STRUCTURE ANALYSES OF SWOLLEN RUBBER–CARBON BLACK SYSTEMS BY USING CONTRAST VARIATION SMALL-ANGLE NEUTRON SCATTERING

INTRODUCTION

In our daily life, rubber–filler systems are commonly used in manufacturing materials such as tire and belts so on.¹ Fillers reinforce rubbers by compounding and improving the mechanical and barrier properties of the rubber compounds. Recently, the reduction of the energy loss in cars has been one of the most important issues in global warming countermeasures. To reduce the loss energy in car driving, it is important to investigate the internal structures formed in tire or rubber–filler systems and to explore the relationship between the physical properties of the systems. Thus, the analyses of the internal structure can contribute to the reduction of the loss of the energy.

Since fillers in polymers form hierarchical structures over a wide length scale, the exploration of the whole structure was rather difficult. Recently, an ultra-small-angle scattering technique has become available for us to measure the structure on the order of micron scale and the combination of various scattering techniques can cover the wide length scale where fillers in polymers forms hierarchical structures. In previous studies,^2–9 the hierarchical structures formed by carbon black (CB) filler in polymers have been investigated by using a combined ultra-small-angle and small-angle scattering method of neutrons and X-rays. These studies explored the network structures of CB having mass-fractal features, the size and shape of aggregation, and surface-fractal features of CB particles. In particular, Marr et al. explored the void structure in polyethylene/CB systems by combing the small-angle scattering method of neutrons and X-rays.¹⁰ Moreover, we have recently characterized the adsorption layers as well as the aggregation of silica particles by using contrast variation small-angle neutron scattering techniques described later.¹¹ Usually, it is difficult to characterize the adsorption layer of polymers from the scattering of bulk systems, since there is no contrast of scattering between adsorption layers and the matrix phase of rubber. However, by swelling the rubber–silica systems, we can induce the contrast between the matrix phase and adsorption layer since the swollen ratio of the absorbed layers is smaller than that of the matrix phase. The contrast variation small-angle neutron scattering technique for the swollen systems enables us to characterize the adsorption layer quantitatively.

In this study, we focus on a rubber–CB particles system and characterize the adsorption layers on CB particles by using the contrast variation small-angle neutron scattering (SANS) technique and characterize the adsorption layers as well as the aggregation of CB particles, network structures in matrix, and so on.

EXPERIMENTAL METHODS

SAMPLES

We used poly(styrene-ran-butadiene) (SBR; Nipol 1502, ZEON Corporation) as rubber. The characteristics of SBR are listed in Table I. The CB used in this study was N339 black. The average size of primary particle of N339 black is estimated to be 13 nm. The composition of sample used in this study is listed in Table II. The compounds were mixed by using Banbury mixer at 150 °C and then molded at 160 °C for 50 min to make the sheets of SBR/CB with 1 mm thickness. The samples were swollen by a mixture of deuterated toluene, (d-tol) and tolene (h-tol) with various composition (d-tol/h-tol = 100/0, 70/30, 50/50, 30/70, and 0/100 by volume fraction). Scattering length density is listed in Table III. After reaching their equilibrium state (typically 12 h), we installed them into quartz cells for further scattering experiments.

Table I Characterization of SBR

View Fullscreen

Table II Composition of Samples Used in This Study

View Fullscreen

Table III Scattering Length Density of Each Component Used in This Study

View Fullscreen

SANS MEASUREMENTS

We conducted SANS measurements with the SANS-J-II spectrometer at Japan Research Reactor-3 (JRR-3) in the Japan Atomic Energy Agency (JAEA), Tokai, Japan.¹² The temperature of samples was set to be 27 °C. The wavelength of the incident beam was 0.65 nm, and its distribution was 0.12. The scattered intensity was detected by a two-dimensional ³He position-sensitive detector. The distances of sample to detector were 2.5 m and 12 m, and thus the observed q range was from 0.07 nm⁻¹ to 0.8 nm⁻¹, where q is a magnitude of the scattering vector defined by

with θ being the scattering angle. The obtained scattering data were circularly averaged and corrected for background scattering of cell, electronic noise of detector, detector sensitivity, and incoherent scattering.

RESULTS AND DISCUSSION

Figure 1 shows the change in the scattering profiles I(q) of the swollen rubber–CB system with the scattering length density of solvent. The scattering intensity increases with h-tol. The q-dependence of I(q) changes with the scattering length density, suggesting that the swelling ratio of solvent to rubber is spatially inhomogeneous and that the network density has spatial fluctuations.

View Figure

• Download as Powerpoint
• Download Figure

We analyzed the scattering profiles by using the contrast variation method developed by Endo.^13–16 The swollen rubber–CB system can be treated as three components systems consisting of polymer, CB, and solvent, and their scattering profiles under incompressible conditions can be described as follows:

Here a_i is the scattering length density of i component (i = P, SBR; C, CB; and T, solvent consisting of d-tol and h-tol). S_ij(q) is partial scattering function defined by

where V is the scattering volume radiated by incident beam and is the fluctuation of volume fraction of i at position . In this experiment, we obtained the vector of scattering intensities from the same samples with n different scattering length density of T. From Eq. 2 can be expressed by

where is a matrix of the difference of the scattering length density and is vector of partial scattering functions. For three component systems, is expressed by

where

is given by

To decompose the scattering intensities into partial scattering functions, we need to calculate the transposed matrix M^T satisfying M^T·M = E by singular value decomposition. By applying M^T to , can be obtained:

Figure 2 shows the calculated partial scattering functions. S_PP (q) and S_CC (q) are positive while S_PC (q) is negative. If there is no adsorption layer and SBR is swollen by toluene homogeneously, S_PC(q) is given by

and the q-dependence of −S_PC(q) is identical with that of S_CC(q). Here φ_SBR is the volume fraction of polymers in swollen network. However, as shown in the plots of S_CC(q), S_PP(q), and −S_PC(q) versus q in double logarithmic scale of Figure 3, the q-dependence of −S_PC(q) is different from S_CC(q) at a lower q-region, indicating that the swollen network is not spatially homogeneous. If the inhomogeneous regions exist without any correlation with CB particles, S_PC(q) expressing the cross-correlation between CB and SBR becomes zero. Thus, we believe that the adsorption layer exits around CB particles.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

To characterize the adsorption layer quantitatively, we calculated the scattering functions for the model consisting of the aggregation of CB particles (region α) and the adsorption layers on the CB particles (region β) with the volume fraction of polymer φ_l, and the matrix region (region γ) with the volume fraction of polymer φ_m as shown in Figure 4. According to the model, the scattering amplitude of CB F_CB(q) and polymer F_P(q) are expressed by

where F_α(q) and F_α+β(q) are, respectively, the scattering amplitudes of region α and regions α + β. The partial scattering function can be described by

The partial scattering function S_CC(q) can be expressed by an object with a sharp interface and a radius of gyration R_g,a as a model of the aggregation of CB particles,^17–19

where R_g,a is the radius of gyration of the aggregation of CB particles. A and B are expressed by

n_a, V_a, and S_a are, respectively, the number of aggregations per unit volume, the volume of a CB aggregation, the volume per a CB aggregation, and the surface area per a CB aggregation. The scattering function from regions α + β is expressed by an object with a radius of gyration R_g,l and a broad interface characterized by interfacial thickness σ:

where C and D are given by

with V_l and S_l being the volume and interfacial area of the regions α + β. We fitted the experimental partial scattering functions with the equation described above. The model functions can be well fitted with the experimental results for S_CC(q) and S_PP(q), indicating that there are adsorption layers on the CB particles. The fitting results yielded the characteristic parameters. We listed them in Table IV. In Table IV, the thickness of the adsorption layer t_I is estimated from

As for S_PC(q), the slight disagreement can be observed around q = 0.2 nm⁻¹. This is because the model does not consider the small fraction of polymer within CB particles well.

View Figure

• Download as Powerpoint
• Download Figure

Table IV Characteristic Parameters Yielded from Fitting

View Fullscreen

First we will compare the volume fraction of polymer φ_m estimated from SANS experiment with that from Q. The value of Q estimated the swollen experiment is 2.44. Solvent cannot swell CB particles but it can swell polymer chains. Thus, φ_m estimated from Q is given by

and agrees with φ_m = 0.32.

Next, let us analyze the structure of the aggregation of CB particles. According to Eq. 18 and the volume fraction of CB particles φ_CB in swollen rubber–CB systems, we can estimate the volume per a CB aggregation V_a from the following equations:

From the swelling experiment, φ_CB = 0.20/2.44 = 0.082. Substituting φ_CB, A, a into Eq. 18, we obtained V_a = 7.9 × 10⁻¹⁷ cm³ and n_a=1.0 × 10¹⁵ cm⁻³. Assuming that the primary particle of CB is a sphere with radius 13 nm, the volume V_CB of a primary particle of CB is 9.2 × 10⁻¹⁸ cm³, and thus there are 8.6 primary particles in an aggregation.

Finally, we will compare the results with those obtained for rubber–silica systems previously.⁴ The thickness of adsorption layers of rubber–CB systems is thicker than that for rubber–silica systems, 5.3 nm. Moreover, there are interface regions or gradients of polymer concentration around the CB aggregates. Although the thickness of the layer would strongly depend on kinds of silica and CB, we can conclude that CB aggregates have a thicker adsorption layer than that of silica aggregates. This may be because the compatibility between CB and rubber is better than that between silica and rubber.

CONCLUSIONS

We have investigated the polymer layers absorbed on CB particles in rubber–CB systems with the contrast variation SANS method. Specimens were swollen by the solvents having various scattering length densities and their SANS intensity was measured. We calculated the partial scattering functions by using singular value decomposition: the scattering function for polymer–polymer correlation S_PP(q), the scattering function for CBCB correlation S_CC(q), and the scattering function for polymer–CB correlation S_PC(q). The analyses of S_PC(q) and S_CC(q) explored the existence of dense polymer layers around CB aggregates. S_CC(q) reflects hierarchical structures formed by CB particles. To characterize the adsorption layer quantitatively, we calculated the scattering functions for the model consisting of the aggregation of CB particles, the adsorption layers on the CB particles, and the matrix region. We used an object with a sharp interface for CB aggregates and an object with a broad interface for the regions adsorption layers and CB aggregates to calculate the partial scattering functions. The model can express the experimental partial scattering functions well, and several characteristic parameters are estimated from the analyses, such as the size of aggregates, the thickness of layers, and the volume fractions of polymer layers and matrix. We will apply this technique to other rubber–CB systems and explore how CB structure affects the adsorption layers.