The Performance of Nickel and Nickel-Iron Catalysts Evaluated As Anodes in Anion Exchange Membrane Water Electrolysis

catalysts

Article The Performance of Nickel and Nickel-Iron Catalysts Evaluated As Anodes in Anion Exchange Membrane Water Electrolysis

Emily Cossar 1, Alejandro Oyarce Barnett 2,3,*, Frode Seland 4 and Elena A. Baranova 1,*

1 Department of Chemical and Biological Engineering, Centre for Catalysis Research and Innovation (CCRI), University of Ottawa, 161 Louis-Pasteur Ottawa, ON K1N 6N5, Canada; [email protected] 2 SINTEF Industry, Sustainable Energy Technology Department, New Energy Solutions Group, NO-7491 Trondheim, Norway 3 Department of Energy and Process Engineering, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway 4 Department of Materials Science and Engineering, Norwegian University of Science and Technology, NO-7491 Trondheim, Norway; [email protected] * Correspondence: [email protected] (A.O.B.); [email protected] (E.A.B.); Tel.: +47-9300-3263 (A.O.B.); +(613)-562-5800 (ext. 6302) (E.A.B.) Received: 29 August 2019; Accepted: 24 September 2019; Published: 27 September 2019

Abstract: Anion exchange membrane water electrolysis (AEMWE) is an efficient, cost-effective solution to renewable energy storage. The process includes oxygen and hydrogen evolution reactions (OER and HER); the OER is kinetically unfavourable. Studies have shown that nickel (Ni)- iron (Fe) catalysts enhance activity towards OER, and cerium oxide (CeO2) supports have shown positive effects on catalytic performance. This study covers the preliminary evaluation of Ni, Ni90Fe10 (at%) and Ni90Fe10/CeO2 (50 wt%) nanoparticles (NPs), synthesized by chemical reduction, as OER catalysts in AEMWE using commercial membranes. Transmission electron microscopy (TEM) images of the Ni-based NPs indicate NPs roughly 4–6 nm in size. Three-electrode cell measurements indicate that Ni90Fe10 is the most active non-noble metal catalyst in 1 and 0.1 M KOH. AEMWE measurements of 2 the anodes show cells achieving overall cell voltages between 1.85 and 1.90 V at 2 A cm− in 1 M KOH at 50 ◦C, which is comparable to the selected iridium-black reference catalyst. In 0.1 M KOH, 2 the AEMWE cell containing Ni90Fe10 attained the lowest voltage of 1.99 V at 2 A cm− . Electrochemical impedance spectroscopy (EIS) of the AEMWE cells using Ni90Fe10/CeO2 showed a higher ohmic resistance than all catalysts, indicating the need for support optimization.

Keywords: nickel; iron; ceria; OER; alkaline exchange membrane; electrolysis; anode

1. Introduction As global warming and climate change concerns continue to rise, the concept of a “hydrogen economy” is becoming more and more important. This ideal is based on using hydrogen (H2) as a clean, renewable fuel [1]. H2 can also be used to store renewable energy through water electrolysis [2]. Water electrolysis utilizing anion exchange membranes (AEMs) is an emerging water electrolysis technology, used for its ability to produce hydrogen both efficiently and at low cost. Compared to traditional alkaline water electrolysis, which employ porous diaphragm separators, solid polymer electrolytes may provide certain advantages, such as lower gas crossover, improved efficiency, higher current densities, differential pressure operation and improved operation dynamics [3]. Unlike other solid polymer electrolyser technologies, such as proton exchange membrane water electrolysis (PEMWE), anion exchange membrane water electrolysis (AEMWE) technology has the potential to

Catalysts 2019, 9, 814; doi:10.3390/catal9100814 www.mdpi.com/journal/catalysts Catalysts 2019, 9, 814 2 of 17

Catalysts 2019, 9, x FOR PEER REVIEW 2 of 17 operate without expensive noble-metal catalysts, such as iridium, ruthenium and platinum, in addition 45 toplatinum, low-cost materials in addition for bipolar to low- platescost materials and current for collectors. bipolar plates AEMWE, and therefore,current collectors. aims to combine AEMWE, 46 thetherefore, low costs aims of alkaline to combine electrolysis the low costs with of the alkaline high e ffielectrolysisciency and with flexibility the high of efficiency the proton and exchange flexibility 47 membraneof the proton (PEM) exchange electrolysis membrane design (PEM) [3]. electrolysis design [3]. 48 TheThe theoretical theoretical thermodynamic thermodynamic potential potential for fo waterr water electrolysis electrolysis is is 1.23 1.23 V V at at room room temperature. temperature. 49 ToTo achieve achieve a lowa low activation activation overvoltage overvoltage during during electrolysis’s electrolysis’s operation, operation, high high performing performing oxygen oxygen and and 50 hydrogenhydrogen evolving evolving catalysts catalysts are are required. required. As As the the goal goal of of water water electrolysis electrolysis is is its its industrialization, industrialization, 51 catalystcatalyst cost cost is imperative.is imperative. As As such, such, the the development development of of active active non-noble non-noble metal metal catalysts catalysts is is crucial crucial to to 52 furtherfurther develop develop and and establish establish AEMWE AEMWE technology. technology The. The number number of of studies studies addressing addressing performance performance 53 improvementsimprovements through through the the development development of of new new AEM AEM materials, materials, catalysts catalysts and and membrane membrane electrode electrode 54 assembliesassemblies (MEAs) (MEAs) have have increased increased in in recent recent years years [4 –[413–].13]. However, However, the the water water splitting splitting performance performance 55 reportedreported for for AEMWE AEMWE is still is still lower lower than than that ofthat PEMWE of PEMWE [14,15 [14,15],], particularly particularly when employingwhen employing non-noble non- 56 metalnoble catalysts metal catalysts and lower and concentrations lower concentrations of alkaline of solutions,alkaline solutions, or water or [16 water]. [16]. 57 TheThe water water electrolysis electrolysis process process occurs occurs through through two two simultaneously simultaneously occurring occurring half-cell half-cell reactions: reactions: 58 thethe oxygen oxygen evolution evolution reaction reaction (OER) (OER) on on the the anode anode and and the the hydrogen hydrogen evolution evolution reaction reaction (HER) (HER) on on the the 59 cathode.cathode. OER OER kinetics kinetics are are more more sluggish sluggish than than HER HER kinetics kinetics [17 [17];]; therefore, therefore, the the performance performance of of water water 60 electrolyserselectrolysers heavily heavily depends depends on on the the OER. OER. Generally, Generally, the the OER OER activity activity of non-noble of non-noble electrocatalysts electrocatalysts is 61 highis high in alkaline in alkaline environments environments [14]. [14]. Non-noble Non-noble metal metal oxides oxides are, therefore,are, therefore, of interest of interest as catalysts as catalysts for 62 AEMWE.for AEMWE. More specifically, More specifically, catalysts catalysts based on basedNi or Co on (hydroxides, Ni or Co (hydroxides, oxides, spinels oxides, and perovskites) spinels and 63 andperovskites) pyrochlores and show pyrochlores good activity show towards good activity the OER towards in alkaline the media. OER in Ni-based alkaline electrocatalysts media. Ni-based 64 haveelectrocatalysts been particularly have wellbeen investigated,particularly well and investigated, include different and ratiosinclude of different Ni–iron (Fe),ratios Ni–chromium of Ni–iron (Fe), 65 (Cr)Ni– andchromium Ni–molybdenum (Cr) and Ni (Mo)–molybdenum oxide catalysts (Mo) oxide [6,18– 23catalysts], amongst [6,18– other23], amongst bimetallics other and bimetallics alloys. 66 Liand et al. alloys. [18] studied Li et alvarious. [18] studied electrodeposited, various electrodeposited, Ni-bimetallic catalysts Ni-bimetallic for OER. catalysts Among for all OER. tested Among metals, all 67 thetested NiFe catalyst metals, showed the NiFe the catalyst highest promotional showed the eff highestect, achieving promotional the lowest effect, overpotential achieving of the 256 lowest mV 2 68 atoverpotential 0.5 A cm− with of 10%256 mV iron at incorporation. 0.5 A cm-2 with Similarly, 10% iron Trotochaud incorporation. et al. [Similarly,20] tested Trotochaud multiple metal et al and. [20] 69 mixed-metaltested multiple oxide metal catalysts and preparedmixed-metal by spin oxide coating catalyst fors OER. prepared Their by study spin showed coating thatfor OER. the Ni Their90Fe10 studyOx 2 70 catalystshowed obtained that the the Ni lowest90Fe10Ox overpotential catalyst obtained of 297 the mV lowest at 1 mA overpotential cm− . of 297 mV at 1 mA cm-2. 71 TheThe most most widely widely accepted accepted description description of of the the Ni Ni oxidation oxidation steps steps in in alkaline alkaline media media is is through through the the 72 BodeBode diagram, diagram, a simplifieda simplified version version of of which which is shownis shown in in Figure Figure1 below 1 below [24 ].[24]. charge Ni α-Ni(OH) γ-NiOOH 2 discharge aging overcharge charge β-Ni(OH) β-NiOOH 2 discharge 73 Figure 1. Nickel oxidation steps in alkaline media. 74 Figure 1. Nickel oxidation steps in alkaline media. In an alkaline environment, nickel is first oxidized to the unstable α-Ni(OH) (about 0.5 V versus 2 − 75 Hg/HgO).In an Prolonged alkaline exposure environment, to an alkalinenickel is environment first oxidized or slight to the anodic unstable polarization α-Ni(OH) brings2 (about it to –0.5 the V 76 stableversusβ-Ni(OH) Hg/HgO).2 phase. Prolonged Further exposure polarization to an alkaline induces environment the oxidative or deprotonation slight anodic polarization of β-Ni(OH) brings2 to 77 β-NiOOHit to the (~0.45 stable V β- versusNi(OH) Hg2 phase./HgO or Further ~1.38 V polarization versus RHE induces at pH 14), the the oxidative active phase deprotonation for OER [17 of]. β- 78 Finally,Ni(OH) further2 to β-NiOOH increasing (~0.45 the V electrode versus Hg/HgO potential or will~1.38 overcharge V versus RHE the nickelat pH 14), catalyst the active and bringphase itfor 79 toOER the γ [17].-NiOOH Finally, phase, further which increasing is believed the electrode to be the potential highest-achievable will overcharge Ni oxidation the nickel phase. catalyst It and is 80 mostbring commonly it to the γ assumed-NiOOH thatphase, the whichβ-NiOOH is believed oxidation to be phase the highest is most-achievable active towards Ni oxidation the OER phase. [17]. It 81 Thisis most oxidative commonly deprotonation assumed process that the to β generate-NiOOH theoxidation catalytic phase species is most for the active OER towards is not particular the OER for[17]. 82 nickel;This itoxidative has been deprotonation reported that cobalt,process iron to generate and manganese-based the catalytic species catalysts for also the deprotonateOER is not particular prior to 83 oxygenfor nickel; evolution, it has been in processes reported that that are cobalt, strongly iron pH-dependentand manganese [25-based–29]. catalysts also deprotonate prior 84 to oxygenCeria (CeO evolution,2) is an in ionically processes conducting that are strongly support pH that-dependent has been [25 widely–29]. studied and applied 85 in variousCeria electrochemical (CeO2) is an applicationsionically conducting due to the support advantageous that has metal been support widely interactions studied and (MSIs) applied that in 86 it mayvarious provide electrochemical [30–32]. The applications MSIs occur throughdue to the the advantageous unique properties metal of support CeO2, such interactions as oxygen (MSIs) storage that 87 it may provide [30–32]. The MSIs occur through the unique properties of CeO2, such as oxygen 88 storage and release properties, and good ionic conductivity [33–35]. Using CeO2 as a catalyst support 89 can modify a catalyst’s dispersion, minimize a catalyst’s agglomeration and increase a catalyst’s

Catalysts 2019, 9, 814 3 of 17

and release properties, and good ionic conductivity [33–35]. Using CeO2 as a catalyst support can modify a catalyst’s dispersion, minimize a catalyst’s agglomeration and increase a catalyst’s surface area. There are very few reports dealing with CeO2 supports for non-noble metal catalysts for OER. Chen et al. [34] reported that the incorporation of CeO2 into a copper oxide catalyst increased OER activity through strong electronic interactions between Ce4+ and Cu ions. They reported an optimal cerium content of 6.9%; further increasing the Ce content resulted in a less optimal Ce4+ distribution, and subsequently lower OER activity. Feng et al. [33] studied the incorporation of CeO2 into FeOOH heterolayered nanotubes and concluded that the enhancement of OER’s performance obtained through the incorporation of ceria into their catalyst was likely the result of the increased FeOOH/CeO2 interfaces as well as the high oxygen storage capacity of the material. Haber et al. [36] developed a Ni0.3Fe0.07Co0.2Ce0.43Ox catalyst exhibiting good activity and stability towards OER. Said catalyst was further studied by Favaro et al. [35], who concluded that the incorporation of ceria into the catalyst does provide unique promotional catalytic properties towards OER. Finally, McCrory et al. [37], showed that their NiCeOx material had the highest overpotential for OER compared to IrOx and NiOx, NiFeOx, amongst other bimetallic electrocatalysts in 1 M NaOH. Preliminary cyclic voltammograms of the effect of ceria incorporation into the Ni catalyst are shown in Figure S1 of the Supplementary Information. With respect to AEMWE studies, Seetharaman et al. [38] studied graphene oxide (GO) modified NiO electrode as an OER catalyst with enhanced electron conductivity and catalytic activity. A NiZnS ternary alloy was used at the cathode. The initial AEMWE performance of the electrodeposited catalysts on Ni foams was evaluated in the study. A Selemion™ AMV membrane (Asahi Glass Co. Ltd.) was used to prepare the membrane electrode assemblies (MEAs) and was sandwiched between the coated, foam electrodes. The AEMWE cell tests were performed using deionized water and concentrations of an alkaline solution (potassium hydroxide (KOH)) ranging from 0 to 5.36 M. They also tested various operating temperatures ranging from 30 to 80 ◦C. Increasing the concentration of the alkaline solution 2 improved the initial cell performance, with current densities of approximately 65 and 140 mA cm− observed for pure water and 5.36 M KOH, respectively, at 1.8 V and 30 ◦C. When using the 5.36 M 2 electrolyte and holding the system at 1.8 V, the current density increased from 100 to 380 mA cm− when increasing the operating temperature from 30 to 80 ◦C. Xiao et al. [39] used electrocatalysts, such as Ni–Fe and Ni–Mo complexes, for the OER and HER in their MEAs. The Ni–Fe electrode was made using a solid-state electrochemical reduction procedure. Specifically, a solution containing Ni and Fe nitrates was sprayed onto a preheated Ni foam substrate and later electrochemically reduced by applying a cathodic current. The resulting loading of the catalysts was 2 40 mg cm− for both electrodes, and the membrane and ionomer used were xQAPS. The performance 2 achieved was 400 mg cm− at 1.85 V in ultra-pure water. Ayers et al. [40] characterized AEMWE using non-noble metal complexes as catalysts for OER. They reported results for the ternary catalysts 30% Ni–Fe–Co/C and 30% Ni–Fe–Mo/C compared to 30% Ni–Fe/C and IrO2. Although they reported very low 2 current densities, between 5 and 40 mA cm− at 1.8 V, the ternary catalysts showed higher performance compared to IrO2 and 30% Ni–Fe/C. Recently, Pavel et al. [41] developed and evaluated AEMWE using low cost transition metal catalysts. The commercial catalysts Acta 4030 (Ni/CeO2-La2O3/C) and Acta 3030 (CuCoOx) (Acta SpA, Italy) were used as HER and OER catalysts, respectively. The loadings of the HER and OER 2 catalysts were 7.4 and 36 mg cm− , respectively. These catalysts were designed to withstand relatively mild alkaline conditions (pH 10–11). The authors explained the effect of HER loading on the kinetic contribution and performance of the AEM electrolysis. Results showed current densities between 60 2 2 and 300 mA cm− at 1.8 V, as the HER catalyst’s loading ranged from 0.6 to 7.4 mg cm− , while the 2 OER catalyst’s loading was kept constant at 36 mg cm− . The aim of the presented work was to synthesize, characterize and electrochemically evaluate nanosized nickel-based electrocatalysts for the OER. Ni, Ni90Fe10 and Ni90Fe10/CeO2 catalysts were synthesized by a scalable method, characterized by scanning and transmission electron microscopy (SEM and TEM), X-ray diffraction (XRD) and electrochemically evaluation using a conventional Catalysts 2019, 9, 814 4 of 17

three-electrode electrochemical cell. The Ni-based electrocatalysts were also evaluated as anodes in real AEMWEs and compared to commercially available noble metal-based anode electrodes. This paper Catalysts 2019, 9, x FOR PEER REVIEW 4 of 17 includes a study on the inﬂuence of the alkaline electrolyte concentration on the catalytic activity of the 142 OER;activity ex-situ, of the three-electrode OER; ex-situ, OERthree- measurementselectrode OERand measurements in situ AEMWE and performancesin situ AEMWE are performances compared at 143 1are M compared and 0.1 M KOHat 1 M concentrations. and 0.1 M KOH concentrations. 2. Results and Discussion 144 2. Results and Discussion 2.1. Scanning and Transmission Electron Microscopy 145 2.1. Scanning and Transmission Electron Microscopy Figure2a,b shows TEM images of the as-synthesized Ni and Ni 90Fe10 nanoparticles (NPs), 146 90 10 respectively.Figure 2a,b Using shows the scale TEM on images the image, of the it was as-synthesized possible to approximate Ni and Ni Fe a particle nanoparticles size of around (NPs), 147 4–6respectively. nm, agglomerated Using the into scale larger on the clusters image, of it NPs.was possible to approximate a particle size of around 4– 148 6 nm, agglomerated into larger clusters of NPs. (a) (b)

149 50 nm

150 FigureFigure 2. 2.TEM TEM of of ( a(a)) Ni Ni and and ( b(b)) Ni Ni9090FeFe1010 nanoparticles (NPs).

151 FigureFigure3 a–d3a–d represents represents SEM SEM images images ofof the the anodes anodes usedused inin thethe AEMWEAEMWE priorprior toto experiments.experiments. 152 LowLow magnitude images images are are displayed displayed to toshow show how how the thebulk bulk electrode electrode surfaces surfaces differ di fromffer each from other. each 153 other.Comparisons Comparisons with withthe Ni, the NiNi,90 NiFe1090 Feand10 and Ni90 NiFe9010/CeOFe10/CeO2 electrodes2 electrodes are are shown shown in in Figure Figure 3b,c,d,3b–d, 154 respectively.respectively. The The Ir Ir black black benchmark benchmark shown shown in Figurein Figure3a, shows 3a, shows a more a porous more porous electrode electrode with a clearerwith a 155 presenceclearer presence of particles of particles making upmaking the electrode up the electrode surface. Althoughsurface. Although the same the ink same preparation ink preparation and electrode and 156 sprayingelectrode procedure spraying was procedure used for allwas electrodes, used for the all Ni-based electrodes, electrode the Ni fabrication-based electrode procedure fabrication may need 157 furtherprocedure optimization. may need Thefurther Ni anode optimization. shows an The almost Ni flakyanode electrode shows an surface, almost which flaky was electrode very similar surface, to 158 resultswhich obtained was very for similar the Ni to90Fe results10 electrode obtained surface. for the The Ni Ni9900FeFe1010 electrode/CeO2 electrode surface. shows The relativelyNi90Fe10/CeO flat2 159 surfaceelectrode with shows large relatively cracks in flat it, whichsurface is with likely large the cracks result ofin havingit, which to is spray likely twice the result as much of having ink onto to 160 thespray membrane twice as much to get ink the sameonto the Ni membrane loading, as to 50 get wt% the of same this catalystNi loading, is CeO as 502 support. wt% of this Furthermore, catalyst is 161 increasingCeO2 support. the magnification Furthermore, of increasing the Ni-based the anodesmagnification did not of present the Ni a-based more detailedanodes did electrode not present surface, a 162 hencemore detailed why they electrode were kept surface, at a lower hence magnification why they were than kept Ir. at a lower magnification than Ir.

Catalysts 2019, 9, 814 5 of 17 Catalysts 2019, 9, x FOR PEER REVIEW 5 of 17

Catalysts 2019, 9, x FOR PEER REVIEW 6 of 17 CatalystsCatalystsCatalysts 2019 2019 2019, ,9 ,9, 9,x ,x xFOR FOR FOR PEER PEER PEER REVIEW REVIEW REVIEW 66 6 of ofof 17 1717 Catalysts 2019, 9, x FOR PEER REVIEW 6 of 17

163

164 FigureFigure 3.3. SEMsSEMs ofof ((aa)) IrIr black,black, ( b(b)) Ni, Ni, ( c()c) Ni Ni9090FeFe1010 andand ((dd)) NiNi9090Fe1010/Ce/CeOO2 anodesanodes of the catalyst coated 165 membranes.membranes. The magnificationsmagnifications ofof the imagesimages areare 30,30, 300,300, 300300 andand 200200 µμm,m, respectively.respectively. 2.2. X-ray Diffraction 166 2.2. X-ray Diffraction The XRD patterns of the Ni and Ni90Fe10 NPs are shown in Figure4. XRD was not used to 167 The XRD patterns of the Ni and Ni90Fe10 NPs are shown in Figure 4. XRD was not used to characterize the supported Ni catalyst, as ceria is also a face centered cubic (FCC) [32], and therefore, 168 characterize the supported Ni catalyst, as ceria is also a face centered cubic (FCC) [32], and therefore, interferes with the main nickel diffraction peaks. As can be seen in Table1, both samples include 169 interferes with the main nickel diffraction peaks. As can be seen in Table 1, both samples include a a mixture of Ni and β-Ni(OH)2—the stable Ni(OH)2 phase. In the Ni90Fe10 XRD pattern, the peak 170 mixture of Ni and β-Ni(OH)2—the stable Ni(OH)2 phase. In the Ni90Fe10 XRD pattern, the peak identified with the circle icon could be Ni (111) or Fe (110) [42,43]. Possible salt contamination can 171 identified with the circle icon could be Ni (111) or Fe (110) [42,43]. Possible salt contamination can occur due to the synthesis method; however, catalyst samples were washed multiple times with water 172 occur due to the synthesis method; however, catalyst samples were washed multiple times with prior to XRD measurements to remove all NaCl. The possible presence of NaCl was ruled out due to 173 water prior to XRD measurements to remove all NaCl. The possible presence of NaCl was ruled out the absence of particular sharp peaks in the XRD spectra [44]. Having both Ni and Fe in the same peak 174 due to the absence of particular sharp peaks in the XRD spectra [44]. Having both Ni and Fe in the position, and the slight shift between the Ni(III) peaks shown in the Ni XRD, indicated the potential 179175179 same peak position, and the slight shift between the Ni(III) peaks shown in the Ni XRD, indicated the 179179 formation of an alloy material. Additionally, the broad peak shown in the Ni90Fe10 pattern reflects the 179176 potential formation of an alloy material. Additionally, the broad peak shown in the Ni90Fe10 pattern 180 Figure 4. XRD results for Ni (black) and Ni90Fe10 (red) NPs. 180180180 very small particle size observedFigureFigureFigure 4. 4. 4. XRD XRD inXRD the results results results TEM for for imagefor Ni Ni Ni (black) (black) (black) of the and and and synthesized Ni Ni Ni9090Fe90FeFe1010 10(red) (red) (red) materials, NPs. NPs. NPs. shown in Figure2. 180177 reflects the very small particleFigure 4. size XRD observed results for in Ni the (black) TEM and image Ni90 Feof10 the (red) synthesized NPs. materials, shown 178181 in Figure 2. Table 1. XRD results [42,43]. 181181181 TableTableTable 1.1. 1. 1. XRD XRD XRD results results results [[42,43]. 42[42,43]. [42,43].,43]. 181 Table 1. XRD results [42,43]. IconIcon on Figure on Figure 4 4 ExperimentalExperimental 2θ 2 θ [deg][deg] CorrespondingCorresponding Compound; Compound; Reported Reported 2θ [deg] 2θ [deg] IconIconIcon on on on Figure Figure Figure 4 4 4 ExperimentalExperimentalExperimental 2θ 2θ 2θ [deg] [deg] [deg] CorrespondingCorrespondingCorresponding Compound; Compound; Compound; Reported Reported Reported 2θ 2θ 2θ [deg] [deg] [deg] Icon on Figure 4 Experimental 2θ [deg] Corresponding Compound; Reported 2θ [deg] Diamond 33.7833.78 β-Ni(OH)β-Ni(OH)2 (100);2 (100); 33.1 33.1 DiamondDiamondDiamondDiamond 33.7833.7833.78 ββ-β-Ni(OH)Ni(OH)-Ni(OH)22 (100); 2(100); (100); 33.1 33.1 33.1 Diamond 33.78 β-Ni(OH)2 (100); 33.1 Pentagon 44.65 Ni (111); 44.45 PentagonPentagonPentagon 44.6544.6544.6544.65 NiNiNiNi (111); (111); (111); (111); 44.45 44.45 44.45 44.45 PentagonPentagon 44.65 Ni (111); 44.45 Ni (111); 44.45 Circle 45.37 NiNiNiNi (111); (111); (111); (111); 44.45 44.45 44.45 44.45 CircleCircleCircle 45.3745.3745.3745.37 NiFe (111); (110); 44.45 44.9 CircleCircle 45.37 FeFeFe (110);Fe (110); (110); (110); 44.9 44.9 44.9 44.9 Fe (110); 44.9 Triangle 59.99 β-Ni(OH)2 (110); 59.1 TriangleTriangleTriangle 59.9959.9959.9959.99 ββ-Ni(OH)β-β-Ni(OH)Ni(OH)-Ni(OH)2 (110);22 (110); 2(110); (110); 59.1 59.1 59.1 59.1 TriangleTriangle 59.99 β-Ni(OH)2 (110); 59.1 Square 70.68 β-Ni(OH)2 (103) ; ~71 SquareSquare 70.6870.68 β-βNi(OH)-Ni(OH)22 (103)2 (103) ; ~71; ~71 Square 70.6870.68 β-Ni(OH)β-Ni(OH)(103); (103) ~71 ; ~71 SquareSquare 70.68 β-Ni(OH)2 2 (103) ; ~71 182 2.3. Electron Energy-Loss Spectroscopy 182182182 2.3.2.3.2.3. Electron Electron Electron Energy Energy Energy--LossLoss-Loss Spectroscopy Spectroscopy Spectroscopy 182 2.3. Electron Energy-Loss Spectroscopy 183 Figure 5 summarized an EELS spectrum taken for the Ni90Fe10/CeO2 material. In Figure 5a, it is 183183183 FigureFigureFigure 5 5 5summarized summarized summarized an an an EELS EELS EELS spectrum spectrum spectrum takentaken taken for for for thethe the Ni Ni Ni9090Fe90FeFe1010/CeO10/CeO/CeO22 material.2material. material. In In In FigureFigure Figure 5a, 5a, 5a, it it it is is is 183184 possibleFigure to 5 see,summarized in orange, an the EELS region spectrum of the takenSTEM for image the Ni where90Fe10 /CeOthe analysis2 material. was In carried Figure 5a,out. it The is 184184184 possiblepossiblepossible to toto see, see,see, in inin orange, orange,orange, the thethe region regionregion of ofof the thethe STEM STEMSTEM image imageimage where wherewhere the thethe analysis analysisanalysis was waswas carried carriedcarried out. out.out. The TheThe 184185 possibleregion into yellowsee, in wasorange, used the for regiondrift correction. of the STEM As shown image inwhere Figure the 5b, analysis within thewas orange carried region, out. The two 185185185 region regionregion in in in yellow yellow yellow was was was used used used for for for drift drift drift correction. correction. correction. As As As shown shown shown in in in Figure Figure Figure 5b, 5b, 5b, within within within the the the orange orange orange region, region, region, two two two 185186 regionsections in yellowwere an wasalyzed; used thefor driftfirst correction.section shows As showna cloudier in Figure area, 5b,while within the thesecond orange section region, shows two a 186186186 sectionssectionssections were werewere an ananalyzed;alyzed;alyzed; the thethe first firstfirst section sectionsection shows showsshows a a a cloudier cloudiercloudier area, area,area, while whilewhile the thethe second secondsecond section sectionsection shows showsshows a a a 186187 sectionsclearer structuredwere analyzed; area. Asthe seenfirst insection Figure shows 5c, the a first cloudier region area, mostly while consisted the second of the section Ni and showsFe metals a 187187187 clearerclearerclearer structured structured structured area. area. area. As As As seen seen seen in in in Figure Figure Figure 5c, 5c, 5c, the the the first first first region region region mostly mostly mostly consisted consisted consisted of of of the the the Ni Ni Ni and and and Fe Fe Fe metals metals metals 187188 clearerin the structuredsample, while area. the As second seen in region Figure mostly 5c, the consisted first region of mostlythe CeO consisted2 support. of This the observationNi and Fe metals is also 188188188 ininin the the the sample, sample, sample, while while while the the the second second second region region region mostly mostly mostly consisted consisted consisted of of of the the the CeO CeO CeO22 s 2s upport.support.upport. This This This observation observation observation is is is also also also 188189 inseen the sample,in the EELS while mapping the second found region in Figuremostly S5 consisted of the Supplementary of the CeO2 support. Information This observation of this article. is also For 189189189 seenseenseen inin in thethe the EELSEELS EELS mappingmapping mapping foundfound found inin in FigureFigure Figure S5S5 S5 ofof of thethe the SupplementarySupplementary Supplementary InformationInformation Information ofof of thisthis this article.article. article. ForFor For 189190 seenTEM, in the STEM EELS and mapping EDX, characterizationfound in Figure S5 of of the the Ni Supplementary90Fe10/CeO2 catalyst, Information see Figures of this S2 article.–S4 in For the 190190190 TEM,TEM,TEM, STEM STEM STEM and and and EDX, EDX, EDX, characterization characterization characterization of of of the the the Ni Ni Ni9090Fe90FeFe1010/CeO10/CeO/CeO22 2catalyst, catalyst,catalyst, see see see Figures Figures Figures S2 S2 S2––S4–S4S4 in in in the the the 190191 TEM,Supplementary STEM and Information. EDX, characterization of the Ni90Fe10/CeO2 catalyst, see Figures S2–S4 in the 191191191 SupplementarySupplementarySupplementary Information. Information. Information. 191 Supplementary Information.

Catalysts 2019, 9, 814 6 of 17 Catalysts 2019, 9, x FOR PEER REVIEW 6 of 17

179 Figure 4. XRD results for Ni (black) and Ni Fe (red) NPs. 180 Figure 4. XRD results for Ni (black) and Ni9090Fe1010 (red) NPs. 2.3. Electron Energy-Loss Spectroscopy 181 Table 1. XRD results [42,43]. Figure5 summarized an EELS spectrum taken for the Ni 90Fe10/CeO2 material. In Figure5a, it isIcon possible on Figure to see, 4 inExperimental orange, the region 2θ [deg] ofthe STEMCorresponding image where Compound; the analysis Reported was carried 2θ [deg] out. The region in yellow was used for drift correction. As shown in Figure5b, within the orange region, Diamond 33.78 β-Ni(OH)2 (100); 33.1 two sections were analyzed; the ﬁrst section shows a cloudier area, while the second section shows a clearerPentagon structured area. As seen44.65 in Figure 5c, the ﬁrst region mostly consistedNi (111); 44.45 of the Ni and Fe metals in the sample, while the second region mostly consisted of the CeO support. This observation is Ni2 (111); 44.45 also seenCircle in the EELS mapping found45.37 in Figure S5 of the Supplementary Information of this article. Fe (110); 44.9 For TEM, STEM and EDX, characterization of the Ni90Fe10/CeO2 catalyst, see Figures S2–S4 in the SupplementaryTriangle Information. 59.99 β-Ni(OH)2 (110); 59.1 Catalysts 2019, 9, x FOR PEER REVIEW 7 of 17

Square 70.68 β-Ni(OH)2 (103) ; ~71

182 2.3. Electron Energy-Loss Spectroscopy

183 Figure 5 summarized an EELS spectrum taken for the Ni90Fe10/CeO2 material. In Figure 5a, it is 184 possible to see, in orange, the region of the STEM image where the analysis was carried out. The 185 region in yellow was used for drift correction. As shown in Figure 5b, within the orange region, two 186 sections were analyzed; the first section shows a cloudier area, while the second section shows a 187 clearer structured area. As seen in Figure 5c, the first region mostly consisted of the Ni and Fe metals 188 in the sample, while the second region mostly consisted of the CeO2 support. This observation is also 189 seen in the EELS mapping found in Figure S5 of the Supplementary Information of this article. For 190 TEM, STEM and EDX, characterization of the Ni90Fe10/CeO2 catalyst, see Figures S2–S4 in the 191 Supplementary Information.

192

193 FigureFigure 5.5. EELS results for Ni 9090FeFe1010/CeO/CeO2. 2(.(a) aThe) The area area of ofSTEM STEM image image that that was was analyzed analyzed in orange, in orange, (b) 194 (showsb) shows the the regions regions within within the the orange orange area area that that were were analyzed analyzed and and ( (cc)) shows shows the the results results of the analysisanalysis 195 ofof thethe twotwo regions.regions.

196 2.4. OER Experiments 197 In the cyclic voltammograms (CVs) shown in Figure 6a, the overpotentials for OER in 1 M KOH 198 can be summarize as ηIr black < ηNi90Fe10/CeO2 < ηNi90Fe10 < ηNi, where the onset potential for OER of Ir occurs 199 at ~0.524 V (~1.449 V versus RHE). It is important to note that the onset potentials are values taken at 200 low currents. In addition to showing the lowest onset potential for OER, the iridium electrode also 201 attains the highest current density and holds the lowest overpotential at 10 mA cm-2, as shown in 202 Table 2. Among the Ni-based electrocatalysts, the Ni electrode shows the highest current density, 203 while the Ni90Fe10/CeO2 sample exhibits the lowest onset potential for the OER at ~0.553 V (~1.478 V 204 versus RHE). However, the Ni90Fe10 sample is the only sample that shows a lowered onset potential, 205 while also attaining a high current density for OER. Additionally, it showed the second lowest 206 overpotential at 10 mA cm-2. From the CV shown in Figure 6b, the reaction overpotentials for OER in 207 0.1 M KOH can be written as ηIr black < ηNi90Fe10/CeO2 < ηNi90Fe10 < ηNi, where the onset potential for OER or 208 Ir occurs at ~0.589 V (1.455 V versus RHE). Although the Ir black still showed the lowest onset 209 potential for the OER in 0.1 M KOH, the catalyst performs rather poorly due to apparently slow 210 kinetics, reaching less than 5 mA cm-2. The Ni90Fe10 sample, on the other hand, continued to show a 211 relatively high current for OER and held an overpotential of 0.404 V at 5 mA cm-2. The Ni90Fe10/CeO2 212 sample still exhibits the lowest onset potential amongst the Ni-based materials at around 0.635 V 213 (1.501 V versus RHE). Figure 6c,d shows Tafel plots obtained from linear sweep voltammograms 214 (LSVs) run at 1 mV s-1 in the OER region. Delineated on the figures are the regions in which the slopes 215 tabulated in Table 2 were calculated. Note that no activity coefficients or exchange current densities 216 were reported, as accurate estimations of a complete kinetic model require very detailed studies, such 217 as the one reported by Reksten et al. [45] for IrxRu(1-x)O2 catalysts in acidic electrolyte.

Catalysts 2019, 9, 814 7 of 17

2.4. OER Experiments In the cyclic voltammograms (CVs) shown in Figure6a, the overpotentials for OER in 1 M KOH can be summarize as ηIr black < ηNi90Fe10/CeO2 < ηNi90Fe10 < ηNi, where the onset potential for OER of Ir occurs at ~0.524 V (~1.449 V versus RHE). It is important to note that the onset potentials are values taken at low currents. In addition to showing the lowest onset potential for OER, the iridium 2 electrode also attains the highest current density and holds the lowest overpotential at 10 mA cm− , as shown in Table2. Among the Ni-based electrocatalysts, the Ni electrode shows the highest current density, while the Ni90Fe10/CeO2 sample exhibits the lowest onset potential for the OER at ~0.553 V (~1.478 V versus RHE). However, the Ni90Fe10 sample is the only sample that shows a lowered onset potential, while also attaining a high current density for OER. Additionally, it showed the second 2 lowest overpotential at 10 mA cm− . From the CV shown in Figure6b, the reaction overpotentials for OER in 0.1 M KOH can be written as ηIr black < ηNi90Fe10/CeO2 < ηNi90Fe10 < ηNi, where the onset potential for OER or Ir occurs at ~0.589 V (1.455 V versus RHE). Although the Ir black still showed the lowest onset potential for the OER in 0.1 M KOH, the catalyst performs rather poorly due to apparently slow 2 kinetics, reaching less than 5 mA cm− . The Ni90Fe10 sample, on the other hand, continued to show 2 a relatively high current for OER and held an overpotential of 0.404 V at 5 mA cm− . The Ni90Fe10/CeO2 sample still exhibits the lowest onset potential amongst the Ni-based materials at around 0.635 V (1.501 V versus RHE). Figure6c,d shows Tafel plots obtained from linear sweep voltammograms 1 (LSVs) run at 1 mV s− in the OER region. Delineated on the ﬁgures are the regions in which the slopes tabulated in Table2 were calculated. Note that no activity coe ﬃcients or exchange current densities were reported, as accurate estimations of a complete kinetic model require very detailed Catalysts 2019, 9, x FOR PEER REVIEW 8 of 17 studies, such as the one reported by Reksten et al. [45] for IrxRu(1 x)O2 catalysts in acidic electrolyte. −

218 Figure 6. (a,b) Stable CVs run at 20 mV s 1 in 1 M and 0.1 M KOH, respectively. (c,d) Tafel plots -1− 219 Figure 6. (a,b) Stable CVs run at 20 mV s in 1 M and1 0.1 M KOH, respectively. (c,d) Tafel plots obtained from LSVs run between [0.3, 0.8] V at 1 mV s− in 1 M and 0.1 M KOH, respectively. Catalysts: 220 obtained from LSVs run between [0.3, 0.8] V at 1 mV s-1 in 1 M and 0.1 M KOH, respectively. Catalysts: Ir black (black), Ni (red), Ni90Fe10 (blue) and Ni90Fe10/CeO2 (green) NPs. 221 Ir black (black), Ni (red), Ni90Fe10 (blue) and Ni90Fe10/CeO2 (green) NPs.

222 Table 2. Tafel slopes and overpotentials at 0.1 and 1 M for Ir black, Ni, Ni90Fe10 and Ni90Fe10/CeO2 223 NPs.

KOH Conc. η @ 5 mA cm-2 η @ 10 mA cm-2 Catalyst Tafel Slope [mV dec-1] [M] [mV] [mV] 1.0 32.0 268 295 Ir Black 0.1 70.6 N/A N/A 1.0 75.5 337 365 Ni 0.1 73.1 424 N/A 1.0 71.9 298 341 Ni90Fe10 0.1 83.3 404 N/A 1.0 70.7 323 369 Ni90Fe10/CeO2 0.1 82.1 424 N/A

224 The results summarized in Figure 6 and Table 2 indicate that the unsupported Ni90Fe10 catalyst 225 is the most promising Ni-based material for OER in alkaline environment, as it shows both a lower 226 onset potential and a higher current density in both 1 M and 0.1 M KOH. Amongst all Ni-based 227 materials, it also shows the lowest overpotential at both 5 and 10 mA cm-2. Similar results have been 228 previously reported, where the addition of Fe increases the OER activity [18,21,23,46,47]. Increased 229 OER overpotentials with decreasing electrolyte concentrations have also been reported 230 [18,21,23,46,47]. When comparing values in Table 2, it is possible to observe that in this case, our Ni- 231 based catalysts experience an increase in overpotential between 87 and 101 mV, when decreasing the 232 electrolyte concentration from 1 M to 0.1 M KOH. As for the OER kinetics, it was noticed that the 233 Tafel slopes obtained in 1 M KOH for our study were sometimes twice as high compared to the values 234 reported in literature [18,20,21,23,46–48]. Discrepancies in the literature are common, with reported 235 values ranging between 40 and 130 mV dec-1 for the Tafel slope of nickel-based oxides, and are likely 236 combinations of the regions of LSV used to calculate the slope; it is well known that there are

Catalysts 2019, 9, 814 8 of 17

Table 2. Tafel slopes and overpotentials at 0.1 and 1 M for Ir black, Ni, Ni90Fe10 and Ni90Fe10/CeO2 NPs.

2 2 KOH Conc. Tafel Slope η @ 5 mA cm− η @ 10 mA cm− Catalyst 1 [M] [mV dec− ] [mV] [mV] 1.0 32.0 268 295 Ir Black 0.1 70.6 N/AN/A 1.0 75.5 337 365 Ni 0.1 73.1 424 N/A 1.0 71.9 298 341 Ni Fe 90 10 0.1 83.3 404 N/A 1.0 70.7 323 369 Ni Fe /CeO 90 10 2 0.1 82.1 424 N/A

The results summarized in Figure6 and Table2 indicate that the unsupported Ni 90Fe10 catalyst is the most promising Ni-based material for OER in alkaline environment, as it shows both a lower onset potential and a higher current density in both 1 M and 0.1 M KOH. Amongst all Ni-based 2 materials, it also shows the lowest overpotential at both 5 and 10 mA cm− . Similar results have been previously reported, where the addition of Fe increases the OER activity [18,21,23,46,47]. Increased OER overpotentials with decreasing electrolyte concentrations have also been reported [18,21,23,46,47]. When comparing values in Table2, it is possible to observe that in this case, our Ni-based catalysts experience an increase in overpotential between 87 and 101 mV, when decreasing the electrolyte concentration from 1 M to 0.1 M KOH. As for the OER kinetics, it was noticed that the Tafel slopes obtained in 1 M KOH for our study were sometimes twice as high compared to the values reported in literature [18,20,21,23,46–48]. Discrepancies in the literature are common, with reported values 1 ranging between 40 and 130 mV dec− for the Tafel slope of nickel-based oxides, and are likely combinations of the regions of LSV used to calculate the slope; it is well known that there are generally two Tafel regions for OER, separated at ~1.5 V versus RHE (~0.575 V versus Hg/HgO in 1 M KOH) [17]. Another possible cause for the diﬀerent Tafel slopes for Ni-based materials could be the state of the oxide surface at the time of the measurement or the scan rate selected. In addition, only a small increase in the Tafel slope was observed for the Ni-based catalysts when decreasing the electrolyte concentration, a clear indication that the kinetics for the OER remain unchanged in this pH region. On the other hand, it seems as though the kinetics of Ir-based catalysts are considerably more affected by the decrease in KOH concentration, more than doubling the Tafel slope from 32 to 1 70 mV dec− , while going from 1 M to 0.1 M KOH, respectively. These results are concurrent with observations reported in literature where Ir electrodes show a rather significant decrease in performance with decreasing electrolyte pH, either due to the development of a poorly conducting oxide film on the surface of the iridium working electrode [49], a change in OER mechanism [49] or possible blocking of the electrode surface due to evolving oxygen [50] However, Pi et al. [51] reported Tafel slopes of 32.7, 38.4 and 40.1 mV 1 1 dec− in 1 M KOH, and 42.1, 48.5, 50.2 mV dec− in 0.1 M KOH for surface-cleaned 3D Ir nanosheets, 3D Ir nanosheets and Ir NPs, respectively. The 3D nanosheets were prepared with the wet-chemical synthesis 1 method. A study by Lyons et al. [50] also reported similar Tafel slopes of 44 and 43 mV dec− in 0.1 and 1 M KOH, respectively, for IrO2 electrodes prepared via thermal decomposition onto a substrate. Finally, 1 Tahir et al. [52], reported a Tafel slope of 65 mV dec− in 1 M KOH for IrO2. Figure6 and Table2 show that the Ni 90Fe10/CeO2 catalyst has a comparable overpotential for OER to those of the unsupported Ni-based materials. Although the current is not as high when including the ceria support, the material still shows promising activity through its Tafel parameters, which are very similar to the rest of the Ni-based materials. It is important to note that the CVs presented in Figure6 were normalized by geometric surface area; normalizing the results by mass of metallic Ni would show that the ceria supported materials are the most promising OER catalysts, as shown in Figure7, where the current densities normalized by geometric surface area and by mass of nickel at 0.8 V versus Hg/HgO (1.725 and 1.666 V versus RHE in 1 and 0.1 M KOH, respectively), are compared. Based on the interesting Catalysts 2019, 9, x FOR PEER REVIEW 9 of 17

237 generally two Tafel regions for OER, separated at ~1.5 V versus RHE (~0.575 V versus Hg/HgO in 1 238 M KOH) [17]. Another possible cause for the different Tafel slopes for Ni-based materials could be 239 the state of the oxide surface at the time of the measurement or the scan rate selected. 240 In addition, only a small increase in the Tafel slope was observed for the Ni-based catalysts when 241 decreasing the electrolyte concentration, a clear indication that the kinetics for the OER remain 242 unchanged in this pH region. On the other hand, it seems as though the kinetics of Ir-based catalysts 243 are considerably more affected by the decrease in KOH concentration, more than doubling the Tafel 244 slope from 32 to 70 mV dec-1, while going from 1 M to 0.1 M KOH, respectively. These results are 245 concurrent with observations reported in literature where Ir electrodes show a rather significant 246 decrease in performance with decreasing electrolyte pH, either due to the development of a poorly 247 conducting oxide film on the surface of the iridium working electrode [49], a change in OER 248 mechanism [49] or possible blocking of the electrode surface due to evolving oxygen [50] However, 249 Pi et al. [51] reported Tafel slopes of 32.7, 38.4 and 40.1 mV dec-1 in 1 M KOH, and 42.1, 48.5, 50.2 mV 250 dec-1 in 0.1 M KOH for surface-cleaned 3D Ir nanosheets, 3D Ir nanosheets and Ir NPs, respectively. 251 The 3D nanosheets were prepared with the wet-chemical synthesis method. A study by Lyons et al. 252 [50] also reported similar Tafel slopes of 44 and 43 mV dec-1 in 0.1 and 1 M KOH, respectively, for 253 IrO2 electrodes prepared via thermal decomposition onto a substrate. Finally, Tahir et al. [52], 254 reported a Tafel slope of 65 mV dec-1 in 1 M KOH for IrO2. 255 Figure 6 and Table 2 show that the Ni90Fe10/CeO2 catalyst has a comparable overpotential for 256 OER to those of the unsupported Ni-based materials. Although the current is not as high when 257 including the ceria support, the material still shows promising activity through its Tafel parameters, 258 which are very similar to the rest of the Ni-based materials. It is important to note that the CVs 259 presented in Figure 6 were normalized by geometric surface area; normalizing the results by mass of 260 metallic Ni would show that the ceria supported materials are the most promising OER catalysts, as 261 shown in Figure 7, where the current densities normalized by geometric surface area and by mass of Catalysts 2019, 9, 814 9 of 17 262 nickel at 0.8 V versus Hg/HgO (1.725 and 1.666 V versus RHE in 1 and 0.1 M KOH, respectively), are 263 compared. Based on the interesting results of the cerium-content optimization reported in the paper 264 byresults Chen of theet al cerium-content. [34], future optimization work on the reported Ni-based in the materials paper by should Chen et includeal. [34], future support work content on the 265 optimizationNi-based materials of the shouldsynthesized include catalysts. support contentFigure S6 optimization in the Supplementary of the synthesized Information catalysts. shows Figure the S6 266 normalizationin the Supplementary by mass Information of metal in showsthe Ni the-based normalization samples. by mass of metal in the Ni-based samples.

267 Figure 7. Comparison between current densities by geometric surface area and by mass of Ni at 0.8 V 268 Figure 7. Comparison between current densities by geometric surface area and by mass of Ni at 0.8 V versus Hg/HgO in (a) 1 M KOH and (b) 0.1 M KOH. Data for Figure7a,b were taken from the CVs 269 versus Hg/HgO in (a) 1 M KOH and (b) 0.1 M KOH. Data for Figures 7a and b were taken from the presented in Figure6a,b, respectively. 270 CVs presented in Figure 6a and b, respectively. 2.5. AEMWE Experiments 271 2.5. AEMWE Experiments Polarization curves and electrochemical impedance spectroscopy (EIS) were carried out 272 in anPolarization AEMWE cell. curves As shownand electrochemical in Figure8a,b, impedance polarization spectroscopy curves in 1 M(EIS) KOH were show carried a significantly out in an 273 AEMWEhigher performance cell. As shown and in lower Figure overpotential 8a,b, polarization for OER curves than in those1 M KOH taken show in 0.1 a significantly M KOH. Achieving higher 2 274 performancecurrent densities and lower of 2 A overpotential cm− at cell voltagesfor OER ofthan 1.85–1.9 those Vtaken in 1 Min KOH0.1 M atKOH. 50 ◦C, Achieving may be regardedcurrent 275 densitiesas an excellent of 2 A cm result,-2 at cell considering voltages of the 1.85 non-noble–1.9 V in 1 metal M KOH nature at 50 of °C, these may catalysts be regarded [2,8 ].as Inan 1excellent M KOH, 276 result,the Ni considering and Ni90Fe10 theelectrodes non-noble show metal similar nature performance of these catalysts to the [2,8]. benchmark In 1 M atKOH, only the twice Ni and the intendedNi90Fe10 2 2 metallic loading (6 mg cm− versus 3 mg cm− for Ir-black). As shown in Table3, the Ni 90Fe10 and Ni90Fe10/CeO2 electrodes have lower overpotentials than the Ni electrode at lower current densities; however, not at higher current densities, in accordance with our ex-situ CV experiments (Figure6a,b). In 0.1 M KOH, the Ni90Fe10 electrode was the only catalyst that could reach an operating current of 2 2 A cm− , without inducing too high an overpotential. Contrary to the non-noble catalyst results, the Ir black benchmark electrode shows a significant decrease in performance at the lower electrolyte concentration. This result not only reflects the three-electrode cell results very nicely, but also shows that Ni-based electrocatalysts, particularly the Ni90Fe10 catalyst, are promising electrocatalysts for use as AEMWE anodes. The Ni90Fe10 catalyst shows similar performance as the Ir noble metal in 1 M KOH and outperforms the noble metal in 0.1 M. The nickel-based materials in this work also show comparable, if not better, performance than other AEMWE tests carried out with similar set ups [12,53]. EIS measurements were carried out to evaluate the overall cell resistance and charge transfer resistance 2 for the single AEMWE cell. Figure8c,d and Table3 show the EIS of each cell set up at 1 A cm − . Although the EIS measurements were done for the full cell, a simple R(QR) circuit was used to fit the experimental data. The R(QR) model is shown in the corners of Figure8c,d, and consists of a solution or overall cell resistance (REL) in series with a parallel combination of a charge transfer resistance (RCT) and a constant phase element (Q). REL is found from the high frequency intersection of the real axis of the fitted curve, while the RCT is represented by the diameter of the ensuing semicircle. All parameters were extracted from fits of the experimental data and reproduced in Table3 under various conditions. Apart from the Ni90Fe10/CeO2, all anodes exhibited similar REL in 1 M KOH. As the REL corresponds to the ohmic resistance of the cell, it was expected that it be similar for the unsupported Ni, Ir and Ni90Fe10 catalysts. It is interesting to observe that the electrolyser cell using the Ni90Fe10/CeO2 shows more than twice the REL compared to the other cells for the same KOH concentration. This increase in ohmic resistance is 2 caused by the rather large amount of less-conductive ceria in the catalytic layer (6 mg cm− for both the Ni catalyst and the CeO2 support), increasing the thickness of the catalytic layer. Incorporating ceria into Catalysts 2019, 9, x FOR PEER REVIEW 10 of 17

277 electrodes show similar performance to the benchmark at only twice the intended metallic loading (6 278 mg cm-2 versus 3 mg cm-2 for Ir-black). As shown in Table 3, the Ni90Fe10 and Ni90Fe10/CeO2 electrodes 279 have lower overpotentials than the Ni electrode at lower current densities; however, not at higher 280 current densities, in accordance with our ex-situ CV experiments (Figure 6a and b). In 0.1 M KOH, 281 the Ni90Fe10 electrode was the only catalyst that could reach an operating current of 2 A cm-2, without 282 inducing too high an overpotential. Contrary to the non-noble catalyst results, the Ir black benchmark 283 electrode shows a significant decrease in performance at the lower electrolyte concentration. This 284 result not only reflects the three-electrode cell results very nicely, but also shows that Ni-based 285 electrocatalysts, particularly the Ni90Fe10 catalyst, are promising electrocatalysts for use as AEMWE 286 anodes. The Ni90Fe10 catalyst shows similar performance as the Ir noble metal in 1 M KOH and 287 outperforms the noble metal in 0.1 M. The nickel-based materials in this work also show comparable, 288 if not better, performance than other AEMWE tests carried out with similar set ups [12,53]. 289 EIS measurements were carried out to evaluate the overall cell resistance and charge transfer 290 resistance for the single AEMWE cell. Figure 8c,d and Table 3 show the EIS of each cell set up at 1 A 291 cm-2. Although the EIS measurements were done for the full cell, a simple R(QR) circuit was used to 292 fit the experimental data. The R(QR) model is shown in the corners of Figure 8c,d, and consists of a 293 solution or overall cell resistance (REL) in series with a parallel combination of a charge transfer 294 resistance (RCT) and a constant phase element (Q). REL is found from the high frequency intersection 295 of the real axis of the fitted curve, while the RCT is represented by the diameter of the ensuing 296 semicircle. All parameters were extracted from fits of the experimental data and reproduced in Table 297 3 under various conditions. Apart from the Ni90Fe10/CeO2, all anodes exhibited similar REL in 1 M 298 KOH. As the REL corresponds to the ohmic resistance of the cell, it was expected that it be similar for 299 the unsupported Ni, Ir and Ni90Fe10 catalysts. It is interesting to observe that the electrolyser cell using 300 the Ni90Fe10/CeO2 shows more than twice the REL compared to the other cells for the same KOH 301 Catalystsconcentration.2019, 9, 814 This increase in ohmic resistance is caused by the rather large amount of10 less of 17- 302 conductive ceria in the catalytic layer (6 mg cm-2 for both the Ni catalyst and the CeO2 support), 303 theincreasing catalyst the therefore thickness affected of the both catalytic the electrode layer. Incorporating bulk conductivity ceria andinto contactthe catalyst resistance therefore between affected the 304 both the electrode bulk conductivity and contact resistance between the anode electrode and the anode electrode and the porous transport layers (PTLs). Nevertheless, the Ni90Fe10/CeO2 anode showed 305 lowerporous charge-transfer transport layers resistances (PTLs). Nevertheles in both 0.1 ands, the 1 M Ni KOH,90Fe10/CeO indicating2 anode good showed kinetics. lower charge-transfer 306 resistances in both 0.1 and 1 M KOH, indicating good kinetics.

307 2 308 FigureFigure 8. 8.( a(,ba,)b Polarization) Polarization curves curves up to up 2 A to cm − 2 Ain 1 cm M-2 and in 0.1 1 M M KOH, and 0.1 respectively. M KOH, ( c respectively.,d) Electrochemical (c,d) impedance spectroscopy in 1 M and 0.1 M KOH, respectively. Catalysts: Ir black (black square), Ni (red 309 Electrochemical impedance spectroscopy in 1 M and 0.1 M KOH, respectively. Catalysts: Ir black circle), Ni90Fe10 (blue triangle) and Ni90Fe10/CeO2 (green inverted triangle) NPs. Experiments were conducted 310 (black square), Ni (red circle), Ni90Fe10 (blue triangle) and Ni90Fe10/CeO2 (green inverted triangle) NPs. at 50 C. 311 Experiments◦ were conducted at 50 °C.

Table 3. Summary of polarization curve and electrochemical impedance results.

2 2 4 KOH conc. E @ 0.4 A/cm E @ 1.4 A/cm REL RCT Y [10− F Catalyst. (1 α) α [M] [V] [V] [mΩ] [mΩ] s − ] 1.0 1.608 1.804 6.0 0.1 27 0.1 3.6 0.1 0.577 0.005 Ir Black ± ± ± ± 0.1 1.691 2.011 11.6 0.2 40.2 0.3 11.5 0.1 0.618 0.007 ± ± ± ± 1.0 1.636 1.815 4.1 0.2 25.2 0.3 12.8 0.8 0.465 0.009 Ni ± ± ± ± 0.1 1.713 1.984 8.0 0.3 36.7 0.5 6.4 0.4 0.478 0.009 ± ± ± ± 1.0 1.622 1.823 6.3 0.2 26.1 0.3 6.6 0.5 0.53 0.01 Ni Fe ± ± ± ± 90 10 0.1 1.659 1.911 9.4 0.1 33.6 0.2 2.8 0.1 0.551 0.005 ± ± ± ± 1.0 1.627 1.854 13.0 0.4 21.2 0.2 10.1 0.3 0.74 0.04 Ni Fe /CeO ± ± ± ± 90 10 2 0.1 1.703 2.023 27.0 0.4 25.6 0.6 0.5 0.1 0.74 0.02 ± ± ± ±

Table3 clearly shows that decreasing that the KOH concentration decreases electrolysis performance for all materials by aﬀecting the charge transfer resistance (RCT) and kinetics of the electrochemical reactions, and the ohmic resistance or electrolyte resistance (REL) through a lack of OH− ions. However, in the case of the Ni90Fe10 and Ni90Fe10/CeO2, the decrease in KOH concentration does not have as large an eﬀect on the activity and kinetics of the OER compared to the Ni and Ir catalysts. The presence of Fe in the catalysts seems to shift the OER onset potential to lower values due to the formation of a NiFe mixed oxyhydroxide (NixFe(1 x)OOH) independent of pH [25]. − 3. Materials and Methods

3.1. Synthesis of Ni-Based Materials by Chemical Reduction

Ni, Ni90Fe10 and Ni90Fe10/CeO2 nanoparticles (NPs) were synthesized by a scalable chemical reduction method in ethanol using sodium borohydride as the reducing agent. First, the nickel chloride Catalysts 2019, 9, 814 11 of 17 hexahydrate (NiCl 6H O, 99.999% purity, Sigma Aldrich, St. Louis, MO, USA) precursor salt was 2· 2 dissolved in ethanol (EtOH, 99%, Greenﬁeld, Grayslake, IL, USA) with magnetic stirring for 45 min at room temperature. Depending on whether the desired compound included Fe or CeO2, or both, the ferrous sulphate heptahydrate (FeSO 7H O, 99%, Sigma Aldrich, St. Louis, MO, USA), precursor 4· 2 ≥ salt and cerium oxide nanopowder support (CeO2, 99.5%, Alfa Aesar, Haverhill, MA, USA ) were dissolved in ethanol using the same procedure. Next, the solutions were mixed together, and magnetic stirring was continued for 5 min to ensure that the solution was homogenous. Sodium borohydride (NaBH , 98%, ACROS, Geel, Belgium) was then dissolved in around 5–10 mL of ultrapure water 4 ≥ (ultrapure water, Milli-Q® Millipore, 18.2 MΩ cm at 293 K), then added to the reaction mixture, which was then further mixed for 30 min to ensure everything was completely reduced. Once the reaction was complete, the nanoparticles were washed three times with ethanol using a centrifuge run at 6000 rpm for 10 min per wash. Finally, the particles were left in a freeze drier overnight (around 15 h) to remove all residual ethanol from the sample. In this study, monometallic Ni nanoparticles were prepared, as were Ni90Fe10 (at%) and Ni90Fe10/CeO2 (50 wt% Ni90Fe10 metal supported on CeO2). Note that all glassware used in the NP synthesis was cleaned using the Aqua Regia procedure (2:1 HCl: HNO3, 37%, 70%, respectively, Sigma Aldrich, St. Louis, MO, USA). This synthesis procedure was easily scaled by adjusting the precursor salt quantities for the new desired material output, and by modifying the ethanol required to dissolve the salts. This ensured a similar concentration of precursor salts throughout all batches.

3.2. Material Characterization

The transmission electron microscopy (TEM) micrographs for the Ni and Ni90Fe10 catalysts were taken on the JEOL JEM 2100F Field Emission Transmission Electron Microscope (FETEM) (Tokyo, Japan) with an operating voltage of 200 kV. The scanning electron microscopy (SEM) micrographs of all AEMWE anodes were taken on the Hitachi Model S-3400N PC-Based Variable Pressure Scanning Electron Microscope (ON, Canada). SEM images were taken at an acceleration voltage of 15 kV. X-ray diffraction (XRD) results for the Ni and Ni90Fe10 catalysts were measured using the Rigaku Ultima IV multi-purpose diffractometer (Tokyo, Japan) using Copper Kα radiation and an x-ray wave length λ, of 1.5418 Å, at 40 kV and 44 mA. The XRD spectra were taken over a 2θ range of 20◦–80◦ at a rate of 1 0.5◦ 2θ min− . For the Ni90Fe10/CeO2 material an Field Electron and Ion Company (FEI, Hillsboro, OR, USA) Titan3 80–300 TEM operated at 300 keV, equipped with a corrected electron optical systems (CEOS) aberration corrector for the probe forming lens and a monochromated field-emission gun, was used to obtain both high-resolution TEM (HRTEM) and annular dark-field (ADF) images. The ADF images were collected using a high-angle annular dark-field (HAADF) Fischione detector in the scanning transmission electron microscopy (STEM) mode. Additionally, energy-dispersive X-ray (EDX) spectroscopy and electron energy-loss spectroscopy (EELS) were carried out on the ceria-supported material using the TEM instrument described above, equipped with an EDX spectrometer (EDAX Analyzer, DPP-II) and an EELS spectrometer with a Gatan Tridiem 866 Image Filter. To optimize the signal intensity, EDX spectra were acquired with the specimen tilted at 15 degrees. EELS spectra were acquired in STEM with a convergence semi angle of 15 mrad and an acceptance semi angle of 40 mrad. Techniques not presented in the main article are presented in the Supplementary Information.

3.3. Ex-Situ Experiments

3.3.1. Cell Set-Up The synthesized materials were ﬁrst studied for the OER in a conventional, two-compartment, three-electrode electrochemical glass cell. The working and counter electrodes (WE and CE) were in the main compartment, while the reference electrode (RE) was in the second compartment. Prior to experiments, the glassy carbon surface of the WE was ﬁrst polished with 30 micron alumina (Al2O3, Catalysts 2019, 9, 814 12 of 17

Carveth Metallurgical Limited, Canada), then with 3 micron alumina (Al2O3, Buehler, Lake Bluff, IL, USA), both of which were mixed with ultrapure water. After polishing, the electrode surface was wiped with ethanol, then ultrapure water. The WE consisted of 10 µL of a catalyst ink deposited on a 0.196 cm2 glassy carbon electrode. The catalyst ink was made with 6 mg catalyst, 1 mL of ultrapure water, 200 µL isopropanol (IPA, 99.9%, Fischer Scientific, Hampton, NH, USA) and 100 µL of Nafion ® (~5%, Sigma Aldrich, St. Louis, MO, USA). Once the ink was made, it was sonicated for around 5 min, then the deposited ink was dried in an oven at 60 ◦C. The CE was a platinum mesh and the RE was mercury/mercury oxide (Hg/HgO) electrode (Koslow Scientific Company, Englewood, NJ, USA). Note that all potentials in this article are reported with respect to the Hg/HgO reference electrode unless otherwise specified. See the Supplementary Information for conversions to RHE and how the overpotentials were calculated. All electrochemistry experiments were carried out using either 1 M or 0.1 M potassium hydroxide solutions (KOH, 85%, Sigma Aldrich, St. Louis, MO, USA) at a room ≥ temperature of around 23 ◦C.

3.3.2. OER Experiments Electrochemical tests were conducted using the Bio-Logic Potentiostat/Galvanostat using the EC Labs software (Bio-Logic Science Instruments, Seyssinet-Pariset, France). The OER tests included 1 a cyclic voltammogram (CV) run between [0.1, 0.8] V for 10 cycles at a scan rate of 20 mV s− , followed 1 by a linear sweep voltammogram (LSV) run between [0.3, 0.8] V at a scan rate of 1 mV s− . The tests mentioned were applied to the nickel-based materials, and an iridium black benchmark (Ir black, 99.8%, Alfa Aesar, Haverhill, MA, USA) catalyst. Before running the CVs and LSVs on the Ni-based materials, the WEs were polarized at 1.3 V for 5 min, and 0.8 V for 10 min [54], to first remove the − − oxides from the surfaces, and then to remove the hydrogen produced in the first polarization step. Only the fifth cycle of all cyclic voltammetry experiments is reported.

3.4. In Situ Experiments

3.4.1. Cell Set-Up The AEMWE cell hardware consisted of a modified 25 cm2 fuel cell hardware (BalticFuelCell Gmbh). The cell has parallel flow fields made with gold coated titanium, which were used as both anode and cathode endplates. The endplate material selection was based on minimizing ohmic losses in the AEMWE cell or rather to minimize the interfacial contact resistances (ICRs). The electrolyser setup also included a 5 L Teflon tank with heaters, which was used as the KOH reservoir. A double 1 headed peristaltic pump was used to pump 300 mL min− of KOH through the AEMWE cell at both anode and cathode. The membrane electrode assembly consisted of a commercial anionic exchange membrane (fumasep FAA-3PE-30, Fumatech, Bietigheim-Bissingen, Germany), onto which the anode and cathode were hand-sprayed with an air brush. The membrane was secured with a gasket, as shown in Figure9. Furthermore, in the interest of conserving material, only 5 cm2 of the 25 cm2 cell was used for the electrode area. The MEAs were assembled in the cell between two, 1 mm thick, commercially available, titanium, porous transport layers (PTLs) (Bekaert, Zwevegem, Belgium). Catalysts 2019, 9, x FOR PEER REVIEW 13 of 17

396 The AEMWE cell hardware consisted of a modified 25 cm2 fuel cell hardware (BalticFuelCell 397 Gmbh). The cell has parallel flow fields made with gold coated titanium, which were used as both 398 anode and cathode endplates. The endplate material selection was based on minimizing ohmic losses 399 in the AEMWE cell or rather to minimize the interfacial contact resistances (ICRs). The electrolyser 400 setup also included a 5 L Teflon tank with heaters, which was used as the KOH reservoir. A double 401 headed peristaltic pump was used to pump 300 ml min-1 of KOH through the AEMWE cell at both 402 anode and cathode. 403 The membrane electrode assembly consisted of a commercial anionic exchange membrane 404 (fumasep FAA-3PE-30, Fumatech, Bietigheim-Bissingen, Germany), onto which the anode and 405 cathode were hand-sprayed with an air brush. The membrane was secured with a gasket, as shown 406 in Figure 9. Furthermore, in the interest of conserving material, only 5 cm2 of the 25 cm2 cell was used 407 forCatalysts the electrode2019, 9, 814 area. The MEAs were assembled in the cell between two, 1 mm thick, commercially13 of 17 408 available, titanium, porous transport layers (PTLs) (Bekaert, Zwevegem, Belgium). Gas Membrane Diffusion Electrode Layer Assembly Bipolar Plate

409

410 FigureFigure 9. 9. CompactCompact Alkaline Alkaline Exchange Exchange Membrane Membrane Water Water Electrolyser Electrolyser Cell Cell Design. Design.

411 TheThe anodes anodes consisted consisted of ofthe thepreviously previously mentioned mentioned Ni-based Ni- materials,based materials, and the Ir-black and the benchmark Ir-black 412 benchmark(Alfa Aesar, (Alfa City, Haverhill, Aesar, City, MA, Haverhill, USA), while MA, the USA), cathode while consisted the ofcathode platinum consisted supported of platinumon carbon 413 supported(Pt/C, 60 wt% on carbon metal on(Pt/C, support, 60 wt% Alfa metal Aesar, onHaverhill, support, Alfa MA, Aesar, USA). Haverhill, The ink for MA, the anodesUSA). The were ink made for 414 thewith anodes a solution were of made 50:50 with by weight a solution of ultrapure of 50:50 water: by weight IPA, whichof ultrapure included water: ~2 wt% IPA, metal which and inc anluded anionic ~2 415 wt%ionomer metal (fumion and an FAA-3-SOLUT-10, anionic ionomer Bietigheim-Bissingen, (fumion FAA-3-SOLUT Fumatech,-10, Bietigheim Germany).-Bissingen, The final Fumatech, electrode 416 Germany).had ~10 wt% The ionomer final electrode acting had as both ~10 awt% catalyst ionomer binder acting and as a both facilitator a catalyst for hydroxide binder and ion a facilitator transport. 417 forMultiple hydroxide sonication ion transport. and ultrasonication Multiple sonication steps were and usedultrasonication to ensure thesteps solution were used was to very ensure fine andthe 418 solutionwell dispersed. was very The fine ink and for well the dispersed. cathodes wereThe ink made for inthe a cathodes similar way were to made the anode in a similar inks; however, way to 419 thethe anode ionomer inks; was however, present the at ~23 ionomer wt% inwas the present final electrode. at ~23 wt% Note in the that final the electrode. ink preparation Note that procedures the ink 420 preparationwere obtained procedures by ink optimization were obtained of theby ink benchmark optimization electrodes; of the benchmark namely, the electrodes; Ir-black anode namely, and the the 421 IrPt-black/C cathode. anode and the Pt/C cathode. 2 422 InIn the the AEMWE AEMWE experiments, experiments, the the anode anode loading loading for for the benchmark was 3 mg cmcm−-2, ,while while the the 2 423 loadingloading for for the the Ni Ni-based-based materials materials was was 6 6mg mg cm cm-2−. The. The cathode cathode loading loading was was held held at ata Pt a Ptloading loading of of1 2 424 mg1 mg cm cm-2 −forfor all all experiments. experiments. Note Note that that those those are are intended intended metal metal loadings. loadings. For For the the actual actual metal metal 425 loadings,loadings, see see Table Table S1 S1 in in the the Supplementary Supplementary Information Information of of this this article. article. Actual Actual metal metal loadings loadings were were 426 calculatcalculateded using using ImageJ ImageJ and and a a differential differential weight weight of of samples samples taken taken from from the the template template used used to to delimit delimit 427 thethe electrode electrode area area during during spraying. spraying. 3.4.2. AEMWE Experiments 428 3.4.2. AEMWE Experiments Before starting electrolysis testing, the cell was mounted into the system, and around 2 L of 1 M 429 Before starting electrolysis testing, the cell was mounted into the system, and around 2 L of 1 M KOH was circulated through the system overnight to exchange the bromine anions in the membrane 430 KOH was circulated through the system overnight to exchange the bromine anions in the membrane with hydroxide ions from the alkaline electrolyte. The system was then cleaned by pumping ultrapure 431 with hydroxide ions from the alkaline electrolyte. The system was then cleaned by pumping water through it. Electrolysis testing was then carried out using the HCP-803 Potentiostat equipped with the EC Labs software (Bio-Logic Science Instruments, Seyssinet-Pariset, France). First, polarization 2 curves were taken, by stepping up the current from 0 to 10 A (0 to 2 A cm− ). Next, electrochemical 2 impedance spectroscopy (EIS) was performed by applying a direct current (DC) of 5 A (1 A cm− ) and an alternating current (AC) of 5% DC, while the frequency ranged from 20,000 to 0.1 Hz. ± The impedance data was fitted to a simple equivalent circuit using magnitude weighting in the Maple application “Complex Nonlinear Least Squares Fitting of Immittance Data,” authored by David Harrington [55]. Polarization and EIS experiments were repeated three times in 1 M KOH, after which the system was flushed with ultrapure water, then all measurements were taken three times in 0.1 M KOH. All AEMWE experiments were performed at 50 2 C, by controlling the temperature in the cell ± ◦ and the KOH storage tank, individually. Catalysts 2019, 9, 814 14 of 17

4. Conclusions In this study, Ni-based materials were synthesized for the promotion of the oxygen evolution reaction for applications in energy storage through anion-exchange, membrane-alkaline water electrolysis (AEMWE). Ni, Ni90Fe10 and Ni90Fe10/CeO2 NPs were synthesized via chemical reduction in ethanol using sodium borohydride. The materials were characterized by TEM imaging and were found to be around 4–6 nm in size. Ex-situ electrochemical experiments were conducted in a conventional three-electrode cell at both 1 M and 0.1 M KOH and results were compared to an Ir-black benchmark electrode. Results indicate that on average, the Ni90Fe10 material is the best-performing material for OER, showing very low onset potentials and excellent catalytic activity. This result was also obtained for the lower electrolyte concentration, wherein the NiFe-based electrode outperformed all catalysts, including the Ir benchmark electrode. 2 In-situ AEMWE experiments were performed in a modiﬁed 25 cm fuel cell for simultaneous O2 and H2 production. At twice the metallic loading compared to the Ir benchmark anode, the Ni-based 2 materials showed state-of-the-art performance, achieving current densities of 2 A cm− at 1.85–1.9 V in 1 M KOH at 50 ◦C. In 0.1 M KOH, the cell set up using the Ni90Fe10 anode attained 1.99 V at 2 A 2 cm− . These are promising results, considering the non-noble metal nature of these catalysts and that long-term stability measurements are on-going. Based on impedance results, the ceria-supported NiFe catalyst showed higher cell resistances (higher REL values), yet showed a low onset potential for the OER, and a low charge transfer resistance, indicating the Ni90Fe10/CeO2 may be used as an anode catalyst in AEMWE if the support is further optimized in terms of electrical conductivity.

Supplementary Materials: The following are available online at http://www.mdpi.com/2073-4344/9/10/814/s1, 1 Figure S1: CVs from 0.1 to 0.8 V of Ni (black), CeO2 (red) and Ni/CeO2 (blue) run at 20 mV s− in 1 M KOH. Figure S2: TEM Images of Ni90Fe10/CeO2 showing (a) mostly the Ni and Fe and (b) mostly the CeO2 support. Figure S3: STEM Images of Ni90Fe10/CeO2 showing (a) mostly the Ni and Fe and (b) mostly the CeO2 support. Figure S4: Spatially-resolved EDX of Ni90Fe10/CeO2. (a) Shows the two selected regions of the STEM image, (b) shows the analysis of Region 1 and (c) shows the analysis of Region 2. Figure S5: EELS mapping of Ni90Fe10/CeO2. (a) Shows the Fe mapping, (b) shows the Ni mapping and (c) shows the analyzed region in orange. Figure S6: Comparison between current densities by geometric surface area and by mass of metal at 0.8 V against Hg/HgO in (a) 1 M KOH and (b) 0.1 M KOH. Table S1: Anode metal loadings on the membrane electrode assemblies. Author Contributions: Material synthesis, characterization, ex-situ oxygen evolution measurements, in-situ electrolysis measurements, writing and editing, E.C. In-situ electrolysis measurements, funding acquisition, supervision, writing and editing, A.O.B. Funding acquisition, supervision, writing and editing, F.S. funding acquisition, supervision, writing and editing, E.A.B. Funding: This research was conducted as part of the Engineered Nickel Catalysts for Electrochemical Clean Energy project administered from Queen’s University and supported by Grant number RGPNM 477963-2015 under the Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Frontiers Program. Additionally, this work was performed within the HAPEEL project “Hydrogen Production by Alkaline Polymer Electrolyte Electrolysis,” financially supported by the Research Council of Norway-ENERGIX program, contract number 268019. The Research Council of Norway is also acknowledged for the support to the Norwegian Fuel cell and Hydrogen Centre and the INTPART project 261620. Acknowledgments: The authors would also like to thank Martin Couillard from the Natural Research Council of Canada for his help in characterizing the Ni90Fe10/CeO2 sample. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Clark, D.; Brandon, N. What’s the “Hydrogen Economy”? The Guardian: London, UK, 2012. 2. Zeng, K.; Zhang, D. Recent progress in alkaline water electrolysis for hydrogen production and applications. Prog. Energy Combust. Sci. 2010, 36, 307–326. [CrossRef] 3. Leng, Y.; Chen, G.; Mendoza, A.J.; Tighe, T.B.; Hickner, M.A.; Wang, C.Y. Solid-state water electrolysis with an alkaline membrane. J. Am. Chem. Soc. 2012, 134, 9054–9057. [CrossRef][PubMed] 4. Varcoe, J.R.; Atanassov, P.; Dekel, D.R.; Herring, A.M.; Hickner, M.A.; Kohl, P.A.; Kucernak, A.R.; Mustain, W.E.; Nijmeijer, K.; Scott, K.; et al. Anion-exchange membranes in electrochemical energy systems. Energy Environ. Sci. 2014, 7, 3135–3191. [CrossRef] Catalysts 2019, 9, 814 15 of 17

5. Cho, M.K.; Park, H.Y.; Lee, H.J.; Kim, H.J.; Lim, A.; Henkensmeier, D.; Yoo, S.J.; Kim, J.Y.; Lee, S.Y.; Park, H.S.; et al. Alkaline anion exchange membrane water electrolysis: Effects of electrolyte feed method and electrode binder content. J. Power Sources 2018, 382, 22–29. [CrossRef] 6. Vincent, I.; Bessarabov, D. Low cost hydrogen production by anion exchange membrane electrolysis: A review. Renew. Sustain. Energy Rev. 2018, 81, 1690–1704. [CrossRef] 7. Phillips, R.; Edwards, A.; Rome, B.; Jones, D.R.; Dunnill, C.W. Minimising the ohmic resistance of an alkaline electrolysis cell through effective cell design. Int. J. Hydrog. Energy 2017, 42, 23986–23994. [CrossRef] 8. Park, J.E.; Kang, S.Y.; Oh, S.H.; Kim, J.K.; Lim, M.S.; Ahn, C.Y.; Cho, Y.H.; Sung, Y.E. High-performance anion-exchange membrane water electrolysis. Electrochim. Acta 2019, 295, 99–106. [CrossRef] 9. Hnát, J.; Paidar, M.; Schauer, J.; Žitka, J.; Bouzek, K. Polymer anion-selective membranes for electrolytic splitting of water. Part II: Enhancement of ionic conductivity and performance under conditions of alkaline water electrolysis. J. Appl. Electrochem. 2012, 42, 545–554. [CrossRef] 10. Chanda, D.; Hnát, J.; Bystron, T.; Paidar, M.; Bouzek, K. Optimization of synthesis of the nickel-cobalt oxide based anode electrocatalyst and of the related membrane-electrode assembly for alkaline water electrolysis. J. Power Sources 2017, 347, 247–258. [CrossRef] 11. Kim, J.H.; Lee, J.N.; Yoo, C.Y.; Lee, K.B.; Lee, W.M. Low-cost and energy-efficient asymmetric nickel electrode for alkaline water electrolysis. Int. J. Hydrog. Energy 2015, 40, 10720–10725. [CrossRef] 12. Schalenbach, M.; Kasian, O.; Mayrhofer, K.J.J. An alkaline water electrolyzer with nickel electrodes enables efficient high current density operation. Int. J. Hydrog. Energy 2018, 43, 11932–11938. [CrossRef] 13. Liu, Z.; Sajjad, S.D.; Gao, Y.; Yang, H.; Kaczur, J.J.; Masel, R.I. The effect of membrane on an alkaline water electrolyzer. Int. J. Hydrog. Energy 2017, 42, 29661–29665. [CrossRef] 14. Katsounaros, I.; Cherevko, S.; Zeradjanin, A.R.; Mayrhofer, K.J.J. Oxygen electrochemistry as a cornerstone for sustainable energy conversion. Angew. Chem. Int. Ed. 2014, 53, 102–121. [CrossRef][PubMed] 15. Marini, S.; Salvi, P.; Nelli, P.; Pesenti, R.; Villa, M.; Berrettoni, M.; Zangari, G.; Kiros, Y. Advanced alkaline water electrolysis. Electrochim. Acta 2012, 82, 384–391. [CrossRef] 16. Faid, A.; Oyarce Barnett, A.; Seland, F.; Sunde, S. Highly active nickel-based catalyst for hydrogen evolution in anion exchange membrane electrolysis. Catalysts 2018, 8, 614. [CrossRef] 17. Fabbri, E.; Habereder, A.; Waltar, K.; Kötz, R.; Schmidt, T.J. Developments and perspectives of oxide-based catalysts for the oxygen evolution reaction. Catal. Sci. Technol. 2014, 4, 3800–3821. [CrossRef] 18. Li, X.; Walsh, F.C.; Pletcher, D. Nickel based electrocatalysts for oxygen evolution in high current density, alkaline water electrolysers. Phys. Chem. Chem. Phys. 2011, 13, 1162–1167. [CrossRef][PubMed] 19. Cho, M.K.; Lim, A.; Lee, S.Y.; Kim, H.; Yoo, S.J.; Sung, Y.; Park, H.S.; Jang, J.H. A review on membranes and catalysts for anion exchange membrane water electrolysis single cells. J. Electrochem. Sci. Technol. 2017, 8, 183–196. [CrossRef] 20. Trotochaud, L.; Ranney, J.K.; Williams, K.N.; Boettcher, S.W. Solution-cast metal oxide thin film electrocatalysts for oxygen evolution. J. Am. Chem. Soc. 2012, 134, 17253–17261. [CrossRef][PubMed] 21. Gong, M.; Dai, H. A mini review of NiFe-based materials as highly active oxygen evolution reaction electrocatalysts. Nano Res. 2015, 8, 23–39. [CrossRef] 22. Long, X.; Ma, Z.; Yu, H.; Gao, X.; Pan, X.; Chen, X.; Yang, S.; Yi, Z. Porous FeNi oxide nanosheets as advanced electrochemical catalysts for sustained water oxidation. J. Mater. Chem. A 2016, 4, 14939–14943. [CrossRef] 23. Gong, M.; Li, Y.; Wang, H.; Liang, Y.; Wu, J.Z.; Zhou, J.; Wang, J.; Regier, T.; Wei, F.; Dai, H. An advanced Ni-Fe layered double hydroxide electrocatalyst for water oxidation. J. Am. Chem. Soc. 2013, 135, 8452–8455. [CrossRef][PubMed] 24. Bode, H.; Dehmelt, K.; White, J. Zur kenntnis der nickelhydroxidelektrode—I.Über das nickel (II)-hydroxidhydrat. Electrochim. Acta 1966, 11, 1079–1087. [CrossRef] 25. Diaz-Morales, O.; Ferrus-Suspedra, D.; Koper, M.T.M. The importance of nickel oxyhydroxide deprotonation on its activity towards electrochemical water oxidation. Chem. Sci. 2016, 7, 2639–2645. [CrossRef][PubMed] 26. Surendranath, Y.; Kanan, M.W.; Nocera, D.G. Mechanistic studies of the oxygen evolution reaction by a cobalt-phosphate catalyst at neutral pH. J. Am. Chem. Soc. 2010, 132, 16501–16509. [CrossRef][PubMed] 27. Takashima, T.; Hashimoto, K.; Nakamura, R. Mechanisms of pH-dependent activity for water oxidation to

molecular oxygen by MnO2 electrocatalysts. J. Am. Chem. Soc. 2012, 134, 1519–1527. [CrossRef][PubMed] Catalysts 2019, 9, 814 16 of 17

28. Lyons, M.E.G.; Cakara, A.; O’Brien, P.; Godwin, I.; Doyle, R.L. Redox, pH sensing and electrolytic water splitting properties of electrochemically generated nickel hydroxide thin ﬁlms in aqueous alkaline solution. Int. J. Electrochem. Sci. 2012, 7, 11768–11795. 29. Lyons, M.E.G.; Doyle, R.L.; Brandon, M.P. Redox switching and oxygen evolution at oxidized metal and metal oxide electrodes: Iron in base. Phys. Chem. Chem. Phys. 2011, 13, 21530–21551. [CrossRef] 30. Panaritis, C.; Edake, M.; Couillard, M.; Einakchi, R.; Baranova, E.A. Insight towards the role of ceria-based supports for reverse water gas shift reaction over RuFe nanoparticles. J. CO2 Util. 2018, 26, 350–358. [CrossRef] 31. Dole, H.A.E.; Baranova, E.A. Ethylene oxidation in an oxygen-deﬁcient environment: Why ceria is an active support? ChemCatChem 2016, 8, 1977–1986. [CrossRef] 32. Rao, G.R.; Mishra, B.G. Structural redox and catalytic chemistry of ceria based materials. Bull. Catal. Soc. India 2003, 2, 122–134.

33. Feng, J.-X.; Ye, S.-H.; Xu, H.; Tong, Y.-X.; Li, G.-R. Design and synthesis of FeOOH/CeO2 heterolayered nanotube electrocatalysts for the oxygen evolution reaction. Adv. Mater. 2016, 28, 4698–4703. [CrossRef] [PubMed] 34. Chen, Z.; Kronawitter, C.X.; Yang, X.; Yeh, Y.W.; Yao, N.; Koel, B.E. The promoting effect of tetravalent cerium on the oxygen evolution activity of copper oxide catalysts. Phys. Chem. Chem. Phys. 2017, 19, 31545–31552. [CrossRef][PubMed] 35. Favaro, M.; Drisdell, W.S.; Marcus, M.A.; Gregoire, J.M.; Crumlin, E.J.; Haber, J.A.; Yano, J. An operando investigation of (Ni-Fe-Co-Ce)Ox system as highly efficient electrocatalyst for oxygen evolution reaction. ACS Catal. 2017, 7, 1248–1258. [CrossRef] 36. Haber, J.A.; Cai, Y.; Jung, S.; Xiang, C.; Mitrovic, S.; Jin, J.; Bell, A.T.; Gregoire, J.M. Discovering Ce-rich oxygen evolution catalysts, from high throughput screening to water electrolysis. Energy Environ. Sci. 2014, 7, 682–688. [CrossRef] 37. McCrory, C.C.L.; Jung, S.; Peters, J.C.; Jaramillo, T.F. Benchmarking heterogeneous electrocatalysts for the oxygen evolution reaction. J. Am. Chem. Soc. 2013, 135, 16977–16987. [CrossRef][PubMed] 38. Seetharaman, S.; Balaji, R.; Ramya, K.; Dhathathreyan, K.S.; Velan, M. Graphene oxide modified non-noble metal electrode for alkaline anion exchange membrane water electrolyzers. Int. J. Hydrog. Energy 2013, 38, 14934–14942. [CrossRef] 39. Xiao, L.; Zhang, S.; Pan, J.; Yang, C.; He, M.; Zhuang, L.; Lu, J. First implementation of alkaline polymer electrolyte water electrolysis working only with pure water. Energy Environ. Sci. 2012, 5, 7869–7871. [CrossRef] 40. Ayers, K.E.; Anderson, E.B.; Capuano, C.B.; Niedzwiecki, M.; Hickner, M.A.; Wang, C.-Y.; Leng, Y.; Zhao, W. Characterization of anion exchange membrane technology for low cost electrolysis. ECS Trans. 2013, 45, 121–130. [CrossRef] 41. Pavel, C.C.; Cecconi, F.; Emiliani, C.; Santiccioli, S.; Scaffidi, A.; Catanorchi, S.; Comotti, M. Highly efficient platinum group metal free based membrane-electrode assembly for anion exchange membrane water electrolysis. Angew. Chem. Int. Ed. 2014, 53, 1378–1381. [CrossRef][PubMed] 42. Sun, Y.P.; Li, X.Q.; Cao, J.; Zhang, W.X.; Wang, H.P. Characterization of zero-valent iron nanoparticles. Adv. Colloid Interface Sci. 2006, 120, 47–56. [CrossRef][PubMed] 43. Hall, D.S.; Lockwood, D.J.; Bock, C.; MacDougall, B.R. Nickel hydroxides and related materials: A review of their structures, synthesis and properties. Proc. R. Soc. A Math. Phys. Eng. Sci. 2015, 471, 20140792. [CrossRef][PubMed] 44. Baranova, E.A.; Le Page, Y.; Ilin, D.; Bock, C.; MacDougall, B.; Mercier, P.H.J. Size and composition for 1–5 nm Ø PtRu alloy nano-particles from Cu Kα X-ray patterns. J. Alloys Compd. 2009, 471, 387–394. [CrossRef]

45. Reksten, A.H.; Thuv, H.; Seland, F.; Sunde, S. The oxygen evolution reaction mechanism at IrxRu1 xO2 − powders produced by hydrolysis synthesis. J. Electroanal. Chem. 2018, 819, 547–561. [CrossRef] 46. Lyons, M.E.G.; Brandon, M.P. A comparative study of the oxygen evolution reaction on oxidised nickel, cobalt and iron electrodes in base. J. Electroanal. Chem. 2010, 641, 119–130. [CrossRef] 47. Lu, X.; Zhao, C. Electrodeposition of hierarchically structured three-dimensional nickel-iron electrodes for eﬃcient oxygen evolution at high current densities. Nat. Commun. 2015, 6, 6616. [CrossRef] Catalysts 2019, 9, 814 17 of 17

48. Tahir, M.; Pan, L.; Idrees, F.; Zhang, X.; Wang, L.; Zou, J.J.; Wang, Z.L. Electrocatalytic oxygen evolution reaction for energy conversion and storage: A comprehensive review. Nano Energy 2017, 37, 136–157. [CrossRef] 49. Cherevko, S.; Geiger, S.; Kasian, O.; Kulyk, N.; Grote, J.-P.; Savan, A.; Shrestha, R.B.; Merzlikin, S.; Breitbach, B.;

Ludwig, A.; et al. Oxygen and hydrogen evolution reactions on Ru, RuO2, Ir, and IrO2 thin ﬁlm electrodes in acidic and alkaline electrolytes: A comparative study on activity and stability. Catal. Today 2015, 262, 170–180. [CrossRef] 50. Lyons, M.E.G.; Floquet, S. Mechanism of oxygen reactions at porous oxide electrodes. Part 2—Oxygen evolution

at RuO2, IrO2 and IrxRu1 xO2 electrodes in aqueous acid and alkaline solution. Phys. Chem. Chem. Phys. 2011, 13, − 5314–5335. [CrossRef] 51. Pi, Y.; Zhang, N.; Guo, S.; Guo, J.; Huang, X. Ultrathin laminar Ir superstructure as highly eﬃcient oxygen evolution electrocatalyst in broad pH range. Nano Lett. 2016, 16, 4424–4430. [CrossRef] 52. Tahir, M.; Mahmood, N.; Zhang, X.; Mahmood, T.; Butt, F.K.; Aslam, I.; Tanveer, M.; Idrees, F.; Khalid, S.;

Shakir, I.; et al. Bifunctional catalysts of Co3O4@GCN tubular nanostructured (TNS) hybrids for oxygen and hydrogen evolution reactions. Nano Res. 2015, 8, 3725–3736. [CrossRef] 53. Vincent, I.; Kruger, A.; Bessarabov, D. Development of eﬃcient membrane electrode assembly for low cost hydrogen production by anion exchange membrane electrolysis. Int. J. Hydrogen Energy 2017, 42, 10752–10761. [CrossRef] 54. Alsabet, M.; Grden, M.; Jerkiewicz, G. Electrochemical growth of surface oxides on nickel. Part 1: Formation

of α-Ni(OH)2 in relation to the polarization potential, polarization time, and temperature. Electrocatalysis 2011, 2, 317–330. [CrossRef] 55. Harrington, D. Complex Nonlinear Least Squares Fitting of Immittance Data. Available online: https: //www.maplesoft.com/applications/view.aspx?SID=154540 (accessed on 26 August 2019).

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).