring. For the ionic crosslinking, 0.5 310

Polymer Testing 67 (2018) 309–321

D.D. Fleischmann et al.

Fig. 3. XNBR networks with different crosslink structures: (a) Polysulfidic links, (b) thioether links, and cluster formation with (c) monovalent and (d) divalent cations.

repeatability. Non-swollen (i.e., untreated) XNBR samples and samples that were subjected to stress-free swelling in deionized water for 4 h were analyzed. This swelling time was chosen due to the similar rate of water uptake for all samples within this time period [17]. Swollen samples were patted with paper towels and immediately subjected to analysis after their withdrawal from water. Mechanical properties of swollen samples were determined by using the cross-section area of the specimens prior to water treatment (according to ASTM D741-06:15.2.1 [42]). Tensile tests were conducted to monitor stress-strain curves and to determine tensile strengths and elongations at break. Cyclic tests were performed to investigate the material behavior upon cyclic loading. Specimens were stretched to a constant strain for ten times. The strain of 600% was chosen as thin films are stretched to high strains in many applications. Each cycle was followed by unloading to 0.01 N. An exemplary hysteresis loop is shown in Fig. 1, which illustrates the energy components of one hysteresis that consists of elastically stored (i.e. reversible) energy and dissipated energy. The total absorbed energy equals the sum of the two components. The energy components were obtained by integrating the loading and unloading curves. Step cycle tests were conducted by stretching specimens to a strain εx (100⋅x %), with x being the cycle number, until break occurred. Again, each cycle was followed by unloading to 0.01 N. The results of step cycle tests were compared to stress-strain curves obtained from monotonic loading by linearly interpolating the unloading curve of each hysteresis up to the point of intersection with the tensile curve (shown in Fig. 2a). The energy components were determined via integration, and the total energy absorbed via tensile testing up to the point of intersection was compared to the energy components of the corresponding hysteresis (displayed in Fig. 2b). The calculations are based on Ref. [43].

and 2 phr ZnO were added to the latex mixtures (except for the reference XNBR-KOH). The used ZnO contents were below the stoichiometric ratio between carboxylic groups and Zn2+ ions. For the sulfur pre-curing, a sulfur dispersion containing elemental sulfur (1 phr) and accelerators were also added to the XNBR latex. All latex mixtures comprising ZnO were pre-cured for 3 h at 50 °C under stirring. After precure of the latex mixture XNBR-ZnO-thiol, photochemicals (1 phr TMPMP, 1 phr photoinitiator and 0.1 phr emulsifier) were added and the mixture was stirred for 1.5 h at room temperature. 2.2.2. Preparation of thin films A conventional coagulant dipping process was used to prepare thin XNBR films. Porcelain formers were immersed into a coagulant bath which contained calcium salts (coagulant), calcium carbonate (release agents) and surfactants, and then heated for 1 min to 100 °C to dry the coagulant. Subsequently, the formers were dipped into the latex mixture for 20 s and dried for 15 min at 100 °C. The films were stripped from the formers after cooling to room temperature. The film thickness ranged from 90 to 140 μm with the applied dipping parameters. 2.2.3. Post-curing with UV-light XNBR-ZnO-thiol samples were photo-cured as described in Ref. [39] by using the radical-mediated thiol-ene reaction which is initiated upon UV exposure [40]. The post-curing reaction was carried out with a medium pressure Ga-doped Hg lamp (Heraeus) under air with a light intensity of 15.7 J cm−2. 2.2.4. Tensile testing and cyclic testing Tensile tests and cyclic tests were performed with a universal testing machine (Instron, model 5500) at 23 °C and a low, constant strain rate of 0.083 s−1 (corresponding to a crosshead speed of 250 mm min−1) was used to avoid temperature increase of the specimens, as thermal effects were not taken into account in this study. Dumbbell shaped specimens were produced with a cutting die according to ASTM D412-D [41]. A minimum of three specimens were used in each test to ensure 311

Polymer Testing 67 (2018) 309–321

D.D. Fleischmann et al.

Fig. 4. Influence of (a) ionic crosslinker content and (b) introduction of covalent crosslinks on the cyclic behavior of untreated XNBR samples.

Table 2 Reversible and dissipated energies per cross-section area obtained from cyclic tests of untreated XNBR samples. Sample name

Reversible energy (MPamm)

Cycle number

1

2

5

10

1

2

5

10

XNBR-KOH

305.1 ± 22.0 457.7 ± 10.6 1162.1 ± 116.5 537.4 ± 9.5 677.7 ± 23.6

238.5 ± 19.0 353.5 ± 6.3 939.1 ± 77.8 412.4 ± 7.7 508.9 ± 16.4

179.4 ± 15.1 266.7 ± 5.2 782.6 ± 45.1 316.0 ± 7.0 388.8 ± 14.1

143.0 ± 13.0 216.3 ± 4.5 696.3 ± 30.6 261.5 ± 6.4 322.2 ± 13.5

254.4 ± 21.9 465.5 ± 25.7 1282.7 ± 79.1 618.6 ± 9.3 858.3 ± 76.6

87.0 ± 6.5 148.7 ± 4.9 374.9 ± 30.2 177.8 ± 4.8 229.8 ± 8.4

47.3 ± 4.5 84.1 ± 1.9 240.0 ± 13.2 100.4 ± 3.1 128.1 ± 4.1

33.2 ± 3.1 60.6 ± 1.5 195.1 ± 7.1 73.4 ± 2.0 92.1 ± 3.3

XNBR-0.5ZnO XNBR-ZnO-S XNBR-ZnO-thiol XNBR-2ZnO

Dissipated energy (MPamm)

312

Polymer Testing 67 (2018) 309–321

D.D. Fleischmann et al.

Fig. 5. Percentage of reversible to total absorbed energy of each hysteresis obtained from cyclic tests of untreated XNBR samples.

of the network density, which was verified by equilibrium swelling experiments and NMR analysis in previous work [17]. This behavior is related to the so-called hopping effect [3,44], which implies ionic rearrangements upon loading. The energy for ion exchange is believed to be lower than the energy required for complete charge separation between the two ion pairs, thus the ionic crosslinks have a higher probability to persist during cyclic testing [45]. The hopping effect is more pronounced in samples with higher amount of ionic crosslinker. Thus, reversible and dissipated energies of all ten cycles follow the trend XNBR-KOH > XNBR-0.5ZnO > XNBR-2ZnO. Most of the ionic rearrangement processes are assumed to be reversible, thus reverted upon stress removal, which explains the high ratio of reversible to total energy of all samples. The ratio of reversible to total energy in the tenth cycle amounts to 81.1% for XNBR-KOH samples. Contrary, the energy ratio decreases with increasing content of ionic crosslinker (to 78.1 and 77.8% for XNBR-0.5ZnO and XNBR-2ZnO samples, respectively), which indicates a slightly higher percentage of irreversible, ionic rearrangement processes. Ionic rearrangement is strongly influenced by the presence of covalent links [3]. As seen in previous analysis [17], the introduction of mono- and polysulfidic links or thioether links increases the crosslink density of XNBR and reduces the hopping effect due to spatial restriction of ionic movements. An increase of reversible and dissipated energies is observed with increasing crosslink density, the highest values being obtained with XNBR-ZnO-S. Regarding the ratio of reversible to total energy, similar values are obtained with XNBR-0.5ZnO and XNBRZnO-thiol. Contrary, an increase of the energy ratio is observed upon sulfur vulcanization (increase from 78.1 to 78.9% in the tenth cycle) which indicates that ions return to their initial arrangement more easily after stress removal due to the spatial restriction. This signifies better cyclic performance, as the amount of energy dissipated during cyclic loading is reduced. As the crosslink density of XNBR-ZnO-S is significantly higher than the crosslink density of XNBR-ZnO-thiol [17], it can be concluded that the amount of covalent linkages needs to be high enough to have a pronounced effect on the cyclic behavior of XNBR and increase the ratio of reversible to total energy.

Table 3 Stress-strain properties of untreated XNBR samples. Sample name

Tensile strength (MPa)

Elongation at break (%)

XNBR-KOH XNBR-0.5ZnO XNBR-ZnO-S XNBR-ZnO-thiol XNBR-2ZnO

19.8 39.4 39.9 26.5 41.4

995 897 672 852 830

± ± ± ± ±

1.4 2.5 2.4 2.3 3.8

± ± ± ± ±

18 14 22 25 41

3. Results and discussion 3.1. Influence of crosslinker type on cyclic behavior of XNBR films The curing procedures described in chapter 2.2 yielded XNBR networks with four different crosslink structures (see Fig. 3). Covalent crosslinks were generated upon sulfur vulcanization or photo-crosslinking, yielding mono- and polysulfidic links or thioether links, respectively. The presence of carboxylate groups was confirmed via FTIR spectroscopy in previous work [17], which are capable of cluster formation with potassium and zinc ions.

3.1.1. Cyclic loading to constant strain Cyclic tests showed that for all analyzed XNBR samples, a stressdrop occurs upon proceeding cycles, which is most significant after the first cycle (see Fig. 4). Consequently, the reversible and dissipated energy components are highest in the first cycle (listed in Table 2). As all samples are subjected to the same elongation of 600% in each cycle, the probability of bond breakage and material softening is highest during the first stretching [20,21]. Thus, the main part of dissipative processes (e.g. disentangling, chain breakage) occurs in the first cycle, which explains the low ratio of reversible to total energy in the first hysteresis (see Fig. 5). The decrease of reversible energies from the first to the second cycle is between 19.2 and 24.9% of initial values whilst the decrease of dissipated energies ranges from 65.8 to 73.2%. After about five cycles, all samples reach a state with repeatable hysteresis loops and a slightly changing ratio of reversible to dissipated energy, with the reversible energy component increasing steadily up to the tenth cycle. The increase of the amount of ionic crosslinker leads to an increase

3.1.2. Tensile tests versus step cycle tests The results of the tensile tests are summarized in Tab. 3 and show 313

Polymer Testing 67 (2018) 309–321

D.D. Fleischmann et al.

Fig. 6. Monotonic testing versus step cycle testing of untreated XNBR samples to assess the influence of (a) ionic crosslinker content and (b) introduction of covalent crosslinks.

step cycle tests (see Fig. 6). Stress levels in loading cycles decrease with proceeding testing which is related to the Mullins effect [23,33]. Up to a strain of 200% to 300%, maxima of step cycle tests overlap with stressstrain curves of monotonic tests. However, with increasing maximum strain, progressive stress-softening due to irreversible material destruction (i.e. disentangling, bond breakage) is observed which is accompanied by a significant deviation from tensile curves. In addition, rupture of samples subjected to step cycle tests occurs at higher strains compared to monotonic tensile tests. Fig. 6a shows that a better matching between the step cycle maxima and the corresponding tensile curve (XNBR-0.5ZnO vs. XNBR-2ZnO) is obtained for samples with rising number of ionic crosslinks, which goes along with more pronounced hopping effect. The presence of covalent links reduces stresssoftening and the samples XNBR-ZnO-thiol and XNBR-ZnO-S show better matching between step cycle maxima and tensile curves with respect to XNBR-0.5ZnO (see Fig. 6b). This trend is related to an increase of the crosslink density on the one hand and the restriction of the

that XNBR-KOH samples exhibit the lowest stress-strain properties. This was expected due to the weak monovalent ionic interactions between potassium and carboxylate ions. The stress-strain behavior is improved by the addition of ZnO because the divalent ionic crosslinker is not only capable of cluster formation but also leads to the generation of zinc dicarboxylate crosslinks [16]. The combined use of ZnO and thiol decreases the tensile strength and the elongation at break with regard to XNBR-0.5ZnO, as the covalent crosslinks generated upon thiol addition impart hindrance to ionic rearrangements. The restriction of the hopping effect by sulfur vulcanization is compensated by the comparatively high density of covalent crosslinks in XNBR-ZnO-S, which results in a reduction of the elongation at break from 897 to 672% whilst maintaining high tensile strength. These trends correspond to the trends seen in our previous work [17], but the absolute values differ due to differences in testing parameters. Stress-strain curves obtained from monotonic loading (which can be seen as single-cycle tests) were compared with hystereses received from 314

Polymer Testing 67 (2018) 309–321

D.D. Fleischmann et al.

Fig. 7. Percentage of reversible to total energy of each hysteresis obtained from step cycle tests of untreated XNBR samples.

Fig. 8. Reversible energy of each hysteresis obtained from step cycle tests in relation to the total absorbed energy obtained from tensile tests of untreated XNBR samples.

their high crosslink density. Upon increase of the strain, disentangling and chain slippage are restricted due to the presence of permanent covalent crosslinks, which causes bond breakage and chain disentangling already during the first few cycles and results in increasing percentage of dissipated energy with proceeding testing. The difference of energy ratios in dependence on the used crosslinker is more pronounced than in cycle tests to constant strain. A decrease of the energy ratio with increasing content of ionic crosslinker is observed, which was also seen in cyclic testing to constant strain and is related to the higher amount of irreversible ionic movements. Contrary to cyclic testing to constant strain, the introduction of covalent links does not increase the energy ratio compared to XNBR-0.5ZnO. The covalent links rather impart hindrance to ionic movements and as the strain is increased in each cycle, bond breakage occurs in every cycle,

hopping effect on the other hand, as ions return to their initial arrangement more easily after stress removal due to spatial restriction introduced by the presence of covalent crosslinks. The ratio of reversible to total energy of each hysteresis of untreated XNBR samples obtained by step cycle tests is displayed in Fig. 7. For most of the XNBR samples, energy ratios reach a maximum in dependence on the cycle number between the third and fifth cycle. The low energy ratio in the beginning is attributed to dissipative processes such as irreversible disentangling and chain slippage, whereas at higher strains, reversible processes (e.g. stretching, ion hopping) occur. Further increase of the strain upon proceeding testing results in bond breakage and thus decreases the ratio of reversible to total energy. An exception of this trend are XNBR-2ZnO and XNBR-ZnO-S which reach the highest percentage of reversible energy in the first cycle owing to 315

Polymer Testing 67 (2018) 309–321

D.D. Fleischmann et al.

Fig. 9. Influence of (a) ionic crosslinker content and (b) introduction of covalent crosslinks on the cyclic behavior of swollen (4 h, water) XNBR samples.

Table 4 Reversible and dissipated energies per cross-section area obtained from cyclic tests of swollen (4 h, water) XNBR samples. Sample name

Reversible energy (MPamm)

Cycle number

1

2

5

10

1

2

5

10

XNBR-KOH

87.8 ± 8.9 220.7 ± 16.6 523.0 ± 22.8 269.5 ± 21.7 238.2 ± 8.0

57.4 ± 6.6 173.7 ± 14.2 456.3 ± 19.8 215.9 ± 20.7 190.1 ± 7.0

35.2 ± 6.1 127.0 ± 13.4 401.6 ± 15.5 170.4 ± 17.8 142.2 ± 5.6

28.3 ± 7.5 102.7 ± 10.4 373.6 ± 13.4 145.7 ± 17.5 119.9 ± 3.6

109.4 ± 0.9 179.7 ± 9.6 370.4 ± 24.8 226.9 ± 12.6 194.6 ± 6.1

29.4 ± 2.4 70.2 ± 5.0 148.9 ± 7.3 83.4 ± 6.5 76.5 ± 1.9

10.5 ± 1.4 38.4 ± 2.3 107.1 ± 5.2 47.2 ± 5.2 43.8 ± 1.2

6.1 ± 1.5 28.2 ± 2.4 94.5 ± 3.8 37.3 ± 4.8 32.2 ± 1.7

XNBR-0.5ZnO XNBR-ZnO-S XNBR-ZnO-thiol XNBR-2ZnO

Dissipated energy (MPamm)

316

Polymer Testing 67 (2018) 309–321

D.D. Fleischmann et al.

Fig. 10. Percentage of reversible to total absorbed energy of each hysteresis obtained from cyclic tests of swollen (4 h, water) XNBR samples.

testing aimed at the investigation of the material properties of XNBR samples exposed to water for a time period of 4 h. Water leads to plasticization due to physical interactions with the polar rubber matrix on the one hand and promotes structural rearrangement processes of ionic linkages and clusters on the other hand. The exposure of XNBR elastomers to water strongly influences structural and mechanical properties [17], and a significant influence can therefore be expected regarding the cyclic performance.

Table 5 Stress-strain properties of swollen (4 h, water) XNBR samples. Sample name

XNBR-KOH XNBR-0.5ZnO XNBR-ZnO-S XNBR-ZnOthiol XNBR-2ZnO a

Tensile strength

Elongation at break

(MPa)

Normalizeda (%)

(%)

Normalizeda (%)

2.6 ± 0.1 5.3 ± 0.2 14.0 ± 1.3 6.5 ± 0.4

13.1 13.5 35.1 24.5

1191 ± 44 920 ± 4 740 ± 25 846 ± 4

119.7 102.6 110.1 99.3

6.3 ± 0.2

15.2

846 ± 16

102.9

3.2.1. Cyclic loading to constant strain The cyclic properties of XNBR are strongly affected by exposure to water as displayed in Fig. 9. It is obvious that swollen samples preserve their hysteretic nature, but they show significantly lower stress values. As a result, reversible and dissipated energies (listed in Tab. 4) are reduced with regard to their original, non-swollen counterparts. Regarding the tenth cycle, with respect to the corresponding non-swollen samples, reversible and dissipated energies of swollen samples are reduced about 44.3–80.2% and 49.2–81.6%, respectively. Higher reduction corresponds to XNBR films that are only ionically crosslinked whereas the introduction of covalent crosslinks increases the retention of material performance upon water exposure. In general, swollen samples have higher ratios of reversible to total energy of the local hysteresis than their non-swollen counterparts (see Fig. 10). The increase of energy ratios upon water exposure shows a positive correlation with the crosslink density. A pronounced increase is seen in the first cycle (5.5–11.1%) whereas energy ratios of non-swollen and swollen films approach each other with proceeding cyclic testing, which results in differences of only 0.8–3.6% in the tenth cycle. It is assumed that the plasticizing effect of water facilitates chain slippage and thus reduces the percentage of dissipated energy in swollen films. After about five cycles, a stable state with repeatable hysteresis loop is obtained and the similar energy ratios of non-swollen and swollen XNBR films at higher cycle numbers show that the influence of water on the ratio of reversible to total energy of the local hysteresis diminishes with proceeding testing. An exception of this trend is XNBR-KOH with a considerable lower energy ratio of swollen films compared to nonswollen ones up to the fifth cycle. The high percentage of dissipated energy of swollen XNBR-KOH samples is related to the labile nature of monovalent crosslinks and the low network density which promotes irreversible structural rearrangements in presence of water.

Referred to tensile properties of non-swollen samples.

causing lower energy ratios. For industrial applications or quality testing of materials, it is of high interest to compare material behavior in relation to the original properties. Therefore, the reversible energy of each hysteresis obtained from step cycle tests is compared to the total absorbed energy obtained from tensile testing (see Fig. 8). The energy ratio decreases with increasing cycle number due to stress-softening on the one hand and the higher amount of dissipated energy due to bond breakage on the other hand. Furthermore, the slopes of the loading cycles continuously change from flat to steep, shifting the point of intersection with the tensile curve (described in section 2.2.4) to lower strains, which consequently reduces the percentage of reversible energy in relation to the total energy of the tensile curve. This effect becomes more obvious in samples with higher content of ionic crosslinker and decreases upon the introduction of covalent links, which is expected to restrict ionic movements at low strains. However, at higher strains, bond cleavage occurs. Consequently, in covalently crosslinked samples, the number of covalent links decreases with repeated stretching and the hopping effect of the ionic clusters becomes a dominating factor again. Thus, all samples with the same amount of ionic crosslinker (i.e. XNBR-0.5ZnO, XNBR-ZnO-S and XNBR-ZnO-thiol) show similar energy ratios after the third cycle. 3.2. Influence of exposure to water Apart from testing of material parameters under air, additional 317

Polymer Testing 67 (2018) 309–321

D.D. Fleischmann et al.

Fig. 11. Monotonic testing versus step cycle testing of swollen (4 h, water) XNBR samples to assess the influence of (a) ionic crosslinker content and (b) introduction of covalent crosslinks.

In a further step, stress-strain curves obtained from tensile tests were compared with step cycle tests (see Fig. 11). Similar to untreated XNBR samples, the maxima of step cycle tests coincide with the stressstrain curves at low strains. However, with increasing strain during testing, gradual stress-softening is observed which is accompanied by a deviation from tensile curves. Contrary to untreated samples, hystereses of swollen samples with higher content of ionic crosslinker do not show better approximation to tensile curves compared to swollen samples with lower amount of ionic bonds (see Fig. 11a). The strong effect of water on the labile monovalent ionic bonds in XNBR-KOH results in a pronounced decrease of the tensile performance, i.e. extremely low tensile strength and high elongation at break. In consequence, deviations of step cycle maxima from the tensile curve are only very small. Compared to fully ionically crosslinked samples, the introduction of covalent crosslinks does not only increase the static durability in water at stress levels below 3 MPa [17] but also significantly improves the

3.2.2. Tensile tests versus step cycle tests As observed in previous work [17], immersion in water causes a substantial deterioration of tensile properties. In particular, tensile strengths of swollen samples range from 13.1 to 24.5% of initial values (Tab. 5). Higher decrease is seen for fully ionically crosslinked XNBR films due to the labile nature of ionic bonds in the presence of water, whereas the introduction of covalent crosslinker increases the retained level of tensile strengths of swollen samples with respect to their nonswollen counterparts. Namely, with regard to initial values prior to water exposure, XNBR-ZnO-thiol and XNBR-ZnO-S reach relative tensile strengths of 24.5 and 35.1% in swollen state, respectively. Regarding elongations at break, relative changes upon swelling do not exceed 10.1% (referred to values of non-swollen films) except for XNBR-KOH with an increase of 19.1%, which is related to the low crosslink density which imparts lower hindrance to disentanglement and chain slippage [46]. 318

Polymer Testing 67 (2018) 309–321

D.D. Fleischmann et al.

Fig. 12. Percentage of reversible to total energy of each hysteresis obtained from step cycle tests of swollen (4 h, water) XNBR samples.

Fig. 13. Reversible energy of each hysteresis obtained from step cycle tests in relation to the total absorbed energy obtained from tensile tests of swollen (4 h, water) XNBR samples.

cyclic performance of XNBR films in presence of water (see Fig. 11b). The improvement of cyclic properties of swollen XNBR films is, together with the tensile performance, in direct correlation with the content of covalent links, i.e. better cyclic performance is achieved with XNBRZnO-S than with XNBR-ZnO-thiol in swollen state. This finding is in contrast to the static durability of XNBR which does not show a positive correlation with the density of covalent crosslinks but rather depends on the chemical structure of crosslinks [17]. The ratio of reversible to total energy of each hysteresis of swollen XNBR samples obtained from step cycle tests is displayed in Fig. 12. Energy ratios of all samples reach a maximum in the fourth or fifth cycle, including XNBR-2ZnO and XNBR-ZnO-S which show different curve characteristics in non-swollen state. This is related to the plasticization effect of water, as chain slippage and disentangling, which is

restricted due to the presence of covalent links in non-swollen state, is facilitated in swollen state. An exception of this trend is XNBR-KOH where the maximum is shifted to higher strains and reached in the ninth cycle. Contrary to non-swollen XNBR films, the difference of energy ratios in dependence on the crosslinker is less pronounced in swollen films, i.e. no remarkable change is observed by going from 0.5 to 2 phr ZnO and by introduction of thioether bonds, whereas a significant increase of the percentage of reversible energy is obtained by sulfur vulcanization. As already observed in cyclic tests to constant strain (chapter 3.2.1), the percentage of reversible energy is higher for swollen films than for their non-swollen counterparts. This is related on the one hand to the fact that dissipative processes can already occur during the immersion step, and on the other hand to the facilitation of structural 319

Polymer Testing 67 (2018) 309–321

D.D. Fleischmann et al.

rearrangements in presence of water. The reversible energy of each hysteresis calculated from step cycle tests is compared to the total absorbed energy obtained from tensile testing in Fig. 13. As already observed in non-swollen samples, energy ratios decrease with increasing cycle number which is related to stresssoftening and bond breakage. The energy ratio increases with rising amount of ionic crosslinker and shows significant further increase upon sulfur vulcanization. Compared to non-swollen samples, the results reveal that swollen samples have higher energy ratios than their nonswollen counterparts, which is related to the lower deviation of step cycle maxima with tensile curves on the one hand and to a higher percentage of reversible energy of swollen samples per se on the other hand due to the presence of water, as also seen in cyclic tests to constant strain. An exception is XNBR-KOH, which exhibits lower energy ratios in swollen than in non-swollen state. The presence of water can cause physical crosslinks (i.e., entanglements that increase the restoring force and thus the percentage of reversibility in non-swollen state) to disentangle, in particular in XNBR-KOH films due to their low crosslink density. In absence of physical network points in swollen XNBR-KOH, the low density of monovalent ionic crosslinks diminishes the reversibility of structural changes during cyclic testing. Disentangling and chain slippage are promoted in the presence of water and due to the low restoring force caused by the low crosslink density of XNBR-KOH, the reversible energy of the hysteresis in relation to the total energy of the tensile curve is comparatively low.

slippage and thus reduces the percentage of dissipated energy. Mechanical properties of XNBR undergo enormous decrease upon exposure to water due to the labile nature of ionic crosslinks, but significant improvement is achieved by the introduction of covalent crosslinks, which not only increase tensile properties but also improve the cyclic performance of swollen films. Both tensile and cyclic properties of swollen XNBR films are in direct correlation with the content of covalent crosslinks. As a result, the best cyclic performance of swollen XNBR films is obtained by sulfur vulcanization (XNBR-ZnO-S). These results clearly reveal that proper design of the network structure of XNBR elastomers is crucial to improve not only the cyclic performance but also the resistance to water.

4. Conclusion

[1] R.J. Wojtecki, M.A. Meador, S.J. Rowan, Using the dynamic bond to access macroscopically responsive structurally dynamic polymers, Nat. Mater. 10 (2011) 14–27, http://dx.doi.org/10.1038/nmat2891. [2] W.J. MacKnight, R.D. Lundberg, Elastomeric ionomers, Rubber Chem. Technol. 57 (1984) 652–663, http://dx.doi.org/10.5254/1.3536023. [3] I. Mora-Barrantes, M.A. Malmierca, J.L. Valentín, A. Rodriguez, L. Ibarra, Effect of covalent cross-links on the network structure of thermo-reversible ionic elastomers, Soft Matter 8 (2012) 5201, http://dx.doi.org/10.1039/c2sm06975j. [4] J.R. Dunn, R.G. Vara, Oil resistant elastomers for hose applications, Rubber Chem. Technol. 56 (1983) 557–574, http://dx.doi.org/10.5254/1.3538141. [5] K. Hashimoto, A. Maeda, K. Hosoya, Y. Todani, Specialty elastomers for automotive applications, Rubber Chem. Technol. 71 (1998) 449–519, http://dx.doi.org/10. 5254/1.3538491. [6] K. Pal, T. Das, S.K. Pal, C.K. Das, Use of carboxylated nitrile rubber and natural rubber blends as retreading compound for OTR tires, Polym. Eng. Sci. 48 (2008) 2410–2417, http://dx.doi.org/10.1002/pen.21196. [7] F. Röthemeyer, F. Sommer, Kautschuktechnologie, second ed., Carl Hanser Verlag, Munich, Germany, 2006. [8] A. González-Jiménez, M.A. Malmierca, P. Bernal-Ortega, P. Posadas, R. PérezAparicio, Á. Marcos-Fernández, P.T. Mather, J.L. Valentín, The shape-memory effect in ionic elastomers: fixation through ionic interactions, Soft Matter 13 (2017) 2983–2994, http://dx.doi.org/10.1039/C7SM00104E. [9] H.P. Brown, Crosslinking reactions of carboxylic elastomers, Rubber Chem. Technol. 36 (1963) 931–962, http://dx.doi.org/10.5254/1.3539642. [10] C.E. Hoyle, C.N. Bowman, Thiol-ene click chemistry, Angew. Chemie - Int. Ed. 49 (2010) 1540–1573, http://dx.doi.org/10.1002/anie.200903924. [11] S. Schlick, Ionomers: Characterization, Theory, and Applications, CRC Press, 1996. [12] A. Eisenberg, B. Hird, R.B. Moore, A new multiplet-cluster model for the morphology of random ionomers, Macromolecules 23 (1990) 4098–4107, http://dx. doi.org/10.1021/ma00220a012. [13] L. Ibarra, M. Alzorriz, Vulcanization of carboxylated nitrile rubber (XNBR) by a mixed zinc peroxide-sulphur system, Polym. Int. 49 (2000) 115–121. [14] L. Ibarra, M. Alzorriz, Ionic elastomers based on carboxylated nitrile rubber and magnesium oxide, J. Appl. Polym. Sci. 103 (2007) 1894–1899. [15] U.K. Mandal, Ionic elastomer based on carboxylated nitrile rubber: infrared spectral analysis, Polym. Int. 49 (2000) 1653–1657. [16] A. Kells, B. Groves, Crosslinking in carboxylated nitrile rubber dipped films, latex, 4th International Conference on Latex and Latex Based Products, Conference Proceedings, Frankfurt, Germany, January 24–25, 2006. [17] D.D. Fleischmann, S. Ayalur-Karunakaran, F. Arbeiter, R. Schaller, A. Holzner, W. Kern, S. Schlögl, Influence of crosslinker and water on mechanical properties of carboxylated nitrile butadiene rubber (XNBR), Polym. Test. 66 (2018) 24–31, http://dx.doi.org/10.1016/j.polymertesting.2018.01.001. [18] A. Gent, Engineering with Rubber: How to Design Rubber Components, third ed., Hanser, Munich, Germany, 2012. [19] W.L. Holt, Behavior of rubber under repeated stresses, Rubber Chem. Technol. 5 (1932) 79–89, http://dx.doi.org/10.5254/1.3539319. [20] L. Mullins, N.R. Tobin, Theoretical model for the elastic behavior of filler-reinforced vulcanized rubbers, Rubber Chem. Technol. 30 (1947) 551–571. [21] L. Mullins, Effect of stretching on the properties of rubber, Rubber Chem. Technol. 21 (1948) 281–300. [22] J.A.C. Harwood, L. Mullins, A.R. Payne, Stress softening in natural rubber

Acknowledgments The research work was performed at the Polymer Competence Center Leoben GmbH (PCCL, Austria) within the framework of the COMET-program of the Federal Ministry for Transport, Innovation and Technology and Federal Ministry for Economy, Family and Youth, with contributions by the Department of Polymer Engineering and Science (Montanuniversitaet Leoben), and Semperit Technische Produkte GmbH. The PCCL is funded by the Austrian Government and the State Governments of Styria (854178), Lower Austria and Upper Austria. References

This study focused on the analysis of the mechanical behavior of XNBR films under cyclic loading in dependence on the network structure. It was shown that the hopping effect, which signifies the exchange of ions under stress, has an enormous effect on the material behavior in ionically crosslinked XNBR materials and becomes more pronounced with rising density of ionic crosslinks. Cyclic tests to constant strain revealed that reversible and dissipated energies show a positive correlation with the content of ionic crosslinker. The main part of ionic rearrangements is assumed to be reversible, which explains the high ratio of reversible to total energy of the local hysteresis of ionically crosslinked XNBR films. The introduction of mono- and polysulfidic links or thioether links increases the crosslink density of XNBR, which results in an increase of reversible and dissipated energies. XNBR-ZnO-S films that have the highest content of covalent crosslinks show improved cyclic performance with a higher percentage of reversible energy, suggesting that the return of ions to their initial arrangement after stress removal is facilitated due to the presence of covalent links that impart spatial restriction to ionic movements. The energy ratio which is obtained from the comparison of reversible energies from step cycle tests with total absorbed energies from tensile tests decreases with increasing cycle number due to stress-softening. This behavior is accompanied by an increasing deviation of step cycle maxima from tensile curves. This trend is more significant in samples with a lower density of ionic bonds and thus a lower extent of ionic rearrangements. The introduction of covalent crosslinks restricts ionic movements, resulting in better approach of step cycle maxima to tensile curves. The presence of water leads to plasticization due to physical interactions with XNBR, strongly governs structural rearrangement processes of ionic bonds and clusters and thus has strong impact on structural and mechanical properties of crosslinked XNBR, which was also observed in previous work [17]. Swollen XNBR samples preserve their hysteretic nature when subjected to cyclic testing but show significantly lower reversible and dissipated energies with regard to their non-swollen counterparts. In general, ratios of reversible to total energy of local hysteresis increase upon exposure to water, which positively correlates with the crosslink density. This finding is related to the plasticizing effect of water, which is assumed to facilitate ionic movements and chain 320

Polymer Testing 67 (2018) 309–321

D.D. Fleischmann et al.

[23]

[24] [25] [26] [27] [28] [29]

[30]

[31]

[32]

[33]

[34]

vulcanizates. Part II. Stress softening effects in pure gun and filler laoded rubbers, J. Appl. Polym. Sci. 9 (1965) 3011–3021. S. Cantournet, R. Desmorat, J. Besson, Mullins effect and cyclic stress softening of filled elastomers by internal sliding and friction thermodynamics model, Int. J. Solids Struct 46 (2009) 2255–2264, http://dx.doi.org/10.1016/j.ijsolstr.2008.12. 025. J. Diani, B. Fayolle, P. Gilormini, A review on the Mullins effect, Eur. Polym. J 45 (2009) 601–612, http://dx.doi.org/10.1016/j.eurpolymj.2008.11.017. J.A.C. Harwood, A.R. Payne, Stress softening in natural rubber vulcanizates. Part IV. Unfilled vulcanizates, J. Appl. Polym. Sci. 10 (1966) 1203–1211. T.J. Pond, Creep behavior of rubbers subjected to repeated loadings, J. Nat. Rubber Res. 4 (1989) 93–106. A. Gent, Engineering with Rubber, third ed., Carl Hanser Verlag, Munich, Germany, 2012. W.D. Callister, D.G. Rethwisch, Materials Science and Engineering. An Introduction, eighth ed., John Wiley & Sons, Inc., New York, US, 2010. W. Chassé, S. Schlögl, G. Riess, K. Saalwächter, Inhomogeneities and local chain stretching in partially swollen networks, Soft Matter 9 (2013) 6943, http://dx.doi. org/10.1039/c3sm50195g. R.E. Webber, C. Creton, H.R. Brown, J.P. Gong, Large strain hysterisis and Mullins effect of tough double network hydrogels, Macromolecules 40 (2007) 2919–2927, http://dx.doi.org/10.1021/ma062924y. W.C. Lin, W. Fan, A. Marcellan, D. Hourdet, C. Creton, Large strain and fracture properties of poly(dimethylacrylamide)/silica hybrid hydrogels, Macromolecules 43 (2010) 2554–2563, http://dx.doi.org/10.1021/ma901937r. A.B. Chai, A. Andriyana, E. Verron, M.R. Johan, A.S.M.A. Haseeb, Development of a compression test device for investigating interaction between diffusion of biodiesel and large deformation in rubber, Polym. Test 30 (2011) 867–875, http://dx.doi. org/10.1016/j.polymertesting.2011.08.009. A.B. Chai, A. Andriyana, E. Verron, M.R. Johan, Mechanical characteristics of swollen elastomers under cyclic loading, Mater. Des. 44 (2013) 566–572, http://dx. doi.org/10.1016/j.matdes.2012.08.027. A. Andriyana, A.B. Chai, E. Verron, M.R. Johan, Interaction between diffusion of palm biodiesel and large strain in rubber: effect on stress-softening during cyclic

[35]

[36]

[37] [38] [39]

[40]

[41] [42] [43]

[44]

[45]

[46]

321

loading, Mech. Res. Commun. 43 (2012) 80–86, http://dx.doi.org/10.1016/j. mechrescom.2012.03.004. A.B. Chai, A. Andriyana, S.Y. Ch, E. Verron, M.R. Johan, An extended two-phase model for Mullins effect in swollen rubber, 18 (2014) 224–227, http://dx.doi.org/ 10.1179/1432891714Z.0000000001036. B. Banerjee, Criteria for the selection of rubbers for coating: physical and chemical characteristics, in: G. Akovali (Ed.), Adv. Polym. Coat. Text, Smithers Rapra Technology Ltd, Shawbury, Shrewsbury, Shropshire, UK, 2012. S.A. Osemeahon, J.T. Barminas, Study of a composite from reactive blending of methylol urea resin with natural rubber, African J. Biotechnol 6 (2007) 810–817. D. Duncan, R. Duncan, R.W. Pewitt, Waterproof and high heat resistant coated gloves, US20090077704A1, 2009. D. Lenko, S. Schlögl, A. Temel, R. Schaller, A. Holzner, W. Kern, Dual crosslinking of carboxylated nitrile butadiene rubber latex employing the thiol-ene photoreaction, J. Appl. Polym. Sci. 129 (2013) 2735–2743, http://dx.doi.org/10.1002/app.38983. C.E. Hoyle, T.Y. Lee, T. Roper, Thiol-enes: chemistry of the past with promise for the future, J. Polym. Sci. Part A Polym. Chem. 42 (2004) 5301–5338, http://dx.doi. org/10.1002/pola.20366. ASTM D412, Standard Test Methods for Vulcanized Rubber and Thermoplastic Elastomers-tension, Beuth Verlag GmbH, Berlin, Germany, 2016. ASTM D471, Standard Test Method for Rubber Property-effect of Liquids, Beuth Verlag GmbH, Berlin, Germany, 2016. R. Calabrò, Mechanical Characterization of Elastomers Under Quasi-static and Dynamic Biaxial Loading Conditions, PhD Thesis (2013) Politecnico di Milano, Italy. W. Cooper, Copolymers of butadiene and unsaturated acids: crosslinking by metal oxides, J. Polym. Sci. 28 (1958) 195–206, http://dx.doi.org/10.1002/pol.1958. 1202811618. S.K. Chakraborty, A.K. Bhowmick, S.K. De, Structure-property relations of carboxylated nitrile rubber, J. Appl. Polym. Sci. 26 (1981) 4011–4020, http://dx.doi. org/10.1002/app.1981.070261204. F. Röthemeyer, F. Sommer, Kautschuktechnologie, third ed., Carl Hanser Verlag, Munich, Germany, 2013.