Yuta Sakakia,
Ryota Usamia,
Atitaya Tohsanb,
Preeyanuch Junkongcd and
Yuko Ikeda*ce
aGraduate School of Science and Technology, Kyoto Institute of Technology, Matsugasaki, Sakyo, Kyoto 606-8585, Japan
bDepartment of Materials and Production Technology Engineering, Faculty of Engineering, King Mongkut's University of Technology North Bangkok, Wongsawang, Bangsue, Bangkok 10800, Thailand
cCenter for Rubber Science and Technology, Kyoto Institute of Technology, Matsugasaki, Sakyo, Kyoto 606-8585, Japan
dResearch Strategy Promotion Center, Kyoto Institute of Technology, Matsugasaki, Sakyo, Kyoto 606-8585, Japan
eFaculty of Molecular Chemistry and Engineering, Kyoto Institute of Technology, Matsugasaki, Sakyo, Kyoto 606-8585, Japan. E-mail: yuko@kit.ac.jp
First published on 19th March 2018
A linear combination fitting in sulfur K-edge X-ray absorption near edge structure (S-XANES) measurements reveals each fraction of monosulfidic, disulfidic and polysulfidic linkages in solvent extracted sulfur cross-linked isoprene rubbers. The sulfidic linkage of a disulfidic type is found for the first time to be dominant when zinc stearate and N-(1,3-benzothiazol-2-ylsulfanyl)cyclohexanamine are used as the activator and accelerator, respectively, for the sulfur cross-linking reaction at 140 °C. The presence of the bridging bidentate zinc/stearate complex as an intermediate for the sulfur cross-linking reaction is supposed to induce the generation of abundant disulfidic linkages in the rubber networks. This unexpected observation is of use for the material design of high performance rubber products with anti-aging and thermal stabilities. S-XANES is a powerful tool that was used to reveal the characteristics of the sulfur cross-linking of rubber. These results will contribute to the development of rubber science and technology.
Regarding the current progress in analytical techniques, it has been accepted that vulcanization leads to the inhomogeneous network structure. Previously, our group reported the effects of the combination and composition of the vulcanizing reagents on the formation of the rubber network structure.9 The results obtained from the small-angle neutron scattering technique indicated that zinc oxide (ZnO) and other reagents are crucial in the control of the two-phase inhomogeneous network structure of vulcanized isoprene rubber, as illustrated in Fig. 1,9 where the network domains of high network-chain density are speculated to be embedded in the mesh network.
![]() | ||
Fig. 1 Proposed two-phase inhomogeneous network structure in sulfur cross-linked isoprene rubber. Reproduced with permission.9 Copyright 2009, American Chemical Society. |
The two-phase network structure was proposed to be initially formed by zinc stearate (ZnSt2) generated by a reaction between ZnO and stearic acid. The ZnSt2 was proposed to further react with the accelerator and sulfur to result in sulfur cross-links and was therefore supposed to be a key factor in controlling the size of the mesh network in the matrix.9 In contrast, sulfur and accelerator were also considered to be adsorbed on the remaining ZnO in the rubber matrix, followed by the sulfur cross-linking reaction around the ZnO particles. Therefore, the amounts of accelerator and sulfur adsorbed on the ZnO were also proposed to control the size of the network domain.9
The formation of a two-phase inhomogeneous network structure was confirmed by employing time-resolved zinc K-edge X-ray absorption fine structure (Zn-XAFS) spectroscopy and differential scanning calorimetry.10 It was detected that the formation of the mesh network occurred a little earlier than that of the network domains and its fraction was unexpectedly observed to be larger than that calculated on the basis of the concentration of ZnSt2. The result was ascribed to the amount of zinc/stearate complex intermediates forming the mesh network.10 These previous studies clearly emphasized that the compositions of vulcanizing reagents play an important role in controlling the two-phase network structure and hence the mechanical properties of vulcanized rubber.
On the other hand, it is also well-known that the compositions of vulcanizing reagents change the sequence of sulfidic linkages in the cross-linking sites in a vulcanization reaction. Particularly, it has been recognized that sulfidic linkages in vulcanizates drastically affect the mechanical properties and thermal resistance of rubber products.5,11–13 For example, a polysulfidic linkage is more easily decomposed by heat and stress than monosulfidic and disulfidic linkages, due to the lower bonding energy of the polysulfidic segment. Generally, sulfur cross-linked rubber has been believed to be composed of a mixture of monosulfidic, disulfidic and polysulfidic cross-linkages with C–C cross-linkings.11,13,14 However, the effect of curing reagents on the sulfidic linkages has not yet been satisfactorily clarified, due to the difficulty of analysis of the linkages in the vulcanizates. So far, a chemical method for the cleavage of sulfidic linkages by lithium aluminium hydride has been utilized for investigating the sulfidic structure in vulcanized rubbers.15 This method is known to not distinguish between cross-links and other sulfurated network features such as pendant accelerator residues and cyclic structures.16 Chemical analysis using reagents of thiols and amines has been used to selectively cleave disulfidic and polysulfidic cross-linked segments.16 However, the handling of these reagents is problematic due to their strong smell and dangerousness. For both chemical methods, the additional effect of the trapped entanglement has not been taken into account for the determination of each concentration of sulfidic linkages. Alternatively, sulfur K-edge X-ray absorption near edge structure (S-XANES) spectroscopy is a tool that can be used to evaluate the sulfidic linkages in vulcanizates, which allows insight into the stereochemistry and electronic state of the absorbing atom.17–22 In addition, this technique gives good structural information as a fingerprint, together with being non-destructive to the sample and it can be used with various types of samples. Synchrotron-based S-XANES measurement under high vacuum is broadly employed to investigate the sulfidic linkage for rubber materials,23–30 where the energy states at the absorption edges are determined.31 Brendebach et al. observed a variation in the cleavage of cyclic sulfur during the vulcanization of carbon black filled conventional rubber.21 Pattanasiriwisawa et al. clarified that sulfidic linkages were controlled by the ratio between the contents of sulfur and accelerator in the conventional vulcanization, effective vulcanization (EV), and semi-EV systems of natural rubber.25 Taweepreda et al. researched the effect of accelerators such as thiuram, thiazole, and dithiocarbamic acid on the sulfidic linkages in the vulcanized squalene.27 However, there has been a request for this S-XANES method to be improved for the quantitative analyses of sulfidic linkages in the vulcanizates.
In this paper, a variation of sulfidic linkages in the solvent-extracted isoprene rubber vulcanizates is intensively studied using S-XANES in order to quantitatively investigate the type of sulfidic linkage in the mesh network in Fig. 1. As a model network for the mesh, a benzothiazolesulfenamide-accelerated curing system with ZnSt2 was utilized for the investigation. An unexpected result is reported in this article, which indicates the unique characteristics of the vulcanizate. The findings are useful for producing important rubber materials with anti-aging and thermal stabilities, as predicted in the ACS News Service Weekly PressPac in 2015.32
Sample code | Pressing time (min) | Network-chain densitya × 104 (mol cm−3) | Degree of swelling by volumeb | Gel fractionc |
---|---|---|---|---|
a Determined by the swelling measurement in toluene at 25 °C.b Determined by the change of volume in toluene before and after swelling.c Determined by the change of weight in toluene before and after swelling followed by drying. | ||||
IR–ZnSt2–CBS–S8-18 | 18 | 0.66 | 7.0 | 0.92 |
IR–ZnSt2–CBS–S8-19 | 19 | 0.69 | 6.9 | 0.97 |
IR–ZnSt2–CBS–S8-22 | 22 | 1.28 | 5.2 | 0.95 |
IR–ZnSt2–CBS–S8-26 | 26 | 1.16 | 5.5 | 0.96 |
IR–ZnSt2–CBS–S8-30 | 30 | 0.84 | 6.3 | 0.94 |
IR–ZnSt2–CBS–S8-40 | 40 | 0.71 | 6.8 | 0.91 |
The XANES spectra were linear-background subtracted and normalized before analysis by the ATHENA (version 0.8.056) software package.35 The quantitative sulfidic linkage analysis, i.e., mono-, di- and polysulfidic linkages, were evaluated by a linear combination fitting method in the energy range of 2461 to 2491 eV. The R-factor is defined as a sum-of-squares measure of fractional misfit showing a degree of best fit; the smaller the R-factor, the better the fitting accuracy.
For S-XANES measurements, the IR vulcanizates of about 10 mm in width × about 20 mm in length × about 0.2 mm in thickness were subjected to the solvent extractions in advance using 30 mL of chloroform/acetone mixture (7/3 vol%, special grade solvents of Wako Pure Chemical Industries, Ltd.), followed by 30 mL of special grade tetrahydrofuran (THF, Wako Pure Chemical Industries, Ltd.). Both extraction steps were performed twice at 25 °C for 24 hours. After the extractions, the sample sheets were rinsed several times using the same solvent for each step. The extracted samples were then dried under air and vacuum for 1 day to remove the solvent. The amounts of extracts (wt%) are summarized in Table 2. No heating was applied for the extraction in order to prevent post-vulcanization at high temperature. These extraction conditions were adequate for removing unreacted S8 and CBS because the solubilities of the reagents at 25 °C were confirmed before the extractions, as follows:30 for S8 and CBS in 100 g of the chloroform/acetone mixture (7/3 vol%) the solubilities were 0.6 g and 5.0 g, respectively; for S8 and CBS in 100 g of THF the solubilities were 2.0 g and 18.2 g, respectively. In the fourth extraction, the extracts of each vulcanizate were subjected to mass spectroscopy with EI mode at 220 °C, where THF was used as the solvent. As shown in Table 2, the amounts of extracts from each sample decreased by repeating the extractions, and trace amounts were observed in each sample after the fourth extraction by THF. Therefore, all unreacted CBS and S8 were judged to be removed almost completely.
Sample code | 1st extractiona | 2nd extractiona | 3rd extractionb | 4th extractionb | ||
---|---|---|---|---|---|---|
Q (wt%) | Q (wt%) | Q (wt%) | Q (wt%) | Mass spectroscopy | ||
CBSc | S8 | |||||
a Solvent extraction using chloroform/acetone mixture (7/3 vol%).b Solvent extraction using THF.c N-(1,3-Benzothiazol-2-ylsulfanyl) cyclohexanamine.d ×: none, T: trace. | ||||||
IR–ZnSt2–CBS–S8-18 | 7.7 | 2.8 | 0.6 | 0.9 | Td | T |
IR–ZnSt2–CBS–S8-19 | 5.2 | 1.5 | 0.5 | 0.6 | ×d | T |
IR–ZnSt2–CBS–S8-22 | 4.5 | 1.3 | 0.1 | 0.4 | × | T |
IR–ZnSt2–CBS–S8-26 | 4.9 | 1.4 | 0.0 | 0.2 | T | T |
IR–ZnSt2–CBS–S8-30 | 5.9 | 1.8 | 0.4 | 0.7 | T | T |
IR–ZnSt2–CBS–S8-40 | 6.5 | 2.2 | 0.3 | 0.5 | T | T |
Fig. 3 shows the XANES spectra of all reference samples plotted for the energy range of 2465–2490 eV with normalized intensity. Their characteristic absorption peaks correspond to their energies at absorption edges and shapes, indicating each local environment of the sulfur atom. The energies at the tops of peaks can be used to assign the sulfidic linkages in the IR vulcanizates. The absorption peaks of the IR–ODS, IR–DBS and IR–S8 were observed at 2473.02, 2472.42 and 2472.24 eV, indicating the mono-, di- and polysulfidic linkages, respectively. The characteristic peaks and shapes of the reference compounds were in good agreement with previous reports.19,31,36 In practice, ZnS is formed during the progress of the vulcanization reaction together with the thermal degradation and oxidation reaction, i.e., the generation of –SO− and –SO42− species; therefore, in this study, ZnS, DBSO and ZnSO4 were selected to assign the characteristic peaks of Zn–S, –SO− and –SO42−, respectively. From Fig. 3, the energies at the tops of the peaks of these sulfur reference species were observed at 2473.58, 2475.70 and 2481.99 eV, respectively.
![]() | ||
Fig. 3 S-XANES spectra of the reference samples; (a) white lines for all reference samples and (b) their magnified region. |
A relationship between the energy of the white line and the heat-pressing time of each sample is shown in Fig. 4(c), in which the dashed and dotted lines represent the energies at the white lines of two reference samples, IR–DBS and IR–S8, suggesting disulfidic and polysulfidic linkages at 2472.42 and 2472.24 eV, respectively. The torque obtained from the cure curve is also plotted as a secondary y-axis with the heat-pressing time. The IR–ZnSt2–CBS–S8 samples prepared by heat pressing for 18, 19, 22, 26 and 30 min show the white lines at about 2472.42 eV, at a similar level to that of the disulfidic linkage. However, the white line of IR–ZnSt2–CBS–S8 prepared by heat pressing for 40 min was substantially higher than 2472.42 eV. These results indicate that disulfidic linkages may be dominantly formed in the mesh network during both the vulcanization and reversion processes. Additionally, the cleavage of polysulfidic linkages to di- and/or monosulfidic linkages is proposed to be significantly accelerated under a long heating time up to 40 min. However, it must be emphasized here that the white lines are represented as a summation of all linkages. Therefore, it is difficult to directly discuss the variation in the length of sulfidic linkages during vulcanization from only the white lines.
On the other hand, the vulcanizate of IR–ZnSt2–CBS–S8-40 showed the presence of di- and monosulfidic linkages without any polysulfidic linkage. As is well understood, the reversion process is generally ascribed to the cleavage of sulfur segments of the cross-linked sites in the network by thermal aging. It is also well-known that the cleavage of the network structure occurs most easily in polysulfidic cross-linking segments, followed by the cleavage of disulfidic linkages. In this study, Fig. 6 clearly and quantitatively supports the decrease in the fraction of polysulfidic linkages with the time of heat pressing and the increase in the fraction of disulfidic linkages from 22 to 30 min. When the cleavage of polysulfidic linkages was mostly finished at about 30 min, there was a decrease in disulfidic linkages, as shown in Fig. 6. Since the thermal stability of the disulfidic linkage is a little weaker than that of the monosulfidic linkage,37 the decrease in disulfidic linkages and increase in monosulfidic linkages were reasonable for heat pressing at 140 °C. The variation in network-chain densities of IR–ZnSt2–CBS–S8 samples supported this consideration. As shown in Table 1, the network-chain densities clearly increased from 18 to 22 min, but decreased after 22 min, where the maximum torque of the cure curve was detected.
In general, zinc sulfide has been recognized as one of the by-products in the sulfur cross-linking reaction.38 In this study, our results support the common opinion that the fraction of zinc sulfide increased with the increase in heat-pressing time. Thus, both the sulfur cross-linking reaction and the decross-linking reaction are concluded to occur even in the reversion process. It is also noted in Fig. 4(a) that the absorption intensities at about 2475 and 2482 eV increased with time, which may be ascribed to the generation of sulfoxide and/or sulfate groups by oxidation.18,21 By fitting, the oxidative products composed of –SO− and –SO42− groups were evidently detected in the samples as trace amounts. Since our samples do not contain any antioxidant reagents, the thin films may have been a little oxidized during the handling and storage of the samples. From these results, it is obvious that the linear combination fitting using the appropriate reference spectra gives us reasonable and quantitative results to monitor the changes in sulfidic linkages as a function of heat-pressing time.
![]() | ||
Fig. 8 Dinuclear type bridging bidentate zinc/stearate complex composed of (Zn2(μ-O2CC17H35)2)2+(OH−)2·XY, where X and Y are water and/or a rubber segment. Reproduced with permission.39 Copyright 2015, American Chemical Society. |
As shown in Fig. 6, the fraction of disulfidic linkages of IR–ZnSt2–CBS–S8-18 was about 59%, although the torque in the cure curve was about 0.17 N m, which was about 43% of the maximum torque of the sample prepared at 22 min in the cure curve shown in Fig. 2. This clearly suggests that the disulfidic linkage was already formed in the first half of the sulfur cross-linking reaction in IR–ZnSt2–CBS–S8. Therefore, the results in this S-XANES study suggest that the intermediate of the bridging bidentate zinc/stearate complex may induce the formation of disulfidic linkages during the vulcanization reaction. As is well-known, the disulfidic linkages are more stable against heat and oxidation compared to the polysulfidic linkages. Therefore, the observed results for the disulfidic linkage in the mesh network formation of this study must be a strategic key for producing high performance rubber products. The other experimental evidence to support this proposal will be reported in the near future.
It was found that IR–ZnSt2–CBS–S8-40 showed higher tensile stresses than IR–ZnSt2–CBS–S8-19, even though their fractions of disulfidic linkages were comparable as shown in Fig. 6. This difference is ascribable to the fractions of polysulfidic and monosulfidic linkages: the fraction of polysulfidic linkages of IR–ZnSt2–CBS–S8-40 was zero and monosulfidic linkages were much higher than that of IR–ZnSt2–CBS–S8-19. The results suggest that the monosulfidic linkage (C–S–C) is much stronger than the polysulfidic linkage (C–Sx–C), due to the higher bonding energy of the C–S linkage. The vulcanizate mainly cross-linked by monosulfidic and disulfidic linkages can load the higher stresses, but give the lower flexibility, compared to the vulcanizate composed of the polysulfidic linkages. This phenomenon was clearly shown in this study by using the S-XANES analyses for the vulcanizates. Due to the absence of ZnO in IR–ZnSt2–CBS–S8 in this study, only the mesh network must have been formed without any network domains that could act as a reinforcing filler. Therefore, the effect of different kinds of sulfidic linkages was apparently detected in the tensile properties of vulcanizates.
This journal is © The Royal Society of Chemistry 2018 |