General rights Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights.

Users may download and print one copy of any publication from the public portal for the purpose of private study or research.

You may not further distribute the material or use it for any profit-making activity or commercial gain

You may freely distribute the URL identifying the publication in the public portal If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from orbit.dtu.dk on: Dec 12, 2021

PBI nanofiber mat-reinforced anion exchange membranes with covalently linkedinterfaces for use in water electrolysers

Najibah, Malikah; Tsoy, Ekaterina; Khalid, Hamza; Chen, Yongfang; Li, Qingfeng; Bae, Chulsung; Hnát,Jaromír; Plevová, Michaela; Bouzek, Karel; Jang, Jong HyunTotal number of authors:12

Published in:Journal of Membrane Science

Link to article, DOI:10.1016/j.memsci.2021.119832

Publication date:2021

Document VersionPeer reviewed version

Link back to DTU Orbit

Citation (APA):Najibah, M., Tsoy, E., Khalid, H., Chen, Y., Li, Q., Bae, C., Hnát, J., Plevová, M., Bouzek, K., Jang, J. H., Park,H. S., & Henkensmeier, D. (2021). PBI nanofiber mat-reinforced anion exchange membranes with covalentlylinked interfaces for use in water electrolysers. Journal of Membrane Science, 640, [119832].https://doi.org/10.1016/j.memsci.2021.119832

Journal Pre-proof

PBI nanofiber mat-reinforced anion exchange membranes with covalently linkedinterfaces for use in water electrolysers

Malikah Najibah, Ekaterina Tsoy, Hamza Khalid, Yongfang Chen, Qingfeng Li,Chulsung Bae, Jaromír Hnát, Michaela Plevová, Karel Bouzek, Jong Hyun Jang,Hyun S. Park, Dirk Henkensmeier

PII: S0376-7388(21)00776-6

DOI: https://doi.org/10.1016/j.memsci.2021.119832

Reference: MEMSCI 119832

To appear in: Journal of Membrane Science

Received Date: 15 June 2021

Revised Date: 23 August 2021

Accepted Date: 2 September 2021

Please cite this article as: M. Najibah, E. Tsoy, H. Khalid, Y. Chen, Q. Li, C. Bae, Jaromí. Hnát, M.Plevová, K. Bouzek, J.H. Jang, H.S. Park, D. Henkensmeier, PBI nanofiber mat-reinforced anionexchange membranes with covalently linked interfaces for use in water electrolysers, Journal ofMembrane Science (2021), doi: https://doi.org/10.1016/j.memsci.2021.119832.

This is a PDF file of an article that has undergone enhancements after acceptance, such as the additionof a cover page and metadata, and formatting for readability, but it is not yet the definitive version ofrecord. This version will undergo additional copyediting, typesetting and review before it is publishedin its final form, but we are providing this version to give early visibility of the article. Please note that,during the production process, errors may be discovered which could affect the content, and all legaldisclaimers that apply to the journal pertain.

© 2021 Published by Elsevier B.V.

Malikah Najibah Investigation, Writing - Original Draft, methodology, formal analysis,

Visualization

Ekaterina Tsoy Methodology Hamza Khalid Investigation, methodology, formal analysis

Yongfang Chen Investigation, methodology

Qingfeng Li Resources, Writing - Review & Editing, Supervision

Chulsung Bae Resources, Writing - Review & Editing Michaela Plevová Investigation, formal analysis

Jaromír Hnát Investigation, Writing - Original Draft, Visualization

Karel Bouzek Writing - Review & Editing, Project administration, Funding acquisition

Hyun S. Park Writing - Review & Editing

Jong Hyun Jang Writing - Review & Editing

Dirk Henkensmeier Conceptualization, Validation, Writing - Original Draft, Writing - Review

& Editing, Visualization, Supervision, Project administration, Funding

acquisition

Term Definition

Conceptualization Ideas; formulation or evolution of overarching research goals and aims

Methodology Development or design of methodology; creation of models

Software

Programming, software development; designing computer programs;

implementation of the computer code and supporting algorithms; testing

of existing code components

Validation

Verification, whether as a part of the activity or separate, of the overall

replication/ reproducibility of results/experiments and other research

outputs

Formal analysis Application of statistical, mathematical, computational, or other formal

techniques to analyze or synthesize study data

Investigation Conducting a research and investigation process, specifically performing

the experiments, or data/evidence collection

Resources

Provision of study materials, reagents, materials, patients, laboratory

samples, animals, instrumentation, computing resources, or other analysis

tools

Data Curation

Management activities to annotate (produce metadata), scrub data and

maintain research data (including software code, where it is necessary for

interpreting the data itself) for initial use and later reuse

Writing - Original

Draft

Preparation, creation and/or presentation of the published work,

specifically writing the initial draft (including substantive translation)

Writing - Review &

Editing

Preparation, creation and/or presentation of the published work by those

from the original research group, specifically critical review, commentary

or revision – including pre-or postpublication stages

Journ

al Pre-

proof

Term Definition

Visualization Preparation, creation and/or presentation of the published work,

specifically visualization/ data presentation

Supervision Oversight and leadership responsibility for the research activity planning

and execution, including mentorship external to the core team

Project

administration

Management and coordination responsibility for the research activity

planning and execution

Funding acquisition Acquisition of the financial support for the project leading to this

publication

Journ

al Pre-

proof

Graphical Abstract

(CH2)5

R = Br: mTPBrR = N(Me)3

+ Br- : mTPN

PBI/mTPN

Journ

al Pre-

proof

1

PBI nanofiber mat-reinforced anion exchange membranes with covalently linked interfaces for 1

use in water electrolysers 2

3

Malikah Najibah,1,2# Ekaterina Tsoy,1,2 Hamza Khalid,1,2 Yongfang Chen,3 Qingfeng Li,3 Chulsung 4

Bae,4 Jaromír Hnát,5 Michaela Plevová,5 Karel Bouzek,5 Jong Hyun Jang,1,2 Hyun S. Park,1,2 Dirk 5

Henkensmeier1,2* 6

7

1. Center for Hydrogen and Fuel Cell Research, Korea Institute of Science and Technology, 8

Seongbukgu, Seoul 02792, South Korea 9

2. Division of Energy & Environment Technology, KIST School, University of Science and 10

Technology, Seongbukgu, Seoul 02792, South Korea 11

3. Department of Energy Conversion and Storage, Technical University of Denmark, 12

Fysikvej, Building 310, 2800 Kgs. Lyngby, Denmark. 13

4. Department of Chemistry and Chemical Biology, Rensselaer Polytechnic Institute, Troy, New 14

York 12180, United States 15

5. Department of Inorganic Technology, University of Chemistry and Technology, 16

Prague, Technická 5, 166 28 Praha 6, Czech Republic 17

*email: henkensmeier@kist.re.kr 18

#main name is Malikah 19

20

Abstract: 21

Anion exchange membranes (AEM) are key components in anion exchange membrane water 22

electrolysers. Recently developed materials are less susceptible to the alkaline degradation of 23

the polymer backbone and quaternary ammonium groups. A remaining challenge is the 24

mechanical stability in contact with hot water and dimensional stability when the temperature 25

of the feed solution changes. One solution is to reinforce membranes with a porous support. 26

Since support materials like PEEK or PTFE have a different swelling behavior than the matrix and 27

no strong interactions with the matrix, voids can form, and gas crossover increases. In this work, 28

we approach this issue by pore filling polybenzimidazole nanofiber mats with the 29

Journ

al Pre-

proof

2

bromomethylated precursor of mTPN, an ultra-stable AEM material. During drying, a covalent 30

interaction between support (PBI amine groups) and matrix (-CH2Br) is established. After 31

quaternization, the optimized PBI/mTPN-50.120 composite membrane still shows a high 32

conductivity of 62 mS cm-1, but 37% reduced length swelling in comparison to the non-reinforced 33

membrane. Tensile strength and Young modulus increase 17% and 56% to 49 MPa and 680 MPa, 34

respectively. In an electrolyser, a stable voltage of 1.98V at 0.25 A cm-2 was achieved, and no 35

change in membrane resistance was observed over the test time of 200 hours (50 °C, 1M KOH, 36

catalysts based on Ni/Fe and Mo). 37

38

Keywords: polybenzmidazole; electrospinning; nanofiber mat; anion exchange membrane; water 39

electrolysis 40

41

1. Introduction 42

Worldwide, societies aim to increase the use of renewable energy, mainly by increasing the use 43

of intermittent energy sources like solar and wind power. To stabilize the electricity grids, to store 44

energy for times when demand exceeds primary supply, and to substitute fossil fuels in the 45

transport sector, hydrogen has been proposed to be a clean and efficient energy carrier. 46

Furthermore, hydrogen can provide a practical means to store large amounts of energy, e.g. in 47

salt caverns.[1] In China in 2016, the amount of abandoned electricity from wind, solar and 48

hydroelectric power amounted to 110,000 GWh, which is estimated to be enough to produce 22 49

billion Nm3 hydrogen.[2] Therefore, hydrogen production by water electrolysis based on 50

renewable energies will play a central role in the future clean energy systems. 51

52

Even though alkaline water electrolysis is a well-established industrial process, the current 53

technology is not suitable for applications in association with renewable energies. Industrial 54

alkaline water electrolysis uses a porous diaphragm as a separator of the gas products, which can 55

cause gas crossover and hence a hydrogen purity issue, especially at differential pressure.[3] 56

Another shortcoming in connection to the diaphragm is the low partial load range (operational 57

at 20-40% full load) and poor dynamics of operation (startup/shutdown and load cycling) of the 58

Journ

al Pre-

proof

3

technology.[4] In addition, the large thickness of the porous diaphragm dictates the inter-59

electrode distance and hence enhances ohmic resistance, which, in turn, limits the operation at 60

low current densities and the operational capacity. 61

62

Proton exchange membrane water electrolysers (PEMWEs) use dense and thin membranes 63

based on perfluorosulfonic acid with lower ohmic resistance. The technology can operate at high 64

current densities with good dynamic performance to match intermittent renewable energy 65

sources. The disadvantage is that the acidic solid electrolyte requires use of precious metals such 66

as platinum and iridium as catalysts. For example, Germany’s Hydrogen Strategy pushes for the 67

installation of 2 GW water electrolysers until 2030.[5] If this is based on a PEMWE technology 68

operating at a voltage of 2V, current density of 2 A cm-2 and an anode catalyst loading of 2 mg 69

iridium cm-2, Germany would need 1000 kg iridium, which is ca 15% of the global yearly supply 70

of the metal.[6] A solution would be the use of an alkaline solid electrolyte based on an anion 71

exchange membrane which combines the advantages of high and dynamic performance with the 72

potential use of non-precious metal catalysts. 73

74

The main problem hindering the wide promotion of alkaline membrane WE is that no commercial 75

membrane fulfills the requirements.[7] For example, membranes from Tokuyama and Fumatech 76

have a relatively low conductivity, whereas highly conductive Sustainion membranes break into 77

pieces in the dry state.[7] Very promising membranes were developed in Chulsung Bae’s group 78

of Rensselaer Polytechnic Institute[8] and are now commercialized as Orion TM1 membranes. In 79

a recent study surveying 50 membranes from 10 organizations, TM1 membranes showed the 80

lowest IEC loss and conductivity loss after 1000 h degradation study at 80 °C in 1M KOH.[9] 81

However, in the alkaline membrane water electrolyser, the fully hydrated material revealed a 82

low mechanical stability.[10] This specific parameter can be improved by reinforcing the 83

membrane with a porous support. 84

85

A learning from the fuel cell field, where PTFE-supported Nafion membranes are used, is that 86

repeated water absorption and desorption delaminates the Nafion matrix from the rigid Teflon 87

Journ

al Pre-

proof

4

support, resulting in voids along the support’s surface, and thus increasing gas crossover. To 88

tackle this issue, Yu et al.[11] and Deborah Jones and colleagues recently developed new support 89

materials based on electrospun PBI nanofiber mats.[12-14] Because Nafion has strongly acidic 90

groups, it protonates PBI, resulting in strong ionic interactions,[15, 16] which strengthen the 91

interface between PBI support and Nafion matrix. 92

93

Electrospun PBI nanofiber membranes are also investigated as separator in lithium ion 94

batteries,[17] as window air filter[18] and as material for fine dust masks[19]. However, a 95

literature search revealed that only one experimental work reports the combination of 96

“nanofiber membrane” (polysulfone co-extruded with quaternized polyetherethersulfone) and 97

“AEM” (quaternized polyethersulfone).[20] 98

99

In this work, we suggest the use of PBI nanofiber mats as a porous support for anion exchange 100

membranes. While the quaternary ammonium groups of the anion exchange membrane (AEM) 101

can form ionic interactions with PBI, stronger covalent links could be formed when the 102

halomethylated precursor polymer is first filled into the pores of a PBI nanofiber mat, and is then 103

allowed to react with the PBI amine groups. Here, we used the bromoalkylated precursor 104

polymer of TM1 membranes (mTPBr),[8] which was shown to form highly alkaline stable 105

membranes, as mentioned above.[9] As the last step, the remaining bromoalkyl groups are 106

quaternized by immersing the membrane in a solution containing trimethylamine. The effect of 107

the porous support is investigated by measuring swelling behavior, conductivity, mechanical 108

properties and the performance in the water electrolyser. 109

110

2. Experimental part 111

2.1 Materials 112

Polybenzimidazole (PBI, Danish Power Systems, Mw = 58,000 g mol-1), N,N-dimethylacetamide 113

(DMAc, 99.8%, Sigma Aldrich), lithium chloride (LiCl, >99%, Sigma Aldrich), tetrahydrofuran (THF, 114

>99%, Daejung Chemicals), sodium hydride (NaH, 60% dispersion in mineral oil), trimethylamine 115

(TMA, 28 wt% in water, Tokyo Chemical Industry), potassium hydroxide (KOH flakes, Daejung 116

Journ

al Pre-

proof

5

Chemicals), potassium carbonate (K2CO3, Sigma Aldrich), sodium chloride (NaCl, Daejung 117

Chemicals), potassium nitrate (KNO3, Daejung Chemicals) were all used as received. mTPBr was 118

synthesized according to the literature.[8] Ni-Fe and Mo based compounds were used as oxygen 119

and hydrogen evolution catalysts respectively. 120

121

2.2. Preparation of PBI nanofiber mats 122

10 wt% PBI and 0.5 wt% LiCl (to suppress polymer-polymer interactions) were dissolved in 89.5 123

wt% DMAc by stirring at 70 rpm for 19 hours at 150 °C under argon atmosphere. For 124

electrospinning (eS-robot electrospinning/spray system ESR200R2, NanoNC NNC-30K-2mA High 125

Voltage Power Supply), 6 ml solution was filled into a syringe, and spun onto a plate covered by 126

aluminium foil at a voltage of 21.1 kV and a flowrate of 0.5 ml h-1. The relative humidity was 127

maintained at above 45%. After 6 hours, the process was stopped, and the nanofiber mat was 128

collected and then dried at 160 °C for 12 hours in vacuum. 129

130

131

2.3 Preparation of reinforced membranes 132

1 mmol of mTPBr and 1 mmol of NaH were dissolved in 4.12 mL of THF. The solution was stirred 133

overnight at room temperature until all was dissolved. Then the solution was cast on a glass plate 134

with a doctor blade adjusted to 200 µm. A PBI mat (10 x 10 cm2) was put on the film, and the 135

solution was cast once more on the PBI mat with the blade gap adjusted to 410 µm. Finally, the 136

membrane was dried under vacuum at 80 °C or 120 °C for 48 h. Dry PBI/mTPBr membranes 137

prepared in this way had a thickness of 50 µm. Samples were noted as PBI/mTPBr-X.Y, with X 138

indicating the targeted thickness (±15% was considered to be within specification), and Y 139

indicating the drying temperature. 140

141

For quaternization, PBI/mTPBr membranes were immersed in 30 ml of TMA solution (28 wt% in 142

water) at room temperature. After 7 days, the PBI/mTPN-membranes were moved to deionized 143

water and kept overnight to remove residual TMA. Cooling the TMA solution in a refrigerator 144

before opening the storage vessel significantly reduced the smell. 145

Journ

al Pre-

proof

6

146

2.4 Membrane characterization 147

Membranes (1 x 4 cm2) were put into vials filled with 1M NaCl solution to exchange the 148

membranes into their chloride form. After 10 hours, solutions were replaced with DI water and 149

kept for another 24 hours at room temperature. Samples were swollen in water at room 150

temperature, and the wet dimensions were measured. Then the samples were dried 1 day at 151

60 °C in the vacuum for 24 hours to get the dry dimensions. To compare values at room 152

temperature and 60 °C, the same samples were swollen again in water at 60 °C, and wet 153

dimensions were measured. From these wet and dry values, the swelling values were calculated. 154

155

Swelling (length)(%) = 𝐿𝑤𝑒𝑡− 𝐿𝑑𝑟𝑦

𝐿𝑑𝑟𝑦 x 100% 156

157

L is the length of the wet and dry state, as indicated. 158

159

The through-plane conductivity was measured by clamping membrane 1 or several stacked 160

samples between two electrode chambers filled with KOH solution. Gold plated metal discs 161

(1.767 cm2 area) were used as electrodes. Resistances were obtained by impedance spectroscopy 162

with a Zahner IM6 potentiostat. Resistances were plotted against the thickness of the stacked 163

membranes, and the slope S of the linear trend was calculated. The conductivity was calculated 164

by the following equation: 165

166

Through-plane conductivity (mS cm-1) = 1

𝑆· 𝐴 167

168

in which A represents the active membrane area. 169

170

The in-plane conductivity was measured in DI water using a Bekktech BT-110 clamp. To exclude 171

carbonates, nitrogen was constantly bubbled through the water, and a voltage of 2V was applied 172

until the resistance reached a constant value.[21, 22] Resistances were measured by impedance 173

spectroscopy (Zahner IM6), and the conductivity was calculated by the following equation: 174

Journ

al Pre-

proof

7

175

In-plane conductivity (mS cm-1) = 𝑑

𝑅𝑚𝑒𝑚 𝑤 𝑡𝑤

176

177

Where d represents distance between the voltage sensing electrodes, w is the width of 178

membrane sample, and tw is membrane thickness after immersing in DI water. 179

180

Mechanical properties were measured with a Cometech QC-508E universal testing machine. 181

Membrane samples were cut into 4x1 cm2 stripes, then immersed in 1M NaCl solution for 10 182

hours to exchange the membranes into their chloride form. After drying at 60 °C in the vacuum 183

oven for 24 hours, samples were clamped between the fixed and moveable grip, and then 184

stretched at 10 mm min-1. 185

186

Scanning electron micrographs of PBI nanofiber mats and reinforced membranes were obtained 187

by FE-SEM Inspect F, 15 kV, after the samples were sputtered with Pt/Pd. Cross-sectional images 188

of PBI nanofiber mats were obtained by ion milling using Ar blade 5000 (Hitachi High-Tech.) with 189

the accelerating voltage of 0.1~8 kV, while reinforced membranes were prepared by freeze-190

breaking after cooling with liquid nitrogen. 191

192

IR spectra were obtained with a Lambda Scientific FTIR 7600 spectrometer equipped with an ATR 193

sample holder. 194

195

IEC values were obtained by UV-VIS spectroscopy of exchanged nitrate ions.[23] Membranes 196

were first immersed overnight in DI water at 80 oC, then moved to 1 M KNO3. Solutions were 197

changed every hour for 5 times. Then samples were immersed in the same concentration of KNO3 198

for another 24 hours to complete the ion exchange. Membranes were rinsed with DI water for 199

48 hours, then subsequently immersed in 20 mL of 0.1 mol L-1 NaCl for 1 hour, another 20 mL of 200

0.1 mol L-1 NaCl for 1 hour, and in 60 mL of 0.1 mol L-1 NaCl for 24 hours. At the last step, the NaCl 201

solutions were combined and analyzed by UV-vis spectroscopy. 202

203

Journ

al Pre-

proof

8

2.5 Catalyst coated membrane electrode assembly preparation 204

Membrane electrode assemblies (MEA) were prepared as catalyst coated membranes (CCM). 205

Non-platinum group metals (non-PGM) based on Ni-Fe and Mo compounds were used as anode 206

and cathode catalyst, respectively (catalyst load 2.5 mg cm-2). A chloromethylated block 207

copolymer of styrene-ethylene-butylene-styrene (PSEBS-CM) dissolved in chloroform was used 208

as a polymer binder of the catalyst layer. After the catalyst layer was deposited, PSEBS-CM was 209

functionalised by 1,4-diazabicyclo[2.2.2]octane (DABCO) functional groups.[24] 210

Prior to the catalyst layer deposition, the membrane was conditioned in the following way: 211

- immersion in 0.1 M NaOH solution for 2 hours in order to swell the membrane and to 212

transform it to the OH- form; 213

- carefully rinsing with deionized water; 214

- immersion into 0.1 M HCl solution in order to change the membrane to the Cl- form and 215

thus to improve its chemical stability prior to the CCM-MEA preparation. 216

To prepare a catalytic ink, 30 mg of catalysts were dispersed in a 5wt.% solution of PSEBS-CM-217

DABCO in chloroform. Volume of solution was calculated as to obtain a catalyst to binder weight 218

ratio (CBR) of 93/7 (45 mg of the 5wt.% solution). Subsequently, additional 9 ml of chloroform 219

was added in order to adjust the ink properties, and the ink was sonicated for 30 minutes in a 220

Bandelin Sonorex Digitec ultrasound bath. 221

222

A CNC (computerized numerical control) unit was used to disperse the ink by means of ultrasound 223

over the membrane surface and thus to produce the CCM-MEA. The membrane was heated to 224

50 °C during the catalyst deposition to accelerate the dispersant evaporation. The catalyst was 225

deposited layer by layer until the mass increase of the dry CCM-MEA corresponded to the desired 226

catalyst load. The prepared CCM-MEA was immersed into 10wt.% solution of DABCO in ethanol 227

for 12 hours in order to transfer the PSEBS-CM polymer binder into the PSEBS-CM-DABCO 228

ionomer binder. After washing the prepared CCM-MEA by deionized water, it was immersed into 229

1 M NaOH for 4 hours prior to the water electrolysis experiment. 230

231

2.6 Water electrolysis 232

Journ

al Pre-

proof

9

The water electrolysis cell consisted of a polyether ether ketone (PEEK) cell body and gold plated 233

Nickel (Ni) current feeders. The sealing was made of expanded PTFE (Fornaxa s.r.o., Czech 234

Republic). Ni foam (INCO Advanced Technology materials Co., Ltd, Dalian) with a geometrical area 235

of 4 cm2 was used as electrode on both sides of the cell. 236

237

Water electrolysis was controlled by an Autolab potentiostat M204 (Metrohm Autolab B. V., 238

Netherland) equipped with FRA module allowing electrochemical impedance measurement (EIS). 239

Short-term characteristics of the prepared CCM-MEA was obtained by recording the load curve 240

of the cell in a potentiostatic regime for the cell voltage range of 1.5 to 2.0 V with a voltage step 241

of 0.05 V. For each step, the cell current was recorded for 60 seconds. The value averaged over 242

the second half of the interval, i.e. from time 31 to 60 seconds, was taken as the cell current load 243

corresponding to the particular cell voltage. Between recording the individual current load steps, 244

the cell voltage of 1.0 V was applied for 60 seconds. This period served to allow the bubbles of 245

the produced gasses to be removed from the electrodes and from the cell as such. Every load 246

curve measurement was repeated three times to verify data reproducibility. Three 247

concentrations of KOH water solutions were used as liquid electrolytes, 0.01, 0.1 and 1.0 M KOH. 248

This allowed to investigate primarily the effect of the liquid electrolyte conductivity on the anion 249

exchange membrane water electrolysis (AEMWE) performance. Operational temperature of 250

50 °C and electrolyte flow rate of 25 ml min-1 were used. EIS was recorded to obtain additional 251

characteristics of the AEMWE cell. The EIS spectra were recorded in the frequency range of 100 252

kHz to 0.01 Hz at cell voltage of 1.8 V. The amplitude of the perturbing voltage signal was set at 253

10 mV. 254

255

A cell stability test was performed at a constant current load of 0.25 A cm-2 at 50 °C using 1.0 M 256

KOH as a liquid electrolyte. The current load was regularly interrupted in order to record the EIS 257

spectra at the cell voltage of 1.8 V, using the same EIS parameters. 258

259

EIS spectra were evaluated using the equivalent electrical circuit approach. The circuit shown in 260

Figure1 was used for this purpose. Here L stands for inductance (Henry, H) corresponding 261

Journ

al Pre-

proof

10

predominantly to the cell electric connections, Rs represents ohmic resistance of the cell (ohm, 262

Ω), including the resistance of the electric connections, current feeders, Ni foam electrodes and 263

electrolyte, Rp corresponds to the charge transfer resistance (ohm, Ω) related to the kinetics of 264

the anode and cathode reactions and CPE indicates the so called constant phase element, i.e. 265

modified capacitance of the electric double layer formed at the interfaces of the electronically 266

and ionically conductive phases (S sn, n is a number between 0 and 1). The constant phase 267

element provides a model representation of the capacitance element with spatially distributed 268

capacitance value. In some cases, the cathode reaction kinetics are significantly faster than the 269

anode one. EIS spectrum then does not show second time constant. In such a case, a simplified 270

equivalent electrical circuit consisting just of the L, Rs, Rp (anode) and CPE (anode) is used to 271

evaluate experimental data. 272

Figure 1: Equivalent electrical circuit used for EIS spectra evaluation. L – inductance (Henry,

H); Rs – resistance of the cell (ohm, Ω); Rp – charge transfer resistance (ohm, Ω); CPE –

constant phase element describing the electrical double layer capacitance (S sn).

273

2.7 Post-mortem analysis 274

A scanning electron microscope (SEM) Hitachi S4700 (Hitachi, Japan) equipped with a silicon 275

drifted detector for energy dispersive spectroscopy (EDS) (ThermoFisher Scientific, USA) was 276

used to evaluate the morphology and composition of the MEA-CCM after the electrolysis stability 277

test. Samples were cut by a scalpel. The sample for cross-section analysis was placed into a holder 278

equipped with a moving part, which pressed the sample against a fixed wall. Samples for surface 279

analysis were fixed to the plane holder by double-sided carbon tape. Prior to SEM-EDS analysis, 280

a 5 nm layer of Au/Pd was sputtered on the samples to increase the surface conductivity. The 281

accelerating voltage was 15 kV. Elemental mapping was used in order to determine the spatial 282

distribution of the selected elements. 283

Journ

al Pre-

proof

11

284

3. Results and Discussion 285

3.1 Membrane Fabrication 286

Membranes were prepared in 4 steps (Figure 2). In the first step, PBI nanofiber mats were 287

prepared by electrospinning. In the second step, mTPBr is casted on a glass plate, a PBI nanofiber 288

mat is deposited on the wet film, and after a short drying time, a second mTPBr layer is cast on 289

top. Then, by heating the membrane, a covalently bound interface between PBI and the mTPBr 290

matrix is formed by reacting PBI amine groups on the nanofiber surface with mTPBr in a 291

nucleophilic substitution reaction. In the last step, the brominated membranes are immersed in 292

trimethylamine solution to transfer all remaining bromide groups to trimethylammonium groups. 293

294

Figure 2: Schematic representation of the PBI/mTPN membrane fabrication and composition. 295

296

The first part of this work was to optimize the preparation of PBI nanofiber mats. Even though 297

such materials are very well known,[25-27] it was necessary to adjust the process parameters to 298

Journ

al Pre-

proof

12

the spinning machine. In the electrospinning process, a polymer solution is pumped through a 299

syringe, which acts as one terminal for the applied voltage. At high voltages, charges move to the 300

surface of the emerging polymer solution droplet. When the electrostatic forces reach a similar 301

magnitude as the surface tension, they deform the droplets at the needle tip into “Taylor cones”. 302

From the tip of the cones, fibers are emitted.[28-30] During their passage through air, repulsive 303

surface charges elongate the fibers, until solvent evaporation (or precipitation of polymer) 304

reaches a point at which the fiber dimensions are fixed. Finally, the fibers are deposited onto the 305

collector. The quality of the nanofibers is influenced by the voltage value. In this work, the voltage 306

was maintained in the range of 19-25 kV. When the value is adjusted below the specified range, 307

the electrostatic force is not strong enough to form the Taylor cone, and large polymer solution 308

droplets appear instead of the desired nanofiber. Otherwise, when the voltage is too high, the 309

Taylor cone is unstable [30] due to the rapid acceleration of charge. 310

311

Other factors influencing the electrospinning process are the distance between the tip and the 312

receiving plate, and temperature and humidity of the surrounding gas phase. For some systems 313

(e.g. cellulose acetate in acetone/DMAc), the fiber diameter was reported to increase with 314

increasing humidity, for other systems, it was reported to decrease (e.g. poly(vinylpyrrolidone) 315

in ethanol).[31, 32] To lower the risk of fire hazards, the machine used in this work was provided 316

with a constant nitrogen stream. The resulting low humidity led to yellow (e.g. dense PBI) areas. 317

This problem was solved by increasing the humidity in the spinning chamber to above 45%, by 318

putting a small ultrasonic humidifier in the chamber and positioning wet tissues around the 319

collector. Presumably, evaporation of DMAc is faster in humid air, because the specific heat 320

capacity of air increases with the relative humidity.[33] Another factor could be that DMAc 321

absorbs water from air, because it is hygroscopic. This should decrease the solubility of PBI. For 322

example, films cast from PBI solutions in DMAC form transparent membranes in dry atmosphere, 323

and opaque membranes in humid atmosphere.[34] 324

325

The size of the obtained nanofiber mats was in the range of 20 cm x 25 cm. SEM analysis of the 326

mats revealed that the fiber mats consisted of individual, not merged fibers (Figure 3). The fiber 327

Journ

al Pre-

proof

13

diameter was below 1 µm (0.51 ± 0.26 µm). The thickness of the prepared PBI nanofiber mats 328

was in the range of 10 µm by thickness gauge (some compression of the mat is expected during 329

the measurement) and 22 µm by SEM. It appears that handling and storage of the prepared mats 330

needs care, to avoid unwanted compression. Alternatively, the mats could also be compressed 331

or calendered to render the mats more robust, however, at the cost of porosity.[35, 36] As shown 332

in Table 1, the density of the fiber mat can be calculated from its weight and dimensions, and [1 333

- (density of porous membrane/density of non-porous PBI)] gives the porosity: 77% by thickness 334

gauge, and 90% by using the thickness obtained from SEM images. Mercury porosimetry gave a 335

porosity in a similar range (84%), and allowed to determine the average pore diameter as 2.4 µm. 336

337

338

339

Figure 3: Photograph (a) and SEM images of electrospun 20 cm x 25 cm PBI nanofiber mats. (b) 340

air side, (c) cross-section, (d) aluminium side. 341

342

a b

c dJo

urnal

Pre-pro

of

14

343

Table 1: Properties of PBI nanofiber mats. 344

Weig

ht

Are

a

Thickne

ss

(gauge)

Thickne

ss

(SEM)

ρ

(gau

ge

base

d)

ρ

(SEM

base

d)

ρ

(den

se

PBI)

Porosi

ty

(gauge

based)

Porosi

ty

(SEM

based)

Porosity

(mercury

porosimet

ry)

Average

pore

diameter

(mercury

porosimet

ry)

0.3

mg

100

m

m2

10 µm 22 µm 0.3 g

cm-3

0.13

3 g

cm-3

1.3 g

cm-3

[37]

77% 90 % 84 % 2.4 µm

345

Membranes were prepared by casting a solution of the brominated precursor polymer, 346

embedding a PBI nanofiber mat, and casting a second layer of the precursor polymer. For the 347

bromoalkyl groups to react with surface amine groups of PBI, the membranes were heated up. 348

The resulting precursor membranes are denoted as PBI/mTPBr-X.Y, with X describing the 349

thickness in µm and y describing the reaction temperature in °C. Finally, the membranes were 350

transferred into AEM by immersion in trimethylamine solution. As can be seen in Figure 4, the 351

color of the nearly white fiber mat changed into the yellow-brown color known for PBI 352

membranes, indicating that the pores were successfully filled. SEM images proved that the two 353

membrane surfaces were dense and smooth. A cross-sectional image revealed a high density of 354

small bright dots in the center of the membrane. These dots show the location of PBI nanofibers. 355

356

Journ

al Pre-

proof

15

357

Figure 4: Photograph (a) and surface (b, d) and cross-sectional (c) SEM images of a nanofiber mat-358

supported anion exchange membranes (PBI/mTPN-50.80). 359

360

Compared to the overall concentration of PBI amine groups and mTPN ammonium groups, the 361

number of PBI-mTPN connecting points is very small. Therefore, the bonding points themselves 362

cannot be well analyzed. As an indirect prove for the successful reaction between the two 363

polymers, PBI-mTPBr membranes were immersed in THF to leach out all mTPBr which did not 364

react with the PBI surface. Such treated membranes again showed the nearly white, opaque 365

appearance of the pristine PBI nanofiber mat. This indicates that no crosslinking side-reactions 366

occurred during the heating step (e.g. by Friedel-Crafts alkylation of phenyl groups in mTPBr by 367

bromoalkyl groups). Analysis by ATR-FTIR (Figure 5, and additional information in the supporting 368

information) proved that not all mTPBr was leached out, indicating that the reaction between 369

a b

c d

Journ

al Pre-

proof

16

mTPBr and PBI was successful: All strong bands of mTPBr are also found in the spectra of PBI-370

mTPBr after leaching with THF. Qualitatively, the spectra also indicate that membranes reacted 371

at 120 °C (PBI/mTPN-50.120) have a higher mTPBr-content after leaching with THF than those 372

reacted at 80 °C. Therefore, the PBI/mTPN-50.120 membranes appear to have a stronger 373

supportmatrix interface than PBI/mTPN-50.80 membranes. 374

375

376

Figure 5: ATR-FTIR spectra of PBI, mTPBr and the composite membrane after leaching mTPBr out 377

with THF (24 h, room temperature); (a) PBI/mTPN-50.80, (b) PBI/mTPN-50.120. Arrows indicate 378

the bands from mTPBr which are still seen after leaching. 379

380

Because the sampling depth in XPS is just about 10 nm, spectra of PBI/mTPBr-50.80 and -50.120 381

after leaching with THF should not show nitrogen signals from PBI, but only signals related to 382

Journ

al Pre-

proof

17

mTPBr. A uniform coating PBI nanofibers with a layer of mTPBr was confirmed by XPS. 383

(Supporting Information, Figure S2). 384

385

3.2 Swelling and conductivity 386

The most important membrane properties are swelling in the area (which can delaminate 387

membrane and electrode and thus reduce the lifetime of the electrolyser) and ion conductivity. 388

Figure 6 shows that the conductivity of PBI-mTPN is always higher than that of dense PBI 389

membranes, which are practically insulating in pure water and in low concentrated KOH solutions. 390

At higher KOH concentrations, PBI gets deprotonated, and potassium polybenzimidazolide is 391

formed. Because of their ionic nature, such membranes easily absorb additional water molecules 392

and potassium and hydroxide ions, rendering them conductive, and the conductivity increases 393

with increasing KOH concentration, up to about 20 wt% (around 4M).[38, 39] 394

395

For ion exchange membranes, two opposing factors need to be considered. (a) With increasing 396

KOH concentration, the osmotic pressure difference between the ionic membrane and the 397

surrounding water phase decreases, reducing the water uptake and thus the conductivity. (b) 398

Although AEMs are generally considered as single ion conductors, which selectively conduct 399

anions, Donnan exclusion is only strictly observed for membranes in very diluted solutions.[40, 400

41] When the KOH concentration increases, also the concentration of potassium ions and (for 401

electroneutrality) hydroxide ions in the membrane increases, which results in increasing 402

conductivity. This means that the conductivity should show an initial drop when the KOH 403

concentration increases, and then an increasing conductivity, because both PBI[39] and mTPN 404

absorb additional potassium and hydroxide ions. This expected trend is observed for all 405

membranes. For PBI/mTPN composite membranes, the resistance should further decrease when 406

the KOH concentration is high enough to render the PBI support conductive, and the 407

conductivities of reinforced and non-reinforced membranes indeed reach similar values in 4M 408

KOH. In comparison to commercial Fumatech FAA3-50 and FAA3-PK-75, the mPBI/mTPN-50.120 409

membrane shows superior conductivity at all tested KOH concentrations. 410

411

Journ

al Pre-

proof

18

Interestingly, the conductivity of PBI/mTPN-50.80 is lower than that of PBI/mTPN-50.120. 412

Hypothetically, this is because evaporation of THF during membrane formation cools down the 413

polymer solution and thus promotes absorption of water from air. At 80 °C, this water is not fully 414

removed and can degrade bromoalkyl groups to alcohol groups. At 120 °C, water traces are 415

efficiently evaporated, and have no time to react with bromoalkyl groups. This should result in 416

lower IEC values for mTPN-50.80, and indeed the IEC values of mTPN-50.80 and mTPN-50.120 417

were 1.69 ± 0.04 and 1.94 ± 0.02 mmol chloride g-1, respectively (Table 2). The IEC of PBI/mTPN-418

50.120 was 1.57 mmol chloride g-1. Assuming a density of 1.2 g cm-3 for mTPN, and a thickness of 419

50 µm, 1 cm2 membrane contains 6 mg mTPN and 0.3 mg PBI. Thus, the expected IEC is 5% lower 420

than 1.94 if no reaction with PBI occurs, and additional 16% lower if all PBI NH groups react with 421

mTPN, i.e. 1.54 mmol chloride g-1. Therefore, the experimental IEC value is in a reasonable range, 422

and supports that some of the bromomethyl groups reacted with amine groups on the PBI surface, 423

possibly accompanied by hydrolysis, because PBI in contact with ambient air contains some water. 424

425

Table 2: IEC values of used materials and membranes 426

Membrane or material IEC

[mmol Cl (g dry membrane)-1]

Comment

mTPN 2.17 Based on chemical structure

PBI 0 Based on chemical structure

mTPN-50.80 1.69 Based on analysis

mTPN-50.120 1.94 Based on analysis

PBI/mTPN-50.120 1.57 Based on analysis

427

428

The conductivity of mTPN-50.120 in pure water (in-plane) was 72 mS cm-1. This is significantly 429

higher than the 54 mS cm-1 reported by Lee et al.[8], presumably because the samples in that 430

work were not purged electrochemically[21,22] and therefore may still have contained traces of 431

(bi)carbonate. A similar observation was also made for other AEM types.[21] 432

Journ

al Pre-

proof

19

433

Figure 6: Conductivity of membranes (in hydroxide form, a) and length swelling of the best 434

performing membrane type (in chloride form, b); data for dense PBI in DI water is from 435

literature[42]; conductivity data for water were measured in-plane, all other data were measured 436

through-plane; (c,d) uptake and swelling in KOH solutions. *The negative thickness swelling 437

values and the related volume swelling values seem to stem from increased waviness in the dry 438

state. 439

440

The swelling behavior of the chloride form membranes was only analyzed for membranes heated 441

at 120 °C, because (a) they showed highest conductivity and (b) FTIR data indicated better 442

interaction between PBI and mTPBr. PBI/mTPN-50.120 swelled 37% less at room temperature 443

and 25% less at 60 °C than mTPN-50.120 (Figure 6b). If membranes are assembled in the wet 444

state and never allowed to dry, the main issue is temperature induced swelling (wet, RT 60 °C), 445

a b

c

in-plane through-plane conductivity

mTPN-50.120

PBI/mTPN-50.120FAA3-50

FAA3-PK-75

0

2

4

6

8

10

12

14

16

18

Le

ng

th s

we

llin

g (

%)

@RT

@60 oC

d

FAA3-50 0.1M 1M KOH

FAA3-PK-75 0.1M 1M KOH

Uptake SR length SR thickness SR volume-10

0

10

20

30

40

50

60

(%)

Uptake SR length SR thickness SR volume-10

0

10

20

30

40

50

60

(%)

mTPN-50.120 0.1M 1M KOH

PBI/mTPN-50.120 0.1M 1M KOH

mTPN-50.120

PBI/mTPN-50.120

mTPN-50.80

PBI/mTPN-50.80

FAA3-50

FAA3-PK-75

dense PBI

0 1 2 3 4

0

20

40

60

80co

nd

uctivity (

mS

cm

-1)

KOH concentration (M)

**

Journ

al Pre-

proof

20

which was 2% for mTPN-50.120 and 2.7% for PBI/mTPN-50.120. In summary, while mTPN-50-120 446

may be considered to be slightly better under optimal conditions, introduction of the PBI 447

nanofiber mats renders the membrane more robust, e.g. if electrolysis cells are not assembled 448

under fully humidified conditions, or if the cells run dry during maintenance work. 449

450

Swelling in KOH solution transfers the membrane into the hydroxide form. In general, the length 451

swelling of increases in the order Cl form (water) < OH form, 0.1M KOH < OH form, 1M KOH 452

(Figure 6c). For mTPN and PBI/mTPN, length swelling is very similar at room temperature in 0.1 453

M KOH and 1M KOH and at 60 °C in the Cl form, indicating that length swelling is restricted to 454

about 10 %. Because the PBI fiber mat does not cover the whole membrane volume, but only a 455

layer in the center of the membrane, it is no surprise that length swelling is restricted, whereas 456

thickness swelling is not much affected – it is practically same for reinforced and non-reinforced 457

membranes in 1M KOH. While thickness swelling can have the unwanted effect that the catalyst 458

layer is pressed into the porous transport layer, strong length swelling and shrinking can be 459

detrimental for the membrane, because expansion leads to increased waviness and shrinking 460

leads to mechanical stresses, the result being cracks, pinholes, and catalyst-membrane 461

delamination. Therefore, the low length and thickness swelling of PBI/mTPN-50.120 is beneficial. 462

463

In comparison with the commercial materials FAA3-50 and FAA3-PK-75 (Figure 6d), the effect of 464

the PBI support on the restriction of swelling is weaker than the effect of the PK support, which 465

is a strong fiber mash. The negative thickness swelling and therefore also very low volume 466

swelling of FAA3-PK-75 was observed by two co-authors and for different samples, and also in 467

the chloride for. Most probably, some waviness in the dry state increased the apparent thickness. 468

469

3.3. Mechanical properties 470

Membranes experience mechanical stresses when they swell and shrink due to changes in 471

humidity and temperature, when the cell is operated under differential pressure (conditions 472

under which the membrane and catalyst layer are pushed into the pores of the porous transport 473

layer) and during handling. Commercial membranes show tensile strength values of 30 MPa 474

Journ

al Pre-

proof

21

(Orion TM1, i.e. mTPN),[8, 43] 25-40 MPa (Fumatech FAA3), 45 MPa (reinforced FAA3), 60 MPa 475

(Aemion membranes) and 96 MPa (Tokuyama’s A201, which is not produced anymore); for 476

Sustainion membranes, values are not reported, because they get very brittle in the dry state and 477

have to be handled in the wet state.[7] 478

479

Figure 7: (a) Tensile strength, Young’s Modulus and elongation at break at ambient and wet (30 °C) 480

conditions; chloride form, measured at 28 °C and 22 %relative humidity, FAA3-50: 26 °C and 481

74 %relative humidity. (b) Stress-strain curves for PBI/mTPN-50.120 and commercial FAA3-PK-75 482

FAA3-PK-75 (old, degraded batch)

b

c

a

Young Modulus (MPa)

Elongation at Break (%)

Tensile Strength (MPa)

mTPN-50.120(wet)

PBI/mTPN-50.120

(wet)

FAA3-50(wet)

FAA3-PK-75(wet)

0

50

100

200

400

600

Journ

al Pre-

proof

22

bought in 2021 (new batch, 74 %rh) and 2017 (old batch, 22 %rh), and (c) photograph of FAA3-483

PK-75 after testing. 484

485

At 28 °C and 22 % relative humidity, mTPN-50.120 showed a tensile strength of 42 MPa (Figure 486

7a). Reinforcement with a PBI nanofiber mat increased the tensile strength of mTPN by 17%. A 487

larger effect was seen for the Young Modulus, which increased 56% from 436 MPa to 680 MPa 488

for PBI/mTPN-50.120. For comparison, Nafion 212 has a Young Modulus of just 160 MPa.[44] The 489

elongation at break decreased from 16% for mTPN to 11% for PBI/mTPN-50.120. FAA3 showed a 490

lower tensile strength than the value found in the literature, possibly because the humidity was 491

above 70% during the measurement. An even stronger effect was observed for all membranes, 492

when the wet samples were analyzed. In general, tensile strength and Young modulus decreased 493

in the wet state, and elongation at break increased. A slightly different behavior is seen for FAA3-494

PK-75, for which the PK support strongly influences the properties, nearly irrespective of the 495

humidity. While the tensile strength of wet FAA3 was 8.6 MPa and that of wet mTPN was 4.9 496

MPa, it was 17.5 MPa for wet PBI/mTPN, and 23 - 26 MPa for FAA3-PK-75 in both wet and dry 497

state. A clear improvement was seen for the Young Modulus in the wet state. The reinforcement 498

increased it from 75 MPa for mTPN (wet) to 240 MPa for reinforced PBI/mTPN-50.120 (wet). 499

500

Figure 7b shows the stress-strain curves of a slightly thinner PBI/mTPN-50.120 membrane (43 501

µm for this sample) and commercial FAA3-PK-75, which is a 75 µm thick AEM reinforced with a 502

porous poly(ether ether ketone). PBI/mTPN has a 2-3 times higher tensile strength than FAA3-503

PK-75. While PBI/mTPN and new FAA3-PK-75 show smooth curves and the membrane has the 504

same appearance before and after testing, an old batch of FAA3-PK-75 which was stored for over 505

4 years cracked in several places, before the sample broke. The reason for this could be 506

degradation of the poly(ether ketone) fiber mash support. But Fumatech also advises that drying 507

of wet membranes may lead to micro-cracks. Therefore, minute repeated swelling and shrinking 508

due to changes in ambient humidity over the years may also have induced micro-cracks. 509

Comparing the photograph of the broken sample (Figure 7c) and the number of spikes in the 510

Journ

al Pre-

proof

23

stress-strain curve reveals that each spike represents a cracking event (see also videos in 511

supporting information). 512

513

In summary, it can be stated that the reinforced membranes show a sufficient tensile strength 514

and Young Modulus. In future work, further improvements seem to be possible by using thicker 515

fiber mats or embedding more than one fiber mat, so that the ratio of support to matrix increases. 516

We expect that this would increase the tensile strength and Young Modulus, and further 517

decrease the length swelling. Due to the high porosity of the PBI nanofiber mat, the negative 518

effect on conductivity should be tolerable. 519

520

3.4 Anion exchange membrane water electrolysis 521

Figure 8 shows the AEMWE performance of PBI/mTPN-50.120 and the evaluated EIS parameters 522

for three different KOH concentrations (0.01, 0.1, 1 M). The comparison of the PBI/mTPN-50.120 523

and mTPN membrane is shown in Figure SX. It confirms the only small effect of the PBI presence 524

on the AEMWE performance. The load curves in Figure 8a indicate strong dependence of the cell 525

performance on the KOH concentration. The highest current density of 0.172 A cm-2 was achieved 526

for the highest concentration of the liquid electrolyte, i.e. for 1M KOH. Direct comparison of the 527

obtained data with the literature is difficult due to the different experimental conditions used by 528

different authors. Nevertheless, according to the review published recently by Hamish et al.[45] 529

current densities ranging from 0.007 to 0.98 A cm-2 at the cell voltage of 1.8 V are reported by 530

the authors for KOH concentrations < 1 M. In the case of the present CCM-MEA based on the 531

PBI/mTPN-50.120 membrane, the maximum current density at cell voltage of 1.8 V is about 0.045 532

A cm-2 with 1 M KOH as liquid electrolyte. However, as pointed out by Hamish et al.,[45] the high 533

current densities reported are typically achieved using PGM catalysts.[46-49] When considering 534

non-PGM catalysts exclusively, the performance of the PBI/mTPN-50.120 membrane is 535

comparable with the literature data.[50-53] Despite the fact that PBI membranes typically need 536

KOH concentrations higher than 1 M KOH,[39, 54-58] Liu at al.[59] applied a Celazole® PBI 537

membrane in their comparative study and achieved a current density of 0.035 A cm-2 at a cell 538

Journ

al Pre-

proof

24

voltage of 1.8 V with 1 M KOH liquid electrolyte at 60 °C. It is thus possible that the PBI support 539

is not fully inert and partially contributes to the AEMWE performance. 540

541

The impact of the increased KOH concentration on the MEA performance was analysed by EIS 542

(Figure 8b). As clearly evident from the figure, increasing KOH concentrations improves both the 543

cell (Rs) and polarisation (Rp) resistances. Rs includes the contributions from several parts of the 544

electrolysis cell, but is primarily dominated by the resistances of the liquid and solid electrolytes. 545

Therefore, changes in Rs do not strictly follow the trend given by the conductivity of the KOH 546

feed solution. While the conductivity of the KOH solution increased by two orders of magnitude 547

from 0.0032 to 0.29 S cm-1 (at 50 °C) when the KOH concentration increased from 0.01 to 1 M, 548

Rs values decreased just by a factor of 16. This is connected with the solid polymer electrolyte, 549

i.e. the anion exchange membrane and ionomer. As demonstrated in Figure 6, the ionic 550

conductivity of the membrane is only partially influenced by the composition of the surrounding 551

liquid electrolyte. In sum, at lower KOH concentrations, Rs is dominated by the liquid electrolyte; 552

at higher KOH concentrations, Rs is limited by the conductivity of the solid electrolyte, and only 553

negligible changes of Rs are observed when the KOH concentration is further increased in this 554

regime.[50,60] 555

It is, however, Rp, which clearly determines the AEMWE performance. When less concentrated 556

KOH solutions are used as a liquid electrolyte, one has Rp >> Rs. This is primarily caused by the 557

limited ionic conductivity of the polymer binder ensuring ionic contact between the three 558

dimensional structure of the catalyst layer and the membrane. Liquid electrolyte of sufficient 559

conductivity can improve this ionic contact and thus improve catalyst utilisation and performance. 560

At lower concentration, the reduced catalyst utilisation increases the local current density in the 561

active regions of the catalyst layer. Consequently, Rp increases. In the case of the 1 M KOH liquid 562

electrolyte, fitting revealed that only one time constant contributed to the EIS spectrum. The 563

oxygen evolution reaction (OER) is generally characterised by the sluggish kinetics when 564

compared to hydrogen evolution reaction (HER). This is also documented by the values of Rp 565

shown in Figure 8b, where HER in general exhibits lower values of polarisation resistance. As the 566

Journ

al Pre-

proof

25

extent of the catalyst layer utilisation increases with increasing KOH concentration, HER kinetics 567

improve to such an extent that they become overlapped in the EIS spectrum by the OER. 568

569

570 571

572

Figure 8: Characteristics of the AEMWE in terms of (A) load curves and (B) system (Rs) and 573

polarisation (Rp) resistances. Operating conditions: PBI/mTPN-50.120 membrane; PSEBS-CM-574

DABCO polymer binder; 2.5 mg cm-2 Ni-Fe and 2.5 mg cm-2 Mo based compounds were used as 575

anode and cathode catalysts respectively; CBR 93/7; Ni foam electrodes of geometrical area 4 576

cm2; 50 °C; liquid KOH electrolyte of concentrations given in the figures (A, B). EIS measured at a 577

cell voltage of 1.8 V. 578

579

Finally, the durability test of the PBI/mTPN-based CCM-MEA was investigated by its galvanostatic 580

polarisation at 0.25 A cm-2. The results are summarised in Figure 9. Figure 9a documents good 581

stability of the cell voltage. The average value of the cell voltage over the 200 hours experiment 582

falls into the interval of (1.98 ± 0.02) V. This corresponds to a fluctuation of ± 1.1 %. At the average 583

value of the cell voltage during stability test short term load curves showed current density of 584

only 0.17 A cm-2. The difference is due to the different regimes of the electrode polarisation, i.e. 585

potentiostatic vs. galvanostatic regime. Observable periodic voltage oscillations are mainly due 586

to changes in the KOH concentration caused by the continuous water depletion by electrode 587

reactions together with evaporation and its regular replenishment to its original concentration. 588

Journ

al Pre-

proof

26

When water is consumed, the KOH concentration increases. This reduces the cell resistance and 589

improves access of hydroxide ions to the anodic catalyst sites. As a result, cell voltage decreases. 590

When the KOH concentration is re-adjusted to 1 M KOH by adding appropriate amount of water, 591

the voltage returns to the original value. 592

593

Further characteristics of the AEMWE cell were again obtained by EIS. Since 1 M KOH solution 594

was used as a liquid electrolyte, the EIS spectrum showed only one time constant and thus, only 595

OER Rp was determined. Figure 9b shows the evolution of Rs and Rp values (fitted to the 596

experimental data) over the time of electrolysis. While the variation of Rs values over time is 597

practically negligible, (0.08 ± 0.01) Ω, Rp was only stable for the first 80 hours. After exceeding 598

this time, it started to rise. The polarisation resistance increase agrees well with the rise of the 599

average cell voltage observed for the same time. Variation of both Rp and cell voltage correspond 600

well to each other confirming the strong impact of the Rp on the cell voltage and at the same 601

time stability of the membrane. The observed Rp fluctuations can be explained by local KOH 602

concentration changes and/or by partial blockage of the catalyst surface by evolving gas bubbles. 603

Journ

al Pre-

proof

27

Figure 9: Stability of the PBI/mTPN-based CCM-MEA in terms of (A) cell voltage and (B) EIS

results. Operating conditions: PBI/mTPN-50.120 membrane; PSEBS-CM-DABCO polymer

binder; 2.5 mg cm-2 Ni-Fe and 2.5 mg cm-2 Mo based compounds as anode and cathode

catalysts respectively; CBR 93/7; Ni foam electrodes of geometrical area 4 cm2; 50 °C; 1 M

KOH liquid electrolyte; 0.25 A cm-2 current density. EIS measured at 1.8 V.

After the stability test, the morphology of the mPBI/mTPN-50.120-based CCM-MEA and the 604

distribution of the catalyst metals over the examined cross-sectional area was analysed by SEM-605

EDS (Figure 10 and Figure S3). The minor presence of the catalyst elements in the membrane is 606

most probably due to the sample preparation method by scalpel cutting. The thicknesses of the 607

catalyst layers are 10 and 4 µm for cathode and anode, respectively. Presumably, some of the 608

catalyst sticked to the porous transport layer when the cell was disassembled (see SEM-EDS 609

analysis of the fresh sample in Figure SX1). But it is nicely seen that neither cathode nor anode 610

side show signs of membrane/electrode delamination even after 200 hours of electrolysis 611

operation and sample preparation by mechanically cutting the MEA. 612

Journ

al Pre-

proof

28

Figure 10: SEM-EDS-analysis of a cross-section of the PBI/mTPN-50.120-based CCM-MEA

after 200 hours electrolysis operation. Acceleration voltage was 15 kV; 600x magnification.

Even though Rp started to increase after 80 hours of electrolysis operation, the cell voltage 613

showed a stable value with only 1.1 % fluctuation around its average value during the 200 hours 614

electrolysis stability test. This proves good stability of the prepared CCM-MEA. In addition, the 615

performance of the water electrolysis during the stability test is promising: The cell voltage 616

reached 1.98 V at 0.25 A cm-2 in 1 M KOH at 50 °C. For comparison, Park et al. reported that a 617

water electrolyser operating with a non-reinforced mTPN membrane (0.5 M NaOH, 50 °C, each 618

electrode 2 mg cm-2 PGM catalyst, AS-4 ionomer binder) required a voltage of 2.2 V to operate 619

at 0.25 A cm-2 (iV curve) and 0.2 A cm-2 at steady-state operation.[10] In comparison to other 620

reported systems using commercial membranes, the performance of 1.98 V at 0.25 A cm-2 does 621

Journ

al Pre-

proof

29

not exceed the state of art materials (see supporting information, Table S1), current densities in 622

the range of up to 500 mA cm-2 can be obtained even without noble metal catalysts.[7] However, 623

the commercially available AEM FAA-3-50 measured under the same conditions achieved the 624

comparable performance of 0.275 A cm-2 at cell voltage 2 V, see Figure S5. It should be stressed 625

that this work was not focused on the cell performance optimization, but on the proof of concept 626

that membranes with an engineered covalently bonded matrix/support interface can be 627

fabricated. Better performance can be reached by further optimization of the MEA setup 628

(matching of membrane and binder) and cell operational conditions targeted to reduce the 629

polarization resistance contributing mainly to the performance losses. 630

631

For future work, we intend to further optimize the PBI nanofiber mats’ robustness, for example 632

by calendering, solvent welding to increase the number of connection points between the fibers, 633

or crosslinking. Second, we plan to test other ion conducting matrices – the reinforcement, 634

especially after increasing its robustness, could well allow the use of materials which are cheaper 635

to produce and have higher IEC and thus conductivity. 636

637

4. Conclusions 638

In this work, polybenzimidazole nanofiber mats were prepared by electrospinning. In a second 639

step, the bromomethylated precursor of highly alkaline-stable mTPN was filled into the pores of 640

PBI nanofiber mats. During drying, covalent bonds were established between the PBI support and 641

the mTPBr matrix. A leaching test in THF supported the formation of covalent bonds. After 642

quaternisation to PBI/mTPN, the reinforced membranes showed the expected behavior: 37% 643

reduced length swelling, 17% increased tensile strength, and 56% increased Young’s Modulus. 644

During the test, PBI/mTPN-50.120 did not change its appearance. This indicated that the PBI 645

nanofiber mat efficiently transfers forces into all directions, and that the interaction between 646

support and matrix is strong. The hydroxide conductivity in DI water decreased only from 72 mS 647

cm-1 to 62 mS cm-1. Finally, a MEA prepared with Ni/Fe and Mo-based catalysts (no PGM) and a 648

SEBS-based binder showed a stable voltage of 1.98V at 250 mA cm-2 in an anion exchange 649

Journ

al Pre-

proof

30

membrane water electrolysis test (1M KOH, 50 °C.), which lasted 200 hours. Furthermore, EIS 650

showed no significant change in the membrane resistance. 651

652

In sum, the PBI/mTPN-50.120-based CCM-MEA utilising PSEBS-CM-DABCO as ionomer binder in 653

the catalyst layer and non-PGM Ni-Fe and Mo based catalysts represents a promising candidate 654

for AEMWE utilising diluted KOH as a liquid electrolyte. 655

656

Acknowledgements 657

This work was done as part of the NEWELY project, and has received funding from NRF (Korea) 658

and from the Fuel Cells and Hydrogen 2 Joint Undertaking under Grant No. 875118. This Joint 659

Undertaking receives support from the European Union’s Horizon 2020 Research and Innovation 660

programme, Hydrogen Europe and Hydrogen Europe research. CB thanks the US Department of 661

Energy ARPA-E (IONICS DE-Ar0000769) for financial support. 662

663

Competing Interests 664

The authors declare no competing financial interest. 665

666

Supporting Information: 667

Videos of stress-strain tests for aged FAA3-PK-75 and PBI/mTPN-50.120. 668

Additional SEM images, XPS data and information on the ATR-FTIR measurements, literature 669

comparison of commercial membranes, and electrolysis data for mTPN and FAA3-50. 670

671

References 672

[1] D.G. Caglayan, N. Weber, H.U. Heinrichs, J. Linßen, M. Robinius, P.A. Kukla, D. Stolten, 673

Technical potential of salt caverns for hydrogen storage in Europe, Int. J. Hydrogen Energy, 45 674

(2020) 6793-6805. https://doi.org/10.1016/j.ijhydene.2019.12.161. 675

[2] M. Huo, The Current Status of Hydrogen and Fuel Cell Development in China, J. Electrochem. 676

Energy Conv. Stor. 17 (2020) 034001. http://dx.doi.org/10.1115/1.4045702. 677

Journ

al Pre-

proof

31

[3] J. Brauns, J. Schönebeck, M. R. Kraglund, D. Aili, J. Hnát, J. Žitka, W. Mues, J.O. Jensen, K. 678

Bouzek, T. Turek, Evaluation of Diaphragms and Membranes as Separators for 679

Alkaline Water Electrolysis, J. Electrochem. Soc. 168 (2021) 014510. 680

https://doi.org/10.1149/1945-7111/abda57 681

[4] M. Carmo, D.L. Fritz, J. Mergel, D. Stolten, A comprehensive review on PEM water electrolysis, 682

Int. J. Hydrogen Energy 38 (2013) 4901 - 4934. http://dx.doi.org/10.1016/j.ijhydene.2013.01.151683

684

[5] https://www.bmwi.de/Redaktion/EN/Publikationen/Energie/the-national-hydrogen-685

strategy.pdf, assessed 2020.09.11. 686

[6] http://www.platinum.matthey.com/documents/new-687

item/pgm%20market%20reports/pgm_market_report_may_19.pdf, assessed 2020.09.11. 688

[7] D. Henkensmeier, Malikah Najibah, C. Harms, J. Žitka, J. Hnát, K. Bouzek, Overview: State-of-689

the Art Commercial Membranes for Anion Exchange Membrane Water Electrolysis, J. 690

Electrochem. Energy Conv. Stor. 18 (2021) 024001. http://dx.doi.org/10.1115/1.4047963. 691

[8] W.-H. Lee, E.J. Park, J. Han, D.W. Shin, Y.S. Kim, C. Bae, Poly(terphenylene) Anion Exchange 692

Membranes: The Effect of Backbone Structure on Morphology and Membrane Property, ACS 693

Macro Lett. 6 (2017) 566−570. http://dx.doi.org/10.1021/acsmacrolett.7b00148. 694

[9] K.M. Meek, C.M. Antunes, D. Strasser, Z.R. Owczarczyk, A. Neyerlin, B.S. Pivovar, High-695

Throughput Anion Exchange Membrane Characterization at NREL, ECS Trans. 92 (2019), 723–731. 696

[10] E.J. Park, C.B. Capuano, K.E. Ayers, C. Bae, Chemically Durable Polymer Electrolytes for Solid-697

State Alkaline Water Electrolysis, J. Power Sour. 375 (2018) 367–372. 698

http://dx.doi.org/10.1016/j.jpowsour.2017.07.090. 699

[11] T.T.L. Yu, S.H. Liu, H.L. Lin, P.H. Su, Nafion/PBI Nanofiber Composite Membranes for Fuel 700

Cells Applications, Proceedings of the ASME 8th international conference on fuel cell science, 701

engineering, and technology 2010, vol 2, pages 631-639. 702

[12] D. Jones, J. Rozière, S. Cavaliere, S. Subianto, S. Burton, “Membrane”, Applicants: Johnson 703

Matthey Fuel Cells Limited, Centre National del la Recherche Scientifique, Universite Montpellier 704

2, Patent WO 2016/020668 A1, first filed 2015.08.04. 705

Journ

al Pre-

proof

32

[13] D. Jones, Reinforced PFSA Membranes with Exceptional Durability, at “Polymers for fuel cells, 706

energy storage, and conversion”, Pacific Grove, USA, 2019.02.24-27. 707

[14] R. Sood, S. Cavaliere, D.J. Jones, J. Rozière, Electrospun nanofiber composite polymer 708

electrolyte fuel cell and electrolysis membranes, Nano Energy 26 (2016) 729–745. 709

http://dx.doi.org/10.1016/j.nanoen.2016.06.027. 710

[15] D. Aili, M.K. Hansen, C. Pan, Q. Li, E. Christensen, J.O. Jensen and N.J. Bjerrum, Phosphoric 711

acid doped membranes based on Nafion®, PBI and their blends – Membrane preparation, 712

characterization and steam electrolysis testing, Int. J. Hydr. Energy 36 (2011) 6985 - 6993. 713

https://doi.org/10.1016/j.ijhydene.2011.03.058. 714

[16] M. Jung, W. Lee, N. Nambi Krishnan, S. Kim, G. Gupta, L. Komsiyska, C. Harms, Y. Kwon, D. 715

Henkensmeier, Porous-Nafion/PBI composite membranes and Nafion/PBI blend membranes for 716

vanadium redox flow batteries, Appl. Surf. Sci. 450 (2018) 301–311. 717

https://doi.org/10.1016/j.apsusc.2018.04.198. 718

[17] S.J. Cho, H. Choi, J.H. Youk, Evaluation of PBI Nanofiber Membranes as a High-temperature 719

Resistance Separator for Lithium-ion Batteries, Fibers Polym. 21 (2020) 993-998. 720

https://doi.org/10.1007/s12221-020-9955-z 721

[18] S. Lee, K.S. Han, M. Kim, M.C. Kim, C.V. Anh, J. Nah, Polybenzimidazole-Benzophenone 722

Composite Nanofiber Window Air Filter with Superb UV Resistance and High Chemical and 723

Thermal Durability, ACS Appl. Mater. Interf. 12 (2020) 23914-23922. 724

https://doi.org/10.1021/acsami.0c03868 725

[19] S. Lee, A.R Cho, D. Park, J.K. Kim, K.S. Han, I.J. Yoon, M.H. Lee, J. Nah, Reusable 726

Polybenzimidazole Nanofiber Membrane Filter for Highly Breathable PM2.5 Dust Proof Mask, 727

ACS Appl. Mater. Interf. 11 (2019) 2750-2757. https://doi.org/10.1021/acsami.8b19741 728

[20] A.M. Park, F.E. Turley, R.J. Wycisk, P.N. Pintauro, Electrospun and Cross-Linked Nanofiber 729

Composite Anion Exchange Membranes, Macromolecules 47 (2014) 227-235. 730

https://doi.org/10.1021/ma401932h 731

[21] N. Ziv, D.R. Dekel, A practical method for measuring the true hydroxide conductivity of anion 732

exchange membranes, Electrochem. Commun. 88 (2018) 109–113. 733

https://doi.org/10.1016/j.elecom.2018.01.021. 734

Journ

al Pre-

proof

33

[22] A. Zhegur-Khais, F. Kubannek, U. Krewer, D.R. Dekel, Measuring the true hydroxide 735

conductivity of anion exchange membranes, J. Membr. Sci. 612 (2020) 118461. 736

https://doi.org/10.1016/j.memsci.2020.118461. 737

[23] F. Karas, J. Hnát, M. Paidar, J. Schauer, K. Bouzek, Determination of the ion-exchange 738

capacity ofanion-selective membranes, Int. J. Hydrogen Energy 39 (2014) 5054-5062. 739

http://dx.doi.org/10.1016/j.ijhydene.2014.01.074. 740

[24] J. Hnát, M. Plevova, R.A. Tufa, J. Zitka, M. Paidar, K. Bouzek, Development and testing of a 741

novel catalyst-coated membrane with platinum-free catalysts for alkaline water electrolysis, Int. 742

J. Hydrogen Energy 44 (2019) 17493-17504. https://doi.org/10.1016/j.ijhydene.2019.05.054. 743

[25] J.-S. Kim, D.H. Reneker, Polybenzimidazole Nanofiber Produced by Electrospinning, Polym. 744

Eng. Sci. 39 (1999) 849-854. https://doi.org/10.1002/pen.11473. 745

[26] S. Anandhan, K. Ponprapakaran, T. Senthil, G. George, Parametric study of manufacturing 746

ultrafine polybenzimidazole fibers by electrospinning, Int. J. Plast. Technol. 16 (2012) 101–116. 747

https://doi.org/10.1007/s12588-012-9036-2. 748

[27] S. Jahangiri, İ. Aravi, L. I. Şanlı, Y. Z. Menceloğlu, E. Özden‐Yenigün, Fabrication and 749

optimization of proton conductive polybenzimidazole electrospun nanofiber membranes, Polym. 750

Adv. Technol. 2018 (29) 594–602. https://doi.org/10.1002/pat.4169. 751

[28] M. Yu, R.H. Dong, X. Yan, G.F., Yu, M.H. You, X. Ning, Y.Z. Long, Recent Advances in Needleless 752

Electrospinning of Ultrathin Fibers: From Academia to Industrial Production, Macromol. Mater. 753

Eng. 302 (2017) 1700002. http://dx.doi.org/10.1002/mame.201700002. 754

[29] N.A.M. Barakat, M. El-Newehy, S.S. Al-Deyab, H.Y. Kim, Cobalt/copper-decorated carbon 755

nanofibers asnovel non-precious electrocatalyst for methanol electrooxidation, Nanoscale Res. 756

Lett. 9 (2014) 2. http://dx.doi.org/10.1186/1556-276X-9-2 757

[30] T. Mazoochi, M. Hamadanian, M. Ahmadi, V. Jabbari, Investigation on the morphological 758

characteristics of nanofiberous membrane as electrospun in the different processing parameters, 759

Int. J. Ind. Chem. 3 (2012) 2. https://doi.org/10.1186/2228-5547-3-2. 760

[31] S. De Vrieze, T. Van Camp, A. Nelvig, B. Hagström, P. Westbroek, K. De Clerck, The effect of 761

temperature and humidity on electrospinning, J. Mater. Sci. 44 (2009) 1357-1362. 762

https://doi.org/10.1007/s10853-008-3010-6. 763

Journ

al Pre-

proof

34

[32] J. Pelipenko, J. Kristl, B. Janković, S. Baumgartner, P. Kocbek, The impact of relative humidity 764

during electrospinning on the morphology and mechanical properties of nanofibers, Pharm. 765

Nanotechn. 456 (2013) 125-134. https://doi.org/10.1016/j.ijpharm.2013.07.078. 766

[33] P.T. Tsilingiris, Thermophysical and transport properties of humid air at temperature range 767

between 0 and 100°C, Energy Conv. Man. 49 (2008) 1098-1110. 768

https://doi.org/10.1016/j.enconman.2007.09.015. 769

[34] X. Li, G. Qian, X. Chen, B.C. Benicewicz, Fuel Cells 13 (2013) 832-842. 770

https://doi.org/10.1002/fuce.201300054. 771

[35] S. Nauman, G. Lubineau, H.F. Alharbi, Post processing strategies for the enhancement of 772

mechanical properties of enms (electrospun nanofibrous membranes): A review, Membranes 11 773

(2021) 39. http://dx.doi.org/10.3390/membranes11010039. 774

[36] T. Wakiya, M. Tanaka, H. Kawakami, Fabrication and electrolyte characterizations of 775

nanofiber framework-based polymer composite membranes with continuous proton conductive 776

pathways, Membranes 11 (2021) 90. http://dx.doi.org/10.3390/membranes11020090 777

[37] X. Li, Structure-Property Relationships in Polybenzimidazole Materials for Gas Separation 778

and Fuel Cell Applications. (Doctoral dissertation), 2014. Retrieved from 779

https://scholarcommons.sc.edu/etd/2671 780

[38] D. Aili, J. Yang, K. Jankova, D. Henkensmeier, Q. Li, From polybenzimidazoles to 781

polybenzimidazoliums and polybenzimidazolides, J. Mater. Chem. A 8 (2020) 12854-12886. DOI: 782

10.1039/d0ta01788d. 783

[39] A. Konovalova, H. Kim, S. Kim, A. Lim, H. S. Park, M. R. Kraglund, D. Aili, J. H. Jang, H.‐J. Kim, 784

D. Henkensmeier, Blend membranes of polybenzimidazole and an anion exchange ionomer 785

(FAA3) for alkaline water electrolysis: Improved alkaline stability and conductivity, J. Membr. Sci. 786

564 (2018) 653–662. https://doi.org/10.1016/j.memsci.2018.07.074. 787

[40] J. Kamcev, D.R. Paul and B.D. Freeman, Ion Activity Coefficients in Ion Exchange Polymers: 788

Applicability of Manning’s Counterion Condensation Theory, Macromolecules 48 (2015) 8011-789

8024. http://dx.doi.org/10.1021/acs.macromol.5b01654. 790

Journ

al Pre-

proof

35

[41] A. Münchinger, K.-D. Kreuer, Selective ion transport through hydrated cation and anion 791

exchange membranes I. The effect of specific interactions, J. Membr. Sci. 592 (2019) 117372. 792

https://doi.org/10.1016/j.memsci.2019.117372. 793

[42] J.A. Asensio, E.M. Sanchez and P. Gomez-Romero, Proton-conducting membranes based on 794

benzimidazole polymers for high-temperature PEM fuel cells. A chemical quest, Chem. Soc. Rev. 795

39 (2010) 3210 – 3239. https://doi.org/10.1039/b922650h. 796

[43] S. Noh, J. Y. Jeon, S. Adhikari, Y. S. Kim, C. Bae, Molecular Engineering of Hydroxide 797

Conducting Polymers for Anion Exchange Membranes in Electrochemical Energy Conversion 798

Technology, Acc. Chem. Res. 52 (2019) 2745-2755. 799

https://doi.org/10.1021/acs.accounts.9b00355. 800

[44] N. Nambi Krishnan, D. Henkensmeier, J. H. Jang, H.-J. Kim, V. Rebbin, I.-H. Oh, S.-A. Hong, S.-801

W. Nam, T.-H. Lim, Sulfonated poly(ether sulfone)-based silica nanocomposite membranes for 802

high temperature polymer electrolyte fuel cell applications, Int. J. Hydrogen Energy 36 (2011) 803

7152-7161. http://dx.doi.org/10.1016/j.ijhydene.2011.03.015. 804

[45] H.A. Miller, K. Bouzek, J. Hnat, S. Loos, C.I. Bernäcker, T. Weißgärber, L. Röntzsch, J. Meier-805

Haack, Green hydrogen from anion exchange membrane water electrolysis: a review of recent 806

developments in critical materials and operating conditions, Sust. Energy Fuels 4 (2020) 2114-807

2133. https://doi.org/10.1039/C9SE01240K. 808

[46] M.K. Cho, H.-Y. Park, H.J. Lee, H.-J. Kim, A. Lim, D. Henkensmeier, S.J. Yoo, J.Y. Kim, S.Y. Lee, 809

H.S. Park, J.H. Jang, Alkaline anion exchange membrane water electrolysis: Effects of electrolyte 810

feed method and electrode binder content, J. Power Sour. 382 (2018) 22-29. 811

https://doi.org/10.1016/j.jpowsour.2018.02.025. 812

[47] A. Lim, H.J. Kim, D. Henkensmeier, S. Jong Yoo, J. Young Kim, S. Young Lee, Y.E. Sung, J.H. 813

Jang, H.S. Park, A study on electrode fabrication and operation variables affecting the 814

performance of anion exchange membrane water electrolysis, J. Ind. Eng. Chem. 76 (2019) 410-815

418. https://doi.org/10.1016/j.jiec.2019.04.007. 816

[48] X. Wu, K. Scott, CuxCo3-xO4 (0 ≤ x < 1) nanoparticles for oxygen evolution in high 817

performance alkaline exchange membrane water electrolysers, J. Mater. Chem. 21 (2011) 12344-818

12351. https://doi.org/10.1039/C1JM11312G. 819

Journ

al Pre-

proof

36

[49] A. Marinkas, I. Struźyńska-Piron, Y. Lee, A. Lim, H.S. Park, J.H. Jang, H.-J. Kim, J. Kim, A. 820

Maljusch, O. Conradi, D. Henkensmeier, Anion-conductive membranes based on 2-mesityl-821

benzimidazolium functionalised poly(2,6-dimethyl-1,4-phenylene oxide) and their use in alkaline 822

water electrolysis, Polymer 145 (2018) 242-251. https://doi.org/10.1016/j.polymer.2018.05.008. 823

[50] J. Hnát, M. Plevová, J. Žitka, M. Paidar, K. Bouzek, Anion-selective materials with 1,4-824

diazabicyclo[2.2.2]octane functional groups for advanced alkaline water electrolysis, Electrochim. 825

Acta 248 (2017) 547-555. https://doi.org/10.1016/j.electacta.2017.07.165. 826

[51] E. López-Fernández, J. Gil-Rostra, J.P. Espinós, A.R. González-Elipe, F. Yubero, A. de Lucas-827

Consuegra, CuxCo3-xO4 ultra-thin film as efficient anodic catalysts for anion exchange 828

membrane water electrolysers, J. Power Sour. 415 (2019) 136-144. 829

https://doi.org/10.1016/j.jpowsour.2019.01.056. 830

[52] S.H. Ahn, B.S. Lee, I. Choi, S.J. Yoo, H.J. Kim, E. Cho, D. Henkensmeier, S.W. Nam, S.K. Kim, 831

J.H. Jang, Development of a membrane electrode assembly for alkaline water electrolysis by 832

direct electrodeposition of nickel on carbon papers, Appl. Cat. B: Environm. 154-155 (2014) 197-833

205. https://doi.org/10.1016/j.apcatb.2014.02.021. 834

[53] L. Zeng, T.S. Zhao, Integrated inorganic membrane electrode assembly with layered double 835

hydroxides as ionic conductors for anion exchange membrane water electrolysis, Nano Energy, 836

11 (2015) 110-118. https://doi.org/10.1016/j.nanoen.2014.10.019. 837

[54] J. Brauns, J. Schönebeck, M.R. Kraglund, D. Aili, J. Hnát, J. Žitka, W. Mues, J.O. Jensen, K. 838

Bouzek, T. Turek, Evaluation of Diaphragms and Membranes as Separators for Alkaline Water 839

Electrolysis, J. Electrochem. Soc. 168 (2021) 014510-014524. https://doi.org/10.1149/1945-840

7111/abda57. 841

[55] L.A. Diaz, J. Hnát, N. Heredia, M.M. Bruno, F.A. Viva, M. Paidar, H.R. Corti, K. Bouzek, G.C. 842

Abuin, Alkali doped poly (2,5-benzimidazole) membrane for alkaline water electrolysis: 843

Characterization and performance, J. Power Sour. 312 (2016) 128-136. 844

https://doi.org/10.1016/j.jpowsour.2016.02.032. 845

[56] D. Aili, A.G. Wright, M.R. Kraglund, K. Jankova, S. Holdcroft, J.O. Jensen, Towards a stable 846

ion-solvating polymer electrolyte for advanced alkaline water electrolysis, J. Mater. Chem. A 5 847

(2017) 5055-5066. https://doi.org/10.1039/C6TA10680C. 848

Journ

al Pre-

proof

37

[57] D. Aili, M.K. Hansen, J.W. Andreasen, J. Zhang, J.O. Jensen, N.J. Bjerrum, Q. Li, Porous 849

poly(perfluorosulfonic acid) membranes for alkaline water electrolysis, J. Membr. Sci. 493 (2015) 850

589-598. https://doi.org/10.1016/j.memsci.2015.06.057. 851

[58] M.R. Kraglund, D. Aili, K. Jankova, E. Christensen, Q. Li, J.O. Jensen, Zero-gap alkaline water 852

electrolysis Using ion-solvating polymer electrolyte Membranes at reduced KOH concentrations, 853

J. Electrochem. Soc. 163 (2016) F3125-F3131. http://dx.doi.org/10.1149/2.0161611jes. 854

[59] Z. Liu, S.D. Sajjad, Y. Gao, H. Yang, J.J. Kaczur, R.I. Masel, The effect of membrane on an 855

alkaline water electrolyzer, Int. J. Hydrogen Energy, 42 (2017) 29661-29665. 856

https://doi.org/10.1016/j.ijhydene.2017.10.050. 857

[60] X. Su, L. Gao, L. Hu, N.A. Qaisrani, X. Yan, W. Zhang, X. Jiang, X. Ruan, G. He, Novel 858

piperidinium functionalized anionic membrane for alkaline polymer electrolysis with excellent 859

electrochemical properties, J. Membr. Sci. 581 (2019) 283-292. 860

https://doi.org/10.1016/j.memsci.2019.03.072. 861

Journ

al Pre-

proof

Highlights

A bromomethylated polymer was filled into the pores of PBI nanofiber mats

Curing led to covalent bonds between porous support and ion conducting matrix

In stress-strain tests, no cracks appeared before failure

Water electrolyzer: 1.98V@0.25 A/cm-2 (50 °C, 1M KOH, catalysts based on Ni,Fe,Mo)

200 hours stability test: voltage and membrane resistance were stable

Journ

al Pre-

proof

Declaration of interests The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper. The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

Journ

al Pre-

proof