lable at ScienceDirect

Polymer 55 (2014) 5389e5395

Contents lists avai

Polymer

journal homepage: www.elsevier .com/locate/polymer

Mechanical properties of polybutadiene reinforced withoctadecylamine modified graphene oxide

Yan Zhang a, James E. Mark a, *, Yanwu Zhu b, Rodney S. Ruoff c, Dale W. Schaefer d, **

a Department of Chemistry, University of Cincinnati, Cincinnati, OH 45221-0172, USAb Department of Materials Science and Engineering and CAS Key Laboratory of Materials for Energy Conversion, University of Science and Technology ofChina, Hefei, Anhui 230026, Chinac Department of Mechanical Engineering and The Texas Materials Institute, University of Texas at Austin, Austin, TX 78712-0292, USAd Department of Biomedical, Chemical and Environmental Engineering, College of Engineering and Applied Science, University of Cincinnati, Cincinnati, OH45221-0012, USA

a r t i c l e i n f o

Article history:Received 26 June 2014Received in revised form18 August 2014Accepted 24 August 2014Available online 3 September 2014

Keywords:PolybutadieneGraphene oxideMechanical property

* Corresponding author. Tel.: þ1 513 556 9292; fax** Corresponding author. Tel.: þ1 513 556 5431; fax

E-mail addresses: markje@ucmail.uc.edu (J.E. Mschaefdw@ucmail.uc.edu (D.W. Schaefer).

http://dx.doi.org/10.1016/j.polymer.2014.08.0650032-3861/© 2014 Elsevier Ltd. All rights reserved.

a b s t r a c t

Octadecylamine-modified graphene-oxide (OMGO) polybutadiene nanocomposites with different OMGOloadings were prepared by solution mixing. The dispersion of OMGO in chloroform is greatly improvedcompared to GO. Toughness and elongation of PBDeOMGO nanocomposites increase by 332% and 191%respectively compared with pure PBD. However, Young's modulus of PBDeOMGO nanocomposite de-creases by 10% at 2-wt% loading. Graphene sheet crumpling accounts for the increased toughness, theabsence of modulus reinforcement and the absence of a Payne effect for PBDeOMGO. The oxidationsusceptibility of PBD is greatly reduced after the addition of OMGO, which is particularly desirable in thetire industry.

© 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Rubbers (elastomers) are an important class of commercialpolymers. Classic elastomers, such as polybutadiene and poly-isoprene are used as general purpose rubbers in high volumeproducts such as tires, hoses, belting, and flexible automotive parts[1]. Rubber materials have been extensively studied because of easyprocessing, flexibility and excellent thermal properties [2e4].However, rubber is commonly used in form of composites sincepure rubber lacks the required mechanical properties such as wearresistance and strength.

Rubber nanocomposites are the class of filled rubbers in whichat least one dimension of the fillers is on the nanometer scale. Mostcommonly used fillers are silica, carbon black, clay and carbonnanotubes [5e7]. Increased modulus is achieved at relative highfiller loading such as 20e50 per hundred rubbers (phr), which canreduce toughness due to defects caused by the fillers [8]. However,the incorporation of small-size fillers in cross-linked elastomersresults in specific nonlinear mechanical behaviors including Payne

: þ1 513 556 9239.: þ1 206 600 3191.ark), dale.schaefer@uc.edu,

effect and Mullins effect. The Payne effect is typically observed atsmall strain. The dynamic storage modulus decreases strongly withincreasing strain amplitude [9]. Mullins et al. first reported that thedegree of softening increases with increasing stiffening ability ofthe fillers [10]. Some Mullins softening is observed in carbon blackand silica filled rubber composites systems [11].

Polybutadiene (PBD), a synthetic rubber, has a higher resistanceto wear over styrene-butadiene rubber and natural rubber, whichare its main competitors in rubber-industry applications due totheir lower glass transition temperatures [12]. PBD is a low costrubber used for soles, gasket, seals and belts [13]. PBD is normallyformulated with fillers, such as silica or carbon black.

Graphene is an emerging filler candidate that has been widelystudied in thermoplastics, but not in elastomers. Graphene showshigh thermal conductivity (5000 W m�1 K�1) [14e16], highestYoung's modulus ever measured (1 TPa) [17] and large theoreticalsurface area (2675 m2 g�1) [18]. High modulus and large surfacearea promise dramatic improvement in mechanical properties,which as yet has not been realized. Within the few published paperon elastomers filed with graphene materials, Araby et al. [19] re-ported that tensile strength of styrene butadiene rubber filled withgraphene increases by 230% using melt compounding. However, inorder to obtain such improvement, a large amount of graphene(24%) was incorporated into rubber, which causes defects inproducts and increases cost.

Y. Zhang et al. / Polymer 55 (2014) 5389e53955390

An effective approach to achieve graphene-based polymercomposites is based on chemical transformation of graphite tographite-oxide (GO), which readily disperses in water and exfoli-ates to form individual, single-layer graphene oxide sheets[20e23]. Further modification of GO is necessary to achieve fullydispersed GO in common organic solvents [24e28]. Here wedemonstrate a modification scheme based on octadecylamine(ODA) modification of GO that greatly improves GO dispersion inchloroform.

We report the synthesis of ODA-modified GO (OMGO) andexamine the mechanical properties of PBDeOMGO nano-composites as a function of OMGO loading. The amino group in theODA modifier reacts with carboxylic acid groups in GO. In contrastto the preponderance of graphene-based thermoplastic compos-ites, we investigate the performance of thermosetting rubber ma-terials. Toughness and elongation at break of PBDeOMGO improveby 332% and 343% respectively at 2-wt% OMGO. However, Young'smodulus of PBDeOMGO decreases by 10% at OMGO loading 2 wt%.

2. Experimental section

2.1. Materials

GO, prepared by a modified Hummers method [20]. ODA and N,N0-diisopropylcarbodiimide (DIC) were purchased from Sigma-eAldrich Co. Chloroform was purchased from Tedia Company Inc.Anhydrous acetonitrile was purchased from Acros Organics. Highcis-1, 4-polybutadiene (c-PBD) was purchased from SigmaeAldrichCo. Dibenzoyl peroxide (BPO) was purchased from Acros Organics.All reagents were used as received.

2.2. Preparation of OMGO

ODA-modified GO (OMGO) was made using the reaction be-tween carboxylic acid groups and epoxy groups fromGO and aminogroups (using DIC to activate carboxylic groups), as shown inScheme 1. The selectivity of carboxylic acid groups and epoxygroups remains an open question [29,30]. Briefly, the desiredamount of GO was dispersed in 40 ml anhydrous acetonitrile fol-lowed by ultrasonication for 1 h. DIC was added into grapheneoxide dispersion followed by reaction for 4 h at 70 �C to activate the

Scheme 1. Synthesis of OMGO with octadecylamine.

carboxylic acid groups. Then ODA was added into the mixture fol-lowed by refluxing at 75 �C overnight under nitrogen atmosphere.After the reaction, the OMGO was purified by washing withdimethyl formamide and acetone successively to remove residualDIC and unreacted amines. Black, solid OMGO was obtained aftervacuum drying at 50 �C overnight.

2.3. Preparation of the unfilled cis-PBD networks

The desired amounts of polymer were first dissolved in chlo-roform. After a clear solution was obtained, 2-wt% BPO was added[31]. The solution was stirred at room temperature for 2 h and thentransferred into Teflon™ dishes that were covered with aluminumfoil for overnight solvent evaporation. Films were pressed atapproximately 1.2 � 104 psi, 130 �C for 2 h. The final samples wereapproximately 1.0 mm thick.

2.4. Preparation of PBDeOMGO nanocomposites

The PBDeOMGO nanocomposites were prepared with variousloadings of OMGO. Firstly OMGO was dispersed in chloroformwiththe aid of ultra-sonication for 1 h to yield a well-dispersed solution.Secondly, thewell-dispersed OMGO solutionwasmixedwith c-PBDwith 0.50, 1.00 and 2.00 weight ratio for 2 h following the sameprocedures as for unfilled c-PBD discussed above.

2.5. Characterization

Fourier transform infrared spectroscopy (FTIR) recorded on aNicolet 6700 (Thermo Scientific) spectrometer was used to char-acterize the chemical structure of GO and OMGO. Samples weremeasured under a mechanical force by pressing the un-exfoliatedsample's surface against a diamond window. FTIR spectra werecollected in the range 4000e450 cm�1.

Thermo-gravimetric analyses (TGA) were done on the GO andOMGO powders using a TA Q500 instrument (TA Instruments)under a nitrogen atmosphere with a heating rate of 5 �C min�1.Degradation temperatures of PBDeOMGO nanocomposites weremeasured using NETZSCH STA 409 instrument under a purge flowof 20ml/min argon at heating rate of 20 �Cmin�1 from 25 to 700 �C.

Powder X-ray diffraction (XRD) measurements were carried outusing a PANalytical X'Pert Pro MPD diffractometer with Cu Ka ra-diation (l ¼ 1.541 Å) at 45 kV and 40 mA. The diffraction angle wasincreased from 5� to 35� with the scanning rate of 0.05�min�1.

Scanning electron microscopy (SEM) was conducted with an FEIPhillips Electroscan XL30 ESEMeFEG microscope using an accel-eration voltage of 15 kV. OMGO suspensions (0.03 mg ml�1) werespin-coated onto a flat aluminum plate at 2000 r.p.m. for 30 s. Thenthe OMGO-coated aluminum plate was mounted on a standardspecimen holder using a double-sided carbon conductive tape forSEM imaging. A transmission electron microscopy (TEM) samplewas prepared by placing a few drops of dispersion onto a laceycarbon film support on a Cu grid. Images were acquired in a JEOL1230 transmission electron microscope operated at 80 kV.

Atomic force microscopy (AFM) images were obtained using aDimension 3100 AFM made by Veeco Instruments Inc., operated intapping mode using Veeco RTESP type silicon cantilevers with aresonance of frequency of 360 kHz. The samples for AFM mea-surements were prepared by ultrasonic treatment of OMGO inchloroform for 1 h, followed by spin-coating OMGO suspensions inchloroform (0.03 mg ml�1) on freshly cleaved mica surfaces at2000 r.p.m. for 30 s and then drying under vacuum at roomtemperature.

The tensile properties were measured using an ESM-301(MARK-10) tensile tester. The experiments were carried out at

Y. Zhang et al. / Polymer 55 (2014) 5389e5395 5391

room temperature at a crosshead speed of 50 mm/min usingdumbbell-shaped specimens with an original length of 40 mm.Three dumbbell-shaped samples were measured and standarddeviation of three measurements is the error bar.

MettlereToledo DMA-861 was used to measure the viscoelasticdata in order to investigate the Payne effect of nanocomposites. Thetests were performed on rectangular samples in shear mode at 5 Hzat different shear amplitude.

Fig. 2. TGA curves of GO and OMGO. GO shows 37% more weight loss compared toOMGO between 100 and 280 �C, indicating COOH groups have been converted afterODA modification.

3. Results and discussions

FTIR data shows that ODA is linked to the GO by the reactionwith carboxylic acid group. Fig. 1 shows the FTIR spectra of GO andOMGO. In the spectrum of GO, broad peak between 3550 and2500 cm�1 (OeH stretching from COOH and OH groups), C]O(1714 cm�1), and aromatic C]C (1595 cm�1) stretches wereobserved. After modification with ODA, there is a peak between3500 and 3000 cm�1, which is attributed to OeH stretching fromOH and two new peaks appear at around 2914 cm�1 and 2839 cm�1,which are assigned to the CeH stretch in themethylene group fromODA [32]. The peak at 1465 cm�1 is due to the asymmetricalbending vibration of methyl group from ODA. The reaction of car-boxylic acid groups with amine groups in the ODA is confirmed bythe observed decrease inwavenumber for the peak of C]O stretch,from 1714 to 1570 cm�1, which overlaps with C]C peak. Since theintensity of peak is proportional to number of functional groups inthe system, the peak at 1570 cm�1 is relatively weak.

TGA was utilized to monitor the decomposition of functionalgroups. Fig. 2 shows the typical TGA curves of GO and OMGO. In theGO curve, relatively low thermal stability was observed. GO startslosing mass below 100 �C (about 7%), which is attributed to theabsorbed water. The 46% weight loss between 100 �C and 280 �C isdue to decomposition of the labile oxygen-containing functionalgroups. Between 280 �C and 600 �C, no obvious weight loss wasobserved indicating all of the oxygen functional groups havedecomposed below 280 �C. The decrease in mass around 600 �C isdue to the pyrolysis of the carbon skeleton of the GO. After modi-fication with ODA, no weight loss is observed in OMGO below100 �C and the weight loss between 100 and 280 �C is 9 wt% whichis 37 wt% lower than that of GO (46 wt%). The 37 wt% difference is

Fig. 1. FTeIR of GO and OMGO. The peaks at 2914 cm�1 and 2829 cm�1 appear aftermodification with ODA assigned to CeH stretch in the methylene group from ODA,which indicates ODA was attached to GO.

attributed to ODA modifier, which degrades above 280 �C. Weconclude that 37 wt% of oxygen functional groups in GO have beenmodified by ODA [33,34]. This value is consistent with the weightloss (38 ± 2 wt%) between 280 �C and 550 �C, which is assigned tothe removal of alkyl chains. The consistency of these weight lossesconfirms the mass loss between 100 and 280 �C is due to ODAmodifier.

XRD was used to determine the morphology and the layerspacing of dry GO and OMGO. Fig. 3 shows a diffraction peak at2q ¼ 10.3� indicating an interlayer distance of 0.86 nm for GO. Aftermodifying GO with ODA, the 10.3� peak almost disappears. How-ever, a broad peak appears with comparable intensity at 22�, whichis closer to the typical (002) diffraction peak of graphite (26.6�),corresponding to 0.40 nm of interlayer distance. The same 22� peakis reported in the work of Dubin et al. [35] and Pei et al. [36]Compared with diffraction peak of graphite, which is extremely

Fig. 3. XRD pattern of GO and OMGO. After modifying GO with ODA, the peak at 10.3�

disappears. The broad peak of OMGO between 15� and 30� indicates that OMGO sheetsare disordered stacks.

Fig. 4. (A): SEM image of OMGO indicating OMGO shows a layered structure; (B): TEM image of OMGO. Extremely thin layers were observed indicating that OMGO was fullydispersed in chloroform.

Fig. 5. AFM images and height profiles of OMGO dispersed at a concentration of 0.03 mg ml�1 in chloroform. 5.0 mm � 5.0 mm scan area was selected to measure the thickness ofOMGO, which is 0.9 nm.

Fig. 6. TGA of PBDeOMGO nanocomposite with different filler loading. At 2 wt%OMGO, the onset temperature of the PBDeOMGO composites is 5 �C higher than purePBD.

Y. Zhang et al. / Polymer 55 (2014) 5389e53955392

sharp and intense, the broad peak of OMGO between 15� and 30�

indicates that OMGO sheets are disordered stacking. Under soni-cation OMGO sheets easily separate into single layers as shown inFig. 4B.

SEM and TEMmeasurements were performed to investigate themorphology of OMGO. Fig. 4A shows the SEM of OMGO at 500 nmscale indicating that OMGO displays a layered structure. The OMGOobservedwith TEM (Fig. 4B) is in the form of flat sheet. In the darkerpart on the right corner is carbon grid. Most of the area on the left islight gray color, which is covered by very thin layer of dispersedOMGO. Fully exfoliated OMGO sheets were achieved after ODAmodification. Some areas are darker, which indicates stacking of theOMGO sheets.

AFM was used to determine the thickness of dispersed OMGO(Fig. 5). The cross-sectional view of the AFM image of OMGO in-dicates that the thickness of OMGO is 0.9 nm. The typical observedmonolayer GO sheet is 0.8 nm, which is larger than the theoreticalgraphite sheet with van der Waals thickness of ca. 0.34 nm becauseof the presence of oxygen functional group above and below the GOplane. So our result indicates that the OMGO platelet is a singlelayer, which can also be seen from TEM image.

TGA was employed to examine the degradation of PBDeOMGOnanocomposites. Fig. 6 is a composite plot of the TGA mass losscurves obtained from the 0, 0.5, 1 and 2 wt% OMGO filled systems.The introduction of OMGO into PBD increases the degradation

Table 1Degradation temperature of PBDeOMGO nanocomposites from Fig. 6.

PBDeOMGO nanocomposites

Loading Pure PBD 0.5 phr 1.0 phr 2.0 phrTd (�C) 380 381 385 385

Fig. 8. Storage modulus (E0) aWs a function of shear amplitude at 5 Hz. The data showalmost no Payne effect (reduction of modulus with strain amplitude). These dataindicate that the crumpled OMGO has a lower modulus than the matrix.

Y. Zhang et al. / Polymer 55 (2014) 5389e5395 5393

temperature (Fig. 6). The derivative of each mass loss curve definesthe degradation onset temperature as shown in Table 1. At 2 wt%OMGO, the onset temperature of the PBDeOMGO composites is5 �C higher than pure PBD.

Different aspects of the mechanical properties of PBDeOMGOnanocomposites were examined. Toughness greatly improveswhile the Young's modulus of PBDeOMGO nanocomposites de-creases slightly. Fig. 7A shows the stressestrain curves ofPBDeOMGO nanocomposites as a function of filler loading. Theareas under the curves are the toughnesses (Fig. 7B) and themaximum extension reflects the elongation of composites (Fig. 7C).The toughness and elongation increase 332% and 191% respectivelyat filler loading of 2.0 wt%. The slope of the curves in the initialsmall strain portion (Fig. 7D) is Young's modulus. The Young'smodulus decreases, which indicates the stiffness of the materialsdoes not increase by adding OMGO as filler. The fact that Young'smodulus decreases and toughness increases by the introduction ofOMGO is due to single-sheet nature of OMGO, which adopts awrinkled morphology that imparts a measure of entropic, rubber-like elasticity to the filler itself [37,38]. That is, because thebending modulus of the sheets is so low, adding filler does notenhance the modulus. In fact, the filler may interfere with cross-linking chemistry leading to decreased modulus. On the otherhand, stretching of the crumpled OMGO sheet adds an extra

Fig. 7. A. Stressestrain curves of PBDeOMGO composites with different filler loading. Fig.PBDeOMGO nanocomposites vs. filler loading. Fig. 7D. Young's modulus of PBDeOMGO co

reinforcement mechanism at large elongations, which improvestoughness and elongation at break.

The samples were tested using dynamic mechanical analysis(DMA) to elucidate the cause of the low Young's modulus. A typicalDMA result on PBDeOMGO nanocomposites deformed in shearmode is shown in Fig. 8. Pure PBD has a larger elastic modulus than

7B. Toughness of PBDeOMGO naocomposites vs filler contents Fig. 7C. Elongation ofmposites vs. filler loading.

Fig. 9. Comparison of the Payne effect for PBDeOMGO and CNFePU. The storagemodulusfor PBDeOMGO is calculated from shearmodulus in Fig. 8 assuming a Poisson ration of 0.5.CNF reinforces PU (and shows a Payne effect) even though PU matrix has a higher zero-strain modulus than PBD. PBD shows neither reinforcement nor a Payne effect.

Fig. 10. (A) Cross-section SEM image of pure PBD at the fracture surface; (B) Cross-section Sobserved in the PBD matrix, which indicates that OMGO is fully exfoliated.

Fig. 11. (A): Pure PBD; (B): 0.5 wt% OMGO; (C): 1 w

Y. Zhang et al. / Polymer 55 (2014) 5389e53955394

PBDeOMGO. The modulus decreases as filler loading increases andreaches minimum at 1.0 wt%, which matches the tensile result(Fig. 7D).Within error, there is no observable Payne effect (decreaseof modulus with strain amplitude).

The Payne effect is necessarily absent in PBDeOMGO becauseOMGO does not reinforce PBD. Fig. 9, compares the strain-dependence of the storage modulus of PBDeOMGO with that ofcarbon-nanofiber (CNF) reinforced polyurethane (PU) [38], whichdoes show a Payne effect. The latter reinforces PU at low strain, butat strain amplitude of ~10%, the modulus of the matrix is recovered.For OMGO the modulus of the matrix is already “recovered” at zerostrain so it cannot display a Payne effect.

The Payne effect is generally attributed to clustering of the fillerparticles induced by van der Waals forces [39]. Clustering is dis-rupted by strain, leading to reduced modulus enhancement and, athigh strain, no modulus enhancement (the Payne effect). Appar-ently the crumpled morphology of OMGO precludes the formationof deck-of-cards configurations required for strong van der Waalsattractions. In the case of CNFs, on the other hand, side-by-sidefiber configurations are present, at least over the persistencelength of the fibers [37,38], which leads to the clustering requiredfor both zero-strain modulus enhancement and the Payne effect.

After the tensile test, the fracture surfaces were examined bySEM. Fig. 10A and B shows the cross-section SEM images at fracturesurface of pure PBD and PBDeOMGO with 0.5 wt% OMGO. No

EM image of PBDeOMGO with 0.5 wt% OMGO at the fracture surface. No aggregation is

t% OMGO; (D): 2 wt% OMGO after 12 months.

Y. Zhang et al. / Polymer 55 (2014) 5389e5395 5395

aggregation is observed in the PBD matrix after the addition ofOMGO, which indicates that OMGO is fully exfoliated.

The degradation of PBDeOMGOnanocomposites was observed attimes up to 12 months after sample prepared. Fig. 11 compares purePBDwithPBDfilledwithOMGOat threedifferentpercentages after12months. PBD has a double bond in the backbone, which is easilyoxidizedbyozonewhenexposed to air [40]. After 12month, pure PBDbecomes yellow, brittle and forms cracks. The addition of OMGOgreatly reduces oxidation. As the OMGO loading increases,PBDeOMGOnanocomposites remainflexible.Nocracking isobservedfor 1 wt% and 2 wt% OMGO. The absence of oxidation in PBDeOMGOnanocompositesmight be due to thefiller-polymer attraction. OMGOhasmanyC]Cmoities,whichcanformpep interactionswithC]C inPBD structure. This observation is very important to industry sincePBD is the most important rubber component in the tire industry.

4. Conclusions

GO was successfully modified by ODA. ODA attaches to surfaceby reaction of surface carboxylic acid groups and epoxy groups onGOwith the amine group on ODA. The introduction of ODA onto GOgreatly improves exfoliation in chloroform and facilitates casting ofPBD composites. OMGO in the PBDeOMGO nanocomposites acts asa spring, which stores energy leading to the 332% improvement intoughness and 191% improvement in elongation. However, Young'smodulus of PBDeOMGO nanocomposites decreases, indicating thecrumpled sheets have a lower modulus than the rubber itself. Sincethere is nomodulus enhancement there is also no filler-induced thePayne effect. The oxidation of PBD decreases after the addition ofOMGO.

Acknowledgments

This work was supported by the National Science Foundationthrough Grant DMR-0803454 (Polymers Program, Division of Mate-rials Research). We thank the Advanced Materials CharacterizingCenter (AMCC) at University of Cincinnati for providing the in-struments and Professor Vesselin Shanov's help on tensilemeasurements.

References

[1] El Fray M, Goettler LA. Application of rubber nanocomposites. In: Rubbernanocomposites. John Wiley & Sons, Ltd; 2010. p. 675e96.

[2] Wang J, Vincent J, Quarles CA. Review of positron annihilation spectroscopystudies of rubber with carbon black filler. Nucl InstrumMethods Phys Res SectB 2005;241(1e4):271e5.

[3] Chen G, Zhao W. Rubber/graphite nanocomposites. In: Rubber nano-composites. John Wiley & Sons, Ltd; 2010. p. 527e50.

[4] Ma J, Zhang L-Q, Geng L. Manufacturing techniques of rubber nanocomposites.In: Rubber nanocomposites. John Wiley & Sons, Ltd; 2010. p. 21e62.

[5] Rohini Thimmaiah S, Siddaramaiah. Investigation of carbon black and meta-kaolin cofillers content on mechanical and thermal behaviors of naturalrubber compounds. J Elastomers Plast 2013;45(2):187e98.

[6] Song M, Wong C, Jin J, Ansarifar A, Zhang Z, Richardson M. Preparation andcharacterization of poly(styrene-co-butadiene) and polybutadiene rubber/claynanocomposites. Polym Int 2005;54(3):560e8.

[7] Zhou D, Mark JE. Preparation and characterization of trans-1,4-polybutadienenanocomposites containing in situ generated silica. J Macromol Sci Pure ApplChem 2004;41(11):1221e32.

[8] Patwardhan S, Taori V, Hassan M, Agashe N, Franklin J, Beaucage G. Aninvestigation of the properties of poly(dimethylsiloxane)-bioinspired silicahybrids. Eur Polym J 2006;42(1):167e78.

[9] Chazeau L, Gauthier C, Chenal JM. Mechanical properties of rubber nano-composites: how, why … and then?. In: Rubber nanocomposites. John Wiley& Sons, Ltd; 2010. p. 291e330.

[10] Mullins L. Effect of stretching on the properties of rubber. Rubber ChemTechnol 1948;21(2):281e300.

[11] Harwood JAC, Payne AR. Stress softening in natural rubber vulcanizates. PartIII. Carbon black-filled vulcanizates. J Appl Polym Sci 1966;10(2):315e24.

[12] Brydson JA. Rubbery materials and their compounds, vol. 401. London:Elsevier Applied Science; 1988.

[13] Hofmann W. Rubber technology handbook. Hanser; 1989.[14] Balandin AA, Ghosh S, Bao W, Calizo I, Teweldebrhan D, Miao F, et al. Superior

Thermal conductivity of single-layer graphene. Nano Lett 2008;8(3):902e7.[15] Cai W, Moore AL, Zhu Y, Li X, Chen S, Shi L, et al. Thermal transport in sus-

pended and supported monolayer graphene grown by chemical vapordeposition. Nano Lett 2010;10(5):1645e51.

[16] Seol JH, Jo I, Moore AL, Lindsay L, Aitken ZH, Pettes MT, et al. Two-dimensionalphonon transport in supported graphene. Science 2010;328(5975):213e6.

[17] Jang B, Zhamu A. Processing of nanographene platelets (NGPs) and NGPnanocomposites: a review. J Mater Sci 2008;43(15):5092e101.

[18] Huang Y, Liang J, Chen Y. An overview of the applications of graphene-basedmaterials in supercapacitors. Small 2012;8(12):1805e34.

[19] Araby S, Zhang L, Kuan H-C, Dai J-B, Majewski P, Ma J. A novel approach toelectrically and thermally conductive elastomers using graphene. Polymer2013;54(14):3663e70.

[20] Stankovich S, Dikin DA, Ruoff RS. Graphene-based composite materials. Na-ture 2006;442(7100):282e6.

[21] Paredes JI, Villar-Rodil S, Martinez-Alonso A, Tascon JMD. Graphene oxidedispersions in organic solvents. Langmuir 2008;24(19):10560e4.

[22] Park S, Bielawski C, Ruoff R. The chemistry of graphene oxide. Chem Soc Rev2010;39(1):228e40.

[23] Haubner K, Murawski J, Olk P, Eng LM, Ziegler C, Adolphi B, et al. The route tofunctional graphene oxide. ChemPhysChem 2010;11(10):2131e9.

[24] Park S, An J, Jung I, Piner RD, An SJ, Li X, et al. Colloidal suspensions of highlyreduced graphene oxide in a wide variety of organic solvents. Nano Lett2009;9(4):1593e7.

[25] Yang H, Li F, Shan C, Han D, Zhang Q, Niu L. Covalent functionalization ofchemically converted graphene sheets via silane and its reinforcement.J Mater Chem 2009;19(26):4632e8.

[26] Chattopadhyay J, Mukherjee A, Chakraborty S, Kang J, Loos PJ, Kelly KF, et al.Exfoliated soluble graphite. Carbon 2009;47(13):2945e9.

[27] Shen J, Li N, Shi M, Hu Y, Ye M. Covalent synthesis of organophilic chemicallyfunctionalized graphene sheets. J Colloid Interface Sci 2010;348(2):377e83.

[28] Zhang Y, Zhu Y, Lin G, Ruoff RS, Hu N, Schaefer DW, et al. What factors controlthe mechanical properties of poly (dimethylsiloxane) reinforced with nano-sheets of 3-aminopropyltriethoxysilane modified graphene oxide? Polymer2013;54(14):3605e11.

[29] Monagle JJ. Carbodiimides. III. Conversion of isocyanates to carbodiimides.Catalyst studies. J Org Chem 1962;27(11):3851e5.

[30] Dreyer DR, Park S, Bielawski CW, Ruoff RS. The chemistry of graphene oxide.Chem Soc Rev 2010;39(1):228e40.

[31] Huang N-J, Sundberg DC. Fundamental studies of grafting reactions in freeradical copolymerization. III. Grafting of styrene, acrylate, and methacrylatemonomers onto cis-polybutadiene using benzoyl peroxide initiator in so-lution polymerization. J Polym Sci Part A: Polym Chem 1995;33(15):2571e86.

[32] Yuan W, Zhao H, Hu H, Wang S, Baker GL. Synthesis and characterization ofthe hole-conducting silica/polymer nanocomposites and application in solid-state dye-sensitized solar cell. ACS Appl Mater Interfaces 2013;5(10):4155e61.

[33] Yang H, Shan C, Li F, Han D, Zhang Q, Niu L. Covalent functionalization ofpolydisperse chemically-converted graphene sheets with amine-terminatedionic liquid. Chem Commun 2009;26:3880e2.

[34] Hu H, Yuan W, Zhao H, Baker GL. A novel polymer gel electrolyte: directpolymerization of ionic liquid from surface of silica nanoparticles. J Polym SciPart A: Polym Chem 2014;52(1):121e7.

[35] Dubin S, Gilje S, Wang K, Tung VC, Cha K, Hall AS, et al. A one-step, sol-vothermal reduction method for producing reduced graphene oxide disper-sions in organic solvents. ACS Nano 2010;4(7):3845e52.

[36] Pei S, Zhao J, Du J, Ren W, Cheng H-M. Direct reduction of graphene oxidefilms into highly conductive and flexible graphene films by hydrohalic acids.Carbon 2010;48(15):4466e74.

[37] Schaefer DW, Justice RS. How nano are nanocomposites? Macromolecules2007;40:8501e17.

[38] Schaefer DW, Zhao J, Dowty H, Alexander M, Orler EB. Carbon nanofibrereinforcement of soft materials. Soft Matter 2008;4(10):2071.

[39] Wang MJ. Effect of polymer-filler and filler-filler interactions on dynamicproperties of filled vulcanizates. Rubber Chem Technol 1998;71(3):520e89.

[40] Cataldo F. The action of ozone on polymers having unconjugated and cross- orlinearly conjugated unsaturation: chemistry and technological aspects. PolymDegrad Stab 2001;73(3):511e20.