STUDYING NR/ORGANO-MONTMORILLONITE NANOCOMPOSITES WITH SILANE COUPLING AGENTS VIA NETWORK VISUALIZATION TEM
ABSTRACT
Organo-montmorillonite (OMMT) increases the elastic modulus of elastomers such as natural rubber (NR) by a very large amount at low strains, but this decreases as the rubber is extended. Silane coupling agents, widely used with silica-filled rubbers, were added to NR/OMMT nanocomposites to increase the effects of OMMT on modulus at high strains. Both bis(triethoxysilylpropyl)tetrasulfide (TESPT) and 3-mercaptopropyl di(tridecan-1-oxy-13-penta(ethylene oxide))ethoxysilane (MPDES) increase tensile modulus significantly at strains greater than 30%. The coupling agents strengthen the rubber–filler interface, reducing cavitation around OMMT particles and preventing NR molecules from sliding at the interface. Evidence for a stronger rubber–filler interaction is provided by measurements of bound rubber content and use of network visualization transmission electron microscopy (NVTEM). OMMT also affects the dynamic properties of NR differently from other fillers. One aspect of this is the appearance of a peak in tan δ between 20 °C and 60 °C, attributed to the glass transition of intercalated and adsorbed NR molecules. The intensity of this peak is diminished by the addition of TESPT or MPDES, implying that they are restricting the intercalation of the rubber between the clay sheets. The coupling agents also have a small effect on vulcanization behavior compared with that of OMMT.
INTRODUCTION
Rubber/clay nanocomposites
In recent years, the field of polymer/clay nanocomposites has become a major area of research. This has included investigation of nanocomposites based on a wide range of elastomeric materials, such as natural rubber (NR),1–8 styrene–butadiene rubber,9 butadiene rubber,9 acrylonitrile–butadiene rubber,10,11 ethylene–propylene–diene monomer rubber,12,13 and epoxidized natural rubber.14 To achieve significant exfoliation of clay in a hydrophobic elastomer, it is generally necessary to modify the clay organically into an organoclay.15 Organoclays are typically produced by exchanging the metal cations between the clay sheets with bulky organic ions, often based on quaternary ammonium salts. These make the clay surface more organophilic and significantly increase the interlayer spacing, both of which facilitate the exfoliation of the clay sheets when mixed with an elastomer.
The overwhelming majority of research into elastomer/clay nanocomposites to date has looked at montmorillonite, which consists of silicate sheets of 50–500 nm diameter and ∼1 nm thickness, stacked face to face. The interlayer spacing of organo-montmorillonite (OMMT) can be up to 4 nm, with the exact distance dependent on the type and quantity of modifying agent used. In recent years, organoclays for use in rubber have been produced using a wide range of clay minerals and modifiers, as discussed during the course of reviews on the whole subject of rubber/clay nanocomposites by Thomas and Stephen16 and by Galimberti.17
Silane coupling agents
It has been recognized for some time that when using silica filler in rubber, it is often beneficial also to add a silane coupling agent. This is because the silica particles are covered with silanol (Si–O–H) groups that make the surface hydrophilic. These produce strong interparticle attractive forces that cause the particles to form large agglomerates composed of many aggregated primary particles. The same effects also lead to poor rubber–filler interaction when mixed into a hydrophobic elastomer such as NR. Silane coupling agents can react with the silanol groups to make the surface hydrophobic, which causes the silica to disagglomerate. The coupling agent can then covalently bond the filler to the rubber network at numerous locations, resulting in improved rubber–filler interaction. Clays also have silanol groups on their surface and so can also react with a silane coupling agent to produce a permanent bond to a rubber.18,19 More recently, some studies have looked at the use of a silane coupling agent with an organoclay filler.20–23 Although organoclays interact better than unmodified clays with hydrophobic elastomers, a silane coupling agent will still substantially affect the physical properties of a rubber/organoclay nanocomposite.
For a silane coupling agent to function properly, it must undergo a rapid chemical reaction with the silicate during mixing. However, the reaction of the coupling agent with the rubber should ideally be postponed until vulcanization begins, as otherwise the processability of the material will deteriorate. With silica, the mixing temperature must be greater than 140 °C for the silica–silane reaction to occur at a suitable rate while not exceeding 160 °C to prevent premature coupling,24 and it must be kept in this range for several minutes. In practice, this can require splitting the internal mixing procedure into two stages, because the heat buildup from mixing rubbers with high silica loading can cause the batch temperature to get excessively high. Once the material has cooled to room temperature after the first mix, it is mixed again to allow any unreacted silane to react with the silica filler. However, the development of high batch temperatures is at least in part a function of the energy required to break apart the strongly bonded silica agglomerates, and so it was believed that it could be possible to use a one-stage mixing process for NR/organoclay/silane nanocomposites without ill effects.
Network visualization transmission electron microscopy
An important tool for understanding the behavior of elastomer/clay nanocomposites is transmission electron microscopy (TEM), because it allows direct visualization of the nanocomposite microstructure. However, standard TEM techniques struggle to distinguish NR/organoclay nanocomposites containing a silane coupling agent from those without. One TEM-based technique that is of use for such materials is network visualization transmission electron microscopy (NVTEM),25 in which styrene is swollen into vulcanizate samples and then polymerized. The sample is then stained with osmium tetroxide, which reacts with the double bonds of the rubber but not with the polystyrene (PS). The osmium stain makes the rubber network visible against the transparent PS background, hence the term network visualization.
When the styrene swells into the vulcanizate, it will disperse throughout the rubber network. However, during polymerization, a constrained phase separation occurs, leading to the PS being located preferentially in certain areas within the rubber. In particular, PS is formed at the rubber–filler interface because of the relatively weak bonds found in this region. This means that the amount of PS in this region is expected to be larger than the quantity of styrene present in the swollen vulcanizate prior to polymerization. Strengthening the rubber–filler interface by using a silane coupling agent reduces the amount of styrene that can swell into this region. This implies an empirical rule that the larger the volume of a PS-filled void or vacuole surrounding an individual filler particle, the weaker its interaction with the rubber matrix.25
EXPERIMENTAL
Materials
The organoclay primarily used in this work was Nanofil® 8, supplied by Rockwood Additives (UK), which is composed of ∼55wt% of montmorillonite and ∼45wt% of the organic modifying agent distearyldimethylammonium. Nanofil® 5, from the same suppliers, was also used, which is composed of ∼65wt% montmorillonite and ∼35wt% distearyldimethylammonium.
The NR used was commercial SMR-L grade from Malaysia. Bis(triethoxysilylpropyl) tetrasulfide (TESPT; Si 69®) and 3-mercaptopropyl di(tridecan-1-oxy-13-penta(ethylene oxide))ethoxysilane (MPDES; VP Si 363®) were supplied by Evonik Industries (Germany; Figure 1). The silica used was Zeosil® 1165MP produced by Rhodia Silica (France), whereas the carbon black used was Sterling® V, an N660 grade black produced by Cabot (Boston, MA). The sulfur, rubber accelerators, and antioxidants (zinc oxide and stearic acid) were commercial rubber chemicals. Styrene and toluene were obtained from Sigma-Aldrich (UK).
Citation: Rubber Chemistry and Technology 86, 4; 10.5254/rct.13.87974
Preparation and physical testing of nanocomposites
All compounds were based on a standard formulation, given in Table I. However, it was necessary to adjust the nominal sulfur content of compounds containing TESPT24 because the polysulfidic groups will add extra sulfur into the formulation. The amount of sulfur used was reduced by 0.1 phr for every 1 phr of TESPT, to maintain the overall level of sulfur at 2.5 phr, as recommended by Evonik. This adjustment was not required when using MPDES.
Each compound is given a designation that specifies the type and amount of filler it contains and also the type and amount of coupling agent if one is used. The filler or coupling agent is given by a letter code (OM = Nanofil 8; OM* = Nanofil 5; CB = carbon black; Si = silica; T = TESPT; M = MPDES) followed by a number indicating the quantity in parts per hundred parts of rubber (phr). This means that, for example, OM-10/T-0.5 contains 10 phr of Nanofil 8 and 0.5 phr of TESPT. The exception is a compound produced without filler or coupling agent, which is designated as “unfilled.”
To distinguish between compounds produced in different batches that had identical formulations, a single-letter batch identifier is included at the end of all compound designations. In addition, the compound designations for batch C also include the procedure for internal mixing used for that compound. The mixing procedures used for all other batches are specified in Table II. A full list of the compounds produced in this work, including the duration of mixing for each, is given in Table III.
The internal mixer used was a Brabender Plasticorder PL-2000 with a 350S mixing head (Banbury-type rotors). The compounds were finalized on a 300 m by 150 mm two-roll mill set to ∼40 °C. All fillers and additives were incorporated during the first internal mixer stage, except for the vulcanization ingredients, which were added during finalization.
Vulcanization behavior was determined using a Monsanto MDR2000E at 150 °C. The compounds were vulcanized to t95 at 150 °C, as determined from the rheometry, in the form of 225 mm × 225 mm × 2 mm flat sheets.
Tensile testing, based on BS ISO 37:2005, was used to characterize the properties of the nanocomposites using dumbbell test pieces that were extended to failure at a cross-head speed of 500 mm/min, giving a strain rate of ∼15–20% per second. Testing was performed using an Instron 5567 equipped with a video extensometer. Three test pieces were tested for each formulation, with the median result reported.
Dynamic tensile properties of the materials over a temperature range from −30 °C to +60 °C were measured using an Instron 1271 servo-hydraulic test machine and an Instron 3119-402 temperature control cabinet. The test pieces had dimensions of 150 mm × 20 mm × ∼2 mm and were gripped with a cross-head separation of 140 mm. Testing was performed using a tensile prestrain of 2.5% and a strain amplitude of 1% at nine different frequencies increasing logarithmically from 0.1 Hz to 10 Hz.
NVTEM SAMPLE PREPARATION AND IMAGING
Vulcanizate samples for NVTEM were extracted in refluxing acetone overnight using a Soxhlet apparatus and dried in vacuo to remove any traces of solvent. The samples were cut to approximately 2 mm × 2 mm × 10 mm in size and then swollen to equilibrium in a solution of 2wt% di-n-butyl phthalate and 1.5wt% benzoyl peroxide in styrene. The polymerization inhibitor (4-tert-butylcatechol) present in the styrene was chemically extracted prior to use and replaced with diphenylpicrylhydrazyl at a concentration of 1 mg per 100 g of styrene. The swollen samples were trimmed and placed inside gelatine capsules, which were then filled with styrene solution and sealed. The capsules were placed in a metal block in an oil bath and heated to 68 °C for 24 h to polymerize the styrene. Ultrathin sections, estimated at between 100 and 150 nm in thickness, were taken from the swollen vulcanizate and embedded in PS with an LKB Ultratome V ultra-microtome at low temperature (∼ −110 °C) using a 45° glass knife set at a shallow clearance angle, collecting the sections on nickel grids. The sections were stained with osmium tetroxide vapor for 1 h and then examined with a Philips CM12 transmission electron microscope operating at 80 kV.
Measurements of bound rubber content
Samples of unvulcanized rubber (2 × ∼0.3 g per compound) were individually swollen in 50 mL of toluene in darkness. The initial samples were swollen for 1 week, whereas later samples were swollen for 3 days. The swollen gel was collected on preweighed lens tissue and weighed immediately. The gel was then dried under vacuum and reweighed. The initial, wet, and dry weights were used to calculate the bound rubber content and the volume fraction of rubber in the swollen gel.
RESULTS AND DISCUSSION
Effect of strain on modulus of NR/OMMT nanocomposites
Figure 2 compares the stress-strain behavior of NR/OMMT nanocomposites with different filler contents with NR vulcanizates containing either carbon black or a silica/silane mix, as well as an unfilled NR vulcanizate. It shows that OMMT has a very large effect on modulus at low strains but that this diminishes substantially as the test piece is extended further. In contrast, carbon black and silica do not have a large effect initially but become much more reinforcing as the material is strained. There is not a very big difference in terms of tensile strength (TS; given in legend of Figure 2) between all of these compounds. The largest TS is seen for OM-10_A, which also has the largest extension at break (EB). Both TS and EB are decreased at higher OMMT contents, due to the rubber becoming stiffer.
Citation: Rubber Chemistry and Technology 86, 4; 10.5254/rct.13.87974
The high initial modulus of the NR/OMMT nanocomposites is believed to be primarily due to a large amount of filler–filler interaction between the highly shaped OMMT particles. The filler–filler interaction is disrupted at relatively low strains, at which point the hydrodynamic effects of the filler become the most important factor. These would still be expected to be greater for OMMT than for carbon black or silica because of its high shape factor. Because both the carbon black and silica overtake the nanocomposites in terms of tensile modulus, it is clear that the hydrodynamic reinforcement provided by the OMMT is not sustainable at high strains. The reasons for this are discussed below.
Nanocomposites containing TESPT: initial work
The initial investigation of the effect of adding TESPT to NR/OMMT nanocomposites used an organoclay content of 5 phr and TESPT contents of 0 phr, 1 phr, or 2 phr. The ratios of TESPT to OMMT of 20wt% and 40wt% are very high compared with the 8wt% typically used for silica-filled compounds. However, the high specific surface area of a well-exfoliated organoclay meant that the necessity of using such high levels could not be dismissed. The mixing procedure used included two internal mixer stages to maximize the yield of the reaction of the silane with the filler.
Tensile Properties
Tensile stress-strain testing of these compounds showed that the TESPT had a noticeable effect on tensile modulus (Figure 3). However, the difference in modulus at low extensions was minimal, with the stress-strain curves diverging at approximately 40% strain. OM-5/T-1_B and OM-5/T-2_B showed identical stress-strain behavior below 200% and only a slight difference at higher strains. This could be due to the additional TESPT forming more bonds between rubber and filler or may simply result from random experimental variation between the two compounds. The TS was also not significantly affected by the addition of TESPT, although the EB did decrease (see the legend of Figure 3).
Citation: Rubber Chemistry and Technology 86, 4; 10.5254/rct.13.87974
The TESPT is not thought to increase tensile modulus by hydrophobization of the filler surface or disagglomeration of the filler particles, as it does with silica. Instead, the effect of the TESPT is believed to be entirely due to the formation of permanent chemical bonds at the rubber–filler interface. One result of this is that the rubber is less capable of dewetting from the filler surface, which would otherwise lead to cavitation around the OMMT particles. Cavities in the rubber will deform relatively easily and so will considerably decrease the modulus of the nanocomposite. Dewetting under strain is expected to be larger with OMMT than with silica because the mixing process aligns the OMMT particles in the plane of the vulcanized rubber sheet. This means that when the test pieces are strained, the stress will be concentrated at the relatively sharp edge of the particles, rather than being distributed over the face. However, because many silanol groups occur around the edge of the sheets, a silane coupling agent can produce a large number of rubber–filler bonds in this region, greatly increasing the resistance of an OMMT nanocomposite to cavitation. This will increase the modulus for a given strain and will also lead to a reduction in EB, as cavitation will mean that the nominal strain at failure is higher. The TS is not affected by the TESPT content, as this depends on intrinsic flaws in the rubber that are much larger than the size of the OMMT particles, with a typical size estimated to be approximately 75 μm.26
In addition to the prevention of cavitation, there is another method by which the TESPT could increase the modulus of a nanocomposite. By fixing some rubber molecules onto the face of the OMMT particles, their ability to slide over the surface in response to an applied stress will be substantially diminished. This is expected to have a relatively small effect on modulus in comparison to cavitation, although that would begin to have an effect at lower strains.
NVTEM Images
The NVTEM images of OM-5_B (Figure 4a) show transparent regions of PS (appearing white) within the osmium-stained rubber network. The PS regions surround the OMMT particles because of the relative weakness of the rubber–filler interface, due to the scarcity of strong bonds between the rubber and the clay surface. The majority of the PS regions are highly elongated and aligned as they reflect the shape and orientation of the OMMT particles within. There is also a small number of much larger PS regions that are believed to occur at flaws in the rubber matrix, possibly associated with residual zinc oxide particles.
Citation: Rubber Chemistry and Technology 86, 4; 10.5254/rct.13.87974
In contrast to the well-ordered appearance of Figure 4a, NVTEM images of both OM-5/T-2_B (Figure 4b) and OM-5/T-1_B showed significant degrees of disorder. The shape of the PS regions varied from almost circular to crescent shaped, with no large-scale alignment present. The difference in behavior is attributed to the premature reaction of the coupling agent with the rubber network, as excessive mixing temperatures break down the polysulfide functionality of the TESPT. Once the clay is coupled to the rubber during mixing, it will be pulled around by the motion of the flowing rubber. This will prevent the clay from becoming well orientated, although the flow process will also cause some localized alignment. It will also lead to unequal stresses at the interface, which will lead to highly distorted PS regions around the filler particles.
Figure 4c shows a TEM image of OM-5/T-1_B that has undergone styrene swelling and polymerization but has not been stained. This allows the effect of the premature coupling on the OMMT particles to be observed much more easily. The clay particles show no overall orientation, although neighboring particles can be well aligned. In a number of cases, particles appear to be bent by the forces transferred from the rubber matrix. This confirms that the distortion of the PS regions reflects that of the OMMT particles, rather than being purely due to the TESPT affecting some property of the rubber alone. Although the premature coupling seen here is not desirable, because it negates any benefits of the clay particles being highly aligned, it does demonstrate that the silane coupling agent can produce strong bonds between OMMT and NR.
Bound Rubber Content
Initial measurements of the bound rubber contents of the compounds in batch B gave results that were too small to measure accurately. To improve this, the swelling period was reduced from 1 week to 3 days, at which time the samples still formed a coherent gel. Table IV shows that the bound rubber content after 3 days was on average more than twice as large in the presence of the silane coupling agent as without it. This demonstrates again that the coupling agent is forming a strong permanent bond between the rubber and the OMMT. The two compounds containing TESPT displayed almost identical levels of bound rubber, indicating that the coupling agent was in excess. However, there was a difference in the volume fraction of rubber in the gel between OM-5/T-1_B and OM-5/T-2_B, indicating that the former swelled less overall than the latter. Although the samples were unvulcanized, the coupling agent could degrade to release sulfur during mixing, which could then form cross-links. If the mixing temperatures were higher for OM-5/T-1_B, then more sulfur would be released and more cross-links could form, which would lead to less solvent swelling into the samples.
Nanocomposites containing TESPT: effects of mixing duration
As sustained high temperatures are required for the silane to react with the silanol functionalities on the organoclay, the reaction will only begin in earnest once sufficient mixing has be done to reach the necessary temperature. This temperature must then be maintained by continued mixing until the reaction is complete. The duration of mixing required for complete reaction had not previously been established, nor was it clear if using two internal mixing stages, as is generally used with silica/silane systems, gives superior results to using a single stage. These aspects were investigated by comparing five NR/OMMT/TESPT compounds with identical formulations produced using different mixing procedures, along with a control containing no TESPT
Tensile Properties
Tensile stress-strain testing of these nanocomposites (Table V) revealed that the variation in M200 and M300 caused by changes in the mixing procedure was small in comparison to the large difference produced by the addition of TESPT. This implies that much of the silane had reacted in all cases, and longer mixing caused a limited amount of additional rubber–filler bonds. For these materials, there was no evidence that the mixing duration was having much effect on either TS or EB.
It is also noticeable in Table V that M20 is not significantly changed by the addition of the coupling agent, although there is a small but significant difference in M100. The behavior of the nanocomposites in this region is shown in more detail in Figure 5, which shows the variation of the tangent modulus, Etan, with strain. Etan–strain curves give a better indication of the filler reinforcement behavior at a specific extension than the stress-strain curves do, and those for the nanocomposites containing TESPT all diverge from that of the control compound OM-10_5min_C between 30% and 40% extension. The first to diverge is OM-10/T-1_7min_C, which Table V shows to have a generally higher modulus than the other materials. All the other nanocomposites containing TESPT have very similar tensile moduli at high strains, although Figure 5 shows that they do have noticeable differences in Etan at low strains. The TESPT appears to have been least effective in OM-10/T-1_5min_C, with the other three NR/OM/TESPT nanocomposites being difficult to distinguish.
Citation: Rubber Chemistry and Technology 86, 4; 10.5254/rct.13.87974
It is apparent from Figure 5 that the primary effect of the TESPT is to reduce considerably the rate at which Etan declines with strain at extensions of more than 30%. There are believed to be two mechanisms by which the coupling agent achieves this. As discussed above, this is attributed to decreases in cavitation and interfacial slippage due to the formation of permanent rubber–filler bonds. The TESPT has little or no impact on the decline in Etan seen below 30% strain, the cause of which is believed to be the disruption of filler–filler interactions that are not significantly affected by the addition of a silane coupling agent.
Nanocomposites containing TESPT: effects of TESPT content
Another important parameter for the use of TESPT was the effect of the silane content and in particular the upper limit on any benefit of adding TESPT. As there are a limited number of active sites on the clay surface available for reaction, it is evident that at some point it will be saturated by silane, at which point adding more will have a minimal effect on the properties of the nanocomposites. The structure of montmorillonite predicts silanol groups to be found only around the edge of the OMMT sheets, although there are likely to be defects in the structure that could lead to an unknown number of facial silanols being present also. Three separate series of NR/OMMT nanocomposites with different TESPT contents were produced, two of which used OM and one that used OM*. The 7 min mixing procedure was used for all of these compounds, as it had given the best results for the previous batch.
Tensile Properties
The tensile properties of these nanocomposites (Table VI) showed good agreement between similar nanocomposites produced in different batches, particularly when looking at M100 and M200, and also with the results given in Table V. It is also noteworthy that the two different OMMT clays gave very similar results despite the difference in their reported levels of organic modification.
Looking at the results from the different batches, it seems that about 4wt%–5wt% of TESPT is required for complete coverage of the active reaction sites on OM and about 5wt%–6wt% for OM*. Although TS did not vary consistently with TESPT content for batch D, the materials in batches E and F did show some signs of increased strength compared with their respective control compounds. This might indicate a slight increase in resistance to crack propagation due to the stronger rubber–filler interface.
NVTEM. —
The previous NVTEM images (Figure 4) showed that the TESPT was reacting with both the organoclay and the rubber. However, there were problems with premature coupling that were presumed to be due to excessive mixing temperature. NVTEM was also performed on batch D to see if this problem recurred. Typical images of the four nanocomposites are shown in Figure 6.
Citation: Rubber Chemistry and Technology 86, 4; 10.5254/rct.13.87974
These NVTEM images showed no sign of the premature coupling that was apparent previously, with all four vulcanizates displaying highly aligned elongated PS regions around the clay particles. Looking at the dimensions of these PS regions can give an indication of the strength of the rubber–filler interaction in the vulcanizate. In the absence of other influences, and assuming that the rubber originally wetted the whole surface of a disc-shaped OMMT particle, the shape of the PS region seen in the NVTEM image should be an ellipse with the foci at either edge of the OMMT particle. The strength of the rubber–filler interaction will determine the amount of swelling at the interface, which will in turn change the aspect ratio of the PS regions. In general, it should be observed that as the interaction of the OMMT with the NR matrix gets stronger, the aspect ratio of the PS regions should increase. Therefore, by comparing the aspect ratios seen for different vulcanizates, it should be possible to assess the strength of the rubber–filler interaction in the compounds. Note that this assumes the bulk cross-link density is the same in all the nanocomposites, which should be the case.
The aspect ratios in Table VII were calculated from 40 measured PS regions in two NVTEM images of OM-10_D and 20 measured PS regions in a single image of each of OM-10/T-0.2_D, OM-10/T-0.5_D, and OM-10/T-1.0_D. There is probably a significant margin of error in these results, but it certainly seems that the two nanocomposites with the highest TESPT content have considerably more elongated PS regions around their OMMT particles. This again suggests that the coupling agent can increase the strength of the rubber–filler interface in an NR/OMMT nanocomposite.
Dynamic Properties
Dynamic tensile testing of the nanocomposites in batch E showed that the tan δ behavior had a significant difference from that of black-filled or silica-filled NR (Figures 7a and 8a). Although the behavior is similar at low temperatures, where it is mostly controlled by the properties of the bulk rubber, between 20 °C and 60 °C there is a tan δ peak seen with OMMT that is not observed with the other fillers. The intensity of this peak can be diminished by the addition of TESPT. The height of this peak does not change with frequency, unlike the situation at low temperatures, where tan δ is much greater at higher frequencies. The dynamic tensile modulus (E*) also fell sharply in this region (Figures 7b and 8b), with the decline beginning slightly earlier for OM-10_E and OM-10/T-0.2_E than for the compounds with higher TESPT contents.
Citation: Rubber Chemistry and Technology 86, 4; 10.5254/rct.13.87974
Citation: Rubber Chemistry and Technology 86, 4; 10.5254/rct.13.87974
Similar tan δ peaks that have been reported for rubber–clay nanocomposites previously3,27 are attributed to rubber being intercalated between the clay sheets. The rubber molecules will be tightly constrained by the clay particles and so will undergo their glass transition at a much higher temperature than the bulk rubber. Below their glass transition, the intercalated rubber will transmit shear force between the clay sheets, making the tactoids behave like rigid blocks. Above their glass transition, the rubber molecules will deform viscously when a shear force is applied, causing energy to be dissipated. This will also lead to a loss of filler rigidity, causing a sudden decrease in E* at the same time, as can be seen in Figures 7b and 8b.
TESPT did not seem to have a consistent effect on E* at low temperatures, where any differences seemed to be due to experimental variation. This might be due to the strain history of the compound or to the test pieces' experiencing slightly different prestrains during testing. The lack of a significant effect on E* is explained by the low strain amplitude (1%) used for testing. As shown by the tensile stress–strain testing, the coupling agent increases modulus only above 30% strain, by reducing cavitation and interfacial slippage, and so no effect would be expected under these testing conditions.
The effect that TESPT had on E* and tan δ can be explained in a number of ways. The first is that the coupling agent is reacting around the edge of the sheets where it sterically hinders the intercalation of the rubber between the clay sheets. Evidence against this comes from X-ray diffraction (XRD), which shows no significant shifts in peak position between nanocomposites with and without TESPT (Figure 9). However, it is conceivable that the displacement of distearyldimethylammonium intercalant by NR molecules does not change the interlayer separation and so would not be picked up by XRD.
Citation: Rubber Chemistry and Technology 86, 4; 10.5254/rct.13.87974
A second explanation is that adding TESPT increases the exfoliation of the clay sheets and therefore reduces the amount of intercalated rubber relative to the amount of bulk rubber. The exfoliation level of the clay cannot be deduced using XRD as there are other factors that also control the intensity of the peaks. The best way to assess exfoliation is using TEM, although this is still more suited to qualitative than quantitative analysis. TEM images of NR/OMMT nanocomposites do not show a clear increase in exfoliation when TESPT is added, although a small effect is possible.
A third possibility is that the tan δ peak is not caused by intercalated NR molecules but rather molecules that are constrained by simultaneous adsorption onto two separate OMMT particles. The coupling agent, by reacting onto and coating the surface of the OMMT particles, prevents NR molecules from being adsorbed onto the surface and so reduces the intensity of the peak. If so, the peak intensity would be expected to rise quickly as the filler concentration increases and particles pack closer together, but this has not yet been tested. It is therefore still unresolved as to how TESPT decreases the intensity of the tan δ peak.
Nanocomposites containing TESPT: vulcanization behavior
The rheometry of the NR/OMMT/TESPT nanocomposites (Table VIII) showed that TESPT has only a minor effect on vulcanization behavior. In addition, the same trends were observed regardless of which type of OMMT was used. This is in stark contrast to the very large effect that organoclays such as OMMT have on vulcanization due to the organic modifier acting as an accelerator.8 There was also a significant random variation in MH-ML, the torque rise upon vulcanization, seen between different batches, although within a batch, the results showed little random error. The TESPT caused a slight increase in MH-ML and in the peak rate of vulcanization, which is believed to be due to the formation of permanent rubber–filler bonds, increasing the effective cross-link density of the vulcanizates. In most cases, the TESPT increases scorch and cure times because the disulfide and short polysulfide units act as sulfur acceptors and hence compete for reactive-free sulfur units.28 The exceptions are in batches E and F, in which the nanocomposites without TESPT showed longer cure times than expected, attributed to experimental variation.
It was also observed that the different mixing times did not have a significant effect on the scorch or cure times. However, there was some variation in MH-ML attributable to the duration of mixing, with longer mixing times leading to higher MH-ML values. This is again attributed to the presence of more rubber–filler bonds due to better reaction of the silane coupling agent.
Nanocomposites containing MPDES
MPDES has two important structural differences from TESPT, as illustrated in Figure 1. One difference is that the polysulfide functionality is replaced by a thiol group, which will be more reactive during vulcanization, although it will not react at all until an accelerator is added. This means that there is much less danger of premature coupling with MPDES than with TESPT. The other difference is that there is only one ethoxy substituent attached to the silicon atom, as opposed to three in TESPT. The other ethoxy groups are replaced by two alkylpolyether chains, which provide steric shielding for the highly reactive thiol group. These side chains also have a secondary function of interacting with the silica surface, through the polyether section, to increase its hydrophobicity.
Vulcanization Behavior
Rheometry of these compounds (Table VIII) showed that MPDES, like TESPT, had a small effect on the vulcanization behavior. It produced a slight decrease in scorch in scorch time, unlike TESPT, which lengthened it. The scorch time is controlled by how quickly the initial curatives are converted into forms that can produce cross-links, and the MPDES increases this rate because the thiol group reacts readily with the CBS accelerator. The cure time is not affected by the MPDES content, nor is the torque rise upon vulcanization. The latter is unexpected, because with TESPT, this was attributed to the formation of rubber–filler bonds that would also be expected with MPDES.
Tensile Properties
MPDES was found to have a significant impact on the tensile properties (Table IX). M100, M200, and M300 were considerably increased by the coupling agent, whereas M20 was not significantly affected in either case. This is the same type of behavior seen with TESPT, although MPDES had a smaller effect on the tensile modulus than TESPT. The likely explanation for this is that the bulky alkylpolyether groups on MPDES take up more space when bonding to the clay, restricting the total number of bonds that can be formed.
As with TESPT, MPDES is believed to act by strengthening the rubber–filler interface, preventing interfacial slippage and cavitation as the rubber is strained. Figure 10 shows how the coupling agent begins to be effective at approximately 25–30% strain, as these mechanisms become important for stress dissipation. It is expected that the filler geometry and the nature of the rubber–filler interaction in the absence of coupling agent are the most important factors governing at what strain a coupling agent becomes effective and not the type of coupling agent used. This explains why both TESPT and MPDES seem to take effect at very similar strains.
Citation: Rubber Chemistry and Technology 86, 4; 10.5254/rct.13.87974
Dynamic Properties
Investigating how the dynamic properties of NR/OMMT/MPDES nanocomposites vary with temperature found both similarities and differences to the behavior seen with TESPT. The intensity of the tan δ peak at 40 °C was again decreased by the coupling agent (Figures 11a and 12a), but in this case, the position of the peak also shifted toward lower temperatures as the quantity of coupling agent increased. Similarly, the drop off in E* seen in this region also begins at lower temperatures as more MPDES is added (Figures 11b and 12b). The most likely reason for this change in behavior is the long alkylpolyether chains found in MPDES, either by changing the amount of NR adsorbed onto the surface or by becoming intercalated between the clay sheets and affecting the glass transition of the intercalated rubber. It could even be that entire MPDES molecules are becoming intercalated, which would also have a significant impact on the dynamic behavior.
Citation: Rubber Chemistry and Technology 86, 4; 10.5254/rct.13.87974
Citation: Rubber Chemistry and Technology 86, 4; 10.5254/rct.13.87974
CONCLUSION
OMMT produces very large increases in tensile modulus at low strains but becomes less effective at high strains due to the rubber dewetting from the filler particles. Adding a silane coupling agent such as TESPT or MPDES results in a significant increase in tensile modulus when the nanocomposite is strained beyond 30–40% while having no effect below this. This is attributed to the coupling agent preventing cavitation by increasing rubber–filler interaction and was supported by use of NVTEM and bound rubber measurements. TESPT produced a greater increase in modulus than MPDES, as its smaller size allows more bonds to be formed between rubber and filler.
NR/OMMT nanocomposites also show an unusual peak in tan δ between 20 °C and 60 °C, attributed to the glass transition of intercalated and adsorbed NR molecules. Both coupling agents reduce the intensity of this peak, which is believed to happen by either increasing exfoliation or blocking the rubber from intercalating. MPDES also shifts the location of the peak to lower temperature, which is thought to be due to partial intercalation of its long alkylpolyether subgroups.
Chemical structures of TESPT and MPDES.
Tensile stress-strain properties of OMMT-filled, carbon black–filled, silica-filled, and unfilled NR.
Tensile stress–strain properties of NR/OMMT nanocomposites with and without TESPT.
NVTEM images of (a) OM-5_B, (b) OM-5/T-2_B, (c) OM-5/T-1_B before staining.
Variation of tangent modulus with strain for NR/OMMT/TESPT nanocomposites with different mixing procedures.
NVTEM images of (a) OM-10_D, (b) OM-10/T-0.2_D, (c) OM-10/T-0.5_D, (d) OM-10/T-1_D.
Variation of (a) tan δ and (b) E* with temperature at 0.1 Hz for NR/OMMT/TESPT nanocomposites.
Variation of (a) tan δ and (b) E* with temperature at 10 Hz for NR/OMMT/TESPT nanocomposites.
X-ray diffraction patterns for NR/OMMT nanocomposites with and without TESPT.
Variation of tangent modulus with strain for NR/OMMT/MPDES nanocomposites with different mixing procedures.
Variation of (a) tan δ and (b) E* with temperature at 0.1 Hz for NR/OMMT/MPDES nanocomposites.
Variation of (a) tan δ and (b) E* with temperature at 10 Hz for NR/OMMT/MPDES nanocomposites.
Contributor Notes