Preparation and Properties of SBR Composites Containing Graphene Nanoplatelets Modified with Pyridinium Derivative

The goal of this work was to study the effect of graphene nanoplatelets (GnPs) modified with ionic liquid (IL) on properties of styrene–butadiene rubber (SBR) composites. GnPs were decorated with IL or were modified in bulk directly during rubber mix preparation. The ionic liquid used was 1-butyl-4-methylpyridinium tetrafluoroborate (BMPFB). The textural properties were studied to confirm surface modification of GnPs with BMPFB. In these investigations, the impact of the concentration of GnPs and the effect of the method of GnPs’ modification with IL on elastomers properties are described. Some thermal measurements (e.g., differential scanning calorimetry and thermogravimetry) were conducted to characterize the thermal stability or the vulcanization process of the investigated samples. Complementary experimental techniques were used to investigate the properties of the obtained elastomers, specifically tensile testing, and electrical and barrier property measurements. The deposition of IL on the GnPs’ surface positively influenced the mechanical and barrier properties of elastomers. However, samples containing graphene nanoplatelets modified from solution were characterized by less electrical conductivity. SEM analysis was additionally performed to investigate GnPs’ dispersion within SBR composites.


Introduction
Styrene-butadiene rubber (SBR) is one of the most widely used polymers in the rubber industry. The processing of polymer materials and possible application of composites, and also of SBR, are significantly dependent of fillers dispersed in the polymer matrix. In order to broaden polymers materials' usefulness, reinforcing fillers must especially be applied. In recent years, much attention has been paid to multifunctional nanofillers, which could not only improve mechanical properties but increase electrical conductivity, as well as guaranteeing low gas permeability and the thermal stability of composites [1,2].
Graphene and other carbon allotropes have been deeply studied and their new possible applications in composites have been investigated [3][4][5][6] due to their exceptional properties [7]. Up to now, numerous researchers worked on graphene-based polymer composites [8][9][10]. However, many difficulties still need to be overcome to obtain graphene-based multifunctional composites that meet requirements. The most important limitations are: the problem of homogenous filler's dispersion with minimal GnPs/BMPFB were prepared through the solution method. The ionic liquid was incorporated on GnPs' surface from acetone, according to the following protocol: GnPs' powder was dispersed easily in (CH 3 ) 2 CO with the chosen IL and then was twice under sonication for 15 min, with 15 min of break. After 48 h, the resultant mixture was irradiated with ultrasound for a quarter of an hour. After 24 h, this suspension re-underwent ultrasonic treatment for 15 min. Subsequently, the resulting product was dried using a vacuum oven at 50 • C, after solvent evaporation. Thus, modified filler was used as the rubber mixes' ingredient. The content of IL incorporated onto GnPs surface was determined using TG curves (described in the Results and Discussion section). It was necessary to develop the rubber mixes' recipes, to ensure that all rubber compounds had equivalent GnPs concentration.
Rubber compounds were prepared (at temperature about 35 • C) using a laboratory two-roll mill (cylinders dimensions were: D = 150 mm and L = 300 mm) (Bridge, UK). This method was chosen to secure good processing and minimize heat buildup in rubber compound during compounds' mixing: (i) the rolls were cooled down by cold water circulating through them, (ii) the friction between the cylinders was 1.1, (iii) the gap between rolls, time of mixing and cutting operations were adjusted. After plasticization of raw rubber, other ingredients necessary to produce rubber mix were added. Rubber sheets with a thickness of about 6-8 mm were prepared. After 48 h-seasoning at room temperature of rubber mixes, their optimal curing time was qualified using rheometric measurements. Afterwards, rubber compounds were moulded in an electrically heated hydraulic press at 160 • C to form the vulcanizates.

Characterization
Methods of investigations, measuring apparatus and conditions, as well as parameters obtained from these measurements, are presented in Table 2. Scheme 1. Diagram of the apparatus for gas permeability measurements of studied styrene-butadiene rubber composites.

Textural Properties of Fillers
The modification of the GnPs' surface with BMPFB from solution was certified by measurements of the textural properties of both pristine GnPs and GnPs/BMPFB. The results are presented in Figures 1  and 2 and Table 3.    Figure 1 compared the adsorption and desorption isotherms of non-modified GnPs or those modified with BMPFB from solution. An isotherm concave to the P/P0 axis adsorption at low pressure was clearly observed for pristine GnPs, which was associated with the presence of some micropores (pore size less or equal 2 nm). The isotherm, here discussed showed also a hysteresis phenomenon that indicates capillary condensation phenomena and the presence of mesopores (pore size between 2 and 50 nm). In this case, some pores exceed certain critical widths (typically Kelvin's radius). Then, in addition to the adsorptive-adsorbent interactions, some adsorbent-adsorbent interactions appear after condensation of molecules under a multilayer form to complete the pore volumes (mesoporous volume).
In the case of the treatment of GnPs with BMPFB, the concave part of the isotherm at very low pressure is not present: a large part of filler's microporosity would therefore be cancelled. Inversely, the treatment shows an increase in total adsorbed volume (reaching value almost two times higher than that of pristine GnPs), see Table 3. It is possible that BMPFB, which is flat around the pyridinium ring, penetrates graphene sheets that have a platelet shape, leading to some interactions between the filler and the charged ionic liquid. Thus, a huge increase in V MAX is observed. The treatment might expand the micropores to achieve the mesoporous size.
The specific surface area of the investigated fillers was determined using the Brunauer-Emmett-Teller BET model [49] and the Rouquerol curve to determine the application domain [50]. GnPs/BMPFB had a more developed surface compared to the unmodified filler ( Table 3). The Dubinin-Astakhov model, which is commonly used, makes it possible to study the microporosity of the carbon [51]. In this last case, we can observe that the microporous part is more important in the pristine GnPs (this could be foreseen by the behaviour at very low pressure). The filler's microporous surface and the microporous volume of both kinds of fillers are given in Table 3.
It was interesting to simulate the pore size distribution to know more completely the effect of the treatment. In order to achieve this, we used the DFT model [52]. For the pristine GnPs, the observed distribution was obtained by using a slit model on the carbon surface (closed porosity, As = 6) ( Figure 2). In this case, we observe four different peaks: d1 = 0.6 nm; d2 = 1.4 nm; d3 = 3 nm and d4 = 5 nm (the large distribution might be attributed to the inter-particular space due to the organization of the particles).
In the case of the modified graphene, two models had to be used to simulate more neatly the isotherm: a cylindrical model at low pressure (P/P0 < 0.1 nm) and, subsequently, a slit model (the separation of both the domains was realized according the mono-model divergence). We can note that the domains approximately separate the micropores and the mesopores. Then, we can observe that the treatment causes the micropores to disappear for the benefit of a small amount of larger cylindrical micropores (2 nm), but for greater benefit of the previous mesopores (d3 and d4) in the pristine carbon. The equivalent inter-particular spaces between both of the materials seems to show that the pores were becoming filled with molecules of modifier (BMPFB was entrapped in filler's pores), resulting in a swelling of the pores. However, as we can observe in Figure 1, the volumes of the principal pores after treatment are largely higher than the ones of the pores in the pristine material. This could be explained by the presence of ultra-micropores (inferior to 0.8 nm) in GnPs. Some divergences can be commonly observed in the analysis of this kind of pore by N 2 adsorption.

Fillers Dispersion
Scanning electron microscopy was useful in estimating of GnPs' agglomeration within composites. Figure 3 shows SEM images of cross-sections of the investigated composites prepared by freeze-fracturing. Samples with 1 or 4 phr of filler were studied. Regardless of the method of GnPs' modification with BMPFB, the tendency for GnPs to stick together and agglomerate was visible. Close-up images revealed stacking of graphene platelets. This disclosed an insufficient level of filler's dispersion. At 4 phr of filler, some carbonous sheets were extracted from the polymer matrix, suggesting poor affinity to the polymer [53] (this is more visible for GnPs + BMPFB, see Figure 3b). This in turn can lead to a weak interphase with some voids in it [54]. However, after the GnPs' modification from solution, SEM images registered some protuberances concealing filler's particles in the polymer matrix. Graphene layers were wrapped by a film of SBR, thus marking a good compatibility between filler and polymer. By modification of GnPs with BMPFB from solution, the interfacial adhesion between rubber and GnPs has been improved. As the result of the above-mentioned GnPs' treatment, no significant interfacial void formation close to the filler particles was exposed. A good compatibility of the filler particles with the polymer is crucial when obtaining polymer composite with satisfactory mechanical and barrier properties.

Thermal Properties of Fillers, BMPFB and SBR Composites
Both TG and DTG analysis were useful tools for determining the thermal behaviour of BMPFB, untreated GnPs and GnPs modified with BMPFB. Some data are presented in Figure 4. In the studied temperature range, graphene was thermally stable, which was manifested by a loss of only 3% of its mass. It can be seen that thermal decomposition of BMPFB is a process with only one stage. This process is almost complete with approximately 100% mass loss. In addition, the highest BMPFB decomposition occurred at 443 • C. Since ionic liquid was settled on graphene surface, thermal stability of this filler was deteriorated by almost several dozen • C when compared to BMPFB alone. This was associated with IL degradation on the filler's surface. It was obvious from the DTG curve that GnPs/BMPFB degradation was one-step process and was manifested as a large endothermic peak at less than 400 • C. In the temperature range of 40-100 • C, there is a loss of functional groups on the GnPs' surface (this is also evidenced by DSC thermograms, see Figure 5). At that time, the greatest thermal effect accompanied the heating of pristine GnPs, in which functional oxygen groups were not involved in interactions with the modifier. The lowest value of the thermal effect related to the loss of functional groups from the filler's surface was noted for GnPs/BMPFB, where the interaction of functional groups-BMPFB was the most intense, see Figure 5.  Knowing the mass loss in a sample of crude graphene, we calculated the mass loss in GnPs/BMPFB associated only with decomposition of BMPFB on the filler's surface. When calculating the BMPFB content in the modified sample, we compared the mass loss of the sample to its total initial mass.
The effect of GnPs on thermal stability of SBR composites was also studied and data are collected in Table 4. Several papers indicated that the addition of GnPs into the polymer can significantly improve the thermal stability of the obtained material [20]. T 02 -initial decomposition temperature of sample, T 50 -temperature for 50% mass loss during sample decomposition (standard deviation of T 02 and T 50 ± 1 • C).
The temperature for 2% mass loss (T 02 ) during sample decomposition was 249 • C for unfilled composite. The application of GnPs modified with BMPFB to fill SBR composites resulted in the increasing of T 02 with increasing filler concentration, suggesting that incorporation of GnPs modified with ionic liquid enhanced the thermal stability of SBR. Compared to unfilled elastomer, when the GnPs' loading was 4 phr, T 02 was only 4 • C higher. It was reported that at 5 phr of GnPs, T 02 equal to 237 • C was recorded. Such deterioration of thermal stability of elastomers could be due to worse filler dispersion in the polymer promoting the effect of composite thermal degradation. Taking into account that the measurement error of GnPs decorated with BMPFB led to a slight increase in the thermal stability of composites in the range of 249-257 • C, then, GnPs with surfaces that were modified with ionic liquid were a physical barrier to the SBR matrix that delayed oxygen permeation and the escape of volatile degradation products [17]. In the case of temperature for 50% mass loss during elastomer decomposition (T 50 ), changes were not so spectacular. Generally, T 50 increased in a matter that was not proportional with filler loading, when GnPs were modified with BMPFB in mass. On the other hand, as the loading of the filler decorated with BMPFB increased, T 50 increased from 305 to 311 • C.

Cure Charcteristics
Curing characteristics including minimum torque (M L ), maximum torque (M H ), scorch and optimal curing time (respectively, TC2 and TC90) are shown for unfilled rubber and rubber mixes filled with GnPs in Table 5. M L is a measure of uncured compounds' viscosity and M H is a measure of the stock modulus of the cured compounds [55]. The introduction of GnPs into rubber mixes caused an increase in M L and M H with filler loading when compared with the control sample, i.e., unfilled. These changes were more pronounced when GnPs were decorated with BMPFB. It is known that an efficient elastomer curing is fundamental to produce composites with satisfactory properties. Some authors noticed that the presence of carbon filler led to TC2 shortening, probably due to an increment in thermal conductivity in presence of those fillers that promoted fulfilment of curing [56]. Our research did not confirm this. Conversely, the addition of modified graphene and its amount did not significantly affect TC2, considering measurement discrepancy. These observations are in line with previous studies regarding carbon-based fillers and their role in delaying the onset of vulcanization [57]. The basic accelerators' additives are absorbed on the filler's surface, which also led to an increase in TC90. This occurs owing to the active centers on the filler's surface, which favor the absorption of curing system compounds, e.g., accelerators of curing. Such behaviour was noticed for carbon fillers [58]. Generally, while an increase in TC2 may be a desirable phenomenon that increases the safety of blends' processing, an increase in TC90 is disadvantageous from a technological point of view. Note that when filler concentration increased to 4 and 5 phr, TC90 shortened, including standard errors (in the case of rubber mixes with GnPs and BMPFB introduced separately into SBR). Such an effect could be assigned to the filler's ability to moderate the vulcanization reaction. Previously, it was postulated that GnPs accelerate the reactions between curing system compounds during active complex formation [35]. Presumably, as GnPs' loading is higher, some part of the filler absorbed curing accelerators and the rest participated in interactions with curing system compounds. Moreover, the presence of GnPs (however modified with BMPFB) caused an increase in torque increment during vulcanization (∆M being the difference between M H and M L , and the measure of crosslink density of elastomers [59]) in relation to unfilled sample. It is probable that the interaction in the filler-polymer interphase was then higher.
Applying DSC analysis (a commonly used technique to study the curing kinetics of polymers [40]), the impact of GnPs modification with BMPFB on the temperature and energetic effects of vulcanization was investigated.
The results for filled compounds and reference sample are shown in Figure 6 and are summarized in Table 5, accounting for standard errors. From the presented data, it appeared that vulcanization of unfilled rubber mix occurred in a range of 154-230 • C. The energetic effect was 10.1 Jg −1 . It should be remembered that lowering the onset vulcanization temperature is beneficial for technological reasons because it allows rubber to be crosslinked at lower temperature. When GnPs were modified with BMPFB through the melt mixing method, the T C and ∆H C of those samples showed no systematic change with filler loading. It is worth noting that among the samples discussed, the lower the amount of GnPs, the lower the onset vulcanization temperature was. When a compound contained only 0.5 phr of filler, this parameter was equal to 144 • C (for SBR/GnPs + BMPFB compound) and 146 • C (for SBR/GnPs/BMPFB compound). However, 5 phr of GnPs in compounds led to an increase in onset vulcanization temperature, with it reaching 159 • C and 158 • C (respectively, for SBR/GnPs + BMPFB and SBR/GnPs/BMPFB). The endset vulcanization temperature was higher by a few • C compared to the value of this parameter characterizing the sample without filler. The exception was a sample with 2 phr GnPs. In the case of samples filled with GnPs decorated with BMPFB, it was noticeable that the onset and endset vulcanization temperature, respectively, were raised and lowered with the increase in filler loading. In turn, ∆H C decreased with higher amounts of GnPs, on which BMPFB was deposited. It is obvious that using different GnPs modified with BMPFB led to no significant changes in the T g values of elastomers in comparison to the unfilled sample (Table 5). These results indicated that the incorporation of GnPs modified with BMPFB did not affect the average relaxation of bulk segments [54]. These results did not demonstrate any tendency regarding filler loading or the methods of its modification. Therefore, from the technological point of view, it can be said that there are no differences in the temperature range in which these elastomers can be used.

Mechanical Performance
The tensile modulus at 100% elongation (M100), tensile strength (TS) and elongation at break (EB) of the investigated elastomers are shown in Table 6. The mechanical strengthening of polymer composites can be achieved, e.g., by filler's functionalization [15]. It is likely that filling the composites with GnPs in the presence of BMPFB would have an effect on the mechanical properties of the investigated samples since IL containing an aromatic part in its structure was treated as a coagent of sulfur curing. Hence, the pyridine rings from IL form hard domains within the polymer structure after curing [60]. If so, BMPFB could participate in the transferring of stress via "filler" particles or act as a stress-redistributor when polymer chains scission occurs. Among the tested samples, the lowest M100 value with regard to standard deviation was recorded for unfilled composite. Higher M100 values were exhibited by composites containing GnPs modified from solution. Modification of GnPs with BMPFB from solution resulted in the enhancing of the filler's roughness and the increasing of its surface area and thus generated, on the filler's surface, a higher number of contact points with the polymer. In order to ensure effective load transfer from GnPs to the polymer, such treatment was advantageous. BMPFB used to modify GnPs hindered their stacking in polymer to a certain extent (see Figure 3), leading to better filler dispersion. The prominent reinforcement effect of GnPs could be attributed to their better dispersion and stronger filler-polymer and filler-filler interactions. Therefore, better filler dispersion led to effective stress transfer at the interfacial region and enhanced mechanical properties of samples. Similar conclusions were drawn in the case of carbon nanotubes [61]. At the highest concentration of GnPs modified with BMPFB from solution, TS was more than 100% higher when compared with sample containing the lowest loading of such filler, and was 2.5-fold higher when compared to unfilled sample. Taking into account the measurements' discrepancy, no significant differences in TS were achieved for vulcanizates containing from 2 to 4 phr of GnPs, regardless of how it was modified. Usually, loading composites with fillers leads to a decrease in EB [59]. For the studied filled samples, the opposite effect was observed. It was noted that graphitic fillers were susceptible to disturbing the physical crosslinking sites of styrenic-based elastomers, thus facilitating polymer chain movement [62]. Herein, the increase in the elongation at break could be explained by the plasticizing effect of the neat GnPs, enabling the polymer chain movement. Even if development of rubber composites with a combination of high strength and higher extensibility simultaneously is a challenge, the results show that SBR composites containing GnPs modified with BMPFB (through two different methods) simultaneously had improved TS and EB. Such an effect may be a sign of additional crosslinks' formation. Owing to this, the mechanical stress acting on the filler was better dispersed in the polymer, and filler particles pulling out from the polymer matrix were reduced. This was in agreement with SEM images. Shore A hardness (H) of all composites are provided in Table 6. As expected, the filled composites had higher H when compared with unfilled sample (taking into account standard error). The more uniform the GnPs' dispersion within the polymer, the harder the composites [12]. Thus, regarding the measurement error, H values that were slightly higher when GnPs were decorated with BMPFB can be explained by stronger interfacial interactions and better filler dispersion.

Electrical Propertiesof SBR Composites
Several factors, e.g., concentration as well as filler dispersion within polymer have crucial meaning for the enhancement of electrical conductivity of polymer composites [63]. One of the ways to increase the electrical conductivity of elastomers is to introduce conductive fillers, for example GnPs [56]. Figure 7 shows the electric volume resistivity (ρ) of SBR composites fabricated with the use of GnPs modified during rubber mix preparation on two-roll mill or from solution. Additionally, reference sample was investigated. Thereby, GnPs' loading in investigated composites changed from 0 to 5 phr (and was equal to 0, 0.23, 0.45, 0.91, 1.36, 1.82 and 2.27 vol.%, respectively). Surprisingly, adding GnPs and BMPFB separately to rubber mix did not affect the electrical conductivity of composites. Despite the presence of ionic liquid and filler with exceptional electron mobility, the ρ of these composites remained practically unchanged compared to unfilled composite. Usually, at higher concentrations of conducting filler and with finer dispersion, the inter-filler distance becomes smaller, which enables the formation of paths for electron mobility [8]. It is worth noting that applying GnPs decorated with BMPFB did not give rise to a significant increase in electrical conductivity. On the contrary, despite better dispersion of these GnPs within SBR, the formation of conduction pathways was difficult. Otherwise, their presence would guarantee an increase in electrical conductivity. It is likely that the mobility of ions of BMPFB, which was immobilized on the surface of GnPs, was then significantly reduced. This way BMPFB deposited on GnPs surface was not active in the electric conductivity process and additionally blocked GnPs, thus preventing the movement of electrons. This is supported by ρ values that were higher than those of composites containing GnPs and BMPFB introduced into rubber mixes separately. ρ was reduced by only a few units with increasing filler loading in SBR/GnP + BMPFB composites. Certainly, even at the high GnPs concentrations used in the investigated SBR vulcanizates, the percolation threshold was not achieved.

Gas Barrier Properties
In this study, the gas barrier properties of SBR composites filled with GnPs as a function of filler loading and the method of GnPs' modification were investigated. The gas permeability coefficient (P) and gas transfer rate (GTR) are collected in Table 7. In the case of unfilled composite, P was 2.57 mole m −2 s −1 Pa −1 and GTR was 2.1 × 10 −4 m 2 s −1 Pa −1 . Taking into account standard errors, the addition of GnPs and BMPFB into composites led to P and GTR increments with increasing filler loading from 0.5 phr to 3 phr, reaching a value of 4.93 mole m −2 s −1 Pa −1 and 4.0 × 10 −4 m 2 s −1 Pa −1 (respectively, for P and GTR). Further increasing of GnPs concentration caused a decrease in P and GTR. This may indicate a deterioration of GnPs' dispersion in composites. At higher GnPs loading, despite the worse dispersion, filler acts as a gas barrier, which extends the diffusion path through polymer. It was discovered previously that fine dispersion of graphene sheets within polymer prolonged tortuous paths for gas diffusion [64]. It appeared to be the case that GnPs decorated with BMPFB had a beneficial effect on the barrier properties of composites. This could indicate better filler dispersion within the polymer matrix.

Conclusions
Graphene nanoplatelets modified with 1-butyl-4-methylpyridinium tetrafluoroborate were prepared using two different strategies of modification. The first strategy involved introducing ionic liquid with filler during the rubber mix preparation (melt-mixing method), and the second strategy involved the incorporation of modifiers onto the filler's surface from solution. This strategy of modification led to significant changes in the textural properties of thus obtained filler in comparison with untreated graphene nanoplatelets. As the specific surface area of modified filler developed its dispersion in the polymer was enhenced thus influencing on some composites' properties. This new strategy of filler modification with ionic liquid was advantageous for load transfer effectiveness. Styrene-butadiene rubber composites containing modified GnPs improved TS and EB simultaneously. Among the two kinds of fillers used, graphene nanoplatelets decorated with ionic liquid had a beneficial effect on barrier properties and led to an enhancement in the thermal stability of composites. Despite its better dispersion in elastomers, no increase in elastomers' electrical conductivity was observed. It is likely that this filler limited the formation of conduction pathways in composites.