Soft Matter View Article Online
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
PAPER
Cite this: DOI: 10.1039/c7sm01262d
View Journal
Quasielastic neutron scattering study of microscopic dynamics in polybutadiene reinforced with an unsaturated carboxylate† Ryo Mashita, ‡a Rintaro Inoue,‡b Taiki Tominaga,c Kaoru Shibata,d Hiroyuki Kishimotoa and Toshiji Kanaya *e We studied the dynamics of zinc diacrylate (ZDA) reinforced polybutadiene rubber (BR) (ZDA/BR) using the quasielastic neutron scattering technique to determine the effect of concentration of ZDA on polymer dynamics. First, we evaluated the temperature dependence of mean square displacements (hu2i) for ZDA/BR with different ZDA volume fractions. hu2i increased with temperature below 170 K, and we observed no significant ZDA volume fraction dependence. However, it increased more steeply above 170 K, and the value of hu2i was smaller for the samples with increasing ZDA fraction. To elucidate the origin of the decrease in hu2i with increasing ZDA content, dynamic scattering laws (S(Q,o)) were analyzed. An increase in the elastic component, an increase in the mean relaxation time, and a broadening of distribution of relaxation time were observed with the increasing volume fraction of ZDA. In addition, the ZDA volume fraction dependence of the elastic component roughly corresponded to that of elastic
Received 27th June 2017, Accepted 19th September 2017 DOI: 10.1039/c7sm01262d
modulus, indicating that the elastic component is related to its mechanical strength. Referring to the previously reported static structure of the present ZDA/BR system, a model for the heterogeneous BR dynamics was proposed. This model assumes the coexistence of immobile, mobile, and interfacial constrained mobile regions. It was found to be appropriate for the explanation of the observed dynamics. We proposed that a network-like structure of the BR having a high crosslinking density around ZDA
rsc.li/soft-matter-journal
aggregates is mainly responsible for the high elastic modulus of ZDA/BR.
1 Introduction Rubber materials utilized in various industrial products, such as automobile tires, airplanes, and base isolation devices, are indispensable in industry and our daily life. Fillers or functional agents are used to enhance the properties of pure rubber, such as mechanical strength, abrasion resistance, and deterioration resistance. The mechanical properties of rubber materials can be enhanced by the addition of crosslinking agents or reinforcing fillers, such as carbon black (CB), silica, or a
SUMITOMO Rubber Industries, LTD, 1-1, 2-chome, Tsutsui-cho, Chuo-ku, Kobe 651-0071, Japan b Research Reactor Institute, Kyoto University, 2, Asashironishi, Kumatori-cho, Sennan-gun, Osaka 590-0494, Japan c Neutron Science and Technology Center, Comprehensive Research Organization for Science and Society (CROSS), Tokai, Ibaraki 319-1106, Japan d Neutron Science Section, Material and Life Science Division, J-PARC Center, JAEA, Tokai, Ibaraki 319-1195, Japan e Material and Life Science Division, J-PARC Center, KEK, 203-1 Shirakata, Tokai, Ibaraki 319-1106, Japan. E-mail: [email protected] † Electronic supplementary information (ESI) available. See DOI: 10.1039/ c7sm01262d ‡ These authors contributed equally to this work.
This journal is © The Royal Society of Chemistry 2017
clay.1 The mechanical strength of polybutadiene rubber (BR) reinforced with zinc diacrylate (ZDA) (ZDA/BR) is known to be considerably higher than that of the typical carbon black or silica reinforced rubber.2–8 The mechanism underlying the high mechanical strength attained when using ZDA remains unknown despite extensive studies. It has been conjectured that the failure to elucidate the reinforcement mechanism in ZDA/BR originates mainly from the lack of detailed information about both its static and dynamic structures. To overcome this shortcoming, first, we studied the structure of ZDA/BR through the complementary use of small-angle X-ray scattering (SAXS) and small-angle neutron scattering (SANS) techniques.8,9 In addition to the hierarchical structures of the ZDA aggregates and BR matrix, a hierarchical BR structure possessing a higher crosslinking density (HC-BR) than that of the BR matrix was observed for ZDA/BR. Furthermore, it is expected that such HC-BR region is mainly segregated on the surface of ZDA aggregates through the contrast variation SANS (CV-SANS) method. It is reported that the dynamics of such a spatially or geometrically confined polymer is different from that of the bulk system.10 Namely, the dynamics of such a confined polymer is severely affected by such a spatially or geometrically confined
Soft Matter
View Article Online
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
Paper
Soft Matter
polymer contribution. As for the BR system, Roh et al. investigated the dynamics of the CB/BR system at one CB concentration through quasielastic neutron scattering (QENS), which can probe dynamics at a microscopic spatial scale.11 They reported the slowing down of the segmental dynamics of matrix BR due to physical/chemical absorption of BR on the surface of CB. It is reported that the mechanical strength of both CB/BR and ZDA/BR is strongly dependent on the concentration of filler or reinforcing agents.8 Hence, the studies focusing on only one filler concentration are not enough for ruling out the dominant factor for both mechanical strength and attributing dynamics. At present, the concentration dependency of reinforcing agents has not been well studied due to the various controlling parameters. Hence, tuning the concentration of reinforcing agents, which is dependent on various fields of application of the BR product, is vital for industrial application. In this manuscript, we try to study the effect of concentration of reinforcing material (ZDA) on the dynamics of ZDA/BR, which exhibits high mechanical strength, through QENS. The mechanisms of various polymer dynamics have been successfully studied with this method.11–19 Based on the previous CB/BR system, the relevance between mechanical strength and dynamics is also discussed.
2 Experimental 2.1
Materials and sample preparation
The details of the materials used in the present study are summarized in Table 1. We used polybutadiene (BR; BR730, JSR Co., Ltd) with a weight-average molecular weight (Mw) of 6.72 105 and a molecular weight distribution index (Mw/Mn) of 2.39. Mn is the number-average molecular weight of the matrix rubber. The percentage of cis, trans, and vinyl in the present BR is 97, 2, and 1%, respectively. Zinc diacrylate (ZDA; Sanceler SR, Sanshin Chemical Industry Co., Ltd) was used as a cross-linker and dicumyl peroxide (DCP; Percmyl D, Nichiyu Co., Ltd) was used as an initiator for the crosslinking reaction. The chemical formula of the ZDA monomer is shown in Fig. 1. BR, ZDA, and DCP were placed into a 6-inch two-roll mill simultaneously, and mixed for 5 minutes at room temperature. After mixing at room temperature, the mixture of BR, ZDA, and DCP was placed into a mold and heated at 443 K for 20 minutes to increase the rate of the chemical reaction. The ZDA monomer possesses a carbon–carbon double bond (CQC) at both of its ends. Firstly, the ZDA monomer is grafted to BR by the reaction between one CQC and BR, and then ZDA aggregates with the size of several nm are subsequently formed by chain
Table 1
Fig. 1
Chemical formula of the zinc diacrylate (ZDA) monomer.
transfer reaction between ZDA monomers. Secondly, the crosslinking of BR proceeds through further chemical reaction between the other CQC end of ZDA and BR. It should be noted that ZDA(0) in Table 1 does not include ZDA, but is crosslinked by DCP. DCP reacts with the CQC of BR, and the generated carbon radicals react with another CQC of BR. Then, intermolecular carbon–carbon bonds (C–C) are formed. ZDA(0) is not composed of neat BR but self-crosslinked BR. Fig. 2 shows the relationship between the elastic modulus obtained by dynamic viscoelasticity measurements and the volume fraction of ZDA at 303 K. The elastic modulus increases dramatically above the volume fraction of 0.05. For clarity, we also added a solid line, which only holds for low ZDA contents in Fig. 2. We then performed QENS measurements at the ZDA volume fractions of 0, 0.03, 0.05, 0.1 and 0.14 to find out the origin of the high mechanical strength of ZDA/BR. The glass transition temperatures (Tg) were measured using differential scanning calorimetry (DSC). DSC curves are shown in ESI,† Fig. SI. Two glass transition temperatures were observed at B170 K and at B360 K. The first Tg (B170 K) was identical to that of neat BR and constant regardless of the volume fraction of ZDA (refer Table 1). However, the second Tg was observed at B360 K in the samples except for ZDA(0). This suggests that the samples crosslinked by ZDA have two regions with Tg of 170 K and 360 K, respectively. The scattering contribution of each component was estimated based on the scattering cross section. For ZDA(14), which contained the highest volume fraction of ZDA among the investigated samples, the proportions of the total scattering cross-sections of BR, ZDA, and DCP were estimated to be 94.7 : 5.0 : 0.3. The large scattering contribution of BR is mainly due to the large incoherent scattering cross section of hydrogen atoms. Thus, the observable scattering intensity for ZDA/BR in this QENS study would be dominated by the BR.
2.2
Measurements
QENS measurements were carried out using an inverted geometry time-of-flight spectrometer (BL02 DNA)20 installed at the
Sample composition details and Tg
ZDA(0)
ZDA(3)
ZDA(5)
ZDA(10)
ZDA(14)
Polybutadiene (vol%) Zinc diacrylate (vol%) Dicumyl peroxide (vol%)
99.5 0 0.5
96.8 2.7 0.5
94.2 5.3 0.5
89.5 10.1 0.4
85.2 14.4 0.4
Glass transition temperature1st Tg (K) 2nd Tg (K)
168.6 1 N/A
170.0 1 353.4 1
170.4 1 359.9 1
170.0 1 360.4 1
170.6 1 362.2 1
Soft Matter
This journal is © The Royal Society of Chemistry 2017
View Article Online
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
Soft Matter
Paper
Fig. 2 Relationship between the elastic modulus and volume fractions of zinc diacrylate (ZDA).
Materials and Life Science Experimental Facility (MLF) in J-PARC, Tokai, Japan. The magnitude of the scattering vector Q (Q = 4p sin y/lf, where 2y and lf = 6.26 Å are the scattering angle and the wavelength of the analyzed neutron, respectively) ranged from 0.12 to 1.78 Å1. To obtain the first Tg of the present ZDA/BR samples, the temperature was varied from 10 K to 330 K at a heating rate of 1 K per min for elastic scattering measurements, as shown later in Fig. 4. On the other hand, dynamic scattering laws S(Q,o), as shown later in Fig. 3, 5, and 6, were measured under isothermal conditions. It should be noted that this temperature range was below the second Tg (B360 K). Each sample with a thickness of 200 20 mm was placed in a double-cylindrical aluminum cell (outer diameter: 14 mm, inner diameter: 13 mm, height: 45 mm) under a helium atmosphere. The dynamic scattering laws S(Q,o) of each sample at 10 K were used to determine the resolution function of the corresponding sample. The energy resolution (dE) defined as the full width at half maximum (FWHM) was estimated to be 4.0 0.2 meV. As seen in ESI,† Fig. SII, the resolution functions were constant regardless of the samples. DSC measurements were carried out using a differential scanning calorimeter (Q200, TA Instruments) in the temperature range from 123 K to 423 K with a heating rate of 10 K per min. Dynamic viscoelasticity measurements were carried out using a dynamic mechanical analyzer (RSA-G2, TA Instruments), at 303 K and 10 Hz.
3 Results and discussion Fig. 3(a) and (b) shows the dynamic scattering laws S(Q,o) with a mean Q of 1.48 Å1 for ZDA(0) and ZDA(14), respectively. The graphs are plotted for several temperatures above the first Tg but below the second Tg, except for 10 K. S(Q,o) apparently consists of elastic and quasielastic components above the first Tg, suggesting that both of the samples ZDA(0) and ZDA(14) have an immobile component. In the case of ZDA(14), this must have originated from the glassy region below the second Tg. On the other hand, ZDA(0) also exhibited the elastic component. This must be due to the structure of BR cross-linked by DCP.
This journal is © The Royal Society of Chemistry 2017
Fig. 3 Dynamic scattering laws S(Q,o) for (a) ZDA(0) at 10 K ( ), 200 K ( ), 250 K ( ), 280 K ( ), 310 K ( ) and 330 K (J) and (b) ZDA(14) at 10 K ( ), 200 K ( ), 250 K ( ), 280 K ( ), 310 K ( ) and 330 K (,), obtained by summing the spectra from Q = 1.38 to 1.58 Å1. The mean value of Q was 1.48 Å1.
Comparison of S(Q,o) for ZDA(0) and ZDA(14) shows that the elastic component is larger and the quasielastic component is narrower in ZDA(14) than ZDA(0). This implies the suppression of motion in BR by the addition of ZDA in the present energy region. To investigate the effect of ZDA on the observable dynamics in more detail, we evaluated the mean-square displacement (hu2i) from the Q2 dependence of the elastic scattering intensity (Iel,T(Q)), assuming the Gaussian approximation.21,22 This is shown by eqn (1). Iel;T ðQÞ ¼ exp u2 Q2 Iel;T¼10K ðQÞ
(1)
Iel,T=10K(Q) is the elastic scattering intensity at 10 K. The left side of eqn (1) approximates the incoherent elastic scattering intensity under the assumption that the structure factor is independent of temperature. The results at 300 K, which is above the first Tg, calculated by eqn (1) are shown in Fig. 4(a). The experimental data conform to calculations of eqn (1), within experimental error. Fig. 4(b) shows the temperature dependence of hu2i on different volume fractions of ZDA for ZDA/BR. In all samples, hu2i increases linearly with temperature in the range of 10 K to 110 K. This
Soft Matter
View Article Online
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
Paper
Fig. 4 (a) Fitting results of Iel,T(Q)/Iel,10K(Q) for ZDA(0) ( ), ZDA(5) ( ) and ZDA(14) ( ) at 300 K using eqn (1) (solid lines). (b) The temperature dependence of hu2i obtained by eqn (1) for ZDA(0) ( ), ZDA(3) (}), ZDA(5) ( ), ZDA(10) ( ) and ZDA(14) ( ). The arrow indicates the first Tg. (c) ZDA volume fraction dependence of the inflexion point of hu2i ( ) and Tg obtained by DSC ( ).
suggests that the motion observed below 110 K is attributed to harmonic oscillation. Changes in the slope of hu2i are observed at 110 K and around the first Tg (B170 K), for all the samples. Referring to the previous works,23,24 the first inflexion point (B110 K) corresponds to the onset of the fast b-process. The second inflexion point (B170 K) clearly indicates the onset of a new motion above the first Tg, which is probably the so-called a-process. Tyagi et al. reported that the relaxation time at around Q = 1.0 Å1 by QENS nicely coincided with those from other
Soft Matter
Soft Matter
dynamical methods.25 Namely, the hierarchical dynamics holds for the a-process. In our previous work,26 we also revealed that excess hu2i, which is the difference of hu2i between two different energy resolutions, exhibits an inversely linear relationship with ln Z (Z: viscosity) by comparison of the results obtained from different energy resolution QENS measurements under the idea of free volume theory. Furthermore, the universal correlation of the structural relaxation time (as well as the viscosity) and the rattling amplitude from glassy to low-viscosity states is fulfilled from MD simulation and the experimental data of several glass-formers and polymers.27 Fig. 4(c) shows the dependence of the second inflexion point (B170 K) of hu2i on the volume fraction of ZDA, which corresponds to the first Tg measured by DSC. It should be noted that the onset temperature is independent of the ZDA volume fraction. In addition, the value of hu2i decreased clearly with the increasing volume fraction of ZDA above the first Tg, especially above B250 K, as shown in Fig. 4(b). In order to interpret the results, we need to recall that the ZDA/BR samples have a tworegion structure defined by the first Tg (B170 K) and the second Tg (B360 K). The fact that the first Tg is independent of the volume fraction of ZDA means that the region between the first and second Tg includes the mobile part, which has the same mobility as neat BR. As for the decrease in hu2i with increasing ZDA volume fraction, there are two plausible reasons. One is the increase of the second (immobile) region, and the other is the increase of the less mobile part in the first (mobile) region. The problem will be discussed based on the analysis of S(Q,o). However, the dynamical heterogeneity increases with the increasing volume fraction of ZDA. To understand the physical origin of the independence of the first Tg and the decrease of hu2i with the increasing volume fraction of ZDA, we analyzed S(Q,o), as shown in Fig. 3. A clear broadening of the central component of S(Q,o) was observed above the first Tg for both ZDA(0) and ZDA(14). In addition, the broadening of S(Q,o) was observed for other samples (data not shown). Based on a previous QENS study on neat BR with similar energy resolution,26 it is expected that the broadening of S(Q,o) must reflect the detection of the a-process. Fig. 5 shows S(Q,o) normalized by the peak intensity of the elastic scattering from different ZDA volume fractions at 310 K in the range Q = 1.48 Å1. Consistent with the decrease of hu2i, with the increase in the ZDA volume fraction, a narrowing of S(Q,o) is observed. To extract the parameters describing dynamical information from the S(Q,o), we performed curve fitting to the observed S(Q,o) at different ZDA volume fractions using an appropriate model function. Richter et al. reported that the intermediate scattering function corresponding to the a-process of neat BR was well described by the Kohlrausch– Williams–Watts (KWW) equation.28,29 Hence, we adopted the KWW function to describe the observed S(Q,o), and the following model function was proposed. n j ko S ðQ; oÞ ¼ A ð1 ecf ÞF exp ðt=tÞb þ ecfdðoÞ RðQ; oÞ
(2)
where A, ecf, t, and b are the amplitude of the relaxation function, the elastic component fraction, the relaxation time,
This journal is © The Royal Society of Chemistry 2017
View Article Online
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
Soft Matter
Paper
Fig. 5 Dynamic scattering laws S(Q,o) for ZDA(0) ( ), ZDA(3) (}), ZDA(5) ( ), ZDA(10) ( ) and ZDA(14) ( ) normalized at the peak top at 310 K in the range Q = 1.48 Å1. The dashed line indicates the instrumental resolution function.
and the distribution of the relaxation function, respectively. It must be kept in mind that we have introduced an elastic component in eqn (2) even at temperatures above the first Tg, because the model did not work well without the elastic component. In addition, the samples have a glassy region at 310 K below the second Tg (B360 K). It should be noted that the KWW function is defined in the time domain, necessitating that we adopt the Fourier-transformed KWW function ðF ðKWWÞÞ in the frequency region.30 R(Q,o) indicates the instrumental resolution function. Fig. 6(a) and (b) shows the results of the fits to S(Q,o) for ZDA(0) and ZDA(14) with eqn (2) at Q = 1.48 Å1 and 310 K, respectively. The S(Q,o)s from both samples are fitted fairly well by eqn (2), supporting the application of an appropriate model function. In the case of ZDA(0), the elastic component was observed although the second Tg was not detected in DSC measurements (Table 1). The elastic component must have originated from the crosslinked structure by DCP. S(Q,o)s are also modelled fairly well by eqn (2) in other Q regions and at other temperatures (see ESI,† Fig. SIII). The dependence of ecf on the volume fraction of ZDA, the average relaxation time hti and b at 310 K in the range Q = 1.48 Å1 are shown in Fig. 7. hti was calculated using eqn (3):30,31 t 1 G (3) hti ¼ b b where G is a gamma function. Summarizing the obtained parameters from the curve fits, the increase in ecf, the decrease of b, and the increase of hti were observed with the increase in the volume fraction of ZDA. The same trends were observed for other Q ranges (see ESI,† Fig. SIV). The elastic fraction comes from the glassy region below the second Tg and/or the confined motion in the mobile region above the first Tg. In the case of ZDA(0), which does not include ZDA, a small amount of the elastic component (B0.02) is observed as seen in Fig. 7(a). This elastic component must be due to the crosslinked structure of BR formed by the
This journal is © The Royal Society of Chemistry 2017
Fig. 6 The fitting results of S(Q,o) for (a) ZDA(0) and (b) ZDA(14) using eqn (2) for Q = 1.48 Å1 at 310 K. Solid lines are the results of calculation with eqn (2). Thick and thin dashed lines are the calculated results obtained by the KWW and elastic components, respectively, in eqn (2).
addition of DCP. The elastic fraction increases with the ZDA volume fraction, suggesting that the elastic fraction mainly originates from the immobile region mediated by the cross links formed by the addition of ZDA. In addition, a deviation from the linear dependence of ecf above the volume fraction of ZDA of 0.05 was seen. Qualitatively, this tendency corresponds to that of elastic modulus as seen in Fig. 2, indicating that the high ecf value observed at high ZDA content is related to high mechanical strength. This will be discussed in a later paragraph based on our previous SANS and SAXS results.8,9 Q dependence of ecf and the results of hypothetical fitting analyses are shown in ESI,† Fig. SV and Table SI, respectively.32–34 As seen in Fig. 7(b), the average relaxation time increases with the volume fraction of ZDA. At the same time, the parameter b, which represents the width of the distribution of relaxation time, decreases with increasing ZDA volume fraction, implying the broadening of the distribution of relaxation time. In other words, the samples become dynamically heterogeneous with the increase of the ZDA volume fraction. In the present analysis, we focused on the dynamics in the mobile region below the second Tg, so that the increase in the dynamic heterogeneity is attributed to the region, but not related to the coexistence of the mobile and the immobile regions.
Soft Matter
View Article Online
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
Paper
Fig. 7 ZDA volume fraction dependence of (a) ecf ( ), (b) hti ( ) and b ( ) at 310 K in the range Q = 1.48 Å1.
To construct a plausible physical picture that could explain the evaluated parameters without inconsistency, we referred to the structural information revealed in our previous work on SAXS and SANS.8,9 From small-angle scattering studies, the coexistence of BR regions having low and high crosslinking densities was revealed. Through DSC measurements a second Tg was also observed at B360 K (Table 1). It should be noted that the highest temperature in the present QENS measurements was lower than this second Tg, suggesting that the first Tg corresponds to the Tg of the low crosslinking density region and the second Tg corresponds to the highly crosslinked region. In the above discussion on S(Q,o), we used the mobile region and the immobile region for the low crosslinking density region and the highly crosslinked region, respectively. Based on this assumption, we considered that the mobile region originates from the low crosslinking density region (LC-BR) and the immobile region originates mainly from the high crosslinking density region (HC-BR), which is still in a glassy state in the present temperature range. As an increase in the elastic fraction was observed with increasing ZDA volume fraction, a physical picture that assumes the coexistence of mobile and immobile regions in the ZDA/BR is seemingly applicable to the observed dynamics. In addition, the ZDA volume fraction dependence of the second Tg obtained by DSC qualitatively reproduced that of ecf (see ESI,† Fig. SVI). If the mobile and immobile regions are dynamically homogeneous for the ZDA/BR, the evaluated relaxation times or distributions of relaxation times must be constant, regardless of the volume fraction of ZDA. However, a
Soft Matter
Soft Matter
slight increase in the relaxation time or broadening of the distribution of relaxation time with the increase in the volume fraction of ZDA was observed (see Fig. 7(b)). Thus, a contributing factor other than the coexistence of mobile and immobile components must be present in the ZDA/BR system. It is known that a polymer near a substrate or an impenetrable boundary exhibits anomalous physical properties.35,36 Hence, it is expected that the dynamics of BR near the HC-BR regions would be strongly perturbed by the HC-BR region in the glassy state, resulting in slower dynamics than that of neat BR. With the increase in the volume fraction of ZDA, the contributions from such interfacial regions near HC-BR would increase. Thus, due to the gradient of dynamics mediated by the ZDA aggregates, more heterogeneous and slower dynamics are observed for ZDA/BR than those of the neat BR. A schematic representation of the ZDA/BR system obtained by the results of our previous work8,9 and this QENS study is shown in Fig. 8. Finally, we compared the present results to previous findings for other rubber systems to find out the origin of the high mechanical strength of ZDA/BR. Roh et al.11 studied the dynamics of the BR reinforced with carbon black (CB) (CB/BR) through neutron back-scattering, and observed both the slowing and broadening of the a-process for CB/BR compared to neat BR. Qualitatively, both our group and Roh’s group observed similar trends, though a drastic difference between the moduli was seen for the two systems. The moduli of ZDA/BR were several times higher than those of CB/BR.8 In both cases, the formation of network-like structures is believed to be responsible for the improved mechanical strength vital for practical application. Accordingly, it is assumed that the difference in the mechanism of interaction between the BR and the reinforcement agents must be related to the difference in mechanical strength. In the CB/BR system, it was considered that a network-like structure of physically adsorbed BR on the surface of CB was related to its mechanical strength.1 In contrast, the formation of covalent bonds between the CQC double bonds in the ZDA aggregates and the BR must effectively contribute to their strong interactions. In addition, with the increasing volume fraction of ZDA, such tightly constrained BR regions must also grow, as described above. Then, the covalently constrained BR on the surface of ZDA (HC-BR) would be expected to act as a kind of cross-linker in ZDA/BR.
Fig. 8 Schematic representation of the ZDA/BR system.
This journal is © The Royal Society of Chemistry 2017
View Article Online
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
Soft Matter
This leads to the formation of a tougher network-like structure than that in CB/BR, in which the physically absorbed BR alone acts as a cross-linker. Formation of such a network-like structure is also supported by the result of ecf as mentioned above. Thus, a more rigid network-like structure is developed in ZDA/BR compared to CB/BR, imparting a higher mechanical strength for the former at a given volume fraction of the reinforcement agent. Namely, it was clearly found that the deviation from the linear dependence of ecf implied the formation of a rigid network-like structure of the immobile component of BR.
4 Conclusions We studied the dynamics of polybutadiene rubber (BR) reinforced with zinc diacrylate (ZDA) (ZDA/BR) by the quasielastic neutron scattering (QENS) technique. The objective of this study was to determine the effect of concentration of reinforcing material on polymer dynamics. The mean square displacements (hu2i) of ZDA/BR for different ZDA volume fractions were evaluated. At temperatures below 170 K, no significant ZDA volume fraction dependence was observed for hu2i. On the other hand, hu2i increased more steeply above 170 K and a clear reduction of hu2i with the increasing volume fraction of ZDA was observed. To understand the origin of the decrease in hu2i with the increasing volume fraction of ZDA, dynamic scattering laws (S(Q,o)) were analyzed with a model function. We assumed the addition of an elastic component and a Fourier-transformed KWW function in the frequency region. With this model function, the observed S(Q,o)s for different ZDA volume fractions were fitted well. From the results of the curve fittings, an increase in the elastic fraction (ecf), an increase of the mean relaxation time (hti), and a broadening of the distribution of relaxation time (b) were observed with the increase in the volume fraction of ZDA. In addition, a deviation from the linear dependence of ecf above the volume fraction of ZDA of 0.05 was seen. This behavior of ecf corresponds to that of elastic modulus, indicating that the high ecf value observed at high ZDA content is related to high mechanical strength. Based on a previously reported static structure, the idea of heterogeneous dynamics, which assumes the existence of immobile, mobile, and interfacial constrained mobile regions, was proposed. This is a consistent model for understanding the observed dynamics. By comparing the reported results for CB/BR with those for ZDA/BR in this study, it was postulated that the difference in the interaction between the BR and the reinforcement agents must be linked to the huge difference in mechanical strength. In the case of ZDA/BR, the existence of chemical bonds between the BR and ZDA contributed to the effective, strong interaction between the BR and the reinforcement agents. Thus, a tougher network-like structure is formed in ZDA/BR than that in CB/BR, in which physical adsorption played the main role. Formation of such a network-like structure was also clearly supported by the resulting dynamics as a function of ZDA concentration. Consequently, a higher mechanical strength was realized for ZDA/BR as compared to that of CB/BR at a given volume fraction of reinforcement agents.
This journal is © The Royal Society of Chemistry 2017
Paper
Conflicts of interest There are no conflicts to declare.
Acknowledgements The QENS measurements on the BL02 DNA instrument installed at J-PARC were performed with the approval of the Materials and Life Science Experimental Facility (MLF) of J-PARC (Proposal No. 2014A0014). This work was supported by ‘‘Photon and Quantum Basic Research Coordinated Development Program’’ from the Ministry of Education, Culture, Sports, Science and Technology, Japan.
References 1 Y. Fukahori, Rubber Chem. Technol., 2007, 80, 701–725. 2 B. Chu, D. Q. Wu, R. D. Lundberg and W. J. MacKnight, Macromolecules, 1993, 26, 994–999. 3 Y. Tsujita, M. Yasuda, M. Takei, T. Kinoshita, A. Takizawa and H. Yoshimizu, Macromolecules, 2001, 34, 2220–2224. 4 M. Al-Hussein, B. G. G. Lohmeijer, U. S. Schubert and W. H. de Jeu, Macromolecules, 2003, 36, 9281–9284. 5 C. N. Zhou, D. C. Chan and I. K. Winey, Macromolecules, 2008, 41, 6134–6140. 6 E. M. Seitz, D. C. Chan, L. K. Opper, W. T. Baughman, B. K. Wagener and I. K. Winey, J. Am. Chem. Soc., 2010, 132, 8165–8174. 7 N. Nagata, T. Sato, T. Fujii and Y. Saito, J. Appl. Polym. Sci., 1994, 53, 103–120. 8 R. Mashita, H. Kishimoto, R. Inoue and T. Kanaya, Polym. J., 2016, 48, 239–245. 9 R. Mashita, H. Kishimoto, R. Inoue and T. Kanaya, Polym. J., 2013, 45, 57–63. 10 G. B. McKenna, Eur. Phys. J.: Spec. Top., 2010, 189, 285–302. 11 J. H. Roh, M. Tyagi, T. E. Hogan and C. M. Roland, Macromolecules, 2013, 46, 6667–6669. 12 J. H. Roh, M. Tyagi, T. E. Hogan and C. M. Roland, J. Chem. Phys., 2013, 139, 134905. 13 V. Arrighi, J. S. Higgins, A. N. Burgess and G. Floudas, Polymer, 1998, 39, 6369–6376. 14 S. Gagliardi, V. Arrighi, R. Ferguson and M. T. F. Telling, Physica B, 2001, 301, 110–114. 15 C. Karlsson, A. S. Best, J. Swenson, W. S. Howells and L. Borjesson, J. Chem. Phys., 2003, 118, 4206. 16 K. Chrissopoulou, S. H. Anastasiadis, E. P. Giannelis and B. Frick, J. Chem. Phys., 2007, 127, 144910. 17 S. Fotiadou, K. Chrissopoulou, B. Frick and S. H. Anastasiadis, J. Polym. Sci., Part B: Polym. Phys., 2010, 48, 1658–1667. 18 P. Akcora, S. K. Kumar, V. G. Sakai, Y. Li, B. C. Benicewicz and L. S. Schadler, Macromolecules, 2010, 43, 8275–8281. 19 K. Sinha, W. Wang, K. I. Winey and J. K. Maranas, Macromolecules, 2012, 45, 4354–4362. 20 K. Shibata, N. Takahashi, Y. Kawakita, M. Matsuura, T. Yamada, T. Tominaga, W. Kambara, M. Kobayashi,
Soft Matter
View Article Online
Paper
21 22 23
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
24 25 26
27 28
Y. Inamura, T. Nakatani, K. Nakajima and M. Arai, JPS Conf. Proc., 2015, 8, 036022. R. Inoue, T. Kanaya, K. Nishida, I. Tsukushi and K. Shibata, Phys. Rev. Lett., 2005, 95, 05102. T. Kanaya and K. Kaji, Adv. Polym. Sci., 2001, 154, 87–141. T. Kanaya, T. Kawaguchi and K. Kaji, J. Chem. Phys., 1993, 98, 8262–8270. J. Colmenero and A. Arbe, Phys. Rev. B: Condens. Matter Mater. Phys., 1998, 57, 13508–13513. M. Tyagi, A. Alegria and J. Colmenero, J. Chem. Phys., 2005, 122, 244909–1-9. T. Kanaya, T. Tsukushi, K. Kaji, J. Bartos and J. Kristiak, Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top., 1999, 60, 1906–1912. L. Larini, A. Ottochian, C. De Michele and D. Leporini, Nat. Phys., 2008, 4, 42–45. D. Richter, M. Monkenbusch, A. Arbe and J. Colmenero, J. Non-Cryst. Solids, 2001, 287, 286–296.
Soft Matter
Soft Matter
29 D. Richter, B. Frick and B. Farago, Phys. Rev. Lett., 1988, 61, 2465–2468. 30 F. Alvarez, A. Alegria and J. Colmenero, Phys. Rev. B: Condens. Matter Mater. Phys., 1991, 44, 7306–7312. 31 T. Kanaya, R. Inoue, M. Saito, M. Seto and Y. Yoda, J. Chem. Phys., 2014, 140, 144906. 32 F. Volino and A. J. Dianoux, Mol. Phys., 1980, 41, 271–279. 33 S. Fujiwara, K. Araki, T. Matsuo, H. Yagi, T. Yamada, K. Shibata and H. Mochiauki, PLoS One, 2016, 11, e0151447. 34 S. Choudhury, P. K. Mondal, V. K. Sharma, S. Mitra, V. G. Sakai, R. Mukhopadhyay and S. K. Pal, J. Phys. Chem., 2015, 119(34), 10849–10857. 35 J. Martin, M. Krutyeva, M. Monkenbusch, A. Arbe, J. Allgaier, A. Radulescu, P. Falus, J. Maiz, C. Mijangos, J. Colmenero and D. Richter, Phys. Rev. Lett., 2010, 104, 197801. 36 T. Glomann, A. Hamm, J. Allgaier, E. G. Hubner, A. Radulescu, B. Faragoc and G. J. Schneider, Soft Matter, 2013, 9, 10559–10571.
This journal is © The Royal Society of Chemistry 2017
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
PAPER
Cite this: DOI: 10.1039/c7sm01262d
View Journal
Quasielastic neutron scattering study of microscopic dynamics in polybutadiene reinforced with an unsaturated carboxylate† Ryo Mashita, ‡a Rintaro Inoue,‡b Taiki Tominaga,c Kaoru Shibata,d Hiroyuki Kishimotoa and Toshiji Kanaya *e We studied the dynamics of zinc diacrylate (ZDA) reinforced polybutadiene rubber (BR) (ZDA/BR) using the quasielastic neutron scattering technique to determine the effect of concentration of ZDA on polymer dynamics. First, we evaluated the temperature dependence of mean square displacements (hu2i) for ZDA/BR with different ZDA volume fractions. hu2i increased with temperature below 170 K, and we observed no significant ZDA volume fraction dependence. However, it increased more steeply above 170 K, and the value of hu2i was smaller for the samples with increasing ZDA fraction. To elucidate the origin of the decrease in hu2i with increasing ZDA content, dynamic scattering laws (S(Q,o)) were analyzed. An increase in the elastic component, an increase in the mean relaxation time, and a broadening of distribution of relaxation time were observed with the increasing volume fraction of ZDA. In addition, the ZDA volume fraction dependence of the elastic component roughly corresponded to that of elastic
Received 27th June 2017, Accepted 19th September 2017 DOI: 10.1039/c7sm01262d
modulus, indicating that the elastic component is related to its mechanical strength. Referring to the previously reported static structure of the present ZDA/BR system, a model for the heterogeneous BR dynamics was proposed. This model assumes the coexistence of immobile, mobile, and interfacial constrained mobile regions. It was found to be appropriate for the explanation of the observed dynamics. We proposed that a network-like structure of the BR having a high crosslinking density around ZDA
rsc.li/soft-matter-journal
aggregates is mainly responsible for the high elastic modulus of ZDA/BR.
1 Introduction Rubber materials utilized in various industrial products, such as automobile tires, airplanes, and base isolation devices, are indispensable in industry and our daily life. Fillers or functional agents are used to enhance the properties of pure rubber, such as mechanical strength, abrasion resistance, and deterioration resistance. The mechanical properties of rubber materials can be enhanced by the addition of crosslinking agents or reinforcing fillers, such as carbon black (CB), silica, or a
SUMITOMO Rubber Industries, LTD, 1-1, 2-chome, Tsutsui-cho, Chuo-ku, Kobe 651-0071, Japan b Research Reactor Institute, Kyoto University, 2, Asashironishi, Kumatori-cho, Sennan-gun, Osaka 590-0494, Japan c Neutron Science and Technology Center, Comprehensive Research Organization for Science and Society (CROSS), Tokai, Ibaraki 319-1106, Japan d Neutron Science Section, Material and Life Science Division, J-PARC Center, JAEA, Tokai, Ibaraki 319-1195, Japan e Material and Life Science Division, J-PARC Center, KEK, 203-1 Shirakata, Tokai, Ibaraki 319-1106, Japan. E-mail: [email protected] † Electronic supplementary information (ESI) available. See DOI: 10.1039/ c7sm01262d ‡ These authors contributed equally to this work.
This journal is © The Royal Society of Chemistry 2017
clay.1 The mechanical strength of polybutadiene rubber (BR) reinforced with zinc diacrylate (ZDA) (ZDA/BR) is known to be considerably higher than that of the typical carbon black or silica reinforced rubber.2–8 The mechanism underlying the high mechanical strength attained when using ZDA remains unknown despite extensive studies. It has been conjectured that the failure to elucidate the reinforcement mechanism in ZDA/BR originates mainly from the lack of detailed information about both its static and dynamic structures. To overcome this shortcoming, first, we studied the structure of ZDA/BR through the complementary use of small-angle X-ray scattering (SAXS) and small-angle neutron scattering (SANS) techniques.8,9 In addition to the hierarchical structures of the ZDA aggregates and BR matrix, a hierarchical BR structure possessing a higher crosslinking density (HC-BR) than that of the BR matrix was observed for ZDA/BR. Furthermore, it is expected that such HC-BR region is mainly segregated on the surface of ZDA aggregates through the contrast variation SANS (CV-SANS) method. It is reported that the dynamics of such a spatially or geometrically confined polymer is different from that of the bulk system.10 Namely, the dynamics of such a confined polymer is severely affected by such a spatially or geometrically confined
Soft Matter
View Article Online
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
Paper
Soft Matter
polymer contribution. As for the BR system, Roh et al. investigated the dynamics of the CB/BR system at one CB concentration through quasielastic neutron scattering (QENS), which can probe dynamics at a microscopic spatial scale.11 They reported the slowing down of the segmental dynamics of matrix BR due to physical/chemical absorption of BR on the surface of CB. It is reported that the mechanical strength of both CB/BR and ZDA/BR is strongly dependent on the concentration of filler or reinforcing agents.8 Hence, the studies focusing on only one filler concentration are not enough for ruling out the dominant factor for both mechanical strength and attributing dynamics. At present, the concentration dependency of reinforcing agents has not been well studied due to the various controlling parameters. Hence, tuning the concentration of reinforcing agents, which is dependent on various fields of application of the BR product, is vital for industrial application. In this manuscript, we try to study the effect of concentration of reinforcing material (ZDA) on the dynamics of ZDA/BR, which exhibits high mechanical strength, through QENS. The mechanisms of various polymer dynamics have been successfully studied with this method.11–19 Based on the previous CB/BR system, the relevance between mechanical strength and dynamics is also discussed.
2 Experimental 2.1
Materials and sample preparation
The details of the materials used in the present study are summarized in Table 1. We used polybutadiene (BR; BR730, JSR Co., Ltd) with a weight-average molecular weight (Mw) of 6.72 105 and a molecular weight distribution index (Mw/Mn) of 2.39. Mn is the number-average molecular weight of the matrix rubber. The percentage of cis, trans, and vinyl in the present BR is 97, 2, and 1%, respectively. Zinc diacrylate (ZDA; Sanceler SR, Sanshin Chemical Industry Co., Ltd) was used as a cross-linker and dicumyl peroxide (DCP; Percmyl D, Nichiyu Co., Ltd) was used as an initiator for the crosslinking reaction. The chemical formula of the ZDA monomer is shown in Fig. 1. BR, ZDA, and DCP were placed into a 6-inch two-roll mill simultaneously, and mixed for 5 minutes at room temperature. After mixing at room temperature, the mixture of BR, ZDA, and DCP was placed into a mold and heated at 443 K for 20 minutes to increase the rate of the chemical reaction. The ZDA monomer possesses a carbon–carbon double bond (CQC) at both of its ends. Firstly, the ZDA monomer is grafted to BR by the reaction between one CQC and BR, and then ZDA aggregates with the size of several nm are subsequently formed by chain
Table 1
Fig. 1
Chemical formula of the zinc diacrylate (ZDA) monomer.
transfer reaction between ZDA monomers. Secondly, the crosslinking of BR proceeds through further chemical reaction between the other CQC end of ZDA and BR. It should be noted that ZDA(0) in Table 1 does not include ZDA, but is crosslinked by DCP. DCP reacts with the CQC of BR, and the generated carbon radicals react with another CQC of BR. Then, intermolecular carbon–carbon bonds (C–C) are formed. ZDA(0) is not composed of neat BR but self-crosslinked BR. Fig. 2 shows the relationship between the elastic modulus obtained by dynamic viscoelasticity measurements and the volume fraction of ZDA at 303 K. The elastic modulus increases dramatically above the volume fraction of 0.05. For clarity, we also added a solid line, which only holds for low ZDA contents in Fig. 2. We then performed QENS measurements at the ZDA volume fractions of 0, 0.03, 0.05, 0.1 and 0.14 to find out the origin of the high mechanical strength of ZDA/BR. The glass transition temperatures (Tg) were measured using differential scanning calorimetry (DSC). DSC curves are shown in ESI,† Fig. SI. Two glass transition temperatures were observed at B170 K and at B360 K. The first Tg (B170 K) was identical to that of neat BR and constant regardless of the volume fraction of ZDA (refer Table 1). However, the second Tg was observed at B360 K in the samples except for ZDA(0). This suggests that the samples crosslinked by ZDA have two regions with Tg of 170 K and 360 K, respectively. The scattering contribution of each component was estimated based on the scattering cross section. For ZDA(14), which contained the highest volume fraction of ZDA among the investigated samples, the proportions of the total scattering cross-sections of BR, ZDA, and DCP were estimated to be 94.7 : 5.0 : 0.3. The large scattering contribution of BR is mainly due to the large incoherent scattering cross section of hydrogen atoms. Thus, the observable scattering intensity for ZDA/BR in this QENS study would be dominated by the BR.
2.2
Measurements
QENS measurements were carried out using an inverted geometry time-of-flight spectrometer (BL02 DNA)20 installed at the
Sample composition details and Tg
ZDA(0)
ZDA(3)
ZDA(5)
ZDA(10)
ZDA(14)
Polybutadiene (vol%) Zinc diacrylate (vol%) Dicumyl peroxide (vol%)
99.5 0 0.5
96.8 2.7 0.5
94.2 5.3 0.5
89.5 10.1 0.4
85.2 14.4 0.4
Glass transition temperature1st Tg (K) 2nd Tg (K)
168.6 1 N/A
170.0 1 353.4 1
170.4 1 359.9 1
170.0 1 360.4 1
170.6 1 362.2 1
Soft Matter
This journal is © The Royal Society of Chemistry 2017
View Article Online
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
Soft Matter
Paper
Fig. 2 Relationship between the elastic modulus and volume fractions of zinc diacrylate (ZDA).
Materials and Life Science Experimental Facility (MLF) in J-PARC, Tokai, Japan. The magnitude of the scattering vector Q (Q = 4p sin y/lf, where 2y and lf = 6.26 Å are the scattering angle and the wavelength of the analyzed neutron, respectively) ranged from 0.12 to 1.78 Å1. To obtain the first Tg of the present ZDA/BR samples, the temperature was varied from 10 K to 330 K at a heating rate of 1 K per min for elastic scattering measurements, as shown later in Fig. 4. On the other hand, dynamic scattering laws S(Q,o), as shown later in Fig. 3, 5, and 6, were measured under isothermal conditions. It should be noted that this temperature range was below the second Tg (B360 K). Each sample with a thickness of 200 20 mm was placed in a double-cylindrical aluminum cell (outer diameter: 14 mm, inner diameter: 13 mm, height: 45 mm) under a helium atmosphere. The dynamic scattering laws S(Q,o) of each sample at 10 K were used to determine the resolution function of the corresponding sample. The energy resolution (dE) defined as the full width at half maximum (FWHM) was estimated to be 4.0 0.2 meV. As seen in ESI,† Fig. SII, the resolution functions were constant regardless of the samples. DSC measurements were carried out using a differential scanning calorimeter (Q200, TA Instruments) in the temperature range from 123 K to 423 K with a heating rate of 10 K per min. Dynamic viscoelasticity measurements were carried out using a dynamic mechanical analyzer (RSA-G2, TA Instruments), at 303 K and 10 Hz.
3 Results and discussion Fig. 3(a) and (b) shows the dynamic scattering laws S(Q,o) with a mean Q of 1.48 Å1 for ZDA(0) and ZDA(14), respectively. The graphs are plotted for several temperatures above the first Tg but below the second Tg, except for 10 K. S(Q,o) apparently consists of elastic and quasielastic components above the first Tg, suggesting that both of the samples ZDA(0) and ZDA(14) have an immobile component. In the case of ZDA(14), this must have originated from the glassy region below the second Tg. On the other hand, ZDA(0) also exhibited the elastic component. This must be due to the structure of BR cross-linked by DCP.
This journal is © The Royal Society of Chemistry 2017
Fig. 3 Dynamic scattering laws S(Q,o) for (a) ZDA(0) at 10 K ( ), 200 K ( ), 250 K ( ), 280 K ( ), 310 K ( ) and 330 K (J) and (b) ZDA(14) at 10 K ( ), 200 K ( ), 250 K ( ), 280 K ( ), 310 K ( ) and 330 K (,), obtained by summing the spectra from Q = 1.38 to 1.58 Å1. The mean value of Q was 1.48 Å1.
Comparison of S(Q,o) for ZDA(0) and ZDA(14) shows that the elastic component is larger and the quasielastic component is narrower in ZDA(14) than ZDA(0). This implies the suppression of motion in BR by the addition of ZDA in the present energy region. To investigate the effect of ZDA on the observable dynamics in more detail, we evaluated the mean-square displacement (hu2i) from the Q2 dependence of the elastic scattering intensity (Iel,T(Q)), assuming the Gaussian approximation.21,22 This is shown by eqn (1). Iel;T ðQÞ ¼ exp u2 Q2 Iel;T¼10K ðQÞ
(1)
Iel,T=10K(Q) is the elastic scattering intensity at 10 K. The left side of eqn (1) approximates the incoherent elastic scattering intensity under the assumption that the structure factor is independent of temperature. The results at 300 K, which is above the first Tg, calculated by eqn (1) are shown in Fig. 4(a). The experimental data conform to calculations of eqn (1), within experimental error. Fig. 4(b) shows the temperature dependence of hu2i on different volume fractions of ZDA for ZDA/BR. In all samples, hu2i increases linearly with temperature in the range of 10 K to 110 K. This
Soft Matter
View Article Online
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
Paper
Fig. 4 (a) Fitting results of Iel,T(Q)/Iel,10K(Q) for ZDA(0) ( ), ZDA(5) ( ) and ZDA(14) ( ) at 300 K using eqn (1) (solid lines). (b) The temperature dependence of hu2i obtained by eqn (1) for ZDA(0) ( ), ZDA(3) (}), ZDA(5) ( ), ZDA(10) ( ) and ZDA(14) ( ). The arrow indicates the first Tg. (c) ZDA volume fraction dependence of the inflexion point of hu2i ( ) and Tg obtained by DSC ( ).
suggests that the motion observed below 110 K is attributed to harmonic oscillation. Changes in the slope of hu2i are observed at 110 K and around the first Tg (B170 K), for all the samples. Referring to the previous works,23,24 the first inflexion point (B110 K) corresponds to the onset of the fast b-process. The second inflexion point (B170 K) clearly indicates the onset of a new motion above the first Tg, which is probably the so-called a-process. Tyagi et al. reported that the relaxation time at around Q = 1.0 Å1 by QENS nicely coincided with those from other
Soft Matter
Soft Matter
dynamical methods.25 Namely, the hierarchical dynamics holds for the a-process. In our previous work,26 we also revealed that excess hu2i, which is the difference of hu2i between two different energy resolutions, exhibits an inversely linear relationship with ln Z (Z: viscosity) by comparison of the results obtained from different energy resolution QENS measurements under the idea of free volume theory. Furthermore, the universal correlation of the structural relaxation time (as well as the viscosity) and the rattling amplitude from glassy to low-viscosity states is fulfilled from MD simulation and the experimental data of several glass-formers and polymers.27 Fig. 4(c) shows the dependence of the second inflexion point (B170 K) of hu2i on the volume fraction of ZDA, which corresponds to the first Tg measured by DSC. It should be noted that the onset temperature is independent of the ZDA volume fraction. In addition, the value of hu2i decreased clearly with the increasing volume fraction of ZDA above the first Tg, especially above B250 K, as shown in Fig. 4(b). In order to interpret the results, we need to recall that the ZDA/BR samples have a tworegion structure defined by the first Tg (B170 K) and the second Tg (B360 K). The fact that the first Tg is independent of the volume fraction of ZDA means that the region between the first and second Tg includes the mobile part, which has the same mobility as neat BR. As for the decrease in hu2i with increasing ZDA volume fraction, there are two plausible reasons. One is the increase of the second (immobile) region, and the other is the increase of the less mobile part in the first (mobile) region. The problem will be discussed based on the analysis of S(Q,o). However, the dynamical heterogeneity increases with the increasing volume fraction of ZDA. To understand the physical origin of the independence of the first Tg and the decrease of hu2i with the increasing volume fraction of ZDA, we analyzed S(Q,o), as shown in Fig. 3. A clear broadening of the central component of S(Q,o) was observed above the first Tg for both ZDA(0) and ZDA(14). In addition, the broadening of S(Q,o) was observed for other samples (data not shown). Based on a previous QENS study on neat BR with similar energy resolution,26 it is expected that the broadening of S(Q,o) must reflect the detection of the a-process. Fig. 5 shows S(Q,o) normalized by the peak intensity of the elastic scattering from different ZDA volume fractions at 310 K in the range Q = 1.48 Å1. Consistent with the decrease of hu2i, with the increase in the ZDA volume fraction, a narrowing of S(Q,o) is observed. To extract the parameters describing dynamical information from the S(Q,o), we performed curve fitting to the observed S(Q,o) at different ZDA volume fractions using an appropriate model function. Richter et al. reported that the intermediate scattering function corresponding to the a-process of neat BR was well described by the Kohlrausch– Williams–Watts (KWW) equation.28,29 Hence, we adopted the KWW function to describe the observed S(Q,o), and the following model function was proposed. n j ko S ðQ; oÞ ¼ A ð1 ecf ÞF exp ðt=tÞb þ ecfdðoÞ RðQ; oÞ
(2)
where A, ecf, t, and b are the amplitude of the relaxation function, the elastic component fraction, the relaxation time,
This journal is © The Royal Society of Chemistry 2017
View Article Online
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
Soft Matter
Paper
Fig. 5 Dynamic scattering laws S(Q,o) for ZDA(0) ( ), ZDA(3) (}), ZDA(5) ( ), ZDA(10) ( ) and ZDA(14) ( ) normalized at the peak top at 310 K in the range Q = 1.48 Å1. The dashed line indicates the instrumental resolution function.
and the distribution of the relaxation function, respectively. It must be kept in mind that we have introduced an elastic component in eqn (2) even at temperatures above the first Tg, because the model did not work well without the elastic component. In addition, the samples have a glassy region at 310 K below the second Tg (B360 K). It should be noted that the KWW function is defined in the time domain, necessitating that we adopt the Fourier-transformed KWW function ðF ðKWWÞÞ in the frequency region.30 R(Q,o) indicates the instrumental resolution function. Fig. 6(a) and (b) shows the results of the fits to S(Q,o) for ZDA(0) and ZDA(14) with eqn (2) at Q = 1.48 Å1 and 310 K, respectively. The S(Q,o)s from both samples are fitted fairly well by eqn (2), supporting the application of an appropriate model function. In the case of ZDA(0), the elastic component was observed although the second Tg was not detected in DSC measurements (Table 1). The elastic component must have originated from the crosslinked structure by DCP. S(Q,o)s are also modelled fairly well by eqn (2) in other Q regions and at other temperatures (see ESI,† Fig. SIII). The dependence of ecf on the volume fraction of ZDA, the average relaxation time hti and b at 310 K in the range Q = 1.48 Å1 are shown in Fig. 7. hti was calculated using eqn (3):30,31 t 1 G (3) hti ¼ b b where G is a gamma function. Summarizing the obtained parameters from the curve fits, the increase in ecf, the decrease of b, and the increase of hti were observed with the increase in the volume fraction of ZDA. The same trends were observed for other Q ranges (see ESI,† Fig. SIV). The elastic fraction comes from the glassy region below the second Tg and/or the confined motion in the mobile region above the first Tg. In the case of ZDA(0), which does not include ZDA, a small amount of the elastic component (B0.02) is observed as seen in Fig. 7(a). This elastic component must be due to the crosslinked structure of BR formed by the
This journal is © The Royal Society of Chemistry 2017
Fig. 6 The fitting results of S(Q,o) for (a) ZDA(0) and (b) ZDA(14) using eqn (2) for Q = 1.48 Å1 at 310 K. Solid lines are the results of calculation with eqn (2). Thick and thin dashed lines are the calculated results obtained by the KWW and elastic components, respectively, in eqn (2).
addition of DCP. The elastic fraction increases with the ZDA volume fraction, suggesting that the elastic fraction mainly originates from the immobile region mediated by the cross links formed by the addition of ZDA. In addition, a deviation from the linear dependence of ecf above the volume fraction of ZDA of 0.05 was seen. Qualitatively, this tendency corresponds to that of elastic modulus as seen in Fig. 2, indicating that the high ecf value observed at high ZDA content is related to high mechanical strength. This will be discussed in a later paragraph based on our previous SANS and SAXS results.8,9 Q dependence of ecf and the results of hypothetical fitting analyses are shown in ESI,† Fig. SV and Table SI, respectively.32–34 As seen in Fig. 7(b), the average relaxation time increases with the volume fraction of ZDA. At the same time, the parameter b, which represents the width of the distribution of relaxation time, decreases with increasing ZDA volume fraction, implying the broadening of the distribution of relaxation time. In other words, the samples become dynamically heterogeneous with the increase of the ZDA volume fraction. In the present analysis, we focused on the dynamics in the mobile region below the second Tg, so that the increase in the dynamic heterogeneity is attributed to the region, but not related to the coexistence of the mobile and the immobile regions.
Soft Matter
View Article Online
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
Paper
Fig. 7 ZDA volume fraction dependence of (a) ecf ( ), (b) hti ( ) and b ( ) at 310 K in the range Q = 1.48 Å1.
To construct a plausible physical picture that could explain the evaluated parameters without inconsistency, we referred to the structural information revealed in our previous work on SAXS and SANS.8,9 From small-angle scattering studies, the coexistence of BR regions having low and high crosslinking densities was revealed. Through DSC measurements a second Tg was also observed at B360 K (Table 1). It should be noted that the highest temperature in the present QENS measurements was lower than this second Tg, suggesting that the first Tg corresponds to the Tg of the low crosslinking density region and the second Tg corresponds to the highly crosslinked region. In the above discussion on S(Q,o), we used the mobile region and the immobile region for the low crosslinking density region and the highly crosslinked region, respectively. Based on this assumption, we considered that the mobile region originates from the low crosslinking density region (LC-BR) and the immobile region originates mainly from the high crosslinking density region (HC-BR), which is still in a glassy state in the present temperature range. As an increase in the elastic fraction was observed with increasing ZDA volume fraction, a physical picture that assumes the coexistence of mobile and immobile regions in the ZDA/BR is seemingly applicable to the observed dynamics. In addition, the ZDA volume fraction dependence of the second Tg obtained by DSC qualitatively reproduced that of ecf (see ESI,† Fig. SVI). If the mobile and immobile regions are dynamically homogeneous for the ZDA/BR, the evaluated relaxation times or distributions of relaxation times must be constant, regardless of the volume fraction of ZDA. However, a
Soft Matter
Soft Matter
slight increase in the relaxation time or broadening of the distribution of relaxation time with the increase in the volume fraction of ZDA was observed (see Fig. 7(b)). Thus, a contributing factor other than the coexistence of mobile and immobile components must be present in the ZDA/BR system. It is known that a polymer near a substrate or an impenetrable boundary exhibits anomalous physical properties.35,36 Hence, it is expected that the dynamics of BR near the HC-BR regions would be strongly perturbed by the HC-BR region in the glassy state, resulting in slower dynamics than that of neat BR. With the increase in the volume fraction of ZDA, the contributions from such interfacial regions near HC-BR would increase. Thus, due to the gradient of dynamics mediated by the ZDA aggregates, more heterogeneous and slower dynamics are observed for ZDA/BR than those of the neat BR. A schematic representation of the ZDA/BR system obtained by the results of our previous work8,9 and this QENS study is shown in Fig. 8. Finally, we compared the present results to previous findings for other rubber systems to find out the origin of the high mechanical strength of ZDA/BR. Roh et al.11 studied the dynamics of the BR reinforced with carbon black (CB) (CB/BR) through neutron back-scattering, and observed both the slowing and broadening of the a-process for CB/BR compared to neat BR. Qualitatively, both our group and Roh’s group observed similar trends, though a drastic difference between the moduli was seen for the two systems. The moduli of ZDA/BR were several times higher than those of CB/BR.8 In both cases, the formation of network-like structures is believed to be responsible for the improved mechanical strength vital for practical application. Accordingly, it is assumed that the difference in the mechanism of interaction between the BR and the reinforcement agents must be related to the difference in mechanical strength. In the CB/BR system, it was considered that a network-like structure of physically adsorbed BR on the surface of CB was related to its mechanical strength.1 In contrast, the formation of covalent bonds between the CQC double bonds in the ZDA aggregates and the BR must effectively contribute to their strong interactions. In addition, with the increasing volume fraction of ZDA, such tightly constrained BR regions must also grow, as described above. Then, the covalently constrained BR on the surface of ZDA (HC-BR) would be expected to act as a kind of cross-linker in ZDA/BR.
Fig. 8 Schematic representation of the ZDA/BR system.
This journal is © The Royal Society of Chemistry 2017
View Article Online
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
Soft Matter
This leads to the formation of a tougher network-like structure than that in CB/BR, in which the physically absorbed BR alone acts as a cross-linker. Formation of such a network-like structure is also supported by the result of ecf as mentioned above. Thus, a more rigid network-like structure is developed in ZDA/BR compared to CB/BR, imparting a higher mechanical strength for the former at a given volume fraction of the reinforcement agent. Namely, it was clearly found that the deviation from the linear dependence of ecf implied the formation of a rigid network-like structure of the immobile component of BR.
4 Conclusions We studied the dynamics of polybutadiene rubber (BR) reinforced with zinc diacrylate (ZDA) (ZDA/BR) by the quasielastic neutron scattering (QENS) technique. The objective of this study was to determine the effect of concentration of reinforcing material on polymer dynamics. The mean square displacements (hu2i) of ZDA/BR for different ZDA volume fractions were evaluated. At temperatures below 170 K, no significant ZDA volume fraction dependence was observed for hu2i. On the other hand, hu2i increased more steeply above 170 K and a clear reduction of hu2i with the increasing volume fraction of ZDA was observed. To understand the origin of the decrease in hu2i with the increasing volume fraction of ZDA, dynamic scattering laws (S(Q,o)) were analyzed with a model function. We assumed the addition of an elastic component and a Fourier-transformed KWW function in the frequency region. With this model function, the observed S(Q,o)s for different ZDA volume fractions were fitted well. From the results of the curve fittings, an increase in the elastic fraction (ecf), an increase of the mean relaxation time (hti), and a broadening of the distribution of relaxation time (b) were observed with the increase in the volume fraction of ZDA. In addition, a deviation from the linear dependence of ecf above the volume fraction of ZDA of 0.05 was seen. This behavior of ecf corresponds to that of elastic modulus, indicating that the high ecf value observed at high ZDA content is related to high mechanical strength. Based on a previously reported static structure, the idea of heterogeneous dynamics, which assumes the existence of immobile, mobile, and interfacial constrained mobile regions, was proposed. This is a consistent model for understanding the observed dynamics. By comparing the reported results for CB/BR with those for ZDA/BR in this study, it was postulated that the difference in the interaction between the BR and the reinforcement agents must be linked to the huge difference in mechanical strength. In the case of ZDA/BR, the existence of chemical bonds between the BR and ZDA contributed to the effective, strong interaction between the BR and the reinforcement agents. Thus, a tougher network-like structure is formed in ZDA/BR than that in CB/BR, in which physical adsorption played the main role. Formation of such a network-like structure was also clearly supported by the resulting dynamics as a function of ZDA concentration. Consequently, a higher mechanical strength was realized for ZDA/BR as compared to that of CB/BR at a given volume fraction of reinforcement agents.
This journal is © The Royal Society of Chemistry 2017
Paper
Conflicts of interest There are no conflicts to declare.
Acknowledgements The QENS measurements on the BL02 DNA instrument installed at J-PARC were performed with the approval of the Materials and Life Science Experimental Facility (MLF) of J-PARC (Proposal No. 2014A0014). This work was supported by ‘‘Photon and Quantum Basic Research Coordinated Development Program’’ from the Ministry of Education, Culture, Sports, Science and Technology, Japan.
References 1 Y. Fukahori, Rubber Chem. Technol., 2007, 80, 701–725. 2 B. Chu, D. Q. Wu, R. D. Lundberg and W. J. MacKnight, Macromolecules, 1993, 26, 994–999. 3 Y. Tsujita, M. Yasuda, M. Takei, T. Kinoshita, A. Takizawa and H. Yoshimizu, Macromolecules, 2001, 34, 2220–2224. 4 M. Al-Hussein, B. G. G. Lohmeijer, U. S. Schubert and W. H. de Jeu, Macromolecules, 2003, 36, 9281–9284. 5 C. N. Zhou, D. C. Chan and I. K. Winey, Macromolecules, 2008, 41, 6134–6140. 6 E. M. Seitz, D. C. Chan, L. K. Opper, W. T. Baughman, B. K. Wagener and I. K. Winey, J. Am. Chem. Soc., 2010, 132, 8165–8174. 7 N. Nagata, T. Sato, T. Fujii and Y. Saito, J. Appl. Polym. Sci., 1994, 53, 103–120. 8 R. Mashita, H. Kishimoto, R. Inoue and T. Kanaya, Polym. J., 2016, 48, 239–245. 9 R. Mashita, H. Kishimoto, R. Inoue and T. Kanaya, Polym. J., 2013, 45, 57–63. 10 G. B. McKenna, Eur. Phys. J.: Spec. Top., 2010, 189, 285–302. 11 J. H. Roh, M. Tyagi, T. E. Hogan and C. M. Roland, Macromolecules, 2013, 46, 6667–6669. 12 J. H. Roh, M. Tyagi, T. E. Hogan and C. M. Roland, J. Chem. Phys., 2013, 139, 134905. 13 V. Arrighi, J. S. Higgins, A. N. Burgess and G. Floudas, Polymer, 1998, 39, 6369–6376. 14 S. Gagliardi, V. Arrighi, R. Ferguson and M. T. F. Telling, Physica B, 2001, 301, 110–114. 15 C. Karlsson, A. S. Best, J. Swenson, W. S. Howells and L. Borjesson, J. Chem. Phys., 2003, 118, 4206. 16 K. Chrissopoulou, S. H. Anastasiadis, E. P. Giannelis and B. Frick, J. Chem. Phys., 2007, 127, 144910. 17 S. Fotiadou, K. Chrissopoulou, B. Frick and S. H. Anastasiadis, J. Polym. Sci., Part B: Polym. Phys., 2010, 48, 1658–1667. 18 P. Akcora, S. K. Kumar, V. G. Sakai, Y. Li, B. C. Benicewicz and L. S. Schadler, Macromolecules, 2010, 43, 8275–8281. 19 K. Sinha, W. Wang, K. I. Winey and J. K. Maranas, Macromolecules, 2012, 45, 4354–4362. 20 K. Shibata, N. Takahashi, Y. Kawakita, M. Matsuura, T. Yamada, T. Tominaga, W. Kambara, M. Kobayashi,
Soft Matter
View Article Online
Paper
21 22 23
Published on 20 September 2017. Downloaded by Gazi Universitesi on 12/10/2017 08:32:19.
24 25 26
27 28
Y. Inamura, T. Nakatani, K. Nakajima and M. Arai, JPS Conf. Proc., 2015, 8, 036022. R. Inoue, T. Kanaya, K. Nishida, I. Tsukushi and K. Shibata, Phys. Rev. Lett., 2005, 95, 05102. T. Kanaya and K. Kaji, Adv. Polym. Sci., 2001, 154, 87–141. T. Kanaya, T. Kawaguchi and K. Kaji, J. Chem. Phys., 1993, 98, 8262–8270. J. Colmenero and A. Arbe, Phys. Rev. B: Condens. Matter Mater. Phys., 1998, 57, 13508–13513. M. Tyagi, A. Alegria and J. Colmenero, J. Chem. Phys., 2005, 122, 244909–1-9. T. Kanaya, T. Tsukushi, K. Kaji, J. Bartos and J. Kristiak, Phys. Rev. E: Stat. Phys., Plasmas, Fluids, Relat. Interdiscip. Top., 1999, 60, 1906–1912. L. Larini, A. Ottochian, C. De Michele and D. Leporini, Nat. Phys., 2008, 4, 42–45. D. Richter, M. Monkenbusch, A. Arbe and J. Colmenero, J. Non-Cryst. Solids, 2001, 287, 286–296.
Soft Matter
Soft Matter
29 D. Richter, B. Frick and B. Farago, Phys. Rev. Lett., 1988, 61, 2465–2468. 30 F. Alvarez, A. Alegria and J. Colmenero, Phys. Rev. B: Condens. Matter Mater. Phys., 1991, 44, 7306–7312. 31 T. Kanaya, R. Inoue, M. Saito, M. Seto and Y. Yoda, J. Chem. Phys., 2014, 140, 144906. 32 F. Volino and A. J. Dianoux, Mol. Phys., 1980, 41, 271–279. 33 S. Fujiwara, K. Araki, T. Matsuo, H. Yagi, T. Yamada, K. Shibata and H. Mochiauki, PLoS One, 2016, 11, e0151447. 34 S. Choudhury, P. K. Mondal, V. K. Sharma, S. Mitra, V. G. Sakai, R. Mukhopadhyay and S. K. Pal, J. Phys. Chem., 2015, 119(34), 10849–10857. 35 J. Martin, M. Krutyeva, M. Monkenbusch, A. Arbe, J. Allgaier, A. Radulescu, P. Falus, J. Maiz, C. Mijangos, J. Colmenero and D. Richter, Phys. Rev. Lett., 2010, 104, 197801. 36 T. Glomann, A. Hamm, J. Allgaier, E. G. Hubner, A. Radulescu, B. Faragoc and G. J. Schneider, Soft Matter, 2013, 9, 10559–10571.
This journal is © The Royal Society of Chemistry 2017
