Journal ofMaterials Chemistry A

PAPER

Publ

ishe

d on

26

Sept

embe

r 20

16. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 0

6/10

/201

6 14

:05:

26.

View Article OnlineView Journal

Anion conductin

aSchool of Materials Science and Engineering

GA 30332-0100, USAbSchool of Chemical and Biomolecular Engi

Atlanta, GA 30332-0100, USA. E-mail: kohl@cU.S. Army Research Laboratory, Adelphi, M

Cite this: DOI: 10.1039/c6ta06653d

Received 4th August 2016Accepted 23rd September 2016

DOI: 10.1039/c6ta06653d

www.rsc.org/MaterialsA

This journal is © The Royal Society of

g multiblock copolymermembranes with partial fluorination and longhead-group tethers

Lisha Liu,a John Ahlfield,b Andrew Tricker,b Deryn Chuc and Paul A. Kohl*b

Anion conductive polymers are of value in anion exchange membrane (AEM) fuel cells. A series anion

conductive polymers composed of partially fluorinated multiblock copoly(arylene ether)s (mPEs) with

long head-group tethers were synthesized via polycondensation and Friedel–Crafts reaction. The

relationship between chemical structure, membrane morphology and physical properties was explored

by varying the length of the hydrophilic and hydrophobic blocks and the number of tethers per

hydrophilic repeat unit in the synthesized multiblock copolymers. Efficient, ion-conductive nano-

channels were formed by using nanophase-separation of the multiblock copolymer to improve channel

formation and ionic conductivity without inducing excess water uptake (WU). The hydrophobicity of the

partially fluorinated backbone further reduced the WU. Doubling the number of head-groups resulted in

more than doubling the hydroxide ion conductivity. From the study of the number of freezable and

bound water molecules per head-group, it was found that bound water played a dominant role in ion

transport, while excess unbound water led to higher unproductive WU. Multiblock copolymer AEMs with

high hydroxide conductivity, up to 119 mS cm�1 were obtained. Polymers with an attractive OH�

conductivity of 94 mS cm�1 at 80 �C, and low WU of 26.7% with a modest ion exchange capacity were

also obtained. The membranes showed excellent alkaline stability due to the use of the long head-group

tether structure and partial fluorination. Less than 1.5% conductivity loss was observed after soaking the

membrane in 1 M NaOH solution at 60 �C for over 1000 h. Membranes with higher ionic conductivity

showed lower oxygen diffusivity and permeability.

Introduction

Fuel cells are electrochemical devices that directly convertchemical energy stored in fuels, such as hydrogen and meth-anol, to electrical energy. Fuel cells have received attention inrecent years owing to their high efficiencies and low emissionscompared to heat engines.1–3 Proton exchangemembrane (PEM)fuel cells have been developed and used in portable electronicdevices and vehicle transportation.2 However, the wide-scalecommercialization of proton conducting membrane fuel cells isimpeded by several factors including sluggish oxygen reductionkinetics at the cathode, high cost of the noble metal platinumcatalyst, complex water management, and fuel crossover fromthe anode to the cathode through the PEM.3 Recently, anionexchange membrane (AEM) fuel cells have attracted attention,which use AEMs as the electrolyte to conduct OH� ions. Thehigh pH environment of the AEM fuel cell mitigates many of the

, Georgia Institute of Technology, Atlanta,

neering, Georgia Institute of Technology,

gatech.edu

aryland 20783, USA

Chemistry 2016

shortfalls encountered with PEM fuel cells including facileoxygen reduction kinetics, potential use of non-precious metalcatalysts, and lower fuel crossover because of the oppositedirection of the ions within the membrane.3–6 However, theperformance of AEM fuel cells is not as good as PEM fuel cellsmainly due to the limitations of current AEM materials,including low ionic conductivity, high water uptake and poorstability of the membrane at high pH.5,7–9

AEMs with ionic conductivity greater than 100 mS cm�1 at80 �C and low water content are desired for efficient fuel celloperation.3,5,10 The ionic conductivity of the membrane isdetermined by two factors: the mobility of hydroxide ions andthe ion exchange capacity (IEC).11 Increasing the IEC byincreasing the number of xed cation head-groups is possible;however, it oen leads to inefficient ion conducting channelswith high water uptake (WU) and low anionmobility. Improvingionic mobility is the preferred route and can be achievedthrough construction of more efficient ion conducting channelswithin the polymer membrane. Multiblock copolymers havebeen reported to have higher ionic conductivity and lower wateruptake than their corresponding random copolymers with thesame IEC.12 The improved ionic conductivity of block copoly-mers is attributed to the formation of efficient, well-developed

J. Mater. Chem. A

Journal of Materials Chemistry A Paper

Publ

ishe

d on

26

Sept

embe

r 20

16. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 0

6/10

/201

6 14

:05:

26.

View Article Online

nano-channels created by phase-segregation.10,13–15 The ionicgroups are locally concentrated in the hydrophilic domainsforming ion-conductive channels. The ionic conductivity andWU of the membranes are greatly affected by the channel size,which is determined by the block length, type of xed cations,and cation density.

Chemical stability at high pH is another obstacle to AEMdevelopment and commercialization.3,5 The stability of AEMshas been studied for different cations and chemical backbones.Benzyl-trimethyl ammonium (BTMA) is the most commoncation group. The degradation routes for BTMA at high pHinclude: (1) b-hydrogen Hofmann elimination,16 (2) directnucleophilic substitution at an a-carbon,17 and (3) eliminationvia ylide formation.18 Other cation groups, such as imidazo-lium,19,20 phosphonium,21 and guanidinium22 have also beenstudied as alternatives to BTMA. A promising structureemerging from recent studies is forming the xed cation at theend of long pendent chains attached to the polymer backbonein a comb-shaped structure23 in place of the vulnerableBTMA groups. Hibbs24 reported that membranes withtrimethylammonium (TMA) cations attached to the backbonewith a hexamethylene spacer showed better stability than themembrane with BTMA cations in 4 M KOH at 90 �C. Mohantyand co-workers25 compared various cations on small moleculesand showed that long alkyl chain tethered quaternary ammo-nium groups had the best alkaline stability. Dang andco-workers26 used tailored side-chains for cation attachment toproduce AEMs with high ionic conductivity and alkalinestability. They showed that a ve-carbon tether performed thebest by comparing different cationic alkyl side chain designs.

Backbone stability at high pH is another important issue.This includes stability of the backbone itself as well as back-bone/head-group interactions. A wide range of polymer back-bone structures including polyphenylenes,24 poly(phenyleneoxide)s,26,27 poly(arylene ether sulfone)s,12,14 and poly(aryleneether)s28 have been investigated. Degradation of AEMs con-taining aryl ether (Csp2–O) bonds under alkaline conditions wasobserved.29,30 Lee et al.31 reported AEMs with good stability andno conductivity loss at 80 �C for a hydrocarbon backbonewithout aryl ether linkages.

In order to combine the advantages of phase separation witha block copolymer structure and improved stability introducedby alkyl spacer linkages between the cation head-group andbackbone, multiblock copoly(arylene ether)s (mPEs) with longhead-group tethers were synthesized in this study. Differenthydrophobic and hydrophilic block lengths containing one ortwo tethers per hydrophilic repeat unit were used. In addition,the partially uorinated backbone structure was designed tofurther reduce water uptake for themembrane by increasing theoverall hydrophobicity. To our knowledge, this is the rst timethe long alkyl spacer structure was combined with multiblockcopolymer backbone in AEM materials. The size of the ionconductive nano-channels, ionic conductivity, water uptake,and the number of freezable water and bound water moleculesper head group were evaluated. The comparison is intendedto help understand the relationship between structure,morphology and physical properties. The properties ofmultiblock

J. Mater. Chem. A

copolymer membranes and pure hydrophilic oligomer were alsocompared. In addition, the oxygen solubility, diffusivity andpermeability of the membranes were analyzed.

Experimental sectionMaterials

N,N0-Dimethylacetamide (DMAc) was obtained from AlfaAesar and dried by vacuum distillation at 130 �C over CaH2.Decauorobiphenyl (DFBP), 4,40-(hexauoroisopropylidene)diphe-nol (HFBPA), 6-bromohexanoyl chloride, AlCl3, 1,2-dichloroethane,triethylsilane, and triuoroacetic acid were also obtained from AlfaAesar and used as-received. 4,40-(9-Fluorenylidene)diphenol (BPFL)was obtained from TCI Co. Ltd and used as-received. Potassiumcarbonate (K2CO3) and dichloromethane (DCM) were purchasedfrom BDH chemicals and used as-received.

Synthesis of hydrophobic and hydrophilic oligomers

The synthetic procedure for the hydrophobic oligomer with x ¼3.1 repeat units is as follows: DFBP (1.67 g, 5 mmol) and HFBPA(1.68 g, 5 mmol) were dissolved in 20 mL DMAc in a 100 mLthree-neck round bottom ask with a condenser under N2

atmosphere at room temperature. K2CO3 (1.66 g, 12 mmol) wasadded to the solution and then the mixture was heated to 80 �Cfor 1 h. HFBPA (0.5 g) was added to the mixture and allowed toreact at 40 �C for 4 h to ensure that all the oligomers werehydroxyl-terminated. The viscous solution was poured intodeionized water to precipitate the product. The white solid wasisolated by ltration, washed three times with deionized waterand dried overnight at 80 �C in a vacuum oven. Hydrophobicoligomers with a different number of repeat units weresynthesized by controlling the reaction time.

Fluoro-terminated hydrophilic oligomer with y ¼ 3.6 repeatunits was synthesized via a similar procedure. BPFL (1.75 g, 5mmol) and DFBP (1.67 g, 5 mmol) were dissolved in 20 mLDMAc under N2 atmosphere. K2CO3 (1.66 g, 12 mmol) wasadded to the solution and the mixture was reacted at 80 �C for 1h. DFBP (0.6 g) was added to themixture and reacted at 40 �C for4 h. Aer precipitation, ltration and drying, a white solid wasyielded. Similarly, hydrophilic oligomers with different amountof repeat units were obtained by varying reaction time.

Synthesis of multiblock copoly(arylene ether)s (mPEs)

The hydrophobic oligomer (1.15 g, 0.5 mmol, x ¼ 3.1) andhydrophilic oligomer (1.33 g, 0.5 mmol, y ¼ 3.6) were dissolvedin 20 mL DMAc at room temperature under N2. K2CO3 (0.15 g,1.1 mmol) was added to the solution and the mixture wasreacted at 80 �C for 4 h. The viscous solution was poured intodeionized water and white solid was precipitated out. Theproduct was ltered, washed with deionized water and driedovernight at 80 �C in a vacuum oven.

Synthesis of BrKC6-mPEs

The synthesis method of BrKC6-mPEs was modied fromliterature.24 Multiblock copolymer mPE-X3.1Y3.6 (1.00 g, 0.74mmol hydrophilic repeat units) was dissolved in 20 mL DCM in

This journal is © The Royal Society of Chemistry 2016

Paper Journal of Materials Chemistry A

Publ

ishe

d on

26

Sept

embe

r 20

16. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 0

6/10

/201

6 14

:05:

26.

View Article Online

a ask under nitrogen atmosphere. The ask was chilled in anice bath. 6-Bromohexanoyl chloride (0.45 mL, 2.96 mmol) andAlCl3 (0.40 g, 2.96 mmol) were added to the solution andwarmed to room temperature. The mixture was reacted for 5 hat room temperature. The solution was poured into 200 mLdeionized water and heated to 60 �C to evaporate the solvent.The product was dissolved in DCM and precipitated in deion-ized water three times to obtain a puried light yellow solid.

Synthesis of BrC6-mPEs

BrKC6-mPE-X3.1Y3.6 (1.14 g, 0.74 mmol hydrophilic repeatunits) was dissolved in 20 mL 1,2-dichloroethane in a three-neck round bottom ask with a condenser at room temperatureunder nitrogen. Triuoroacetic acid (3.1 mL, 40.7 mmol) andtriethylsilane (0.59 mL, 3.7 mmol) were added to the solution.The solution was heated to reux for 24 h. Then it was cooled toroom temperature and poured into a NaOH solution (1.63 gNaOH, 200 mL deionized water). The mixture was heated to 80�C to evaporate the solvent. The product was isolated by ltra-tion. The product was dissolved in DCE and precipitated indeionized water three times producing a white solid product.

Membrane casting, quaternization and ion exchange

BrC6-mPEs (0.20 g) was dissolved in 5 mL 1,2-dichloroethaneand the resulting solution was ltered through a 0.45 mmpoly(tetrauoroethylene) (PTFE) membrane syringe lter intoa 4 cm diameter aluminum dish. The solvent was evaporated ina tube furnace at 40 �C under nitrogen for 18 h. The free-standing membrane was about 40 mm thick.

The membrane was quaternized as follows: immersion in45 wt% trimethylamine aqueous solution at room temperaturefor 48 h. Aer being removed from the solution, the quaternizedmembrane with a bromide counter ion was washed withdeionized water three times. The membrane was soaked in 1 MKOH solution under nitrogen for 24 h to exchange the bromideions for hydroxide ions. Aer being washed with deionizedwater three times, the membrane was stored in deionized water.

Nuclear magnetic resonance (NMR) spectra and gelpermeation chromatography (GPC)

The chemical structure of the oligomers and polymers wereanalyzed by a variety of NMR techniques, including one-dimensional 1H NMR and 19F NMR obtained with a VarianMercury Vx 400 MHz spectrometer, two-dimensional hetero-nuclear single quantum coherence (HSQC) and one-dimen-sional nuclear Overhauser effect (NOE) obtained with a Bruker500 MHz spectrometer. HSQC and NOE NMR spectra were usedto identify the position of the tether attachment on the polymerbackbone. Chloroform-d was used as the solvent for non-ionicsamples and N,N-dimethylformamide-d7 was used for polarsamples. Quantitative 19F NMR spectra were collected at376.273 MHz with a 12.5 s relaxation delay. The HSQC analysisemployed 8 scans, 256 increments along t1, 1024 data pointsalong t2, and 145 Hz as one-bond coupling constant. The one-dimensional NOE spectra used the DPFGSE-NOE pulse

This journal is © The Royal Society of Chemistry 2016

program, a mixing time of 0.2 s, a relaxation delay of 2 s, and anacquisition time of 3.28 s.

The molecular weight of the polymers was determined by gelpermeation chromatography (GPC) (Shimadzu) equipped withan LC-20 ADHPLC pump and a refractive index detector (RID-10A, 120 V). Tetrahydrofuran (THF) was used as the eluent witha ow rate of 1.0 mL min�1 at 35 �C. The molecular weight wasmeasured by a calibration curve based on polystyrenestandards.

Ionic conductivity

The ionic resistance of the membranes was measured witha four-probe, in-plane electrochemical impedance spectrometerover the frequency range from 1 Hz to 1 MHz with a PAR 2273potentiostat. All samples were tested in HPLC-grade waterunder nitrogen to minimize the effect of CO2. The samples wereequilibrated for 30 min prior to measurement. The in-planeionic conductivity was calculated using eqn (1).

s ¼ L

WTR(1)

In eqn (1), s is the ionic conductivity in S cm�1, L is thelength between sensing electrodes in cm,W and T are the widthand thickness of the membrane in cm, respectively, and R is theresistance measured in ohms.

Ion exchange capacity (IEC), water uptake (WU), hydrationnumber (l), number of freezable water (Nfree) and bound, non-freezable water (Nbound) molecules

The ion exchange capacity was calculated using NMR datawhich is discussed in detail in the next section. The wateruptake of the membranes was calculated by eqn (2).

WUð%Þ ¼ Mw �Md

Md

� 100 (2)

In eqn (2), Md is the dry mass of the membranes measuredaer being dried in vacuum for 24 h and Mw is the wet mass ofthe membranes without surface water. Both dry and wetmembranes were in the OH� form and measured at roomtemperature. The number of water molecules per ionic group,hydration number l, was calculated using eqn (3).

l ¼ 1000 �WU%

IEC � 18(3)

The number of freezable water (Nfree) and bound water (ornon-freezable water) (Nbound) were determined by differentialscanning calorimetry (DSC). DSC measurements were carriedout on a DSC Q200 (TA Instruments). The membrane sampleswere fully-hydrated by soaking in deionized water at least forone week. Aer the water on the membrane surface was wipedoff, a 3 to 5 mg sample was quickly sealed in an aluminum pan.The sample was cooled to �50 �C and heated at a rate of 5 �Cmin�1 under N2 (20 mL min�1). The quantity of freezable andnon-freezable water was determined by eqn (4)–(6).32–34

J. Mater. Chem. A

Journal of Materials Chemistry A Paper

Publ

ishe

d on

26

Sept

embe

r 20

16. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 0

6/10

/201

6 14

:05:

26.

View Article Online

Nfree ¼ Mfree

Mtot

� l (4)

Mfree is the mass of freezable water and Mtot is the total mass ofwater absorbed in the membrane. The weight fraction offreezable water was calculated by eqn (5).

Mfree

Mtot

¼ Hf

�Hice

ðMw �MdÞ=Mw

(5)

Hf is enthalpy obtained by the integration of the DSC freezingpeak and Hice is enthalpy of fusion for water, corrected for thesubzero freezing point according to eqn (6).

Hice ¼ Hoice � DCpDTf (6)

DCp is the difference between the specic heat capacity of liquidwater and ice. DTf is the freezing point depression.

Morphological characterization

The surface morphology of the membranes was analyzed bytapping mode atomic force microscopy (AFM) with an AsylumResearch instrument. The probe (AC240TM-R3, Olympus) witha cantilever spring constant of 1.5 N m�1 was used to takeimages of the samples at ambient temperature and 50% relativehumidity. The Amp InvOLS was 109 nm V�1. The scanningfrequency was 1 Hz. The measurements were conducted underthe same conditions for each sample to keep consistency.

Thermal stability and mechanical properties

Thermal stability of the hydroxide form of the membranes wasanalyzed by thermogravimetric analysis (TGA) on a TA Instru-ments Q50 analyzer. The thermal degradation was evaluated ata heating rate of 5 �C min�1 up to 800 �C in nitrogen.

The stress–strain relationship was investigated by dynamicmechanical analysis (DMA) using TA Instruments Q800 withcontrolled force mode. Rectangular membrane samples werefully hydrated and tested using tension clamps aer removingsurface water at 100% relative humidity. The experimentalparameters were set as: preload force 0.05 N, isothermal at30 �C, soak time 1 min, force ramp rate 0.5 N min�1 and upperforce limit up to 18 N.19

Alkaline stability

The alkaline stability of the membranes was evaluated bysoaking the OH� form membranes into 1 M NaOH in a Teonlined Parr reactor at 60 �C for up to 1000 h to measure thechanges in ionic conductivity. Before measurement, eachmembrane was thoroughly washed with deionized water. Theionic conductivity was determined in HPLC-grade water at roomtemperature.

Oxygen solubility, diffusivity and permeability

The oxygen transport properties, including solubility, diffusivityand permeability, were measured using an intelligent gravi-metric analyzer by Hiden Analytical Ltd. (IGA-1). A membranesample was placed in the chamber which was then evacuated

J. Mater. Chem. A

for 24 h. Oxygen gas was introduced into the chamber in stepincrements of 250 mbar from vacuum to 1 bar. At each step, themass was recorded until equilibrium was reached at eachpressure to give the amount of oxygen absorbed by the sampleat a given pressure. The oxygen solubility S (mass% per bar) ofthe membrane was calculated by the weight percent increasefrom vacuum to 1 bar oxygen pressure. The mass vs. time datafor the step from vacuum to 250 mbar was used to estimate theoxygen diffusion coefficient, or diffusivity D (cm2 s�1), in themembrane according to eqn (7).

Mt

MN

¼ 2

l

�D

p

�12

t12 (7)

Mt is the mass of the membrane and absorbed oxygen at time t,MN is the mass at equilibrium and l is half of the thickness ofthe membrane in cm. The oxygen permeability, P, in mol (barcm s)�1 was calculated by eqn (8).

P ¼ rS

Moxygen

D (8)

r is the density of the membrane (g cm�3), which was measuredby pycnometer to be 1.1 g cm�3, and Moxygen is the molecularweight of oxygen.

Results and discussionSynthesis and characterization of hydrophobic andhydrophilic oligomers and multiblock copolymers

A series of mPEs multiblock copolymers were synthesized viaa polycondensation reaction of individually prepared hydroxyl-terminated hydrophobic oligomers and uorine-terminatedhydrophilic oligomers, as shown in Scheme 1. The number ofrepeat units of each oligomer was controlled by the reactiontime. The reaction temperature was held below 80 �C in orderto avoid cross-linking with uorine atoms on the side ofDFBP.35 End-capping of oligomers with excess DFBP for thehydrophilic oligomers and excess HFBPA for the hydrophobicoligomers was performed at 40 �C in order to avoid the cross-linking reaction. The number of repeat units (x) of thehydroxyl-terminated hydrophobic oligomers was calculatedaccording to the ratio of the integration of peaks a and a0 in the1H NMR spectra shown in Fig. 1a. Similarly, the length of F-terminated hydrophilic oligomers (y) was determined by theratio of peaks 2 and 3 shown in Fig. 1b. The F-terminatedoligomers were converted into the ion conducting hydrophilicblocks aer attachment of the long tether and ionic head-groups. The number-average molecular weight (Mn) andpolydispersity (PDI) of mPEs were obtained by GPC, as shownin Table 1. The target molecular weight of mPEs was >50kg mol�1 in order to obtain stable, free-standing, solvent-castmembranes.

The long tethers connecting the ionic head-groups to thepolymer backbone were attached to mPEs backbones viaa Friedel–Cras acylation reaction with 6-bromohexanoylchloride and AlCl3. The successful attachment of the tetherswas conrmed by the 1H NMR spectrum of the acylation

This journal is © The Royal Society of Chemistry 2016

Scheme 1 Synthesis of multiblock copoly(arylene ether)s with long alkyl tethers for ionic head-groups.

Paper Journal of Materials Chemistry A

Publ

ishe

d on

26

Sept

embe

r 20

16. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 0

6/10

/201

6 14

:05:

26.

View Article Online

product BrKC6-mPEs, as shown in Fig. 2b. A new peak i at2.95 ppm appeared aer acylation compared to the 1H NMRspectrum of mPEs backbone structure, as shown in Fig. 2a. Thenew peak is attributed to methylene protons adjacent to benzyl

This journal is © The Royal Society of Chemistry 2016

carbonyl group. The appearance of the new peak at 7.95 ppm,the shi of peak h from 7.77 ppm to 7.81 ppm, and the disap-pearance of peak f at 7.28 ppm also conrmed the successfulacylation reaction.

J. Mater. Chem. A

Fig. 1 (a) 1H NMR spectrum of OH-terminated hydrophobic oligomer,and (b) 19F NMR spectrum of F-terminated hydrophilic oligomer.

Table 1 Structural characteristics of mPEs multiblock copolymers

Multiblock copolymera xb yc Nd Mne PDIf

H-1 — 8 — 18 K 1.81X3.1Y3.6 3.1 3.6 18.0 88.6 K 2.27X5.4Y7 5.4 7 8.0 68.2 K 1.97X3.1Y8 3.1 8 7.2 55.9 K 1.83X5.9Y5 5.9 5 7.8 59.0 K 2.22

a H-1: pure hydrophilic oligomer; X: hydrophobic block; Y: hydrophilicblock. b x: number of repeat units in hydrophobic blocks. c y: number ofrepeat units in hydrophilic blocks. d N: number of repeat units in blockcopolymer. e Mn: number-average molecular weight from GPC. f PDI:polydispersity index.

Fig. 2 1H NMR spectra of (a) multiblock copolymer mPEs, (b) BrKC6-mPEs, and (c) BrC6-mPEs.

Journal of Materials Chemistry A Paper

Publ

ishe

d on

26

Sept

embe

r 20

16. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 0

6/10

/201

6 14

:05:

26.

View Article Online

Since there was no change in the peak positions and peakareas for the aryl protons a, b, c, and d in Fig. 2a and b, thesesites are not possible locations for attachment of the tether. Dueto the strong electron-withdrawing effect of the triuoromethylgroups, the aryl protons a and b adjacent to them are unfavor-able sites for the Friedel–Cras reaction. This was shownseparately by the unsuccessful Friedel–Cras acylation of thepure hydrophobic oligomers. Aryl protons c and d on thehydrophilic blocks are sterically hindered. In addition, they aredeactivated by the electron-withdrawing effect of uorine atomson DFBP monomers.

Thus, the tethers could only be attached at sites e, f, g or h onthe backbone, shown in Fig. 2a. Two-dimensional HSQC and

J. Mater. Chem. A

one-dimensional NOE NMR spectra, as shown in Fig. 3a and b,respectively, were used to identify the location of the tetherattachment. In order to simplify the experiment, the hydro-philic oligomer with one tether aer Friedel–Cras reaction wasused as the sample.

In the HSQC spectrum, the cross peaks show the coupling ofcarbon and proton bonded nuclei. Cross peaks 1 and 2 showthat the corresponding protons Ha were linked to two differentcarbons Cc and Cf. Thus, the peaks for these protons in the 1HNMR spectrum are an overlapping peak composed of a singletand a doublet peak. Similarly, cross peaks 3 and 4 show that thepeaks of the corresponding protonsHb in the 1H NMR spectrumare composed of two overlapping doublet peaks. The protonsHc

for cross peaks 5 and 6 were linked to carbon Cd and Ce, andshow a doublet and triplet overlapped peak. The proton corre-sponding to cross peak 7 is a singlet peak and only linked to

This journal is © The Royal Society of Chemistry 2016

Fig. 3 (a) Two-dimensional HSQC, and (b) one-dimensional NOENMR spectra.

Fig. 4 1H NMR spectra of (a) BrKC6–BrC6-mPEs, and (b) 2BrC6-mPEs.

Paper Journal of Materials Chemistry A

Publ

ishe

d on

26

Sept

embe

r 20

16. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 0

6/10

/201

6 14

:05:

26.

View Article Online

carbon Cg. There are two possible locations for the tetherattachment which could show this set of peaks in the 1H NMRspectrum: f or g site.

In order to further identify the tether attachment location, for g, the one-dimensional NOE spectrum, which identiesspatially close atom pairs, was obtained. In Fig. 3b, the uppertrace is the 1H NMR spectrum of the aromatic protons of one-tether hydrophilic oligomer, and the lower trace is the corre-sponding NOE difference spectrum obtained from irradiationof proton d at 7.23 ppm. It shows that NOE can be seen betweenthe proton of a singlet peak at 7.98 ppm and proton d, whichmeans that these protons have spatial proximity. This provesthat the Friedel–Cras reaction occurred at the f site instead ofthe g site. The resulting singlet peak is attributed to aryl protone’ formed at the expense of the proton at f due to the attachmentof tethers. The attachment at g site cannot result in the observedsinglet peak.

The number of tethers attached to each monomer wascalculated from the ratio of the integrated peaks i and the sumof the aromatic protons, shown in Fig. 2b. It was found that onlyone tether was attached per hydrophilic repeat unit, eventhough there was excess 6-bromohexanoyl chloride reactantduring the Friedel–Cras reaction. This is due to the strongelectron-withdrawing effect of the ketone group within thetether, which reduced the reactivity of aromatic rings.

This journal is © The Royal Society of Chemistry 2016

As suggested by Hibbs,24 protons in the a-position of ketonesare known to be acidic due to the formation of enolate ions, sothe a-protons on the side chains might lead to degradationreactions under alkaline conditions. It was also shown that thecorresponding reduction product was more stable in the alka-line stability test. The ketone groups of side chains in BrKC6-mPEs were reduced by triethylsilane to methylene groups forbetter alkaline stability. The appearance of peak p (2.60 ppm),assigned to the methylene group adjacent to aromatic ring, andthe shi of peak i, assigned to the a-protons, from 2.95 ppm to1.60 ppm, in Fig. 2c were clear evidence for the reduction of theketone on the tether.

Aer the ketone groups were reduced to methylene groups,the electron-withdrawing effect of the tether was eliminated andthe aromatic rings were again susceptible to the Friedel–Crasacylation reaction. Fig. 4a and b show the 1H NMR spectra of theFriedel–Cras reaction product with new tethers, BrKC6–BrC6-mPEs, and the corresponding reduction product 2BrC6-mPEs,respectively. In Fig. 4a, the 1 : 1 ratio of the integration of peakareas of peak i0, assigned to the new methylene group adjacentto benzyl ketone group, and peak p shows that a second tetherwas attached to the hydrophilic oligomer repeat unit on theremaining f site.

The one-tether BrC6-mPEs and two-tether 2BrC6-mPEs werequaternized by reaction with trimethylamine (TMA). Fig. 5shows the 1H NMR spectra of the quaternized multiblockcopolymer Q-mPEs with one tether. The appearance of peak q,assigned to methyl groups of trimethyl quaternary ammoniumions, conrms the successful quaternization reaction. In

J. Mater. Chem. A

Fig. 5 1H NMR spectrum of Q-mPEs.

Journal of Materials Chemistry A Paper

Publ

ishe

d on

26

Sept

embe

r 20

16. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 0

6/10

/201

6 14

:05:

26.

View Article Online

addition, the 9 : 2 ratio of peak q to peak p (protons of methy-lene groups on the tether) shows that the methylene bromidewas fully quaternized.

Morphology

AFM phase images of the membranes (500 � 500 nm) wereobtained for morphology analysis. Hydrophilic-hydrophobicphase separation was found for all multiblock copolymermembranes. Each had a different domain size and size distri-bution, as shown in Fig. 6. The dark regions correspond to thehydrophilic domains and the bright regions correspond to thehydrophobic domains.14,15 The hydrophilic domains formedcontiguous ion conductive nano-channels for ion transport.14,15

The resulting properties of the membranes, such as ionicconductivity and water uptake, were the result of the ion-channel size, density, and mobility of anions within the chan-nels. The properties are summarized in Table 2. The sizes of theionic channels were measured to be in the range of 7.8 to22.5 nm. This channel size was determined by the length of the

Fig. 6 AFM phase images of mPEs membranes.

J. Mater. Chem. A

hydrophobic and hydrophilic blocks, tether size and number,and microstructure of the membrane. For example, mPE-X3.1Y3.6-1 and mPE-X3.1Y3.6-2 had the same backbone structurewith one and two tethers per monomer unit, respectively. Themore tethers in mPE-X3.1Y3.6-2 resulted in a larger channel size.The channel size of mPE-X3.1Y8-2 was even larger, because it hada longer hydrophilic block. In addition, the pure hydrophilicpolymer based membrane, H-1, also showed a certain degree ofphase segregation. However, the multiblock copolymermembranes consistently had a more uniform distribution ofchannel sizes.

Ion exchange capacity (IEC) and ionic conductivity

The ion exchange capacity was calculated by quantitative NMRanalysis of the number of tethers with respect to the number ofmonomer units and assuming 100% quaternization of the alkylbromide tether. The IEC was calculated by eqn (9).

IEC ¼ Ntethery

Mb

(9)

Ntether is the number of tethers attached to each hydrophilicrepeat unit, y is the number of repeat units for the hydrophilicblock, andMb is themolecular weight of block copolymer repeatunit for the overall IEC or the molecular weight of hydrophilicblock for local IEC.

As shown in Table 2, the local IEC in the hydrophilic blockwas the same for each of the one-tether or two-tether mPEscopolymers. The introduction of hydrophobic blocks results inthe overall IEC lower than the local IEC. The IEC was higher fortwo-tether block copolymer than one-tether ones with the samebackbone structure. For example, the IEC of mPE-X5.4Y7-2 was

This journal is © The Royal Society of Chemistry 2016

Table 2 Channel size, IEC, ionic conductivity, water uptake, hydration number, Nfree and Nbound of mPEs membranes

MembraneaChannelsizeb (nm)

Overall IECc

(meq. g�1)Local IECd

(meq. g�1)

Ionicconductivitye

(mS cm�1)Wateruptakef (%)

Hydrationnumber, l Nfree NboundR. T. 80 �C

H-1 16.5 � 3.9 1.25 1.25 13.1 36.1 35.9 16.9 5.1 11.7X3.1Y3.6-1 7.8 � 1.2 0.66 1.25 16.4 51.5 5.6 4.7 0.2 4.5X5.4Y7-1 9.7 � 1.6 0.73 1.25 14.1 34.7 8.0 6.1 0.5 5.6X5.4Y7-2 22.5 � 2.9 1.30 2.08 38.2 119.7 50.8 21.7 11.9 9.8X3.1Y8-2 15.3 � 1.3 1.56 2.08 23.1 94.0 26.7 9.5 0.6 8.9X3.1Y3.6-2 12.2 � 1.4 1.19 2.08 25.8 85.0 25.0 11.7 1.8 9.9X5.9Y5-2 11.3 � 1.9 1.10 2.08 22.1 66.7 19.6 9.9 0.5 9.4

a 1, 2: the number of tethers on each hydrophilic repeat unit. b Channel size: calculated by the average size measured on AFM phase image. c OverallIEC: with respect to the multiblock copolymer. d Local IEC: with respect to the hydrophilic block. e Ionic conductivity: tested in hydroxide ion form.f Water uptake: measured at room temperature in hydroxide ion form.

Paper Journal of Materials Chemistry A

Publ

ishe

d on

26

Sept

embe

r 20

16. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 0

6/10

/201

6 14

:05:

26.

View Article Online

higher than mPE-X5.4Y7-1, because each tether had an ionichead-group.

Fig. 7 shows the hydroxide conductivity of the membranesfrom 20 �C to 80 �C. Generally, the ionic conductivity increaseswith temperature and follows an Arrhenius relationship. Theslope of the Arrhenius plot corresponds to the activation energyfor ionic conductivity. The mPE membranes shows a similaractivation value, about 16 to 17 kJ mol�1. Typically, the ionicconductivity increased with IEC.5 With the same backbonestructure, mPE-X5.4Y7-2 showed higher ionic conductivity thanmPE-X5.4Y7-1, because mPE-X5.4Y7-2 had a higher IEC. In addi-tion, it is noted that the ionic conductivity of mPE-X5.4Y7-2 wasmore than twice of that of mPE-X5.4Y7-1, even though thenumber of head-groups of mPE-X5.4Y7-2 only doubled. Thisdemonstrates a non-linear relationship between the number ofhead groups and ionic conductivity. With more head groups onthe same backbone, the multiblock copolymer formed largernano-channels for ion transport, resulting in higher ionic

Fig. 7 Arrhenius plot of ionic conductivity vs. inverse temperature.

This journal is © The Royal Society of Chemistry 2016

conductivity. However, even with the same number of head-groups per hydrophilic repeat unit, the mPEs multiblockcopolymers had different ionic conductivities. This is becausethe mPEs multiblock copolymers formed different size nano-channels for ion transport due to the different size of the phasescaused by the different ratio of hydrophobic to hydrophilicblock lengths. For example, the multiblock copolymermembrane mPE-X3.1Y3.6-1 had a lower IEC than that of purehydrophilic polymer based H-1, but higher ionic conductivity,which shows the efficiency of conductive nano-channels for iontransport. For all tested mPEs membranes, the larger the iontransport channel size in the multiblock copolymer, the higherthe resulting ion conductivity. The mPE-X5.4Y7-2 had the high-est ionic conductivity, up to 119 mS cm�1 at 80 �C, however, notthe highest IEC. This did not follow the typical relationshipbetween IEC and ionic conductivity for random copolymers. Inmultiblock copolymers, the ion mobility within the channelsalso greatly affects the ionic conductivity. In other words, theoptimum polymer structure will optimize the size of ion chan-nels and mobility of hydroxide within the channel.

Water uptake (WU), hydration number (l), number offreezable water molecules (Nfree) and bound, non-freezablewater molecules (Nbound)

Water uptake is an important property for AEMs. AEMs need tomaintain a certain amount of water for ion hydration andtransport. However, excess, unbound water is not productiveand swells the membrane, leading to poor performance of themembrane electrode assembly (MEA) and ooding of the ionconductive channels. In order to reduce the WU of AEMs,a uorinated backbone was used in the mPEs multiblockcopolymer to introduce greater hydrophobicity. The wateruptake of the multiblock copolymer membranes, shown inTable 2, was less than the corresponding pure hydrophilicpolymer. For example, H-1 had 35.9% WU and mPE-X3.1Y3.6-1had much lower value of 5.6%. Similar to ionic conductivity,water uptake of mPEs membranes was affected by both the IECand nano-channel size: higher IEC led to higher WU, and

J. Mater. Chem. A

Fig. 8 Relationship of ionic conductivity at 80 �C and water uptake ofmPEs membranes.

Journal of Materials Chemistry A Paper

Publ

ishe

d on

26

Sept

embe

r 20

16. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 0

6/10

/201

6 14

:05:

26.

View Article Online

excessively large channels led to excessive WU. The properchannel size was important to acquire optimized WU. The two-tether mPEs membranes had higher WU than the one-tethermembranes. The mPEs membranes with the same number oftethers showed different WU due to different ion conductivechannel sizes.

The hydration number is the number of water molecules perionic head-group (cation/anion pair). As shown in Table 2,multiblock copolymers had lower hydration numbers than thehydrophilic polymer H-1, except mPE-X5.4Y7-2, which showsthat the efficiency of water utilization in the multiblock copol-ymer membrane for ion transport was higher. In other words,the multiblock copolymer membranes could efficiently utilizewater in ion transport due to the well-developed nanophaseseparation. The reason for the high hydration number of mPE-X5.4Y7-2 was that the ion conductive channel size was too large,which resulted in higher WU and more free water.

To gain a deeper insight into water mobility and its effect onion transport in AEMs, the state of water in the membranes,freezable and bound water (or non-freezable water), was inves-tigated by DSC. Non-freezable water is bound to an ion or polarpolymer segment and shows no characteristic thermal transi-tion in DSC. Freezable water is associated with ion exchangesand frozen at subzero temperatures. The sum of the numbers offreezable water and non-freezable water is the hydrationnumber. The number of freezable water molecules (Nfree) andnon-freezable water molecules (Nbound) per head-group (orcation-hydroxide ion pear) are summarized in Table 2. It showsthat all the two-tether multiblock copolymer mPEs had a similaramount of bound water, 9 to 10, per ionic pair, or 4.5 to 5 boundwater per cation or anion, regardless of overall WU. So thedifference in hydration numbers was only due to the differencein free, unbound water. The number of bound water per head-group for the one-tether mPEs membrane was lower because ofthe small ion conductive nano-channel size, which limited thewater absorption and mobility in the membrane. Excess,unbound water could lead to ooding of the ion conductivechannels, which reduces the efficiency of ion transport. Asshown, the membranes could still maintain high ionicconductivity even with little freezable, unbound water, such aswith mPE-X3.1Y8-2 and mPE-X3.1Y3.6-2. This implies that thedominant ion transport occurs close to the ion exchange groupsthrough the assistance of non-freezable, bound water on theions.

Fig. 8 shows the relationship between ionic conductivity andWU for mPEs membranes. It is desirable to have sufficientbound water for ion hydration and transport without excess,unbound water, the upper le region of Fig. 8. Among all mPEsmembranes, mPE-X5.4Y7-2 had the highest ionic conductivity,119.7 mS cm�1 at 80 �C, but relatively high water uptake, 50.8%.Although this value of WU is not excessive,3 it shows there were11.9 unbound water molecules per head-group. There were only9.8 water molecules per head-group needed for hydration. Thereason for the excess free water was the large size of the nano-channels. Membrane mPE-X3.1Y8-2 had lower water uptake,26.7%, and high ionic conductivity up to 94.0 mS cm�1 at 80 �C.There were 8.9 bound water molecules per head-group, about

J. Mater. Chem. A

the same as that for mPE-X5.4Y7-2, however, it had virtually nofree water, only 0.6 water molecules per anion/cation pair. Thisis due to the optimized size of the nano-channels. In compar-ison, the pure hydrophilic polymer, H-1 membrane, showed lowionic conductivity and high water uptake, which is in the lowerright region in Fig. 8. This is due to the lack of ion-channelformation and inefficient use of the ionic head-groups. Thus,phase segregation within the multiblock copolymer was bene-cial to the improvement in ion conductivity and reduced WU.

Alkaline stability

Long-term alkaline stability of AEMs is of great importance forfuel cell applications. The mPEs membranes with long tethersshowed outstanding alkaline stability in strong base solutions.The alkaline stability of mPE-X3.1Y3.6-2 and mPE-X5.9Y5-2membranes was evaluated by measuring the drop in ionicconductivity due to soaking in 1 M NaOH at 60 �C. As shown inFig. 9, nearly no change in ionic conductivity was observed aer1000 h. The drop in ionic conductivity for mPE-X3.1Y3.4-2and mPE-X5.9Y5-2 were 1.2% and 3.0%, respectively. Thisresult conrmed that long alkyl tether for quaternary ammo-nium ionic exchange groups greatly improves the alkalinestability.23,24

Thermal stability and mechanical properties

Fig. 10 showed the thermal degradation behavior of mPEsmembranes in the hydroxide form via TGA. All of themembranes showed a similar tendency. The initial weight lossbelow 100 �C corresponds to evaporation of water in themembrane. The mPEs membranes showed a three-stage weightloss. The rst stage from 180 �C to 220 �C was due to thedegradation of the quaternary ammonium groups. The secondstage from 220 �C to 400 �C was due to the degradation of thelong alkyl side chains. All the membranes underwent polymerbackbone degradation above 400 �C. This trend in degradation

This journal is © The Royal Society of Chemistry 2016

Fig. 9 Alkaline stability of the mPEs membranes in 1 M NaOH solutionat 60 �C.

Fig. 10 TGA curves of mPEs membranes under nitrogen atmosphere.

Table 3 Mechanical properties of mPEs membranes

MembraneTensile strength(MPa)

Young'smodulus (GPa)

Elongation atbreak (%)

H-1 29.4 1.6 2.3X3.1Y3.6-1 33.4 0.8 5.5X5.4Y7-1 27.7 2.3 1.7X5.4Y7-2 12.3 0.5 3.4X3.1Y8-2 19.1 1.3 2.1X3.1Y4-2 24.8 0.9 2.7X5.9Y5-2 21.2 1 2.5

Table 4 Oxygen transport properties of Nafion® and mPEs blockcopolymersa

S(mass%) D (cm2 s�1) P (mol (bar cm s)�1)

H-1 1.11 3.0 � 10�11 1.1 � 10�12

X3.1Y3.6-1 0.90 1.3 � 10�8 4.0 � 10�10

X5.4Y7-1 0.92 8.5 � 10�9 2.7 � 10�10

X5.4Y7-2 0.81 9.0 � 10�10 2.5 � 10�11

X3.1Y8-2 0.64 5.9 � 10�9 1.3 � 10�10

X3.1Y3.6-2 0.71 5.2 � 10�9 1.3 � 10�10

X5.9Y5-2 1.14 1.0 � 10�8 3.9 � 10�10

Naon® 0.27 1.3 � 10�7 1.2 � 10�9

Tokuyama A201 0.38 2.0 � 10�9 2.6 � 10�11

a The density of Naon® and Tokuyama A201 were taken as 1.1 g cm�3

for calculation.

Paper Journal of Materials Chemistry A

Publ

ishe

d on

26

Sept

embe

r 20

16. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 0

6/10

/201

6 14

:05:

26.

View Article Online

was consistent with reported multi-block copolymermembranes with long alkyl side chains.13,19

Good mechanical stability is essential for the fabrication ofMEAs and withstanding the pressure difference between twosides of the membrane during fuel cell operation. Themechanical properties were obtained from stress–strain curveof mPEs membranes, as summarized in Table 3. The tensilestrength of the membranes was in the range of 12.3 to 33.4 MPa,and the Young's modulus was in the range of 0.5 to 2.3 GPa,indicating that the membranes had adequate mechanicalproperties for use as polymer electrolytes. The tensile strengthdecreased with the increase in the number of tethers andhydrophilic content. This was likely due to the increase of watercontent, which might act as a plasticizer15 in the membranesand disintegrate the originally tightly arranged polymer chains.

This journal is © The Royal Society of Chemistry 2016

Oxygen transport property

The oxygen solubility, diffusivity and permeability of all themembrane samples, Naon®, and Tokuyama A201 aresummarized in Table 4. Low oxygen diffusivity and permeabilityare desired for AEMs in order to get minimized oxygen crossoverduring fuel cell operation. As shown in Table 4, the oxygensolubility of the multiblock copolymer membranes was higherthan that of the proton conductive Naon. However, the oxygendiffusivity and permeability of the multiblock copolymers wasless than that of Naon. Compared with the commercial AEMmaterial Tokuyama A201, the oxygen solubility of multiblockcopolymer membranes was higher. The oxygen diffusivity andpermeability were on the same order of magnitude, 10�9 cm2

s�1 and 10�10 to 10�11 mol (bar cm s)�1, respectively. It is notedthat the oxygen transport properties of mPE-X5.4Y7-2 were betterthan A201 membrane. By comparing mPE-X3.1Y3.6-1 and mPE-X3.1Y3.6-2 or mPE-X5.4Y7-1 andmPE-X5.4Y7-2, it can be concludedthat the two tether mPEs membrane had lower oxygen diffu-sivity and permeability than the one tether membrane. Themembranes with higher ionic conductivity and water uptake,such as mPE-X5.4Y7-2 and mPE-X3.1Y3.6-2, showed lower oxygendiffusivity and permeability. The pure hydrophilic oligomermembrane showed the lowest oxygen permeability. Thissuggests that the hydrophobic domains facilitate oxygentransport or that the channels for oxygen transport wereimpeded by water.

J. Mater. Chem. A

Journal of Materials Chemistry A Paper

Publ

ishe

d on

26

Sept

embe

r 20

16. D

ownl

oade

d by

Geo

rgia

Ins

titut

e of

Tec

hnol

ogy

on 0

6/10

/201

6 14

:05:

26.

View Article Online

Conclusions

A systematic study of the effect of the hydrophilic and hydro-phobic block lengths and ion exchange capacity of partiallyuorinated multiblock copolymer mPEs with long head-grouptethers was undertaken to explore the relationship of thechemical structure, morphology and properties of the mPEsAEMs. The chemical structure was studied via one-dimensional1H NMR, 19F NMR, NOE, and two-dimensional HSQC NMRspectra. Only one tether could be attached per hydrophilicrepeat unit at each time. The formation of ion conductive nano-channels for hydroxide ion transport due to nanophase sepa-ration of the multiblock copolymers greatly improved the ionicconductivity and reduced the water uptake. Multiblock copol-ymer mPE-X5.4Y7-2 showed the highest ionic conductivity, 119mS cm�1 at 80 �C, but not the highest IEC, because it formedefficient channels. The ratio of ionic conductivity to wateruptake of the multiblock copolymers was high. For example,mPE-X3.1Y8-2 had very high ionic conductivity up to 94.0 mScm�1 at 80 �C, but only 26.7% water uptake. From DSCmeasurements of the number of freezable water and boundwater molecules, the number of bound water molecules per ionof two-tether polymers was 4.5 to 5. In addition, the boundwater played the dominant role in the hydroxide ion transportwithin the channels. The multiblock copolymer AEM showedgood thermal, mechanical stability and excellent alkalinestability. The ionic conductivity was hardly changed aersoaking the membrane in 1 M NaOH solution at 60 �C for1000 h. Oxygen transport properties were also investigated. ThemPEs AEM with higher ionic conductivity showed lower oxygendiffusivity and permeability, which means that the oxygencrossover problem was less severe.

Acknowledgements

We thank Shengming Li for the assistance of AFM test andgratefully acknowledge the nancial support of the ArmyResearch Laboratory.

Notes and references

1 B. C. H. Steele and A. Heinzel, Nature, 2001, 414, 345–352.2 S. J. Peighambardoust, S. Rowshanzamir and M. Amjadi, Int.J. Hydrogen Energy, 2010, 35, 9349–9384.

3 J. R. Varcoe, P. Atanassov, D. R. Dekel, A. M. Herring,M. A. Hickner, P. A. Kohl, A. R. Kucernak, W. E. Mustain,K. Nijmeijer, K. Scott, T. Xu and L. Zhuang, Energy Environ.Sci., 2014, 7, 3135–3191.

4 J. R. Varcoe and R. C. T. Slade, Fuel Cells, 2005, 5, 187–200.5 M. A. Hickner, A. M. Herring and E. B. Coughlin, J. Polym.Sci., Part B: Polym. Phys., 2013, 51, 1727–1735.

6 J. Zhou, K. Joseph, J. M. Ahleld, D.-Y. Park and P. A. Kohl, J.Electrochem. Soc., 2013, 160, F573–F578.

7 G. Merle, M. Wessling and K. Nijmeijer, J. Membr. Sci., 2011,377, 1–35.

8 G. Couture, A. Alaaeddine, F. Boschet and B. Ameduri, Prog.Polym. Sci., 2011, 36, 1521–1557.

J. Mater. Chem. A

9 Y.-J. Wang, J. Qiao, R. Baker and J. Zhang, Chem. Soc. Rev.,2013, 42, 5768–5787.

10 N. Li andM. D. Guiver,Macromolecules, 2014, 47, 2175–2198.11 J. Pan, C. Chen, Y. Li, L. Wang, L. Tan, G. Li, X. Tang, L. Xiao,

J. Lu and L. Zhuang, Energy Environ. Sci., 2014, 7, 354–360.12 D.-Y. Park, P. A. Kohl and H. W. Beckham, J. Phys. Chem. C,

2013, 117, 15468–15477.13 L. Li, X. Yue, W. Wu, W. Yan, M. Zeng, Y. Zhou, S. Liao and

X. Li, RSC Adv., 2016, 6, 41453–41464.14 E. A. Weiber, D. Meis and P. Jannasch, Polym. Chem., 2015, 6,

1986–1996.15 A. N. Lai, L. S. Wang, C. X. Lin, Y. Z. Zhuo, Q. G. Zhang,

A. M. Zhu and Q. L. Liu, ACS Appl. Mater. Interfaces, 2015,7, 8284–8292.

16 A. C. Cope and E. R. Trumbull, in Organic Reactions, JohnWiley & Sons, Inc., 2004, DOI: 10.1002/0471264180.or011.05.

17 S. Chempath, J. M. Boncella, L. R. Pratt, N. Henson andB. S. Pivovar, J. Phys. Chem. C, 2010, 114, 11977–11983.

18 S. Chempath, B. R. Einsla, L. R. Pratt, C. S. Macomber,J. M. Boncella, J. A. Rau and B. S. Pivovar, J. Phys. Chem. C,2008, 112, 3179–3182.

19 A. H. N. Rao, S. Nam and T.-H. Kim, J. Mater. Chem. A, 2015,3, 8571–8580.

20 F. Gu, H. Dong, Y. Li, Z. Si and F. Yan,Macromolecules, 2013,47, 208–216.

21 S. Gu, R. Cai, T. Luo, K. Jensen, C. Contreras and Y. Yan,ChemSusChem, 2010, 3, 555–558.

22 D. S. Kim, C. H. Fujimoto, M. R. Hibbs, A. Labouriau,Y.-K. Choe and Y. S. Kim, Macromolecules, 2013, 46, 7826–7833.

23 P. Jannasch and E. A. Weiber, Macromol. Chem. Phys., 2016,217, 1108–1118.

24 M. R. Hibbs, J. Polym. Sci., Part B: Polym. Phys., 2013, 51,1736–1742.

25 A. D. Mohanty and C. Bae, J. Mater. Chem. A, 2014, 2, 17314–17320.

26 H.-S. Dang and P. Jannasch,Macromolecules, 2015, 48, 5742–5751.

27 Z. Yang, J. Zhou, S. Wang, J. Hou, L. Wu and T. Xu, J. Mater.Chem. A, 2015, 3, 15015–15019.

28 M. Shimada, S. Shimada, J. Miyake, M. Uchida andK. Miyatake, J. Polym. Sci., Part A: Polym. Chem., 2016, 54,935–944.

29 C. Fujimoto, D.-S. Kim, M. Hibbs, D. Wrobleski andY. S. Kim, J. Membr. Sci., 2012, 423–424, 438–449.

30 S. A. Nunez and M. A. Hickner, ACS Macro Lett., 2013, 2, 49–52.

31 W.-H. Lee, A. D. Mohanty and C. Bae, ACS Macro Lett., 2015,4, 453–457.

32 S. J. Lue and S.-J. Shieh, J. Macromol. Sci., Part B: Phys., 2009,48, 114–127.

33 B. Mecheri, V. Felice, Z. Zhang, A. D'Epifanio, S. Licoccia andA. C. Tavares, J. Phys. Chem. C, 2012, 116, 20820–20829.

34 A. L. Moster and B. S. Mitchell, J. Appl. Polym. Sci., 2009, 113,243–250.

35 D. S. Kim, G. P. Robertson, Y. S. Kim and M. D. Guiver,Macromolecules, 2009, 42, 957–963.

This journal is © The Royal Society of Chemistry 2016