lable at ScienceDirect

Polymer 93 (2016) 53e60

Contents lists avai

Polymer

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

Synthesis and gas permeability of highly elasticpoly(dimethylsiloxane)/graphene oxide composite elastomers usingtelechelic polymers

Heonjoo Ha a, Jaesung Park a, KiRyong Ha b, Benny D. Freeman a, Christopher J. Ellison a, *

a McKetta Department of Chemical Engineering, University of Texas at Austin, Austin, TX, 78712, USAb Department of Chemical Engineering, Keimyung University, Daegu, 704-701, Republic of Korea

a r t i c l e i n f o

Article history:Received 14 January 2016Received in revised form5 April 2016Accepted 10 April 2016Available online 11 April 2016

Keywords:Telechelic polymerGraphene oxideBarrier membraneElastomerAmine-epoxy reaction

* Corresponding author.E-mail address: ellison@che.utexas.edu (C.J. Ellison

http://dx.doi.org/10.1016/j.polymer.2016.04.0160032-3861/© 2016 Elsevier Ltd. All rights reserved.

a b s t r a c t

This study illustrates that amine functional groups on the ends of telechelic poly(dimethylsiloxane)(PDMS) can undergo post-processing reactions with surface epoxy groups on graphene oxide (GO) toform a robust elastomer during simple heating. In these materials, GO acts both as a nanofiller whichreinforces the mechanical properties and participates as a multifunctional crosslinker, thereby promotingelastic properties. Experiments indicate that the telechelic PDMS/GO elastomer is highly crosslinked (e.g.,more than 75 wt % is a non-dissolving crosslinked gel) but highly flexible such that it can be stretched upto 300% of its original length. Finally, the PDMS/GO elastomer was tested as a single gas barrier mem-brane and gas permeability was decreased ~45% by incorporating 1 wt % (0.43 vol %) of GO, therebyhighlighting its potential use in practical applications.

© 2016 Elsevier Ltd. All rights reserved.

1. Introduction

In recent years, flexible polymers have become increasinglyimportant for many applications that impact daily life, includingwearable devices [1e3], flexible displays [4e6], and devices formonitoring physiological signals [7e9], to name a few. These high-end applications typically require the use of an elastomer or rubber,such as crosslinked poly(dimethylsiloxane) (PDMS), to impartflexibility and mechanical robustness in addition to comfort; all areimportant features for interfacing with the human body. Whilethere exist commercial choices for PDMS elastomer precursors,such PDMS precursors are generally high in viscosity, limited toseveral grades and the recipe for preparing the final elastomer canbe complex [10e12]. Therefore, the development of new methodsfor forming elastomers with only a few simple components whichexhibit unique combinations of properties (e.g., electricallyconductive but flexible andmechanically robust, etc.) could be veryuseful.

Herein, we report a simple process for creating a strong andflexible elastomer from a reactive mixture of graphene oxide (GO)

).

and functional telechelic (i.e., both ends functionalized) PDMS. Theresulting composite has excellent mechanical properties but alsoserves as an effective and potentially economical gas barriermembrane for industrial use. For context, materials presenting ahigh barrier to oxygen permeation are typically very important forthe applications listed above because flexible devices often containorganic components to promote flexibility that can be particularlysusceptible to oxidative degradation.

For a number of reasons, the telechelic PDMS/GO compositeelastomer described in this study can be considered to be uniquewhen compared alongside other crosslinked PDMS materials. Forinstance, a distinguishing characteristic of the present elastomer isthat it not only possesses the traditional properties of crosslinkedPDMS materials, such as flexibility and good mechanical integrity,but it can also be made by simple and scalable methods using twolow-cost materials: graphite and telechelic polymer. In contrast,conventionally crosslinked PDMS elastomers are typically synthe-sized by one of three methods: platinum-catalyzed addition, vinyl-peroxide, or condensation curing [13]. All of these methods requirea catalyst to perform the reaction, accompanied with a range ofseveral different chemical additives, sometimes including fillerssuch as silica [11,12]. Many prior studies have investigated thetuning of physical properties, including mechanical, electrical, and

H. Ha et al. / Polymer 93 (2016) 53e6054

dielectric properties, of PDMS materials by adjusting the crosslinkdensity [14], adding fillers such as silica [15], clays [16], carbonnanotubes [17], graphene derivatives [18] or other polymers [19].Generally, these additional additives are most commonly physicallyintermixed with PDMS with no chemical bonding between thePDMS and additives. In contrast, the preparation of telechelicPDMS/GO composite elastomers in this study incorporates a GOfiller that also participates as a multi-functional crosslinker, whichfacilitates formation of the final covalently bondedmacromolecularnetwork.

While neat PDMS is highly gas permeable, the addition ofimpermeable platelet-like fillers such as GO can dramaticallyinhibit gas permeation. Recent studies of gas barrier membranesusing different graphene derivatives have focused on coating thesurface of flexible substrates (such as PDMS films or sheets) with athin layer of graphene or GO that is typically weakly bound bynoncovalent adhesion forces [6,20,21]. Graphene coated mem-branes could be susceptible to sections of the coating delaminatingfrom the flexible substrate after repeated stress, bending or up-stream/downstream pressure cycles [6,20,22]. Other potentialsources of damage from environmental exposure of the activegraphene or GO coating layer, such as irreversible adsorption ofundesired contaminants during prolonged use, could be possible. Incontrast to this coating approach, the elastomer described hereinforms a covalent macromolecular network between the flexiblecomponent (PDMS) and the gas impermeable component (GO)producing a dense and homogeneous barrier membrane. In sum-mary, from a membrane engineering perspective, we believe thiscomposite provides for the possibility of a cost-effective, highperformance gas barrier membrane.

2. Experimental section

2.1. Materials

In order to synthesize PDMS/GO elastomers, three componentswere combined: telechelic PDMS as the base material, GO as both afiller and multifunctional crosslinker, and tetrahydrofuran (THF;Fisher Chemical) as a solvent. Aminopropyl terminated telechelicPDMS with a viscosity of 1000 cSt (Mn ¼ 25,000 g/mol, DMS-A31,henceforth referred to as N1000) and 2000 cSt (Mn ¼ 30,000 g/mol, DMS-A32, henceforth referred to as N2000) were purchasedfrom Gelest and used without any further purification. For com-parison, silanol terminated PDMS with a viscosity of 1000 cSt(Mn ¼ 26,000 g/mol, DMS-S31) and epoxypropoxypropyl termi-nated PDMS with a viscosity of 15 cSt (Mn ¼ 550 g/mol, DMS-E11)were also purchased from Gelest and used as received. GOs weresynthesized by the modified Hummer's method as describedelsewhere [23], where sulfuric acid, potassium permanganate,hydrogen peroxide, and hydrochloric acid were purchased fromSigmaeAldrich and used as-received. Pre-oxidized graphite fromBay Carbon Inc. (SP-1) was used as the starting material to syn-thesize GOs. The nominal number average particle size of the SP-1graphite was approximately 19.3 mm [24], and it has been reportedby others that the GO sheet size reduces to an average value of8.3 mm when this grade of graphite is oxidized using the modifiedHummer's method with an oxidation temperature of 35 �C [25].

2.2. Fabrication of PDMS/GO elastomers

First, a known amount of telechelic PDMS (for 1 wt % GO com-posites, 990 mg) was dissolved in THF (6 ml) while stirring toensure complete dissolution of the polymers. Then, dried GO(10 mg) was suspended in the solutionwith vigorous stirring, and ahomogeneous dispersion of GO in this solution was promoted by

sonication using a 400 W probe sonicator (Branson Digital Sonifier450) with 10% amplitude for 10 min (24 kJ) in an ice bath. It isnoteworthy to point out that, although THF is not a good solvent forGOs, amine terminated telechelic PDMS can effectively act as asurfactant to assist in dispersing GOs homogeneously in THFthrough hydrogen bonding between amine ends and oxygen con-taining groups on GO surfaces. A good dispersion after sonicationwas confirmed by examining the residual liquid film in the vialwhile rolling the vial containing the solution and after pouring themajority of the solution into a PTFE dish for solution casting. Ineither case, the liquid film in the vial was essentially transparent(although brownish-black from the GO content) for all solutionconcentrations and no aggregates could be identified with thenaked eye. This solutionwas immediately poured into a Teflon dishand covered in order to slowly evaporate the solvent thereby so-lution casting a film at room temperature. Solvent was evaporatedfor at least 2 days and then the sample was vacuum dried at roomtemperature for an additional day to make a homogenous PDMS/GO uncrosslinked liquid sol. In order to form a crosslinked elas-tomer, the sol precursor was heated to 160 �C in a vacuum oven for24 h and cooled slowly to room temperature before use.

For one of the gas permeation tests, a reference sample (neatPDMS elastomer) was synthesized by reacting aminopropylterminated telechelic PDMS and epoxypropoxypropyl terminatedtelechelic PDMS. A detailed explanation of how this freestanding,neat PDMS elastomer was synthesized is summarized in the sup-porting information.

2.3. Characterization of PDMS/GO elastomers

The chemical crosslinking reaction between PDMS and GOs wasinvestigated using Fourier transform infrared spectroscopy (FTIR;Thermo Nicolett, 6700) with a scan size (resolution) of 2 cm�1

collected over 256 scans per sample. Solvent uptake and gel contentof each composite elastomer were calculated using Equations (1)and (2) below where ‘w’ indicates weight, according to ASTMD2765 test method C. THF was used as the solvent to dissolve awayunreacted soluble polymers.

Solvent Uptake ð%Þ ¼ wafter swell �wbefore swell

wbefore swell� 100 (1)

Gel Content ð%Þ ¼ wafter swell and dried in vacuum

wbefore swell� 100 (2)

All rheological experiments were conducted on a shearrheometer (TA Instruments, AR-2000EX) using an 8 mm parallelupper plate and a Peltier lower plate fixture for temperature controlwith a gap of 500 mm. First, a strain sweep was conducted toidentify the linear viscoelastic regime at 25 �C in the range of0.01e10% strain at a frequency of 1 Hz. A frequency sweep was thenconducted at 25 �C in the range of 0.01e10 Hz with 0.1% strain.Temperature sweeps were conducted from 25 �C to 100 �C with5 �C step changes using both 0.1% strain and a frequency of 1 Hz.Stress relaxation tests were conducted by subjecting the sample toan initial strain of 10% and 20%. In order to measure the mechanicalproperties of the composite elastomers, microtensile specimenswere prepared by solution casting on a mold (i.e., to produce a dogbone sample with a gauge length of 22 mm, width of 4.8 mm andthickness of about 0.7 mm) satisfying ASTM D1708-13 standards.An Instron (model 5966) equipped with a 1 kN load cell was usedwith a strain rate of 10 mm/min (i.e., at our gauge length this is anominal strain rate equivalent to 50 mm/min for ASTM D638standard samples, which is the recommended speed for rubbersamples) and samples were tested in triplicate.

H. Ha et al. / Polymer 93 (2016) 53e60 55

Thermal properties of the elastomers were characterized bydifferential scanning calorimetry (DSC; Mettler Toledo DSC1).Heating and cooling rates of 10 �C/min were used for all experi-ments, and the second heating curve was used for the glass tran-sition temperature (Tg) analysis to erase thermal history. GOs werefurther characterized using X-ray photoelectron spectroscopy (XPS;Kratos) to compare the composition of the GOs before and afterannealing. CasaXPS software (v 2.3.16) was used to perform curvefitting in which a Shirley background was assumed. The dispersionof GOs in the composite material was determined by optical mi-croscopy (OM; Olympus BX60) using a microtomed sample on a400 mesh Cu grid. Sectioning was performed by a benchtopmicrotome (RMC Products, PT-XL Power Tome) with a diamondknife where the temperature of the diamond knife and the sampleholder were both set to �150 �C. According to the images (sup-porting information), all of the composite materials considered inthis study showed good dispersion of GOs without any noticeablelarge aggregates.

2.4. Single gas permeation measurements

Single gas, including H2, O2, N2, CH4 and CO2, transport prop-erties were measured using the constant volume, variable pressuremethod at 35 �C with ultra-high purity grade gases from Airgas. A1000 psig pressure transducer (Honeywell Sensotec, Model STJE)was used tomeasure the upstream pressure in the system. A 10 Torrcapacitance manometer (MKS, Baratron 626 A) was used to mea-sure the downstream pressure, and the downstream pressure waskept below 10 Torr using a vacuum pump. All data were recordedusing National Instruments Lab-VIEW software. Permeability,which is an intrinsic property of a specific material to a specificpermeate, was expressed in barrer units, where 1 barrer equals to10�10 cm3(STP) cm/cm2 s cmHg. The average thickness of thesample was measured using a dial gauge (Mitutoyo, Absolute dig-imatic indicator) by sandwiching a sample in between two thinmetal plates (each metal plate thickness ¼ 0.05 mm). Metal plateswere used to prevent samples from being compressed by the tip ofthe dial gauge during measurement. Thicknesses for the neat PDMSelastomer and telechelic PDMS N1000/1 wt % GO elastomer were0.998 ± 0.021 mm and 0.354 ± 0.010 mm, respectively.

3. Results and discussion

To fully understand the chemical nature of the thermallycrosslinked telechelic PDMS/GO elastomers, a fundamental inves-tigation regarding the reaction between amine terminated tele-chelic PDMS and GO was pursued first.

3.1. Proposed reaction and evidence of crosslinking

As shown in Scheme 1 (a), only twomaterials were used to formthe highly elastic elastomer, the first being amine terminated tel-echelic PDMS, which contains a primary amine attached to eachend of the polymer chain. The second material, GO, containsnumerous oxygen-containing functional groups, including hy-droxyls, carboxylic acids, and epoxides, on the surface of the GOsheets [26]. XPS data supports the fact that hydroxyls and epoxidesare present in larger proportion than other functionalities (seeFigure S2). Once the two materials are mixed together and suffi-cient energy is applied, in this case heat, an epoxide on GO can reactwith the nearest primary amine on PDMS by a ring-opening reac-tion. Such a reaction results in the formation of ethanolamine co-valent bonds between the telechelic polymer ends and the epoxygroups on the surface of GO, as shown in Scheme 1 (b). Thisnucleophilic substitution reaction has been proposed and

demonstrated previously by other groups when functionalizing GOsurfaces with amines [27,28] as well as for synthesizing chemicallycrosslinked nylons [29,30], epoxy composites [31e33], and poly-mer/GO papers [34,35]. However, to the best of our knowledge, noresearch implementing telechelic polymers with GOs to form acontrollable, highly elastic composite material has been reported[20]. It is worth noting that we have also confirmed this reactionproceeds similarly for PDMS materials containing secondary amideend groups [36].

There are three possible scenarios for amine terminated tele-chelic PDMS to react with GOs: two ends react with two differentGO sheets, one end reacts with one GO sheet while the other endforms a dangling strand and remains unreacted, or two ends reactwith the same GO sheet forming a loop. Considering the high gelcontent obtained for various samples (which will be discussed inmore detail in a later section), it is postulated that the majority ofthe reaction takes place by forming a crosslink between twodifferent GO sheets, since the latter two possibilities would notcontribute to enhancing the gel content and mechanical propertiesof the composite. Given that reaction is only possible at chain ends,telechelic polymers offer the opportunity to produce elastomerswith a predefined and tailorable molecular weight betweencrosslink points. Considering that this reaction only involvespolymer end-groups, the versatility and universality of this cross-linking reaction could be easily demonstrated by applying it toother repeat unit chemistries extending this approach to polymersthat are amorphous or semicrystalline, hydrophobic or hydrophilic,synthesized by either addition or step growth polymerizations, etc.Finally, due to the simplicity of the reaction and the fact that nobyproducts are formed during crosslinking, it is expected that thisfacile method could be scalable to mass produce composite mate-rials in a cost-effective way.

As shown in Fig. 1, simply annealing the solution cast telechelicPDMS/GO composite film in vacuum for 24 h resulted in a materialthat was freely standing, highly flexible and mass producible indifferent sizes and shapes (note the larger size in Fig. 1 (f)). Thetransformation of a viscous sol to a self-supporting solid elastomerprovided indirect evidence of a reaction forming covalent bondsbetween the polymer and GO. For comparison, a silanol terminatedtelechelic PDMS/GO composite remained a soluble viscous liquideven after annealing at high temperature (Fig. S1). This clearlydemonstrates the importance of complementary reactive func-tionalities for performing the reaction that leads to the finalelastomers.

While the occurrence of the crosslinking reaction is evidencedby the images in Fig. 1, absorbance spectra obtained using FTIRprovide more direct evidence of the chemical changes associatedwith this reaction (Fig. 2). To understand the crosslinking reactionand formation of the ethanolamine linkage, the telechelic PDMS/GO mixture was investigated before and after annealing. As ex-pected, themajority of the intense peaks in the spectrawere relatedto the PDMSmain chain repeat units; detailed peak assignments forFig. 2 (a) are tabulated in the supporting information in Table S1. Forthe N1000 telechelic PDMS used in Fig. 2, with amine ends presentat approximately 0.6 mol % (i.e., 0.12 wt %), the 1500e1700 cm�1

region of the spectra was examined in further detail in Fig. 2 (b).Before annealing, the peak associated with NeH bending wasassigned at 1600 cm�1 for both N1000 PDMS and N1000 PDMS/1 wt % GO sol samples. However, after annealing the sol to crosslinkthe sample producing an elastomer, a significant fraction of theNeH bending peak disappeared, indicating that the majority of theprimary amine end groups were reacted during this step. In addi-tion, due to a known thermal annealing induced reduction reactionof GO to reduced GO, a stronger C]C ring stretching peak at1580 cm�1 was formed for the N1000 PDMS/1 wt % GO elastomer

Scheme 1. (a) Chemical structures of the two materials used in this study. Each material's active sites are capable of reacting with complementary functional groups (colored) onthe other component. (b) The anticipated reaction mechanism that produces elastomers in this study. No other byproducts are formed during the crosslinking reaction.

Fig. 1. Images of N1000/1 wt % GO elastomers. (a) As prepared material is not only (b) freestanding, but also (c) bendable and (d) flexible. (e) The material is elastic enough to makea knot. (aee) is a N1000/1 wt % GO elastomer disc 38 mm in diameter, while (f) is 10 cm in diameter.

H. Ha et al. / Polymer 93 (2016) 53e6056

sample. A separate case study on the effect of heat on the trans-formation of GO to reduced GO is provided in Fig. S2. Because theFTIR spectra after extended annealing were nearly identical tothose after 24 h of annealing at 160 �C, it can be concluded that thereaction was essentially completed within 24 h.

Another method to confirm the formation of chemically cross-linked elastomers is to measure the swelling ratio and gel contentof the composite material. By controlling the content of GO and theMn of the PDMS polymer incorporated in this system, the gel

crosslink density can be altered effectively (Table 1). For example,increasing the GO content resulted in the introduction of morecrosslinking sites and, as a result, dramatically increased the gelcontent and decreased the solvent uptake of the elastomers. On theother hand, as the Mn of the PDMS was increased from 25 to 30 kg/mol, the mol fraction of the amine end groups was reduced from0.6% to 0.5%. Table 1 shows that the degree of swelling expectedlyincreased and gel fraction decreased; the likelihood of aminesreacting with the surface epoxy groups of GO would be expected to

Fig. 2. FTIR spectra before (sol) and after (elastomer) thermal annealing. (a) Full scan and (b) selected region from 1500 to 1700 cm�1. FTIR spectra have been shifted vertically forclarity but are otherwise on the same scale.

Table 1Solvent uptake and gel content for elastomer samples measured as a function of PDMSmolecular weight and GO content. The indicated error is standard deviation of at least 5separate samples.

Sample Mn,PDMS

(g/mol)Solvent uptake(%)

Gel content(%)

Soluble fraction(%)

N1000/1 wt % GO 25,000 711 ± 26 77.0 ± 1.7 ~23N1000/2 wt % GO 25,000 334 ± 17 89.2 ± 0.8 ~11N2000/1 wt % GO 30,000 836 ± 31 72.1 ± 1.8 ~28N2000/2 wt % GO 30,000 421 ± 4 84.4 ± 1.3 ~16

H. Ha et al. / Polymer 93 (2016) 53e60 57

decrease in this situation reducing crosslinking density. An asso-ciated factor is that the mobility of a higher Mn polymer could alsobe slightly lower due to increased entanglements and viscosity.

It is noteworthy that the viscosity of the telechelic PDMScomponent, which is directly related to the Mn of the polymer,plays an important role in forming a homogeneous crosslinkedmaterial following solution casting and thermal annealing. Lowviscosity telechelic PDMS, such as 50 cSt and 100 cSt, formed amaterial showing a gradient in elasticity through the film thicknessdirection (data not shown). This occurs as GOs (i.e., the highlyfunctional crosslinkers) settled to the bottom of the PTFE dish,resulting higher crosslink density at the bottom while essentiallyleaving the top as a viscous amine terminated PDMS liquid. Asimilar phenomena was observed previously by Clarizia and co-workers for PDMS/zeolite composite materials [37]. This can beavoided by using telechelic PDMS with higher molecular weightand viscosity to slow or prevent settling of GO; these observationscould be exploited for tailoring these materials further in thefuture.

3.2. Physical properties of telechelic PDMS/GO elastomers

Even though extensive information on the formation of thecrosslinked network was obtained from FTIR and gel contentmeasurements, rheological and mechanical properties provideadditional understanding of both the elastomer and its precursor.Thus, the dynamic modulus for PDMS N1000/1 wt % GO sol andelastomer samples were investigated in more detail as a function ofthe % strain (Fig. 3 (a)), frequency (Fig. 3 (b)), temperature (Fig. 3(c)), and during stress relaxation for 10% and 20% strain (Fig. 3 (d)).

As shown in Fig. 3 (a)-(c), the elastomer exhibited a storagemodulus (G0) that was approximately an order of magnitude higherthan the lossmodulus (G00) over awide range of strains, frequencies,and temperatures, indicative of stable, solid-like viscoelasticbehavior. In contrast, the PDMS/GO uncrosslinked sol exhibited

liquid-like behavior (G00 being higher than G0) over the range ofinterest and even showed considerable changes across both thefrequency and temperature sweep, which is clearly distinguishablefrom that of the elastomer sample. A stress relaxation test alsoconfirmed the elastomers as viscoelastic solids, which is in agree-ment with the aforementioned rheological observations shown inFig. 3 (a)e(c). After relatively short relaxation times, the stressrelaxation curve reached an asymptotic equilibrium stress ofapproximately 20 kPa. On the other hand, the uncrosslinked liquidsol composites exhibited complete stress relaxation within a fewseconds.

The mechanical integrity of the elastomer was also investigatedusing a universal tensile test. As expected, increasing the GO con-centration resulted in a higher tensile strength while reducing theelongation at break (see stressestrain curve in Fig. 4 and datatabulated in Table 2). It is noteworthy that the precursor PDMS andPDMS/GO sol are liquids, and it was therefore impossible toperform such experiments on these samples. This result highlightsthe effect of chemical crosslinks between the polymer and GO,where other studies on physically mixed thermoplastic poly-urethane/octadecylamine functionalized GO composites showedno reinforcement at loading levels below 2.5 vol % [38]. Althoughthe gel content for high Mn PDMS (N2000) composites was slightlylower than that of low Mn PDMS (N1000, related data in Table 1),the N2000 samples exhibited moderately higher tensile strength.We hypothesize that a major contributing factor could be related totrapped entanglements between the reacted PDMS chain ends thatcould act as physical crosslinks within the network increasing thetensile strength.

In general, the physical and mechanical properties of the tele-chelic PDMS/GO elastomers are very important aspects whenconsidering their use as gas barrier membranes. Depending on thespecific application, dense polymer gas barrier membranes may besubjected to high pressures (over 10 atm) with high gas fluxpenetrating through the thin film membrane [39]. It is critical for

Fig. 3. Comparison in rheological properties between N1000/1 wt % GO uncrosslinked sol and elastomer samples. (a) Oscillatory strain sweep, (b) oscillatory frequency sweep, (c)oscillatory temperature sweep, and (d) stress relaxation results. During the oscillatory frequency sweep in (b), the N1000/1 wt % sol sample's raw phase reached the instrumentlimit (>150�). Therefore, data in the region of uncertainty is not shown.

Fig. 4. Representative stressestrain curves of telechelic PDMS/GO elastomers.

Table 2Tensile properties of telechelic PDMS/GO elastomers depending on the Mn of thetelechelic PDMS and GO content. The indicated error is standard deviation of at least3 separate samples.

Sample Tensile strength at break (MPa) Elongation at break (%)

N1000/1 wt % GO 0.15 ± 0.02 280 ± 12N1000/2 wt % GO 2.20 ± 0.17 136 ± 17N2000/1 wt % GO 0.77 ± 0.02 232 ± 8N2000/2 wt % GO 2.33 ± 0.40 98 ± 16

H. Ha et al. / Polymer 93 (2016) 53e6058

the membrane to withstand both external pressure and gas flowwithout forming pinholes or rupturing completely.

3.3. Application as an effective gas barrier membrane

There is a strong demand for enhancing the gas barrier prop-erties of polymer nanocomposite materials for prolonging lifetimesof organic electronics or packaged food, among other applications.One of the common approaches to improve the gas barrier prop-erties is to optimize the dispersion of the filler material usingsuitable processing procedures. Most of the polymer/GO mem-brane research has involved preparation of samples by solutionmixing, in situ polymerization, and melt mixing [20]. While thehigh barrier performance of thesematerials has been demonstratedpreviously by physically mixing two or more components to obtainsynergistic properties, a number of recent studies have focused oncoating a thin layer of graphene or graphene derivative on top of aflexible substrate, or preparing a free-standing graphene derivativemembrane [21]. Such approaches accordingly concentrate thegraphene forming a denser layer of gas impermeable filler but theycan introduce new challenges such as requiring careful handling toprevent the filler from fracturing or delaminating from the sub-strate. Exploiting the fact that the telechelic PDMS/GO elastomersexhibit excellent mechanical integrity with the PDMS componenteffectively encapsulating and protecting the gas impermeable GOfrom external stress and environmental damage, we envisionedthat this material would performwell as a gas barrier or separationmembrane.

PDMS (and closely related derivatives) are common materials

Fig. 5. Single gas permeability of (a) neat PDMS and (b) telechelic PDMS N1000/1 wt % GO telechelic PDMS/GO elastomers.

H. Ha et al. / Polymer 93 (2016) 53e60 59

used in commercial gas/vapor separationmembranes [39]. It is wellaccepted that by incorporating less permeable fillers (in this caseGO) into the composite matrix, the gas barrier properties of thematerial should increase by introducing tortuous pathwaysthrough which the gas permeates [38,40e42]. To quantitativelymeasure the permeation of various gases, a single gas permeationtest was performed, as shown in Fig. 5 and tabulated in Table S2. Asa reference case, neat PDMS elastomer was synthesized and tested(Fig. 5 (a)), which exhibited identical gas permeability performanceto that reported previously by others [39]. For the various gasestested, the gas permeability for both neat PDMS and telechelicPDMS/GO composite elastomer increased in the order of N2, O2, H2,CH4, and CO2. Strongly sorbing penetrants, such as CO2, had thetendency to plasticize the polymer matrix and thereby slightly in-crease the permeability with increasing pressure. Other low-sorbing gas molecules had less effect on permeability with in-creases in feed pressure [39]. Reacting 1 wt % GO (0.43 vol %) withtelechelic PDMS to form the elastomeric matrix resulted in anaverage reduction of ~45% in N2, O2, CH4, and CO2 single gas per-meabilities. Osman and coworkers achieved a 30% reduction in O2transmission rate at 3 vol % for physically mixed polyurethane/alumino-silicate nanocomposites [43] while others have shown a30e50% reduction in single gas permeation for polymermixedwith1.6e1.8 vol % thermally reduced graphene oxide composites(Table S3, comparing the gas barrier performance of differentcomposite systems) [44]. Finally, as shown in Table 3, the gasselectivity values for the telechelic PDMS/GO composite elastomerswere more or less similar to that of the neat PDMS elastomer.

One could speculate that the thermal properties of the neatPDMS and telechelic PDMS/GO elastomer could contribute to dif-ferences in their gas permeability data, especially given that thereaction between PDMS and GO might be anticipated to increasethe glass transition temperature; Fig. S3 shows that the glasstransition temperature (Tg) and melting temperature (Tm) of bothmaterials were essentially identical. This clearly indicates that thesignificant decrease in gas permeability upon the addition of GO ismostly due to the impermeable GO dispersed phase, which

Table 3Gas selectivities of neat PDMS and N1000/1 wt % GO elastomer. Selectivity values arecalculated based on 10 atm permeability values.

Material CO2/N2 CO2/CH4 O2/N2

Neat PDMS 10 3 2N1000/1 wt % GO 10 4 2

constrains gas molecules forcing them to migrate through irregularamorphous regions of the composite material, and not by thereduction in chain segment mobility in the interstitial amorphousphase. It is evident that incorporation of higher loadings of GO(under the constraint that higher levels of this reactive GOcomponent still allows for formation of mechanically robust elas-tomers) and enhanced alignment of the dispersed GO plateletswould provide better gas barrier performance. Research related tothese aspects, accompanied with efforts to model the relevantprocesses, is currently being investigated and will be reported inthe near future.

4. Conclusions

A facile method to prepare elastic polymer/GO composite ma-terials using telechelic polymers was developed. By taking advan-tage of simple chemistry between amine terminated telechelicpolymers and GO, it was possible to transform liquid-like precursormaterials into solid, robust elastomers by incorporating less than 1vol % GO into the system. The GO acted both as a filler material thatsignificantly reinforced the mechanical properties and as a multi-functional crosslinker forming chemical bonds with the polymerestablishing the final macromolecular network. The elastomersprepared in this work were highly crosslinked (>75 wt % gel) andcould be extended up to 300% when mechanically stretched. Therheological studies revealed that the elastomer is stable to externalstress over a wide range of strain, frequency, and temperature. Theelastomer showed ~45% reduction in various gas permeabilities byincorporating 0.43 vol% of GO, which highlights the use of thismaterial as an effective gas barrier membrane. Furthermore, thissimple, potentially cost-effective materials design approach shouldenable more versatile usage of polymer/GO composite elastomersin many commercial applications, including as gas barrier or sep-aration membranes.

Author contributions

The manuscript was written through contributions of all au-thors. All authors have given approval to the final version of themanuscript.

Notes

The authors declare no competing financial interest.

H. Ha et al. / Polymer 93 (2016) 53e6060

Acknowledgment

The authors gratefully acknowledge the partial financial supportfrom theWelch Foundation (grant #F-1709) and the Korea CCS R&DCenter (KCRC) grant funded by the Ministry of Science, ICT& FuturePlanning from the Korean government (grant#2015M1A8A1053569). We would also like to acknowledge theKwanjeong Educational Foundation for their financial support of H.Ha.

Appendix A. Supplementary data

Supplementary data related to this article can be found at http://dx.doi.org/10.1016/j.polymer.2016.04.016.

References

[1] J. Zang, S. Ryu, N. Pugno, Q. Wang, Q. Tu, M.J. Buehler, et al., Multifunctionalityand control of the crumpling and unfolding of large-area graphene, Nat. Mater12 (2013) 321e325, http://dx.doi.org/10.1038/nmat3542.

[2] S. Bae, Y. Lee, B.K. Sharma, H. Lee, J. Kim, J. Ahn, Graphene-based transparentstrain sensor, Carbon 51 (2013) 236e242, http://dx.doi.org/10.1016/j.carbon.2012.08.048.

[3] W. Zeng, L. Shu, Q. Li, S. Chen, F. Wang, X.-M. Tao, Fiber-based wearableelectronics: a review of materials, fabrication, devices, and applications, Adv.Mater 26 (2014) 5310e5336, http://dx.doi.org/10.1002/adma.201400633.

[4] M.-C. Choi, Y. Kim, C.-S. Ha, Polymers for flexible displays: From materialselection to device applications, Prog. Polym. Sci. 33 (2008) 581e630, http://dx.doi.org/10.1016/j.progpolymsci.2007.11.004.

[5] K.S. Kim, Y. Zhao, H. Jang, S.Y. Lee, J.M. Kim, K.S. Kim, et al., Large-scale patterngrowth of graphene films for stretchable transparent electrodes, Nature 457(2009) 706e710, http://dx.doi.org/10.1038/nature07719.

[6] K. Choi, S. Nam, Y. Lee, M. Lee, J. Jang, S.J. Kim, et al., Reduced water vaportransmission rate of graphene gas barrier films for flexible organic field-effecttransistors, ACS Nano 9 (2015) 5818e5824, http://dx.doi.org/10.1021/acsnano.5b01161.

[7] C.-C. Yang, Y.-L. Hsu, A review of accelerometry-based wearable motion de-tectors for physical activity monitoring, Sensors 10 (2010) 7772e7788, http://dx.doi.org/10.3390/s100807772.

[8] J. Yang, D. Wei, L. Tang, X. Song, W. Luo, J. Chu, et al., Wearable temperaturesensor based on graphene nanowalls, RSC Adv. 5 (2015) 25609e25615, http://dx.doi.org/10.1039/C5RA00871A.

[9] Y. Wang, L. Wang, T. Yang, X. Li, X. Zang, M. Zhu, et al., Wearable and highlysensitive graphene strain sensors for human motion monitoring, Adv. Funct.Mater 24 (2014) 4666e4670, http://dx.doi.org/10.1002/adfm.201400379.

[10] D.R. Paul, J.E. Mark, Fillers for polysiloxane (“silicone”) elastomers, Prog.Polym. Sci. 35 (2010) 893e901, http://dx.doi.org/10.1016/j.progpolymsci.2010.03.004.

[11] R.R. LeVier, T.H. Lane, M.C. Harrison, R.R. Cook, What is silicone? J. Clin. Epi-demiol. 48 (1995) 513e517, http://dx.doi.org/10.1016/0895-4356(94)00207-7.

[12] K. Efimenko, W.E. Wallace, J. Genzer, Surface modification of sylgard-184poly(dimethyl siloxane) networks by ultraviolet and ultraviolet/ozone treat-ment, J. Colloid Interface Sci. 254 (2002) 306e315, http://dx.doi.org/10.1006/jcis.2002.8594.

[13] Silicone Rubber Chemical Crosslinking Principle http://www.wacker.com/cms/en/products/product_groups/silicone_rubbers_1/vernetzung_1.jsp.

[14] A. Lamberti, M. Di Donato, A. Chiappone, F. Giorgis, G. Canavese, Tunableelectromechanical actuation in silicone dielectric film, Smart Mater Struct. 23(2014) 105001, http://dx.doi.org/10.1088/0964-1726/23/10/105001.

[15] A. Camenzind, T. Schweizer, M. Sztucki, S.E. Pratsinis, Structure & strength ofsilica-PDMS nanocomposites, Polymer 51 (2010) 1796e1804, http://dx.doi.org/10.1016/j.polymer.2010.02.030.

[16] S.D. Burnside, E.P. Giannelis, Synthesis and properties of new poly(-dimethylsiloxane) nanocomposites, Chem. Mater 7 (1995) 1597e1600, http://dx.doi.org/10.1021/cm00057a001.

[17] P.-C. Ma, N.A. Siddiqui, G. Marom, J.-K. Kim, Dispersion and functionalizationof carbon nanotubes for polymer-based nanocomposites: A review, ComposPart A Appl. Sci. Manuf. 41 (2010) 1345e1367, http://dx.doi.org/10.1016/j.compositesa.2010.07.003.

[18] Z. Wang, J.K. Nelson, H. Hillborg, S. Zhao, L.S. Schadler, Graphene oxide fillednanocomposite with novel electrical and dielectric properties, Adv. Mater 24(2012) 3134e3137, http://dx.doi.org/10.1002/adma.201200827.

[19] P. Du, X. Lin, X. Zhang, Tunable electrical and mechanical responses of PDMSand polypyrrole nanowire composites, J. Phys. D. Appl. Phys. 46 (2013)195303, http://dx.doi.org/10.1088/0022-3727/46/19/195303.

[20] B.M. Yoo, H.J. Shin, H.W. Yoon, H.B. Park, Graphene and graphene oxide and

their uses in barrier polymers, J. Appl. Polym. Sci. 131 (2014) 39628, http://dx.doi.org/10.1002/app.39628.

[21] H. Huang, Y. Ying, X. Peng, Graphene oxide nanosheet: an emerging starmaterial for novel separation membranes, J. Mater Chem. A 2 (2014) 13772,http://dx.doi.org/10.1039/C4TA02359E.

[22] L. Gomez De Arco, Y. Zhang, C.W. Schlenker, K. Ryu, M.E. Thompson, C. Zhou,Continuous, highly flexible, and transparent graphene films by chemical vapordeposition for organic photovoltaics, ACS Nano 4 (2010) 2865e2873, http://dx.doi.org/10.1021/nn901587x.

[23] H. Ha, K. Shanmuganathan, C.J. Ellison, Mechanically stable thermally cross-linked poly(acrylic acid)/reduced graphene oxide aerogels, ACS Appl. MaterInterfaces 7 (2015) 6220e6229, http://dx.doi.org/10.1021/acsami.5b00407.

[24] Bay Carbon Inc., SP-1 Graphite Powder http://www.baycarbon.com/SP1Summary.htm.

[25] D.-W. Kang, H.-S. Shin, Control of size and physical properties of grapheneoxide by changing the oxidation temperature, Carbon Lett. 13 (2012) 39e43,http://dx.doi.org/10.5714/CL.2012.13.1.039.

[26] D.R. Dreyer, S. Park, C.W. Bielawski, R.S. Ruoff, The chemistry of grapheneoxide, Chem. Soc. Rev. 39 (2010) 228e240, http://dx.doi.org/10.1039/b920539j.

[27] A.B. Bourlinos, D. Gournis, D. Petridis, T. Szab�o, A. Szeri, I. D�ek�any, Graphiteoxide: chemical reduction to graphite and surface modification with primaryaliphatic amines and amino acids, Langmuir 19 (2003) 6050e6055, http://dx.doi.org/10.1021/la026525h.

[28] N.I. Kovtyukhova, P.J. Ollivier, B.R. Martin, T.E. Mallouk, S.A. Chizhik,E.V. Buzaneva, et al., Layer-by-layer assembly of ultrathin composite filmsfrom micron-sized graphite oxide sheets and polycations, Chem. Mater 11(1999) 771e778, http://dx.doi.org/10.1021/cm981085u.

[29] R.B. Prime, E. Sacher, Kinetics of epoxy cure: 2. The system bisphenol-Adiglycidyl ether/polyamide, Polymer 13 (1972) 455e458, http://dx.doi.org/10.1016/0032-3861(72)90113-9.

[30] K.-C. Chiou, F.-C. Chang, Reactive compatibilization of polyamide-6 (PA 6)/polybutylene terephthalate (PBT) blends by a multifunctional epoxy resin,J. Polym. Sci. Part B Polym. Phys. 38 (2000) 23e33, http://dx.doi.org/10.1002/(SICI)1099-0488(20000101)38, 1<23::AID-POLB3>3.0.CO;2-Y.

[31] S.H. Ryu, J.H. Sin, A.M. Shanmugharaj, Study on the effect of hexamethylenediamine functionalized graphene oxide on the curing kinetics of epoxynanocomposites, Eur. Polym. J. 52 (2014) 88e97, http://dx.doi.org/10.1016/j.eurpolymj.2013.12.014.

[32] W. Hung, C. Tsou, M. De Guzman, Q. An, Y. Liu, Y. Zhang, et al., Cross-linkingwith diamine monomers to prepare composite graphene oxide-frameworkmembranes with varying d -spacing, Chem. Mater 26 (2014) 2983e2990,http://dx.doi.org/10.1021/cm5007873.

[33] S. Cholake, M. Mada, R. Raman, Y. Bai, X. Zhao, S. Rizkalla, et al., Quantitativeanalysis of curing mechanisms of epoxy resin by mid- and near - Fouriertransform infra red spectroscopy, Def. Sci. J. 64 (2014) 314e321, http://dx.doi.org/10.14429/dsj.64.7326.

[34] S. Park, D. a. Dikin, S.T. Nguyen, R.S. Ruoff, Graphene oxide sheets chemicallycross-linked by polyallylamine, J. Phys. Chem. C 113 (2009) 15801e15804,http://dx.doi.org/10.1021/jp907613s.

[35] H. Liu, T. Kuila, N.H. Kim, B.-C. Ku, J.H. Lee, In situ synthesis of the reducedgraphene oxideepolyethyleneimine composite and its gas barrier properties,J. Mater Chem. A 1 (2013) 3739, http://dx.doi.org/10.1039/c3ta01228j.

[36] H. Ha, K. Shanmuganathan, Y. Fei, C.J. Ellison, Thermal stimuli-responsivebehavior of pyrene end-functionalized PDMS through tunable P-P in-teractions, J. Polym. Sci. Part B Polym. Phys. 54 (2016) 159e168, http://dx.doi.org/10.1002/polb.23805.

[37] G. Clarizia, C. Algieri, E. Drioli, Filler-polymer combination: a route to modifygas transport properties of a polymeric membrane, Polymer 45 (2004)5671e5681, http://dx.doi.org/10.1016/j.polymer.2004.06.001.

[38] K. Nawaz, U. Khan, N. Ul-Haq, P. May, A. O'Neill, J.N. Coleman, Observation ofmechanical percolation in functionalized graphene oxide/elastomer compos-ites, Carbon 50 (2012) 4489e4494, http://dx.doi.org/10.1016/j.carbon.2012.05.029.

[39] T.C. Merkel, V.I. Bondar, K. Nagai, B.D. Freeman, I. Pinnau, Gas sorption,diffusion, and permeation in poly(dimethylsiloxane), J. Polym. Sci. Part BPolym. Phys. 38 (2000) 415e434, http://dx.doi.org/10.1002/(SICI)1099-0488(20000201)38, 3<415::AID-POLB8>3.0.CO;2-Z.

[40] E.L. Cussler, S.E. Hughes, W.J. Ward, R. Aris, Barrier membranes, J. Memb. Sci.38 (1988) 161e174, http://dx.doi.org/10.1016/S0376-7388(00)80877-7.

[41] N.K. Lape, E.E. Nuxoll, E.L. Cussler, Polydisperse flakes in barrier films, J. Memb.Sci. 236 (2004) 29e37, http://dx.doi.org/10.1016/j.memsci.2003.12.026.

[42] G. Choudalakis, A.D. Gotsis, Permeability of polymer/clay nanocomposites: Areview, Eur. Polym. J. 45 (2009) 967e984, http://dx.doi.org/10.1016/j.eurpolymj.2009.01.027.

[43] M. a. Osman, V. Mittal, M. Morbidelli, U.W. Suter, Polyurethane adhesivenanocomposites as gas permeation barrier, Macromolecules 36 (2003)9851e9858, http://dx.doi.org/10.1021/ma035077x.

[44] H. Kim, A. a. Abdala, C.W. Macosko, Graphene/polymer nanocomposites,Macromolecules 43 (2010) 6515e6530, http://dx.doi.org/10.1021/ma100572e(For Table of contents use only).