Use of TBzTD as Noncarcinogenic Accelerator for ENR / SiO 2 Nanocomposites : Cured Characteristics , Mechanical Properties , Thermal Behaviors , and Oil Resistance

This study reported the use of tetrabenzylthiuram disulphide (TBzTD) as a noncarcinogenic accelerator in a traditional sulfur curing system of epoxidized natural rubber (ENR)/nanosilica (nSiO 2 ) composites. ENR used in this work was synthesized via in situ epoxidation of natural rubber (NR) in the presence of performic acid generated from the reaction of formic acid and hydrogen peroxide at 50C for 8 h to acquire the epoxide content of about 40mol%. Accordingly, the resulting ENR was referred to as ENR 40.The curing characteristics, mechanical properties, thermal behaviors, dynamicmechanical properties, and oil resistance of ENR 40/nSiO 2 nanocomposites filled with three loadings of nSiO 2 (1, 2, and 3 parts per hundred parts of rubber) were investigated and compared with NR and neat ENR 40. The results revealed that the scorch and cure times of ENR 40/nSiO 2 nanocomposites were slightly longer than those of NR but slightly shorter than those of ENR 40. The tensile properties and tear strength for both before and after aging of all ENR 40/nSiO 2 nanocomposites were higher than those of ENR 40, while the glass transition temperature, storage modulus at −65∘C, thermal stability, and oil resistance of ENR 40/nSiO 2 nanocomposites were higher than those of NR and ENR 40.


Introduction
Natural rubber (NR) is one of the most important natural biosynthesis polymers and renewable resources that possesses excellent mechanical properties owing to its ability to crystallize upon stretching (strain-induced crystallization), but its thermal, oxidation, and oil resistance cannot compete with some special purpose synthetic rubbers [1][2][3][4][5][6][7][8][9][10][11][12].This is a normal consequence of its highly unsaturated hydrocarbon chain structure and nonpolar nature [2][3][4].However, the certain properties of NR can be improved via chemical modifications at the carbon-carbon double bonds (C=C) in its isoprene unit to produce functionalized specialty rubbers with an expanded range of physical and chemical properties.The controlled partial epoxidation of NR to form epoxidized NR (ENR) is one of the important methods that introduce a new reactive group into the NR backbone, leading to the new and useful properties.This involves the random introduction of polar epoxide (oxirane) groups onto the C=C in the repeat unit of NR by the two consecutive reactions.The first one included the "in situ" formed peracid generated from the endothermic reaction of hydrogen peroxide (H 2 O 2 ) and the acid, followed by the exothermic epoxidation reaction, leading to an increased polarity and a reduction in the C=C in the rubber chains [5,8,9,[12][13][14][15][16][17][18].As a consequence, ENR would have oil and oxidation resistance comparable to some of the synthetic rubbers, depending on the extent of epoxidation (mol% epoxidation) [5,9,10,12].It has been reported that ENR with 50 mol% epoxidation exhibits oil resistance similar to that of medium-acrylonitrile-content nitrile rubber and superior to that of chloroprene rubber [9,12,19].As the level of epoxidation increases more than 50 mol%, the ENR products become harder and of less resilience and elasticity at room temperature (RT), whereas the products of lower epoxide content are typical elastomers with less oil resistance [8,10,17].In this study, ENR with medium epoxide content (approximately 40 mol%) was synthesized from concentrated NR latex to impart higher oil resistance compared to the NR while maintaining both elasticity and strain crystallization of the products.The resulting ENR was then vulcanized by a traditional sulfur curing system using an organic accelerator in a usual manner applicable for NR.Tetrabenzylthiuram disulphide (TBzTD, C 30 H 28 N 2 S 4 , Scheme 1) was selected for using as a safe (noncarcinogenic) accelerator to reduce the generation of hazardous nitrosamines in rubber compounds.Nitrosamines are generally generated from the rubber products during mixing, curing, and storage when using amine containing accelerators such as tetramethylthiuram disulfide (TMTD) [7,20].With this respect, the TBzTD, a fast curing accelerator with no secondary amine, has been developed to replace the TMTD, resulting in the nitrosamine free cure.
Recently, reinforcement of rubbers with inorganic nanoparticles, even at a very low loading, can significantly improve their mechanical and thermal properties [9,10,19,[21][22][23][24][25].To achieve the desired properties, the nanofillers must sufficiently disperse in the rubber matrix, as poor dispersion leads to the obvious decrease in the performance properties of the nanocomposites.Nanosilica (nSiO 2 ) is an inorganic filler with a large number of hydroxyl (-OH) groups on its surface and thus provides strong filler-filler interactions to generate nSiO 2 aggregates and agglomerates.However, the epoxide groups of ENR can effectively cause a specific interaction between nSiO 2 and ENR via hydrogen bonding, leading to a decrease in filler-filler interaction and an increase in compatibilization [9,[22][23][24].With this respect, ENR nanocomposites can be directly prepared using nSiO 2 as a nonblack reinforcement without the aid of any coupling agents or pretreatments [9,12].Therefore, their unique properties can be combined in order to improve the application properties of the resulting nanocomposites.To the best of our knowledge, use of the TBzTD as the nitrosamine free accelerator in the sulfur curing of the ENR/nSiO 2 nanocomposites has never been reported.Therefore, the mechanical properties and thermal and dynamic mechanical properties, as well as oil resistance of the prepared ENR/nSiO 2 nanocomposites, were investigated and compared with those of the neat NR and ENR.

Preparation and Characterization of ENR.
The ENR was synthesized via in situ latex epoxidation of NR by performic acid generated from the reaction between HCOOH and H 2 O 2 .Prior to epoxidation, NR latex was diluted with distilled water to reduce the DRC from 60% to 20% and then stabilized with the nonionic surfactant Terric 16A-16 (3 parts by weight per hundred of rubber, phr) at RT under stirring for 1 h.The assigned amounts of HCOOH (1 M of isoprene unit) and H 2 O 2 (0.75 M of isoprene unit) were then added to the stabilized NR latex.The HCOOH was fed dropwise over 10-15 min at 40 ∘ C, followed by raising the temperature to 50 ∘ C within 10 min prior to adding the H 2 O 2 over 5-10 min with continuous stirring.The reaction was allowed to proceed at 50 ∘ C for 8 h.The obtained products were sampled and coagulated by an excess 95% (v/v) methanol, filtered, washed well with distilled water, soaked in 10% (w/w) sodium carbonate solution for 30 min, washed again until being neutral, pressed into a thin sheet, and dried at 60 ∘ C for 48 h until at a constant weight before analysis.
The mol% epoxidation of the resulting ENR was examined by Fourier transform infrared (FT-IR) spectroscopy on a Nicolet 6700-FT-IR (USA) over a frequency range of 400-4000 cm −1 and by proton nuclear magnetic resonance spectroscopy ( 1 H-NMR) on an Avance III HD 500 MHz NMR spectrometer (Bruker, Germany).The mol% epoxidation was calculated from (1) using data from the FT-IR peaks at 870 and 835 cm −1 [14,26].
where  870 and  835 are the absorbance at wavenumbers 870 and 835 cm −1 , respectively.
To estimate the mol% epoxidation from the 1 H-NMR analysis, two characteristic peaks were considered in relation with (2) [8,9,13,24]: where A 2.7 and A 5.14 are the integrate areas under the 1 H-NMR peaks at about 2.7 ppm, representing the epoxide proton, and at about 5.14 ppm, representing an olefinic proton, respectively.The glass transition temperature ( g ) of the sample was evaluated by differential scanning calorimetry (DSC) on a Mettler Toledo DSC 1 STARe System (Switzerland) at a heating/cooling rate of 5 ∘ C min −1 under a N 2 flow of 20 mL min −1 from −80 to 20 ∘ C.

Preparation and Characterization of ENR/nSiO 2 Nanocomposites.
The as-synthesized ENR was compounded with nSiO 2 (0, 1, 2, and 3 phr) by latex compounding.The obtained nanocomposites were sampled and coagulated by an excess 95% (v/v) methanol, filtered, washed well with distilled water, soaked in 10% (w/w) sodium carbonate solution for 30 min, and washed again until being neutral.The ENR/nSiO 2 nanocomposites were mixed with ZnO (5 phr), stearic acid (2 phr), TBzTD (6 phr), and sulfur (0.5 phr) in an internal mixer (MX500-D75L90, Thailand) at 70 ∘ C for 9 min and then sheeted on an open two-roll mill at RT for 1 min.The cure characteristics of the different stocks were determined using a moving die rheometer (MDR, A0225-rheo Tech Techpro, USA) at 150 ∘ C according to ASTM D5289.The respective scorch time ( s2 ), cure time ( 90 ), minimum torque ( L ), maximum torque ( H ), and torque difference ( H −  L ) were obtained from the rheograph.The cure rate index (CRI) was then calculated from (3) [1,7,18]: The ENR well-mixed compound was then press-cured into 3 mm thick sheet using a compression molding machine (LP-S-20, LabTech Engineering, Thailand) at 150 ∘ C according to the t 90 obtained from the MDR.For comparison, NR vulcanizate was also prepared using the same sulfur curing process as above.

Testing and Characterization of the Vulcanizates.
The tensile properties (tensile strength, modulus at 300% strain (M300), and elongation at break) of the vulcanizates were measured on the standard dumbbell-shape specimen according to ISO 37 Type II using an Instron universal testing machine (5840 series single column systems, USA) with a 1 kN load cell and at a crosshead speed of 500 mm min −1 .The tear strength was measured according to ASTM D624 on an angle-shaped sample with the same universal testing machine at a strain rate of 500 mm min −1 .At least five specimens were examined for each test.
The thermal aging of the sample was performed in a universal oven (Model UF55, Germany) with air circulating system at 70 ∘ C for 72 h according to ASTM D573.After aging, the specimen was conditioned at RT for at least 16 h prior to determining the tensile properties and tear strength in the same manner as the unaged specimen.The percentage retention in properties was calculated from the following equation [4]: Value after aging Value before aging The thermal stability of the sample was measured by means of thermogravimetric analysis (TGA) on a Mettler Toledo, TGA/SDTA 851 e instrument (Switzerland), operating under a N 2 atmosphere at a heating rate of 20 ∘ C min −1 within the temperature range of 50-1,000 ∘ C. The dynamic mechanical analysis (DMA) of the sample was performed on a Mettler Toledo DMA/SDTA 861 e instrument (Switzerland) under a shear mode at a constant frequency of 1 Hz and a heating rate of 3 ∘ C min −1 over a temperature range of −50 to 50 ∘ C and under a N 2 flow.
The oil resistance was determined by measuring the percentage volume change of the sample after a 72 h immersion period in two types of oil, ASTM oil number 1 and number 3, at 100 ∘ C. Prior to immersion, the test specimen was weighted in both air and distilled water at RT.After immersion, the test specimen was removed from the respective oils, cooled to RT by transferring to a cool and clean test oil for about 30 min, and then dipped quickly in acetone at RT and wiped lightly with filter paper to remove the excess oil from the surface prior to weighing again in both air and distilled water.The percent change in volume was then evaluated using (5) according to ASTM D471-79: where M 1 and M 2 are the initial mass (g) of the specimen in air and distilled water, respectively, and M 3 and M 4 are the mass (g) of the specimen in air and distilled water after immersion, respectively.

Results and Discussion
3.1.Characterization of Unvulcanized Samples. Figure 1 shows the representative FT-IR spectra of the samples over a frequency range of 4000-400 cm −1 .The characteristic peaks of NR (Figure 1(a)) that appeared at 3037, 2850, 1660, 1450, 1375, and 835 cm −1 are assigned to the =C-H stretching, C-H stretching, C=C stretching, C-H bending of CH 2 , C-H bending of CH 3 , and =C-H bending, respectively [3,8,15,24].The new characteristic peaks that appeared at 1240 and 870 cm −1 (Figure 1(b)) are assigned to the symmetric stretching and asymmetric stretching of epoxide ring, respectively [5,8,15,16,18,[24][25][26][27].This confirmed the formation of epoxide or oxirane rings from the reaction of performic (generated in situ by the reaction of HCOOH and H 2 O 2 ) with the C=C on the NR molecules.The absorbance at 870 cm −1 was then compared with that of the residue olefinic peak (C=C cis) at 835 cm −1 to evaluate the mol% epoxidation.In addition, the broad peak at about 3400 cm −1 (Figure 1(b)) indicates the presence of -OH groups in the ENR molecules, resulting from a partial opening of epoxide ring during ENR preparation [5,15,16,24].Figures 1(c) and 1(d) also show broad peaks at about 3400 cm −1 , which are mainly owing to the O-H stretching vibration of -OH groups on nSiO 2 surfaces and from the opening of epoxide ring, and so the peak size increased with increasing nSiO 2 loading.The peak at 1590 cm −1 is attributed to the bending vibration of the O-H bond in the adsorbed water molecules on nSiO 2 surfaces.
The intensive peaks at about 1083 cm −1 , together with the less intensive peaks at 800 and 455 cm −1 , are attributed to the asymmetric stretching vibration, symmetric stretching vibration, and asymmetric bending vibration of Si-O-Si, respectively [24][25][26].
The presence of epoxide rings in ENR molecules was also confirmed by the 1 H-NMR analysis.The 1 H-NMR spectrum of the synthesized ENR sample revealed two peaks at 2.7 and 1.3 ppm (Figure 2), which represented the proton attached to the epoxide ring and the proton from a methyl group attached to the epoxide ring, respectively, while the peak at 5.14 ppm represented the olefinic proton of the unreacted NR.In addition, the peak at 3.7 ppm was attributed to some functional groups such as -OH and hydrofuran obtained from the opening of epoxide rings [8,9,14,24].This is in agreement with the result of the FT-IR analysis that was previously mentioned.
DSC analysis was also used as a quantitative measurement for the mol% epoxidation from the increased  g of the synthesized ENR compared with that of the pure NR [5,[12][13][14].It has been reported that, for the ENR with side ring-opening reaction, each mol% epoxidation gives a 1.15 ∘ C increase in  g [12].The increased  g of ENR was owing to the polar epoxide group that provided strong intermolecular interaction and so restricted the mobility of the ENR chains [5,16,24].The DSC thermograms of the unvulcanized samples based on the second heating scan performed between −80 and 20 ∘ C are shown on the left side of Figure 3 and the obtained  g values are summarized in Table 1.As can be seen,  g of the synthesized ENR (−26.2 ∘ C) was higher than that of the NR (−70 ∘ C) about 43.8 ∘ C.Although  g of ENR was much higher than that of NR, it was not significantly changed after adding different nSiO 2 loadings, suggesting that the presence of nSiO 2 had a small influence on  g of the nanocomposites.
The MDR rheographs of the samples conducted at 150 ∘ C are shown in Figure 4, and the main cure characteristics ( s2 , t 90 , CRI,  L ,  H , and  H −  L ) are all summarized in Table 2.The sulfur vulcanization with unsaturated rubbers generally occurred through a free radical reaction initiated by accelerator.The resulting sulfur radicals further attack reactive allylic hydrogens in rubber chains to form sulfur crosslinks, and so the vulcanization depends upon the number of the allylic hydrogens [28]. s2 is the time required for the torque value to increase two units from  L , which corresponds to the premature vulcanization of rubber compound, while t 90 is the vulcanization time for optimum cure torque [1,3].As can be seen, NR exhibited shorter  s2 and t 90 than ENR 40 after being cured under the same condition.This is because NR possessed a higher level of unsaturation and reactive sites than ENR 40, and so the crosslinks started and reached the optimum cure torque in a shorter time [1,18,28].Moreover, the high content of benzylthiuram substituents in the TBzTD accelerator may provide a steric hindrance to the limited crosslink sites of the ENR 40.However, the ENR 40 had higher CRI than the NR because the isolated C=C in ENR reacted more rapidly than the contiguous C=C in NR, and also owing to the activation of the adjacent C=C by the epoxide group in the ENR [9,10,13]. s2 and t 90 of the NR were also shorter than those of the ENR 40/nSiO 2 nanocomposites.This again attributed to the higher level of the reactive sites in the NR molecules.Compared with the neat ENR 40, the nanocomposite containing 1 phr nSiO 2 had lower  s2 and t 90 because of the self-crosslinks or additional crosslinks from the specific interaction between the epoxide groups in the ENR 40 and -OH groups on the nSiO 2 surface through hydrogen bonding [9,[22][23][24]29].However, the CRI of the ENR 40 nanocomposite with 1 phr nSiO 2 was lower than that of the neat ENR 40, indicating a slower rate of vulcanization.This is attributed to the adsorption of the basic accelerator on the -OH groups of the nSiO 2 surface and so lowered the vulcanization rate of the nanocomposites [9,25].Further increase in the nSiO 2 loading to 2 phr resulted in longer  s2 and t 90 , and a lower CRI because of the higher level of the adsorbed accelerator on the nSiO 2 surface, which subsequently reduced the accelerator efficiency.However, at 3 phr nSiO 2 ,  s2 and t 90 were again shorter and the CRI was again increased owing to the agglomeration of nSiO 2 particles in the nanocomposite that decreased the free Si-OH groups and so reduced the accelerator being adsorbed on the nSiO 2 surface.
L is commonly related to the viscosity of the uncured compound, while  H is regarded to the elastic modulus of the vulcanizate [3,9,10]. H −  L is a measure of the dynamic shear modulus, which indirectly indicates the extent of crosslinking (crosslink density) of the vulcanizate [9,10,18,25].The  L ,  H , and  H −  L values of the ENR 40 were found to be lower than those of the NR (Table 2), indicating the lower viscosity of the uncured ENR 40 and also the lower elastic modulus and crosslink density of the ENR 40 vulcanizate.The lower viscosity of the uncured compound may be owing to the possible opening of epoxide rings and the degradation of ENR 40 molecules either during epoxidation or during compounding at high temperature, while the lower elastic modulus and crosslink density of the ENR 40 vulcanizate may be owing to the lower number of the C=C in the ENR 40 compared to that in the NR [9].However, the  L values of ENR 40 and its nanocomposites were nearly the same, indicating that the nSiO 2 at a very low loading had no effect on the viscosity of the uncured compounds, while the  H and  H −  L values of the nanocomposites noticeably increased compared to those of the ENR 40 as a consequence of the strong interaction between ENR 40 and nSiO 2 as previously mentioned.

Characterization of Vulcanized Samples.
The mechanical properties of the samples both before and after aging at 100 ∘ C for 72 h in air are summarized in Table 2.It is seen that the tensile properties (tensile strength, M300, and elongation at break) and tear strength of the NR were all higher than those of the ENR 40.This is owing to the higher ability of NR to undergo strain crystallization as a result of its high molecular stereoregularity and also owing to the higher elasticity of NR that provided superior mechanical properties.Moreover, the epoxidation resulted in a reduced tendency to crystallize at high elongation of the synthesized ENR 40.Table 2 also shows a reduction in all the studied mechanical properties of NR and ENR 40 after aging as a result of the molecular chain scission and/or crosslink breakage at high temperature and oxygen content [3,4].It is seen that the percentage retentions of tensile strength, M300, elongation at break, and tear strength of the NR were all lower than those of the ENR 40.This is because of a higher susceptibility to thermal oxidative degradation of NR, which was mainly owing to the high level of the C=C in the main chain that gave rise to the lower retention values [4].Therefore, changes in the mechanical properties of rubber products depended on the extent of such reactions that limited their useful lifetime at elevated temperature in the presence of atmospheric oxygen.However, the high rubber-rubber interaction could improve the aging resistance of the ENR 40.
The effects of nSiO 2 on the mechanical properties of ENR 40 nanocomposites are also shown in Table 2 and Figure 5.As can be seen, the tensile properties and tear strength of the ENR 40 nanocomposites with three loadings of nSiO 2 (1, 2 and 3 phr) were all higher than those of the neat ENR 40, owing to the reinforcing effect of nSiO 2 , but at 2 phr nSiO 2 was the highest for all the mechanical properties.This may be attributed to a better dispersion of nSiO 2 in the ENR 40 matrix, providing a greater filler-rubber interaction between the silanol and epoxide groups.This is an essential role for achieving the good mechanical properties.However, the decrease in the mechanical properties of the nanocomposite at 3 phr nSiO 2 may be owing to the agglomeration of an excess nSiO 2 , leading to a poor dispersion of nSiO 2 in the ENR 40 matrix, a reduction in contact area between nSiO 2 and ENR 40, and also a filler-rubber interaction.However, the percentage retention in mechanical properties of all the ENR 40 nanocomposites was higher than that of the neat ENR 40, owing to a very high thermal resistance of the nSiO 2 [25,26].
The DSC curves of the vulcanized NR, ENR 40, and ENR 40/nSiO 2 nanocomposites are shown on the right side of Figure 3, and the obtained  g values are also presented in Table 1.The  g values of all the vulcanizates were higher than those of the corresponding unvulcanized samples according to the formation of crosslinks in the rubber matrix that restricted the flexibility of the rubber chains; hence  g increased.In addition, all the  g values of the vulcanized samples followed the same trend as those of the unvulcanized samples (Table 1).Among the nanocomposites, the ENR 40 filled with 2 phr nSiO 2 exhibited the lowest  g owing to a better dispersion of nSiO 2 that caused a decrease of the ENR 40 being trapped within the nSiO 2 network and so resulted in a higher mobility of the rubber chain segments; hence  g decreased [30,31].
Figure 6 shows the storage modulus (  ), loss modulus (  ), and loss tangent (tan ) curves obtained from the DMA analysis of the samples, and their   (at −65 ∘ C),  g , and tan  max (height of tan  peak) values are also listed in Table 1.According to Figure 6(a), the ENR 40 and ENR 40/nSiO 2 nanocomposites had much higher   values than the NR.The higher   value of ENR 40 is again owing to the strong intermolecular force between epoxide groups that restricted the mobility of the ENR 40 chain segments.Additionally, all the ENR 40/nSiO 2 nanocomposites exhibited higher   values than the neat ENR 40.This may be owing to the good interaction between the nSiO 2 and ENR 40 and also to the high stiffness of nSiO 2 that restricted the mobility of the rubber chains as previously discussed.According to Figures 6(b) and 6(c), the  g values derived from the temperatures at the   and tan  peaks followed the same behavior as that obtained from the DSC curves (Figure 3).It is seen that the  g values of the ENR 40 and ENR 40/nSiO 2 nanocomposites were also much higher than those of the NR according to the same reason as previously discussed.However, the tan  max values of the ENR 40 and ENR 40/nSiO 2 nanocomposites were all lower than those of the NR.This indicated that the ENR 40 and ENR 40/nSiO 2 nanocomposites were stiffer than the NR as a consequence of the strong intermolecular attraction or rubber-filler interaction as previously mentioned.
The thermal stability of the samples was evaluated by TGA, and their representative TG and derivative thermogravimetry (DTG) curves are shown in Figure 7.The derived  onset , T 50% ,  end set ,  max , and % char values are also listed in Table 1.All the samples had similar TG and DTG curves and each one presented a two-step degradation.The degradation of the first step for all samples observed at 226-252 ∘ C ( onset 1 ), 287-296 ∘ C ( end set 1 ), and 268.5-284.8∘ C ( max 1 ) was attributed to the low molecular weight compounds such as nonrubber components and impurities as well as additives [15,16] and has a weight loss of 6.7-8.1%,whereas the second step observed at 362.7-379.7 ∘ C ( onset 2 ), 419.4-435 ∘ C ( end set 2 ), and 385-401 ∘ C ( max 2 ) was attributed to the main degradation of rubber hydrocarbon, corresponding to the weight loss of 76.2-87.1%.Additionally, T 50% for all samples was in the range of 388.4-410.6 ∘ C. As can be seen,  onset 2 , T 50% ,  end set 2 , and  max 2 of the ENR 40 were all higher than those of the NR, indicating that the thermal International Journal of Polymer Science stability of ENR 40 was improved.This is owing to the strong intermolecular attraction in ENR 40 caused by the insertion of oxygen atoms in the NR molecules and also to the high susceptibility to degradation of NR as previously mentioned.Moreover, the degradation temperatures from Table 3 also revealed that the addition of nSiO 2 to the ENR 40 exhibited a dose-dependent increase in the thermal stability of the nanocomposites by shifting  onset 2 , T 50% ,  end set 2 , and  max 2 towards higher temperatures compared to those of the neat ENR 40.This is owing to the high thermal stability of nSiO 2 accompanied with a strong filler-rubber interaction through the hydrogen bonding that caused the degradation to occur at a higher temperature and that also hindered the mass transport during the thermal degradation process [22][23][24][25].It was also found that the char formation of the nanocomposites (13.4-16.6 wt%) remarkably increased with increasing nSiO 2 loading compared to that of the neat NR and ENR 40 (both at 6 wt%) as shown in Figure 7 and Table 1.This is because nSiO 2 is an inorganic nanofiller that can withstand high temperature and remain as residue even at 1,000 ∘ C and so delays and reduces the thermal degradation of the vulcanizates by restricting the permeability of volatile decomposition products out of the samples [4,25].
The percentage volume change of the NR, ENR 40, and ENR 40/nSiO 2 nanocomposites after being immersed in ASTM oil number 1 (aniline point ∼124 ∘ C) and ASTM oil number 3 (aniline point ∼90 ∘ C) at 100 ∘ C for 72 h is presented in Table 4.It is seen that the percentage volume  change of NR in both oils was much higher than that of the ENR 40 and ENR 40/nSiO 2 nanocomposites, indicating that the epoxidation significantly improved their oil resistance.Moreover, the ENR 40/nSiO 2 nanocomposites exhibited a nSiO 2 load-dependent higher oil resistance than the neat ENR 40 according to the interaction between the epoxide groups of the ENR 40 and the Si-OH groups on the nSiO 2 and also to their higher crosslink density that limited the penetration of oils into the rubber matrix.However, the volume change of the samples in ASTM oil number 3 was much higher than that in ASTM oil number 1.This is owing to the fact that the lower the aniline point the more severe the swelling action.

Conclusions
In this work, the ENR 40 was successfully synthesized via in situ performic epoxidation of NR in the latex state using a fixed molar ratio of HCOOH/H 2 O 2 to isoprene unit of 1/0.75 in the presence of a surfactant at 50 ∘ C for 8 h.Beside the epoxide groups, the ring-opened structures were also observed in the as-synthesized ENR 40.The resulting ENR 40 was melt mixed with three loadings of nSiO 2 (1, 2, and 3 phr) and then vulcanized as usual in a sulfur curing system using TBzTD as a noncarcinogenic accelerator, which resulted into nitrosamine free cure.The curing characteristics, mechanical properties, thermal behaviors, dynamic mechanical ) was slightly higher.The mechanical properties (tensile properties and tear strength) of ENR 40 for both before and after aging at 100 ∘ C for 72 h were found to be improved as a result of the incorporation of nSiO 2 (all, but 2 phr optimal), while the % retention in all mechanical properties of ENR 40/nSiO 2 nanocomposites was higher than that of NR and ENR 40 in a dose-dependent manner.The  g ,   at −65 ∘ C, thermal stability, and oil resistance (in ASTM Oil number 1 and number 3) of ENR 40 were also enhanced with the addition of nSiO 2 (all, but 1 phr optimal).
UK) and H 2 O 2 (35%) manufactured by Chem-Supply Pty Co. (Australia) were purchased from Myscience Co. (Thailand) and RCI Lab Scan Co. (Thailand), respectively.Nano-SiO 2 having particle size of about 12 nm and specific surface area of 175-225 m 2 g −1 manufactured by Sigma Aldrich (Singapore) was purchased from A.C.S. Xenon Co. (Thailand).Sodium carbonate and methanol were purchased from Merck Thailand Co. (Thailand) and RCI Lab Scan Co. (Thailand), respectively.ASTM oil number 1 (IRM 901) and ASTM oil number 3 (IRM 903) were purchased from Chemical Innovation Co. (Thailand).All materials were used as received without further purification.

Table 4 :
Oil resistance of NR, ENR 40, and ENR 40/nSiO 2 nanocomposites., and oil resistance of the prepared ENR 40/nSiO 2 nanocomposites were investigated and compared with NR and neat ENR 40.The results revealed that  s2 and t 90 of ENR 40/nSiO 2 nanocomposites were slightly longer than those of NR but slightly shorter than those of ENR 40.However, the CRI of all ENR 40/nSiO 2 nanocomposites was lower than that of NR and ENR 40, while the crosslink density ( H −  L properties