Influence of chemical composition and sequence length on the transport properties of proton exchange membranes

Influence of Chemical Composition and SequenceLength on the Transport Properties of ProtonExchange Membranes

ABHISHEK ROY,1 MICHAEL A. HICKNER,2 XIANG YU,1 YANXIANG LI,1 THOMAS E. GLASS,1 JAMES E. MCGRATH1

1Macromolecules and Interfaces Institute; Virginia Polytechnic Institute and State University, Blacksburg, Virginia 24061

2Chemical and Biological Systems Department, Sandia National Laboratories, Albuquerque, New Mexico 87185

Received 5 January 2006; revised 22 March 2006; accepted 5 April 2006DOI: 10.1002/polb.20859Published online in Wiley InterScience (www.interscience.wiley.com).

ABSTRACT: One of the integral parts of the fuel cell is the proton exchange mem-brane. Our research group has been engaged in the past few years in the synthesis ofseveral sulfonated poly(arylene ether) random copolymers. The copolymers were var-ied in both the bisphenol structure as well as in the functional groups in the back-bone such as sulfone and ketones. To compare the effect of sequence length, multi-block copolymers based on poly(arylene ether sulfone)s were synthesized. This paperaims to describe our investigation of the effect of chemical composition, morphology,and ion exchange capacity (IEC) on the transport properties of proton conductingmembranes. The key properties examined were proton conductivity, methanol perme-ability, and water self diffusion coefficient in the membranes. It was observed thatunder fully hydrated conditions, proton conductivity for both random and blockcopolymers was a function of IEC and water uptake. However, under partiallyhydrated conditions, the block copolymers showed improved proton conductivity overthe random copolymers. The proton conductivity for the block copolymer series wasfound to increase with increasing block lengths under partially hydrated conditions.VVC 2006 Wiley Periodicals, Inc. J Polym Sci Part B: Polym Phys 44: 2226–2239, 2006

Keywords: proton exchange membrane; sulfonated polymer; block copolymer;direct methanol fuel cell; nuclear magnetic resonance

INTRODUCTION

Fuel cells are electrochemical devices that con-vert chemical energy directly into electricalenergy.1 Proton exchange membrane fuel cells(PEMFC) have shown promise as alternativeportable, automotive, and stationary power sour-ces.1,2 One of the important components in aPEMFC is the proton exchange membrane(PEM), which serves as the electrolyte that trans-

fers protons from the anode to the cathode andseparates the fuel and oxidizer. The current stateof the art PEMs are perfluorosulfonic acid mem-branes such as Nafion1 manufactured by DuPont.These membranes show excellent chemical andelectrochemical stability as well as high protonconductivity with relatively low water uptake on amass basis. However, Nafion suffers from disad-vantages, including high cost, high methanol per-meability (DMFCs), and limited operating temper-ature (80 8C)3,4, due to its depressed hydrated arelaxation temperature.5

Our research group has been engaged in thepast few years in the synthesis and character-ization of biphenol-based partially disulfonated

Correspondence to: J. E. McGrath (E-mail: [email protected])

Journal of Polymer Science: Part B: Polymer Physics, Vol. 44, 2226–2239 (2006)VVC 2006 Wiley Periodicals, Inc.

2226

poly(arylene ether sulfone) random copolymers aspotential PEMs.6–11 This series of polymers arecalled BPSH-xx [Fig. 1(a)], where BP stands forbiphenol, S is for sulfonated, and H denotes theproton form of the acid where xx represents thedegree of disulfonation. Along with the BPSH se-ries, hydroquinone (HQ)-based partially disulfo-nated poly(arylene ether sulfone)12 [HQSH-xx,Fig. 1(b)] random copolymers and partially disul-fonated poly(arylene ether ketone) randomcopolymers13 [B-ketone-xx and PB-diketone-xxseries, Fig. 1(c)] were synthesized. The B-ketone-xx series has a mono ketone functional group perrepeat unit and the PB-diketone-xx series has adiketone functional group per repeat unit. Inaddition to the random copolymers, a series ofmultiblock copolymers [Block BisAF-BPSH(x:y)Kseries, Fig. 1(d)] was also synthesized.14 In themultiblock copolymer series, BisAF (fluorinatedbisphenolA) is the hydrophobic unit and BPSHwith 100% degree of disulfonation is the hydro-philic unit. Here x and y represent the blocklength in g/mol of hydrophobic:hydrophilic unitsrespectively.

The copolymers were varied in both chemicalcomposition and ion concentration to obtain desir-able properties. The transport of water and pro-tons in these systems is both of fundamental in-terest in terms of structure/property relation-

ships for PEMs and technological importance inthe design of new materials for fuel cells. This pa-per aims to describe our investigation of the effectof chemical composition, sequence length, mor-phology, and ion concentration on the transportproperties for proton conducting membranes. Thekey properties examined were proton conductiv-ity, methanol permeability, and the self diffusioncoefficient of water in the membrane.

Proton Conductivity

Proton transport in aqueous environments hasbeen studied extensively and can be described bytwo principal mechanisms.15–17 The first is theGrotthuss18 mechanism where the protons aretransferred down a chain of correlated hydrogenbonds forming and reforming hydronium cations(H3O

þ). The second mechanism as described byKreuer19 is referred to as the vehicle mechanism.It is believed to occur by bulk diffusion of a protonand its associated water molecules or ‘‘vehicle’’ inan H3O

þ, H5O2þ, or other H2nþ1O

nþ cation spe-cies. Hence, the self diffusion coefficient of wateris of great importance in understanding thetransport behavior of protons as shown by bothKreuer et al.16 and Zawodzinski et al.20 Boththese reports have provided evidence for the vehi-cle mechanism dominating at low water contents

Figure 1. (a–d) Copolymer chemical structures synthesized in this work (a) BPSH-xx, (b) HQSH-xx, (c) B-ketone-xx and PB-diketone-xx, (d) Block BisAF-BPSH(x:y)K.

PROPERTIES OF PROTON EXCHANGE MEMBRANES 2227

Journal of Polymer Science: Part B: Polymer PhysicsDOI 10.1002/polb

(or hydration number k) where the diffusion coef-ficients of protons and water are similar: waterand protons move in concert. At high water con-tents, the Grotthuss mechanism dominates pro-ton conduction where the computed diffusioncoefficient of protons is higher than that of themeasured water self-diffusion coefficients: pro-tons are moving faster across the membrane thanthe water.

To understand the effect of chemical structureand chemical compositions of the copolymer onproton transport, we have chosen to study protontransport over a range of hydration levels. It hasbeen reported in the literature that for Nafion,both the water self diffusion coefficients and pro-ton conductivity increase with increase in hydra-tion levels.20 Similar observations have beenfound for the BPSH type random copolymer mem-branes.21 As reported earlier, sulfonic acid-con-taining ionomers tend to phase separate into anano-phase separated hydrophilic and hydropho-bic domain morphology with an increase in hydra-tion levels.7,10,22 In other words, the hydratedhydrophilic domains interconnect to form an asso-ciated, percolated type morphology. However, atlow hydration levels most of the water is tightlyassociated with sulfonic groups and has a low dif-fusion coefficient. This tends to form an isolateddomain morphology. Thus, although there may besignificant concentrations of protons, the trans-port is limited by the discontinuous morphologicalstructure. With an increase in hydration levels,water assisted percolated structure ensuresproper connectivity between the ionic domains.This is marked by the increase in diffusion coeffi-cient of water as transport of water is facilitatedthough the interconnected channels.

The challenge lies in how to modify the chemis-try of the polymers to obtain significant protonconductivity at low hydration levels where theself-diffusion coefficient of the water may still bequite low due to its association to the acid groups.High conductivity at low degrees of disulfonationand low hydration levels may be possible if onecan alter the sequential chemical structure of thepolymer backbone to produce block copolymers.

Block copolymers23 consist of two or morechemically dissimilar backbone segments (i.e.blocks) that are chemically conjoined throughcovalent bonds in the same chain. In many caseswhere the blocks are immiscible, this leads toimproved morphological control by tailoring thechemical composition, molecular weight, and vol-ume fraction of the blocks, and thus the creation

of new materials with properties that were notpossible in a single phase. In a block ionomerPEMs, the ionic groups act as proton conductingsites, while the nonionic component providesdimensional strength and, in the case of DMFCs,may serve as a barrier for methanol transport.Relatively few attempts have been made in thedevelopment of block copolymer ionomers havingthermally and chemically stable aromatic back-bone.14,24–26 Synthesis of such systems is madepossible by step growth polymerization, that ishydrophobic and hydrophilic telechelic oligomerswith appropriate end groups may be copolymer-ized to obtain multiblock, or segmented, copoly-mer ionomers. We have recently synthesized a se-ries of multiblock ionomers having highly fluori-nated hydrophobic blocks and poly(arylene ethersulfone) hydrophilic blocks with 100% degree ofsulfonation [Fig. 1(d)].14,26 Unlike the BPSHcopolymers, where the sulfonic acid groups arerandomly distributed in pairs along the chain,the multiblock copolymers will feature an orderedsequence of hydrophilic and hydrophobic seg-ments. If connectivity is established between thehydrophilic domains in these multiblock copoly-mers, they will not need as much water to form apercolated structure, and hence may show muchbetter proton conductivity than the random co-polymers (with similar ionic compositions) underpartially hydrated conditions.

Methanol Permeability

Methanol permeability is a critical transportproperty when considering a new PEM for use inliquid-fed direct methanol fuel cells. Methanolthat is unoxidized at the anode can ‘‘crossover’’though the membrane and be oxidized at thecathode. This methanol short circuit decreasesthe fuel efficiency of the system, lowers the cellvoltage by causing a mixed potential at the cath-ode, and increases cell heating. Many novelPEMs have low methanol permeability, but theirproton conductivity tends to decrease as well.Membranes with low crossover and high conduc-tivity are sought while still maintaining the abil-ity to form robust junctions at the membrane–electrode interface.

Two common methods for measuring the meth-anol permeability of PEMs are a membrane sepa-rated cell27 and an open circuit electrochemicalcrossover measurements in a DMFC.28 The twomethods have been shown to be approximatelyequivalent.29

2228 ROY ET AL.


Self Diffusion Coefficient of Water

One of the most important techniques for mea-suring self diffusion coefficient of water is bypulsed-field gradient spin echo nuclear magneticresonance (PGSE NMR).30–34 One of the mainadvantages of this method is that the measureddiffusion coefficients can be related to the meansquare displacement of the water molecule, or theeffective diffusion length scale30,31,35–37 becausethe time over which the diffusion measurement ismade is known. The displacement is given by eq 1

h�ri ¼ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi2Deff�

pð1Þ

for a gradient in one direction. This will provideinsight for probing the influence of the restrictedmorphology on the transport properties.

EXPERIMENTAL

Materials

Highly purified 4,40-dichlorodiphenylsulfone (DCDPS)was provided by Solvay Advanced Polymers. BP andHQ were obtained from Eastman Chemical. 4,40-Hexafluoroisopropylidenediphenol (6F-BPA), re-ceived from Ciba, was purified by sublimation.The ketone monomers 4,40-difluorobenzophenone(DFBP) and 1,4-bis(p-fluorobenzoyl)benzene (PBFB)were purchased from Aldrich and used as received.All these monomers were dried under vacuumprior to use. Disulfonated comonomers (SDCDPS,SDFBP, SPBFB) were synthesized according tothe modified literature methods8 and dried undervacuum before copolymerization. The solventN,N-dimethylacetamide (DMAc, Fisher) and N-methyl-2-pyrrolidinone (NMP, Fisher) was vac-uum-distilled from calcium hydride onto molecu-lar sieves. Potassium carbonate (Aldrich) wasdried in vacuum before copolymerization. Tolu-ene, sodium chloride, 30% fuming sulfuric acid,and methanol were obtained from Aldrich andused as received. The fuming sulfuric acid wasfurther analyzed for active SO3 concentration.38

Nafion1 112 and Nafion1 117 were obtained fromElectroChem.

Synthesis of Disulfonated Poly(arylene Ether)Random Copolymers (BPSH-xx, HQSH-xx,B-Ketone-xx, and PB-Diketone-xx)

Two series of disulfonated poly(arylene ether sul-fone) random copolymers were synthesized via

aromatic nucleophilic substitution as reportedearlier.8,12 One typical copolymerization for aBPSH-32 copolymer was as follows. 4,40-Biphenol(1.8621 g, 10 mmole), 4,40-dichlorodiphenyl sul-fone (1.9528 g, 6.8 mmole), and 3,30-disulfonated4,40-dichlorodiphenyl sulfone (1.5721 g, 3.2 mmole)were copolymerized in dry DMAc solvent undernitrogen. Potassium carbonate (1.15 equiv) wasadded and toluene (DMAc/toluene ¼ 2/1 v/v) wasused as an azeotropic agent to dehydrate the sys-tem. The HQSH series copolymers were synthe-sized with HQ instead of biphenol. The two seriesof disulfonated poly(arylene ether sulfone) copoly-mers were designiated as BPSH-xx and HQSH-xx(acid form) [Fig. 1(a,b)].

The copolymerization procedures for polyke-tones were similar to those of polysulfones,although the monomers’ reactivities are different.The detailed procedure was also introduced inthe literature.13 The PB-diketone-xx series [Fig.1(c)] copolymers were based on monomers 6F-BPA, PBFB, and SPBFB, while the B-ketone-xxseries [Fig. 1(c)] copolymers were based on 6F-BPA, DFBP, and SDFBP. Because the fluorinatedmonomers have higher reactivity, high molecularweight copolymers can be achieved within shorterreaction times.

Synthesis of the Hydrophilic/HydrophobicMultiblock Copolymers

The syntheses of hydrophobic and hydrophilicoligomers and multiblock copolymers were car-ried out as reported earlier.14 A typical polymeri-zation procedure was as follows: BPA (1.174 g,5.142 mmol) was added to a three-necked round-bottomed flask equipped with a mechanical stir-rer, condenser, nitrogen inlet, and a Dean-Starktrap. NMP (10 mL) was added and the mixturewas stirred until dissolved. K2CO3 (1.183 g,7.20 mmol) and toluene (5 mL) were added andthe system was dehydrated at 150 8C. Then thereaction bath was cooled to 50 8C and decafluoro-biphenyl (2.046 g, 6.124 mmol) was added. Poly-merization was allowed to proceed at 110 8C for5 h. The reaction mixture was isolated by precipi-tation into a H2O/methanol (50/50 v/v) mixture.Biphenol (0.412 g, 2.213 mmol), SDCDPS (0.912 g,1.856 mmol), and NMP (10 mL) were charged toanother three-necked round-bottomed flask.The mixture was stirred until dissolved, thenK2CO3 (0.430 g, 3.12 mmol) and toluene (5 mL)were added. After dehydration at 150 8C, thepolymerization proceeded at 190 8C for 16 h.



Then, the reaction bath was cooled to 80 8C, andthe perfluorinated hydrophobic telechelicoligomer (1.050 g, 0.350 mmol) was dissolved inNMP and added to the same reaction flask. Thebath temperature was raised to 95 8C and kept atthis temperature for 9 h. The reaction mixturewas precipitated into isopropanol to obtain abrownish fibrous polymer. A series of blockcopolymers BisAF-BPSH(x:y)K (Fig. 1) were syn-thesized, differing in ion exchange capacity (IEC)and block length.

Membrane Preparation

The salt form copolymers were redissolved inDMAc to afford transparent 5 wt % solutions,which were then cast onto clean glass substrates.The films were slowly dried for 2 days with infra-red heat at gradually increasing temperatures,and then dried under vacuum at 110 8C for 2days. The membranes were converted to theiracid form by boiling in 0.5M H2SO4 for 2 h, andwere then boiled in deionized water for 2 h toremove any residual acid. Membranes werestored in deionized water until it is used for meas-urements.

Proton Conductivity

Proton conductivity at 30 8C at full hydration (inliquid water) was determined in a window cell ge-ometry20 using a Solartron 1252 þ 1287 Imped-ance/Gain-Phase Analyzer over the frequencyrange of 10 Hz to 1 MHz following the procedurereported in the literature.39 In determining pro-ton conductivity in liquid water, membranes wereequilibrated at 30 8C in DI water for 24 h prior tothe testing. The temperature range chosen forcalculation of activation energy for proton trans-port was from 30 to 80 8C. For determining protonconductivity under partially hydrated conditions,membranes were equilibrated in a humidity-tem-perature oven (ESPEC, SH-240) at the specifiedRH and 80 8C for 24 h before each measurements.

Water Uptake

The water uptake of the membranes was deter-mined gravimetrically.10 The water uptake of themembranes was calculated according to eq 2,where massdry and masswet refer to the mass ofthe wet membrane and the mass of the dry mem-brane, respectively.

water uptake % ¼ masswet �massdrymassdry

� 100 ð2Þ

The hydration number (k), number of water mole-cules absorbed per sulfonic acid, can be calculatedfrom the mass water uptake and the ion contentof the dry copolymer as shown in eq 3

k ¼ masswet �massdry=MWH2O

IEC�massdryð3Þ

where MWH2O is the molecular weight of water(18.01 g/mol) and IEC is the ion exchangecapacity of the dry copolymer in equivalents pergram.


Methanol permeability of the membranes wasdetermined by measuring the crossover currentin a DMFC at open circuit. The measurementwas performed in an identical manner to Renet al.28

Pulsed-Field Gradient Spin Echo NuclearMagnetic Resonance

Water self diffusion coefficients were measuredusing a Varian Inova 400 MHz (for protons) nu-clear magnetic resonance spectrometer with a30 G/cm gradient diffusion probe. A total of 16points were collected across the range of gradientstrength and the signal to noise ratio enhancedby coadding 4 scans. The standard stimulated-echo NMR pulse sequence is shown in Figure 2.

Figure 2. Pulse sequence schematic for PGSE NMR experiments.

2230 ROY ET AL.


The measurement was conducted by observingthe echo signal intensity (A) as a function of thegradient strength. The diffusion coefficient (D)was determined by fitting the data to the eq 431,36

AðgÞ ¼ AðoÞ exp½�c2Dg2d2ð�� d=3Þ� ð4Þ

where A is the NMR signal intensity (A) as afunction of gradient strength, c is the gyromag-netic ratio (26,752 rad G�1 s�1 for protons), d islength of the gradient pulse, D is the timebetween gradient pulse.

Membrane samples of approximately 5 mm� 15 mm � 150 lm were equilibrated in liquidwater for at least 24 h. The samples wereremoved from the liquid water, blotted to removedroplets, quickly inserted into the NMR tube, andimmediately measured over a span of about5 min. Measurements were repeated by reim-mersing the sample in DI water, waiting at least30 min, and then repeating the transfer and mea-surement process. Separate measurements werecollected with different times between the gradi-ent pulses.

NMR Spectroscopy, Gel PermeationChromatography, Intrinsic Viscosity andAtomic Force Microscopy Characterization

1H and 19F NMR analysis were conducted on aVarian Unity 400 spectrometer. All spectra wereobtained from a 10% solution (w/v) in a DMSO.d6

solution at room temperature. Gel permeationchromatography (GPC) experiments were per-formed on a liquid chromatograph equipped witha Waters 1515 isocratic HPLC pump, WatersAutosampler, Waters 2414 refractive index detec-tor, and Viscotek 270 RALLS/ viscometric dualdetector. NMP (containing 0.05M LiBr) was usedas the mobile phase. The column temperaturewas maintained at 60 8C because of the viscousnature of NMP. Both the mobile phase solventand sample solution were filtered before introduc-tion to the GPC system. Molecular weights weredetermined from universal calibration plot usingpolystyrene as standard. Intrinsic viscositieswere determined in 0.05M LiBr NMP at 25 8Cusing a Cannon Ubbelholde viscometer. Atomicforce microscopy characterization (AFM) imageswere taken using Digital Instruments Dimension3000 with a microfabricated cantilever. The forceconstant was 40 N/m.

RESULTS AND DISCUSSIONS

NMR Spectroscopy, GPC, Intrinsic Viscosity, andAFM Characterization of the Multiblock Copolymer

Figure 3(a) shows the 1H NMR spectra of themultiblock copolymer BisAF-BPSH(8:8)K and thecorresponding sulfonated hydrophilic telechelicoligomer.14 Figure 3(b) shows the 19F NMR spec-trum of BisAF-BPSH(8:8)K and that of the corre-sponding fluorinated hydrophobic telechelicoligomer. The number-average molecular weightsof both telechelic oligomers can be calculatedusing end-group analysis in the NMR spectra andwere found to be in good agreement with the tar-get values. The end-group peaks that are presentin the spectra of telechelic oligomers are notobserved in those of the multiblock copolymer.This confirms the high conversion in the blockcopolymerization. The molecular weights of themultiblock copolymers were confirmed by GPCand intrinsic viscosity data as given in Table 1.

Figure 4 shows the tapping mode AFM imagesof BisAF-BPSH(5:5)K multiblock copolymer, withthe dark regions representing the hydrophilicdomains. Although no orientated morphology canbe observed, there exist sharp nanophase sepa-rated hydrophilic clusters, which may be inter-connected by narrow hydrophilic channels.

Proton Conductivity and Water Uptake

It has been widely reported in the literature thatproton conductivity for sulfonated polymersdepends both on the water uptake and on the IECof the material.7 In addition to IEC, and wateruptake on a mass basis, the hydration number (k)is widely used to compare membranes of differentpolymer backbone architectures. Previous studiesperformed in our laboratory indicated that theproton conductivity scales linearly with the IECfor the BPSH copolymers at fully hydrated condi-tions.6,10 Similar trends were observed for HQSH,PB, and B series as given in Figure 5 and Table 2.

The increase in proton conductivity can beexplained on the basis of an increase in the con-centration of protons from the acid sites and cor-responding increase in water uptake. In additionto proton conductivity, it has also been reportedthat water uptake for these sulfonic acid-contain-ing copolymers increases with an increase inIEC.5 It can be argued that at higher wateruptake, water facilitates proton transport eitherby an increase in Grotthuss hopping (dynamic,



Figure 3. (a) 1H NMR spectra of hydrophilic telechelic oligomer (top) and multi-block (bottom) BisAF-BPSH(8:8)K copolymer. (b) 19F NMR spectra of hydrophobictelechelic oligomer (top) and multiblock BisAF-BPSH(8:8)K (bottom) copolymer.

2232 ROY ET AL.


long-range hydrogen-bonded water molecule chains)or by promoting the formation of percolated struc-ture. This is supported by the decrease in activa-tion energy for proton transport with an increasein IEC and k as given in Table 2. The activationenergies were calculated from the temperaturedependence of proton conductivity using a simpleArrhenius analysis. The temperature dependenceplots of proton conductivity for the HQSH-xx andPB-diketone-xx series are shown in Figures 6 and7, respectively.

To study the effect of the chemical structureson proton conductivity, the proton conductivityfor different polymers with the same IEC wascompared. The families of copolymers in this

study have similar sulfonic acid-containing moi-eties and hence the acidity of these copolymers isassumed to be similar.

Figure 8 represents the plots of proton conduc-tivity versus hydration number for the BPSH,HQSH, PB-diketone, and B-ketone copolymers atan IEC of 1.4 mequiv/g. It is apparent from thefigure that the proton conductivity at fullyhydrated conditions for the copolymers with simi-lar acidity is a linear function of hydration num-ber. With increase in water uptake, proton trans-port is facilitated, irrespective of the IEC value.Thus the chemical structures of these copolymersplay a role in controlling the hydration levels,which in turn dominates the proton conductivity.

Table 1. Block Lengths, IV, Mn, IEC, Water Uptake, and Proton Conductivity of the MultiblockBisAF-BPSH(x:y)K Copolymers

PolymerBlock Lengthsa

(g/mol) IECb IVcMn (g/mole)from GPC

Water Uptake(mass %)

Protond

Conductivity(mS/cm)

BisAF-BPSH(3:3)K 3k:3k 1.6 0.60 30K 71 130BisAF-BPSH(5:5)K 5k:5k 1.4 0.54 30K 58 104BisAF-BPSH(8:8)K 8k:8k 1.1 0.53 37K 42 90

a Block lengths are expressed in the form hydrophobic:hydrophilic.b Measured from 1H NMR.c Measured at 25 8C in 0.05M LiBr NMP.d Measured in liquid water at 30 8C.

Figure 4. Tapping mode AFM height image (left), phase image (right) of blockBisAF-BPSH(5:5)K indicating nano-phase separation between hydrophilic and hydro-phobic domains.



Now let us consider the case of the multiblockcopolymers. Table 1 shows the proton conductiv-ity and water uptake for the multiblock copoly-mers at different IECs and block lengths. It canbe concluded that at fully hydrated conditions,the chemical structure or the block length has lit-tle influence on the proton transport. Similar tothe random copolymers, proton conductivity in-creases with increase in IEC and water uptake ina systematic fashion.

Until now, only the proton conductivity underfully hydrated conditions has been discussed andthe concentration of the ionic groups and hydra-tion number were found to be the primary effect

in this regime. At partially hydrated conditions,proton conductivity for BPSH11 and other randomcopolymers was found to be strongly dependenton the hydration levels.

Figure 9 shows the proton conductivity as afunction of relative humidity, respectively, forHQSH 30, PB-diketone 50, and Nafion 117. Below50% RH, the proton conductivity decreases signif-icantly for the random sulfonated aromaticcopolymers. The fact is supported by the decreasein activation energy for proton transport with anincrease in RH as shown in Figure 10 for HQSH30. For Nafion, the extent of decrease in conduc-tivity at low RHs is not as severe as that meas-ured for random copolymers. This is related tothe unique chemical structure of the Nafion,which consists of highly flexible side chain hydro-

Table 2. Water Uptake, Proton Conductivity at 30 and 80 8C and Activation Energies for Proton Transport forHQSH-xx, PB-Diketone-xx, and B-Ketone-xx Random Copolymers

SampleIEC

(mequiv/g)Water

Uptake (k)

ProtonicConductivity

(mS/cm) at 30 8C

ProtonicConductivity

(mS/cm) at 80 8C

ActivationEnergy

Ea(kJ/mole)

HQSH 30 1.6 29 100 170 8.6HQSH 25 1.4 24 80 140 9.0HQSH 20 1.2 14 70 122 10.6PB-Diketone 50 1.4 26 71 141 10.4PB-Diketone 40 1.2 12 40 90 14.8PB-Diketone 30 0.9 9 10 24 15.6B-Ketone 50 1.7 32 90 156 8.5B-Ketone 40 1.4 18 73 130 10.3B-Ketone 30 1.1 13 23 40 12.9

Figure 5. Proton conductivity for HQSH-xx, B-ke-tone-xx, and PB-diketone-xx increases with increasein IEC.

Figure 6. Proton conductivity for HQSH-xx randomcopolymers follows Arrhenius relationship with tem-perature.

2234 ROY ET AL.


philic sulfonic acid groups and a hydrophobicfluorinated flexible backbone. These chemicalfeatures apparently promote strong nano-phaseseparation between distinctly hydrophilic andhydrophobic domains even in random ionomers3

and allow proton transport between the intercon-nected hydrophilic domains even at low hydrationlevels. Hence, by synthesizing a predefined nano-phase segregated ionomeric block copolymerusing an aromatic backbone, one may obtain im-proved proton conductivity under partially hy-drated conditions.

Figure 11 shows the proton conductivity as afunction of relative humidity for the block BisAF-

BPSH(x:y)K series with differing block lengthsand Nafion 117. The proton conductivity of theblock BisAF-BPSH(x:y)K series increases withincreasing block lengths under partially hydratedconditions. This is in contrast to fully hydratedconditions where proton conductivity was a func-tion of IEC and hydration number rather than

Figure 7. Proton conductivity for PB-diketone-xxrandom copolymers as a function of temperature.

Figure 8. Proton conductivity for copolymers scaleslinearly with hydration number at IEC of 1.4 mequiv/g.

Figure 9. Proton conductivity as a function of rela-tive humidity for Nafion 117, HQSH-30, and PB-dike-tone-50 at 80 8C.

Figure 10. Activation energy for proton conductionfor HQSH-30 random copolymer decreases with in-crease in relative humidity over a temperature rangeof 30–80 8C.



block length. The block BisAF-BPSH(8:8)K sam-ple has the highest proton conductivity amongthe block BisAF-BPSH(x:y)K copolymers acrossthe range of relative humidity, but the lowestwater uptake and IEC as shown in Table 1. Withan increase in block length, the extent of nano-phase separation and the connectivity betweenthe hydrophilic domains increases. This de-creases the morphological barrier for protontransport and as a result, the proton conductivityincreased. Under partially hydrated conditions,connectivity between the hydrophilic domainsseems to be the driving force for proton transport.

Self Diffusion Coefficient of Water

The self-diffusion coefficient for water was meas-ured by PGSE NMR. All of the samples weremeasured under fully hydrated conditions at30 8C.

Figure 12 shows the self diffusion coefficient ofwater versus IEC for the random and the multi-block copolymers. For the random copolymers,particularly for the HQSH series, the diffusioncoefficient increases with an increase in IEC andso with water uptake. Surprisingly, for the blockBisAF-BPSH(x:y)K series, the diffusion coefficientincreased with an increase in block length irre-spective of the IEC values. This clearly empha-sizes the importance of connectivity between thehydrophilic domains on the transport properties.

Probing the morphology of the interconnectedchannels between the hydrophilic domains willprovide a better understanding about the trans-

port mechanism. As mentioned earlier, mea-suring the self diffusion coefficient of water byPGSE NMR technique allows relating the self dif-fusion coefficients of water to the mean squaredisplacement of the water molecules. The effec-tive diffusion time (and hence the length overwhich the diffusion coefficient is measured) canbe varied by varying the (D � d/3) term in theNMR pulse sequence. It has been reported byGebel and coauthors40 that the self diffusion coef-ficient of water in ion containing copolymersdecreased with diffusion time and became con-stant after a particular diffusion time. The re-stricted diffusion can be related to a morphologi-cal barrier to the transport of water moleculesover a length scale given by eq 1.

Figure 13 shows the self diffusion coefficientsof water versus the effective diffusion time forNafion112, block BisAF-BPSH(8:8)K, PB-dike-tone 50, and HQSH 25. The value for the diffu-sion coefficient is high at shorter diffusion times.With an increase in diffusion time, the diffusioncoefficient decreases. This is more pronounced forthe case of random copolymers, HQSH 25, andPB-diketone 50. For Nafion and block BisAF-BPSH(8:8)K, the diffusion coefficient became con-stant with an increase in diffusion time after a

Figure 11. Proton conductivity for block BisAF-BPSH(x:y)K under partially hydrated conditionincreases with increase in block length at 80 8C.

Figure 12. Self diffusion coefficient for water as afunction of IEC for the random and block copolymersand Nafion 112.

2236 ROY ET AL.


particular timescale. The water experiences lessof a morphological barrier to the translationalmotion for the case in the block copolymers thanfor random copolymers. Presence of this low mor-phological barrier eases proton transport underpartially hydrated conditions for the block copoly-mers. It is to be noted that the diffusion coeffi-cient values for the block BisAF-BPSH(8:8)K issignificantly higher than that of Nafion and therandom aromatic copolymers.


One of the major disadvantages of using Nafionas a PEM in DMFC type fuel cell is its high meth-anol permeability. In comparison to Nafion,BPSH copolymers have shown much lower meth-anol permeability.10 This is true in the case forHQSH-xx, B-ketone-xx, and PB-diketone-xx se-ries also.

Figure 14 shows methanol permeability as afunction of IEC for Nafion117, HQSH-xx, B-ke-tone-xx, and block BisAF-BPSH(x:y)K seriesmeasured at 80 8C. Similar to conductivity data,the methanol permeability increases with anincrease in IEC. Permeability is a function of boththe diffusion coefficient for methanol transportand its solubility in the membrane. At higherIECs, greater water self-diffusion coefficients inthe membrane may promote higher methanol dif-

fusion and higher methanol permeability. Al-though a high diffusion coefficient for water is de-sirable for high proton conductivity, it is not de-sirable for lowered methanol permeability. Thereis a tradeoff between proton conductivity andmethanol permeability when designing a newPEM for DFMC applications. The random copoly-mers with the aromatic backbone and low self dif-fusion coefficient of water particularly at lowerIECs offer resistance to the diffusion of the meth-anol. This leads to lower methanol permeabilitythan Nafion, thus making aromatic hydrocarbonmembranes one of the most promising candidatesfor DMFC applications.9 As discussed earlier, theblock copolymers are found to have higher selfdiffusion coefficient of water at higher blocklengths. However, this is not reflected in the val-ues for methanol permeability as given in Figure14. Like random copolymers, methanol perme-ability of the block copolymers tends to dependmore on the IEC and water uptake. Hence, themechanisms of water and methanol transportmay be different in the case for block copolymers.Hydrophobic segments present in the blockcopolymers offer resistance to the diffusion ofmethanol and thus decrease the methanol perme-ability. In ionomeric block copolymers, the hardsegments may lower the methanol permeability,while the soft segments provide pathways forwater and proton transport.

Figure 13. Self diffusion coefficient of water in thepolymer samples as a function of effective diffusiontime indicating presence of a lower morphological bar-rier to transport for Nafion112 and block BisAF-BPSH(8:8)K.

Figure 14. Methanol Permeability for HQSH-xx, B-ketone-xx, block BisAF-BPSH(x:y)K series, andNafion117 as a function of IEC at 80 8C.



CONCLUSIONS

The main objective of the paper was to investi-gate the influence of the chemical structures andcompositions of novel PEMs on transport proper-ties for fuel cell applications. Disulfonated poly(arylene ether sulfone) and poly(arylene ether ke-tone)-based statistical or random copolymerswere synthesized and the copolymers were variedin their ionic content. To compare the effect ofchemical structures and sequence lengths, ioncontaining multiblock poly(arylene ether sulfone)copolymers were also synthesized. Proton conduc-tivity, methanol permeability, and the self diffu-sion coefficient of water were determined underdifferent experimental conditions. Proton conduc-tivity when measured under fully hydrated condi-tions showed that for both the random and blockcopolymers, the value increased with increase inIEC and water uptake. This is consistent with thetheory of vehicle mechanism for proton transport,where water acts as a vehicle for proton trans-port. The influence of hydration levels in theseion-containing copolymers on proton transport ismore evident under partially hydrated condi-tions. For the random copolymers, the proton con-ductivity was found to decrease drastically with adecrease in hydration level. In the case of multi-block copolymers and for Nafion, the extent ofdecrease in proton conductivity with hydrationlevels is not so severe. It was also observed thatthe proton conductivity under partially hydratedconditions increased with sequence or blocklengths.

Synthesizing ion containing nano-phase sepa-rated block copolymers, one can develop intercon-nected hydrophilic domains. This reduces themorphological barrier for proton transport andhence proton transport through these intercon-nected channels can be possible even in the ab-sence of water. The extent of interconnectivity isexpected to increase with an increase in blocklengths as evident from the self diffusion coeffi-cient values of water. Measurements of the selfdiffusion coefficients of water showed that for theblock copolymers, the value increases with anincrease in block lengths irrespective of decreasein IEC and water uptake values. However, forrandom copolymers, the value increases with IECand water uptake. Thus, there is a significant dif-ference in proton and water transport betweenblock and random copolymers. Diffusion coeffi-cient values were also measured as a function ofdiffusion time. For the random copolymers, the

value was found to decrease sharply with time,suggesting a strong morphological barrier totransport. Such a dependency was not observedfor Nafion and the blocks. The diffusion coeffi-cient reached a constant value after a particulardiffusion time.

Methanol permeability was found to increasewith an increase in IEC, water uptake, and selfdiffusion coefficient of water for the randomcopolymers. For multiblock copolymers, the valuewas not found to depend upon self diffusion coeffi-cient of water. The hydrophobic segments presentin the block copolymers offer resistance to the dif-fusion of methanol and thus decreases the metha-nol permeability. Thus, in ionomeric block copoly-mers, the ‘‘hard’’ hydrophobic segments lower themethanol permeability while the ‘‘soft’’ hydro-philic segments provide pathways for water andproton transport. Further papers documentingthese concepts will be forthcoming.

The authors thank the National Science Foundation‘‘Partnership for Innovation’’ Program (HER-0090556)and the Department of Energy (DE-FC36-01G01086)for support of this research effort. The authors wouldalso like to thank UTC Fuel Cell (No. PO3561) fortheir support.

REFERENCES AND NOTES

1. Winter, M.; Brodd, R. J. Chem Rev 2004, 104,4245–4269.

2. Whittingham, M. S.; Savinell, R. F.; Zawodzinski,T. Chem Rev 2004, 104, 4243, 4244.

SYMBOLS

jp Proton fluxCp Concentration of protonsDp Diffusion coefficient of protonsZp Protonic chargeC Electrostatic potentialR Gas constantF Faradayr Proton conductivityA cross-sectional areak Hydration numberA NMR signal intensity.c Gyromagnetic ratiod Length of gradient pulseD Time between gradient pulseDr Mean square displacementDeff Effective self diffusion coefficient of water

2238 ROY ET AL.


3. Mauritz, K. A.; Moore, R. B. Chem Rev 2004, 104,4535–4585.

4. Yeo, R. S. J Electrochem Soc 1983, 130, 533–538.

5. Kim, Y. S.; Dong, L. M.; Hickner, M. A.; Glass, T.E.; Webb, V.; McGrath, J. E. Macromolecules2003, 36, 6281–6285.

6. Wang, F.; Hickner, M.; Kim, Y. S.; Zawodzinski, T.A.; McGrath, J. E. J Membr Sci 2002, 197, 231–242.

7. Hickner, M. A.; Ghassemi, H.; Kim, Y. S.; Einsla,B. R.; McGrath, J. E. Chem Rev 2004, 104, 4587–4611.

8. Wang, F.; Hickner, M.; Ji, Q.; Harrison, W.;Mecham, J.; Zawodzinski, T. A.; McGrath, J. E.Macromol Symp 2001, 175, 387–395.

9. Kim, Y. S.; Sumner, M. J.; Harrison, W. L.; Riffle,J. S.; McGrath, J. E.; Pivovar, B. S. J ElectrochemSoc 2004, 151, A2150–A2156.

10. Kim, Y. S.; Hickner, M. A.; Dong, L. M.; Pivovar,B. S.; McGrath, J. E. J Membr Sci 2004, 243,317–326.

11. Kim, Y. S.; Wang, F.; Hickner, M.; McCartney, S.;Hong, Y. T.; Harrison, W.; Zawodzinski, T. A.;McGrath, J. E. J Polym Sci Part B: Polym Phys2003, 41, 2816–2828.

12. Roy, A.; Einsla, B.; Harrison, W.; McGrath, J. E.Am Chem Soc Div Fuel Chem Preprints 2004, 49,614–615.

13. Li, Y. X.; Mukundan, T.; Harrison, W. L.; Hill, M.;Sankir, M.; Yang, J.; McGrath, J. E.; Am ChemSoc Div Fuel Chem Preprints, 2004, 49, 536, 537.

14. Xiang, Y.; Roy, A.; McGrath, J. E. PreprintsSymp – Am Chem Soc Div Fuel Chem 2005, 50,577–578.

15. Kreuer, K. D. Chem Mater 1996, 8, 610–641.16. Kreuer, K. D. Solid State Ionics 2000, 136, 149–

160.17. Paddison, S. J. Annu Rev Mater Res 2003, 33,

289–319.18. Van Grotthuss, C. J. D. Ann Chim 1806, 58, 54.19. Kreuer, K. D.; Rabenau, A.; Weppner, W. Angew

Chem Int Ed Engl 1982, 21, 208, 209.20. Zawodzinski, T. A.; Neeman, M.; Sillerud, L. O.;

Gottesfeld, S. J Phys Chem 1991, 95, 6040–6044.

21. Dong, L. M.; Kim, Y. S.; Wang, F.; Hickner, M.;Glass, T. E.; McGrath, J. E. Abstr Pap Am ChemSoc 2003, 225, U631.

22. Zawodzinski, T. A.; Davey, J.; Valerio, J.; Gottes-feld, S. Abstr Pap Am Chem Soc 1993, 205,PMSE-75.

23. Noshay, A.; McGrath, J. E. Block Copolymers:Overview and Critical Survey; Academic Press:New York, 1977; p 91.

24. Ghassemi, H.; Ndip, G.; McGrath, J. E. Polymer2004, 45, 5855–5862.

25. Ghassemi, H.; McGrath, J. E. Polymer 2004, 45,5847–5854.

26. Ghassemi, H.; Zawodzinski, T.; McGrath, J. E.Polymer, in press.

27. Pivovar, B. S.; Wang, Y. X.; Cussler, E. L.J Membr Sci 1999, 154, 155–162.

28. Ren, X. M.; Springer, T. E.; Zawodzinski, T. A.;Gottesfeld, S. J Electrochem Soc 2000, 147, 466–474.

29. Hickner, M. PhD Thesis, Virginia PolytechnicInstitute and State University, Blacksburg, Vir-ginia, August 2003.

30. Stejskal, E. O.; Tanner, J. E. J Chem Phys 1965,42, 288–292.

31. Zawodzinski, T. A.; Springer, T. E.; Neeman,M.; Sillerud, L. O. Isr J Chem 1992, 32, 281–289.

32. Matsukawa, S.; Yasunaga, H.; Zhao, C.; Kuroki,S.; Kurosu, H.; Ando, I. Prog Polym Sci 1999, 24,995–1044.

33. Matsukawa, S.; Ando, I. Macromolecules 1999,32, 1865–1871.

34. Zhao, C. H.; Matsukawa, S.; Kurosu, H.; Ando, I.Macromolecules 1998, 31, 3139–3141.

35. Ghi, P. Y.; Hill, D. J. T.; Maillet, D.; Whittaker, A.K. Polymer 1997, 38, 3985–3989.

36. Ghi, P. Y.; Hill, D. J. T.; Whittaker, A. K. Bioma-cromolecules 2002, 3, 554–559.

37. Ghi, P. Y.; Hill, D. J. T.; Whittaker, A. K. Bioma-cromolecules 2002, 3, 991–997.

38. Sankir, M.; Bhanu, V. A.; Harrison, W. L.;Ghassemi, H.; Wiles, K. B.; Glass, T. E.; Brink,A. E.; Brink, M. H.; McGrath, J. E. J ApplPolym Sci, 2006, 100, 4595–4602.

39. Springer, T. E.; Zawodzinski, T. A.; Wilson, M. S.;Gottesfeld, S. J Electrochem Soc 1996, 143, 587–599.

40. Rollet, A. L.; Blachot, J.; Delville, A.; Diat, O.;Guillermo, A.; Porion, P.; Rubatat, L.; Gebel, G.Eur Phys J E 2003, 12, S131–S134.



Documents

Influence of chemical composition and sequence length on the transport properties of proton exchange membranes