Co-compatibilising effect of carbon nanotubes and liquid isoprene rubber on carbon black filled natural rubber/polybutadiene rubber blend

Abstract

Bis-(triethoxysilylpropyl)-tetrasulfane functionalised carbon nanotubes (t-CNTs) were used as compatibiliser along with liquid isoprene rubber (LIR) in the natural rubber (NR)/polybutadiene rubber (BR) blend. Their reinforcing and compatibilising effects were evaluated by mechanical, fatigue crack growth resistance properties and blend homogeneity. Scanning electron microscope and transmission electron microscope showed enhanced interfacial adhesion between the binary rubber phases and improved dispersion of the minor phase in the rubber blend respectively with the co-existence of LIR and carbon nanotubes. The tensile strength of the carbon black (CB) filled NR/BR blend reached its optimum when 3 phr CB was replaced with an equal amount of t-CNTs in the presence of 7 phr LIR, while the fatigue crack growth resistance property achieved its maximum in the presence of 3 phr LIR. This interesting co-compatibilisation behaviour of t-CNTs and LIR suggests that t-CNTs have a better effect than CB with the assistance of LIR, which is an effective plasticiser in the NR/BR blend.

Keywords

Compatibilisation Carbon nanotubes Natural rubber Liquid isoprene rubber

Introduction

Elastomer blends are frequently used in the rubber industry to achieve the best balance among physical properties, processability and cost. However, most of the elastomer blends are incompatible. The incompatible blends are usually characterised by multiphase morphology, narrow interphase, inferior physical and chemical interactions across the phase boundaries and poor mechanical properties, which limit their application in the industry. As a traditional rubber blend, natural rubber/polybutadiene rubber (NR/BR) blend, which is also an incompatible rubber pair, has gained great attention in both tire industry and scientific research. The effect of thermal aging and blend ratio on the physical properties,1^–3 the effect of additive type4^–6 and the influences of mixing procedures and additive incorporation sequence on the NR/BR blend properties7, 8 were investigated. However, the great challenge is how to improve compatibilisation of the rubber pair.

Compatibilisation strategies for immiscible polymer blends have extensively been investigated in the past decade. The most conventional method involves the use of homopolymer,9, 10 block or graft co-polymers,11^–16 whose blocks are structurally similar or thermodynamically miscible with the heterogeneous components. Moreover, there is also early explored compatibilisation method using isotropic nanofiller such as carbon black (CB)17^–19 and nanosilica.20, 21 However, recent studies have reported that the compatibilisation effect of anisotropic nanofillers such as layered silicate on the immiscible polymer blend is better than the isotropic one.22^–27 The drastic reduction of the characteristic size of the polymer domains and good mechanical properties in comparison to the incompatibilised blend were found in these studies. Two possible mechanisms were proposed to understand the compatibilising effect of inorganic fillers. The first is mechanism I (thermodynamic compatibility): with a large specific area and high aspect ratio, anisotropic inorganic nanofillers can absorb the polymer chains on its surface to gain stabilising energy, which makes the overall free energy of mixing ΔG_m negative and thermodynamically driven compatibility between immiscible components.10 The second is mechanism II (kinetic compatibility): the selective localisation of nanofiller in the polymer interface decreases interfacial tension and prevents coalescence of the domains during melt mixing, which kinetically improves the compatibility between the binary phases.7

As anisotropic one-dimensional nanofiller, carbon nanotube (CNT) is recognised as a promising reinforcement for the manufacture of high performance, multifunctional composites due to their extraordinary properties such as excellent Young's modulus, good flexibility and high electrical and thermal conductivity. Currently, however, there are only few papers dealing with the compatibilisation effect of CNTs in the polymer blend.28^–30 In addition, liquid isoprene rubber (LIR), as a plasticiser with low molecular weight, is commonly used in NR, BR and other low polarity rubber. The addition of the plasticiser LIR, which has similar molecular structure and good miscibility with binary blends, improves the processing and reduces the interfacial energy of the rubber blend, which is beneficial to the even migration and dispersion of the filler in the binary blend.

However, to the best of our knowledge, no work has concerned the co-compatibilisation effect of CNTs and LIR in the NR and BR blend up to now.

In the present work, bis-(triethoxysilylpropyl)-tetrasulfane (TESPT) functionalised CNTs (t-CNTs) along with LIR are used as both compatibiliser and nanoreinforcement to modify the NR/BR blends. A co-compatibilising effect of t-CNTs and LIR was observed, and the co-compatibilising behaviour was studied.

Experimental

Materials

Natural rubber (CSR5) was purchased from Yunnan Natural Rubber Industry Company Limited (Kunming, China). Polybutadiene rubber (BR9000, cis-1,4-polybutadiene, 96%) was supplied by Dushanzi Petrochemical Company (Xinjiang, China). Liquid isoprene rubber (LIR 50) was purchased from Kuraray Co., Ltd. Carbon black (N330) with a size of 22–24 nm and a surface area of about 75–95 m² g⁻¹ was provided by the Institute of Carbon Black (Zigong, China). Hydroxyl multiwalled CNTs (CNTs-OH, ∼3 wt-%OH) with a diameter of 10–20 nm were obtained from Chengdu Institute of Organic Chemistry, Chinese Academy of Sciences. Bis-(triethoxysilylpropyl)-tetrasulphide was provided by Nanjing Shuguang Chemical Company (Nanjing, China). Ethanol was purchased from Chengdu Haihong Chemical Reagent Company (China). Other reagents including zinc oxide, stearic acid, sulphur, antioxidant (4010NA) and accelerators, including N-cyclohexyl-2-benzothiazole-sulphenamide and 2-mercaptobenzothiazole, are all commercially available.

Fabrication of t-CNTs

The synthesis of t-CNTs shown in Fig. 1 was performed in ethanol solution, which has been reported elsewhere.31 In a typical procedure, 20 g CNTs-OH was dispersed in a mixed solution of 180 mL ethanol, 2 g TESPT and 20 mL deionised water. The pH value of the solution was adjusted to ∼4·5 with acetic acid. The mixture was then transferred into a three-neck flask equipped with a condenser and agitated for 8 h at 80°C. It was then washed with deionised water by filtration and dried in the vacuum oven at 80°C for ∼12 h. Thus, the TESPT treated CNTs-OH (t-CNTs) containing polysulphide bonds were obtained.

Figure 1

Schematic of modification process of CNTs-OH with TESPT

Preparation of NR/BR samples

The NR, BR and LIR were first mixed in a Haake Rheomix 600 internal mixer at 60°C and 30 rev min⁻¹ for 10 min. Then, the compounds were plunged into a two-roll mill and milled for an additional 5 min at 60°C, during which the t-CNTs, CB and curing agents were sequentially added. The blends were pressed into sheets of 2 mm thickness at 150°C under a pressure of 10 MPa according to respective optimum cure time t₉₀ and then compression moulded at room temperature. Experimental compositions are shown in Table 1.

Table 1

Recipe for preparation of NR/BR composites*

Components/phr	C0L0	C3L0	C0L3	C3L3	C0L5	C3L5	C0L7	C3L7
NR	80	80	80	80	80	80	80	80
BR	20	20	20	20	20	20	20	20
ZnO	5	5	5	5	5	5	5	5
SA	2	2	2	2	2	2	2	2
Antioxidant (4010NA)	3	3	3	3	3	3	3	3
CBS	1·5	1·5	1·5	1·5	1·5	1·5	1·5	1·5
M	0·1	0·1	0·1	0·1	0·1	0·1	0·1	0·1
S	3	3	3	3	3	3	3	3
CB	25	22	25	22	25	22	25	22
t-CNTs	0	3	0	3	0	3	0	3
LIR	0	0	3	3	5	5	7	7

*NR: natural rubber; BR: polybutadiene rubber; SA: stearic acid; CBS: N-cyclohexyl-2-benzothiazole-sulphenamide; M: 2-mercaptobenzothiazole; CB: carbon black; t-CNTs: functionalised carbon nanotubes; LIR: liquid isoprene rubber.

Characterisation

Fourier transform infrared (FTIR) spectra

Fourier transform infrared was performed on a Nicolet 560 FTIR spectrometer for CNTs-OH and t-CNTs. Before the FTIR test, the t-CNTs were extracted by a Soxhlet extractor in ethanol for 48 h. The FTIR measurements were conducted in transmission mode in KBr pellets. The spectra were collected from 4000 to 400 cm⁻¹ with a resolution of 4 cm⁻¹ over scans.

Scanning electron microscopy (SEM)

The blend morphology was examined by SEM using an inspect F model FEI apparatus operating at an accelerating voltage of 10 kV. The samples were fractured in liquid nitrogen and then coated with gold.

Transmission electron microscopy (TEM)

The samples for TEM analysis were prepared by ultracryomicrotomy using a Leica EM UC6/FC6 to get cryosections of 70–80 nm thickness. The microscopic study was performed using an FEI Tecnai G² F20 S-Twin TEM operating at an accelerating voltage of 200 kV.

Dynamic mechanical analysis (DMA)

A TA Q800 DMA instrument was used in the tension mode on a sample of ∼25 mm in length, 4 mm in width and 2 mm in thickness. Temperature–time scans were carried out at a frequency of 1 Hz in the temperature range of −120–30°C at a heating rate of 3 K min⁻¹. The loss modulus peak area was calculated by integrating the loss modulus versus temperature curve in the corresponding temperature range.

Mechanical properties

The tensile measurements were performed according to China standard GB/T528-1998. The tensile strength and elongation at break were measured in a tensile Instron 5567 tester at a strain rate of 500 mm min⁻¹ at room temperature. The result was the average values of five specimens.

Fatigue testing

Fatigue tests were conducted on edge notched rectangular specimens.32 The specimens had a gage length of 89·0 mm, a width of 25·4 mm, a thickness of 1·8 mm and an initial crack of length ∼5·0 mm cut into the edge of the rubber specimens. The specimens were subjected to fatigue testing with a displacement controlled tension mode in MTS810 at room temperature. The maximum strain was 20%, and the minimum strain was 0. The waveform of the fatigue cycle was sinusoidal at a frequency of 3 Hz, and the R ratio for the test was set at 0. The R ratio is defined as the ratio of minimum to maximum deformation of rubber during cycles. During the tests, the machine was periodically stopped to record the number of cycles and the length of the crack as measured by a vernier calliper.

Results and discussion

Modification of CNTs

The schematic illustration of the modification process of CNTs-OH with TESPT is shown in Fig. 1. The FTIR spectra of the CNTs-OH and t-CNTs are shown in Fig. 2. Table 2 lists the FTIR peak attributions. The band at ∼3436 cm⁻¹ can be attributed to –OH stretching vibrations. Compared with CNTs-OH, t-CNTs exhibit the characteristic peak of silicon in the low wavenumber range. Two weak peaks at 1066 and 560 cm⁻¹ can be ascribed to C–O–Si and Si–OH stretching vibrations respectively. The FTIR results indicated that after the reaction with TESPT, the polysulphide group was bonded to the surface of CNTs.

Figure 2

Fourier transform infrared spectra of a CNTs-OH and b t-CNT

Table 2

Fourier transform infrared results for t-CNTs

Wavenumbers/cm⁻¹	Group	Vibration mode
3436	–OH	υ(O–H)
2923	–CH₃	υ(C–H)
1380	–OH	σ(–OH)
1295	C–O	υ(C–O)
1066	O–Si	υ(O–Si)
560	O–Si–O	σ(O–Si–O)

Phase morphology

The cryogenitically fractured surfaces of the samples are shown in Fig. 3. Figure 3a shows many large particles with poor interfacial adhesion in the fractured surface of sample C0L0. The particles with average size of tens of micrometres should be the immiscible polybutadiene rubberrish blend. The phase separation can be obviously observed. Upon the addition of 7 phr LIR, the particle size in C0L7 significantly decreases compared with C0L0, as shown in Fig. 3b. However, the surface is still coarse with visible interfaces, indicating the expected insufficient miscibility. Figure 3c shows that sample C3L0 with immiscible BR particles in the fracture surface has a similar structure to C0L0, which also suggests the insufficient compatibility. With the co-existence of t-CNTs and LIR, the fracture surface of the sample shows a decreased trend in BR phase size and improved interface adhesion between the NR and BR phases for samples C3L0, C3L3, C3L5 and C3L7, suggesting that the compatibilising effect is enhanced with increasing LIR content.

Figure 3

Images (SEM) of NR/BR/CB blend fracture surfaces with different CNTs and LIR content

To further testify the co-compatibilisation of LIR and t-CNTs, TEM analysis was conducted for the selected samples C0L0 and C3L7 to observe the dispersed BR phase and t-CNTs in the blend. Figure 4a gives a view of a two-phase structure, where the BR domains (dark region) dispersed in the NR matrix, which is attributed to the immiscibility of the binary rubber phase. Figure 4b shows that the interface of the two phases is vague, and the large BR domains in Fig. 4a disappear and the BR phase is evenly dispersed in the NR matrix with the simultaneous addition of t-CNTs and LIR. These suggest that the improved compatibilisation of the NR/BR blend was gained by the co-existence of LIR and t-CNTs. Furthermore, it can be clearly seen in Fig. 4c that the individual CNT travels through the CB cluster, suggesting its good dispersion in the rubber blend.

Figure 4

Image (TEM) of a C0L0, b C3L7 and c higher magnification of b

Dynamic mechanical analysis

The tan δ curves for 80NR/20BR/22CB/3t-CNTs with different loadings of LIR are shown in Fig. 5a. The tan δ curve of the NR rich phase shows a peak at ˜−46°C due to the α transition arising from the segmental motion. This corresponds to the glass transition temperature T_g of NR in the blend. The T_g of BR rich phase is at −92°C, as shown by a tan δ peak in the curve. There are two relaxation peaks corresponding to two phases, which suggest the immiscibility of the binary components. However, there is no relaxation peak corresponding to LIR, indicating good miscibility of LIR with both NR and BR phases. It is noteworthy that the tan δ_max peak value for the NR rich phase increases with increasing LIR content, whereas it decreases for the BR rich phase, as shown in Table 3. The area under the loss modulus curve was calculated by integrating the loss modulus versus temperature curve in the temperature range of −110 to −70°C, corresponding to the glass transition process of the BR rich phase, as shown in Table 3. It shows that the loss modulus peak area shifts to the lower value with an increase in the LIR content in the 80NR/20BR/22CB/3t-CNTs blends. The intensity of the tan δ peak and the loss modulus peak area at the glass transition temperature are considered to reflect the mobility of the molecular chain segments.33, 34 The decrease in the intensity of the tan δ peak and loss modulus peak area of the BR phase with increasing LIR content indicates the decrease in mobility of the BR molecular chain segments, which suggests the high interfacial interaction between matrix NR and minor BR phase with the assistance of LIR and t-CNTs, which act as a bridge connecting the interface of the binary phases. The interfacial interaction can drive the miscibility of the BR with the NR matrix.21

Figure 5

Dynamic mechanical analysis curves of C0L0, C3L0, C3L3, C3L5, C0L7 and C3L7 blends

Table 3

Dynamic mechanical analysis data from Fig. 5

Components	T_g NR/°C	T_g BR/°C	tan δ_max NR	tan δ_max BR	Loss modulus peak area of BR
C0L0	−46·8	−92·3	1·73	0·048	4692
C3L0	−46·0	−92·2	1·46	0·054	5007
C3L3	−46·9	−93·6	1·52	0·050	4646
C3L5	−48·5	−95·4	1·53	0·049	4369
C0L7	−48·8	−94·1	1·76	0·047	4553
C3L7	−46·7	−93·8	1·71	0·048	4341

Moreover, the loss modulus peak area corresponding to the minor phase can be employed to evaluate the minor phase morphology in the polymer blend.35 At a fixed proportion of the minor phase in the blend, a lower loss modulus peak area means an even particle size and a better dispersion of the minor phase in the matrix. The loss modulus peak area shifts to the lower value with an increase in the LIR content in the 80NR/20BR/22CB/3t-CNTs blends for samples C3L0 C3L3, C3L5 and C3L7, indicating the improved compatibilisation of the blend at a higher LIR content. Moreover, for samples C0L0 and C3L0 with equal proportion of BR, the loss modulus peak area of the BR rich phase increases with the addition of t-CNTs in the absence of LIR; however, it decreases in the presence of LIR when comparing samples C0L7 and C3L7. This should be attributed to the co-compatibilising effect between LIR and t-CNTs. As a plasticiser with a similar molecular structure to the binary rubber phase, LIR can reduce the Mooney viscosity of the rubber blend and improve the dispersion of t-CNTs in the blend, which is assumed to provide affluent opportunity for t-CNTs to localise in the interface. With a high aspect ratio and high specific surface energy, t-CNTs surround the minor phase to adsorb the molecular chains, and this originates a strong interfacial interaction to confine the movement of the molecular chain. Theoretically, in a three-component system, the free energy of mixing ΔG_m is given by26, 36 (1) where A and B denote the polymer pairs, and S is the compatibiliser. The system is thermodynamically stable (i.e. ΔG_m<0) in the presence of compatibiliser when the polymer blend is immiscible with a positive value of ΔG_AB (i.e.ΔG_AB>0 and ΔG_AS, ΔG_BS<0). S plays a role as a compatibiliser by adsorbing A and/or B polymers on its surface to get the stabilising energy. Evidently, with a large specific surface area per unit weight and the well dispersion within the binary blend in the presence of LIR, t-CNTs act as the compatibiliser by adsorbing NR and BR polymer chains on their surface. Therefore, the co-existence of t-CNTs and LIR exhibits high co-compatibilising efficiency, whereas the 80NR/20BR/25CB blends with and/or without LIR appear with almost no compatibilising effect.

Mechanical property

The mechanical properties of samples C3L0, C3L3 C3L5 and C3L7, including the tensile strength and elongation at break, are shown in Table 4. It shows that the increase in LIR content enhances the overall mechanical properties in the presence of 3 phr t-CNTs due to a good co-compatibilisation of t-CNTs and LIR. The tensile strengths of C0L0, C0L3, C0L5 and C0L7 without t-CNTs do not change too much with respect to the LIR content except the increase in the elongation at break, which is a normal phenomenon due to the plastic effect of LIR. Moreover, the tensile strength comparison under the same LIR loading levels between the ternary (80NR/20BR/25CB) and quadruple (80NR/20BR/22CB/3CNTs) blends shows that the quadruple blends exhibit better performance than the ternary except in the absence of LIR.

Table 4

Mechanical properties of rubber blends

Component	Tensile strength/MPa	Elongation at break/%
C0L0	22·4±1·3	423±16
C3L0	21·6±2·5	443±20
C0L3	22·9±0·7	529±22
C3L3	24·3±2·2	515±17
C0L5	22·1±2·0	510±25
C3L5	25·5±0·8	541±11
C0L7	23·3±2·2	508±22
C3L7	27·3±0·9	549±14

The explanation based on the phenomenon may be that with a high aspect ratio and high specific surface energy, t-CNTs tend to entangle and agglomerate in the rubber matrix and are even harder to achieve an even migration and distribution in the heterogeneous polymer blend, which lead to a poor performance. However, the addition of the plasticiser LIR, which has a similar molecular structure and good miscibility with binary blends, improves the processing and reduces the interfacial energy of the two phases, which are beneficial to the even migration and dispersion of t-CNTs in the binary blend.

Furthermore, polysulphide groups in t-CNTs are easily broken down to produce free radicals at a high temperature in the process of vulcanisation. The radicals are active enough to capture the hydrogen in the rubber molecular chains and turn into mercapto groups in the end of the t-CNTs.31 The mercapto groups can be easily bonded to diene rubber molecules via addition reaction, which is schematically illustrated in Fig. 6. This observation indicates that t-CNTs can play a dual role not only as a bridge connecting the interface between the phases to transfer load but also in providing stabilising energy to promote the miscibility of the immiscible blend.

Figure 6

Schematic illustration of possible chemical reaction between t-CNTs and rubber blend

Crack growth resistance

Tearing energy is considered as the driving force for crack propagation. It is defined as the energy released per unit area of crack surface growth37 (2) where T is the tearing energy, W is the total elastic energy stored in the specimen and A is the area of the one fracture surface of the crack. The subscript l denotes differentiation with constant displacement of the boundaries over which forces are applied. The tearing energy expression for a specimen with the edge notched geometry was given by Rivlin and Thomas37 (3) where w is the recoverable strain energy density per unit volume, c is the crack length and k is the strain dependent term. Lake38 proposed an approximate relation for k (4) where λ = 1+ϵ denotes the extension ratio, and ϵ = strain. Thus, T of the edge notched specimens under uniaxial stretching can be expressed as follows (5) Lake and Lindley39 identified four distinct regimes of fatigue crack growth behaviour based on the maximum energy release rate per cycle T for R = 0 (the R ratio is defined as the ratio of minimum/maximum deformation of rubber during cycles) in rubber. The power law relationship for regime 3 (Ref. 39) between dc/dn and T can be expressed as follows (6) Figure 7 shows the crack length versus fatigue cycle plots for the NR/BR blend of different components. The crack propagation rate dc/dn can be obtained by differentiating the trend line's equation in Fig. 7. The T values were determined from equation (5) at different crack lengths. Figure 8 displays the relationship between dc/dn and T on a logarithmic scale, which obeys a power law dependence. The fatigue parameters are listed in Table 5. Exponent B and constant A are the two crack growth parameters determined from Fig. 8. Exponent B is the slope of the line, and the logarithm of constant A represents the intercept of the line in the Y axis. The lower values of exponent B and constant A for the NR/BR/CB/t-CNTs composites denote a stronger resistance to crack growth at a given T value compared with NR/BR/CB blends. For the NR/BR/CB/CNT composites with various contents of LIR, the resistance to crack growth at a given T reaches the best for C3L3 and then becomes poor as the LIR content increases, but they are still better than NR/BR/CB/CNTs without LIR. Moreover, when comparing C0L0 with C0L7, the resistance to crack growth at a given T decreased with the addition of LIR in the absence of t-CNTs. However, the reverse trend was found when comparing the crack growth resistance of C3L0 with C3L7 as shown in Table 5. This also implies the co-compatibilising efficiency of LIR and t-CNTs in the NR/BR blend.

Figure 7

Crack length versus fatigue cycles for NR/BR blend

Figure 8

Crack growth rate dc/dn versus tearing energy

Table 5

Fatigue parameters of NR/BR blends

Components	A	B
C3L0	3·1	1·07
C3L3	0·90	0·99
C3L5	1·72	0·97
C3L7	2·73	0·96
C0L0	9·24	1·04
C0L7	11·9	1·02

Conclusions

The mechanical properties and crack growth resistance of vulcanised CB filled NR/BR blends were improved by the simultaneous addition of t-CNTs and LIR. It is the result of the co-compatibilisation effect of t-CNTs and LIR in the rubber blend. Scanning electron microscopy results show that the co-existence of t-CNTs and LIR in the NR/BR/CB blend gives a better interfacial adhesion than the other blends. Transmission electron microscopy analyses show that the dispersion of the minor phase in the matrix is improved, and the minor phase size significantly reduces with the co-existence of t-CNTs and LIR in the blend. The DMA reveals that the increase in LIR content in the NR/BR/CB/t-CNTs/LIR blends decreases the mobility of the BR molecular chains, and this reflects the strong interfacial interaction between the binary phases, which is the main cause for driving miscibility. Liquid isoprene rubber as a plasticiser can reduce the Mooney viscosity of the rubber blend, which benefits the dispersion of t-CNTs in the blend, which is assumed to provide opportunity for t-CNTs to interact with the binary rubber phase. This shows the co-compatibilising effect of CNTs and LIR in the NR/BR blend.

Footnotes

Acknowledgements

Financial support from the National Basic Research Program of China (973 Program, grant no. 2007CB714701) is acknowledged.

References

Bristow

, Metherell

: Rubber Ind. , 1973, 7, 252.

Mallik

, Shah

: ‘Indian rubber’, 31; 1983, Thane, Manufacturers Research Association.

Budrugeac

: Polym. Degrad. Stab. , 1992, 38, 165.

Bhagawan

, Tripathy

: Mater. Chem. Phys. , 1987, 17, 415.

Kim

, Kim

, Kang

, Choe

: Macromol. Res. , 2006, 14, 187.

Jurkowska

, Jurkowski

, Nadolny

: Eur. Polym. J. , 2006, 42, 1676.

Castro

, Suarez

, Nunes

: J. Appl. Polym. Sci. , 2004, 94, 1575.

Clarke

, Clarke

, Freakley

: Rubber Chem. Technol. , 2001, 74, 1.

Machado

, Lee

: Polym. Eng. Sci. , 1994, 34, 59.

10.

Kwei

, Frisch

, Vogel

: Macromolecules , 1977, 10, 157.

11.

Barentsen

, Heikens

, Piet

: Polymer , 1974, 15, 119.

12.

Fayt

, Teyssie

: J. Polym. Sci. B: Polym Phys. , 1981, 19B, 1269.

13.

Schwarz

, Barlow

, Paul

: J. Appl. Polym. Sci ., 1988, 35, 2053.

14.

Hlavata

, Horak

, Hromadkova

, Lednicky

, Pleska

: J. Polym. Sci. B: Polym. Phys. , 1999, 37B, 1647.

15.

Hlavata

, Horak

, Lednicky

, Hromadkova

, Pleska

, Zanevskii

: J. Polym. Sci. B: Polym. Phys. , 2001, 39B, 931.

16.

Radonjic

, Musil

, Smit

: J. Appl. Polym. Sci. , 1998, 69, 2625.

17.

Callan

, Hess

, Scott

: Rubber Chem. Technol. , 1971, 44, 814.

18.

Sircar

, Lamond

: Rubber Chem. Technol. , 1973, 46, 178.

19.

Clarke

, Clarke

, Sutherland

: Plast. Rubber Compos. , 2001, 30, 39.

20.

Liu

, Kontopoulou

: Polymer , 2006, 47, 7731.

21.

Zhang

, Yang

, Fu

: Polymer , 2004, 45, 1913.

22.

Yurekli

, Karim

, Krishnamoorti

: Macromolecules , 2003, 36, 7256.

23.

Yurekli

, Karim

, Krishnamoorti

: Macromolecules , 2004, 37, 507.

24.

Gelfer

, Song

, Liu

, Hsiao

, Chu

: J. Polym. Sci. B: Polym. Phys. , 2003, 41B, 44.

25.

Wang

, Zhang

, Fu

: Macromol. Rapid Commun. , 2003, 24, 231.

26.

Ray

, Pouliot

, Bousmina

, Utracki

: Polymer , 2004, 45, 8403.

27.

Ray

, Bousmina

: Macromol. Rapid Commun. , 2005, 26, 450.

28.

Cheng

, Sahool

, Juay

: J. Nanosci. Nanotechnol. , 2010, 10, 5242.

29.

, Zhang

, Yu

: Biomacromolecules , 2009, 10, 417.

30.

, Wang

, Su

: J. Appl. Polym. Sci. , 2010, 107, 4070.

31.

Yan

, Wu

, Zhan

, Xia

: Plast. Rubber Compos. , 2009, 38, 7.

32.

Schubel

, Gdoutos

, Daniel

: Theor. Appl. Fract. Mech. , 2004, 42, 149.

33.

Roberts

: in ‘The physics of glassy polymers’, (ed. , Haward

R N

); 1973, New York, Wiley.

34.

Perera

: Eur. Polym. J. , 2001, 37, 167.

35.

: in ‘The principle of polymer blending and modification’, to be published.

36.

Lipatov

: ‘Polymer reinforcement’; 1995, Toronto, Ont., ChemTech Publishing.

37.

Rivlin

, Thomas

: J. Polym. Sci. , 1953, 10, 291.

38.

Lake

: Rubber Chem. Technol. , 1995, 68, 435.

39.

Lake

, Lindley

: Rubber Chem. Technol. , 1966, 39, 348.