application of carbon nanotubes in polymer · pdf filepolymer electrolyte based fuel cells...

1Application of carbon nanotubes in polymer electrolyte based fuel cells

© 2011 Advanced Study Center Co. Ltd.

Rev. Adv. Mater. Sci. 29 (2011) 1-14

Corresponding author: W. Zhang, e-mail: [email protected]

APPLICATION OF CARBON NANOTUBES IN POLYMERELECTROLYTE BASED FUEL CELLS

Wei Zhang and S. Ravi P. Silva

Nanoelectronics Centre, Advanced Technology Institute, University of Surrey, Guildford, Surrey GU2 7XH,United Kingdom

Received: February 11, 2011

Abstract. Polymer electrolyte based fuel cells (PEFCs) are always in the forefront of fuel cellrevolution. Recently a wide variety of application of carbon nanotubes (CNTs) in PEFC compo-nents has been exploited. The impetus is to improve the PEFC performance by taking advan-tages of CNTs’ extraord]nary phys]cal, chem]cal and electron]c propert]es. Here]n, we br]eflyreview these efforts with an attempt to obtain a better understanding on the role of CNTs inPEFCs, and this article is structured as the following: the contribution of CNTs is first addressedin terms of improving the mechanical strength and proton conductivity of polymer electrolytemembrane; their role in electrocatalysis is then discussed with respect to facilitating the utiliza-tion of noble metal catalysts (platinum) and exploring the platinum-free catalysts; the consider-ation of CNTs as hydrogen storage materials is also highlighted. Based on the literatures stud-ied, CNTs demonstrate great potential as multifunctional materials in improving PEFC perfor-mance.

1. INTRODUCTION

With the ever-growing concerns about the environ-ment pollutions and the exhaustion of energy re-sources, and the commercialization of fuel cell au-tomobiles, there is an urgent need for the develop-ment of reliable, high performance fuel cells thatare operable at low temperature. As such, polymerelectrolyte-based fuel cells (PEFCs), which can beoperated at around 100 °C, have attracted particu-lar attention recently. PEFCs are also distinguish-able because they can provide high power densitywith a short start-up time. PEFCs include protonexchange membrane fuel cells (PEMFCs) usinghydrogen gas as the fuel source and direct metha-nol fuel cells (DMFCs) which generate power usingliquid methanol as the fuel source. Owing to theiradvantageous features, PEFCs have been consid-ered as the promising power sources for portable,automobile, and stationary applications [1-4]. De-

spite the tremendous development of PEFCs, chal-lenges are still there, hampering the large scaledcommercialization of PEFCs. The first barrier is theirhigh cost. For instant, the cost target of the USDepartment of Energy (US D.O.E.) for PEMFC stackis 30 $ per kW by 2015, which is less than one thirdof its current cost (110 $ per kW) [5]. High loadingof the noble metal catalysts on the electrodes andthe use of perfluorosulfonic acid membrane areamong the main contributors to the current highcost. The durability of fuel cell is another propertyneeding to be improved. The major modes of fail-ures are the breakdown of polymer electrolyte mem-brane (PEM) and the loss in active surface area ofmetal catalysts with operating time. To find a suit-able material that is capable of adsorbing and re-leasing large quantities of hydrogen at ambient tem-perature and pressure easily and reliably representsanother challenge. Undoubtedly, the success in over-

2 W. Zhang and S. Ravi P. Silva

Fig. 1. Chemical structure of Nafion.

coming these barriers will witness the proliferationof PEFCs.

On the other hand, carbon nanotubes (CNTs)have set themselves apart as a new kind of materi-als due to their distinct properties. Recently, a num-ber of investigations have been carried out to incor-porate CNTs into the components of fuel cells toimprove the performance and lower the cost. CNTshave high strength and toughness to weight char-acteristics, which stimulates researchers to employthem as reinforcing fillers to improve the mechani-cal strength of PEM; their high surface area andthermal conductivity make them useful aselectrocatalyst supports; they can also be appliedin gas diffusion layers and current collectors be-cause of their high electrical conductivity; their lay-ered graphene tubular structure enables them to bea potential candidate for hydrogen storage. In thisarticle, we will briefly review the recent communica-tions, mainly published in 2005 and after, and dis-cuss the results obtained; this may help us obtaina better understanding on the multifunctional role ofCNTs and improve the fuel cell performance by tak-ing advantage of the unique properties of CNTs.

2. APPLICATION OF CNTs IN PEM

PEM is one of the key components in PEFCs, whichnot only restricts the performance of fuel cell butalso the overall cost. An ideal PEM should be pro-ton conductive, facilitating the transport of protonfrom anode to cathode; but yet electron insulating,avoiding the occurrence of electrical short circuit-ing. It should also have low permeability to the feed-ing fuels, such as hydrogen and methanol, and highchemical stability to withstand the harsh chemicalconditions that the fuel cell is undergoing underoperation. Additionally, it needs to be mechanicallystrong enough to withstand the stress of fabricationand the dimension change (swelling and contrac-tion) due to the change of water content as well.Owing to the presence of sulfonic acid groups,perfluorosulfonate ionomer, such as Nafion (see Fig.1 for its chemical structure), possesses high pro-

ton conductivity. And it is also relatively mechani-cally and chemically stable. Therefore, it seems tomatch most of the requirements as a PEM mate-rial, and is a promising candidate. Indeed, its appli-cations in fuel cells have been explored extensivelyin the past few years.

2.1. Improvement of mechanicalproperties

The strength of the sp² carbon-carbon bonds g]vesCNTs amaz]ng mechan]cal propert]es. The Young’smodulus of the best nanotubes can be as high as1000 GPa which is approximately 5 fold higher thansteel; the tensile strength of nanotubes can be upto 63 GPa, around 50 fold higher than steel. Theseproperties, coupled with their lightness, endow CNTsgreat potential to be applied as reinforcing fillers toimprove the mechanical strength of host matrices[6]. On the other hand, Nafion membranes merelyhave a moderate mechanical stability, especially athigh temperature [7,8]. This may incur problems inthe fabrication of membrane; furthermore, when thefuel cell is in operation, the swelling and contrac-tion of Nafion membrane initiated by the variation inhydration level may change the dimension of mem-brane and affect the fuel cell performance and work-ing life. As a matter of fact, the breakdown of PEMis one of the major modes of failures of fuel cellsystems. To hurdle these obstacles, great effortshave been devoted to improving the mechanicalstrength of Nafion by incorporating CNTs [9].

Liu et al. reported that the incorporation of 1 wt.%CNTs into Nafion could decrease the dimensionalchange of membrane, whilst maintain the protonconductivity [10]. Based on this work, they subse-quently proposed a sandwich structured electrolytemembrane with a platinum (Pt)/CNT/Nafion layerinterpolated between two plain Nafion layers [11].The beauties of this design are (i) as an catalyst, Ptcan facilitate the recombination of H2 and O2 thatpermeate from electrodes in the electrolyte layerand produce H2O, and thus this layer has the self-humidifying function; (ii) as reinforcing fillers, CNTscan improve the mechanical strength of the mem-brane; (iii) as electrical insulators, the two Nafionlayers can remove the worries about the happeningof short circuiting between two electrodes.Thomassion et al also demonstrated that the incor-poration of 2 wt.% carboxyl acid group decoratedmulti-walled carbon nanotubes (MWCNTs) intoNaf]on could ]ncrease the Young’s modulus up to160%, but still maintain the similar level of ionicconductivity with respect to pure Nafion membrane


[12]. By blending 3 wt.% H2O2/NaOH oxidizedMWCNTs into Nafion membrane, the tensile strengthincreased from 18.5 MPa to 28.6 MPa and the elon-gation at break increased from 112% to 142% [13].

It is worth stressing that care must been takenwhen CNTs are used to reinforce Nafion membranethat is going to be applied as ionic conductive mem-brane in fuel cell. This is correlated with anotherd]st]nct property of CNTs—excellent electr]cal con-ductivity. If the quantity of CNTs involved is suffi-ciently high, the membrane will become electronconductive and completely lose its function as theelectron barrier. It seems 2-3 wt.% can be the safethreshold value [12,13], but this can really truly serveas the referenced number and the true value for anyspecific system can only be established by trialsand errors, because the development of electronconductive network by CNTs in polymer matrix isalso controlled by the aspect ratio of CNTs, theirintrinsically electrical property, their dispersion andorientation state in polymer matrix, their interactionswith the matrix at the interfaces, and the phaseseparation effect of a third phase, if applied [14].

2.2. Effect on proton conductivity

The proton conductive ability of polymer electrolytelayer is one of the key parameters that govern theperformance of fuel cell. As such, the effect of in-corporation of CNTs on the proton conductivity ofNafion has become the research topic of many re-searchers.

Cele et al. found that the introduction of HNO3-oxided 1 wt.% MWCNTs or hexadecylamine-functionalized MWCNTs into Nafion decreased itsionic conductivity by a small amount, particularly,in the case of amine functionalized MWCNTs, whichmight be ascribed to the presence of amine surfac-tant cha]ns on the tubes’ outer layer and the poordispersion of CNTs in Nafion [7]. By covalently graftinghydrophilic layers composed of poly(oxyalkylene)diamines and tetraethyl orthosilicate-reinforced polysiloxane in a layer-by-layer manner,MWCNTs were functionalized with polysiloxane,which were then blended with Nafion to form a pro-ton-conducting membrane [15]. The proton conduc-tivity of these membranes kept stable up to 130 °C;unlike pure Nafion, the proton conductivity of whichdropped significantly when the temperature was over100 °C. The authors ascribed this to the strongerbinding of water inside the membrane, particularlythe stronger bond]ng of water on –CH2CH2O- and –SO3-NH3

+ linkages. This finding is exciting, sincethe proton conductivity of Nafion increases expo-

nentially with increasing temperature and it is likelythat fuel cell performance will be improved by oper-ating at elevated temperatures. Another bonus ofoperation at elevated temperatures is its beneficialeffect on electrode kinetics, promoting the carbonmonoxide (CO) tolerance of electrocatalyst and low-ering the loading of precious metal, for instant, Pt[16]. Nafion is conductive to proton in that its sul-fonic acid groups provide the active sites for thehopping of proton upon hydration [17]. Hence, theincrease of sulfonic acid group concentration could,in principle, increase the proton conductivity ofNafion. By mixing the sulfonic acid functionalizedsingle walled carbon nanotubes (SWCNTs) andNafion, followed by heat-cast at 70 °C, a decoratedNafion membrane was prepared and it was observedthat the proton conductivity was enhanced by oneorder of magnitude [8].

Despite its excellent ionic conductivity, Nafionis costly and can only work efficiently with a highlevel of humidity [18]. These limitations are the driv-ing forces for exploring other cheaper ionic conduc-tors that may be used to replace Nafion [19-21].Kannan et al. implanted 2-aminoethylphosphonicacidfunctionalized MWCNTs into polybenzimidazolemembrane and observed that the proton conductiv-ity of membrane was improved by 50% (from 0.07 Scm-1 to 0.11 S cm-1) [22]. By the solution castingmethod, Zhou et al. manufactured a series of com-posite membranes consisting of sulfonatedMWCNTs and sulfonated poly (ether sulfone etherketone ketone) [23]. It was reported that the intro-duction of sulfonated MWCNTs could increase theproton conduct]v]ty from 2.94×10-3 S cm-1 to 4.39×10-3 S cm-1, depending on the quantity of MWCNTsinvolved; while the reverse trend was observed forraw MWCNTs, where the proton conductivity droppeda bit on the blending of MWCNTs. Irrespective ofthese efforts, Nafion is still preferred in terms of ionicconductivity, which is generally 10 fold higher thanthese proposed polymers [12,13,15,24,25].

3. APPLICATION OF CNTs ASCATALYST SUPPORTS

One dominant obstacle to the commercialization offuel cells is their high cost. Even though Pt is usedcurrently as the benchmark catalyst, it has to beultimately eliminated from the catalyst layer con-sidering the very low abundance of Pt on the earthcrust (3.7×10-6) and its fluctuating cost [26]. It hasbeen estimated that the fuel cell electrode encom-passing the carbon backing layer and Pt preciousmetal represents about 40% of the total cost of a


fuel cell system [27]. Thus, there is a clear demandand urgency to reduce the quantity of Pt applied,which is intimately related to reducing a major costcontributor of fuel cells.

As innovative catalyst supports, CNTs havedrawn a great deal of attention due to their uniquegeometric shape, excellent mechanical and ther-mal properties, high electric conductivity, large sur-face areas, and fascinating chemical stability. Ascompared with the most widely used Vulcan XC-72R carbon black (CB) support that has an elec-tronic conductivity of 4.0 S cm-1 and specific sur-face area of 237 m2 g-1 [28], CNT has extremelyhigher electronic conductivities of 104 S cm-1 andsignificantly high specific surface areas of 200-900m2 g-1 [29]. Furthermore, Vulcan XC-72R has a largeratio of micropores that are smaller than 2 nm,whereas the CNT has no micropores smaller than 2nm. For the Vulcan XC-72R support, the Ptnanoparticles may sink into the micropores andbecome inaccessible for triple-phase-boundary for-mation, reducing the Pt utilization [30]; for the CNTsupport, however, no Pt particles may sink in themicropores and be wasted [31]. Additionally, VulcanXC-72 undergoes corrosion, particularly sensitive toperoxide species environment, resulting in the ag-gregation and dissociation of Pt nanoparticles [32];whilst the chemical inertness of CNTs can offer an-other bonus on this issue [33]. Therefore, in the pastfew years, continuous efforts have been devoted toincreasing the utilization efficiency of Pt and reduc-ing Pt loading by taking advantage of the uniqueproperties of CNTs.

3.1. Improvement of the utilization ofPt catalysts

3.1.1. Commercial Pt black catalysts

Girishkumar et al. demonstrated that SWCNTs couldserve as an excellent support to anchor Pt blackparticles and carry out electrochemical oxidationand reduction reactions effectively, similar to theexisting commercial CB support [34]. Films ofSWCNTs and commercial Pt black were sequen-tially cast on a carbon fiber electrode (CFE) using asimple electrophoretic deposition procedure. Elec-trochemical impedance spectroscopy revealed thatthe CNT-based electrodes exhibited an order ofmagnitude lower charge-transfer reaction resistancefor the hydrogen evolution reaction than did the com-mercial CB-based electrodes, 44 vs. 496 . ThePEM assembly fabricated using the CFE/SWCNT/Pt electrodes (as depicted in Fig. 2) was evaluatedusing a fuel cell testing unit operating with H2 andO2 as input fuels at 60 °C. The maximum powerdensity obtained using CFE/SWCNT/Pt electrodesas both the anode and the cathode was 20% higherthan that using the CFE/CB/Pt electrodes. By us-ing a thin film rotating disk electrode, Kongkan et alevaluated the oxygen reduction reaction at Pt blackparticle/SWCNTs catalysts [35]. Compared with acommercial Pt/CB catalyst, Pt/SWCNT films caston a rotating disk electrode exhibited a lower onsetpotential, i.e. 10 mV shift, for oxygen reduction. Whatis more, in durability tests, SWCNT exhibits agreater stability to anchor Pt particles; this is be-

Fig. 2. SWCNT-based PEM assembly for a H2/O2-based fuel cell [34]. Reprinted with permission from G.Girishkumar, M. Rettker, R. Underhile, D. Binz, K. Vinodgopal, P. McGinn and P. Kamat, Langmuir 21(2005) 8487, © 2005 Amer]can Chem]cal Soc]ety.


Fig. 3. Schematic of the fuel cell architecture. E-TEK 20 wt % Pt/XC-72 is used in the catalyst layerof anode. The cathode contains the thin film Pt/MWNT in the catalyst layer. The Nafion NRE-212 isused as a PEM. The assembly of the carbon back-ing layer and microporous layer on both anode andcathode are SGL Carbon Group products (GDL25CC) [37]. Reprinted with permission from J. M.Tang, K. Jensen, M. Waje, W. Li, P. Larsen, K.Pauley, Z. Chen, P. Ramesh, M. E. Itkis, Y. Yanand R. C. Haddon, J. Phys. Chem. C 111 (2007)17901, © 2007 Amer]can Chem]cal Soc]ety.

cause SWCNTs can minimize the Pt aggregationeffect during long term usage.

3.1.2. Pt catalysts produced bychemical reduction ofhexachloro-platinic acid(H2Pt Cl6)

Ethylene glycol (EG) reduction

EG is one of the most popular agents that are em-ployed to reduce H2PtCl6 to obtain Pt particles. Liet al. developed a partially oriented MWCNT film forPEMFC usage [36]. Pt/MWCNTs were prepared byEG reduction of H2PtCl6; the obtained catalysts werethen filtrated with nylon. This obtained film was sub-sequently transferred to a Nafion membrane by ahot-press method. A PEMFC with the oriented CNTfilm as the cathode achieved higher single-cell per-formance than those with CB and a disordered CNT-film-based cathode in 0-0.2 A cm-2 current densityrange. This was proposed to be associated with thealigned arrangement of CNTs, which is favourablefor the improvement of fuel cell performance in that(i) the electrical conductivity of CNTs is much higherin axial direction than in radical direction, and there

is no energy loss when electron transfer along thetubes; (ii) higher gas permeability is expected withthe oriented CNT film, promoting fuel utilization; (iii)an oriented film may exhibit superhydrophobicity,facilitating water removal within the electrode andimproving mass transport in PEMFCs. Using ananalogous filtration method, Tang et al. prepared athin 1.3 mm catalyst layer with Pt nanoparticles(2.1 nm) supported on 360 m long MWNTs (Pt/MWNT) [37]. When this catalyst layer was incorpo-rated into the cathode (see Fig 3), a peak powerdensity of 431 mW/cm2 (at a current density of 1.15A/cm2) was achieved in a H2/O2 fuel cell operatingat 70 °C. The merit of this technique is no additionalionomer (e.g. Nafion) is applied in the catalyst layer,since such a thin catalyst layer allows for a highersurface area to volume ratio, giving smaller travel-ling distances for protons to percolate to the cata-lyst surface sites from the PEM.

As stated previously, in a general polyol process,CNTs are first subjected to oxidation pretreatmentusing a strong acid, such as HNO3 or a mixture ofHNO3/ H2SO4, so as to remove impurities and gen-erate suff]c]ent amounts of funct]onal groups, e.g. –OH, –COOH, –CO, etc. on the surfaces. Thesesurface functional groups have stronger attractionforces toward metal ions than bare CNT surfacesand some even have ion exchange capabilities, suchas carboxylic acid groups. Therefore, they are be-lieved to work as metal-anchoring sites and facili-tate metal nuclei formation and electrocatalystsdeposition [38]. Yue et al. developed a reductionroute for the direct immobilization of Pt nanoparticleson nitrogen-containing carbon nanotubes (NCNTs)without the necessity of pre-surface modification onCNTs [39]. NCNTs w]th n]trogen content of 3–5%were synthesized at 700 °C by the chemical vapordeposition (CVD) process. Pt nanoparticles withdiameters ranging from 3-9 nm were supported onNCNTs by the EG reduction. The Pt/NCNTs cata-lyst demonstrated better electrocatalytic activity overPt/CNTs catalyst for methanol oxidation accordingto the cyclic voltammetry analyses. This can beinterpreted by the good dispersion of Pt nanoparticleson the NCNTs due to the N-participation as well asthe homogeneous n-type or metallic behaviour ofNCNTs.

It is also recognized that the dispersion state ofCNTs in solution affects the deposition of metal cata-lysts. When suspended in solution, pristine CNTslike to form large bundles due to the strong van derWaals interactions between them. Such a bundlingeffect leads to a poor dispersion of metal particleson CNTs and limits the overall electrocatalytic ac-


tivity. To increase the active surface area of CNTsand facilitate the diffusion of reactant species,Kongkanand et al suggested to wrap and disperseCNTs with poly(sodium 4-styrenesulfonate) first, andthen deposit Pt particles on polymer wrapped CNTsby the EG reduction of PtCl6

2- [40]. The obtainedCNT supported Pt showed 10 fold high peak currentthan the commercial CB supported Pt in the oxida-tion of methanol at the comparable Pt loading (30wt.% Pt).

Hydrogen (H2) reduction

Suitable Pt particles are required in fuel cell appli-cations; this can be attributed to the structure sen-sitive nature of the chemical reaction and the factthat particles with different sizes will have differentdominant crystal planes and hence the differentintercrystallite distances, which may influence theadsorption of feeding fuels. To deposit small Pt par-ticles on CNTs, CNTs were first functionalized witharyl dizaonium salt and then treated with concen-trated H2SO4, the experimental procedure is diagram-matically shown in Fig. 4 [41]. These attached sul-fonic acid groups provided active sites for Pt ion(PtCl6

2-) adsorption; after the reduction with H2 at600 °C, Pt particles with the diameters of 2-2.5 nmwere uniformly deposited on CNTs. The small sizeof Pt nanoparticles can be attributed to the pres-ence of bulky functional groups on CNTs which actas molecular sites for the Pt ion adsorption and alsoavoid nonuniform Pt deposition on the CNT surface.

Fig. 4. Scheme for Pt deposition on CNTs on carbon paper [41]. Reprinted with permission from M. M.Waje, X. Wang, W. L] and Y.Yan, Nanotechnology 16 (2005) S395, © 2005 IOP Publ]sh]ng, Ltd.

A membrane and electrode assembly (MEA) with aPt/CNT electrode as cathode and an E-TEK Pt/Vulcan XC-72 electrode as anode offered better per-formance than a conventional E-TEK MEA.

By H2 reduction of H2PtCl6, Kaempgen et al de-posited Pt particles on SWCNTs and developed aSWCNT network based multifunctional electrode[42,43]. With respect to the commercially availableamorphous carbon based electrode, the conductiv-ity of this electrode was increased by more thantwo orders of magnitude (2000 S cm-1 vs. 2 S cm-1),whereas the thickness and mass were decreasedby more than one order of magnitude (10-20 m vs.400 m and 0.7 mg cm-2 vs. 24 mg cm-2, respec-tively). Although the performance of this electrodein H2/O2 fuel cell is still inferior to the commercialamorphous carbon electrode, this technique doespave ways to make highly conductive but extremelythin and light gas diffusion electrodes by using CNTnetworks.

The incorporation of nitrogen also contributes tothe improvement of catalyst performance. NCNTswere synthesized by impregnating polyvinylpyrroli-done inside the alumina membrane template andsubsequent carbonization of the polymer. Pt par-ticles were then dispersed onto the nanotubes bythe H2 reduction of H2PtCl6. The NCNT supportedelectrodes showed a ten fold increase in the cata-lytic activity in methanol oxidation compared withthe commercial E-TEK electrode; the peak currentdensity values were 13.3 mA cm-2 and 1.3 mA cm-2,respectively [44]. The authors believed that the ni-


trogen functional group on the CNT surface intensi-fies the electron withdrawing effect against Pt andthe decreased electron density of Pt facilitates theoxidation of methanol.

Sodium borohydride (NaBH4)reduction

NaBH4 offers another option as a reducing agent toreduce H2PtCl6. Reddy et al prepared SWCNT sup-ported Pt catalysts by the reduction of H2PtCl6 withNaBH4-NaOH [45]. Pt particles with a size of about3-5 nm were uniformly deposited on SWCNTs. Again,prior to the metal deposition, SWCNTs were treatedwith HNO3 to obtain anchoring sites for metal pre-cursor salt. The PEMFC with both the anode andcathode containing 50 wt.% Pt/SWNT and 50 wt.%Pt/C nanocomposite gives a better current density(at 540 mV) of 485 mA cm-2 and power density of262 mW cm-2 than E-TEK/E-TEK one (258 mAcm-2 and 139 mW cm-2 respectively).To improve theutilization of noble Pt catalysts in PEFCs, Du et al.modified Pt/CNT catalysts by tethering sulfonic acidgroups onto the surface of CNT supports by ther-mal decomposition of ammonium sulphate,(NH4)2SO4, and in situ radical polymerization of 4-styrenesulfonate [46]. The electrodes with the Pt/CNT catalysts sulfonated by the in situ radical po-lymerization of 4-styrenesulfonate exhibited betterperformance than those with the unsulfonated coun-terparts, mainly resulting from the easier access ofprotons and the well dispersed distribution of Ptcatalysts. The electrodes with the Pt/CNT catalystssulfonated by the thermal decomposition of(NH4)2SO4, however, did not yield the expected per-formance as in the case of CB supported Pt cata-lysts, presumably due to the significant agglomera-tion of Pt particles on the CNT surface at high tem-peratures. This result indicates, for the Pt/CNT-based PEFCs, the sulfonation of CNTs can be anefficient approach to improve performance and re-duce cost. However, the introduction of sulfonic acidgroup will unavoidably increase the hydrophilicity ofCNTs and decrease their electrical conductivity, af-fecting the water management and electron trans-port property. To balance these effects and achievean optimum overall performance, an optimum sul-fonation degree should be targeted.

3.1.3. Pt-alloy catalysts

Pt-based electrocatalysts display the significantactivity and necessary stability in the harsh envi-ronment of the PEFCs. However, the hydrogen-rich

feeding gas in PEMFCs, produced by reforming orpartial oxidation of hydrocarbons, usually containsCO as well; for DMFCs, the partial oxidation ofmethanol can also result in the presence of CO.The existence of CO will create troubles for the Ptcatalyst based full cell systems, as CO can adsorbstrongly on the Pt surface and block the active sitesfrom adsorbing fuels (H2 or methanol), deterioratingthe performance of electrode. One possible solu-tion for this CO poisoning problem involves the useof CO-tolerant electrocatalysts formed by alloyingPt with a second transition metal (e.g. Ruthenium,Ru, Molybdenum, Mo) [47,48]. Among all the bi-nary alloys, the Pt-Ru has shown the most promis-ing performance for both the hydrogen oxidation re-action in the presence of CO and the oxidation ofmethanol. This is because Ru can form oxygen-ated species at a lower potential than Pt and pro-mote the oxidation of CO to CO2 [49].

By H2 reduction of H2PtCl6 and ruthenium chlo-ride (RuCl3), Carmo et al. prepared MWCNT,SWCNT, and CB (Vulcan XC 72) supported Pt-Rucatalysts [50]. When the feeding H2 in PEMFCswas mixed with 100 ppm CO, Pt-Ru catalysts thatwere supported on both CNTs and CBs still exhib-ited good performance, demonstrating high CO tol-erance. The power densities accomplished on asingle DMFC exceeded 100 mW cm-2 at 90 °C and0.3 MPa and the activity of the anodes followed thedescending sequence of Pt-Ru/MWNT, Pt-Ru/CB,and Pt-Ru/SWNT. Similarly, Pt-Ru/MWCNTs cata-lyst was prepared by the NaBH4 reduction of H2PtCl6and RuCl3. When the anode and cathode of DMFCswas fabricated using Pt-Ru/MWNT and 1:1 Pt/MWNT+Pt/C electrocatalyst, respectively, a maxi-mum power density of 39.3 mW cm-2 at 80 °C (witha methanol flow rate of 1 ml/min) was obtained [51].Using the same reducing agent, Prabhuram et alalso synthesized Pt-Ru/MWCNTs catalysts [52]. Ina single DMFC, the power density yielded usingMWCNT supported Pt-Ru was 35-39% higher thanthat using CB (Vulcan XC-72) supported Pt-Ru. Theyattributed this to the influence of supporting materi-als on the crystalline nature of Pt and proposed thatthe distinctive Pt crystallite phases, i.e., Pt (110),on the Pt-Ru particles supported on the MWCNTsis likely to be the reason for enhancing the activityof the methanol oxidation. Li et al compared thecatalytic performance of Pt-Ru loaded SWCNTs,DWCNTs, MWCNTs in DMFCs [53]. Pt-Ru/CNTswere prepared by the EG reduction of H2PtCl6 andRuCl3, the obtained catalysts were then filtrated withnylon and transferred to a Nafion membrane by ahot-press method, as illustrated in Fig. 5. The


Pt-Ru/DWNTs catalyst showed the highest perfor-mance for methanol oxidation reaction in rotatingdisk electrode experiments and as an anode cata-lyst in DMFC single cell tests. It is notable that theDMFC single cell with Pt-Ru/DWNTs (50 wt.%, 0.34mg Pt-Ru/cm2) produces a 68% enhancement ofpower density, and at the same time, an 83% re-duction of Pt-Ru electrode loading when comparedto Pt-Ru/C (40 wt.%, 2.0 mg Pt-Ru/cm2).The rea-son underlying this is still not fully understood andthe authors tentatively associated this with the uniqueproperties of DWCNTs such as electrical conduc-tivity, surface area, and diameter etc, with respectto other types of CNTs.

One challenge facing the preparation of Pt-Rualloy based catalysts is to obtain the proper atomicratio of Pt and Ru, since Pt and Ru ions or saltscan not be reduced with the same rate at the samepH. For example, the pH value for Pt deposition canbe as high as 12, but Ru salts will precipitate underthis condition. Jeng et al succeeded in preparingPt-Ru/CNTs catalysts with the desired atomic ratioof Pt and Ru (1:1) by creating a deposition environ-ment on the CNT surfaces near the isoelectric point[54]; at this specific point, the composition of formedPt-Ru catalyst can be controlled by tailoring thecomposition of metal ions in the deposition solu-tion. When the synthesized catalyst was used asan anode catalyst in a single cell DMFC operatingat 60 °C, a maximum power density of 62 mWcm-2 was obtained around 0.21 V with a currentdensity of 300 mA cm-2.

3.2. Application of CNTs in theexploration of Pt- free catalysts

Apart from the improvement of Pt utilization effi-ciency, the entire replacement of Pt with othercheaper catalysts provides another alternative ap-proach to reduce the prohibitive price of fuel cellsystem.

Molybdenum (Mo) is cheaper and more abun-dant than Pt, and more important molybdenum car-bide (Mo2C) possesses similar electronic states asPt. Based on these facts, Mo2C/CNTs has beensuggested as the anode catalyst by Nakamura etal [55,56]. Ultroviolet light was first applied to cre-ate defects on CNTs; after adsorption on CNTs,molybdenum dioxo acetyl acetonate was reducedwith H2 and carburized in CH4. In contrast to Pt-basedelectrodes, these Mo2C/CNT electrodes need towork at a much higher overvoltage. Although theiractivity may be further improved, at current stage,they still can not compete with Pt catalysts.

Transition metal-nitrogen-carbon nanotube (TM/N/CNT) catalyst is another focus of recent research.The active site for oxygen reduction reaction on suchkind of catalyst containing iron is proposed to com-prise iron and nitrogen and can be expressed bythe formula FeNx (x = 2,4) [57]. Using ferrocene asan iron additive catalyst and ammonia as the N-dopant, Yang et al. prepared vertically aligned CNTswith built-in FeN4 sites by CVD process [58]. Theobtained CNTs exhibited catalytic activity in an O2-saturated acidic electrolyte, analogous to the con-

Fig. 5. Schemat]c ]llustrat]on of Pt”Ru/CNTs-f]lm-based membrane electrode assembly for DMFC [53].Reprinted with permission from W. Li, X. Wang, Z. Chen, M. Waje and Y. Yan, J Phys Chem B 110 (2006)15353, © 2006 Amer]can Chem]cal Soc]ety.


ditions on the cathode of PEFCs. X-ray absorptionspectroscopy revealed that the active site is com-posed of a divalent iron coordinated with fournitrogens in a slightly distorted squareplanar con-figuration. Prehn et al produced NCNTs in a CVDmethod by adding acetonitrile as nitrogen sourceand found that these obtained NCNTs exhibitedrather low catalytic activity in oxygen reduction re-action [59]. However, by impregnating CNTs withiron-tetramethoxyphenylporphyrin chloride and thenheating, the electocatalystic activity was enhancedgreatly, reflecting the beneficial effect of iron. Ac-cording to density functional theory, Titov et al. con-ducted the ab initio studies on the energetics andthe geometries of the metal-nitrogen sites in CNTs[60]. The results showed that the iron atom bindingenergy in both Fe-2N and Fe-4N configurations ismuch higher than that on pristine CNTs, suggestingthat N stabilizes Fe on CNTs. For Fe-2N configura-tions, the binding energy is generally lower than thatin Fe-4N configurations. In the case of Fe-4N con-figurations, theoretical results showed that thepyridinic configurations are energetically preferredover pyrrolic configurations. The results also illus-trated the preferable incorporation of Fe-4N sitesinto large-diameter CNTs with regard to small-diam-eter ones. These results suggested the iron atomsare embedded into the surface of CNTs. However, itneeds to be pointed out that the role of metal in theoxygen reduction reaction is still a controversial is-sue.

Matteret al. synthesized [61] NCNTs by the py-rolysis of acetonitrile at 900 °C using iron/nickelparticles supported on alumina as catalysts. Themost active catalysts in oxygen reduction reactionwere prepared over alumina impregnated with up to2 wt.% iron, but the catalysts that were preparedby the acetonitrile pyrolysis over alumina withoutany metal doping still had significant activity. Thelack of correlation between iron content and activitymakes them suggest that an iron phase formedduring the heat treatment is not the primary sourceof the oxygen reduction reaction activity. Similarly,by pyrolyzing a suitable polymer precursor, includ-ing poly(phenylacetylene), poly(4-vinylpyridine),poly(3-methylpyrrole), poly(2-methyl-1-vinylimidazole), and poly(p-pyridazine-3,6-diyl),within the cylindrical and uniform-sized pores of analumina membrane, Rao et al. synthesized NCNTswithout the involvement of any metal particles [62].All obtained NCNTs exhibited a certain extent ofcatalytic activity, demonstrating that the present ofmetal phase is not a prerequisite to obtain catalytic

activity. They further studied the effect of N contentin carbon lattice on the electrocatalytic activity, andfound poly (2-methyl-1-vinylimidazole) gave the op-timum catalytic performance. As the surface N con-tent in the obtained CNTs using the aforementionedpolymers is 0, 4.3, 5.6, 8.4, and 10.7 at.%, respec-tively. They concluded that there exists an optimumN concentration (8.4 at.%,), and ascribed this to ahigher density of pyridinic-type nitrogen active sites.Analogously, NCNTs were prepared by the pyroly-sis of acetonitrile over cobalt catalysts at differenttemperatures. A detailed analysis of the Co 2p spec-tra and N 1s spectra of NCNTs indicated that a verysmall amount of cobalt nitride was likely to bepresent in the sample. On the other hand, the au-thors did not observe any voltammetric peak thatwould correspond to the oxidation of metallic cobaltor reduction of cobalt oxide on the surface of theCNTs. Transmission electron microscopy studiesof the NCNTs showed that the cobalt particles arecovered by at least a few graphene layers, indicat-ing that the cobalt nanoparticles are not accessibleby the electrolyte. Based on these facts, the au-thors suggested that cobalt is not part of the activesites. The experimental results also revealed thatthe NCNT samples prepared at 550 °C exh]b]tedhigher activity in contrast to the samples preparedat 750 °C, wh]ch could be attr]buted to h]gher amountof pyridinic groups presented in the former samples[63].

Despite these arguments regarding the virtualrole of metal in the oxygen reduction reaction, thereseems to be consensus in literatures that pyridine-type nitrogen can be considered to be responsiblefor the high catalytic activity; this is somewhat inagreement with the N 1s X-ray photoelectron spec-tra on NCNTs, where three principle types of nitro-gen coordination, namely pyridinic, pyrolic, andquaternary, can be identified [64,65]. In terms ofelectrocatalytic activity, TM /N/CNT catalysts remaintoo low to be considered for real fuel cell applica-tions, in general only about one tenth of the activityof Pt-based catalysts [66].

As discussed above, it is an efficient and effec-tive way to decrease the Pt loading by dispersingPt particles on CNTs to increase Pt utilization effi-ciency. In general, small particle size and uniformdispersion will result in high electrocatalytic activ-ity; the incorporation of nitrogen into CNTs can alsoimpose a positive effect. Additionally, both the typeof CNTs and their arranged manner on substratehave a role to play with respect to the catalyticalactivity.


Ranges of CNT charality and radiusFracture Strained No strain or Self collapse

self collapse

Single-wall CNTs (7,7) and down (8,8) to (18, 18) (19,19) to (30, 30) (31,31) and up <0.47 nm 0.54-1.2 nm 1.3-2.0 nm >2.1 nm

Bundle of (7,7) and down (8,8) to (21,21) (22,22) to (30,30) (31,31) and upsingle-wall CNTs <0.47 nm 0.54-1.4 nm 1.5-2.0 nm >2.1 nmDouble-wall CNTs (11,11) and down (12, 12) to (31,31) (32,32) to (36,36) (37,37) and up(inner tube) <0.75 nm 0.81-2.1 nm 2.2-2.4 nm >2.5nmTriple-wall CNTs (16,16) and down (17,17) to (41,41) -- (42,42) and up(innermost tube) <1.1 nm 1.2-2.8 nm >2.8 nm

Table 1. Behaviour of different carbon nanotubes to achieve the goals of hydrogen storage (62 kg H2/m3 and

6.5 wt.%) [70].

4. APPLICATION OF CNTs INHYDROGEN STORAGE

Developing safe, cost-effective, and practical meansof storing hydrogen is crucial for the advancementof fuel cell technologies [67], because hydrogen hasa much higher per unit mass energy density thantypical hydrocarbon fuel, but a poorer per unit vol-ume energy density [68]. This high occupied spacewill create problems for mobile applications, and theUS D.O.E. has set up the criteria with respect tothe hydrogen storage density (9 wt.%, in 2015) forthe practical applications of fuel-cell powered ve-hicles [5]. CNTs have been viewed as a possiblesolution to the storage of hydrogen in fuel cell-pow-ered vehicles due to their low density, high strength,and hydrogen adsorption characteristics.

Hydrogen can be adsorbed by CNTs in aphysisorption manner, that is to say, hydrogen istrapped inside the cylindrical structure of thenanotubes or in the interstitial sites betweennanotubes. By studying the H2 adsorption in the(4,4) armchair SWCNTs with density function theorymethod, Mahdavian pointed out the efficient hydro-gen storage in CNTs requires the control of tubesize, pressure, and temperature [69]. Chen et alestablished a continuum mechanics model to studythe hydrogen storage in SWCNTs, MWCNTs andthe bundle of SWCNTs [70]. The results also indi-cated the importance of tube size in affecting hy-drogen storage and the authors concluded that tinyCNTs can not achieve the goals of hydrogen stor-age without fracture, see the 2nd column in Table 1;small CNTs are strained during hydrogen storage(the 3rd column); medium CNTs can achieve theabove goals without the strain and do not self col-lapse (the 4th column); and large CNTs may self

collapse upon the release of hydrogen (the 5th col-umn).

Doping of certain metal ions was also predictedto be beneficial for the hydrogen storage in CNTs. Afirst-principles study suggested that one SWCNTcoated with single titanium (Ti) atom can bind up tofour hydrogen molecules and at high Ti coverage, aSWCNT can absorb up to 8 wt.% hydrogen [71].Similarly, density-functional calculations of the ad-sorption of molecular hydrogen on the external sur-face of a (4,4) CNT, undoped and doped with lithium(Li), have been carried out by Cabria et al [72]. Thecalculations predicated that the physisorption bind-ing energy of hydrogen on Li-doped CNTs is almostdoubled when compared with that on pure CNTs,54-84 meV/molecule vs. 147-173 meV/molecule.The analysis of the charge density showed the ori-gin of this increase: the charge transfer from the Liatom to the CNT induces a larger electron densityof the surface region near the Li impurity, whichenhances the interaction between the hydrogenmolecule and the Li-doped surface, resulting in theenhancement of hydrogen storage ability. To furtherincrease the adsorption of hydrogen, Tylianakis etal. built up a nice 3D carbon nanostructure modelwith large surface area and tunable pore size [73].The model consisted of parallel graphene layers atvariable distance, stabilized by CNTs placed verti-cally to the graphene planes (see Fig. 6). In thisarchitecture, CNTs supported the graphene layerslike pillars and assembled a 3D block reducing theempty space between the graphene layers by fillingit with CNTs. When this material was doped with Liions, it could store up to 41 g H2/L under ambientconditions, only slightly lower than the volumetriccapacity (45 g H2/L) required to be matched in 2010by the US D.O.E.


Fig. 6. Snapshot from the GCMC simulations of Li doped pillared structure at 77K and 3 bar. Hydrogenmolecules are represented in green while Li atoms are in purple [73]. Reprinted with permission from D. E.Tyl]anak]s and G. E. Froudak]s, Nano Lett 8 (2008) 3166, © 2008 Amer]can Chem]cal Soc]ety.

In the above mentioned studies, the interactionbetween H2 and CNTs is dominated by weak vander Waals forces, and thus only a limited quantityof hydrogen can be stored. As such, chemisorptionof hydrogen on CNTs has been suggested, wherethe molecular hydrogen is dissociated with the aidof catalyst and the resulting atomic hydrogen sub-sequently forms bond with the unsaturated carbonbonds along the tube.

Obviously, this method offers safety benefit be-cause the hydrogen atoms are bonded to carbonatoms, rather than freely floating as a potentiallyexplosive gas. By applying an atomic hydrogenbeam to SWCNT thin film, new C-H bonds werecreated, and such C-H bonds could be completelybroken by heating to 600 °C, reflecting the revers-ible nature of this process[74]. The authors demon-strated an approx]mately 65±15 at.% hydrogena-tion of carbon atoms in the SWCNTs, which isequ]valent to 5.1± 1.2 wt.% hydrogen capac]ty. Th]sresult is marvellous, but challenges still exist. Thestored hydrogen could be released by breaking theC-H bond at above 600 °C, but two cycles of hydro-genation and de-hydrogenation appeared to causedefects in the tubes. Not to mention, ideally, thehydrogen should be released at 50 to 100 °C. Add-ing metal catalysts and adjusting the radius of thetubes are potential solutions.

The theoretical and experimental studies indi-cate CNTs hold a great promise for future mobilehydrogen storage. So far, however, no practical prod-uct is envisionable, even at laboratory scale. To turnthis potential into a reality, it is essential for physi-cist, chemist and engineer to come together andcooperate with each other.

5. SUMMARY

The overall performance of PEFCs is controlled bydifferent processes like diffusion of the feeding fuelto the electrolyte-catalyst interface, electrochemi-cal reaction at the catalyst-electrolyte interface,transport of charges by the current collector, protontransport across the membrane, and the hydrationof electrolyte membrane etc. As illustrated, CNTshave demonstrated great performance in PEFC com-ponents. When blended with Nafion to make PEM,CNTs can improve the mechanical strength of themembrane; if functionalized with sulfonic acid group,they can improve the proton conductivity as well.Pt/CNTs and Pt-Ru/CNTs exhibit higherelectrocatalytic activity than the CB supported coun-terparts; the doping of nitrogen also has a benefi-cial role to play with respect to catalytic activity;CNTs have also been adopted to explore Pt-freecatalysts, e.g. TM/N/CNTs. New CNT based mod-elling structures show high hydrogen storage abil-ity, which almost meets the requirement set by theUS D.O.E.

Even though CNTs have been proved to be valu-able in improving the PEFC performance, few chal-lenges still need to be overcome before the wide-spread use of CNTs in PEFCs. One challenge arisesfrom the inhomogeneous nature of as-grown CNTs.Regardless of the growth methods, the obtainedCNTs are always with scattering helicity, number oflayers, diameter, length, and topological defects.This, in turn, will affect their electrical properties,chemical stability and interactions with other spe-cies, leading to the difficulty in predicting and ma-nipulating their performance in PEFCs. CNT purity


is another concern. Depending on the growth methodand growing conditions, as-grown CNTs are usuallyaccompanied with impurities, such as amorphouscarbon, graphite particles and metal particles. Theknowledge on how these impurities can affect thePEFC performance is still not fully established.Additionally, so far only limited data is available re-garding the toxicity of CNTs, and still not much isknown about their impact on biological systems in-cluding humans.

In brief, the research efforts, made so far, havelead to the successful incorporation of CNTs intoPMFC components; the outcome is the improvedperformance of PEFCs in many aspects. This in-deed raises the hope for the proliferation of PEFCsby combining the virtues of fuel cells with CNTs,and it is likely that the further advances in CNT growthand purification techniques will accelerate this pro-cedure.

ACKNOWLEDGEMENT

We would like to thank Dr. Martin Blissett for hisuseful discussions.

REFERENCES

[1] S. Litster and G. McLean // J. Power Sources130 (2004) 61.

[2] S. M. Haile // Acta. Mater. 51(2003)5981.[3] V. Mehta and J.S. Cooper // J. Power Sources

114 (2003) 32.[4] A. F. Ghenciu // Curr. Opin. Solid State Mater

6 (2002) 389.[5] Multi-Year Research, Development and

Demonstration Plan: Planned programactivities for 2005-2015, US Department ofEnergy, http:// www.eere.energy.gov/hydrogenandfuelcells/mypp/. Accessed 01 Oct2010.

[6] M. Yu, O. Lourie, M.J. Dyer, K. Moloni, T.F.Kelly and R.S. Ruoff // Science 287(2000)637.

[7] N.P. Cele, S.S. Ray, S.K. Pillai,M. Ndwandwe, S. Nonjola and L. Sikhwivhilu //Fuel Cells 1 (2010) 64.

[8] R. Kannan, B.A. Kakade and V.K. Pillai //Angew Chem. Int. Ed. 47 (2008) 2653.

[9] N. Cele and S.S. Ray // Macromol. Mater.Eng. 294 (2009) 719.

[10] Y. Liu, B. Yi, Z. Shao, D. Xing and H. Zhang// Electrochem. Solid-State Lett. 9(2006):A356.

[11] Y. Liu, B. Yi, Z. Shao, L. Wang, D. Xing andH. Zhang // J. Power Sources 163 (2007)807.

[12] J.M. Thomassin, J. Kollar, G. Caldarella,A. Germain, R. Jerome and C. Detrembleur //J. Membrane Sci 303 (2007) 252.

[12] L. Wang, D.M. Xing, H.M. Zhang, H.M. Yu ,Y.H. Liu and B.L. Yi // J. Power Sources 176(2008) 270.

[14] W. Zhang, A.A. Dehghani-Sanij and R.S.Blackburn // J. Mater. Sci. 42 (2007)3408.

[15] W. Chen, J. Wu and P. Kuo // Chem. Mater.20 (2008) 5756.

[16] D.H. Jung, S.Y. Cho, D.H. Peck, D.R. Shinand J. S. Kim // J. Power Sources 118 (2003)205.

[17] R. O’Hayre, S. Cha, W. Colella and F.B.Prinz, Fuel Cell Fundamentals (John Wiley& Sons, New York, 2006).

[18] Q. Li, R. He, J.O. Jensen and N.J. Bjerrum //Chem. Mater. 15 (2003) 4896.

[19] R. He, Q. Li, G. Xiao and N.J. Bjerrum //J. Membrane Sci. 226 (2003) 169.

[20] X. Xiao, H. Zhang, E. Scanon , L.S.Ramanathan, E.W. Choe, D. Rogers,T. Apple and B. C. Benicewicz // Chem.Mater. 17 (2005) 5328.

[21] K.D. Kreuer, S.J. Paddison, E. Spohr andE.M. Schuster // Chem. Rev. 104 (2004)4637.

[22] R. Kannan, P.P. Aher, T. Palaniselvam,S. Kurungot, U.K. Kharul and V.K. Pillai//J. Phys. Chem. Lett. 1 (2010) 2109.

[23] W. Zhou, J. Xiao, Y. Chen, R. Zeng, S. Xiao,H. Nie, F. Li and C. Song // Polym. Adv.Technol. (2010) Doi: 10.1002/pat.1666.

[24] Y. Tominaga, I. Hong, S. Asai and M. Sumita// J. Power Sources 171 (2007) 530.

[25] H. Lin, T. L. Yu and F. Han // J. Polym. Res.13 (2006) 379.

[26] Platinum Today, http://www.platinum.matthey.com/pgm-prices/price-charts. Accessed1 Oct 2010.

[27] H. Tsuchiya and O. Kobayashi // Int.J. Hydrogen Energy 29 (2004) 985.

[28] W. Li, C. Liang, W. Zhou, J. Qiu, Z. Zhou,G. Sun and Q. Xin // J. Phys. Chem. B 107(2003) 6292.

[29] P. Serp, M. Corrias and P. Kalck// Appl.Catal. A: Gen 253 (2003) 337.

[30] Z.X. Liang and T.S. Zhao //J. Phys. Chem.C 111 (2007) 8128.


[31] K. Lee, J.J. Zhang, H.J. Wang and D.P.Wilkinson // J. Appl. Electrochem. 36 (2006)507.

[32] X. Sun, R. Li, D. Villers, J. P. Dodelet andS. Desilets // Chem. Phys. Lett. 379 (2003)99.

[33] X. Wang, W. Li, Z. Chen, M. Waje and Y. Yan// J. Power Sources 158 (2006) 154.

[34] G. Girishkumar, M. Rettker, R. Underhile,D. Binz, K. Vinodgopal, P. McGinn andP. Kamat // Langmuir 21 (2005) 8487.

[35] A. Kongkan, S. Kuwabata, G. Girishkumarand P. Kamat// Langmuir 22 (2006) 2392.

[36] W. Li, X. Wang, Z. Chen, M. Waje and Y. Yan// Langmuir 21 (2005) 9386.

[37] J.M. Tang, K. Jensen, M. Waje, W. Li,P. Larsen, K. Pauley, Z. Chen, P. Ramesh,M.E. Itkis, Y.Yan and R.C. Haddon //J. Phys. Chem. C 111 (2007) 17901.

[38] N. Rajalakshmi, H. Ryu, M.M. Shaijumonand S. Ramaprabhu // J. Power Sources 140(2005) 250.

[39] B. Yue, Y. Ma, H. Tao, L. Yu, G. Jian,X. Wang, X. Wang, Y. Lu and Z. Hu//J. Mater. Chem. 18 (2008) 1747.

[40] A. Kongkanand, K. Vinodgopal, S. Kuwabataand P.V. Kamat // J. Phys. Chem. B 110(2006) 16185.

[41] M.M. Waje, X. Wang, W. Li and Y. Yan //Nanotechnology 16 (2005) S395.

[42] M. Kaempgen, M. Lebert, S. Roth, M. Soehnand N. Nicoloso// J. Power Sources 180(2008) 755.

[43] M. Kaempgen, M. Lebert, N. Nicoloso andS. Roth // Appl. Phys. Lett. 92 (2008)094103.

[44] T. Maiyalagan, B. Viswanathan and U.V.Varadaraju // Electrochem. Comm. 7 (2005)905.

[45] A.L.M. Reddy and S. Ramaprabhu // J. Phys.Chem. C 111 (2007) 16138.

[46] C.Y. Du, T.S. Zhao and Z.X. Liang // J. PowerSources 176 (2008) 9.

[47] H.F. Oetjen, V.M. Schmidt, U. Stimming andF. Trila // J. Electrochem. Soc. 143 (1996)3838.

[48] A. Gasteiger, N. Markovic and P.N. Ross //J. Phys. Chem. 99 (1995) 16757.

[49] M.Watanabe, M. Uchida and S. Motoo //J. Electroanal. Chem. 99 (1986) 311.

[50] M. Carmo, V.A. Paganin, J.M. Rosolen andE.R. Gonzalez // J. Power Sources 42 (2005)169.

[51] N. Jha, A.L.M. Reddy, M.M. Shaijumon,N. Rajalakshmi and S. Ramaprabhu // Int.J. of Hydrogen Energ 33 (2008) 427.

[52] J. Prabhuram, T.S. Zhao, Z.K. Tang, R. Chenand Z.X. Liang // J. Phys. Chem. B 110(2006) 5245.

[53] W. Li, X. Wang, Z. Chen, M. Waje andY. Yan // J Phys Chem B 110 (2006): 15353.

[54] K. Jeng, C. Chien, N. Hsu, S. Yen, S. Chiou,S. Lin and W. Huang // J Power Sources 160(2006) 97.

[55] T. Matsumoto, Y. Nagashima, T. Yamazakiand J. Nakamura // Electrochem Solid-StateLett 9 (2006) A160.

[56] E. Yoo, Y. Nagashima, T. Yamazaki,T. Matsumoto and J. Nakamura // Polym AdvTechnol 17 (2006) 540.

[57] M. Lefèvre and J.P. Dodelet //J Phys ChemB 106 (2002) 8705.

[58] J. Yang, D. Liu, N.N. Kariuki and L.X. Chen //Chem Commun (2008) 329.

[59] K. Prehn, A. Warburg, T. Schilling, M. Bronand K. Schulte // Compos Sci Technol 69(2009) 1570.

[60] A. T]tov, P. Zapol, P. Kral, D. L]u, H. Idd]r,K. Baishya and L. A. Curtiss // J Phys ChemC 113 (2009) 21629.

[61] P.H. Matter, E. Wang, M. Arias, E.J.Biddinger and U. S. Ozkan// J Phys ChemB 110 (2006) 18374.

[62] C.V. Rao, C.R. Cabrera and Y. Ishikawa //J Phys Chem Lett 1 (2010 ) 2622.

[63] S. Kundu, T.C. Nagaiah, W. Xia, Y. Wang,S.V. Dommele, J.H. Bitter, M. Santa,G. Grundmeier, M. Bron, W. Schuhmann andM. Muhler // J Phys Chem C 113 (2009)14302.

[64] S. Maldonado, S. Morin and K.J. Stevenson// Carbon 44 (2006) 1429.

[65] J.D. Wiggins-Camacho and K.J. Stevenson //J Phys Chem C 113 (2009) 19082.

[66] H.A. Gasteiger, S.S. Kocha, B. Sompalli andF.T. Wagner // Appl Catal B: Environ 56(2005) 9.

[67] R. Coontz and B. Hanson // Science 305(2004) 957.

[68] J. McCarthy, Hydrogen as a Fuel (StanfordUniversity, 2008), http://www-formal.stanford.edu/jmc/progress/hydrogen.html. Accessed 1 Oct 2010.

[69] L. Mahdavian // International Journal of GreenNanotechnology: Materials Science &Engineering 1(2009) M61.


[70] Y.L. Chen, B. Liu, J. Wu , Y. Huang, H. Jiangand K.C. Hwang // J Mech Phys Solids 56(2008) 3224.

[71] T. Yildirim and S. Ciraci // Phys Rev Lett 94(2005) 175501.

[72] I. Cabria and M.J. López // J Chem Phys 123(2005) 204721.

[73] D.E. Tylianakis and G.E. Froudakis // NanoLett 8 (2008) 3166.

[74] A. Nikitin, H. Ogasawara, D. Mann,R. Denecke, Z. Zhang, H. Dai, K. Cho andA. Nilsson // Phys Rev Lett 95 (2005)225507.

application of carbon nanotubes in polymer · pdf filepolymer electrolyte based fuel cells...

Documents