Implementing MOF-derived Nickel Catalysts as Anion Exchange Membrane Fuel Cell Anode

EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476255 B0203 Page 1/10 B0203 Implementing MOF-derived Nickel Catalysts as Anion Exchange Membrane Fuel Cell Anode Frédéric Jaouen* (1), Simon Amigues (1), Nicolas Bibent (1), Eliot Petitdemange (1), Laetitia Dubau (2), Frédéric Maillard (2) (1) ICGM, Univ. Montpellier, CNRS, ENSCM, 34293 Montpellier/France; (2) Univ. Grenoble Alpes, Univ. Savoie Mont Blanc, CNRS, Grenoble INP, LEPMI, 38000 Grenoble/France; *Contact corresponding authors: www.EFCF.com/ContactRequest Abstract Due to their acidic environment, the present generation of proton-exchange membrane fuel cells (PEMFC) requires precious metal catalysts, in particular platinum, which is an obstacle to their large-scale deployment. In contrast, anion-exchange membrane fuel cells (AEMFCs) operate at high pH, facilitating the use of sustainable precious metal free catalysts. While high performance has been reached with platinum-group-metal (PGM) free cathodes in AEMFCs, the replacement of PGM-based anodes by PGM-free ones is a challenge. Nickelbased catalysts are the most promising PGM-free hydrogen oxidation reaction (HOR) catalysts in alkaline medium, but nickel suffers from early surface passivation at potentials above ca 0.1 V vs. the reversible hydrogen electrode (RHE), which in turn blocks the HOR. In order to mitigate the surface oxidation of nickel at HOR potentials, two main approaches are studied in the literature: optimization of nickel intrinsic properties by alloying with other earth abundant elements, or core@shell nanostructuration of nickel by a protective carbon shell. We will report on recent results obtained with this second approach, and the preparation of core@shell Ni@NC catalysts, comprising a nitrogen-doped carbon shell, via the annealing of Ni-based MOFs. The latter were obtained either by autoclave or by a solvent-free mechanochemical synthesis. Our observations confirm the significant effect of pyrolysis atmosphere (NH3, H2, N2 and combinations) on the HOR activity. Other synthesis parameters will be shown to play a key role on the MOF structure and morphology as well as on the HOR activity of final materials. The materials were characterized by X-ray diffraction, electron microscopy, nitrogen adsorption, X-ray photoelectron spectroscopy (XPS), and electrochemical techniques, including AEMFC tests. The presentation will correlate the material’s structure to their activity and stability. In particular, it will be shown that Ni-based anodes require a break-in protocol to show any performance in AEMFC. Left) TEM micrograph Showing the nickel-core N-C shell. Right) Example of Performance achieved with a Ni anode and Pt/C cathode in AEMFC 20 nm 0100 200 300 400 500 600 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Voltage (V) Current Density (mA cm-2) 226 mW cm-2 H2/O2 - 100kPa - Ni@NC - Pt/C 0 50 100 150 200 250 Power Density (mW cm-2) EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476255 B0203 Page 2/10 Introduction Due to their acidic environment, the present generation of proton-exchange membrane fuel cells (PEMFC) requires precious metal catalysts, in particular platinum, which is an obstacle to their large-scale deployment. In contrast, anion-exchange membrane fuel cells (AEMFCs) operate at high pH, facilitating the use of sustainable precious metal free catalysts [1]. While high performance has been reached with platinum-group-metal (PGM) free cathodes in AEMFCs 2-5, the replacement of PGM-based anodes by PGM-free ones is a challenge 1, 6. Nickel-based catalysts are the most promising PGM-free hydrogen oxidation reaction (HOR) catalysts in alkaline medium, but nickel (Ni) suffers from early surface passivation at potentials above ca 0.1 V vs. the reversible hydrogen electrode (RHE), which in turn blocks the HOR. In order to mitigate the surface oxidation of Ni at HOR potentials, two main approaches are studied in the literature: optimization of Ni intrinsic properties by alloying with other earth abundant elements 7, or core@shell nanostructuration of Ni by a protective carbon shell 8, 9-10. 1. Scientific Approach The present work focuses on the synthesis, characterization and implementation at the anode of AEMFCs of Ni-based catalysts comprising a metallic Ni core and a nitrogen-doped carbon shell (referred to as NC). The work leans on our strong experience in developing FeN-C catalysts synthesized from the thermal transformation of metal organic frameworks (MOFs) [5], and on the recent work from Yushan Yan’s and Xile Hu’s groups, where a NiMOF was shown to spontaneously form Ni@NC nanostructures upon annealing, leading to high HOR activity [8]. We adopted a mechanochemistry approach for a fast and soft synthesis method for synthesizing Ni-MOFs. X-ray diffraction (XRD) shows that different NiMOFs are produced compared to the autoclave method adopted in [8], and that different NiMOFs are obtained using different Ni salts, even when the nature of the ligand and the Ni/ligand ratio are fixed. After an optimization of different parameters of the synthesis, a highly active Ni@NC catalyst was obtained, which was implemented at the anode of an AEMFC. We report on the issue of Ni-based catalysts oxidation/passivation and the related break-in issue in AEMFC. Developing a break-in protocol is a necessary step without which an AEMFC comprising a Ni-based anode simply does not deliver any power. 2. Experiments 2.1. Ni@NC synthesis Ni@NC catalysts were synthetized in two steps. First, a MOF precursor was synthetized by ball milling, via an adapted reported method [11]. Briefly, a given mass of the ligand (1,3,5benzenetricarboxylic acid, H3BTC) and of Ni(II) acetate tetrahydrate was added in the grinding bowl so as to result in a molar ratio of Ni/Ligand of 3/2. A low amount of water was also introduced in the bowl and the mix was milled for 30 min at 600 rpm. When different Ni salts were used (Ni(C2H3O2)2‧4 H2O) (Ni(II) acetylacetonate Ni(C5H7O2)2, Ni(II) chloride hexahydrate Ni(Cl)2‧6H2O, Ni(II) nitrate hexahydrate Ni(NO3)2‧6 H2O, Ni(II) sulfate hexahydrate Ni(SO4)‧6 H2O), the mass of Ni salt was modified so as to keep the same molar ratio of Ni/ligand of 3/2. The MOF product was extracted with anhydrous ethanol and centrifugated. The product was then dried overnight. The second step consists in a pyrolysis of the MOF by an adapted reported method [8]. 550 mg of Ni-MOF was deposited in a quartz crucible and introduced in the centre of a quartz tube. The tube was connected to a gas flow line, sealed and placed in a tubular oven. The oven was first purged with N2 flow for 5 min, EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476255 B0203 Page 3/10 and heated to 175°C at a ramp of 10 °C min-1 under a 400 mL min-1 flowrate of N2/H2 (95:5, v/v). Beyond 175 °C, the gaseous atmosphere varied for different syntheses, but was either a N2/NH3 mix, a N2/H2 mix or a N2/H2/NH3 mix. To form the N2/NH3 mix (80:20, v/v), 100 mL min-1 of pure NH3 and 400 mL min-1 of pure N2 was introduced whereas the H2/N2 flowrate was turned off. To form the N2/H2/NH3 mix (76:4:20, v/v/v), 100 mL min-1 of pure NH3 was introduced in addition to the flow of 400 mL min-1 of N2/H2 (95:5, v/v). The furnace was then heated to 390°C at 10 °C min-1 under one of the gas mix described above. The temperature was maintained at 390°C for X min (X = 30, 60, 90 or 120 min) before cooling down under N2 atmosphere for 30 min. 2.2. Electrochemical characterization in rotating disk electrode The catalytic ink was prepared by mixing 4.0 mg of catalyst powder, 0.182 mL of isopropanol, 0.358 mL of ultrapure water, 4.7 μL of Nafion solution (5 wt.%) and dispersed by ice-cooled ultrasonication. The experimental setup consisted of a PTFE based electrochemical cell with a three-electrode setup, utilizing a commercial RHE (Gaskatel HydroFlex) as the reference electrode, and a glassy carbon disk electrode tip as a support for the HOR catalytic film (working electrode). The electrolyte was 0.1 M KOH (≥85%). 8 µL of the ink was deposited to achieve 0.3 mgNi@NC cm-2 for all catalysts. All data was recorded at a rotating speed of 1600 rpm. 2.3 Gas diffusion electrode preparation for AEMFC Gas diffusion electrodes (GDEs) were prepared by spray deposition. The ink was prepared in a Agate mortar and pestle. Typically, the anion-exchange ionomer (AEI) was first grinded, and then the catalytic power was added and grinded for 5 min. For Ni@NC anodes, 25 wt.% of Vulcan XC-72R carbon black was added and grinded 5 minutes. Next, 2.5 ml of ultrapure water was added and grinded for 5 more minutes. Then, 5 ml of isopropanol was added slowly while continuously grinding. 17.5 ml of isopropanol was used to completely transfer the slurry from the mortar to a glass vial. The slurry was dispersed for 1 hour in an ice-cooled ultrasonic bath prior to spraying. The GDEs were obtained by hand spray deposition of the catalytic ink onto 25 cm2 of gas diffusion layers (GDL) using a plate heated to 70°C to evaporate the solvent. Lastly, the prepared GDEs were cut into four 4.84 cm2 samples suitable for the fuel cell hardware. The AEI used was ETFE poly-VBC TMA grafted powder [12]. The GDL was from Toray (TGP-H-060 with 5 wt.% PTFE wet proofing). 2.4 Testing in AEMFC The AEMFCs were controlled by an 850 Fuel Cell Test System (Scribner) and the electrochemical measurements were recorded with a SP-150 potentiostat (Bio-Logic) associated to a booster (VMP3B-20, 20A/20V) and monitored using the EC-Lab software (Bio-Logic). The AEM was from Versogen (PiperION™). It was ion-exchanged by at least 24 h of immersion in a N2-bubbled 1.0 M KOH aqueous solution. Before assembling the fuel cell hardware (Scribner), the membrane and the GDEs were soaked in 1.0 M KOH for 60 min while changing the 1.0 M KOH aqueous solution every 20 min. The anode and cathode GDEs and AEM were mounted in the fuel cell hardware with 5 cm2 serpentine flow field and PTFE gaskets. After assembly, the cell was heated to 60°C under 0.3 L min-1 of N2 on both sides and the cell and humidifier’s temperature was stabilized for 1 h. 3. Results EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476255 B0203 Page 4/10 In a first step, we studied the effect of the nature of the Ni salt on the structure of the resulting MOFs prepared by ballmilling, and on the activity of the resulting Ni@NC catalysts after annealing, while keeping all other synthesis conditions fixed. The five Ni salts were Ni(II) acetylacetonate, Ni(II) chloride, Ni(II) nitrate, Ni(II) sulfate, some containing also water molecules as described in section 2.1. Figure 1 shows that the XRD patterns of the resulting MOFs are different for each Ni salt, except for Ni(II) acetate tetrahydrate and Ni(II) acetylacetonate sharing the same XRD pattern. According to a literature search, only the structure corresponding to the XRD pattern of these two MOFs has been reported and identified, corresponding to a Ni3(BTC)2•12 H2O formula/structure 13. Figure 1. XRD patterns of the MOFs prepared by planetary ballmilling of Ni salt and H3BTC, changing only the nature of the Ni salt used for their preparation. In brief, these results highlight the profound changes of the nature and/or hydration level of the Ni salts on the MOF structures obtained by planetary ballmilling, with further work needed to identify some of the obtained crystalline structures. The Ni@NC catalysts obtained by annealing these different MOFs in a N2/H2/NH3 atmosphere at 390 °C also show very different activities for the HOR in alkaline medium, as seen in Figure 2. While the catalysts derived from MOFs prepared from Ni acetate and Ni acetylacetonate show high and similar HOR activity, those prepared from Ni sulfate and Ni chloride show almost no activity and the catalyst derived from the Ni-MOF prepared from Ni nitrate shows intermediate HOR activity. Therefore, there seems to be a correlation between the Ni-MOF structure and the derived catalyst’s properties, as suggested by the same XRD patterns of the Ni-MOFs made from Ni acetylacetonate and Ni acetate and their similar HOR activities. A precise identification of the MOF structures is however needed to identify the Ni-MOF descriptor(s) leading to high activity. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476255 B0203 Page 5/10 Figure 2. HOR polarization curves of the Ni@NC catalysts derived by annealing at 390 °C the MOFs prepared from different Ni salts. Annealing in N2/H2/NH3 atmosphere at 390 °C for 1 h. Measurements performed at room temperature, in H2-saturated 0.1 M KOH electrolyte, rotation rate 1600 rpm, catalyst loading 0.3 mg cm-2, scan rate 5 mV s-1. The curves are corrected for Ohmic drop. In a third step we investigated the effect of the annealing atmosphere on the HOR activity, fixing the temperature to 390 °C and 1 h duration and fixing the Ni salt to Ni(II) acetate. Three atmospheres were studied, labeled as N2/NH3, N2/H2, and N2/H2/NH3. The XRD patterns of the catalysts are similar and identify only the metallic Ni fcc phase (Figure 3). XPS reveals increasing N content in the order N2/H2 (0.5 at %), N2/H2/NH3 (4.2 at. %) and N2/NH3 (7.4 at %). While nitrogen (from NH3) seems thus critical for high activity, the HOR activity does not simply correlate with the total N content, as the catalyst annealed in N2/H2/NH3 is more active than the one annealed in N2/NH3 (see later). The type of N functionalities that are present also differ, as highlighted by the fitting of the N1s narrow scan spectra (Figure 4). The HOR activities of the three resulting catalysts are significantly different, with the Ni@NC prepared in N2/H2/NH3 demonstrating the highest activity (in the microkinetic region of ca 0.00-0.05 V vs. RHE), followed by the one prepared in N2/NH3, while the one prepared in N2/H2 has low HOR activity (Figure 5). The results suggest a correlation between the N content in the Ni@NC and their HOR activity. The effect of these atmospheres on the HOR activity is in line with a previous publication 1. Figure 3. XRD patterns of the Ni@NC catalysts prepared by annealing at 390 °C for 1 h a same Ni-MOF (prepared from Ni acetate) in three different annealing atmospheres. EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476255 B0203 Page 6/10 Figure 4. XP narrow scan spectra of the N1s region of the Ni@NC catalysts prepared by annealing at 390 °C for 1 h a same Ni-MOF (prepared from Ni acetate) in three different annealing atmospheres. Figure 5. HOR polarization curves of the Ni@NC catalysts derived by annealing at 390 °C for 1 h a same Ni-MOF (prepared from Ni acetate) in three different annealing atmospheres. Electrochemical measurements performed at room temperature, in H2-saturated 0.1 M KOH electrolyte, rotation rate 1600 rpm, catalyst loading 0.3 mg cm-2, scan rate 5 mV s-1. The curves are corrected for Ohmic drop. In a fourth step, we investigated the effect of the pyrolysis duration at 390 °C, with a fixed annealing atmosphere (N2/H2/NH3) and fixed Ni salt (Ni acetate). The annealing duration once the oven reaches 390 °C was varied from 30 to 120 min. The shape of the polarization curves was similar and akin to the red curve in Figure 5, but the slopes in the microkinetic region varied, revealing differences in HOR activity. Figure 6 reports the HOR activity trend as a function of the annealing duration, with the HOR activity assessed from the linear slope in the microkinetic region of the polarization curves. The figure shows an important effect of the annealing duration and that a necessary duration at 390 °C is required to reach high HOR activity. With the investigated durations, an optimum is seen at 90 min, although the EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476255 B0203 Page 7/10 true optimum may be found between 30 and 120 min. XRD of the catalysts did not reveal a trend between the average particle size of metallic Ni and the HOR activity (average Ni crystallite size of 13 nm after 30 min, 17 nm after 60 min, 15 nm after 90 min and 17 nm after 120 min). We are currently investigating if the electrochemical surface area of Ni correlates with the HOR activity. Its measurement is however not trivial due to the presence of a carbon shell on top of the Ni particles. Figure 6. Effect of annealing duration on the HOR activity of the Ni@NC catalysts derived by annealing at 390 °C in N2/H2/NH3 atmosphere a same Ni-MOF (prepared from Ni acetate). Electrochemical measurements performed at room temperature, in H2-saturated 0.1 M KOH electrolyte, rotation rate 1600 rpm, catalyst loading 0.3 mg cm-2, scan rate 5 mV s-1. The curves are corrected for Ohmic drop. The HOR activity was assessed from the fitted slope of the polarization curve in the microkinetic region of -0.02 to + 0.02 V vs. RHE. With the optimized catalyst from the present study (Ni acetate salt, N2/H2/NH3 atmosphere, 90 min annealing duration), we then worked on its implementation at the anode of anionexchange membrane fuel cells (AEMFCs). A recurrent and dramatic issue observed regardless of the anode preparation, ionomer/catalyst ratio or type of AEM used, was the very low open circuit voltage (OCV, i.e. the cathode - anode difference) of single-cell AEMFC fed with pure H2 at anode and pure O2 at cathode, when the anode was based on the Ni@NC material (Figure 7). As seen in Figure 7, the OCV only transiently increases up to ca 0.6 V when switching from N2/N2 to H2/O2 feeds and then slowly decreases to ca 0.2 V. This is not due to the setup but specific to the Ni@NC anode, as the same experiment performed with a state of art PtRu/C anode shows a fast increase and steady OCV of ca 1.03 V achieved after only 3 min of gas switch. As the cathode is based on a Pt/C catalyst, the results suggest that the Ni@NC anode keeps a high OCP of ca 0.7-0.8 V vs. RHE in such conditions, instead of reaching 0 V vs. RHE as expected for an anode comprising an HOR-active catalyst. In turn, this can be explained if the Ni@NC catalyst is not HOR-active in such conditions, which might be the case if the Ni surface is passivated by a thin Ni(hydr)oxide layer. Ni oxidation is a recognized issue and known to block the HOR in alkaline medium. The HOR polarization curves in RDE (such as Figure 2) show a decline in current above 0.3 V vs. RHE assigned to Ni surface oxidation/passivation, and the observed plateau at E < 0.3 V vs. RHE has a lower value than expected for HOR at 1600 rpm in such conditions (ca 3 mA cm-2), suggesting that even at low potential, the Ni surface may be partially oxidized. Thus, for a Ni@NC anode at 0.7-0.8 V vs. RHE, the Ni surface is expectedly fully oxidized and thus inactivated for the HOR. The high OCP of Ni@NC thin films is also observed initially in RDE, EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476255 B0203 Page 8/10 but the presence of a reference electrode in such a setup allows to easily activate the Ni@NC thin film by applying a low potential of 0 V vs. RHE or even slightly negative. This however cannot be done in an AEMFC where only the cell voltage is measured. Figure 7. The issue of open circuit voltage of AEMFC with Ni-based anode. Pt/C cathode, O2 and H2 flows on cathode and anode, 60 °C cell temperature. The anode was either prepared from a PtRu/C commercial catalyst or from the optimized Ni@NC catalyst from the present study. Anode: Ni@NC prepared from Ni acetate, annealing in N2/H2/NH3 atmosphere for 90 min, 7.0 mg Ni@NC cm-2, AEM was PiperION-A20-HCO3 from Versogen, and cathode was based on 40 % Pt/C, 0.2 mgPt cm-2. PtRu/C anode was based on 40%Pt and 20%Ru on carbon, 0.5 mgPtRu cm-2. The presentation will show how the Ni@NC anode can be activated in AEMFC through different ways, and that the initially oxidized surface state of Ni is responsible for the low OCV (high anode OCP) issue. Figure 8 shows a representative AEMFC polarization curve obtained at our laboratory with a Ni@NC anode. Figure 8. AEMFC polarization curve of Ni@NC from Ni acetate salt and annealed 90 min. Anode loading 6.8mgcata cm-2. Cell temperature was 60 °C, 100% RH and 100 kPa gauge for both anode and cathode, H2 and O2 400 mL/min flowrate, scan rate of 5 mV s-1, the 0100 200 300 400 500 600 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Voltage (V) Current Density (mA cm-2) 226 mW cm-2 H2/O2 - 100kPa - Ni@NC - Pt/C 0 50 100 150 200 250 Power Density (mW cm-2) EFCF 2025: Low-Temp. Fuel Cells, Electrolysers & H2 Processing 1 – 4 July 2025, Lucerne Switzerland https://doi.org/10.5281/zenodo.17476255 B0203 Page 9/10 anion-exchange membrane was Versogen PiperION-A60-HCO3 60µm, cathode: 40 % Pt/C, 0. 22 mgPt cm-2. Acknowledgements The authors acknowledge funding by the French National Research Agency through the DEEP project (grant number ANR-21-CE05-0021). F.J., E.P and N.B also acknowledge financial support from the Clean Hydrogen Partnership and its members under the grant agreement SUSTAINCELL (Project No. 101101479). We thank John Varcoe (Univ. Surrey, UK) and his group for providing the ETFE-based anion-exchange ionomer used for preparing the GDEs. Disclaimer: Co-funded by the European Union. Views and opinions expressed are however those of the author(s) only and do not necessarily reflect those of the European Union or the Clean Hydrogen Partnership. Neither the European Union nor the granting authority can be held responsible for them. References [1] Horie Adabi Firouzjaie and William E. Mustain, Catalytic advantages, challenges, and priorities in alkaline membrane fuel cells, ACS Catal. 10 (2020) 225-234. [2] Ying Wang et al, Synergistic Mn-Co catalyst outperforms Pt on high-rate oxygen reduction for alkaline polymer electrolyte fuel cells. Nature Commun. 10 (2019) 1506. [3] K. Ehelebe et al, Fuel cell catalyst layer evaluation using a gas diffusion electrode half cell: oxygen reduction reaction on Fe-N-C in alkaline media, Electrochem. Commun. 116 (2020) 106761. [4] Horie Adabi et al, High-performing commercial Fe-N-C cathode electrocatalyst for anion-exchange membrane fuel cells, Nature Energy 6 (2021) 834-843. [5] Hori Adabi, Pietro G. Santori, et al, Understanding how single-atom site density drives the performance and durability of PGM-free Fe-N-C cathodes in anion exchange membrane fuel cells, Materials Today Adv. 12 (2021) 100179. [6] Elena S. Davydova, Sanjeev Mukerjee, Frédéric Jaouen, Dario R. Dekel, Electrocatalysts for hydrogen oxidation reaction in alkaline electrolytes, ACS Catalysis 8 (2018) 6665-6690. [7] Aaron Roy et al, Nickel-copper supported on carbon black hydrogen oxidation catalyst integrated into an anion exchange membrane fuel cell, Sustain..Energy Fuels 2 (2018) 2268-2275. [8] Weiyan Ni et al, An efficient nickel hydrogen oxidation catalyst for hydroxide exchange membrane fuel cells, Nature Mater. 21 (2022) 804-810. [9] Yunfei Gao et al, Improving the antioxidation capability of the Ni catalyst by carbon shell coating for alkaline hydrogen oxidation reaction, ACS Appl. Mater. Interfaces 12 (2020) 31575-31581. [10] J. Wang et al, Ultrafine nickel nanoparticles encapsulated in N-doped carbon promoting hydrogen oxidation reaction in alkaline media, ACS Catal. 11 (2021) 74227428. [11] Ren Zhang et al, Ultrafast synthesis of Ni-MOF in one minute by ball milling, Nanomaterials 8 (2018) 1067. [12] S. D. Poynton et al, Preparation of radiation-grafted powders for use as anion exchange ionomers in alkaline polymer electrolyte fuel cells, J. Mater. Chem. A 2 (2014) 5124-5130.