Structure analysis of swollen rubber carbon black systems by contrast variation SANS

Transcription

1 Nippon Gomu Kyokaishi, No. 1, 010, pp Structure analysis of swollen rubber carbon black systems by contrast variation SANS M. Takenaka 1, S. Nishitsuji 1, S. Fujii 1, N. Amino, Y. Ishikawa, D. Yamaguchi 3, and S. Koizumi 3 1 Kyoto University, Kyoto , Japan The Yokohama-rubber o. Ltd.,Kanagawa , Japan 3 Advanced Science Research entre, Japan Atomic Energy Agency, Tokai, Ibaraki , Japan Selected from International Polymer Science and Technology, 38, No. 3, 011, reference NG 1/1/390; transl. serial no Translated by K. Halpin Introduction Elucidation of the structure of the filler aggregate and the filler-polymer interface, and the correlation of this structure with performance, is a key topic in tyre development and research aimed at obtaining higher performance from filled rubber systems [1]. We have been exploring the filler aggregate structure in filled rubber systems by determining the structure from the micron scale to the nanometre scale with a combination of neutron and X-ray scattering techniues [, 3]. Shinohara et al. have been following a similar approach in ultra-small angle X-ray scattering studies [4]. The results have revealed that the filler forms aggregates of 10 nm order, which then form a network of mass fractal structure, thus generating a hierarchical structure. It was expected that these findings would serve to clarify the relation between the structure and mechanical behaviour of the filled system. However, mechanical behaviour is not just affected by the structures mentioned; it is also affected by the more microscopic interfacial structure. The cohesion between rubber and filler in the interfacial zone greatly influences the mechanical properties of filled rubber systems. However, scattering in the bulk system cannot identify how the rubber adsorbs on the filler surface since the adsorption layer and the unadsorbed matrix consist of the same rubber, with the result that scattering contrast is nonexistent. To resolve this difficulty, we swelled a silica filled rubber system in solvents of different scattering length density and analysed scattering by the contrast variation small angle neutron techniue, as already reported [5, 6]. The results revealed the existence at the silica surface of a region of low degree of swelling associated with strong adsorption to the silica, and it was thus possible to obtain a uantitative estimate of the thickness of the polymer adsorption layer. The study reported here shows that contrast variation in small angle neutron scattering is an eually effective tool for the analysis of interfacial structure in carbon black (B) filled systems. Experimental Materials The study used styrene butadiene random copolymer (SBR: Nipol 150) as the rubber component of the B filled system. The molecular properties of the SBR are listed in Table 1. The rubber was compounded in the SBR/B/zinc oxide/stearic acid ratio 100/50/3/1 and mixed for 5 min in a 1.7L Banbury mixer to obtain a master batch. The mixer rotor speed was adjusted to give a rubber temperature of 150 at dumping. Accelerator and sulphur were then added to the master batch at 1.5 parts by weight each per 100 parts by weight of SBR, using a two-roll mixing mill. The B used was N339 black. N-tert-butyl--benzothiazyl sulphenamide (TBBS) was used as the accelerator. The uncured compound obtained was press vulcanised for 50 min at 160 to obtain 1 mm thick sheet. The samples thus prepared were soaked for 1 hours in a solvent of deuterated 011 Smithers Rapra Technology T/1

2 Table 1. haracterisation of SBR Polymers Mw Mw/Mn w PS, (%) a Vinyl content (%) b SBR 5.0x a Weight fraction of styrene content b Vinyl content in butadiene seuence toluene (d-tol) and toluene (h-tol) mixed at various volume fractions (d-tol/h-tol = 100/0, 70/30, 50/50, 30/70, 0/100) to produce euilibrium swelling and then loaded into a uartz cell with solvent and examined by small angle neutron scattering (SANS). chemical properties [10]. Hence, by using a mixture of deuterated solvent and non-deuterated solvent for swelling and examining the neutron scattering at different mixing ratios, we can clarify the dependence of scattering on the scattering length density of the solvent. Analysis of this dependence allows separation of the spatial distribution in contrast and spatial distribution in structure, facilitating structural analysis of the adsorption layer. The techniue is called contrast variation small angle neutron scattering, and the following paragraphs describe the actual method of analysis. Results and discussion SANS measurements The measurements were made in the focusing polarised neutron ultra-small angle scattering spectrometer (SANS- J-II) at the Japan Atomic Energy Research Institute JRR-3M facility [7]. Two sample-to-detector distances of.5 m and 10 m were used so as to broaden the accessible range of wavenumber. The beam wavelength was l = 0.65 nm, for which the observable range was 0.04<<0.7 nm -1. The observation times were 1800 s and 600 s, respectively, at.5 m and 10 m. All measurements were conducted at room temperature. The data obtained were processed by circular averaging, transmittance correction, and correction for blank cell scattering; the data were then reduced to absolute intensities using the scattering from aluminium. orrections were also made for incoherent scattering. Dependence of SANS profile on d-tol fraction Figure 1 shows the intensity I() of small angle scattering for the B filled system swollen with the d tol/h-tol solvent mixed in different proportions. The wavenumber () dependence of the scattering function is observed to vary when the scattering length density of the solvent is varied by adjustment of the d-tol to h-tol ratio. is defined by the following euation: = (4p/l)sin(/) (1) where is the scattering angle. The variation in dependence with solvent mixing ratio suggests a spatial inhomogeneity in degree of swelling in the rubber component, i.e. the degree of swelling is different in the zone around the B particles and in the matrix, implying the presence of an adsorption layer. ontrast variation small angle neutron scattering We attempted structural analysis of the various components by the contrast variation techniue, as developed by Endo et al. [8-10]. As noted above, the adsorbed rubber layer is undetectable from bulk scattering. However, if the filled rubber is swollen with solvent, the partial degree of swelling of the adsorption layer decreases when the rubber adsorbs strongly to the filler surface. Hence, scattering contrast develops owing to the difference in degree of swelling between the adsorption layer and the unadsorbed matrix of crosslinked rubber. Differences in contrast can then be used to determine the existence or otherwise of an adsorption layer from scattering experiments. However, the system has three components: rubber, B and solvent. The spatial distribution of contrast in scattering and spatial distribution of structure are therefore extremely difficult to separate. SANS is used to resolve this difficulty. By substituting hydrogen with deuterium in SANS, it is possible to alter the contrast in scattering the neutrons without significantly affecting Figure 1. Scattering profiles for SBR-B system swollen by various composition of d-tol/h-tol. omposition are indicated in the Figure T/ International Polymer Science and Technology, Vol. 38, No. 6, 011

3 alculation of partial scattering function in contrast variation SANS Since the fractions of accelerator and sulphur are very much smaller than those of the rubber component, the B or the solvent, we can treat the system as a threecomponent system of rubber-b-solvent. The scattering function I() of the system may hence be expressed by the euation [5, 6]: ( ) S PP ( ) + ( a P a S ) a a S ( ) S () I() = a P a S + a a S ( )S P ( ) The uantity a i in Euation () is the scattering length density of the component i (subscript i=p: SBR, i=: B, i=s: toluene). S PP () and S () are the respective scattering functions of SBR and B attendant on autocorrelation, while S P () is the partial scattering function in cross-correlation between SBR and B. The functions are defined as follows: S ij ( ) = 1 V δφ i ( r )δφ j r 1 ( ( )) exp i ( r r 1 ) d rd r 1 where V is the volume irradiated by the incident beam, df i (`r) is the variation in volume fraction of component i from the mean at location `r. Defining Da i as the difference in scattering length density (contrast) between i (P or ) and toluene, Da i = a i a S and Euation () can be represented by the matrix: I( ) = Δa ( p Δa p Δa Δa c ) S PP S P S () (3) (4) S PP S P = S 1 Δa p Δa p 3 Δa p 1 Δa 1 Δa Δa p Δa Δa 3 Δa 3 p Δa 3 Δa 1 I 1 I I 3 we may use the inverse matrix of the contrast matrix for the scattering matrix. However, since the experimental data contain errors, S ij cannot be determined with any precision by this method. To determine the partial scattering functions with good precision, therefore, the number of data points is increased and the matrix pseudoinverse found by singular value decomposition to determine S ij ; in other words, the partial scattering function is found from the product of the pseudoinverse and the scattering function vector [I 1 (), I (),...I n ()] determined from the n (n>3) experiments in which the scattering length density of toluene was varied. 1 Δa 1 Δa S PP Δa S P = p Δa p Δa Δa S n Δa n Δa 1 I 1 I I n () Figure plots the partial scattering functions obtained by singular value decomposition from the scattering functions in Figure 1. If an adsorption layer did not Table. Scattering length density of each componenent used in this study toluene d-toluene SBR B a i (cm - ) 9.41x x x x10 10 (6) (7) The scattering length densities of the components are set out in Table. If the proportion of d-tol to h-tol is varied and n scattering measurements are made under conditions of differing a S, the scattering function vector [I 1 (), I (),...I n ()] at wavenumber will be: I 1 1 Δa 1 Δa c S I Δa = p Δa p Δa Δa PP c S P I n () S n Δa n Δa c (5) In principle, S ij may be found by conducting an experiment with toluene mixtures of three different scattering length densities and solving the three-element simultaneous linear euations at each. Thus, putting Figure. Partial scattering function of SBR-B systems obtained from scattering profiles from Figure Smithers Rapra Technology T/3

4 exist and the solvent were uniformly distributed in the SBR phase, then: S P ( ) = φ p S ( ),S PP ( ) = φ p S ( ) where f p is the volume fraction of SBR in the SBR (plus toluene) phase. As the structure function of the B particles, S () is necessarily positive; hence, if an adsorption layer is absent, S P () becomes negative; at the same time, the functions -S P (), S (), and S PP () must have the same form on a double logarithmic plot of the partial scattering function versus wavenumber. As shown in Figure, S P () takes negative values in the experimental plots. However, as shown by the double logarithmic plot in Figure 3, S PP (), -S P () and S () show different -dependence in the small angle region, indicating that the solvent has not swollen the SBR phase uniformly. The difference in -dependence at small angles (8) presumably reflects the appearance of an effect due to an adsorption layer in the cross-correlation term. This is similar to the trend observed in silica filled systems in previous reports. Identification of adsorption layer by analysis of partial scattering functions To deduce information on the adsorption layer from the partial scattering functions, the same procedure was followed as for analysis of silica filled systems. Thus, we derived the partial scattering functions by considering a model in which an SBR adsorption layer surrounded the B particles (Figure 4). Supposing an adsorption layer of SBR (b region) of volume fraction f l is present in the B aggregate structure (a region) and surrounded by a matrix (g region) of volume fraction f m, the partial scattering functions may be represented as follows [5,6,1]: S ( ) = F α ( ) (9) S P ( ) = ( φ 1 φ m )F α+β ( )F α ( ) φ 1 F α ( ) (10) S PP ( ) = ( φ 1 φ m )F α+β ( ) φ 1 F α ( ) (11) Figure 3. Partial scattering function of SBR-B systems and their fitting results with model functions (solid lines) where F a () is the structural amplitude of the a region, i.e. the amplitude of the B particle aggregate structure, and F a+b () is the structural amplitude of the a and b regions combined. As a structural model of the B particle aggregate, a structure of radius of gyration R g,a and a smooth interface was used [13-15]. The relation S ()= S a ()= F a () may then be represented as: Figure 4. Schematic graph of the model of SBR-B system swollen by solvent T/4 International Polymer Science and Technology, Vol. 38, No. 6, 011

5 S α ( ) + B erf R g,α / 6 ( ) = A exp R g,α / 3 { ( )} 3 / 4 (1) where R g,a denotes the radius of gyration of the aggregate structure. A and B are A = KV (13) B = πas / V (14) and K, V and S are respectively the proportionality multiplier, the volume of the aggregate, and the surface area of the aggregate. The model used [13-15] for the whole structure combining regions a and b was a structure with a radius of gyration R g,l and smooth interface, in which the interfacial region has an interface thickness s. The scattering function of the structure S a+b ()= F a+b () may be represented: ( ) S α=β ( ) = exp R g,l / 3 +D { erf ( R g,l / 6) } 3 4 / where and D are respectively exp ( σ ) (15) = KV 1 (16) D = πs 1 / V l (17) and V l and S l are respectively the volume and surface area of the whole structure combining the a and b regions. The results of curve fitting with the above euations are shown by the solid lines in Figure 3. The results are a good representation of the partial scattering functions, demonstrating the validity of the models. Table 3 sets out the parameters as evaluated by fitting. Thus, from the radius of gyration R g,a of the a region, the size of the aggregate structure of B particles was found to be 41.4 nm. The radius of gyration R g,l of the whole region combining the a and b regions is clearly larger than the Table 3. haracteristic parameters yielded from fitting R g, a (nm) R g,l (nm) t I (nm) f m f l A B D s (nm) 6.47x x x x radius of gyration R g,a of the a region comprising the B phase, and the volume fraction f l of polymer in the b region is greater than the volume fraction f m of polymer in the g region. It was hence clear that an adsorption layer of rubber is present around the B particles. An adsorption layer thickness of 9.7 nm was obtained from the difference in R g,a and R g,l. Furthermore, the interfacial thickness of s=5.7 nm revealed the existence of a polymer concentration gradient between the b region and g region. omparing the above results with the silica filled system previously studied, it is evident that the thickness of 5.3 nm for the adsorption layer in the silica filled system is less than the thickness in the B filled system; and the silica aggregate structure (radius of gyration 3.7 nm) is smaller than the B aggregate structure. Because the respective structures will inevitably differ greatly with the silane coupling agent and species of B, it is impossible to generalise, but it would appear that the B system has a thicker adsorption layer and larger aggregate structure. onclusions Using contrast variation small angle neutron scattering, the study sought to establish uantitatively the presence of a rubber adsorption layer at the carbon black aggregate surface in carbon black filled systems by examining carbon black filled rubber swelled with solvents of different scattering length density obtained by mixing toluene and deuterated toluene. An adsorption layer of around 10 nm was shown to exist at the filler surface. It also proved possible to shed light on the B particle aggregate structure. Future studies will examine differences in the adsorption layer among different carbon blacks. References 1. Nippon Gomu Kyokai Ed., "Shimban Gomu Gijutsu no Kiso", Nippon Gomu Kyokai, Tokyo (005).. Koga, T., Takenaka, M., Aizawa, K., Nakamura, M., Hashimoto, T., Langmuir, 1, (005). 3. Koga, T., Hashimoto, T., Takenaka, M., Aizawa, K., Amino, N., Nakamura, M., Yamaguchi, D., Koizumi, S., Macromolecules, 41, 453 (008). 4. Shinohara, Y., Kishimoto, H., Inoue, K., Suzuki, Y., Takeuchi, A., Uesugi, K., Yagi, N., Muraoka, K., Mizoguchi, T., Amemiya, Y., J. Appl.ryst., 40, s397 (007). 5. Takenaka, M, Nishitsuji, S, Amino, N, Ishikawa, Y., Yamaguchi, D., Koizumi, S., Macromolecules, 4, 308 (009). 011 Smithers Rapra Technology T/5

6 6. Takenaka, M.,Nishitsuji, S., Amino, N., Ishikawa, Y., Yamaguchi, D., Koizumi, S., Hashimoto, T., Nippon Gomu Kyokaishi, 81, 334 (008). 7. Koizumi, S., Iwase, H., Suzuki, J., Oku, T., Motokawa, R., Sasao, H., Tanaka, H., Yamaguchi, D., Shimizu, H. M., Hashimoto, T., J. Appl. ry St., 40, s474 (007). 8. Endo, H., Allgaier, J., Gompper, G., Jakobs, B., Monkenbusch, M., Richter, D., Sottmann, T., Strey, R., Phys. Rev. Lett.. 85, 10 (000). 9. Endo, H., Mihailescu, M., Monkenbusch, M., Allgaier, J., Gompper, G., Richter, D., Jakobs, B., Sottmann, T., Strey, R., Grillo, I., J. hem. Phys.,115, 580 (000). 10. Endo, H., Schwahn, D., olfen, H., J. hem. Phys., 10, 9410 (004). 11. Miyazaki, S., Endo, H., Karino, T., Haraguchi, K., Shibayama, M., Macromolecules, 40, 487 (007). 1. Koubunshi Gakkai Ed., "Shin Koubunshi Jikkengaku 6 Koubunshi no Kouzou ()", Kyoritsu- Shuppan, Tokyo (1997). 13. Beaucage, G., Schaefer, D.W., J. Non-rystal. Solids, 17, 797 (1994). 14. Beaugcage, G., J. Appl.ryst., 9, 134 (1996). 15. Beaugcage, G., J. Appl.ryst., 8, 717 (1995). T/6 International Polymer Science and Technology, Vol. 38, No. 6, 011