Improvement of mechanical performance of solution styrene butadiene rubber by controlling the concentration and the size of in situ derived sol–gel silica particles

Abstract

Incorporation of precipitated silica rubber compounds is now a standard technology in energy efficient green tire manufacturing process. In the present work silica nanoparticles were generated by an in situ sol–gel technique in solution styrene butadiene rubber (SSBR) using tetraethylorthosilicate (TEOS) as silica precursor. A full control over the amount of silica and size of the silica particles is realized while maintaining a good distribution and dispersion level of the nanosilica particles. In the present case the amount of synthesized silica were varied from 10–50 phr (parts per hundred) and the particles size was found to remain in the range of 200–300 nm. The in situ silica based rubber composites offer better processing characteristics as compared with standard commercial precipitated silica at the same loading level. The reinforcing character of in situ silica was further enhanced by the use of a silane coupling agent, i.e. bis[3-(triethoxysilyl)propyl]tetrasulfide (TESPT). Various instrumental techniques, such as tensile test, dynamic mechanical analysis and strain sweep analysis, abrasion test, rebound test and Mooney viscometry reveal superior mechanical performance and good processability of the in situ silica composites. Scanning and transmission electron microscopy studies indicate that the particle size distribution remains in the same range irrespective of the concentration of the silica particles.

---

Introduction

Reinforcement of rubbers by addition of fillers is a most common technique to exploit different soft polymeric materials in several technological applications starting from high performance vehicle tires to floor mats. Plenty of organic and inorganic materials can be utilized to improve the performance of rubbers. Carbon black is one of the major reinforcing materials holding its share since the nineteenth century. Reinforcement by carbon black is mainly achieved through high surface area, smaller particle size and higher surface activity of carbon black particles that result in a strong rubber–filler interaction.^1–4 Especially in research for the past few decades, nanofillers from inorganic origin such as precipitated or fumed silica, nanoclay,^5,6 layered double hydroxide⁷ and halloysite nanotubes⁸ have been used as reinforcing agents in rubber compounds. These inorganic nanofillers offer very good mechanical reinforcement even at low volume fractions. Amongst various inorganic fillers, precipitated silica is widely used and well accepted in the field of tire technology. The reinforcing capability of silica can only be utilized when a silane coupling agent is used in the rubber compounding process.^9,10 Silica filled tires have become more popular in terms of low rolling resistance, improved wet grip and skid resistance compared to conventional carbon black filled tires.^11,12 Even though silica technology is successfully being practiced in industries, some issues are still encountered while using silica as fillers in rubbers. One of the major issues is poor nanoscale dispersion of the silica in highly filled rubber composites.^13–16 To address this problem the generation of silica particles by a sol–gel method offers new possibilities to achieve reinforcing silica particles inside the rubber matrix by simple chemical reactions.^17,18 The schematic of such a sol–gel reaction and silica particle formation is given in Scheme 1. Several techniques are developed to produce sol–gel silica particles inside the polymer which are mainly based on latex blending,^19,20 solution technique²¹ and swelling of rubber matrix in silica precursors.^17,22 Mostly in latex blending technique, the final composites show poor particle growth and mechanical properties. Swelling technique has been explored very rapidly in the last two decades because of its ease of processing, controlled particles growth and low agglomeration tendencies. Recently in situ silica nanoparticles are prepared by swelling of pre-molded unvulcanized rubbers with particular dimension on TEOS and immersed spontaneously in a catalyst solution.^23,24 In most of the literature sol–gel silica was synthesized by swelling vulcanized or unvulcanized rubber in tetraorthosilicate (TEOS) based precursor solution.^{17,19,22,23,25–27} Generating silica particles in a vulcanized rubber material cannot be a good practice for industries as it deforms the shape of the final rubber product.

Scheme 1 Silica particle formation through sol–gel reaction.

In the present study we have followed the sol–gel route to generate silica nanoparticles inside an uncrosslinked SSBR matrix to overcome the above-mentioned issues. In present case the amounts as well as the size of the silica particles are controlled by tuning of reaction conditions.

Experimental

Materials

Solution styrene butadiene rubber (SSBR) BUNA 2525-0 VSL HM containing 25% vinyl content, 25% styrene content, N-cyclohexyl-2-benzothioazolesulfonamide (CBS) and diphenyl guanidine (DPG) were kindly provided by Lanxess chemicals Ltd, Germany. Zinc oxide, stearic acid, sulfur, tetrahydrofuran (THF) and N-butylamine were obtained from Acros Organics, Germany. The silica precursor tetraethoxyorthosilicate (TEOS) with a purity of 99% was purchased from Sigma Aldrich. The silane coupling agent bis[3-(triethoxysilyl)propyl]tetrasulfide (TESPT) and precipitated silica (Ultrasil-VN3) were kindly supplied by Evonik Industries (Essen, Germany) having a purity of 99%.

Preparation of silica–rubber composites

The preparation of SSBR–silica composites was carried out in two stages. In the first stage, silica–rubber masterbatches were prepared. To a rubber solution with 30 g of SSBR in 300 ml of THF, 0.1 mol of TEOS and 0.2 mol of water were added in a round bottom flask. Then 0.025 mol of n-butylamine was added as a catalyst and the whole homogenous mixture was stirred and refluxed for 4 h at 60 °C. The obtained white viscous solution mixture was ultrasonicated for 10 min to avoid the pre-agglomeration of silica particles. The sonicated solution was then slowly poured into 900 ml of ethanol to solidify and precipitate the silica–rubber phase immediately. The precipitated mass was collected by a simple filtering process, washed with ethanol and dried for 48 h at 30 °C in a hot air oven to evaporate the trace solvents. The dried mass contained only the SSBR matrix with sol–gel silica nanoparticles. By varying the quantity of TEOS and water (with the same mole ratio of 1 : 2) different volume fractions of silica in rubber were prepared and quantified by TGA.

In the second stage the rubber–silica masterbatches were compounded with other rubber chemicals. A two-step mixing process was followed to incorporate the rubber chemicals. In the first step, along with the SSBR/in situ-silica masterbatch, an appropriate amount of raw SSBR was added into a internal mixer to achieve the fill factor of the mixing chamber (Haake Rheomix 600P, fill factor of 0.7 which corresponds to 56 cm³ volume). After sufficient mixing, other ingredients were sequentially added: zinc oxide (2 phr), stearic acid (3 phr) and the required amount of TESPT coupling agent (1 phr of TESPT for every 10 phr of silica). The compounding was performed at 110 °C and 80 rpm for 6 min and dumped at an average temperature of ∼140 °C. In a second step, the vulcanizing chemicals CBS (1.4 phr), DPG (1.7 phr) and sulfur (1.4 phr) were added at a two-roll mill (Polymix-110L, Servitec Maschinen Service GmbH, Wustermark, Germany) at 50 °C with a constant friction ratio of 1 : 1.2. The samples were then allowed to mature for 24 h at room temperature. The matured samples were subjected to rheometric study by a rubber process analyzer (Scarabaeus SIS-V50, Scarabaeus GmbH, Wetzlar, Germany) to find the optimum cure time. The compounds were vulcanized into sheets of 2 mm thickness at 160 °C by the use of a compression molding press.

In the sample abbreviations (e.g.: i10, i10s, i30, i30s etc.), ‘i’ represents in situ silica compound, the number is the amount of silica in phr (parts per hundred of rubber) and the suffix ‘s’ represent the compounds containing TESPT silane coupling agent.

Characterization of silica–SSBR nanocomposites

The amount of silica in the rubber compounds was investigated by thermogravimetry (TGA Q 5000 from TA instruments, New Castle, DE, USA) at a heating rate of 20 K min⁻¹ under nitrogen atmosphere up to 600 °C and then under oxygen atmosphere from 600 to 800 °C. From the samples containing only silica and rubber, the quantitative content of silica and the conversion of silica from the precursor can be calculated by the following simple expressions:

To study the silica formation in rubber, solution samples were taken at different periods of time and poured immediately into ethanol. The solidified mass was dried at 30 °C for 2 h and used for further Fourier-transform infrared (FTIR) analysis. To separate the in situ prepared silica powder the final reaction mixture was poured into THF and washed by means of centrifugation process for 5 cycles. Attenuated total reflection (ATR) FTIR spectra were taken using a Vertex 80v spectrometer (Bruker) equipped with both HgCdTe-detector and Golden Gate ATR-unit (Specac). The spectral region was 4000–600 cm⁻¹ and 4 cm⁻¹ spectral resolution was applied. 100 scans were co-added to every spectrum. To compare the spectra properly the data were normalized using the band of CH₂-stretching vibration at 2915 cm⁻¹ as a reference (internal thickness band).²⁸

Filler flocculation characteristics and vulcanization kinetics of the rubber compounds were studied using a Scarabaeus SIS-V50 rubber process analyzer. A three-stage flocculation study was carried out at 120 °C and 1.67 Hz. In the first step a high dynamic strain of ∼25% was applied to destroy all the filler agglomerates in the sample. In second stage, a low dynamic shear strain (1.4%) was applied for 2 h to study the reformation of the filler–filler networks. In the third stage dynamic strain from 0 to 70% was applied to study the Payne effect.²⁹ The flocculation tendency of silica particles and their strain dependency (Payne effect) were calculated by the following equations:

Amplitude of Payne effect = G′_0% − G′_70% (4)

where, G′_120min is the shear modulus at 120 min and G′_0min is the shear modulus during the beginning of the experiment, G′_70% is the modulus at 70% dynamic shear strain and G′_0% is the modulus at ∼0% dynamic strain. Vulcanization kinetics of the rubber–silica composites were conducted in the same instrument in isothermal time sweep mode at 160 °C for 60 min.

Measurements of the viscosity of unfilled and filled rubber compounds were carried out in a Mooney viscometer (Montech Rheotechologies, Germany). A conventional large rotor was used and the measurement conditions were ML (1 + 4) at 100 °C. The Mooney viscosity of the rubber compounds were noted at the 4th minute of the measurement. The gradients of shear induced viscosity drop (thixotropic behavior) of in situ silica included rubber compounds were calculated as the difference between the Mooney units at 0 and 4 min.

The tensile tests were performed with DIN S2 dumbbell specimens as per the DIN 53504 using Zwick/Roell-Z010 material testing machine with an optical elongation sensor at a cross head speed rate of 200 mm min⁻¹ at room temperature. The hardness of rubber composites was measured by a Bareiss Shore-A hardness tester. The dynamic mechanical analysis of different silica filled composites was performed on a dynamic mechanical thermal spectrometer (Gabo Qualimeter, Ahlden, Germany, model Eplexor-150N) in tension mode. The temperature sweep experiments were performed at a frequency of 10 Hz between −60 and 80 °C with a heating rate of 2 K min⁻¹, 0.5% dynamic strain and 1% static strain. The amplitude sweep measurements were performed on Eplexor-2000 N in tension mode at room temperature, at a constant frequency of 10 Hz, 60% pre-strain and dynamic strain from 0.01–30%. Hydrodynamic reinforcement values for the samples were calculated using the Guth–Gold model (eqn (5)). It is assumed that at high values of dynamic strain there would be no filler–filler interaction and the dynamic modulus of the composite will be only influenced by hydrodynamic reinforcement.³⁰ The experimentally obtained strain sweep values and extrapolated higher strain values are further fitted with the Kraus model (eqn (6)) to understand the quantitative information about filler dispersion and filler–filler networks from the strain sweep measurements:

E_c = E_m (1 + 1.25φ + 14.1φ²) (5)

E′(γ_c) = E′_∞ + (E′_i − E′_∞)/(1 + (γ/γ_c)^2m (6)

Here, E_c and E_m are the low-strain dynamic elastic moduli of the composite and pure matrix obtained from dynamic strain sweep measurements, φ is the volume fraction of silica presented in a vulcanizate, γ_c is a critical strain amplitude defining the maximum breakdown of filler–filler network (or strain at which the maximum modulus reduction occurs), γ is the tensile strain amplitude, E′_i and E′_∞ are the initial and final modulus of rubber composites, respectively (here E′_∞ is the modulus extrapolated by Guth–Gold function), and m is a strain amplitude (strain sensitivity) constant.

The SEM micrographs were taken using an Ultra plus electron microscope from Carl Zeiss NTS GmbH, Oberkochen, Germany, 3 kV, 30 μm aperture size, SE2 detector. The rubber samples were subjected to cryofracture after exposure in liquid nitrogen. The fractured surface was further sputter coated with 3 nm platinum using BAL-TEC SCD 500 sputter coater and then examined under the SEM at zero tilt angles. The morphology of silica in rubber matrix was investigated by using TEM model JEM 2010 with 120 kV acceleration voltage and bright field illumination. The ultra-thin sections of silica composites were prepared by ultra-microtomy (Leica Ultracut UCT, Leica microsystems GMbH, Wetzlar, Germany) at −120 °C.

The rebound resilience properties of composites towards free drop pendulum impact force were determined by a resilience tester (Bareiss, Germany) according to DIN 53512, ISO 4662 method at 20 and 60 °C. A circular specimen with 60 mm diameter and 5 mm thickness was utilized for the resilience measurements. The heat build-up experiments were conducted by a dynamic mechanical thermal spectrometer (Eplexor-2000 N, Gabo, Ahlden, Germany). The samples were preconditioned at 50 °C for 30 min. The measurement was performed with 1 MPa preload and 4.45 MPa of dynamic compression load for 25 min as per DIN 53533 standards. Solid cylindrical samples of 25 mm height and 17 mm diameter were used for analysis. The heat generated inside the sample was measured by a sharp tip thermocouple and calculated by eqn (7):

Heat build up (°C) = T₂ − T₁ (7)

where, T₁ and T₂ are the core temperatures of the sample before and after the experiment.

The abrasion experiment was performed by a DIN rotating drum abrader according to DIN 53516 standard procedure. Disc samples of 16.1 mm diameter and 4.2 mm thickness were used for the abrasion test. The abrasion resistance index (ARI) and relative volume loss (ΔV_rel) of SSBR gum and silica filled composites was calculated by eqn (8) and (9):

where, Δm_r is the mass loss of the standard rubber #1 test piece in mg, ρ_r the density of standard rubber #1 in g cm⁻³, Δm_t the mass loss of the test rubber in mg, ρ_t is the density of the test rubber in g cm⁻³, Δm_const is the measured mass loss in standard reference rubber in mg, Δm_r is the mass loss of standard rubber test piece in mg, Δm_t is the mass loss of test rubber piece in mg, and ρ_t is density of test rubber in mg cm⁻³.

The crosslink density of gum and in situ silica filled SSBR composites were investigated by equilibrium swelling method. The volume fraction of rubber (v_r) is determined³¹ by eqn (10):

where, w_i, w_s and w_d are the initial weight of samples before swelling, swollen and dried samples after swelling, respectively, w₀ is the equilibrium weight of solvent absorbed by the samples, w₀ = w_s − w_d, f_ins is the weight fraction of the insoluble substances such as silica and zinc oxide. The classical Flory–Rehner equation is utilized with regard to the volume fraction to calculate the crosslinking density (v_FR = 1/M_c). Two types of network models³² have been proposed to analyze the network structures of cross-linked polymers which are the affine (eqn (11)) and the phantom network (eqn (12)) model; the equations are as follows:hereby, ρ_r is the density of rubber, V_s is the molar volume of toluene, χ is the Flory–Huggins interaction parameter (0.413) for SBR–toluene and f is the crosslink functionality.

Results and discussion

Time dependent silica conversion study

The time dependent silica particle growth inside SSBR matrix investigated by thermogravimetric analysis is depicted in Fig. 1a. The final weight obtained above 700 °C is considered as the amount of silica present in the rubber.¹⁷ Fig. 1a shows the increase in amount of residue with time (every 30 min) indicating the conversion of TEOS into silica. Fig. 1b represents the amount of silica in the masterbatch with time calculated as per eqn (1) and the conversion of silica obtained from eqn (2). The amount of residue obtained in the first 30 min is ∼25% and the conversion is ∼60%. An overall conversion rate of ∼98% is observed within 240 min. For the samples with the highest conversion, almost 2–4% weight loss is observed between 100 and 350 °C and this may be due to the evaporation of trace catalyst, solvent and TEOS.¹⁹ Fig. 2a depicts ATR-FTIR spectra of raw SSBR and TEOS used for the in situ silica synthesis. For raw SSBR, bands are observed at 909 and 967 cm⁻¹, due to wagging vibrations of cis-1,2 and trans-1,4 –CH butadiene groups,³³ respectively. In the case of TEOS, a strong band observed at 1070 cm⁻¹ represents the stretching vibration of Si–O–C groups. At the same time, multiple bands are observed in the 3000–2800 cm⁻¹ region for all pure substances, which pertains to the –CH₂– and –CH₃ stretching vibrations.³⁴

Fig. 1 (a) Time-dependent silica conversion from TEOS inside the SSBR matrix analyzed by thermogravimetric analysis. (b) Percentage of silica conversion and residue with respect to reaction time.

Fig. 2 (a) ATR-FTIR spectra of pure chemical substances used for in situ silica synthesis. (b) Normalized ATR spectra of the solidified reacting mixture as kinetic data of sol–gel reaction. (c) ATR spectra of pure in situ silica, commercial silica powder and the final reaction mixture at 240 min.

The qualitative kinetics of silica particle growth inside the rubber matrix analyzed by ATR-FTIR technique is depicted in Fig. 2b. The new growing band at 1044 cm⁻¹ as a function of reaction time can be clearly seen and represents the Si–O–Si stretching vibration of silica oxide that is derived from the TEOS.³⁵ The silica band at 1044 cm⁻¹ becomes more intense gradually which confirms the silica concentration increase in the SSBR.³⁶ The second broad band appears in the silica–rubber system around 3310 cm⁻¹, which is not observed for pure TEOS as well as for raw rubber. This broad band corresponds to the stretching vibrations of hydroxyl groups (–OH) from the silica surface or could also be due to water adsorbed on the surface of the generated silica particles.³⁷ Fig. 2c demonstrates a comparison of ATR spectra of pure in situ silica powder and the final reaction mixture (at 240 min) with the ATR-FTIR spectrum of a commercial silica powder. The spectral coincidence of self-synthesized and commercial silicas (bands appeared around 1044 cm⁻¹) confirms the expected synthesis in the rubber system and shows that SSBR rubber is highly filled with silica.

The quantitative determination of in situ silica present in silica–rubber masterbatches estimated by thermogravimetric analysis is presented in Fig. 3. Higher amounts of silica could be produced with the use of higher amounts of TEOS and water. In Fig. 3 the amount of residue increases with the increase in amount of TEOS and water. The final weight residue obtained at 800 °C is considered as the amount of silica present in the rubber masterbatch. Therefore, from the final weight the amount of in situ silica present in the system is calculated by eqn (1) and converted into ‘phr’.³⁸ To cross check the amount of in situ silica in the masterbatches, commercial precipitated silica compounds are prepared with similar volume fraction and subjected to thermal analysis (samples designated as ‘x’ in Fig. 3). The experimental result reveals a similar temperature loss profile for both the samples with approximately the same final weight loss. This further confirms the estimated amount of in situ silica in the masterbatches to be around 10–50 phr. Fig. 3 demonstrates no weight loss for commercial silica compounds at 100–350 °C affirming the presence of small amount of catalyst, solvent and TEOS in the in situ silica compounds.¹⁹

Fig. 3 Thermogravimetric analysis of in situ and commercial silica–rubber masterbatches.

Characterization of the rubber compounds

Fig. 4 compares the Mooney viscosities of commercial and in situ silica compounds and extracted data are summarized in Table 1. Gum SSBR shows a lowest value of 59 MU and increases significantly with increase in silica content. In situ silica based SSBR compounds exhibits lower Mooney viscosity compared to commercial silica compounds.³⁹ A lower Mooney viscosity is always preferred as such compounds display good rheological properties and ease of processing. The Mooney viscosity of 50 phr commercial silica (without plasticizers or oil) compound is too high and could not be evaluated as it exceeds the measuring capability of the instrument. The commercial silica particles form a filler–filler network structure resulting in an increase in Mooney viscosity. In contrast, in situ silica particles form less networks or weak networks and are freely mobile when subjected to shear. Such characteristics could be the reason for the lower viscosity of the in situ compounds. Interestingly, the gradient of viscosity drop under shear stress of in situ and commercial silica compounds are very similar for the same volume fraction.⁴⁰ The filler flocculation and filler–filler networks in silica–rubber masterbatches are investigated to understand the agglomeration and aggregation tendency of the in situ silica particles during processing. It is well known that SSBR and many general purpose rubbers are hydrophobic and that silica is hydrophilic in nature. Due to the difference in surface energies between rubber and silica, the silica particles tend to aggregate. Such a tendency of fillers to form aggregates in the rubber is called filler flocculation.²⁹ This affects the processability and final properties of the compounds to a large extent. Formation of filler networks significantly influences the elastic modulus of highly filled rubber compounds. To understand such effects, dynamic strain sweep measurements are often conducted. The dynamic filler flocculation and strain dependent filler–filler network characteristics are depicted in Scheme 2.

Fig. 4 Effect of Mooney viscosity with different silica fraction in SSBR at 100 °C.

Table 1 Mooney viscosity of varies silica filled SSBR compounds

Sample	Commercial silica			In situ silica
	MU0	MU4	MU0–4	MU0	MU4	MU0–4
Gum	80	59	21	80	59	21
10	126	81	45	100	66	34
20	147	103	44	123	82	41
30	177	131	46	156	110	46
40	229	183	46	177	131	46
50	NA	NA	NA	224	147	77

---

Scheme 2 A schematic representation of filler flocculation and filler–filler network breakdown under dynamic strain sweep measurements.

Fig. 5 shows the three different stages involved in the filler flocculation study. In a first stage, a constant 25% dynamic strain is applied for 5 min assuming a complete destruction of the filler networks. The dynamic shear modulus is higher for higher volume fractions of silica due to a hydrodynamic effect. The experiment is then continued to the second stage with low strain amplitude of 1.4% for 120 min to facilitate filler network reformation. With time, a slow increase in the shear moduli (G′) of the filled compounds is evident in Fig. 5. The increase in G′ is truly dependent on the volume fraction of the filler and the filler–filler aggregation characteristics. Therefore, high silica content would result in a higher aggregation tendency. Finally in the third stage, the compounds are subjected to strain sweep measurements (Payne effect up to 70% dynamic strain) and the experimental results shows that highly filled silica rubber compounds exhibit faster reduction of the G′ values. The rate of flocculation and the amplitude of Payne effect are calculated quantitatively as per eqn (3) and (4), respectively, and depicted in Fig. 6.

Fig. 5 Dynamic filler flocculation and strain sweep measurements.

Fig. 6 Flocculation and dependent Payne effect of in situ silica–SSBR compounds.

Vulcanization characteristics of in situ silica SSBR composites are summarized in Table 2 and their respective cure curves are given in Fig. 7. The unfilled SSBR gum compound exhibits higher scorch time (t₂) and optimum cure time (t₉₀), compared to in situ silica filled compounds. Interestingly, the scorch times and optimum cure times of the in situ silica filled composites without silane coupling agent are very similar for all filler contents. The result clearly shows the vulcanization behavior of SSBR not being affected by the increasing amounts of in situ silica. Such results are unique as compared to commercial silica filled compounds which retard the vulcanization reaction at higher filler concentrations, leading to higher scorch times and optimum cure times, due to its acidic nature.⁴¹ Since in situ silica particles are generated through base catalysis sol–gel reaction, it does not affect the vulcanization kinetics even at higher concentrations. Meanwhile, in the presence of silane coupling agent the scorch times are similar, whereas the composites show higher optimum cure times with increasing silane concentration.

Table 2 Vulcanization characteristics of SSBR/in situ silica composites

Sample	t2/min	t90/min	S′min /dN m	S′max/dN m	CRI/min^−1
Gum	2.35	5.56	0.78	9.34	31.15
i-10	2.05	4.11	1.06	10.95	48.54
i-20	2.11	4.00	1.42	13.43	52.91
i-30	1.59	4.19	1.96	17.7	38.46
i-40	2.05	4.11	2.31	20.21	48.54
i-50	1.59	4.10	3.11	25.58	39.84
i-10s	1.56	5.33	0.99	11.6	26.52
i-20s	2.23	9.23	1.32	15.38	14.28
i-30s	2.01	11.12	1.58	20.03	10.76
i-40s	2.17	12.10	2.01	24.69	10.07
i-50s	1.59	13.45	2.89	30.98	8.43

---

Fig. 7 Vulcanization curves of SSBR/in situ silica composites in (a) absence and (b) presence of silane coupling agent.

The cure rate index (CRI) of the compounds decrease with higher dosage of the silane coupling agent. It is evident that the rate of cure is affected by the incorporation of silane, as higher CRI values are observed in the absence of the silane coupling agent. Therefore, in situ silica compounds show faster cure characteristics in the absence of silane.⁸ The maximum torque (S′_max) increases with the amount of in situ silica in both pristine as well as silane assisted composites. Nevertheless, a higher maximum torque is observed for silane assisted in situ silica composites. Additionally, pristine in situ silica compounds exhibit reversion signifying a higher proportion of polysulfidic crosslink bridges.

Mechanical and dynamic mechanical properties

The stress–strain characteristics of in situ silica in the presence and absence of silane coupling agent are analyzed by tensile measurements and results are tabulated in Table 3 and depicted in Fig. 8. The tensile strength and modulus of SSBR composites are improved significantly by the in situ silica. The stress–strain behavior of the filled composites in the absence of silane coupling agent show a gradual increase in moduli and tensile strength compared to the unfilled vulcanizate. However, the stress–strain plots do not follow any pattern at higher silica loadings, indicating higher anisotropy in the composites. Nevertheless, the composites with silane coupling agent offer better filler–polymer interactions, exhibiting huge improvements in moduli and tensile strength. Unfortunately, as a consequence of better filler–polymer interaction by silane, the stiffness of the matrix increases, compromising the elongation at break. Moreover, the stress–strain plots of in situ silica composites with silane display higher modulus in the low strain region also behaving like Hookean materials. To quantify the amount of reinforcement enhanced by the silica particles, the reinforcing efficiency (RE)¹⁴ is estimated by:

Table 3 Stress–strain properties of SSBR/in situ silica composites^a

Sample	M100%/MPa	M200%/MPa	M300%/MPa	TS/MPa	RE (%)	EB (%)	HRD (shore A)
a M-modulus at, TS-tensile strength, RE-reinforcing efficiency, EB-elongation at break, HRD-hardness.
Gum	0.87	1.76	—	2.72	—	290	42
i-10	1.27	2.69	4.42	4.65	4.59	310	48
i-20	1.87	3.73	5.40	6.10	6.25	340	53
i-30	2.10	3.59	4.73	5.88	5.53	390	56
i-40	2.87	4.87	5.87	5.92	7.24	308	60
i-50	2.93	6.98	—	7.53	6.38	449	63
i-10s	1.08	2.13	3.72	5.58	2.43	390	48
i-20s	1.56	3.62	6.45	6.53	4.38	305	54
i-30s	2.95	6.95	—	9.72	9.57	280	59
i-40s	4.25	—	—	10.98	12.58	205	61
i-50s	7.53	—	—	12.5	21.31	180	68

---

Fig. 8 Tensile properties of SSBR/in situ silica composites in (a) absence and (b) presence of silane coupling agent.

The dynamic mechanical properties of unfilled SSBR and its in situ silica composites studied over a temperature range of −60 to 80 °C are depicted in Fig. 9 and various important parameters are summarized in Table 4. Fig. 9a and c illustrates the gradually increasing storage moduli of the in situ silica composites according to the silica fraction. From Table 4, improvements in the dynamic mechanical properties of in situ silica composites due to enhanced filler–polymer interaction by the silane coupling agent could be visualized. Fig. 9b and d shows the tan δ plot for the in situ silica SSBR composites. It is well known that addition of silica in rubber reduces tan δ_max values and further reduction would be observed upon incorporation of silanes.⁴² In Fig. 9b the glass transition temperature (T_g) is slightly shifted to higher temperatures for all the composites. As is evident in Fig. 9d, i.e. in the presence of silane coupling agents, the T_g shifts are significant.¹⁴ This result confirms that the silica–rubber interface is substantially improved by effective silanization as well as by formation of strong nanolayers of immobilized rubber chains around the silica surface.²⁹ Thus restricted mobility of rubber chains by a silane modified silica surface causes the shift in T_g.⁴³ From the temperature sweep studies, in situ silica reinforces the SSBR matrix very well and the dynamic behavior of the composites are further improved with the presence of coupling agents.¹⁸

Fig. 9 Dynamic mechanical temperature sweep analysis of in situ silica–SSBR nanocomposites: (a), (c) storage modulus and (b), (d) tan δ graph for SSBR/in situ silica composites in the absence and presence of coupling agent.

Table 4 Dynamic mechanical properties of SSBR/in situ silica composites

Sample	E′ @ 0 °C/MPa	E′ @ 25 °C/MPa	E′ @ 60 °C/MPa	tan δmax	Tg/°C
Gum	2.85	2.44	2.14	1.78	−27.8
i-10	3.29	2.90	2.59	1.74	−27.5
i-20	4.88	4.13	3.79	1.65	−27.0
i-30	6.58	5.58	5.05	1.56	−26.7
i-40	9.05	7.41	6.48	1.51	−27.1
i-50	12.78	9.81	7.97	1.49	−27.5
i-10s	3.68	3.12	2.89	1.72	−27.0
i-20s	6.06	5.33	5.06	1.61	−26.8
i-30s	7.20	6.12	5.74	1.53	−26.5
i-40s	10.2	8.32	7.26	1.51	−26.0
i-50s	13.2	10.85	9.75	1.41	−25.2

---

The dynamic mechanical strain sweep measurements, also known as the ‘Payne effect’, are performed to understand the filler–filler interaction and strain induced softening of the composites. The obtained experimental results are fitted with the Kraus equation (eqn (6)) by extrapolating to 1000% strain; where only hydrodynamic reinforcement exists (see eqn (5)). The results are plotted in Fig. 10 and quantitative information about the Payne effect is tabulated in Table 5.

Fig. 10 Strain dependency of dynamic elastic modulus for the SSBR/in situ silica composites (symbols represent experimental data and lines represent the fitted Kraus equation).

Table 5 Calculated strain dependent dynamic properties of composites by the Kraus function^a

Sample	(E′i − E′∞)/MPa		Critical strain (γc) (%)		Amplitude constant (m)
	In situ w/o silane	In situ with silane	In situ w/o silane	In situ with silane	In situ w/o silane	In situ with silane
a w/o = without.
Gum	0.22	—	—	—	—	—
10	0.42	1.04	55.06	53.53	0.54	0.53
20	1.70	2.12	53.24	53.38	0.50	0.48
30	3.30	4.60	42.97	41.93	0.48	0.46
40	4.98	5.99	39.48	40.85	0.49	0.45
50	7.23	8.34	30.59	35.92	0.47	0.43

---

In Fig. 10, the solid lines represent samples without silane and dotted lines represent the presence of silane coupling agents in the Kraus fitting function.⁴⁴ As the volume fraction of silica is increased in the rubber matrix, well percolated filler–filler networks are formed. With the incorporation of silane coupling agent, the in situ silica rubber compounds display an increase in the dynamic storage modulus. This effect is contrary to commercial precipitated silica filled composites for which silane incorporation reduces the storage modulus.¹⁸ Similar effects (increase in dynamic modulus) are usually observed for carbon nanotube filled systems.^45,46 However, the Payne effect is observed for the in situ silica compounds both in the presence and absence of silane coupling agent. At the same time, the critical strain value (γ_c) of the composites reduces with increase in silica fraction. Such behavior indicates a higher dependency of the filler networks towards dynamic strain as the filler fraction increases. The silane incorporated composites show less strain dependency by displaying higher critical strain values, meaning the filler networks are less susceptible to the dynamic strain. Accordingly, silane incorporation significantly improves the dynamic performance of the composites by improving the filler–polymer interaction. The amplitude constant (m) constantly decreases with increase in filler content and lies in between 0.43 to 0.54.⁴⁴

Rubber performance

Some important rubber performance parameters such as rebound resilience, heat buildup, abrasion and swelling in solvent is also evaluated and obtained results are depicted in Fig. 11 and 12. At room temperature, the resilience values are found to be reducing with increase in the amount of fillers (Fig. 11a). At the same time, resilience properties are greatly improved with the addition of silane in the in situ silica composites;¹⁹ even the high temperature (60 °C) resilience properties are improved.

Fig. 11 (a) Rebound resilience characteristics, (b) heat build-up properties, (c) relative volume loss, and (d) abrasion resistance index (ARI) % of in situ silica filled SSBR composites.

Fig. 12 (a) Effect on crosslink density of SSBR with respect to silica fraction and silane. (b) Correlation between T_g and crosslink density.

The heat build-up properties of in situ silica filled SSBR composites in the presence and absence of silane coupling agent are investigated as per eqn (7) and are plotted in Fig. 11b. The unfilled gum vulcanizate shows the least heat build-up inside the core of the sample. At the same time, as the amount of silica fraction increases in the SSBR, the degree of heat generation and build-up inside the composites is increased significantly. Inclusion of rigid particles in a viscoelastic material would exhibit excessive heat generation by increasing the internal friction due to poor filler polymer interaction and a lower heat loss.⁴⁷ Evidently, incorporation of silane coupling agent in the silica filled system reduces heat generation and increases the heat loss of the material by improving the filler–polymer interaction due to silanization.¹⁴

The abrasion properties of in situ silica filled SSBR compounds are plotted in Fig. 11c and d. The abrasion resistance index (calculated by eqn (8)) clearly depicts the improved resistance of the composites towards friction with the addition of in situ silica. At the same time, addition of silane coupling agent results in better resistance towards abrasion. The silane improves the abrasion characteristics through better filler–polymer interaction. The relative volume loss (calculated by eqn (9)) is also significantly reduced by increasing the silica fraction and silane coupling agent.

The crosslink densities of gum and in situ silica filled SSBR composites are calculated by equilibrium swelling based Flory–Rehner equations, eqn (11) and (12). The estimated crosslink densities based on affine and phantom network models are depicted in Fig. 12a. The crosslink density of SSBR gradually increases with the volume fraction of silica. Also, the silane coupled in situ silica composites exhibit higher crosslink densities, compared to the unsilanized composites.¹⁴ The enhanced crosslink density of silane incorporated SSBR composites is due to the adsorbed silane coupling agent on the surface of the filler and improved filler–rubber interaction. This improves the filler–polymer interaction through covalent bonding between the rubber and silica surface.⁴⁸ The crosslink densities predicted by the affine network model are lower than by the phantom network model in both silane and silane free systems. Fig. 12b shows the plot of crosslink density obtained from the phantom network model against the glass transition temperature (T_g) measured by the dynamic mechanical analyzer (DMA). Linearity is observed between crosslink density and T_g with increasing volume fraction of filler. In the absence of silane the obtained slope is 2.0, however with the incorporation of silane a higher slope of 3.5 is observed. This confirms that crosslink density and T_g of composites are mainly dependent on the silanization reaction as well as filler volume fraction. This correlation further evidences the enhanced silica–rubber interaction by the silane.

Morphology

To understand the in situ silica–rubber interaction and interfacial adhesion in the presence and absence of silane coupling agent, the fracture surface of the composites are analyzed by SEM. In Fig. 13a, the 30 phr in situ silica composite without silane shows aggregated particles with an average particle size of around 200–300 nm. The fracture surface shows some holes, which represent the poorly adhered particles being pulled out from the matrix during cryo-fracture. In Fig. 13b, in the presence of coupling agent, the in situ silica the particles are fully covered by rubber layers. The micrograph shows the superior adhesion of the in situ silica fillers with the rubber matrix as a benefit of the silane coupling agent. The particle size and dispersion of silica nanoparticles in rubber matrix is one of the most critical parameters which determine the overall performance of the nanocomposite.

Fig. 13 SEM micrographs of cryo-fracture failure surface investigation of in situ silica composites in (a) absence and (b) presence of TESPT silane coupling agent at low and high magnifications.

Aggregates and agglomerates of filler particles in a rubber matrix act as stress concentration points resulting in early failure. Fig. 14 displays the TEM images of silane coupled SSBR/in situ silica composites. The silica particles are observed to be generally individual and large in size with an average particle size of around 200 to 300 nm. Silane coupling agents improve the dispersion of silica particles although some particles formed are interconnected during synthesis.

Fig. 14 TEM images of silane modified SSBR/in situ silica composites.

Conclusions

In situ synthesis of silica nanoparticles in SSBR matrix was achieved successfully by a sol–gel method. In situ silica composites offered improved dynamic mechanical properties, lower heat buildup, higher rebound resilience, better processability and abrasion characteristics in the presence of silane coupling agent. In comparison to existing literature, the current approach promises (a) an improved state of silica dispersion, (b) higher amounts of silica loading into the rubber without sacrificing the processability, (c) a better way to control and customize the particle size and distribution, (d) good industrial feasibility, and (e) a possible way to save mixing energy due to easy incorporation. Currently, the focus is being laid to understand the reinforcement mechanism of in situ silica in the rubber on the basis of (i) surface activity,²⁴ (ii) particle distribution, (iii) trapped rubber chains inside the silica particles¹² and (iv) the physical and chemical interactions. Detailed study of the results will be communicated in our forthcoming publications.

Acknowledgements

The authors are thankful to Dr K. J. Eichhorn for FTIR discussion and Mrs S. Krause for TGA analysis.

Notes and references

1. E. M. Dannenberg, Rubber Chem. Technol., 1975, 48, 410–444.
2. J.-B. Donnet and A. Voet, Carbon black: physics, chemistry, and elastomer reinforcement, M. Dekker, 1976.
3. G. Kraus, Rubber Chem. Technol., 1978, 51, 297–321.
4. M. Klüppel and G. Heinrich, Rubber Chem. Technol., 1995, 68, 623–651.
5. S. Rooj, A. Das, K. W. Stöckelhuber, D.-Y. Wang, V. Galiatsatos and G. Heinrich, Soft Matter, 2013, 9, 3798.
6. A. Das, R. Jurk, K. W. Stöckelhuber and G. Heinrich, Macromol. Mater. Eng., 2008, 293, 479–490.
7. S. B. Eshwaran, D. Basu, S. R. Vaikuntam, B. Kutlu, S. Wiessner, A. Das, K. Naskar and G. Heinrich, J. Appl. Polym. Sci., 2015, 132, 41539.
8. V. Raman, S. Rooj, A. Das, K. Stöckelhuber, F. Simon, G. Nando and G. Heinrich, J. Macromol. Sci., Part A: Pure Appl. Chem., 2013, 50, 1091–1106.
9. U. Goerl, A. Hunsche, A. Mueller and H. Koban, Rubber Chem. Technol., 1997, 70, 608–623.
10. S. Wolff, Rubber Chem. Technol., 1996, 69, 325–346.
11. N. Vleugels, W. Pille-Wolf, W. K. Dierkes and J. W. M. Noordermeer, Rubber Chem. Technol., 2015, 88, 65–79.
12. E. Miloskovska, C. Friedrichs, D. Hristova-Bogaerds, O. Persenair, M. van Duin, M. R. Hansen and G. de With, Macromolecules, 2015, 48, 1093–1103.
13. F. Bohin, I. Manas-Zloczower and D. Feke, Rubber Chem. Technol., 1994, 67, 602–609.
14. A. Hashim, B. Azahari, Y. Ikeda and S. Kohjiya, Rubber Chem. Technol., 1998, 71, 289–299.
15. P. Cassagnau, Polymer, 2003, 44, 2455–2462.
16. J. Ramier, C. Gauthier, L. Chazeau, L. Stelandre and L. Guy, J. Polym. Sci., Part B: Polym. Phys., 2007, 45, 286–298.
17. S. Kohjiya and Y. Ikeda, Rubber Chem. Technol., 2000, 73, 534–550.
18. A. Das, R. Jurk, K. Werner Stöckelhuber and G. Heinrich, J. Macromol. Sci., Part A: Pure Appl. Chem., 2007, 45, 101–106.
19. S. Kohjiya, K. Murakami, S. Iio, T. Tanahashi and Y. Ikeda, Rubber Chem. Technol., 2001, 74, 16–27.
20. S. Prasertsri and N. Rattanasom, Polym. Test., 2011, 30, 515–526.
21. B. Chaichua, P. Prasassarakich and S. Poompradub, J. Sol-Gel Sci. Technol., 2009, 52, 219–227.
22. Y. Ikeda and S. Kohjiya, Polymer, 1997, 38, 4417–4423.
23. E. Miloskovska, E. Nies, D. Hristova-Bogaerds, M. van Duin and G. de With, J. Polym. Sci., Part B: Polym. Phys., 2014, 52, 967–978.
24. E. Miloskovska, M. R. Hansen, C. Friedrich, D. Hristova-Bogaerds, M. van Duin and G. de With, Macromolecules, 2014, 47, 5174–5185.
25. L. Bokobza and J.-P. Chauvin, Polymer, 2005, 46, 4144–4151.
26. Y. Ikeda, S. Poompradub, Y. Morita and S. Kohjiya, J. Sol-Gel Sci. Technol., 2008, 45, 299–306.
27. Y. Ikeda, A. Tanaka and S. Kohjiya, J. Mater. Chem., 1997, 7, 1497–1503.
28. J. L. Koenig, Spectroscopy of polymers, Elsevier, 1999.
29. K. W. Stöckelhuber, A. S. Svistkov, A. G. Pelevin and G. Heinrich, Macromolecules, 2011, 44, 4366–4381.
30. G. Heinrich, M. Klüppel and T. A. Vilgis, Curr. Opin. Solid State Mater. Sci., 2002, 6, 195–203.
31. J. L. Valentín, I. Mora-Barrantes, J. Carretero-González, M. A. López-Manchado, P. Sotta, D. R. Long and K. Saalwächter, Macromolecules, 2010, 43, 334–346.
32. A. Vieyres, R. Pérez-Aparicio, P.-A. Albouy, O. Sanseau, K. Saalwächter, D. R. Long and P. Sotta, Macromolecules, 2013, 46, 889–899.
33. V. M. Litvinov and P. P. De, Spectroscopy of rubbers and rubbery materials, iSmithers Rapra Publishing, 2002.
34. G. Socrates, J. Am. Chem. Soc., 1995, 117, 1671.
35. J. Martinez, F. Ruiz, Y. V. Vorobiev, F. Pérez-Robles and J. González-Hernández, J. Chem. Phys., 1998, 109, 7511–7514.
36. M. L. Kralevich and J. L. Koenig, Rubber Chem. Technol., 1998, 71, 300–309.
37. A. Beganskienė, V. Sirutkaitis, M. Kurtinaitienė, R. Juškėnas and A. Kareiva, J. Mater. Sci., 2004, 10, 287–290.
38. Y. Ikeda and Y. Kameda, J. Sol-Gel Sci. Technol., 2004, 31, 137–142.
39. S. Mihara, R. Datta and J. Noordermeer, Rubber Chem. Technol., 2009, 82, 524–540.
40. N. K. Dutta and D. Tripathy, J. Appl. Polym. Sci., 1992, 44, 1635–1648.
41. P. Pal and S. De, Rubber Chem. Technol., 1982, 55, 1370–1388.
42. Y. Ikeda, A. Tanaka and S. Kohjiya, J. Mater. Chem., 1997, 7, 455–458.
43. G. Tsagaropoulos and A. Eisenberg, Macromolecules, 1995, 28, 6067–6077.
44. G. Heinrich and M. Klüppel, in Filled Elastomers Drug Delivery Systems, Springer, 2002, pp. 1–44.
45. A. Das, K. Stöckelhuber, R. Jurk, M. Saphiannikova, J. Fritzsche, H. Lorenz, M. Klüppel and G. Heinrich, Polymer, 2008, 49, 5276–5283.
46. K. Subramaniam, A. Das and G. Heinrich, Compos. Sci. Technol., 2011, 71, 1441–1449.
47. M. P. Wagner, Rubber Chem. Technol., 1976, 49, 703–774.
48. S.-J. Park and K.-S. Cho, J. Colloid Interface Sci., 2003, 267, 86–91.

---