IMPROVING THE PROPERTIES OF NR/NBR BLEND BY INTRODUCING INTERFACIAL CROSSLINKS USING BISMALEIMIDE DURING THE INITIAL PHASE OF ACCELERATED SULFUR CURING
ABSTRACT
To develop a technologically compatible blend of NR and NBR is always a challenge due to their polarity mismatch. As a result, the physico-mechanical properties of their blends are generally poor. To address this issue, an attempt was made to increase the uniform distribution of crosslinks across the blend phases at the time of molding at 170°C. A cure composition consisting of sulfur (S) and a delayed action accelerator (N-cyclohexyl-2-benzothiazole sulfenamide [CBS]) has been designed to co-crosslink both phases of the blend simultaneously. The tensile properties, particularly the tensile strength (TS) of the blend cured by this method, were superior (∼371% greater) than the TS of the blend cured using a combination of S/CBS and an ultrafast accelerator (tetramethylthiuram disulfide [TMTD]). A bifunctional maleimide (Maleide F) was also used in conjunction with S/CBS in the curing recipe to further improve the distribution of sulfidic crosslinks by reducing the interfacial tension between the NR and NBR phases via Alder-ene reaction.
INTRODUCTION
Blending different rubbers is one of the most economically viable techniques to tailor the unique properties of individual rubbers for specific applications.1–4 For example, NR is known for its higher gum tensile strength (TS) than any other synthetic elastomer because of its unique strain-induced crystallization (SIC) behavior.5–9 Moreover, NR possesses good tackiness, green strength, and low-temperature flexibility. However, NR exhibits poor heat aging resistance and resistance to hydrocarbon oils. By contrast, nitrile rubber (NBR) is famous for its oil swelling resistance because of the presence of polar acrylonitrile (ACN) content.10 Therefore, NBR is the primary material of choice for products such as oil seals, gaskets, and oil transportation hoses.11 Furthermore, owing to polarity, NBR exhibits a greater dielectric constant (>10) than any other nonpolar rubbers. The high dielectric constant of NBR enables it to be used as an effective actuation material in the fabrication of artificial muscles.12,13 However, unlike NR, NBR lacks the ability to show SIC behavior due to its irregular microstructure. As a result, the TS of the unfilled NBR is often lower than that of the NR. Thus, the main intention of blending NR and NBR is to bring the above-mentioned unique qualities of these individual elastomers into their blend vulcanizate to show high TS, good low-temperature flexibility, and superior oil resistance. However, in reality, it is very challenging to develop a technologically compatible blend due to the interfacial tension caused by the extreme polarity mismatch between NR and NBR. Poor compatibility between the blend’s components will result in a phase-separated morphology in the final mix, which will have inferior physico-mechanical properties.14
One of the extensively used scientific methods for increasing compatibility is the use of a compatibilizer, which can increase the interfacial adhesion between the blend’s components.15,16 A compatibilizer might be a graft or block copolymer,17–20 a low molecular weight chemical,21–24 or modified filler.25–27 Depending on the mechanism, the process of compatibilization could be reactive or nonreactive. Nonreactive compatibilization, also known as physical compatibilization, occurs when a graft or block copolymer achieves interfacial adhesion by solubilizing its individual blocks in blend phases with equivalent solubility or affinity. Reactive compatibilization, or in situ compatibilization, refers to the process whereby the compatibility of the blend components is improved through the in situ creation of graft or block copolymers by the combination of the reactive blend phases.
To increase the compatibility of NR/NBR blends, many types of compatibilizing agents have been used. Studies by Nah and colleagues examined the morphological, rheological, and dynamic mechanical properties of a 50/50 blend of NR/NBR by using a trans-polyoctylene rubber (TOR) as a compatibilizer.28–30 They observed that the incorporation of TOR reduces the size of the NBR phase, which was distributed as a dispersed phase in the continuous NR phase. They also determined that TOR was situated at the interface of NR and NBR with the aid of dynamic mechanical analysis (DMA). Tetrachlorophthalic anhydride was studied by Nashar et al. to enhance the compatibility and dielectric characteristics of a 25/75 NBR/NR blend.31 To increase the compatibility of a 20/80 NR/NBR blend. Sirisinha et al. added various amounts of maleated ethylene propylene diene rubber (EPDM-g-MA) and maleated ethylene octene copolymer (EOR-g-MA), but they noticed that neither EPDM-g-MA nor EOR-g-MA worked well as a compatibilizing agent.32 Instead, because of the low amount of grafted maleic anhydride, these materials migrated more toward the dispersed phase (NR) and increased its size by increasing viscosity. Fly ash particles with and without chloroprene rubber (CR) and epoxidized natural rubber (ENR) were used as compatibilizers in a study by Kantala et al. to examine the reinforcing effect of an NR/NBR blend.33 They concluded that the blend’s mechanical properties were improved by both CR and ENR; however, ENR was a better compatibilizer than CR. Kumari et al. investigated the vapor permeation characteristics of NR/NBR membranes without and with poly (ethylene-co-vinyl acetate [EVA]) as a compatibilizer and observed excellent vapor permeation resistance at 6 phr loading of EVA compatibilizer.34 Angnanon et al. prepared a graft copolymer of styrene (ST) and ACN on to NR via solution polymerization and tested the resultant NR-g-(ST-co-ACN) as compatibilizer on a 50/50 NR/NBR blend.35 They claimed that the mechanical and oil resistance properties of the blend significantly improved due to the excellent compatibilizing action of NR-g-(ST-co-ACN). Zhao et al. investigated the use of ENR with different levels of epoxidation as a compatibilizer in a ternary blend of 70/30/10 NR/NBR/ENR.36 They reported that the mentioned ternary blend exhibited improved TS and tear strength compared with that of the 70/30 NR/NBR blend. Moolsin et al. prepared methyl methacrylate-grafted NR initiated by benzoyl peroxide as (NR-g-PMMA<BPO>) and potassium persulfate as (NR-g-PMMA<PPS>) and investigated their effects as compatibilizers on a 50/50 blend of NR/NBR.37 They observed that both the TS and tear strength of the blend were considerably improved with the addition of 10 phr of (NR-g-PMMA<BPO>) as the compatibilizer. Moreover, they could see that the IRM 903 oil resistance of the blend got remarkably improved after the incorporation of 7.5 phr of (NR-g-PMMA<BPO>). In addition to the use of compatibilizing agents, the selection of proper curatives; their kinetics of mixing, dispersion, and distribution in the rubber matrices; and the rate of curing are also very important to achieve good compatibility between the blend phases. For example, if the rate of curing of individual rubber components in a blend to an applied curing agent is different, the resultant blend may have uneven crosslinked phases that leads to inferior physical properties. Therefore, a suitable curing agent or cure composition that can co-crosslink the rubber phases of the blend at the same rate is essential to have a fully crosslinked blend with uniform distribution of crosslinks across the blend phases.38,39
In this work, an attempt has been made to enhance the properties of a 50/50 blend of NR and NBR by uniformly crosslinking the blend phases by using S and a delayed action accelerator (n-cyclohexyl-2-benzothiazole sulfenamide [CBS]) along with a bismaleimide known as N,N′ meta-phenylene dimaleimide. The 50/50 blend ratio was selected to retain the unique characteristics of the individual elastomers in equal proportions in the resulted blend. Numerous studies have used this chemical, under the commercial name HVA-2, as a compatibilizer, crosslinking agent, or co-agent in a variety of thermoplastic and elastomer blend systems. For example, Hassan et al. noticed that HVA-2 acted as an effective compatibilizer for improving the tensile and impact properties of a ternary blend of polypropylene/natural rubber/linear low-density polyethylene (PP/NR/LLDPE).40 Similarly, the compatibilizing effect of HVA-2 was tested by Celestino et al. in an acrylic rubber/NBR blend.41 For wire and cable applications, Alwaan et al. used HVA-2 as a compatibilizer in a metalocene LLDPE/ENR-50 blend.42 HVA-2 was discovered to operate as a crosslinker in the HDPE/NR/thermoplastic tapioca starch blend by Kahar et al.43 Soares et al. discovered that HVA-2, in conjunction with dicumyl peroxide (DCP), can crosslink the PP/NBR blend.44 The effect of HVA-2/DCP combination on the dynamic vulcanization of a PP/SBR blend was examined by Leite et al.45 The mechanical, dynamic mechanical, and Fourier-transform infrared spectroscopy investigations revealed that HVA-2 acted as a curing co-agent as well as a compatibilizer between PP and SBR.
Marek et al. examined the use of N,N′ meta-phenylene dimaleimide (trade name Maleide F [MF]) as a co-curing agent in combination with an accelerated S system on a 50/50 blend of NR/CR46 and NR/BIIR.47,48 In this work, we used the ability of this chemical to undergo an Alder-ene reaction with the unsaturated elastomers (NR and NBR) in the blend components to enhance their interfacial adhesion. To the best of our knowledge, there are no detailed descriptions in the literature concerning how the physical properties of the NR/NBR blend improved after crosslinking via a combination of MF and an accelerated S system.
EXPERIMENTAL
materials
The materials used were NR [standard Vietnamese rubber with a Mooney viscosity ML(1 + 4) @ 100°C = 60 ± 5], obtained from Binh Phuoc, Vietnam, under the trade name SVR CV60, and NBR [Europrene GRN 1945 with an ACN content of 19 wt% and a Mooney viscosity ML(1 + 8) @ 100°C = 45] as the base elastomers. MF, a combination of 75% N,N′-meta phenylene dimaleimide, and a 25% blending agent, was procured from Krata Pigment, Tambov, Mentazhnikov, Russia. Other ingredients such as S, CBS, stearic acid, and zinc oxide (ZnO) were purchased from Sigma-Aldrich, Prague, Czech Republic.
preparation of rubber compounds
The rubber compounds were prepared per the formulations displayed in Table I. All of the compounds were prepared using a 0.4 L mini lab Banbury mixer (Farrel Pomini, GmbH & Co. KG, Duisburg, Germany). A fill-factor of 0.75 was taken for the efficient mixing of the ingredients. The individual rubbers were masticated together for 1 min at 50°C under 60 rpm. ZnO and stearic acid were then added and the mixing was continued for up to 5 min. Next, the mix is taken out of the Banbury mixer and homogenized for 1 min by using a two-roll mill. The homogenized batch was again taken back into the Banbury mixer, where the curatives (S, CBS, and MF) were added and mixed for 3 min at 50°C under 60 rpm. After the mixing, the whole mix was discharged and again homogenized using a two-roll mill for 5 min to get a better dispersion of the curatives into the rubber blends. Finally, the mix was molded into sheets with a thickness of 2 mm by using a compression molding heat press (LaBEcon 300, Fontijne Presses, Rotterdam, The Netherlands) based on the rheometer cure data at 170°C.
CHARACTERIZATION TECHNIQUES
Cure Characteristics
A moving die rheometer (MDR-3000, MonTech, Buchen, Germany) as per ASTM Standard D 5289 was used to measure the curing parameters such as maximum torque (MH), minimum torque (ML), the difference between maximum and minimum torque (ΔM), scorch time (TS2), optimum cure time (T90, the time required for the torque to reach 90% of the maximum torque obtained from Eq. 1) of the rubber compounds at 170°C for 1 h. The cure rate index (CRI), a measure of the rate of curing, was calculated using Eq. 2 as follows:
cure–strain sweep analysis
To understand the strength of the cured network of the blend due to the formation of the anticipated bismaleimide crosslinks in the early phase of the curing process, the shear storage modulus (G′) of the blend was monitored immediately after the curing. For this, a cure test was first performed on the specified sample at 170°C up to the predetermined testing time by using a Premier rubber process analyzer (RPA; Alfa Technologies, Wilmington, DE, USA). The cured sample is then cooled down to 40°C in the RPA die, and a strain sweep was performed from 0.5 to 100% at a constant frequency of 1 Hz.
dma
Strain Amplitude Sweep Test.
—To corroborate the results of the cure–strain sweep obtained from RPA, similar samples (10 mm × 5 mm × 2 mm) cured as per their T90 were subjected to a dynamic mechanical analyzer (model details) in tension mode, and we measured the G′ by applying a strain amplitude from 2 to 220 μm, which corresponds to a dynamic strain of 0.02–2.2% at a constant frequency of 1 Hz at 40°C.
Temperature Sweep Test.
—The temperature sweep test was performed on selected samples having the above-mentioned sample dimensions in tension mode from −100 to +100°C at a heating rate of 2°C/min by applying a constant amplitude of 0.33 μm at a frequency of 1 Hz.
equilibrium swelling behavior
The swelling behaviors of the blends were examined by measuring the swelling index (SI, %), solvent uptake (Q, mole %), and volume fraction of rubber in the swollen mass (Vr) by using Eq. 3, 4, and 5, respectively.46 For that, the blend samples of dimensions 20 mm × 30 mm × 2 mm were allowed to swell in the selected solvents until they achieved the equilibrium swelling state. The swollen samples taken out and weighed immediately after removing the solvent retained on the surface of the samples by using blotting paper. Next, the swollen samples were dried and weighed again to calculate the volume fraction of rubber as follows:
where Wi is the initial weight of the sample, Ws is the swollen weight, Wd is the dried weight, ρr is the density of the rubber blend, ρs is the density of the solvent, and Mw is the molecular weight of the solvent.
tensile properties
The stress–strain behavior and the corresponding tensile properties of the vulcanizates were measured using a universal testing machine (Testometric M350, Testometric Company, Ltd., Rochdale, UK). The testing was performed under ambient conditions at a crosshead speed of 500 mm/min as per ISO 37 by using an S2-type specimen with a thickness of 2 mm. The results were reported for an average of six tested specimens. The tensile properties of selected samples were also measured after aging them at 100°C for 72 h by using a forced air circulating oven.
hardness testing
Cured samples having smooth surfaces were used to measure the indentation hardness using a Shore A hardness tester (Bareiss durometer, Oberdischingen, Germany) as per ASTM Standard D 2240. Indentations were made on different areas of the samples by applying constant pressure for 3 s. Six readings were taken from different areas of the sample, and the average value is reported.
RESULTS AND DISCUSSION
curing analysis of nr-s and nbr-s without and with mf
The curing of an elastomeric blend system is critical for achieving good physico-mechanical properties. If the cure system exhibits preferential curing behavior toward one phase of the blend, the crosslink distribution will be uneven, resulting in more crosslinked and less crosslinked blend phases. Therefore, the cure system designed for the blend must have the ability to crosslink both phases of the blend simultaneously and uniformly. This is only feasible if the rates of curing of the different rubber components in the blend with the curing agent used are the same. Taking these parameters into account, a cure composition, as shown in Table I, was developed and tested the curing behaviors of the individual elastomers and their blends. Figure 1a, b depicts the cure curves of NR-S and NBR-S without and with 1 phr of MF. Table II summarizes their curing properties. NR-S cures quickly and has reached its full extent of cure (maximum torque) within 5 min of curing. Afterward, reversion causes a rapid declination in the rheometric torque until the end of the given curing time. The cure rate curves given in the figure as an inset, in contrast, clearly illustrate that the rate of curing was slightly slowed in NR-SMF1. Therefore, NR-SMF1 took approximately 10 min to reach its maximum torque. However, the intensity of reversion (%) in NR-SMF1 was substantially weaker than in NR-S. The intensities of reversion in NR-S and NR-SMF1 at different intervals from the time of their respective maximum torques (S′max) are displayed in Table III and computed using Eq. 6 as follows:49
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
NR-S displayed 14.38% reversion after merely 10 min of cure from its maximum torque (S′max + 10 min). The amplitude of reversion increased to 29.42% as the curing period was increased to 45 min from S′max (S′max + 45 min). NR-SMF1 follows the same reversion pattern as NR-S. However, the amount of reversion in NR-SMF1 was substantially lower than in NR-S. After integrating just 1 phr of MF, the reversion observed in NR-S at S′max + 10 min was reduced to 3.42%. As the curing time of NR-SMF1 extended to 45 min beyond S′max, it experienced a reversion of just 11.05%. This significant reduction in the severity of reversion is attributed to the formation of thermally stable bismaleimide crosslinks between NR chains during the initial stage of the accelerated S curing process. The Alder-ene reaction between the maleimide moieties of MF and the unsaturated NR chains was proposed as one of the plausible crosslinking mechanisms between NR and MF.50 According to this mechanism, the unsaturated segments in the maleimide moieties of MF abstract allyl hydrogen atoms from the NR chains and get sandwiched between the NR chains in the form of crosslinks (Figure 2). The curing curve of NBR-S does not show a distinct reversion; rather, it shows a slight marching modulus curing behavior beyond 30 min of curing with the same composition of the curing system as used in NR-S. As observed in NR-SMF1, the speed of curing in NBR-SMF1 was also considerably reduced compared with that of NBR-S, which is evident from the cure rate curves of these compounds depicted as an inset of Figure 1b. One of the probable reasons for the low speed of curing in NBR-SMF1 can be attributed to the interruption of the S curing process due to the early formation of few numbers of bismaleimide crosslinks via the above-mentioned Alder-ene reaction. Here, the Alder-ene reaction is taking place by the abstraction of the allyl hydrogen atoms from the NBR chains by the maleimide moieties of MF (Figure 3).
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
To support the proposed reaction mechanisms in Figures 2 and 3, the curing behaviors of pristine NR and NBR with 1 phr of MF alone were evaluated at 170°C for 1 h. The results are depicted in Figure 4. Both NR-MF1 and NBR-MF1 exhibited a considerable increment in the rheometric torque within a short period, even though their extent of crosslinking was poor. For instance, to produce 0.5 dNm of torque due to crosslinking, NR-MF1 took 3.3 min. However, to reach the same level of torque, NBR-MF1 took 4.71 min. These experimental observations confirm that MF can crosslink both the NR and NBR via the Alder-ene reaction. From Figure 1a, b, it has already been noticed that the curing curves of both NR-S and NBR-S get disturbed and slightly delayed in the presence of MF. This is possible only if the curing of NR or NBR with MF takes place at a faster rate (or at the same rate) compared with that of the curing reactions of NR-S or NBR-S. To check this, the cure rate curves of NR-S/NR-MF1 and NBR-S/NBR-MF1 were evaluated (Figure 5a, b), and they confirmed that the rates of cure of NR-MF1 and NBR-MF1 were slightly faster than the respective NR-S and NBR-S at the beginning of the reaction, even though the extent of cure offered by MF alone was poor. These experimental results further endorse the proposed Alder-ene reaction during the curing of NR-SMF1 and NBR-SMF1.
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
curing behavior of the blend nr/nbr-s without and with mf
Based on the curing behaviors of NR-S and NBR-S without and with MF, an attempt was made to increase the interfacial adhesion between NR and NBR in a 50/50 blend via curing by using the same composition of accelerated S and 1 phr of MF. Represented in Figure 6 are the curing curves and the corresponding rate curves of the blends NR/NBR-S and NR/NBR-SMF1 at 170°C for 1 h. Their cure characteristics are also depicted in Table II. The curing patterns of the blend without and with MF1 were comparable with those of NR-S and NR-SMF1, although NR/NBR-SMF1 (T90 = 5.27 min) requires more time to optimum cure than NR/NBR-S (T90 = 3.65 min). However, the percentage of reversion (see Table III) in NR/NBR-S and NR/NBR-SMF1 was significantly lower than the corresponding percentages for NR-S and NR-SMF1. For example, NR/NBR-S showed a reversion of 11% at S′max + 10 min, which was 23.5% lower than the reversion of its individual constituents, NR-S, under the same curing condition. Similarly, the reversion at S′max + 10 min for NR/NBR-SMF1 was 25.4% lower than the corresponding NR-SMF1. As the curing time extended to 45 min beyond S′max, the magnitude of reversion in NR/NBR-S and NR/NBR-SMF1 was 33.2 and 45.7% lower than the corresponding NR-S and NR-SMF1, respectively. One of the reasons for the higher reversion resistance in NR/NBR-S compared with NR-S is the inclusion of co-crosslinked NR chains with thermally more stable NBR chains via S atoms (mono-, di-, or polysulfidic) in the network. In contrast to the networks of NR/NBR-S, the feasibility of interchain crosslinks with MF via the Alder-ene reaction (Figure 7) can be attributed to the enhanced reversion resistance in NR/NBR-SMF1. The curing curves of virgin NR, NBR, and their 50/50 blend with MF alone (Figure 4) show that the rate and extent of reaction in NR/NBR-MF1 are significantly higher than that of NR-MF1 and NBR-MF1. This experimental result is one of the most compelling pieces of evidence for the creation of interchain crosslinks between NR and NBR via MF during the initial phase of accelerated S curing of NR/NBR-SMF1.
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
cure–strain sweep analysis using rpa
According to the literature, the network formed after curing the rubber compound with a conventional S vulcanization (CV) system has 95% poly- and disulfidic crosslinks and 5% monosulfidic crosslinks.51 Therefore, the network formed in T90-cured NR/NBR-S samples may be largely composed of poly- and disulfidic crosslinks. However, it is believed that some C–C linkages were formed between NR and NBR during the early curing phase of NR/NBR-SMF1. Because of the high bond dissociation energy (∼350 kJ/mole) of C–C crosslinks, the cured networks made of C–C crosslinks are thermally and physically more stable than networks created with sulfidic crosslinks. Therefore, it is fair to assume that the physical strength of the network formed after the curing of NR/NBR-SMF1 will be greater than that of NR/NBR-S. To check this assumption, a strength analysis in terms of shear G′ was performed on NR/NBR-S and NR/NBR-SMF1 by subjecting them to a combination of cure and strain sweep tests using RPA as per the procedure described in the experimental section.52Figure 8a, b shows the RPA curing curves of NR/NBR-S and NR/NBR-SMF1 for 5 min, as well as their shear G′ as a function of dynamic strain from 0.5 to 100%. The curing rate in NR/NBR-S was faster, and the torque due to curing was greater from the beginning to the end of the curing time. However, due to the projected growth of MF-based C–C crosslinks in NR/NBR-SMF1 at the start of curing, the amount of sulfidic crosslinks generated within the given 5 min of curing may be rather low. If the curing torque is regarded as a metric of network stiffness, the 5 min–cured NR/NBR-S should have a greater G′ than the NR/NBR-SMF1 under the imposed shear deformation. However, the G′ of NR/NBR-SMF1 was much larger than that of NR/NBR-S over the full range of the applied oscillatory shear strain, from 0.5 to 100%. This implies that the shear force necessary to deform the 5 min–cured NR/NBR-SMF1 was greater than that required to deform NR/NBR-S. This RPA finding adds to the evidence that few numbers of strong bismaleimide-based C–C crosslinks emerge in the NR/NBR-SMF1 network during its initial stage of curing via Alder-ene reaction before advancing to sulfidic crosslink formation.
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
dma
The T90-cured NR/NBR-S and NR/NBR-SMF1 were subjected to a tension mode amplitude deformation sweep by using DMA to validate their shear G′ behavior observed during the dynamic strain sweep analysis in RPA. Figure 9 shows the dependences of storage modulus (E′) of NR/NBR-S and NR/NBR-SMF1 as a function of amplitude sweep from 2 to 220 μm (0.02–2.2%). Both NR/NBR-S and NR/NBR-SMF1 demonstrated a linear response of E′ over the whole strain sweep range. This is because the applied deformation occurred in the linear viscoelastic zones of these materials. However, as seen in the RPA strain–sweep analysis, the magnitude of E′ for NR/NBR-SMF1 was greater than that for NR/NBR-S throughout the deformation range. This indicates that the formation of bismaleimide crosslinks strengthens the cured network of NR/NBR-SMF1. As seen in the insets of Figure 9, bismaleimide crosslinks can develop between the NR chains, the NBR chains, and the NR and NBR chains within the cured network of NR/NBR-SMF1. All of these crosslinks may contribute to the strengthening of the NR/NBR-SMF1 network, requiring greater force to create a deformation.
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
It has been proven that the macromolecular chains of rubbers can have substantial degrees of motion before crosslinking. However, the addition of crosslinks significantly reduces the mobility of the molecular chains. The amount of crosslinks (crosslink density) and the type of crosslinks (–C–C– or sulfidic) between the rubber chains further influence chain mobility. In general, the larger the density of crosslinks, the lower the chain mobility. The C–C crosslinks generated by the direct combining of macro radicals by the action of an organic peroxide have the lowest chain mobility around the crosslinks. If the crosslinks generated between the chains are sulfidic in nature, the order of chain mobility will be highest in polysulfidic crosslinked chains and lowest in monosulfidic crosslinked chains. Of course, the mobility of disulfidic crosslinked chains will fall somewhere between poly- and monosulfidic crosslinked chains.53–55 However, the addition of multifunctional compounds such as bismaleimide might cause diverse chemical alterations in the macromolecular chains.56 For example, from the knowledge of the curing behaviors of neat NR, NBR, and NR/NBR with MF alone (Figure 5), it is a reasonable assumption that the MF can produce a cured network of NR/NBR comprising a mixture of (a) MF crosslinked NR chains, (b) MF crosslinked NBR chains, (c) MF interlinked NR and NBR chains, (d) MF grafted NR chains, and can MF grafted NBR chains (Figure 10). All of these elastic (crosslinks) and nonelastic (grafted units) alterations can have a variety of effects on the overall properties of the cured rubber. Higher numbers of interchain crosslinks (c) indicate greater compatibility between NR and NBR, which may improve overall blend properties. The blend phases will be unevenly crosslinked if crosslinks (a) or (b) are larger, resulting in poor blend properties. The dangling grafted units, on the one hand, diminish the flexibility of the grafted chains, but on the other hand, they provide more free space within the cured network.
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
To understand more about the molecular morphologies of the blends, a temperature sweep test was performed from −100 to +100°C and evaluated their viscoelastic properties. Represented in Figure 11a, b are the temperature-dependent storage moduli and the respective tan δ curves of the T90-cured NR-S, NBR-S, and NR/NBR-S. The storage modulus curve of NR-S exhibited a distinct plateau region from −100 to −68°C, well above its glass transition temperature (Tg). It then showed a single-stage declination due to glass transition with increasing temperature. Similarly, the E′ curve of NBR-S also exhibited a plateau region from −100 to −54°C and then a single-stage declination, as in the case of NR-S. The modulus curve of NR/NBR-S occupied a position in between NR-S and NBR-S. However, after showing a plateau region from −100 to −64°C, the E′ curve of the blend NR/NBR-S exhibited a two-step declination with increasing temperature due to the immiscible nature of the blend. The temperature-dependent damping (tan δ) curves of NR-S exhibited a well-defined peak with a height of 2.79 and a Tg of −55.0°C. Similarly, NBR-S exhibited a tan δ peak at −38.43°C, with a peak height of 1.56, which indicates its Tg. The high tan δ of NR suggests that it has a higher damping efficiency than NBR. However, the 50/50 NR/NBR-S revealed two peaks: one peak at −56.60°C and a second peak at −38.39°C, corresponding to the Tg of NR and NBR in their blend, respectively. It is worth noting that the tan δ peak height of NR in the blend was reduced from 2.79 to 0.48, and the peak position was changed to a low temperature regime by 1.6°C. Similarly, the peak height of NBR in the blend was lowered from 1.56 to 1.12, with no noticeable change in peak position. This implies that after blending, the damping efficiencies of both NR and NBR have decreased. However, the damping efficacy of NR phase in the blend was significantly lower than that of NBR phase. This also suggests that the mobility of NR chains in the blend is severely limited, most likely due to the inclusion of NBR chains with bulky ACN groups in NR. In short, the significant reduction in the tan δ peak height of these individual elastomers confirms that the applied curing system can impart a certain level of technological compatibility between NR and NBR in their 50/50 blend while preserving the individual elastomeric characteristics.
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
The DMA study on NR/NBR-SMF1 was also performed to figure out the network characteristics of the blend after curing it with MF. The temperature dependency values of E′ and the associated tan δ peaks of the T90- and 60 min–cured NR/NBR-S and NR/NBR-SMF1 are shown in Figure 12a, b. The E′ curves of both the T90- and the 60 min–cured NR/NBR-SMF1 were greater than the E′ curves of NR/NBR-S over the whole temperature range of −100 to 100°C. This may be due to the enhanced crosslink density (ν) in NR/NBR-SMF1 due to the formation of bismaleimide crosslinks in its network. To confirm this, the crosslink densities of the T90- and 60 min–cured blends were calculated from their respective E′ values in the rubbery plateau per Eqs. 7 and 8 derived from the theory of rubber elasticity as follows:57where E′r is the storage modulus of the blends measured at 25°C from the rubbery plateau, T is the absolute temperature, R is the universal gas constant, ρ is the density of the rubber blends, and Mc is the number average molar mass of the elastically effective chains between the crosslinks. Represented in Table IV are the E′ values of NR/NBR-S and NR/NBR-SMF1 at 25°C and their respective crosslink densities. The crosslink densities of the T90- and 60 min–cured NR/NBR-SMF1 were significantly greater than those of the respective NR/NBR-S, as expected. It is worth noting that the 60 min–cured NR/NBR-S exhibited a 48% loss of crosslink density compared with the T90 min–cured NR/NBR-S due to reversion. However, compared with the T90-cured NR/NBR-SMF1, the 60 min–cured NR/NBR-SMF1 lost only 18% of its crosslink density. This demonstrates the better thermal stability of the bismaleimide-based cured network of NR/NBR-SMF1.
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
To obtain more insight into the molecular dynamics of the crosslinked network, the tan δ peaks of the T90- and 60 min–cured blends were meticulously analyzed. The height of the tan δ peak responsible for the Tg of the NR phase in the T90-cured NR/NBR-SMF1 was approximately 6% lower than NR/NBR-S. Moreover, the peak position was also shifted to a lower temperature region. The lowering of peak height can be attributed to the enhanced crosslink density due to the formation of the bismaleimide crosslinks between the NR chains and also between the NR and NBR chains (in addition to the sulfidic crosslinks). The shifting of the peak position to the lower temperature regime indicates that the chains of the NR phase in NR/NBR-SMF1 can have considerable freedom of mobility. The presence of some nonelastic grafted units of MF (Figure 10) in the NR segments of the blend might be one of the probable reasons for this observation. However, the tan δ peak corresponding to the Tg of the NBR phase in the T90-cured NR/NBR-SMF1 was higher than the T90-cured NR/NBR-S, with a slight peak shifting toward the lower temperature region. One of the reasons for this might be the presence of uncrosslinked MF grafted NBR chains within the network, maybe for the same reasons why there appears to be a shift in the tan δ peak position toward the lower temperature region.
Figure 12b shows that, in contrast to the T90-cured blends, the peak height corresponding to the Tg values of the respective NR and NBR phases in the 60 min–cured NR/NBR-SMF1 is lower than the corresponding blend phases in the 60 min–cured NR/NBR-S. This indicates that extending the curing time may introduce certain structural changes within the network of the blends. The breaking and subsequent rearrangements of the more flexible polysulfidic (–Sx–; x > 3 crosslinks created within and also in between the blend components into relatively less flexible di (–S–S–)- and mono (–S–)-sulfidic crosslinks can be considered one of such structural changes. Moreover, the prolonged curing time may also introduce more bismaleimide bridges within and between the blend components of NR/NBR-SMF1. This newly created bismaleimide crosslink may aid in maintaining a higher crosslink density than the 60 min–cured NR/NBR-S. All of these structural changes, as well as the creation of extra bismalimide bridges in the respective blend phases of NR/NBR-SMF1, significantly constrain their chain mobility, resulting in decreased tan δ peak heights.
swelling analysis
Swelling is essentially a diffusion process in which the relatively small solvent molecules penetrate into the cured network of the elastomer when they come into contact with each other. From the knowledge of the swelling parameters such as SI, Q, and Vr, one can understand how good the cured network is at resisting solvent penetration. In general, the higher the crosslinking density, the lower the penetration of the solvent molecules into the cured network and hence the lower the swelling index. However, the swelling behavior of an immiscible rubber blend in a selected solvent is not very straightforward. In most cases, the crosslinking of an immiscible rubber blend may not show a uniform distribution of crosslinks across the blend phases because of poor interfacial interaction. The differences in the rate of curing of the individual rubber components in the blend with an applied curing system may also cause uneven crosslink distribution. In such cases, it is very difficult to know which phase of the blend is crosslinked more (swelling less) and which phase is marginally crosslinked (swelling more). By considering these factors, two solvents, toluene and acetone, have been selected to study the swelling behavior of NR/NBR-S and NR/NBR-S MF1. This selection was made based on the solubility parameters (δ) of the solvents (δtoluene = 18.2 MPa1/2 and δacetone = 19.9 MPa1/2) and the rubbers (δNR = 17.1 MPa1/2 and δNBR = 19.2 MPa1/2).58,59 These δ values indicate that toluene can cause a substantial degree of swelling in the cured network of NR, NBR, and their blends NR/NBR. The degree of swelling of the cured NR in acetone is expected to be poor because of the significant difference in their solubility parameters due to the polarity mismatch. However, acetone can cause substantial swelling in the cured network of NBR.
Depicted in Table V are the SI, Q, and Vr obtained after equilibrium swelling of the samples in toluene and acetone. The cured network of NR-S exhibited approximately 82.53% swelling in toluene. However, NBR-S exhibited a swelling of 86.86%, which was approximately 4% higher than the swelling of NR-S in toluene. This may be due to the relatively low crosslinking density of NBR-S. To check this further, the equilibrium swelling of NBR-ST in toluene was measured, and it showed a value of approximately 81.13%. This means that TMTD improves the overall crosslinking density of NBR-ST. The swelling behavior of 50/50 NR/NBR-ST was also measured to know whether TMTD can selectively crosslink the NBR phase of NR/NBR-ST. Interestingly, NR/NBR-ST exhibited only 82.51% swelling in toluene, which was close to the swelling behavior of NR-S and was considerably lower compared with NBR-S. Moreover, the percentage swelling of the T90-cured NR/NBR-ST was also lower than that of the T90-cured NR/NBR-S. To get more insight, the swelling behavior of these samples in acetone was also analyzed. As expected, NR-S exhibited a high swelling resistance in acetone, with a swelling index of just 9.98%. This can be attributed to the huge polarity mismatch between NR and acetone. By contrast, NBR-S exhibited an equilibrium swelling of 66.89%. However, the equilibrium swelling of NBR-ST was reduced to 58.47%, indicating that TMTD improves the overall crosslinking density of NBR-ST. Similarly, the swelling index of the T90-cured NR/NBR-ST was approximately 4% lower than the swelling index of the T90-cured NR/NBR-S. From these equilibrium swelling data, it is reasonable to assume that the NBR phase of the T90-cured NR/NBR-ST would be in a more crosslinked state than the same in the T90-cured network of NR/NBR-S. To further understand this swelling behavior, the curing behaviors of NBR-S, NBR-ST, NR/NBR-S, and NR/NBR-ST were analyzed together as depicted in Figure 13a, b. It can be seen that TMTD improves the state of cure and also accelerates the cure rate of NBR-ST. However, the presence of TMTD does not enhance the state of cure; rather, it improves only the rate of cure in NR/NBR-ST compared with NR/NBR-S. From these curing behaviors, it is clear that the presence of TMTD leads to a rapid selective crosslinking of the NBR phase of the blend NR/NBR-ST. Because TMTD encourages the crosslinking of the NBR phase, the chances of interfacial crosslinking between NR and NBR in NR/NBR-ST will be negligible even if the overall extent of crosslinking is high. However, the possibilities of interfacial crosslinking between NR and NBR in NR/NBR-S are expected to be high even if the overall extent of crosslinking is relatively low. One probable reason for this would be the more or less comparable curing rates of the components NR and NBR in NR/NBR-S with the given curing system (S/CBS).
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
The percentage swelling and the solvent uptake of the T90-cured NR/NBR-SMF1 in toluene were lower than those of the T90-cured NR/NBR-S. In connection with this, the Vr of NR/NBR-SMF1 in its equilibrium swollen mass with toluene was relatively higher than the corresponding NR/NBR-S. As shown in Figure 8, the possibilities of bismaleimide crosslinks between the NR chains and also between the NR and NBR chains (interfacial crosslinks) in addition to the sulfidic crosslinks can be attributed to the enhanced swelling resistance of NR/NBR-SMF1 in toluene. However, the percentage swelling of the T90-cured NR/NBR-SMF1 in acetone was higher (49.36%) than the corresponding T90-cured NR/NBR-S (41.61%) in acetone. The higher swelling index of the T90-cured NR/NBR-SMF1 in acetone is mainly due to the swelling of its NBR phase. From the DMA, it has been noticed that the tan δ peak corresponding to the NBR phase of the T90-cured NR/NBR-SMF1 was higher than the tan δ peak corresponding to the NBR phase of the T90-cured NR/NBR-S. The presence of certain uncrosslinked bismaleimide grafted NBR chains in the network of NR/NBR-SMF1 was considered one of the promising reasons for this behavior. The uncrosslinked MF grafted NBR chains may create more free volume in the cured network, through which more solvent molecules can go inside the network. As observed in the T90-cured samples, the percentage swelling of 60 min–cured NR/NBR-S in toluene was higher than that of 60 min–cured NR/NBR-SMF1. On the contrary, the swelling indexes of the 60 min–cured NR/NBR-S and NR/NBR-SMF1 in acetone were almost similar. It is worth noting that the swelling index of the 60 min–cured NR/NBR-SMF1 in acetone was approximately 11.3% lower than the T90-cured NR/NBR-SMF1. This means that the solvent (acetone) resistance of the NBR phase of the 60 min–cured NR/NBR-SMF1 improved. This observation was in line with the crosslink densities of these blends obtained from the DMA analysis (Table IV). Extending the curing time in NR/NBR-SMF1 may lead to the formation of more numbers of bismaleimide crosslinks of the types (a) and (b) in both phases and also the interfacial crosslinking (c), as explained in Figure 10. As a result, the overall swelling resistance of the 60 min–cured NR/NBR-SMF1 will always be higher than that of the 60 min–cured NR/NBR-S. The lower heights of the tan δ peaks corresponding to the NR and NBR phases of the 60 min–cured NR/NBR-SMF1 compared with the tan δ peaks of the respective phases in the T90-cured NR/NBR-S further endorse this swelling behavior.
tensile properties
The stress–strain behaviors of the representative samples of the T90-cured NR-S, NBR-S, NBR-ST, NR/NBR-S, NR/NBR-ST, and NR/NBR-SMF1 are represented in Figure 14a, b. Their tensile properties collected from an average of six tested specimens are displayed in Table VI. NR-S exhibited a TS of 9.2 MPa with an elongation at break (EB) of 689%. However, with the same curing system, NBR-S exhibited a low TS of 2.9 MPa with an EB of 789%. The TS of NBR-ST (2.32 MPa) was more or less equal to the TS of NBR-S. However, its EB was drastically reduced to 348%. This might be due to the increased crosslink density in the cured network of NBR-ST caused by the action of TMTD. The hardness (45.8 Shore A) of NBR-ST compared with NBR-S (35.7 Shore A) also supports the increased level of crosslink density in NBR-ST. It was interesting to note that the blend NR/NBR-ST exhibited a TS of just 3.5 MPa with an EB of 517%. By contrast, the TS of NR/NBR-S was 16.5 MPa, which was approximately 371% higher than NR/NBR-ST. In addition to this, the TS of NR/NBR-S was 80% higher than the TS of NR-S, 469% higher than the TS of NBR-S, and 617% higher than the TS of NBR-ST. The EB (938%) of NR/NBR-S was also quite good. The uniform distribution of crosslinks across the blend phases of NR/NBR-S might be one of the probable reasons for its improved TS. The proportionate curing rates of the blend components (NR and NBR) to the applied curing system at the given temperature (170°C) might be the reasons for the uniform distribution of the crosslinks within the network of NR/NBR-S. On the other hand, from the curing behaviors of NBR-S, NBR-ST, NR/NBR-S, and NR/NBR-ST (Figure 13a, b), it has been identified that the presence of TMTD enables crosslinking the NBR phase of NR/NBR-ST at a faster rate than the NR phase. As a result, more NBR chains get crosslinked quickly, leading to an uneven distribution of crosslinks across the blend phases. This brings more incompatibility between the blend components in NR/NBR-ST, which leads to inferior tensile properties in TS and EB.
Citation: Rubber Chemistry and Technology 96, 4; 10.5254/rct23.948326
The TS of the T90-cured NR/NBR-SMF1 was approximately 12% higher than NR/NBR-S. Moreover, a considerable improvement in EB (9% higher than NR/NBR-S) was also noticed. This indicates that the initially formed few numbers of bismaleimide crosslinks (before the sulfidic crosslinks) in NR/NBR-SMF1 may reduce the interfacial tension between the NR and NBR phases. This may further enhance the uniform distribution of sulfidic crosslinks across the blend phases than it was in the NR/NBR-S, as pictorially demonstrated in the inset of Figure 14b. Therefore, the reduction in the interfacial tension together with a more uniform distribution of the sulfidic crosslinks in the blend phases contributes to the enhanced TS in NR/NBR-SMF1.
The prolonged curing time (60 min) of both NR/NBR-S and NR/NBR-SMF1 showed reversion due to the breakdown of the initially generated polysulfidic crosslinks from their network structure. However, due to the presence of bismaleimide crosslinks in the network structure, the intensity of reversion during curing of NR/NBR-SMF1 was significantly lower than that of NR/NBR-S (see Table III). To further understand how extended curing time impacts mechanical performance, the tensile characteristics of NR/NBR-S and NR/NBR-SMF1 were evaluated after 60 min of molding, and the findings are also shown in Table VI. The 60 min–molded NR/NBR-S displayed a TS of 4.4 MPa, which was 73.2% lower than the TS (16.4 MPa) of the T90-molded NR/NBR-S. The EB was also lowered by 12% from the EB (938%) of NR/NBR-S as its molding time was extended to 60 min. This decrease in property verifies the breakage of the initially produced sulfidic crosslinks in NR/NBR-S as a result of prolonged curing. The breakdown (due to thermal degradation) of the rubber chains during overcuring might additionally contribute to the poor TS and EB. Similarly, compared with the TS (18.36 MPa) and EB (1025%) of the T90-molded samples, the TS and EB of the 60 min–molded NR/NBR-SMF1 was lowered by 39.5 and 15%, respectively. However, it is worth noting that the retention of TS (60.5%) of the 60 min–molded NR/NBR-S MF1 was higher than the retention of TS (26.8%) of the 60 min–molded NR/NBR-S. This shows that the network of NR/NBR-SMF1 preserves some integrity (due to higher crosslink density; see Table IV) even after a long curing period, owing to the presence of thermally stable bismaleimide crosslinks, as illustrated in Figure 10. The lower solvent uptake (higher Vr) of the 60 min–molded NR/NBR-SMF1 over NR/NBR-S obtained from the swelling experiment supports this presumption. In addition, a good correlation was also observed between the tensile properties of the overcured (60 min–molded) samples and the samples after the thermal aging at 100°C for 72 h. For example, approximately 38% retention in the TS was noticed after aging NR/NBR-SMF1. At the same time, the retention in TS of NR/NBR-S was just 24.3% after its aging under the same condition. This finding confirms the ability of NR/NBR-SMF1 to endure severe heat degradation, most likely due to strong interfacial interaction and greater compatibility between the components in the NR/NBR-SMF1 blend via bismaleimide crosslinks.
CONCLUSIONS
This research demonstrated that curing a 50/50 NR/NBR blend with S and a CBS/TMTD combination resulted in preferential crosslinking of the NBR phase, which leads to poor physico-mechanical blend properties. However, the TS of the blend after curing with S and CBS was approximately five times higher than the blend cured with S/CBS/TMTD. A reasonably comparable curing rate of the blend components (NR and NBR) toward S/CBS was found as the cause of the uniformly cured blend phases with better tensile characteristics. The use of bismaleimide (MF) along with S/CBS imparted several added advantages to the NR/NBR blend. The improved reversion resistance during curing and the enhanced overall crosslinking density were some of the benefits of using MF as a crosslinking agent along with S/CBS. Above all, the TS of the NR/NBR cured with S/CBS/MF was 11% higher than the blend cured with only S/CBS. The ability of MF to reduce the interfacial tension through the formation of crosslinks at the interface of NR and NBR via the Alder-ene reaction was recognized as the reason behind the improved tensile strength. Better TS retention of the overcured (60 min) NR/NBR blend with S/CBS/MF demonstrates that MF improves blend compatibility and has the ability to preserve network integrity via the proposed interfacial crosslinking.
Curing behaviors of (a) NR-S and NR-SMF1 and (b) NBR-S and NBR-SMF1.
Proposed curing mechanism between NR and MF.
Proposed curing mechanism between NBR and MF.
Curing curves of pristine NR, NBR, and NR/NBR with 1 phr of MF.
Curing rate curves of (a) NR-S and NR-MF1 and (b) NBR-S and NBR-SMF1.
Curing curves and the rate curves of NR/NBR-S and NR/NBR-SMF1.
Proposed interchain crosslinking between NR and NBR via MF.
Curing curves at 170°C, 5 min (a) and the corresponding G′ vs strain sweep curves at 40° C (b) of NR/NBR-S and NR/NBR-SMF1.
Storage modulus, E′ vs amplitude sweep curves of NR/NBR-S and NR/NBR-SMF1 at 40°C.
Proposed network structure of NR/NBR blend with MF (a) crosslinked MF between NBR chains, (b) crosslinked MF between NR chains, (c) crosslinked MF between NBR and NR chains, (d) NBR chains with grafted MF, and (e) NR chains with grafted MF.
E′ (a) and tan δ (b) vs temperature curves of NR-S, NBR-S, and NR/NBR-S.
E′ and tan δ vs temperature curves of NR/NBR-S and NR/NBR-SMF1 (a) T90 min–cured samples and (b) 60 min–cured samples.
Curing curves of (a) NBR-S and NBR-ST and (b) NR/NBR-S and NR/NBR-ST.
Stress–strain behaviors of (a) NR-S, NBR-S, and NBR-ST and (b) NR/NBR-S, NR/NBR-ST, and NR/NBR-SMF1.
Contributor Notes