PERFORMANCE COMPARISON OF VARIOUS SURFACE MODIFYING AGENTS ON PROPERTIES OF SILICA-FILLED CHLOROPRENE RUBBER
ABSTRACT
Influence of the surface modifying agents (SMAs) polyethylene glycol (PEG), 3-aminopropyl triethoxysilane (APTES), and bis-(3-triethoxysilylpropyl)tetrasulfide (TESPT) on properties of silica-filled chloroprene rubber (CR) was investigated. Results reveal that the presence of SMAs greatly improves mechanical and dynamic properties of the silica-filled CR because of the reduced filler–filler interaction and improved rubber–filler interaction as evidenced by the Payne effect and bound rubber content results, respectively. When compared at the same SMA dosage, TESPT gives the best overall vulcanizate properties. This effect is attributed to high coupling efficiency of TESPT and its ability to donate sulfur atoms during vulcanization, leading to the improved filler dispersion, rubber–filler interaction, and cross-link density. APTES gives high coupling efficiency, but because of the lack of sulfur atoms, its performance is slightly inferior to that of TESPT. In contrast, PEG can only reduce filler–filler interaction, with no significant impact on rubber–filler interaction and cross-link density; therefore, PEG provides lower property improvement.
INTRODUCTION
Chloroprene rubber (CR) is widely used in many applications in the automotive and construction industries,1 as adhesives, and in liner pads on missile launchers.2 The cross-linking reaction of CR is different from that of other diene rubbers because of the presence of electronegative chlorine atoms that inhibits the electrophilic substitution reaction ordinarily found in the cross-linking mechanism of other unsaturated polymers.3 A typical cure system to produce cross-linked CR contains zinc oxide/magnesium oxide, a curing accelerator, stearic acid, and additive.3 The use of CR in the engineering products requires the addition of reinforcing filler, particularly carbon black (CB) or precipitated silica (PSi). Because of their special surface characteristics, both fillers can form agglomerates, but the basis of such agglomeration is not similar, leading to a difference in dispersion ability. The filler–filler interaction of CB is mainly generated through relatively weak van der Waals forces that can be readily broken during mixing. In contrast, mixing of PSi is more problematic because of a large number of highly polar silanol groups on its surface; thus, agglomeration is caused by hydrogen bonding in addition to van der Waals forces and other physical interactions, leading to a much stronger filler–filler interaction.4 However, PSi is classified as an environmentally friendly filler that could be used as a substitution for CB in many applications and, unlike CB, the characteristics of PSi are independent of oil resources.5–7 Despite its dispersion difficulty, it has been reported that the presence of silica in CR could also give rise to additional cross-links via the chemical reaction between the silanol groups on the silica surface and the allylic chlorine atoms in CR.8,9 These reports confirm that silica interacts with CR in a manner that is different from the of CB, the most commonly used filler. Furthermore, PSi properly treated with coupling agents is capable of offering superior dynamic mechanical properties of rubber composites to CB, that is, lower rolling resistance and greater wet skid.10,11 The use of PSi has therefore gained much attention from researchers, particularly as it relates to tire tread applications.12–14 Although PSi is more compatible with polar rubbers such as CR9 and NBR,15 the use of PSi without surface modifying agents (SMAs) still results in poor dispersion and consequently poor mechanical properties.16 In addition, alkaline accelerators could be adsorbed on the acidic surface of PSi, thereby negatively affecting cure properties.4,17 To overcome such problems, the modification of PSi surfaces is inevitable, and such modifications could typically be achieved by the incorporation of glycol, coupling agents, or both. Polyethylene glycol [PEG; H-(OC2H4)n-OH] has long been used as an SMA for PSi because of its hydrophilicity and low cost. Silane coupling agents (SCAs), with the general formula YSi(X)3, where X and Y stand for a hydrolysable group and a non-hydrolyzable organo-functional group capable of interacting with PSi and rubber, respectively, generally provide greater benefits over the glycol compounds.18,19 The surface treatment of PSi with SCAs via a silanization reaction is generally carried out through hydrolysis and condensation reactions. There are two methods for such treatment: pretreatment and in situ. In the pretreatment method PSi is pretreated before mixing and the in situ method the silanization reaction between PSi and SCA takes place during mixing.
Although many findings have revealed superior coupling efficiency of bis-(3-triethoxysilylpropyl) tetrasulfide (TESPT) to other SMAs, most of these findings deal with the sulfur-curable diene rubbers. In contrast, little attention has been given to the comparison of coupling efficiency of TESPT and other SMAs in metal-curable CR. In our previous study,20 PSi was pretreated with various SCAs before being mixed with CR. We found that 3-aminopropyltriethoxy silane (APTES) and TESPT are capable of reducing the filler–filler interaction and enhancing the rubber–filler interaction to a greater extent than 3-chloropropylthriethoxy silane leading to superiorities in processability and mechanical properties of the silica-filled CR.
This work aims to compare and investigate the effect of the SMAs PEG, APTES, and TESPT on the mechanical and dynamic properties of PSi-filled CR in which the in situ technique was used for surface treatment of PSi.
EXPERIMENTAL
Materials
CR (grade W) was purchased from DuPont Dow Elastomer Co., Ltd. (Wilmington, DE, USA). PSi (Tokusil 233) was manufactured by Tokuyama Siam Silica Co., Ltd. (Rayong, Thailand). The SCAs APTES and TESPT were supplied by Evonik Co., Ltd. (Essen, Germany). PEG, with average molecular weight of 4000 Da, was purchased from Dow Chemical Co. (Essen, Germany). MgO, N-(1,3-dimethylbutyl)-N′-phenyl-p-phenylenediamine (6-PPD), ZnO, ethylene thiourea (ETU), stearic acid, and sulfur were purchased from local suppliers. All chemicals were used as received without further purification.
Preparation and testing of rubber
Rubber compounds were prepared using a laboratory-scale Rheomix 90 internal mixer (Haake, Karlsruhe, Germany). The mixing conditions were as follows: fill factor = 0.8, chamber temperature = 50 °C, rotor speed = 40 rpm, and mixing time = 20 min. The compounding recipe is given in Table 1. A two-stage mixing technique was used for preparing the compounds. In the first mixing stage, MgO, PSi, SCA, or PEG; stearic acid; and 6-PPD were incorporated sequentially into CR matrix, with a total mixing time of 20 min. In the second mixing stage, ZnO, ETU, and sulfur were charged into the rubber mix and then further mixed for 7 min to achieve good dispersion and distribution of all ingredients. After mixing, the compounds were sheeted on a two-roll mill and kept at room temperature at least 16 h before testing. The extent of filler–filler interaction (filler transient network) was evaluated using a RPA2000 rubber process analyzer (Alpha Technologies, Akron, OH, USA). Storage modulus (G′) of the rubber compounds was measured at swept strain from 0.56 to 100% at 100 °C and 1.7 Hz. The difference in G′ at low (0.56%) and high (100%) strains (ΔG′), widely known as the Payne effect, was used to represent the degree of filler–filler interaction.21,22 For a measurement of bound rubber content (BRC), small pieces of uncured specimens (∼1 g) were extracted with 150 mL of toluene for 7 d at room temperature. The insoluble portion was then filtered and dried in an oven at 80 °C for 24 h. The percentage of BRC was determined using Eq. 1:
where Wfg is the weight of filler–rubber gel; W is the weight of the test specimen; and mf and mp are the weights of filler and polymer in the rubber compound, respectively.
Measurement of BRC was also carried out under ammonia atmosphere to assess the extent of chemical rubber–filler interaction as described previously.23,24
Cure behavior was monitored with a RheoTECH MD+ moving die rheometer (MDR; TechPro, Cuyahoga Falls, OH, USA) at 155 °C. To prepare cured specimens, the uncured compounds were compression molded using a hydraulic hot press under a pressure of 16 MPa at 155 °C for the cure time (tc99) as predetermined from the MDR. Hardness of the vulcanizates was measured using a durometer with Shore A scale (Cogenix Wallace, Burlington, UK) as per ISO 7619-1. Tensile properties were determined using an Instron 5566 universal testing machine (Norwood, MA, USA) according to ISO 37 (die Type 1). Compression set test was conducted per ISO 815-1 at 100 °C for 22 h. Heat build-up (HBU) test was performed using a model II flexometer (BF Goodrich Instruments, Carolina, NC, USA) according to ISO 4666-3 at 30 Hz under static force of 245 N with a full stroke of 4.45 mm for 25 min. The value of tan δ at 25 °C was measured using a dynamic EPLEXOR 25 N mechanical thermal analyzer (Gabo, Ahlden, Germany) under the static strain, dynamic strain, and frequency of 1%, 0.1%, and 10 Hz, respectively. Cross-link density of the cured specimens was determined by a swelling technique in toluene at room temperature for 7 d. Then, the swollen specimen was removed from the toluene, and the excess toluene was blotted using a towel paper. Thereafter, the specimen was weighed accurately with the weighing bottle. Cross-link density is inversely proportional to the swelling ratio (Q), calculated using Eq. 2:
where W1 and W2 represent the weights of unswollen and swollen test specimens, respectively.
The degree of filler dispersion was examined using scanning electron microscopy (SEM) with a JSM 6400 microscope (JEOL, Tokyo, Japan). The fractured surface was prepared by a cryogenic fracturing technique. The test specimen was later placed on a stub, and its surface was sputter coated with gold to prevent electron bombardment. To determine the aggregate size of PSi, the SEM images were analyzed using an Image-Pro® Express version 6 (MediaCybernetics, Inc., Bethesda, MD, USA).
RESULTS AND DISCUSSION
Generally, the magnitude of discrepancy in G′ at low (0.56%) and high (100%) strains or ΔG′, widely known as the Payne effect, is used as a measure of filler transient network magnitude in rubber compounds.22 As evidenced in Figure 1, the systems treated with SMAs show significantly lower magnitude of filler network than the untreated system. This response is attributed to the developed interaction between silanol groups on the PSi surfaces and hydroxy groups of PEG or alkoxy groups of SCAs, resulting in a reduction of hydrophilicity and H-bond among PSi aggregates. The plasticization effect cause by the presence of liquid SMAs might also facilitate filler incorporation during the mixing process, leading to the longer time available for filler dispersion (at a given total mixing time) and thus the improved state of mix. TESPT and APTES gave similar magnitudes of the Payne effect and were slightly lower than that offered by PEG.
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
Figure 2 shows BRC results of the PSi-filled CR compounds treated with various SMAs. When measured under normal conditions, the BRC values of all compounds are relatively high, ranging from 37 to 46%. The high BRC in the untreated system could be explained by the high extent of dipole–dipole interaction between the polar groups on the PSi surfaces and Cl atoms on the CR backbone.20 Obviously, the addition of PEG gives no significant effect on BRC. The BRC, however, slightly increases in the presence of SCAs. Such increase is thought to arise from the enhanced rubber–filler interaction by SCAs as shown in the proposed mechanisms in Scheme 1 and Scheme 2 for APTES and TESPT, respectively. In this aspect, TESPT gives a slightly higher BRC than APTES. However, by measuring BRC under ammonia conditions, a significant drop of BRC is observed in all systems, suggesting the ability of ammonia to weaken the rubber–filler interaction in silica-filled CR. It is widely known that ammonia could cleave physical linkages between silica and rubber because of the stronger silica–ammonia interaction than silica–rubber interaction.23,24 The BRC of all systems remains relatively high, indicating a strong interaction between CR and PSi surfaces, even in the absence of SMA. However, when tested in the presence of ammonia, APTES shows slightly higher BRC than TESPT, possibly due to the high reactivity of the amino end chain of APTES toward CR molecules.
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
Figure 3 shows Q value of the CR-cured specimens measured under different test conditions. Results reveal that the magnitudes of the swelling ratio are in the following order: untreated system > PEG-treated system > APTES-treated system > TESPT-treated system. In general, the cross-linking reaction of CR in the presence of ETU and ZnO is initiated by the rearrangement of 1,2-isomer and followed by the substitution of the chlorine atom with oxygen from ZnO, creating an active site on the CR chain. The sulfur atom of ETU can then react with the CR chain to induce cross-linking.3,25,26 Apparently, the untreated system shows the highest Q, indicating its lowest cross-link density, possibly due to the highest adsorption level of curatives (i.e., ZnO and ETU) on the PSi surfaces. By treating with PEG, the amount of adsorbed curatives on PSi surfaces is reduced, leading to a slight increase in cross-link density. Results also reveal that the presence of SCAs could further enhance the cross-link density, especially in the TESPT-modified system. In addition to the high surface wetting efficiency of SCAs, the greatest cross-link density found in the TESPT-modified system might arise from the dissociation of sulfur linkage of TESPT during the curing process.3,27,28 When tested under ammonia conditions, Q is increased in all systems. The results imply that the swelling behavior is governed by not only cross-link density but also rubber–filler interaction, which could somehow be cleaved by the exposure to ammonia environment.23,29
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
Figure 4 shows SEM micrographs of the PSi-filled CR with and without SMAs. Figures 4a and Figure 5 show PSi agglomerates with broad size distribution in the untreated PSi-filled CR. As evidenced in Figures 4b–d and 5, with PSi surface modification, the mean particle size, and distribution of PSi agglomerates are reduced. Such decreases are mainly due to the reduced filler–filler interaction via the reaction between the alkoxy or hydroxy groups of SMAs and the silanol groups on PSi surfaces. The results agree well with the Payne effect results as discussed previously (Figure 1).
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
Influences of the SMAs on hardness and tensile properties of the PSi-filled CR vulcanizates are illustrated in Figures 6–9. The CR filled with untreated PSi shows the lowest hardness (Figure 6) and 100% modulus (Figure 7), due to its relatively low cross-link density as discussed above. Among all SMAs, TESPT gives the highest hardness and modulus, followed by APTES and then PEG. The explanation is given by the combined effects of increased cross-link density and improved rubber–filler interaction. Although APTES gives relatively high rubber–filler interaction, the APTES-treated system possesses lower modulus and hardness than the TESPT-treated system, which could be explained by its lower cross-link density. Evidently, the PEG-treated system shows the lowest hardness and modulus compared to other treated systems. It is understandable because PEG helps improve mainly the PSi dispersion with no marked enhancements in cross-link density and rubber–filler interaction. As demonstrated in Figure 8, tensile strength of the PSi-filled CR is significantly increased in the presence of SMAs. Surprisingly, the APTES-modified system gives the highest tensile strength, a strength that is significantly higher than the TESPT-treated system. This unexpectedly low tensile strength found in the TESPT-treated system could be explained by the excessive cross-link density as a result of sulfur released from TESPT during the curing process. It is well known that elongation at break depends strongly on cross-link density, that is, the higher the cross-link density, the lower the elongation at break.10,21 As a consequence, the TESPT-treated system gives the lowest elongation at break due to its lowest swelling degree, as evidenced in Figure 9. The low swelling degree could be caused by the high cross-link density in association with high bound rubber content. Also, elongation at break of the PEG-treated system is slightly higher than that of the untreated system, despite its lower swelling degree. The enhanced filler dispersion offered by PEG could be used to explain the results.
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
Figure 10 reveals the compression set results as affected by the SMAs. The CR filled with untreated PSi has the poorest compression set, possibly due to (i) its poorest filler dispersion and (ii) the lowest cross-link density and bound rubber content as indicated by the highest swelling degree. By treating PSi surfaces with PEG, a slight improvement in compression set is observed that is thought to be the consequence from reduced filler–filler interaction, improved filler dispersion, or both. The greater extent of improvement is found in the SCA-treated systems. Apart from the improved filler dispersion, the enhanced cross-link density and the improved rubber–filler interaction play important roles on properties of the systems modified with SCAs. Both TESPT and APTES give comparable compression set.
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
Figure 11 exhibits HBU results of the CR filled with untreated and treated PSi. The HBU results are in the following order: untreated-system > PEG-treated system > APTES-treated system ∼ TESPT-treated system. Evidently, the result trend is in good accordance with that of the compression set because both compression set and HBU are known to be directly related to viscous response of rubber bulk. Thus, similar explanation is applied.
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
Figure 12 shows tan δ at 25 °C of the CR filled with untreated and treated PSi. Results demonstrate that tan δ decreases in the presence of SMAs and may be attributed to the combination of enhanced filler dispersion and increased cross-link density and rubber–filler interaction. Among the three SMAs used in this work, TESPT and APTES show relatively low tan δ, possibly because of their greater ability to enhance the magnitudes of cross-link density and rubber–filler interaction.
Citation: Rubber Chemistry and Technology 90, 1; 10.5254/rct.16.83777
In summary, compared with the results reported previously20 in which the pretreatment technique was used for PSi treatment, we found that both treatment techniques give comparable levels of rubber–filler interaction and degree of filler dispersion. The in situ technique, however, offers the vulcanizates with slightly greater cross-link density and thus modulus and hardness than the pretreatment technique.
CONCLUSIONS
The influence of SMAs on properties of the PSi-filled CR was investigated. Viscoelastic, mechanical, and dynamic properties of the CR vulcanizates were compared. Results showed that both PEG and SCAs are capable of improving PSi dispersion. SCAs improved the mechanical and dynamic mechanical properties of the CR vulcanizates to a greater extent than PEG. Compared with APTES, TESPT provided higher hardness and modulus with comparable elasticity and value of tan δ at 25 °C. APTES, however, gives greater tensile strength and elongation at break than TESPT.
Payne effect magnitude of the PSi-filled CR compounds treated with various SMAs.
BRC of the PSi-filled CR incorporated with various SMAs.
Proposed mechanism of the PSi-filled CR coupling with APTES.
Proposed mechanism of the PSi-filled CR coupling with TESPT.
Swelling ratio of the PSi-filled CR measured under different test conditions.
SEM images of the PSi-filled CR vulcanizates incorporated with (a) untreated PSi; (b) PEG-treated PSi; (c) APTES-treated PSi; and (d) TESPT-treated PSi.
Particle size distribution curves of PSi dispersion in CR matrix treated with various SMAs.
Hardness of the PSi-filled CR treated with various SMAs.
M100 of the PSi-filled CR treated with various SMAs.
Tensile strength of the PSi-filled CR treated with various SMAs.
Elongation at break of the PSi-filled CR treated with various SMAs.
Compression set of the PSi-filled CR treated with various SMAs.
HBU of the PSi-filled CR treated with various SMAs.
Tan δ at 25 °C of the PSi-filled CR treated with various SMAs.
Contributor Notes