novel multiblock-co-ionomers as potential polymer ... ke… · novel multiblock-co-ionomers as...

Novel Multiblock-co-Ionomers as PotentialPolymer Electrolyte Membrane Materials

FRANK SCHONBERGER, JOCHEN KERRES

Institute for Chemical Process Engineering, University of Stuttgart, Boblinger Strasse 72, Stuttgart 70199, Germany

Received 2 May 2007; accepted 22 June 2007DOI: 10.1002/pola.22269Published online in Wiley InterScience (www.interscience.wiley.com).

ABSTRACT: In this contribution a series of novel multiblock-co-ionomers consisting ofhydrophobic (partially fluorinated) and hydrophilic (sulfonated) domains has beenprepared and characterised in terms of their applicability as fuel cell membranes.The synthesis of these multiblock-co-ionomers is a four-step procedure including (1)the sulfonation of the monomer 4,40-difluorodiphenylsulfone, (2) the preparation ofhydrophilic telechelic macromonomers by molecular-weight controlled step-growthpolycondensation of the sulfonated monomer with various bis(thio)phenols, (3) thepreparation of a hydrophobic telechelic macromonomer and (4) the coupling of bothtelechelic macromonomers to yield microphase-separated block-co-ionomers. Thisstudy focuses on the investigation of the influence of various linkage groups andatoms within the hydrophilic domains of the multiblock-co-ionomers. Both the tele-chelic macromonomers and the multiblock-co-ionomers were structurally investigatedby 1H- and 19F-NMR spectroscopy and gel permeation chromatography (GPC). Allmultiblock-co-ionomers of this series could be cast into membranes and their mem-brane properties (ion-exchange capacity, specific resistance, swelling ratio, wateruptake, thermal and oxidative stability) were measured and discussed in dependenceof the various linkage groups within the hydrophilic domains. VVC 2007 Wiley Periodicals,

Inc. J Polym Sci Part A: Polym Chem 45: 5237–5255, 2007

Keywords: Carother’s equation; direct methanol fuel cell; gel permeation chroma-tography (GPC); macromonomers; multiblock-co-ionomers; NMR; polymer electrolytemembrane fuel cell; step-growth polycondensation

INTRODUCTION

Polymer electrolyte membranes (PEMs) havebeen on focus of research and development formany years as one of their applications is theuse in polymer electrolyte membrane fuel cells(PEMFCs) or direct methanol fuel cells (DMFCs),which are promising, environmentally friendlyand efficient power sources. The membrane ma-terial has to fulfil a number of criteria to with-stand the harsh conditions during operation andto ensure an economically reasonable running ofthe system. Beside high protonic and zero elec-

tronic conductivity, the membrane has to possesslow permeability to fuel and oxidant, low watertransport through diffusion and electro-osmosis,oxidative and hydrolytic stability, good mechani-cal properties in both the dry and the hydratedstate, and finally economically reasonable produc-tion and maintenance costs.1,2 The current state-of-the-art PEM material is polyperfluoroalkylsul-fonic acid3 (e.g., Nafion1, Flemion1, or Dow1)having good mechanical, thermal, and chemicalstability as well as good protonic conductivity attemperatures below 80 8C.4 However, majordrawbacks of this material are the high meth-anol-crossover that significantly lowers the effi-ciency of a DMFC system, the reduction inconductivity at higher temperatures and the rela-tively low glass transition temperature (ca.

Correspondence to:F. Schonberger (E-mail: [email protected])

Journal of Polymer Science: Part A: Polymer Chemistry, Vol. 45, 5237–5255 (2007)VVC 2007 Wiley Periodicals, Inc.

5237

110 8C in the dry state and considerably lower inthe hydrated state) limiting the operation tem-perature of fuel cells.5 Additionally the compli-cated production process, which includes stronglyhazardous intermediates, demands for highsafety and environment precautions.6 During thelast decades much effort in designing alternativePEM materials has been made and reviewed inliterature by a lot of different authors.1,3,6–9

Many different families of polymers and differentstrategies of introducing ionogenic groups (sul-fonic and/or phosphonic acids10–13) have beenexplored. Sulfonated poly(arylene ether)s andpoly(arylene ether sulfone)s are considered asgood candidates in terms of high thermal, oxida-tive, hydrolytic, and mechanical stability as wellas reduced methanol permeability.4 An evidenttrend in the design of novel PEM materials andthus the tailoring of properties is the control ofpolymer architecture, including the number andposition of sulfonic acid groups as well as themorphology. Beside the preparation of statisticalco-ionomers starting from a mixture of sulfo-nated/unsulfonated difluoromonomers and appro-priate bisphenol monomers4,14–17 there are a fewexamples for designing morphologically con-trolled and microphase-separated sulfonated poly-mers. An interesting approach on this kind ofpolymers has recently been done by Holdcroftet al. who reported on highly fluorinated comb-shaped copolymers18 and ionic graft polymers19

to mimic the principal architecture of polyper-fluoroalkylsulfonic acids with improved mechani-cal integrity at higher temperatures and lower tomoderate glass transition temperatures comparedwith Nafion1. The preparation of sulfonatedmultiblock-co-ionomers for fuel cell membranesconsisting of hydrophilic (sulfonated) and hydro-phobic segments, in which the first one serves forproton transport and the latter one for increasingmechanical stability in comparison with the re-spective homo-ionomers [by lowering the wateruptake (WU)] is sparsely reported in literature.Nonfluorinated sulfonated multiblock-co-poly(ary-lene ether ketone)s were synthesised and char-acterized with respect to fuel cell applications byScherer and coworkers.20 Meier-Haack and co-workers delved into the synthesis of nonfluorinatedsulfonated multiblock-co-poly(ether sulfone)s.21

Recently, Miyatake and coworkers could dem-onstrated that sulfonated multiblock-co-poly-imides have superior proton conductivity, espe-cially at low humidity, when compared withthe corresponding sulfonated statistical co-poly-imides.22 McGrath’s group published the synthe-

sis and characterization of a multiblock-co-poly(arylene ether sulfone) consisting of highly hy-drophobic partially fluorinated segments and sul-fonic acid groups in the hydrophilic segmentbeing located meta to a sulfone bridge.23–25 Theelectron-withdrawing effects (-M and -I effects) ofthe sulfone bridge guarantee a high acid strengthof the ionogenic groups,26–28 whereas the highlyfluorinated hydrophobic segments in the multi-block architecture counteract the inherent tend-ency of the sulfonic acid group to attract watermolecules, which should be favorable for highproton conductivity and mechanical stability.

In this contribution a series of novel partiallyfluorinated multiblock-co-ionomers has beensynthesised and the influence of the chemicalbridging groups (E and X) in the hydrophilicblock has been comparatively investigated (cf.Scheme 1 for notations). The multiblock archi-tecture was synthetically built up by combiningF-terminated hydrophobic (5) and OH- or SH-terminated hydrophilic telechelic macromono-mers (1a–4a), which were prepared from thecorresponding monomers in the course of a nucle-ophilic aromatic polycondensation. Macromono-mer 5 was chosen to provide highest possiblehydrophobicity for the hydrophobic block andto avoid disadvantages2,26 raised from other acti-vated partially fluorinated monomers (like car-bonyl groups in a highly fluorinated environment).

The length and thus the degree of polymer-ization (DP) of both the hydrophobic (m) and thehydrophilic (n) block, respectively, was fixed tom ¼ 8 and n ¼ 4, respectively, to achieve ion-exchange capacities (IEC) comparable to that ofNafion1. The DP for those step-growth polymer-isations was adjusted by the stoichiometricimbalance r of the two monomers according toCarother’s Equation.29,30 The calculation of thisstoichiometric imbalance r is exemplified for thesynthesis of the F-terminated hydrophobic tele-chelic macromonomer 5 in the following. Itshould be noted that—unlike a system withbifunctional monomers with different functionalgroups (A–B, e.g., hydroxyl- or amino acids)—adifference between the number average DP Xnh iand the DP in a system with two bifunctionalmonomers of the type A–A and B–B exists. Forthe considered case of the bifunctional mono-mer 2,2-bis(4-hydroxyphenyl)hexafluoroisopro-pane (A–A) and decafluorobiphenyl (that canbe regarded as bifunctional monomer because ofthe much higher reactivity of the fluorines in4,40-position than in 2,20- or 3,30-position, B–B)

5238 SCHONBERGER AND KERRES

Journal of Polymer Science: Part A: Polymer ChemistryDOI 10.1002/pola

this difference between Xnh i and the DP (m) fol-lows the simple correlation given in the follow-ing equation.

Xnh i ¼ 2DP ¼ 2m ¼ 16 ð1ÞThe DP and the targeted molecular weight

(Mtarget) are related by eq 2 (Mru ¼ molecularweight of the repeating unit)

Mtarget ¼ DP3Mru ¼ 83 630:34 Da � 5043Da

ð2ÞFinally the stoichiometric imbalance r, which

is the ratio of monomer A–A to monomer B–B,can be calculated by the following equation.

r ¼ nA�A

nB�B¼ Xnh i � 1

Xnh i þ 1¼ 15

17¼ 0:882 ð3Þ

EXPERIMENTAL

Instrumentation

Ion-exchange capacities (IECdirect and IECtotal)were determined by titration. Membranes in theHþ form were immersed in saturated sodiumchloride solution (NaCl) for 24 h to convert theminto the Naþ form. The exchanged Hþ ions werethen titrated with 0.1 M NaOH to the equiva-lent point (IECdirect). After that a defined excessof NaOH was added and this solution was back-titrated with 0.1 M HCl (IECtotal). The specificresistance of membranes was determined byelectrochemical impedance spectroscopy, using amethod described in literature,31 using 0.5 MHCl as electrolyte at 25 8C on IM6 Model of

Zahner Elektrik. The WU of the samples wasdetermined after equilibrating in water ofdefined temperatures (25, 40, 60, and 90 8C). Af-ter 48 h the sample was removed from the watersolution, quickly dry wiped, and immediatelyweighed (mwet). Then the sample was dried toweight constancy at 90 8C and weighed onceagain (Mdry). The WU can be obtained from thesetwo values according to the following equation.

WUð%Þ ¼ mwet �mdry

mdry3 100 ð%Þ ð4Þ

The procedure for the determination of theswelling degree is similar and is calculated fromthe length in the dry (ldry) and in the wet (lwet)state as follows.

Swelling degreeð%Þ ¼ lwet � ldryldry

3 100 ð%Þ ð5Þ

Thermal stability of the membranes wasdetermined by thermogravimetry (TGA, Netz-sch, model STA 449C) with a heating rate of20 8C/min under an atmosphere enriched withoxygen (65–70% O2, 35–30% N2). The decompo-sition gases were further examined in a coupledFTIR spectrometer (Nicolet Nexus FTIR spec-trometer) to identify the splitting-off tempera-ture of the sulfonic acid group (TSO3H onset), forwhich the asymmetric stretching vibration ofthe S¼¼O group at 1352–1342 cm�1 was used.

Isothermal TGA measurements (at 200, 220,and 240 8C) were done on the same machine(Netzsch, model STA 449C) and under the sameoxygen-enriched atmosphere. Membrane pieces

Scheme 1. Overview of the synthesised multiblock-co-ionomers (n ¼ 4 and m ¼ 8;n, m ¼ degree of polymerization).

MULTIBLOCK-co-IONOMERS AS POTENTIAL PEM MATERIALS 5239


were dried in a vacuum oven at 90 8C for 16 hbefore measurement.

Nuclear magnetic resonance spectra wererecorded using a Bruker Avance 400 spectrometerat a resonance frequency of 400 or 250 MHz for1H, 188 MHz for 19F and 250 MHz for 1H,13C-HSQC (heteronuclear single quantum coherence).

The morphology of the multiblock-co-ionomerswas determined by transmission electron mi-croscopy (TEM), using a Philips CM10 TEM. Tohave a better contrast in the images, mem-branes were pretreated in 1 M Pb(NO3)2 at90 8C for 24 h. Subsequently excessive Pb(NO3)2was washed out by repeated immersion of themembrane pieces in water at 60 8C. Then themembranes were dried in an oven at 90 8C for24 h. The further preparation of sample forTEM has already been described earlier.31

Gel permeation chromatography (GPC) wascarried out with an Agilent Technologies GPCsystem (Series 1200), using the universal cali-bration method and a refractive index detector(Shodex RI71).

Materials

4,40-Dihydroxydiphenylsulfide, 4,40-sulfanediyl-bis-thiophenol, 2,2-bis(4-hydroxyphenyl)sulfone,2,2-bis(4-hydroxyphenyl)hexafluoroisopropane,decafluorobiphenyl, and anhydrous potassiumcarbonate were purchased at Aldrich and weredried at 60 8C in a vacuum oven (100 mbar) for 16h before use. Disodium-4,40-difluorodiphenylsul-fone-3,30-disulfonate is not commercially availableand was synthesised from 4,40-difluorodiphenylsul-fone (ABCR). All other reagents were supplied byAldrich and used as received.

Disodium-4,40-difluorodiphenylsulfone-3,30-disulfonate

This monomer was synthesized from 4,40-difluorodiphenylsulfone following a procedurefrom literature,32,33 which was optimized interms of the recrystallization reagent.

About 25.71 g (98.33 mmol) 4,40-difluorodiphe-nylsulfone were dissolved in 60 mL of fumingsulphuric acid (30% SO3) and stirred for 6 h at110 8C. The solution was allowed to cool to roomtemperature and then poured into 400 mL of icewater. Then 180 g sodium chloride was added toprecipitate a white product. The raw productwas filtered, dried at 90 8C in a vacuum ovenand redissolved in 600 mL of deionized water.After the pH value of this solution was adjustedto pH ¼ 8 by 2 M NaOH, another 180 g were

added to isolate the product. This product wasrecrystallized from methanol/water (8/1) firstand then from methanol/water (8/2). The prod-uct was dried at 90 8C in a vacuum oven beforecharacterisation. Yield: 82%.

ELEM. ANAL for C12H6O8S3F2Na2�2H2O [exper-imental values (theoretical values)]: C, 29.33(29.15); H, 2.38 (2.04); S, 19.10 (19.46); Cl, 0.44(0.00).

1H NMR (200 MHz, DMSO-d6, d): 8.19 (dd, J1

¼ 6.33 Hz, J2 ¼ 2.52 Hz, 3H, 1H), 8.02 (m, 5H,1H), 7.46–7.50 (m, 6H, 1H), 3.44 (s, 7H, 2H). 13CNMR (100 MHz, DMSO-d6, d): 161.25 (d, 1JC��F

¼ 259.5 Hz, 1C), 136.48 (d, 4JC��F ¼ 18.6 Hz,4C), 136.25 (d, 2JC��F ¼ 3.3 Hz, 5C), 131.29 (d,2JC��F ¼ 9.9 Hz, 2C), 128.46 (d, 3JC��F ¼ 4.6 Hz,3C), 118.61 (d, 3JC��F ¼ 24.3 Hz, 6C).

Preparation of the OH-Terminated HydrophilicTelechelic Macromonomers 1a–4a

The hydrophilic telechelic macromonomers 1a–4awere synthesised from disodium-4,40-difluoro-diphenylsulfone-3,30-disulfonate and the respectivebisphenols according to a standard nucleophilicdisplacement polycondensation whereof the DPwas adjusted to n ¼ 4 by using a stoichiometricimbalance r of n ¼ 0.778.

About 33.653 mmol of the bisphenol com-pound (7.3454 g 4,40-dihydroxydiphenylsulfidefor 1a, 8.4267 g 4,40-sulfanediyl-bis-thiophe-nol for 2a, 8.4223 g 2,2-bis(4-hydroxyphenyl)sul-fone for 3a and 11.3151 g 2,2-bis(4-hydroxyphe-nyl)hexafluoroisopropane for 4a) were dissolvedin NMP (62.5 mL for 1a, 66 mL for 2a and 3a,72.5 mL for 4a) in a 250-mL three-neck flaskequipped with magnetic stir bar, argon inlet,reflux condenser, and mercury bubbler. After themonomer had been dissolved completely 13.82 g(0.100 mol) potassium carbonate were addedand the mixture was heated to 80 8C for 7 h.Then a solution of 12.0000 g (26.182 mmol) ofdisodium-4,40-difluorodiphenylsulfone-3,30-disulf-onate in NMP (same amounts as aforemen-tioned) was dropped into the reaction mixtureand the temperature was raised to 130 8C forfurther 24 h. The reaction mixture was pouredinto 1 L iso-propanol to precipitate the product,which was then separated and dissolved in



hydrochloric acid (10 wt %) at 60 8C for 4 h.Finally this solution was dialyzed for 48 h andthe desired product was recovered after the

water was evaporated. The product was dried at110 8C for 16 h in a vacuum oven before charac-terization and further use.

Telechelic 1a

1H NMR (DMSO-d6, 400 MHz, d): 8.29 (s, 5H,8H, 13H, 1.25 H), 7.86 (d, 3JH��H ¼ 8.3 Hz, 9H,15H), 7.83 (d, 3JH��H ¼ 6.9 Hz, 7H), 7.36 (d,3JH��H ¼ 7.9 Hz, 12H, 2.00 H), 7.29 (d, 3JH��H ¼8.3 Hz, 3H, 0.44 H), 7.15 (d, 3JH��H ¼ 8.2 Hz, 1H,

0.43 H), 7.04 (d, 3JH��H ¼ 7.8 Hz, 11H), 7.01 (d,3JH��H ¼ 8.8 Hz, 10H, 14H), 6.97 (d, 3JH��H ¼ 8.2Hz, 4H), 6.91 (d, 3JH��H ¼ 8.5 Hz, 6H, 0.24 H),6.82 (d, 3JH��H ¼ 8.4 Hz, 2H, 0.41 H). Integralvalues for the overlapping signals: cf. Figure 2.

1H NMR (DMSO-d6, 400 MHz, d): 8.17 (s,5H, 8H, 13H, 1H), 7.62 (fine structure not recog-nizable, 7H, 9H, 15H, 1H), 7.53–7.42 (signal

heap, 2H, 3H, 4H, 11H, 12H, 4.18H), 7.30 (d,3JH��H ¼ 7.7 Hz, 1H 0.57H), 6.90–6.82 (finestructure not recognizable, 6H, 10H, 14H, 1H).

1H NMR (DMSO-d6, 400 MHz, d): 8.32 (s, 5H,8H, 13H, 1H), 7.97–7.83 (superposition of peaks,3H, 7H, 9H, 12H, 15H, 3.30H; doublet of 12H canstill be seen at 7.90 with 3JH��H ¼ 8.4 Hz), 7.73(d, 3JH��H ¼ 8.7 Hz, 2H, 0.34H), 7.23 (d, 3JH��H

¼ 8.4 Hz, 6H, 10H, 14H, 1.00H), 7.16–7.01 (super-position of peaks, 11H, 4H, 2.36H; doublet of 11Hcan still be seen at 7.10 with 3JH��H ¼ 8.6 Hz),6.91 (d, 3JH��H ¼ 8.7 Hz, 1H, 0.36H).

Telechelic 2a

Telechelic 3a

Telechelic 4a



1H,13C-HSQC (DMSO-d6, 250 MHz, d): 8.32(s, 5H, 8H, 13H, 1H), 128.64 (5C, 8C, 13C), 7.90(d, 3JH��H ¼ 6.3 Hz, 7H, 9H, 15H, 0.97H), 130.57(7C, 9C, 15C), 7.35 (d, 3JH��H ¼ 7.5 Hz, 3H, 12H,2.26), 131.73 (3C, 12C), 7.11 (d, 3JH��H ¼ 6.7 Hz,4H, 11H, 2.27H), 119.53 (4C, 11C), 7.10 (d, 3JH��H

¼ 6.6 Hz, 6H, 10H, 14H, 0.99H), 120.94 (6C, 10C,14C), 7.10 (d, 3JH��H ¼ 7.1 Hz, 2H, 0.57H),131.61 (2C), 6.82 (d, 3JH��H ¼ 8.7 Hz, 1H, 0.55H).

Preparation of the F-Terminated HydrophobicTelechelic Macromonomer 5

The hydrophobic telechelic macromonomer 5was synthesised under similar conditions as thecorresponding polymer.34 Divergent from thatprocedure, the stoichiometric ratio was fixed to r¼ 0.882 to obtain m ¼ 8 and the reaction timewas reduced to 3 h.

About 26.6900 g (79.38 mmol) 2,2-bis(4-hydroxyphenyl)hexafluoropropane and 30.0708(90.00 mmol) decafluorobiphenyl were dissolvedin 420 mL N,N-dimethylacetamide (DMAc) in a1000-mL three-neck flask equipped with an ar-gon inlet, a mechanical stirrer, a reflux con-denser and a mercury bubbler. 30.68 g (0.222mol) of potassium carbonate were added afterthe monomers had been dissolved completelyand the reaction mixture was stirred at 80 8Cfor 3 h. Then the reaction mixture was pouredinto 2 L of stirred water, the precipitation wasfiltered, washed twice with water and methanoland dried at 75 8C in a vacuum oven for 16 h.

1H NMR (200 MHz, CDCl3, d): 7.07 (d, 3JH��H

¼ 8,4 Hz, 1H), 7.43 (d, 3JH��H ¼ 8,8 Hz, 1H). 13CNMR (50 MHz, CDCl3, d): 157.60, 145.03 (d,1JC��F ¼ 253 Hz), 141.80 (d, 1JC��F ¼ 253 Hz),135.03–134.41 (nJ, n � 2, not further analysable),132.04, 128.97, 115.51, 103.57–102.91. 13C NMRsignals of the hexafluoroisopropylidene group are

not visible because of the strong coupling with 19F(quartet, septet). 19F NMR (188 MHz, CDCl3, d):�64.35 (6.00 F), �137.74 (4.46 F), �150.19 (0.25F), �152.65 (4.06 F), 160.64 (0.50 F).

Polycondensation of the TelechelicMacromonomers

The nucleophilic displacement polycondensationreactions for the preparation of 1b–4b were car-ried out with the same starting concentrationsof monomers [13 wt % in N-methylpyrrolidinone(NMP)], the same threefold excess of potassiumcarbonate related to the OH-terminated te-lechelic hydrophilic macromonomer (1a–4a)and the same starting temperature. Since theincrease in viscosity of the different reactionmixture was dramatically different indicatingvarious reactivities of the four macromonomers1a–4a, there was a need to adapt the dilutionand temperature for each reaction batch individ-ually. The individual reactions conditions aresummarized in Table 5. An explicit procedure isgiven in the following for the preparation of themultiblock-co-ionomer 2b.

About 8.5090 g (13.499 mmol) of the hydrophobictelechelic macromonomer 5 and 8.0000 g (13.449mmol) of the hydrophilic telechelic macromonomer2a were dissolved in 108 mL NMP under argonatmosphere and intensive stirring. This mixturewas heated to 80 8C for 2 h. Then 5.5971 g (40.497mmol) potassium carbonate was added. The reac-tion mixture was stirred at 80 8C and diluted by 20mL NMP after 90, 135, 150, 180, 210 min, and byfurther 80 mL NMP after 225 min. Then it wasstirred for further 15 min and then poured into 2 Lof distilled water. The product was separated by fil-tration, digested in 500 mL of hydrochloric acid (10wt %) at 60 8C for 4 h. Finally the product was fil-tered and dried at 90 8C in a vacuum oven for 16 h.

Multiblock-co-Ionomer 1b

19F NMR (188 MHz, DMSO-d6, d): �59.01(5F, 6 F), �133.59 [2F, 3F, 7F, 8F, 4.53 F (calcu-lated with m ¼ 8: 4.50 F)], �149.05 [1F, 4F, 6F,9F, 4.53 F (calculated with m ¼ 8: 4.50 F)]. 1HNMR (250 MHz, DMSO-d6, d): 8.26 (s, 3H, 1H),

7.83 (d, 3JH��H ¼ 8.1 Hz, 4H, 1H), 7.37–7.34[signal heap, 2H, 5H, 7H, 8H, 6.13H (calculatedfor n ¼ 4 and m ¼ 8: 7.50 H)], 7.04–6.97 [signalheap, 1H, 6H, 9H, 10H, 3.50H (calculated for n ¼4 and m ¼ 8: 6.50 H)].



1H NMR (250 MHz, DMSO-d6, d): 8.17 (s, 3H,1H), 7.65–7.37 [signal heap, 1H, 2H, 4H, 6H, 7H,8H, 9H, 10H, 11H, 10.02H (calculated for n ¼ 4

and m ¼ 8: 14.00 H)], 6.89–6.83 (d, 3JH��H ¼7.90 Hz, 5H, 1H).


19F NMR (188 MHz, DMSO-d6, d): �59.00 (5F,6 F), �133.47 [2F, 3F, 7F, 8F, 4.57 F (calculatedwith m ¼ 8: 4.50 F)], �148.97 [1F, 4F, 6F, 9F,4.63 F (calculated for m ¼ 8: 4.50 F)]. 1H NMR(250 MHz, DMSO-d6, d): 8.33 (s, 3H, 1H), 7.93

(signal heap, 2H, 4H, 7H, 8H, 3.19H (calculatedfor n ¼ 4 and m ¼ 8: 2.50H), 7.38–7.10 [signalheap, 1H, 6H, 9H, 10H, 11H, 5.91H (calculated forn ¼ 4 and m ¼ 8: 6.50 H)].

19F NMR (188 MHz, DMSO-d6, d): �58.80 –�59.02 (1F, 2F, 7F, 1 F), �133.59 [4F, 5F, 9F, 10F,0.44 F (calculated for m ¼ 8: 0.46 F)], �149.03[3F, 6F, 8F, 11F, 0.45 F (calculated for m ¼ 8: 0.46F)]. 1H NMR (250 MHz, DMSO-d6, d): 8.30 (s,

3H, 1H), 7.89 (d, 3JH��H ¼ 7.9 Hz, 4H, 1H), 7.38–7.34 [signal heap, 2H, 5H, 7H, 8H, 5.86H (calcu-lated for n ¼ 4 and m ¼ 8: 7.50 H)], 7.04–6.97[signal heap, 1H, 6H, 9H, 10H, 3.53H (calculatedfor n ¼ 4 and m ¼ 8: 6.50 H)].



Preparation of Model Polymer 8

About 12.3197 g (49.20 mmol) of 4,40-sulfane-diyl-bis-thiophenol and 16.7060 g (50.00 mmol)of decafluorobiphenyl were dissolved in 140 mLNMP under argon atmosphere and intensivestirring. Then 19.32 g (0.140 mmol) potassiumcarbonate was added and the reaction mixturewas stirred at room temperature for 6 h (consid-

erable increase in viscosity). This mixture waspoured into 2 L of methanol under stirring andthe polymer was filtered, washed with methanoltwice and finally dried at 90 8C in a vacuumoven for 16 h.

1H NMR (200 MHz, CDCl3, d): 7.38 (d, 3JH��H

¼ 8.30 Hz, 1H), 7.27 (d, 3JH��H ¼ 7.27 Hz, 1H).19F NMR (188 MHz, CDCl3, d): �131.88 (1 F),�136.97 (1 F).



Membrane Preparation

Membranes were cast into aluminium bowlsfrom their NMP/DMAc solutions (10 wt %) anddried at 130 8C at atmospheric pressure forat least 3 h. All membranes were soaked in10 wt % HCl at 90 8C for 48 h and subsequentlywashed with deionized water. After that, themembranes were immersed in deionized waterat 60 8C for another 48 h.

RESULTS AND DISCUSSION

Characterization of the Telechelic Macromonomersand Multiblock-co-Ionomers

Both the hydrophilic and the hydrophobic tele-chelic macromonomers were synthesised bystep-growth polycondensation, using the respec-tive calculated stoichiometric imbalance r. 1HNMR and 1H,13C-HSQC-NMR (heteronuclearsingle quantum coherence nuclear magnetic res-onance) spectroscopy was used as characterisa-tion tool in the case of the OH-terminatedhydrophilic macromonomers (1a–4a) whereas19F NMR spectroscopy served for the elucidationof the F-terminated hydrophobic macromonomer5. NMR spectroscopy could also be used for theestimation of the number average molecularweight Mn, which is exemplified for the endgroup analysis of 1a (1H NMR) and 4a (1H,13C-HSQC-NMR) as representatives for the hydro-philic telechelic macromonomers and of 5 (19FNMR). The number average molecular weightMn

NMR values are compared with the calculatedones and those (Mn

GPC) obtained from GPC inTable 1. Figure 2 shows the 1H NMR spectrumof the OH-terminated hydrophilic telechelicmacromonomer 1a. Signals were assigned by

comparing the chemical shifts of the monomerand of the corresponding homo-ionomer. The po-lymerization degree n (n ¼ x þ 1, x is the poly-merization degree chosen in Figure 2 for thepeak assignment, the real polymerization degreen is higher for one repeating unit) was esti-mated from the ratio of signal intensities to be3.33 according to the following equations.

x ¼ Ið12HÞ2Ið3HÞ ¼

2

2 � 0:43 ¼ 2:33 ð6Þ

n ¼ xþ 1 ¼ 3:33 ð7Þ

In case of the OH-terminated hydrophilicmacromonomer 4a the one-dimensional 1H NMRspectrum could not reveal the chemical struc-ture in detail because of an overlapping ofpeaks. The determination of the chemical struc-

Figure 1. Molecular weight distribution of multi-block-co-ionomer 3b in comparison with the corre-sponding macromonomers 3a and 5, determined byGPC. [Color figure can be viewed in the online issue,which is available at www.interscience.wiley.com.]

Table 1. Comparison of the Calculated Number Average Molecular Weight (Mn) with the Experimental ValuesObtained from End Group Analysis by NMR (Mn

NMR) or from GPC (MnGPC)

Telechelic Macromonomer Mntarget [Da] Mn

NMR [Da] MnGPC [Da] PDI

1a 2,549 2,101 2,990 1.712a 2,677 2,443 4,275 1.303a 2,677 2,490 6,063 1.264a 3,022 3,240 7,405 1.545 5,043 5,270 2,990 2.50

Mntarget, calculated number average molecular weight; Mn

NMR, number average molecular weight determined by endgroup analysis using NMR spectroscopy; Mn

GPC, number average molecular weight determined by GPC; PDI, polydispersity indexdetermined by GPC.



ture was carried out by 1H,13C-HSQC. Figure 3shows the assignment of the peaks togetherwith the chemical shifts and the integrals. Thenumber average molecular weight was calcu-

lated by the ratio of signal intensities asdescribed earlier for 1a.

19F NMR spectroscopy gave information on theend groups and thus the number average molecular

Figure 2. 1H NMR spectrum of the hydrophilic telechelic macromonomer 1a.

Figure 3. 1H-, 13C-HSQC spectrum of the OH-terminated telechelic hydrophilicmacromonomer 4a.



weight of macromonomer 5 as depicted in Figure 4.The peaks at �64.35, �137.78, and �152.65 ppmcan be assigned to the chain fluorine atoms 3F, 2F,and 1F in accordance with the observed chemicalshift for the corresponding homo-polymers.26 Thelittle fronting observed for all those peaks and thusthe increased integrals of the peak at �137.78 and�152.65 ppm might be attributed to the modifiedchemical environment at the end of the chains andcaused from the fluorine atoms 2F, 4F, and 6F.Finally the remaining two signals at �150.19 and�160.64 ppm were attributed to be caused by thefluorine atoms 7F and 8F, which is supported by thereported chemical shift values of decafluorobi-phenyl.35 Using this assignment the polymerizationdegree can be calculated to m ¼ 8.36 correspondingto a number average molecular weight of Mn ¼5270 Da, which is in good accordance with the cal-culated value Mn ¼ 5043 Da for the used stoichio-metric imbalance.

The comparison in Table 1 between the experi-mental Mn

NMR and MnGPC values with the calcu-

lated Mntarget reveals a certain discrepancy for

these two methods in the characterization of thetelechelic macromonomers 1a–4a and 5. Obvi-ously the calculations of Mn from the NMR spec-tra match the Mn

target values quite well whereasthose obtained from GPC analysis show abroader variation. It can be further seen fromTable 1 that the Mn

GPC values for the hydrophilictelechelic macromonomers (1a–4a) are higherbut for the hydrophobic one (5) lower than thecalculated value Mn

target. This variation mightbe associated with deviations from the Mark–

Houwink relation for polymers with molarmasses below 10–20 kDa (depending on the poly-mer class). The validity of the applied universalcalibration is based on the Mark–Houwink coeffi-cient, which are derived by assuming a Gaussiandistribution for the polymer coils.36–39 Thisassumption might not necessarily be fulfilled forall polymer classes below a certain molecularweight. Therefore NMR spectroscopy is thoughtto be more precise for the determination of Mn

for the telechelic macromonomers. It can be con-cluded from Table 1 that both the hydrophilic tel-echelic macromonomers 2a–4a and the hydropho-bic telechelic macromonomer 5 were formed inthe desired and afore calculated number averagemolecular weight Mn

target. Only 1a shows aslightly higher Mn

NMR value as calculated, whichcorresponds to a block length of n ¼ 5.5 insteadof n ¼ 4. Despite a relatively good accordancebetween the presented Mn

NMR values andthe calculated Mn

target one has to consider thatthe applied classical theory of step-growth poly-merisation, which is mainly based on Carother’sand Flory’s work does not consider contributionsof cyclization reactions and does not differentiatebetween kinetically and thermodynamically con-trolled step-growth polymerisations.40,41 Figure 1reveals at the example of the hydrophilic tele-chelic macromonomer 3a and the hydrophobictelechelic macromonomer 5 that oligomericimpurities (shoulder and/or little peaks) are pres-ent, which might also include cyclic oligomersapart from linear ones. However the interferingeffect of eventually existent cyclic oligomers on

Figure 4. 19F NMR spectrum of the F-terminated hydrophobic telechelic macromo-nomer 5. [Color figure can be viewed in the online issue, which is available atwww.interscience.wiley.com.]



the coupling reaction between the hydrophilicand the hydrophobic telechelic macromonomer isthought to be rather low because of the lowerreactivity of fluorine atoms in 2,20- and 3,30-posi-tion in cyclic oligomers compared with the 4,40-position in linear oligomers. The possible physicalincorporation of cyclic oligomers in the growingchain network during the coupling reactionwould influence the morphological structure butit might also contribute to an enhanced mechani-cal stability due to the entanglement of thesecyclic oligomers among each other and with thegrowing chain. Since the scope of this work is notthe precise investigation of the step-growth poly-merization by itself, which would require morepowerful analysis tools, such as matrix-assistedlaser desorption/ionization time-of-flight massspectroscopy (MALDI/TOF-MS),42,43 the follow-ing sections deal with the characterization ofmultiblock-co-ionomers.

The coupling products 1b–4b of the polycon-densation between the hydrophilic and hydro-phobic macromonomers were investigated bothby GPC analysis and NMR spectroscopy (1H, 19FNMR). The molecular weight distribution ofmultiblock-co-ionomer 3b is depicted in Figure 1together with the distributions of the corre-sponding macromonomers 3a and 5. The cou-pling reaction was obviously successful althoughthere is still some need for further optimizationin terms of the broad distribution. Table 2 liststhe GPC results of the coupling products.Assuming that the ideal composition is com-prised of alternating hydrophilic A and hydro-phobic B blocks with block lengths n(A) ¼ 4 andm(B) ¼ 8 the number of repeating AB units canbe estimated (p, Scheme 1). With the exceptionof ionomer 2b the corresponding values lie inthe range between 6 and 8. Further informationon the structural composition was gained from

Table 2. Overview of the Molecular Weight Distribution of the Multiblock-co-Ionomers 1b–4b, Determinedby Gel Permeation Chromatography (GPC)

Multiblock-co-Ionomer Mn

GPC [Da] PDIMn

calculated perAB-Block [Da]

Number of RepeatingAB-Blocks

1b 59,992 2.86 7,592 7.902b 9,924 2.46 7,720 1.293b 46,388 2.99 7,720 6.014b 61,281 4.19 8,065 7.60

Mncalculated per repeating AB-Block, calculated number average molecular weight of a repeating unit [AB-block with A ¼

hydrophilic block (assuming n ¼ 4) and B ¼ hydrophobic block (assuming m ¼ 8)].

Figure 5. 19F NMR spectrum of multiblock-co-ionomer 1b. [Color figure can beviewed in the online issue, which is available at www.interscience.wiley.com.]



the 19F NMR spectra of the products 1b–4b. Asexpected the 19F-NMR signals of the end groupatoms (4F–8F, Fig. 4) of the telechelic hydropho-bic macromonomer 5 disappear after reactionwith the telechelic hydrophilic macromonomers1a, 3a, and 4a, which is exemplified in Figure 5for 1b as representative for this group. In con-trast to all other ionomers 2b is an exception interms of a consistent and clearly interpretable19F NMR spectrum (Fig. 6). The signal at�63.52 ppm can be assigned to the fluorineatoms of the hexafluoroisopropylidene bridginggroup. To understand this spectrum more clearlythe effects of substitution in 4,40-position of theoctafluorobiphenylene moiety (Scheme 2) on thechemical shift were studied by comparison withliterature26,35 and by preparing the polymeric

model compound 8 from decafluorobiphenyl and4,40-sulfanediyl-bis-thiophenol. Scheme 3 revealsthat the chemical shifts of the fluorine nuclei 2Fand 3F are very similar for the both differentpoly(arylene ether)s 6 and 7.2,34 By substitutingthe ether bridge by a thioether bridge (8) adownfield shift of the corresponding signals isobserved. Using this information the two majorgroups of signals (�138.16, �153.59 ppm, and�132.24, �137.17 ppm) in Figure 6 can beattributed to be caused by fluorine atoms of thedifferently (O�� or S��) substituted octafluorobi-phenylene moieties. The origin of the signals at�141.50 and �160.51 ppm as well as the should-ers of the main peaks at �136.16 and �153.59ppm is a bit dubious. Although the position andintensity of four peaks (8F, 9F, 10F, and 11F) fit

Figure 6. 19F NMR spectrum of multiblock-co-ionomer 2b. [Color figure can beviewed in the online issue, which is available at www.interscience.wiley.com.]

Scheme 2. Definition of terms used in the text.



quite well to the end group fluorine atoms thereis a little doubt since the peak at �141.50 ppm(7F) would have been downfield-shifted for about8.5 ppm relative to the signal observed for deca-fluorobiphenyl and the macromonomer 5. A littledifference in the chemical shift in dependence ofthe solvent (CDCl3, DMSO-d6) was reported in

the literature for decafluorobiphenyl in the orderof magnitude of 1 ppm34 (Scheme 3). However,dipole–dipole or van der Waals interactions of thesolvent with the multiblock-co-ionomer especiallywith its end groups could lead to sterically preferredorientations and could thus cause the observed shiftin the resonance frequency of 7F in 1b.

By using this assignment of peaks (with theintensities given in Fig. 6) the realized AnBm

structure would differ from the ideal repeatingstructure (where A is the hydrophobic multi-block with n ¼ 4 and B the hydrophobic multi-block with m ¼ 8) that is expected by assuminga statistically controlled course of the poly-condensation reaction between the two mostencountered macromonomers (A with n ¼ 4 andB with m ¼ 8). Table 3 summarizes the theoreti-cal intensities for the first A4B8 multiblock struc-tures and compares them with the experimen-tally obtained values. The experimental integralsof 1F and 2F are considerably higher when com-pared with the intensities of 6F and 7F asexpected for an A4B8 multiblock structure. Itcan be concluded from Figure 6 and Table 3 thatthe ideal and most likely A4B8 composition isformed to a much lower extent than expected,which might have one or both of the followingreasons:

1. The reactivity of the shorter B (m < 8)macromonomers is higher than for m � 8

2. Trans(thio)etherification occurs between OH-or SH-terminated macromonomers with (thio)ether bridges in the growing chain: R��O��R1

þ R2��OHÐ R��O��R2 þ R1��OH.

Transetherification is usually used for thesynthesis of polyethers45–48 and may occur dur-ing step-growth polycondensations leading to

Scheme 3. Influence of the effects of substitution in4,40-position of the octafluorobiphenylene moiety onthe chemical shift in CDCl3 and in DMSO-d6 (inbrackets). Values are taken from literature25,34,44 (for6, 7, and 9) and measured for the model polymer 8.

Table 3. Theoretical Intensity in the 19F NMR Spectrum of 1b for the First A4B8 Block Structures

Signal At Caused ByExperimental

Intensity

Theoretical Intensitya

AB ABA BAB ABAB ABABA BABAB

�63.52 5F 6.00 6.00 6.00 6.00 6.00 6.00 6.00�132.24 2F 0.70 0.25 0.50 0.25 0.38 0.50 0.33�137.17 to�138.78

1F, 3F, 7F, 9F, 10F 5.68 4.50 4.50 4.50 4.50 4.50 4.50

�141.50 12F 0.23 0.13 0 0.13 0.06 0 0.08�153.59 4F, 6F, 8F 4.54 4.00 4.00 4.00 4.00 4.00 4.00�160.51 11F 0.45 0.25 0 0.25 0.13 0 0.17

a Calculated by assuming n ¼ 4 for the hydrophilic A-Block and m ¼ 8 for the hydrophobic B-Block.



the observed microstructure. However, a moredetailed study on the microstructure would re-quire more powerful characterization tools espe-cially (MALDI/TOF-MS)42,43 which is not an in-house technique.

Further evidence of deviations from the idealA4B8 structure could be obtained from the 1HNMR spectra. As shown for ionomer 1b in Fig-ure 7 and listed for the other ionomers 2b–4b inthe Experimental Section there are discrepan-cies between the calculated and the experimen-tal integral values. These discrepancies mightbe caused by the same reasons as discussedbefore for the 19F NMR spectrum of 1b.

Membrane Characterisation

TEM was used to stress the formation of multi-block-co-ionomers with a microphase-separated

morphology. Figure 8 represents the TEMimages of 3b in two different magnifications. At52,0003 magnification a more or less homogene-ous surface is visible (Fig. 8, left). The cavityprobably arises from a nonideal evaporation ofthe solvent during the membrane casting pro-cess. A more structured morphology is observedat the magnification of 145,0003 (Fig. 8, right),where the dark regions can be interpreted aslead sulfonate groups and the light domainsmight correspond to the hydrophobic multi-blocks of the multiblock-co-ionomer.

The series of these multiblock-co-ionomerswas fully characterized in terms of suitabilityfor application as proton exchange membrane infuel cells. Table 4 gives an overview of IEC, thespecific resistance and the thermal stability ofthe synthesised multiblock-co-ionomers in com-parison to the corresponding values of Nafion1

Figure 7. 1H NMR spectrum of multiblock-co-ionomer 1b. [Color figure can beviewed in the online issue, which is available at www.interscience.wiley.com.]

Figure 8. TEM images of membrane 3b with a magnification of 352,000 (left) and3145,000 (right).



117. It clearly can be seen from Table 4 that theexperimental IECdirect and IECtotal values arealmost identical among each other (which isexpected for a non-crosslinked membrane26) andthat they slightly differ with the calculated IECvalue (IECcalculated) for the proposed structure (n¼ 4 and m ¼ 8) which might be connected withthe reasons discussed in the previous section.The specific resistance of all multiblock-co-ionomer membranes is in the same order ofmagnitude with that of Nafion1 117. Ionomer4b shows a slightly lower calculated IEC valueand therefore a decreased number of proton con-ductive groups per weight unit so that a littlehigher proton resistance for this ionomer isexpected when compared with the other threeones which is indeed in accordance with thepresented data. The specific resistance of theremaining ionomers 1b–3b varies with the elec-tronic effects (-I and -M effects) of the bridging

groups as expected. The lowest specific resist-ance is measured for ionomer 3b, which has themost electron-withdrawing bridging groups (X ¼O, E ¼ SO2), followed by ionomer 1b with X ¼O and E ¼ S and ionomer 2b with X ¼ S andE¼ S. Thermal stability of these multiblock-co-ionomers is estimated from TGA/FTIR couplingby using the asymmetric S¼¼O stretching vibra-tion at 1352–1342 cm�1 in the FTIR spectrumas detection for the onset of the splitting-offtemperature of the sulfonic acid group. The TGAtraces of the multiblock-co-ionomers 1b–4b aredepicted in Figure 9 and compared with the cor-responding one for Nafion1 117. All synthesisedmultiblock-co-ionomers show the typical andexpected course for sulfonated arylene main-chain ionomers and exhibit two distinct thermaldegradation steps, of which the first weight lossstep at around 220 8C is mainly attributed tothe loss of sulfonic acid groups and the secondweight loss step at roughly 480 8C, which refersmainly to the decomposition of the main chain.In case of Nafion1 117 a seamless transitionbetween these two degradation steps takesplace. Because of the similarity of all TGAtraces and the detected onset of the splitting-offtemperature of the SO3H group in this series ofmultiblock-co-ionomers a reliable trend for thepreferred bridging groups E and X can not berecognized. It can be assumed that the influenceof the bridging groups on the stability of theSO3H group is rather low. To estimate the rela-tively thermal stability of the novel multiblock-co-ionomers in comparison to Nafion1 117 moreprecisely, isothermal TGA measurements werecarried out for Nafion1 117 and membrane 2bat three different temperatures (200, 220, and240 8C, cf. Fig. 10). The initially steep weightloss within the first hour of isothermal heatingcan be led back to the release of residualabsorbed water and is observed for all investi-

Table 4. Overview of the Ion-Exchange Capacities (IECs), the Specific Resistance and the Thermal Stabilityof the Multiblock-co-Ionomers

Multiblock-co-Ionomer

IECdirect

[mmol/g]IECtotal

[mmol/g]IECcalculated

[mmol/g]Specific Resistance

q [X cm]TSO3H onset

[8C]a

1b 1.13 1.20 1.05 4.02 226.02b 0.92 0.92 1.04 4.17 225.43b 1.07 1.07 1.04 3.56 218.54b 0.87 0.87 0.99 5.26 214.1Nafion1 117 0.90 0.90 – 6.59 228.7

a Obtained from TGA/FTIR coupling (first traces of SO2 in the IR spectrum were used for evaluation).

Figure 9. TGA traces of the multiblock-co-ionomersin comparison with Nafion1 117. [Color figure can beviewed in the online issue, which is available atwww.interscience.wiley.com.]



gated samples. There is a clear dependence ofthermal stability of the applied temperature,which is more pronounced in case of Nafion1

117. While Nafion1 117 is very stable at 200 8C,the gradient of the isotherms at 220 and 240 8Cbecomes more negative indicating an increasedthermal decomposition at these higher tempera-tures under the oxidising atmosphere. Thermaldegradation of Nafion1 117 has already beeninvestigated in air by Savinell et al. who pro-posed a decomposition mechanism commencingat the sulfonic acid end groups with the loss ofSO2 and HO radicals under the formation ofperfluoroalkyl radicals, which initiate furtherreactions.49 These authors reported a decomposi-tion temperature of 270 8C in air. The earlierdecomposition in our experiments might beattributed to the stronger oxidising conditions(65–70% O2) in the TGA measurements. In con-trast to Nafion1 117 the investigated multi-block-co-ionomer 2b does not show such a dis-tinct change in the gradient of the isothermsindicating a higher thermal stability of thesulfonic acid group attached at the aromaticphenylene rings when compared with thatof Nafion1. Figure 11 highlights the isothermsof the novel multiblock-co-ionomers 1b–4band Nafion1 117 and verifies the behaviour,observed and discussed before for 2b, for theentire series.

Mechanical properties of proton exchangemembranes are closely related to the WU andswelling ratio; the WU of 1b–4b in comparison

with Nafion1 117 is shown in Figure 12 and theswelling degree in Figure 13. Although thehydration behavior of all synthesised mem-branes is higher than that of Nafion1 117, a sys-tematic dependence on the type of bridgingatoms or groups can be seen. Both the WU andthe swelling ratio of the membrane with thethio ether bridges E ¼ S and X ¼ S (2b) are thehighest within this series. Lower WU and lowerswelling ratio can be realized by substitutingone of the thio ether bridges (E ¼ S) by an etherbridge (E ¼ O). Membrane 1b exhibits a clearlydecreased WU and swelling ratio compared with2b in spite of having the lower specific resist-ance. The substitution of the second thio etherbridging group (X ¼ S) in 1b by the sulfonebridging group (X ¼ SO2) in 3b seems to haveonly minor impact on the hydration behaviour.These observations might be explained by thedifferent electronic environments of E and X inthe hydrophilic multiblock of the multiblock-co-ionomer. The bridging group X is part of a sym-metrically 4,40-disubstituted bisphenol moietywhereas E is the linkage group between thebisphenol and the biphenyl moiety and is there-fore adjacent to chemically different neighbour-ing groups with distinct electronic effects. Incase of the linkage group E the strongly elec-tron-deficient sulfonic acid group in meta-posi-tion to E causes a higher polarity of bonds inthe nearer environment of this bridging groupin comparison to those in the neighbouring of X.Additionally the polarisability of C��S is higherthan that of C��O, which explains the higher

Figure 11. Isothermal TGA measurements ofNafion1 117 and of the novel multiblock-co-ionomers.[Color figure can be viewed in the online issue, whichis available at www.interscience.wiley.com.]

Figure 10. Comparison of thermal stability ofNafion1 117 with that of multiblock-co-ionomer 2b atdifferent temperatures. [Color figure can be viewed inthe online issue, which is available at www.interscience.wiley.com.]



susceptibility of 2b to water in comparison to 1band 3b. As shown by means of ionomer 4b theincorporation of the highly hydrophobic bridginggroup X ¼ C(CF3)2 can further improve the WUand swelling degree.

To be able to estimate the oxidative stabilityof these novel membranes in comparison toNafion1 117 the weight loss in hydrogen perox-ide solution (5 wt %) at 60 8C in dependence oftime was determined. The results are graphedin Figure 14. Although there is still a need forimproving the oxidative stability of the synthe-sised multiblock-co-ionomers in general (e.g.,

namely by increasing their molecular weightand decreasing their polydispersity), a cleartendency of oxidative stability in dependence ofthe bridging atoms or groups E and X can beidentified. The multiblock-co-ionomer membrane2b was completely dissolved after 24 h whereasall other membranes were still integrative afterthat time of exposition to hydrogen peroxide so-lution. As discussed before for the hydrationbehaviour, the polarisability of the C��S bondcould also be a reason for the worse oxidativestability of 2b. The weight loss of 1b and 3bincreases nearly linearly with increasing time,whereas ionomer 4b shows no significant degra-

Figure 14. Weight loss in hydrogen peroxide solu-tion of the novel multiblock-co-ionomers in compari-son to Nafion1 117. [Color figure can be viewed in theonline issue, which is available at www.interscience.wiley.com.]

Table 5. Overview of the Reaction Conditions forthe Preparation of the Different Multiblock-co-Ionomers

Block-co-Iononer

ReactionTemperature

[8C]ReactionTime [h]

Degree ofDilutiona

1b 80 18 0110 3120 2

2b 80 3 1.673b 80 21 0.944b 80 4 3.20

a The degree of dilution is defined as the ratio of the vol-ume of NMP added during reaction to the starting volume ofNMP.

Figure 12. Water uptake behavior of the multi-block-co-ionomer membranes 1b–4b. [Color figure canbe viewed in the online issue, which is available atwww.interscience.wiley.com.]

Figure 13. Swelling ratio of the multiblock-co-ionomer membranes 1b–4b. [Color figure can beviewed in the online issue, which is available at www.interscience.wiley.com.]



dation after 6 h. After that time the weight lossof 4b increases from nearly 0 to 30% andremains at this value up to the end of the testafter 48 h. This simple test can admittedly notbe used for an accurate study of degradation,but it can serve as a preselection for the mostpromising membranes in terms of oxidative sta-bility. Using the results of this test a series witha rough estimation of oxidative stability of theprepared membranes can be given as follows: 4b� 3b > 1b > 2b. Finally it can be stated thatthe membranes 1b, 3b, and 4b are promisingcandidate for the use in a PEMFC or DMFC.

CONCLUSIONS

In this contribution the influence of the chemi-cal nature of bridging atoms or groups in thehydrophilic segments of multiblock-co-ionomershas been comparatively investigated in termsof their properties as fuel cell membranes. Itturned out that not only the chemical nature ofthe bridging atoms or groups is crucial for theproperties of the membranes, but also their posi-tion and electronic environment. The thio etherlinkage group (E in Scheme 2) between twoasymmetrically substituted arylene rings seemsto be a weak point in the polymer architecture.Ionomers with symmetrically substituted thioether linkages on the other hand (X in Scheme2) entail similar properties as corresponding ion-omers with sulfone linkages although their oxi-dative stability seems to be slightly lower. Butthe alleged infeasibility of the synthesised ion-omer with the highest possible set of thio etherlinkages might be compensated by the selectiveoxidation of the thio ether bridges to sulfonebridges provided that the crystallinity of theionomer does not change dramatically andcauses brittleness. The principal possibility ofselective oxidation of this kind of bridges hasalready been discussed in literature50,51 andpatents.44,52–54 Further membrane developmentin our group will also be done by the variationof multiblock length of the hydrophilic andhydrophobic multiblock for the most promisingsystems.

The authors would like to thank the LandesstiftungBaden–Wurttemberg for financial support of thisresearch effort in the project ‘‘Direktmethanol-Mikro-brennstoffzellenstack mit neuen MEA.’’

REFERENCES AND NOTES

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

2. Schonberger, F.; Kerres, J. In 4th InternationalSymposium on High-Tech Polymer Materials, Bei-jing, China, May 14–20, 2006, p 35.

3. Li, Q.; He, R.; Jensen, J. O.; Bjerrum, N. J ChemMater 2003, 15, 4896.

4. Li, Y.; Wang, F.; Yang, J.; Liu, D.; Roy, A.; Case, S.;Lesko, J.; McGrath, J. E. Polymer 2006, 47, 4210.

5. Miyatake, K.; Wanatabe, M. J Mater Chem 2006,16, 4465.

6. Kerres, J. A. J Membr Sci 2001, 185, 3.7. Kreuer, K. D. J Membr Sci 2001, 185, 29.8. Lee, J. S.; Quan, N. D.; Hwang, J. M.; Lee, S. D.;

Kim, H.; Lee, H.; Kim, H. S. J. Ind Eng Chem2006, 12, 175.

9. Rikukawa, M.; Sanui, K. Prog Polym Sci 2000,25, 1463.

10. Miyatake, K.; Hay, A. S. J Polym Sci Part A:Polym Chem 2001, 39, 3770.

11. Lafitte, B.; Jannasch, P. J Polym Sci Part A:Polym Chem 2005, 43, 273.

12. Jakoby, K.; Peinemann, K. V.; Nunes, S. P. Macro-mol Chem Phys 2003, 204, 61.

13. Parcero, E.; Herrera, R.; Nunes, S. P. J MembrSci 2006, 285(1/2), 206.

14. Kim, Y. S.; Sumner, M. J.; Harrison, W. L.; Riffle,J. S.; McGrath, J. E.; Pivovar, B. S. J Electro-chem Soc 2004, 151, A2150.

15. Harrison, W. L.; Wang, F.; Mecham, J. B.; Bhanu,V. A.; Hill, M.; Kim, Y. S.; McGrath, J. E. J PolymSci Part A: Polym Chem 2003, 41, 2264.

16. Gao, Y.; Robertson, G. P.; Guiver, M. D.; Mikhai-lenko, S. D.; Li, X.; Kaliaguine, S. Macromole-cules 2005, 38, 3237.

17. Liu, B.; Robertson, G. P.; Guiver, M. D.; Sun, Y.-M.; Liu, Y.-L; Lai, J.-Y.; Mikhailenko, S.; Kalia-guine, S. J Polym Sci Part B: Polym Phys 2006,44, 2299.

18. Norsten, T. L.; Guiver, M. D.; Murphy, J.; Astill,T.; Navessin, T.; Holdcroft, S.; Frankamp, B. L.;Rotello, V. M.; Ding, J. Adv Funct Mater 2006,16, 1814.

19. Ding, J.; Chuy, C.; Holdcroft, S. Adv Funct Mater2002, 12, 389.

20. Shin, C. K.; Maier, G.; Andreaus, B.; Scherer, G.G. J Membr Sci 2004, 245, 147.

21. Taeger, A.; Volen, C.; Lehmann, D.; Lenk, W.;Schlendstedt, K.; Meier-Haack, J. MacromolSymp 2004, 210, 175.

22. Asano, N.; Miyatake, K.; Watanabe, M. J PolymSci Part A: Polym Chem 2006, 44, 2744.

23. Ghassemi, H.; Harrison, W.; Zawodzinski, T. A.;McGrath, J. E. Polym Prepr 2004, 45, 68.

24. Roy, A.; Hickner, M. A.; Yu, X.; Li, Y.; Glass, T.E.; McGrath, J. E. J Polym Sci Part B: PolymPhys 2006, 44, 2226.



25. Ghassemi, H.; McGrath, J. E.; Zawodzinski T. A.,Jr. Polymer 2006, 47, 4132.

26. Schonberger, F.; Hein, M.; Kerres, J. Solid StateIonics 2007, 178, 547.

27. Kim, Y. S.; Einsla, B.; Sankir, M.; Harrison, W.;Pivovar, B. S. Polymer 2006, 47, 4026.

28. Xing, D.; Kerres, J. Polym Adv Technol 2006,17(7/8), 591.

29. Kunz, M. J.; Hayn, G.; Saf, R.; Binder, W. H. JPolym Sci Part A: Polym Chem 2004, 42, 661.

30. Jurek, M. J.; McGrath, J. E. Polymer 1989, 30,1552.

31. Kerres, J.; Ullrich, A.; Haring, T.; Preidel, W.;Baldauf, M.; Gebhardt, U. J New Mater Electro-chem Syst 2000, 3, 229.

32. Harrison, W. L.; Wang, F.; Mecham, J. B.; Bhanu,V. A.; Hill, M.; Kim, Y. S.; McGrath, J. E. J PolymSci Part A: Polym Chem 2003, 41, 2264.

33. Shin, C. K.; Maier, G.; Scherer, G. G. J MembrSci 2004, 245, 163.

34. Kerres, J. A.; Xing, D.; Schonberger, F. J PolymSci Part B: Polym Phys 2006, 44, 2311.

35. Fircks, G. V.; Hausmann, H.; Francke, V.; Gunter,H. J Org Chem 1997, 62, 5074.

36. Mori, S. Anal Chem 1981, 53, 1817.37. Weiss, A. R.; Cohn-Ginsberg, E. J Polym Sci Part

A-2: Polym Phys 1970, 8, 148.38. Mahabadi, H. K.; O’Driscoll, K. F. J Appl Polym

Sci 1977, 21, 1283.39. Mahabadi, H. K. J Appl Polym Sci 1985, 30, 1535.

40. Kricheldorf, H. R. Macromol Symp 2003, 199, 1.41. Kricheldorf, H. R. Macromol Rapid Commun

2003, 24, 359.42. Yang, J.; Nonidez, W. K.; Mays, J. W. Int J Polym

Anal Char 2001, 6, 547.43. Wilczek-Vera, G.; Danis, P. O.; Eisenberg, A. Mac-

romolecules 1996, 29, 4036.44. Zierer, D.; Scheckenbach, H.; Dierolf, M. EP

1,044,999 A2, 2000.45. Maes, C.; Devaux, J.; Legras, R.; Parsons, I. W.;

McGrail, P. T. J Polym Sci Part A: Polym Chem1994, 32, 3171.

46. Rao, M. R.; Rao, V. L. J Appl Polym Sci 1999, 74,3425.

47. Jayakannan, M.; Ramakrishnan, S. MacromolChem Phys 2000, 201, 759.

48. Behera, G. C.; Ramakrishnan, S. J Polym SciPart A: Polym Chem 2004, 42, 102.

49. Samms, S. R.; Wasmus, S.; Savinell, R. F. J Elec-trochem Soc 1996, 143, 1498.

50. Li, Z.; Ding, J.; Robertson, G. P.; Guiver, M. D.Macromolecules 2006, 39, 6990.

51. Schuster, M.; Kreuer, K.-D.; Anderson, H. T.;Maier, J. Macromolecules 2007, 40, 598.

52. Zierer, D.; Bruck, M.; Scheckenbach, H. EP1,045,064 A1, 2000.

53. Kratz, T.; Kagermeier, S.; Zeiss, W. WO 99/61513,1999.

54. Schuster, M.; Kreuer, K.-D.; Thalbitzer, A.; Maier,J. WO 2006/094767 A1, 2006.



novel multiblock-co-ionomers as potential polymer ... ke… · novel multiblock-co-ionomers as...

Documents