at SciVerse ScienceDirect

Polymer 54 (2013) 3605e3611

Contents lists available

Polymer

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

What factors control the mechanical properties of poly(dimethylsiloxane) reinforced with nanosheets of 3-aminopropyltriethoxysilane modified graphene oxide?

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

aDepartment of Chemistry, University of Cincinnati, Cincinnati, OH 45221-0172, USAbDepartment of Materials Science and Engineering and CAS Key Lab of Materials for Energy Conversion, University of Science and Technology of China,Hefei, Anhui 230026, ChinacDepartment of Mechanical Engineering and the Texas Materials Institute, University of Texas at Austin, Austin, TX 78712-0292, USAd School of Energy, Environment, Biological and Medical Engineering, College of Engineering and Applied Science, University of Cincinnati, Cincinnati,OH 45221-0012, USA

a r t i c l e i n f o

Article history:Received 17 December 2012Received in revised form22 April 2013Accepted 26 April 2013Available online 3 May 2013

Keywords:CompositeGraphene oxideMechanical

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

E-mail addresses: dale.schaefer@uc.edu (D.W. Sch(J.E. Mark).

0032-3861/$ e see front matter � 2013 Elsevier Ltd.http://dx.doi.org/10.1016/j.polymer.2013.04.057

a b s t r a c t

Graphite oxide (GO) was modified using 3-aminopropyltriethoxysilane (APTES) to investigate the impactof dispersion and interfacial bonding on the mechanical properties of reinforced silicone elastomer, poly(dimethylsiloxane), PDMS. Although a 71% enhancement of Young’s modulus was observed (comparableto thermoplastics) at a loading of 3.0 wt% APTES-modified GO (A-GO), the observed enhancement factoris just 8% of that expected for randomly dispersed perfect graphene sheets. We attribute this less-than-ideal enhancement to crumping and break-up of the graphene sheets caused by ultrasonic dispersion. Inthe absence of TEOS crosslinker, the A-GO PDMS melt does not solidify so modulus enhancement is notdue to crosslinking through the APTES-modified filler particles. Modulus enhancement due to rubber-like elasticity of the filler itself, however, may be active. Characterization indicates that amino groupsare chemically attached to the surface of GO in A-GO. The improved dispersibility enhances thetoughness, but only at low loadings. In fact, at 3-wt% loading the toughness is less than that of pure PDMSdue to loss of elongation caused by filler aggregation.

� 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Graphene, a monolayer of carbon atoms arranged into a two-dimensional honeycomb lattice, has become the focus of intensiveresearch in the last few years [1e11]. The outstanding properties ofgraphene include high thermal conductivity (5000 Wm�1 K�1)[12e14]; highest Young’s modulus ever measured (1 TPa) [8], largetheoretical surface area (2675 m2 g�1) [15] and high mobility ofcharge carriers at room temperature (200,000 cm2 V�1 s�1) [16].All these properties make graphene promising for many applica-tions, such as polymer nanocomposites [2,4,7,8,17e19], ultra-capacitors [20] and solar cells [21,22], among others.

The use of graphene as nanofiller in a polymer matrix is of in-terest since graphene has excellent mechanical properties, elec-trical properties and a large theoretical surface area. An effective

: þ1 206 600 3191.: þ1 513 556 9239.aefer), markje@ucmail.uc.edu

All rights reserved.

approach to achieve graphene-based polymer composites is basedon chemical transformation of graphite to graphite oxide (GO),which readily disperses in water and exfoliates to form individual,single-layer graphene oxide sheets [2,23e25]. Engineering poly-mers, however, typically are not soluble in water. Further modifi-cation of GO is necessary to achieve fully dispersed GO in commonorganic solvents [26e29]. We use a modification scheme based on3-aminopropyltriethoxysilane (APTES) modification of GO.

The literature is replete with evidence of superior properties ofgraphene-based polymer composites. For example, poly (vinylalcohol) (PVA)/GO nanocomposites were made by a water solutionprocessingmethod and a 62% improvement of Young’smoduluswasachieved at 0.7 wt% of GO [30]. Although the mechanical propertiesof PVA nanocomposites were greatly improved, the modulus dataimply that the aspect ratio is about 128, which is inconsistent withmorphology of the sheets, whose aspect ratio is around 3000.Similar inconsistency was found in the work by Shi et al. [31].

Although most literature papers do not provide thestructural data needed to evaluate the true performance of

Y. Zhang et al. / Polymer 54 (2013) 3605e36113606

composites, compilations of modulus enhancement data [32]imply that graphene-based composites perform substantiallybelow expectations based on idealized flat sheets. Schaefer andJustice attribute the reduced performance to crumpling of thefiller, which leads to a reduced effective aspect ratio [33]. Based onour TEM data, we present evidence that this mechanism is activein our elastomer composites.

In contrast to the preponderance of graphene-based thermo-plastic composites, we study reinforcement of a silicone rubber.Reinforcement of soft materials is more promising than hard ma-terials because, due to the crumpling effect discussed above, someelastic energy is stored as rubber-like elasticity in the filler itself[33]. That is, the filler itself can behave as an elastomer.

Silicone, with repeat unit [eSiRR0Oe], is one of the mostimportant non-carbon backbone polymers. Due to many unusualproperties, such as high thermal stability, unusually large perme-ability, low surface tension and good water resistance [34,35] sili-cone polymers have application in areas ranging from artificialorgans and facial reconstruction in medical applications to high-temperature lubricants [36,37]. Poly(dimethylsiloxane) (PDMS),[eSi (CH3)2Oe] was selected due to its extremely low glass tran-sition temperature (Tg, �125 �C) and high thermal stability [35]. Inthe unfilled state, however, PDMS has poor mechanical strength e

hence the need for reinforcement.In this paper, we report the first synthesis of APTES-modified GO

(A-GO). The amino group in the APTES modifier reacts with car-boxylic acid groups in GO. There is only one other report of APTESmodification in the literature [38]. That report examines a differentform of carbon (reduced graphene oxide) in a thermoset. We drawdifferent conclusions regarding nature of both the APTESecarbonand the APTESepolymer linkages in A-GO.

The origin of the beneficial effects of APTES modification is notobvious. APTES could crosslink the PDMS chains through itspendant ethoxy groups. APTESmight also improve linkage between

Scheme 1. Synthesis of

the filler and the polymer. Finally, APTES might simply improvethe dispersion of the filler in the polymer. By examining the me-chanical properties with and without TEOS crosslinker, we showthat the modulus enhancement is a physical rather than a chemicaleffect. We conclude that the enhanced dispersion is the dominantmechanism.

2. Experimental

Materials. Hydroxyl-terminated PDMS (Mn ¼ 15,000 g/mol),tetraethyl orthosilicate (TEOS) and stannous-2-ethyl-hexanoate(SNB1100) were obtained from Gelest Inc. APTES and N, N0-diiso-propylcarbodiimide (DIC) were obtained from Aldrich Company,Inc. and chloroform was purchased from the Tedia Company, Inc.GO was prepared according to the modified Hummers Method[2,39]

Preparation of A-GO. APTES-modified GO (A-GO) was madeusing the reaction between carboxylic acid groups and aminogroups (using DIC to activate amino groups) [40], as shown inScheme 1. This synthesis was carried out at 70 �C for 4 hanhy-drously under nitrogen to prevent the hydrolysis of the alkoxylgroups in the silane [41]. After the reaction, the A-GO was purifiedby washing with dimethyl formamide and acetone successively.Black solid A-GO was obtained after drying in the vacuum at 50 �Covernight.

Preparation of PDMS/A-GO composites. A-GO was dispersed inchloroform to generate a homogenous dispersion after sonicationfor 1 h. PDMS/A-GO composites were then prepared with variousloadings of A-GO pre-dispersed in chloroform. Specifically 0.50,1.00, 2.00 and 3.00 weight percent relative to the weight of PDMS,which was also dissolved in chloroform. TEOS was used as a cross-linking reagent (TEOS: PDMS ¼ 1:2 M ratio in order to assure thehydroxy-terminated PDMS fully cross-linked). The cross-linkingreaction [34] was catalyzed by 0.5 wt% SNB1100 based on the

A-GO with APTES.

Fig. 2. TGA curves of GO and A-GO. GO shows 16% more weight loss compared to A-GObetween 100 and 280 �C, indicating COOH groups have been converted after APTESmodification.

Y. Zhang et al. / Polymer 54 (2013) 3605e3611 3607

combined weight of PDMS and TEOS. The mixed suspensions werecast in Teflon� molds. Solid composites were formed after 0.5 h.The PDMS/A-GO composites were harvested after one day in themold.

3. Characterization

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

Thermo-gravimetric analyses (TGA) were done on the GO and A-GO powders using a TA Q500 instrument (TA Instruments) under anitrogen atmosphere with a heating rate of 5 �C min�1.

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�A) at 45 kV and 40 mA. The diffraction angle wasincreased from 5� to 30� with the scanning rate of 0.05� min�1.

Scanning electron microscopy (SEM) was conducted with an FEIPhillips Electroscan XL30 ESEM-FEG microscope using an acceler-ation voltage of 15 kV. A-GO suspensions (0.03mgml�1) were spin-coated onto a flat aluminum plate at 2000 r.p.m. for 30 s. Then theA-GO coated aluminum plate wasmounted on a standard specimenholder using a double-sided carbon conductive tape for SEM im-aging. A transmission electron microscopy (TEM) sample was pre-pared by placing a few drops of dispersion onto a lacey carbon filmsupport on a Cu grid. Images were acquired in a JEOL 1230 trans-mission 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 A-GO inchloroform for 1 h, followed by spin-coating A-GO 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 strength was measured using an ESM-301 (MARK-10) tensile test tester. The experiments were carried out at room

Fig. 1. FT-IR of GO and A-GO. The peaks at 2922 cm�1 and 2886 cm�1 appear aftermodification with APTES. Meanwhile, the peak at 1714 cm�1 shifts to 1637 cm�1, whichindicates APTES is attached to GO through the amine group.

temperature at a crosshead speed of 50 mm/min using dumbbell-shaped specimens with an original length of 40 mm.

4. Results and discussions

FTIR data show that the APTES is linked to the GO through theamine group by the reaction with carboxylic acid group. Fig. 1shows the FTIR spectra of GO and A-GO. In the spectrum of GO, eOH (3118 cm�1), CaO (1714 cm�1), and aromatic CaC (1595 cm�1)stretches were observed. After modification with APTES, two newpeaks appear at around 2922 cm�1 and 2886 cm�1, which areassigned to the CeH stretch in the methylene group from APTES.There was a significant decrease in the wavenumber for the peak ofCaO stretch, from 1714 to 1637 cm�1, presumably due to the re-action of carboxylic acid groups with amine groups in the APTES.The wavenumber decreases because CaO group is connected to e

NH2 after the reaction and the resonance effect between CaO ande

NH2 increases the CaO bond length and reduces the frequency ofabsorption. Another peak at 1062 cm�1 assigned to SieOeC was

Fig. 3. XRD pattern of GO and A-GO. After modifying GO with APTES, the peak at 10.3�

disappears. The absence of interlayer reflection for A-GO implies that the sheets arefully exfoliated.

Fig. 4. (A): Comparison of the dispersion GO (left) and A-GO (right) observed after several days after dispersion by sonication for 1 h; (B): SEM image of A-GO, showing the wrinkledmorphology. The size of A-GO flake is around 3 mm; (C): TEM image of A-GO. The 100 nm bar is approximately the distance between folds, which we identify as the persistence length.

Fig. 5. AFM images and height profiles of A-GO dispersed at a concentration of 0.03 mg ml�1 in chloroform. Flat A-GO flake was selected to study the thickness, which is 2 � 0.3 nm.

Y. Zhang et al. / Polymer 54 (2013) 3605e36113608

Fig. 6. XRD patterns of A-GO, PDMS and PDMS/A-GO composites with different A-GOloading. As A-GO loading increases to 2 wt%, the intensity of the PDMS amorphouspeak (centered at 12�) decreases, indicating that A-GO intercalates into PDMS network.At 3 wt% this trend reverses indicating reduced intercalation due to aggregation.

Fig. 7. Young’s modulus enhancement factor of PDMS/A-GO composites as a functionof the A-GO volume fraction. The line is a fit to the combined data.

Y. Zhang et al. / Polymer 54 (2013) 3605e3611 3609

observed in A-GO spectrum, confirming that the ethoxy groups stillexist in A-GO.

TGA shows two stages of decomposition, one due to functionalgroups and the other due to pyrolysis of the carbon skeleton. Fig. 2shows the typical TGA curves for GO and A-GO. GO shows relativelylow thermal stability and starts losing mass below 100 �C (about7%), which is attributed to the absorbed water. The 46% weight lossbetween 100 �C and 280 �C is due to decomposition of the labileoxygen-containing functional groups [42,43]. Between 280 �C and600 �C, no obvious weight loss was observed indicating all of theoxygen functional groups have decomposed below 280 �C. Thedecrease in mass around 600 �C is due to the pyrolysis of the carbonskeleton of the GO [44]. For A-GO, the mass loss below 100 �C isabout 4% which is lower than that of GO, indicating lower hydro-philicity after modification with APTES.

The modification level was determined from the TGA data usingthe method of Yang et al. [27] and Dubin et al. [45] to identify thepercentage of different functional groups. The mass loss for A-GObetween 100 �C and 280 �C is 30 wt% compared to 46 wt% for GO.The 16-wt% difference is attributed to APTES modifier, which de-grades above 280 �C. We conclude that 16 wt% of oxygen-containing functional groups in GO have been modified by APTES.This value is consistent with the mass loss (16% � 2%) between280 �C and 600 �C, which is attributed to the removal of alkyl chains[44], corresponding to conversion of 16 wt% of the oxygen-containing functional groups in GO. The consistency of thesevalues justifies the attribution of the mass loss between 100 and280 to loss of the APTES modifier.

XRD was used to study the layer spacing of dry GO and A-GO. InFig. 3, the diffraction peak at 2q ¼ 10.3�, corresponding to aninterlayer distance of 0.86 nm, is significantly larger than that ofnatural graphite at 26� (Fig. S1 in Supporting Information) butwithin the range expected for GO, where the interlayer spacingdepends on relative humidity [46,47]. After modifying grapheneoxide with APTES, the peak at 10.3� disappears. The broad peakwith relatively low intensity between 15� and 25� observed forboth GO and A-GO is due to the graphene oxide flakes. The absenceof interlayer reflection for A-GO implies that the sheets are fullyexfoliated.

After sonication for 1 h, the A-GO dispersion in chloroformbecame ‘dark’ and remained stable for several days (Fig. 4A, right). Incontrast, GO remained undispersed at the bottom of the bottle(Fig. 4A, left). SEM and TEM measurements were performed tofurther investigate the morphology of A-GO. Fig. 4B shows the SEMimageofA-GO. The image shows thewrinkled sheets. And the size ofthe A-GO flakes is around 3 mm. From the TEM image of A-GO(Fig. 4C), exfoliated A-GO sheets are observed indicating greatlyenhanced dispersion. Some of the exfoliated sheets overlap withother sheets, which results in darker areas in the TEM image. Somesmallerflakes are observed since the sampleswere sonicated for 1 h.

An A-GO flat flakewas selected to determine the thickness usingAFM in taping mode as shown in Fig. 5. The cross-sectional view ofthe AFM image of A-GO indicates that the thickness of A-GO is2 nm. The typical observed monolayer graphene sheet is around0.8 nmwhich is larger than the theoretical graphite sheet with vander Waals thickness of ca. 0.34 nm [44] because of the presence ofoxygen functional group above and below the GO plane. So ourresult indicates that the platelet of A-GO consists of four layers ofgraphene sheets or less [48].

XRDwas used to study the effect of A-GO in PDMSmatrix, whichindirectly shows that A-GO starts to aggregate in PDMS at relativelyhigh loading (Fig. 6). The pure PDMS has a broad XRD peak between10� and 15� centered at around 12�, corresponding to the diffrac-tion from amorphous PDMS phase, which is similar to a previousreport [49]. When the amount of A-GO is less than 3 wt%, the

amorphous peak for A-GO totally disappears and the intensity ofthe amorphous peak for PDMS composites decreases compared topure PDMS. These observations indicate that A-GO disrupts whatlittle order there is in PDMS. A-GO intercalates into PDMS networkduring the cross-linking reaction, restricting the mobility of poly-mer chain [50]. At the highest loading the trend reverses becausethe A-GO is poorly dispersed [51] as anticipated on the basis of thetoughness data shown below.

Mechanical properties of the PDMS/A-GO nanocompositesgreatly improve as the filler loading increases. Fig. 7 shows thetensile Young’s modulus enhancement of (E/E0) versus A-GO vol-ume fraction (4). Young’s modulus increases from 0.25 MPa inPDMS to 0.46 MPa in A-GO/PDMS (3.0 wt %). This change isattributed to the uniformly dispersed of A-GO in the PDMS matrix.The line in Fig. 7 is a linear fit to the combined data. For platelets, E/E0 is expected to be proportional to filler volume fraction, 4 [33],which was calculated from the density of PDMS ¼ 0.965 g/cm3 anddensity of A-GO ¼ 2.2 g/cm3 [52]

Fig. 8. a. Stressestrain curves of PDMS/A-GO composites with different filler contents. The elongation of composites at 3 wt% decreases due to filler aggregation in composites. b.Toughness of PDMS/A-GO composites as a function of the A-GO volume fraction. The observation of toughness below that of the matrix is attributed to defects induced by theaggregated filler at high loading.

Y. Zhang et al. / Polymer 54 (2013) 3605e36113610

EE0

¼ 1þ a4þ. (1)

The parameter a is the graphene aspect ratio. For the data inFig. 8, a ¼ 58/0.93 ¼ 62. Assuming a sheet thickness of 2 nm,this aspect ratio implies a 1-dimensional persistence lengthof 124 nm, which is substantially smaller than the aspect ratio(1500) estimated from the SEM flat-plate width and AFMthickness.

The discrepancy between the aspect ratio determined from themodulus enhancement and that of flat graphene sheets is attrib-uted to the wrinkled morphology of graphene sheets and thebreakup of large sheets caused by sonication [53]. In this situationthe aspect ratio in equation (1) is expected to be the determinedby persistence length rather than the ideal sheet size [33]. Asshown in Fig. 4B and C, A-GO sheets are not stiff but displaywrinkled morphology. The persistence length is roughly the meandistance between folds [33]. A 100-nm bar is included in Fig. 4 forvisual comparison with the fold distance. The fold distance iscomparable to the 124-nm estimate based on equation (1), con-firming the importance of crumpling on modulus. In addition,many of the A-GO sheets are broken into small pieces as shown inthe TEM image (Fig. 4C) caused by 1-h sonication, which alsocontributes to the reduced experimental modulus enhancementfactor.

In the absence of a quantitative measure of persistence length inthe composite itself, it is not clear if rubber-like elasticity of thefiller [33] is playing a role. Undoubtedly filler elasticity is importantbecause dispersion likely reduces the persistence length below thatobserved in Fig. 4C, which means that reinforcement due tobending of the sheets is inadequate to explain the observedenhancement factor.

Since APTES has potentially reactive ethoxide groups, thequestion arises as to whether the reinforcing effect is due toenhanced crosslinking of PDMS through the APTES functionalgroups. To elucidate this issue, two PDMS samples cross-linkedwith TEOS and A-GO were prepared separately. When PDMS iscross-linked using TEOS (1:6 M ratio to PDMS) a solid elastomerresults within 24 h. For comparison, PDMS “crosslinked” with 0.1 gA-GO, which contains the same number of ethoxide group, remainsa liquid after 72 h. The amount of A-GO was calculated based onTGA, which indicates A-GO contains 16 � 2 wt% of APTES. Theseobservations show that A-GO crosslinking makes no contributionto the modulus of the composites.

Toughness shows enhancement but only at low loadings. Fig. 8ashows the stressestrain curves of PDMS/A-GO nanocomposites as afunction of loading. The areas under the curves were used as thetoughness measure. The toughness (Fig. 8b) reaches a maximum at0.5 wt % filler content. At higher loading toughness decreases due toreduced elongation as has been observed in many classes nano-composites. In our case, however, the modulus falls below that ofthe matrix, which indicates premature failure at low elongationsdue to defects (stress concentrators) induced by aggregated fillerparticles.

5. Conclusions

GO was successfully modified by APTES. APTES attaches by re-action of surface carboxylic acid groups on GO with the aminegroup on APTES. The introduction of APTES onto GO improves GOexfoliation in chloroform leading to a stable dispersion, which fa-cilitates the dispersion of GO in PDMS composites. Although eth-oxide moieties are available on APTES, the crosslinking of PDMSthrough the filler is negligible, which indicates that reinforcementis a physical not chemical phenomena. The modulus data show thatcrumpling of the filler plays an important role in reinforcement.Although a 71% improvement of Young’s modulus was observedthis value is substantially below that expected for ideal flat-sheets.The observed enhancement is consistent with a 124 nm persistencelength of the crumpled filler sheets, which is comparable to thedistance between folds observed by TEM on flakes deposited fromchloroform solution.

Acknowledgments

The authors are grateful to the financial support provided JEMby the National Science Foundation through Grant DMR-0803454(Polymers Program, Division of Materials Research). We thank Dr.Necati Kaval for his help in FTIR characterization and ProfessorVesselin Shanov’s help on tensile measurements. Professor Zhu andRSR appreciate support from the University of Texas throughstartup funds to RSR.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.polymer.2013.04.057.

Y. Zhang et al. / Polymer 54 (2013) 3605e3611 3611

References

[1] Geim AK, Novoselov KS. Nat Mater 2007;6(3):183e91.[2] Stankovich S, Dikin DA, Ruoff RS. Nature 2006;442(7100):282e6.[3] Park S, Ruoff RS. Nat Nanotechnol 2009;4(4):217e24.[4] YangY,Wang J, Zhang J, Liu J, YangX, ZhaoH. Langmuir 2009;25(19):11808e14.[5] Rao CNR, Sood AK, Subrahmanyam KS, Govindaraj A. Angew Chem Int Ed

2009;48(42):7752e77.[6] Zhao X, Zhang Q, Chen D, Lu P. Macromolecules 2010;43(5):2357e63.[7] Salavagione HJ, GomezMA,Martinez G. Macromolecules 2009;42(17):6331e4.[8] Jang B, Zhamu A. J Mater Sci 2008;43(15):5092e101.[9] Lomeda JR, Doyle CD, Kosynkin DV, Hwang W-F, Tour JM. J Am Chem Soc

2008;130(48):16201e6.[10] Zhu Y, Murali S, Cai W, Li X, Suk JW, Potts JR, et al. Adv Mater 2010;22(35):

3906e24.[11] Geim AK. Science 2009;324(5934):1530e4.[12] Balandin AA, Ghosh S, Bao W, Calizo I, Teweldebrhan D, Miao F, et al. Nano

Lett 2008;8(3):902e7.[13] Cai W, Moore AL, Zhu Y, Li X, Chen S, Shi L, et al. Nano Lett 2010;10(5):

1645e51.[14] Seol JH, Jo I, Moore AL, Lindsay L, Aitken ZH, Pettes MT, et al. Science

2010;328(5975):213e6.[15] Huang Y, Liang J, Chen Y. Small 2012;8(12):1805e34.[16] Xiang Q, Yu J, Jaroniec M. Chem Soc Rev 2012;41(2):782e96.[17] Liang J, Xu Y, Huang Y, Zhang L, Wang Y, Ma Y, et al. J Phys Chem C

2009;113(22):9921e7.[18] Park S, Dikin D, Nguyen S, Ruoff R. J Phys Chem C 2009;113(36):15801e4.[19] Ramanathan T, Stankovich S, Dikin DA, Herrera-Alonso M. Nat Nanotechnol

2008;3(6):327e31.[20] Stoller MD, Park S, Zhu Y, An J, Ruoff RS. Nano Lett 2008;8(10):3498e502.[21] Valentini L, Cardinali M, Bon Bittolo S, Bagnis D, Verdejo R, Lopez-

Manchado MA, et al. J Mater Chem 2010;20(5):995e1000.[22] Hong W, Xu Y, Lu G, Li C, Shi G. Electrochem Commun 2008;10(10):1555e8.[23] Paredes JI, Villar-Rodil S, Martinez-Alonso A, Tascon JMD. Langmuir

2008;24(19):10560e4.[24] Park S, Bielawski C, Ruoff R. Chem Soc Rev 2010;39(1):228e40.[25] Haubner K, Murawski J, Olk P, Eng LM, Ziegler C, Adolphi B, et al. Chem-

PhysChem 2010;11(10):2131e9.

[26] Park S, An J, Jung I, Piner RD, An SJ, Li X, et al. Nano Lett 2009;9(4):1593e7.[27] Yang H, Li F, Shan C, Han D, Zhang Q, Niu L. J Mater Chem 2009;19(26):

4632e8.[28] Chattopadhyay J, Mukherjee A, Chakraborty S, Kang J, Loos PJ, Kelly KF, et al.

Carbon 2009;47(13):2945e9.[29] Shen J, Li N, Shi M, Hu Y, Ye M. J Colloid Interface Sci 2010;348(2):377e83.[30] Liang J, Huang Y, Zhang L, Wang Y, Ma Y, Guo T, et al. Adv Funct Mater

2009;19(14):2297e302.[31] Xu Y, Hong W, Bai H, Li C, Shi G. Carbon 2009;47(15):3538e43.[32] Kim H, Abdala AA, Macosko CW. Macromolecules 2010;43(16):6515e30.[33] Schaefer DW, Justice RS. Macromolecules 2007;40(24):8501e17.[34] Yuan QW, Mark JE. Macromol Chem Phys 1999;200(1):206e20.[35] Mark JE, Allcock HR, West R. Inorganic polymers. 2nd ed. New York: Oxford

University Press, Inc; 2005.[36] Lewicki JP, Liggat JJ, Patel M. Polym Degrad Stab 2009;94(9):1548e57.[37] Gao Z, Nahrup JS, Mark JE, Sakr A. J Appl Polym Sci 2005;96(2):494e501.[38] Wang X, Xing W, Zhang P, Song L, Yang H, Hu Y. Compos Sci Technol

2012;72(6):737e43.[39] Hummers WS, Offeman RE. J Am Chem Soc 1958;80(6). 1339e1339.[40] Monagle JJ. J Org Chem 1962;27(11):3851e5.[41] Kathi J, Rhee K. J Mater Sci 2008;43(1):33e7.[42] Hu Y, Shen J, Li N, Shi M, Ma H, Yan B, et al. Polym Compos 2010;31(12):

1987e94.[43] Zhu Y, Stoller MD, Cai W, Velamakanni A, Piner RD, Chen D, et al. ACS Nano

2010;4(2):1227e33.[44] Yang H, Shan C, Li F, Han D, Zhang Q, Niu L. Chem Commun 2009;26:3880e2.[45] Dubin S, Gilje S, Wang K, Tung VC, Cha K, Hall AS, et al. ACS Nano 2010;4(7):

3845e52.[46] Bissessur R, Scully SF. Solid State Ionics 2007;178(11e12):877e82.[47] Ansari S, Giannelis EP. J Polym Sci Part B: Polym Phys 2009;47(9):888e97.[48] Akhavan O. Carbon 2010;48(2):509e19.[49] Liu L, Jiang Z, Pan F, Peng F, Wu H. J Membr Sci 2006;279:111e9.[50] Li R, Liu C, Ma J. Carbohydr Polym 2011;84(1):631e7.[51] Ahmed R, Ahmed A, Tarek A, Nasser S. Polym Adv Technol 2007;18(6):

463e71.[52] Zheng D, Tang G, Zhang H-B, Yu Z-Z, Yavari F, Koratkar N, et al. Compos Sci

Technol 2012;72(2):284e9.[53] Moniruzzaman M, Winey KI. Macromolecules 2006;39(16):5194e205.