IrO2 coated TiO2 core-shell microparticles advance performance of low loading proton exchange membrane water electrolyzers

Journal Pre-proof IrO2 coated TiO2 core-shell microparticles advance performance of low loading proton exchange membrane water electrolyzers Chuyen Van Pham (Conceptualization) (Methodology) (Validation) (Formal analysis) (Investigation) (Writing - original draft) (Writing review and editing) (Visualization), Melanie Buhler ¨ (Conceptualization) (Methodology) (Validation) (Formal analysis) (Investigation) (Writing - original draft) (Writing - review and editing) (Visualization), Julius Kn¨oppel (Methodology) (Formal analysis) (Investigation) (Writing - review and editing) (Visualization), Markus Bierling (Investigation) (Formal analysis) (Writing - review and editing), Dominik Seeberger (Investigation) (Formal analysis) ´ (Writing - review and editing), Daniel Escalera-Lopez (Formal analysis) (Writing - review and editing) (Visualization), Karl J.J. Mayrhofer (Conceptualization) (Formal analysis) (Resources) (Writing - review and editing), Serhiy Cherevko (Conceptualization) (Formal analysis) (Resources) (Writing - review and editing), Simon Thiele (Conceptualization) (Formal analysis) (Resources) (Writing review and editing) (Supervision) (Project administration) (Funding acquisition)

PII:

S0926-3373(20)30177-6

DOI:

https://doi.org/10.1016/j.apcatb.2020.118762

Reference:

APCATB 118762

To appear in:

Applied Catalysis B: Environmental

Received Date:

26 September 2019

Revised Date:

10 February 2020

Accepted Date:

14 February 2020

Please cite this article as: Van Pham C, Buhler ¨ M, Kn¨oppel J, Bierling M, Seeberger D, ´ Escalera-Lopez D, Mayrhofer KJJ, Cherevko S, Thiele S, IrO2 coated TiO2 core-shell microparticles advance performance of low loading proton exchange membrane water electrolyzers, Applied Catalysis B: Environmental (2020), doi: https://doi.org/10.1016/j.apcatb.2020.118762

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2020 Published by Elsevier.

IrO2 coated TiO2 core-shell microparticles advance performance of low loading proton exchange membrane water electrolyzers Chuyen Van Phama,1,*, Melanie Bühlera,b,1, Julius Knöppelc,d, Markus Bierlingc,d, Dominik Seebergerc,d, Daniel Escalera-Lópezc, Karl J. J. Mayrhoferc,d, Serhiy Cherevkoc, and Simon Thielea,b,c,d*

ro of

[a] Electrochemical Energy Systems, IMTEK - Department of Microsystems Engineering, University of Freiburg, Georges-Koehler-Allee 103, 79110 Freiburg, Germany. [b] Hahn-Schickard, Georges-Koehler-Allee 103, 79110 Freiburg, Germany. [c] Forschungszentrum Jülich GmbH, Helmholtz Institute Erlangen-Nürnberg for Renewable Energy (IEK-11), Forschungszentrum Jülich, Egerlandstr. 3, 91058 Erlangen, Germany. [d] Department of Chemical and Biological Engineering, Friedrich-Alexander-Universität ErlangenNürnberg, Egerlandstr. 3, 91058 Erlangen, Germany. *Corresponding authors. E-mails: [email protected] (S. Thiele), [email protected] (C.V. Pham). 1

Chuyen Van Pham and Melanie Bühler contributed equally.

Jo

ur

na

lP

re

-p

Graphical abstract

1

Highlights 

A novel, scalable synthesis route for obtaining IrO2 coated TiO2 core-shell microparticles based on electrostatic interaction of Ir-precursor and the surface of TiO2 particles .



High catalyst-dispersed electrodes achieved by combination of the unique microstructural IrO2@TiO2 catalyst with a new electrode configuration, catalyst coated titanium porous transport layer.



A high performance of proton exchange membrane water electrolysers (PEMWEs) using IrO2@TiO2 OER catalyst with low Ir-loading of 0.4 mg cm-2, outperforming both unsupported IrO2 and supported IrO2/TiO2 commercial benchmark catalysts.



The potential of our novel IrO2@TiO2 catalyst for industrial PEMWE applications was demonstrated in this work,

ro of

given the facile synthesis route and high PEMWE performance at low Ir loadings.

Abstract

lP

re

-p

Herein we present novel IrO2 coated TiO2 core-shell microparticles (IrO2@TiO2) as an oxygen evolution reaction catalyst. We compare the IrO2@TiO2 catalyst to commercial TiO2 supported IrO2 catalyst (IrO2/TiO2) and pure IrO2 catalyst powder. A stability analysis via on-line inductively coupled plasma mass spectrometry based on the S-number, a descriptor considering both energy efficiency and catalyst utilization efficiency, shows that the IrO2@TiO2 catalyst shows high potential for practical applications. This was further confirmed by full-cell tests showing superior performance of the IrO2@TiO2 catalyst with moderate and low loadings of 1.2 mgIr cm-2 and 0.4 mgIr cm-2, respectively. The core-shell catalyst is synthesized via facile route suitable for large quantities. Moreover, stable inks from the synthesized catalyst powder make this system appealing for large scale manufacturing of cells. Given the facile synthesis route, high activity, and good stability, the IrO2@TiO2 catalyst is potentially suitable for industry proton exchange membrane water electrolysis application.

1. Introduction

na

Keywords: iridium oxide catalyst; oxygen evolution reaction; proton exchange membrane; water electrolysis; core shell catalyst.

Jo

ur

Wind and solar energy technologies are increasingly important for a sustainable energy supply. However, both technologies suffer from intermittent natural supply, which makes energy storage e. g. in the form of a chemical fuel, such as hydrogen a necessity. Among water electrolysis technologies, proton exchange membrane water electrolyzers (PEMWEs) are expected to play an important role due to their ability for fast load change, high power and energy efficiency, and high gas purity.[1] Due to the sluggish kinetics of the oxygen evolution reaction (OER) and the corrosive operation conditions at the anode, so far only IrO2 is known to be a suitable catalyst that meets the criteria of activity and stability for PEM OER catalysts.[2] However, the scarcity and high cost of this precious metal are challenges for sustainable development of PEMWEs. A development of alternative non-noble OER catalysts has only achieved limited successes so far.[3–5] Extensive research has been therefore devoted to improve the effectiveness of Ir based OER catalysts. The research activities so far can be categorized into two categories: enhanced intrinsic activity of the catalysts by material design on the one hand and by advanced electrode morphologies and cell configurations on the other hand. The intrinsic activity of OER catalysts has been dramatically improved over the last few years. One of the 2

most active OER catalysts are mixed oxides of IrO2 and RuO2 with different Ir:Ru ratios (IrxRu1xO2).[6,7] As a result, high performing PEMWEs have been achieved even with low OER catalyst loadings (e.g. 0.2-0.5 mg cm-2).[8–10] By this way, the loading of precious Ir metal can be reduced. However, the stability of IrxRu1-xO2 is still a concern due to the dissolution of the Ru component under typical operation conditions.[11]

na

lP

re

-p

ro of

At catalyst layer scale, using catalyst supports could largely enhance catalyst utilization[12] and improve the interfacial and structural stability of catalyst layers at low loadings.[13] However, again the highly corrosive conditions of the PEMWE anode restrict the support selection. Several candidates have been tested as supports for OER catalysts, including NbO2,[14] antimony doped tin oxide (ATO),[15] fluorine doped tin oxide (FTO),[16] TaC,[17] doped TiO2,[18] TiN,[19] Ti metal,[12] and TiO2.[20] Of those, TiO2[20] and FTO[16] are promising due to their stability. Millet et al. reported the application of micrometer sized Ti metal particles as support for IrO2 catalysts by physically mixing the two components.[12] This strategy improved the dispersion of IrO2 within the anodes, leading to an unprecedented high catalyst utilization, and allowed achieving 1 A cm-2 at 1.73 V with low catalyst loadings of 0.12 mgIr cm-2 on the anode. This demonstrated the necessity of catalyst support for reducing Iridium loading. Still, Ti metal micro particles might be oxidized gradually in electrolysis working conditions, increasing the internal electrical resistance of the anode. In spite of low conductivity of TiO2, catalysts using TiO2 support exhibit good performance.[14,20] TiO2 is thus a promising OER catalyst support for large scale application, given its high stability, low cost, abundance, and a mature production industry. The electronic interaction between oxide support and IrO2 could improve the stability of supported catalysts[21] and even, according to some reports, enhance intrinsic activity of IrO2.[22] Mazur et al.[23] studied the use of IrO2/TiO2 supported catalyst, synthesized by a modified Adam fusion method, and found an enhanced performance compared to unsupported IrO2 catalyst. Bernt et al. extensively optimized catalyst layers of IrO2/TiO2 supported catalyst (Umicore), targeting a maximal catalyst utilization to reduce the Ir loading.[14] The study found that with this catalyst, the optimal Ir loading is in the range of 1-2 mg cm-2 to obtain optimal overall performances. At loadings lower than 0.5 mgIr cm-2, the catalyst layers were too thin and structurally not a homogeneous layer anymore. Generally, due to the low conductivity of the TiO2 support, large amounts of IrO2 catalyst are needed to form an electrically percolating structure.[20,23] The authors suggested that a core-shell design of IrO2/TiO2 could increase the IrO2 utilization, maximizing the IrO2 dispersion the catalyst layer.

Jo

ur

Elsewhere, core-shell catalysts with non-noble metal cores of such as Ni or Sr, and IrO2 rich shells have been developed, allowing high precious metal utilizations.[24–26] Unfortunately, in these approaches only a portion of the Ir core was replaced with Ni or Sr, moreover Ni and Sr are likely dissolved in acidic solution during operation.[27] So far, core-shell structured catalysts with a core of stable nonprecious metals and a nanoscale-thin shell of high crystalline IrO2 have not been realized due to catalyst – metal oxide support interaction leading to the disturbance of crystal structure of the IrO2 shells.[28] In this context, we introduce an approach that combines an advanced structural IrO2 catalyst with a porous transport electrode (PTE) configuration to form a catalyst layer with optimal IrO2 catalyst dispersion. The catalyst consists of IrO2 coated TiO2 core-shell microparticles (IrO2@TiO2) with 50 wt. % of IrO2 synthesized by a facile method. To form a catalyst layer, IrO2@TiO2 microparticles were directly deposited on a titanium porous transport layer (PTL). This allowed an optimal electrical contact of IrO2@TiO2 microparticles with the large pore sized titanium PTL current collector. As a 3

result, PEMWEs with an anodic catalyst loading of 0.4 mgIr cm-2, using the IrO2@TiO2 catalyst, achieved 1 A cm-2 at 1.67 V, outperforming both unsupported and supported commercial catalysts from Alfa Aesar and Umicore, respectively. Despite relatively low Ir loading, this performance is amongst the best reported in literature.[12,13]

2. Experimental 2.1 IrO2@TiO2 catalyst

-p

ro of

2.1.1 Synthesis For IrO2@TiO2 synthesis, 0.448 g TiO2 (rutile particles, size < 5 µm, Sigma Aldrich) and 0.957 g H2IrCl6.4H2O were added in a glass vial containing a solvent mixture of 20 ml ethanol, 40 µl acetic acid, and 20 µl H2O. The closed vial containing reaction solution was then sonicated for 10 min to obtain a homogeneous dispersion. The mixture was kept stirring at 100 °C (oil bath temperature) while the vial was opened for solvent evaporation until the content was dried. The dried powder was placed into a tubular furnace and heated up to 500 °C at a rate of 5 °C min−1 and held at 500 °C for 30 min in air, resulting in ca. 0.900 g IrO2@TiO2 powder. The powder was ground for 30 min using an agate mortar to obtain fine IrO2@TiO2 powder. This method delivered a 100 % yield from the metal precursors. The final IrO2@TiO2 consisted of 50 wt. % IrO2 and 50 wt. % TiO2, calculated based on the weight of the final product and the TiO2 precursor.

𝐾

lP

re

2.1.2 Physical characterization Samples for scanning electron microscopy (SEM) and energy-dispersive X-ray (EDX) measurements were prepared by drop-casting the dispersions of TiO2 and IrO2@TiO2 powder in isopropanol (IPA) onto silicon substrates, respectively. The SEM imaging and EDX characterizations were acquired using a Tescan VEGA SBU with EDAX EDX-System. Transmission electron microscopy (TEM) samples were prepared by dip-coating TEM grids with respective dispersions of TiO2 and IrO2@TiO2 powder in IPA, followed by drying in air. TEM images were obtained using a Talos L120C (FEI). Powder X-ray diffraction measurements (XRD) were acquired using a Bruker D8 Advance with Cu K1 X-ray source (wavelength: 0.154056 nm). Crystallite sizes are calculated using the Scherrer equation:

na

𝐵(2) = 𝐿∗𝑐𝑜𝑠 , where B is the crystallite size; K is the shape factor, being 0.9;  is the X-ray

ur

wavelength in nm ( = 0.154056 nm for Cu K1); L is the full width at half maximum (FWHM) of the peak in radian; and  is the Bragg angle of the peak.

Jo

Zeta potential was measured and particle size measurements were performed by dynamic light scattering (DLS), on three samples. I) Pristine TiO2: 0.0045 g TiO2 was added to a solution of 2 µl H2O and 2 ml ethanol, pH = 7.0. II) Surface charge modified TiO2: 0.0045 g TiO2 was added to a solution of 4 µl acetic acid, 2 µl H2O, and 2 ml ethanol, pH = 1.5. III) IrCl6@TiO2 core-shell sample: 0.0096 g mH2IrCl6.4H2O and 0.0045 g TiO2 were added to a solution of 4 µl acetic acid, 2 µl H2O, and 2 ml ethanol, pH = 1.4. The three received dispersions were ultra-sonicated for 10 min, and kept stirring overnight (12 h) on a magnetic stirring bar. Zeta potential and size measurements were performed using a Zetasizer Nano ZSP (Malvern Panalytical). Measurement parameters were chosen as follows: refractive index R = 2.58 and absorption k = 0.001 for TiO2 rutile at laser wavelength  = 633 nm of the Zetasizer device; dispersant refractive index RI = 1.361, viscosity cP = 1.08, and dispersant dielectric constant  = 24.3 for ethanol solvent.

4

-p

ro of

2.1.3 Electrochemical characterization Electrochemical stability: The SFC-ICP-MS setup is the same as described in ref.[27] The SFC cell was connected to the counter electrode (graphite rod) and the reference electrode (Metrohm Ag/AgCl) compartments and an ICP-MS (Perkin-Elmer NexION 300) with Tygon tubing. Samples were prepared by suspending the catalysts in a H2O solution with 12.5 % IPA. Nafion® perfluorinated resin solution (5 wt. % Sigma Aldrich) was added so that Nafion made up 20 wt. % of the suspensions solid content. The suspensions were sonicated for 10 min (4 s pulse, 2 s pause) and dropcasted with 0.2 µl per spot on a polished glassy carbon plate. The resulting Ir loading on the formed catalyst spots of Ø ~ 1.3 mm was 10 µgIr cm-2 for IrO2@TiO2 and Alfa Aesar as well as 50 µgIr cm-2 for the IrO2/TiO2 Umicore Elyst catalyst. The higher loading employed for the IrO2/TiO2 Umicore Elyst catalyst aimed to prevent glassy carbon plate corrosion during the galvanostatic hold. This corresponded to total Ir loadings of 0.133 µg and 0.663 µg respectively. The spots were rinsed with water and located by a vertical camera. The cell (Ø 2 mm opening) was placed on the spots to perform electrochemical measurements. The spots were positioned by a vertical camera. More details on the SFC-ICP-MS setup and measurements can be found elsewhere in more detail.[29] SFC measurements were performed in Ar-saturated 0.1 M HClO4 (Merck, Suprapur®) diluted with ultrapure water. ICP-MS measurements were performed using 10 µg L-1 187Re and 45Sc as internal standards for Ir and Ti respectively. The flow rate through the cell was 205 µl min-1. Currents were normalized to the total Ir loading of the spots. 2.2 Membrane electrode assembly preparation and characterization

Jo

ur

na

lP

re

2.2.1 Catalyst inks: preparation and stability test The goal of this work was to evaluate the potential of the novel in-house synthesized IrO2@TiO2 catalyst for PEMWE. To be able to compare the influence of microstructure and membrane electrode assembly (MEA) performance with other benchmark catalyst powders, the fabrication (catalyst ink mixing and spray parameters) and catalyst layer composition (Nafion content, noble metal loading) of the PTE catalyst layers were kept as close as possible. An individual optimization of all the process parameters for every different catalyst material for ultrasonic spray coating was beyond the scope of this work. The IrO2 catalyst layer was studied extensively in our previous work and was optimized regarding reproducibility and performance. [30] Three different OER catalyst powders, which were used to form the porous transport electrodes in the current work, were the synthesized IrO2@TiO2 particles and two commercially available catalyst materials. The commercial catalysts were 75 wt. % Ir supported on TiO2 (IrO2/TiO2,Umicore) and unsupported IrO2 (Premion, Alfa Aesar). The solvents for the IrO2/TiO2 and IrO2 ink were DI-water and IPA (1:1). For the IrO2@TiO2 ink, DI-water and methanol were used in a 1:3 ratio, since using IPA instead of methanol resulted in a less homogeneous catalyst ink mixture. All catalyst inks contained 1 wt. % solids, which consisted of 99 wt. % catalyst material and 1 wt. % Nafion (D520, FuelCellStore). The catalyst powder was weighted in a glass bottle prior to adding the solvents and finally Nafion. After adding a new component, the ink was stirred for a short time. After all components were added, the ink was continuously stirred overnight after an additional 30 min mixing step using an ultrasonic tip (Hielscher UIS250L, 90 % amplitude, continuous mode, 0.55 W). The bottle was placed in an ice bath during ultrasonication and stirred continuously. After being stirred overnight, the ultrasonication step was repeated prior to spray coating. Besides comparing the IrO2@TiO2 catalyst with the IrO2 and IrO2/TiO2 catalyst, a fourth catalyst ink with just mixing the same IrO2 and TiO2 particles (IrO2+TiO2) which were used to synthesize the 5

ro of

IrO2@TiO2 particles was prepared for spray coating. A stable dispersion of catalyst particles in a matrix of solvents and Nafion was required to prevent particle precipitation in the long tubing of the spray coater, when the ink was pumped from syringe reservoir towards the ultrasonic nozzle. Since the IrO2+TiO2 ink showed insufficient ink stability due to particle precipitation already during ink preparation, no PTE and subsequently MEA could be fabricated with that catalyst. To quantify the unstable character of the IrO2+TiO2 based ink, a turbiscan stability index (TSI) measurement was performed. The prepared IrO2@TiO2 and IrO2+TiO2 inks were transferred to glass vials (length = 45 mm) and placed in the measurement chambers of a TURBISCAN TOWER (Formulaction, France). The temperature in the chambers was set to 25 °C. After an equilibration time of 5 min, the transmission and backscattering signals of the luminescent diode (λAir = 880 nm) were collected at an angle of 180° and 45°, respectively. The signals were collected over the full length of the sample, acquiring transmission and backscattering data every 40 µm. Scans were taken over a course of 24 h. After the first measurement, the samples were scanned every 20 minutes during the first 2 h, every hour the following 10 hours and every 2 hours during the last 12 hours. The transmission and backscattering data was recorded by the device software, in the range from the bottom of the vial to the meniscus of the ink.

Jo

ur

na

lP

re

-p

2.2.3 Fabrication of porous transport electrodes To prepare the porous transport electrodes, the catalyst inks containing IrO2/TiO2, IrO2@TiO2 and unsupported commercial IrO2 were spray coated on top of 4 cm2 fiber sintered titanium substrates (Bekaert, 1 mm thick, 57 % porosity) using an Exacta Coat (Sono-Tek) ultrasonic spray coater. The titanium substrates were placed on a hot plate set to 120 °C. An ultrasonic nozzle of the type AccuMistTM (48 kHz) was used. A meander shaped pattern (pitch of 1.5 mm) was sprayed at a speed of 170 mm s-1, a flow rate of 0.45 ml min-1 and the ultrasonic power of the nozzle set to 5 W. The height of the nozzle was 37 mm and the shaping air was set to 0.6 kPa. The noble metal loading was controlled several times during spray coating via weighing a 1 cm2 reference metal sheet on a microscale (Sartorius ME 36S) which was spray-coated in parallel to the titanium substrates. The syringe of the spray coater was not fully filled with the ink to prevent particle sedimentation during spray coating. The way of refilling the syringe more often whilst stirring the bigger part of the ink on a separate magnetic stirrer, led to more homogeneous spray patterns due to a stable smaller portion of the ink in the syringe of the spray coater. An additional magnetic stirrer in the syringe helped to keep the ink homogeneous. Since still some precipitation of catalyst material was observed in the syringe of the spray-coater, the final Nafion vs. catalyst ratio in the catalyst layers was measured by a thermogravimetric analysis (TGA, Netsch STA 449F5) of the sprayed catalyst layers. It revealed an actual Nafion content of around 5 wt. %. For TGA, the Nafion part in the sprayed catalyst layer was burned according to Feng et al.[31] under air atmosphere to analyze the loss of weight. The heating rate was 5 K min-1 up to a temperature of 1000 °C, which was held for one hour. Due to the discrepancy between ‘as mixed’ and final catalyst layer, we suggest TGA analysis as a tool for controlling the results of the spraycoating procedure. 2.2.4 Structural characterization of porous transport electrodes The surface and cross-sections of the different porous transport electrodes were studied with a focused ion beam (FIB) SEM (Zeiss Crossbeam 540 with GEMINI II). By this method the homogeneity and porosity of the deposited catalyst layers were studied to explain influences on the polarization

6

behavior. The interfacial area between titanium fibers and catalyst layer was analyzed to explain possible influences on the electrical resistance.

ur

na

lP

re

-p

ro of

2.2.5 MEA testing For PEMWE cell tests, the fabricated porous transport electrodes were pressed against a Nafion N212 or N115 membrane (FuelCellStore) and a carbon cloth based gas diffusion electrode (0.5 mg cm-2 (60 %) Pt/C (SL-GDE, FuelCellsEtc). The geometric active area of all MEAs used in this work was 4 cm2. The test-cell was designed in-house and consisted basically of the MEA sandwiched between two titanium flow fields, two copper plates for the electrical connection towards the potentiostat and two aluminium end plates to fix the inner cell parts. The test cell was tightened with 8 screws at 8.5 Nm. Two heating elements inserted in the end plates were set to 80°C. The titanium flow fields were designed with a parallel finger structure. The porous transport electrodes at the anode side were placed in a 1 mm thick PTFE frame, the cathode electrode was placed in a 150 µm thick PTFE frame. The PTFE frames were used to set the compression level of the porous supports of the electrodes. The titanium material was not significantly compressed, but the thickness of the carbon cloth based cathode was reduced by almost 40 %. The test cell was connected to a peristaltic pump (Ismatec IP 65) which supported both the anode and cathode separately with DI-water at a flow rate of 40 ml min-1. The tubing at the inlets of the test cell were immersed in a water bath (Lauda Ecoline 003, 88 °C) to pre-heat the inflowing water and therefore prevent temperature gradients within the cell. Two in-line temperature sensors in the tubing at the in- and outlet of the anode side, as well as one temperature probe inserted in the flow field were assembled. After a stable temperature of 80 °C was observed at all temperature sensors, the conditioning of the MEA was started with a potentiostat type 857 from Scribner: the voltage was varied for 15 times from 1.4 V – 2.2 V in 200 mV steps (30 s per step). Afterwards, a polarization curve was recorded via controlling the current density. Every step was held for 120 s and started in a range of 0 A – 40 mA in 10 mA steps. When proceeding to higher current densities, the step size was increased to 50 mA in the range of 50 mA - 1 A and to a step size of 250 mA between 1.25 A and 3 A. From 3.5 A - 6 A the step size was 500 mA. Above 7 A, the step size was increased to 1 A. The frequency to measure the high frequency resistance (HFR) was 1 kHz. The corresponding voltage of the polarization curve at a set current point was an average value during the final 10 % of the holding time. The polarization curves were analyzed according to the contributions of the kinetic overpotential at low current densities, ohmic losses at moderate current densities towards mass transport losses when high current densities were applied.[32] To study differences in the reproducibility of different types of porous transport electrodes, the average cell voltage and standard deviation was calculated for the three samples tested per configuration.

Jo

3. Results

3.1 IrO2@TiO2 catalyst 3.1.1 Synthesis and characterization of IrO2@TiO2 catalyst The synthesis process was based on two steps: forming H2IrCl6 shell on TiO2 core and transforming H2IrCl6 shell into IrO2 layer via pyrolysis (Figure 1). To make the first step possible, the surface charge of TiO2 in ethanol needed to be converted from negative to positive for the attachment of [IrCl6]2anions via adjusting the solution pH to a value lower than point of zero charge, e.g. 6.[33] This was accomplished by adding acetic acid and H2IrCl6 acid, so that the Zeta potential of TiO2 particles changed to positive value of +28 mV from -4.2 mV of unmodified TiO2 (Figure 1a and b). Therefore, 7

lP

re

-p

ro of

[IrCl6]2- anions were electrostatically attracted towards the surface of TiO2 particles, and changed their surface charge again to negative value of -42.5 mV (Figure 1c). Zeta potential values were averaged over three measurements (see Figure S1 for details). With the highly charged surfaces, the dispersion of TiO2 particles with [IrCl6]2- shells (IrCl6@TiO2) improved significantly compared to the unmodified TiO2 particles. Due to agglomeration, the average particle size of pristine TiO2 in ethanol solution was 4.5 µm as measured by dynamic light scattering (DLS) (Figure S2). Conversely, the average size of IrCl6@TiO2 was 1.1 µm, which was consistent with SEM result for single particles, indicating that IrCl6@TiO2 particles were well dispersed in ethanol. We assume that first H2IrCl6 layers that were formed based on electrostatic interaction between [IrCl6]2- anions and positively charged surface of TiO2 served as seeds for a continuous growth of H2IrCl6 shells around the TiO2 particles upon solvent evaporation (Figure 1d). The conversion of H2IrCl6 shell to an IrO2 layer was performed based on a chemical reaction, H2IrCl6 + O2  IrO2 + 2HCl + 2Cl2, that occurred via pyrolysis at 500 °C for 30 min in air (Figure 1e). EDX analysis of the IrO2@TiO2 powder revealed the atomic Ir/Ti ratio to be  19.4/80.6 corresponding to a weight ratio Ir/Ti of  49.2/50.8, which is close to the target ratio (Figure S3). This confirmed a  100 % yield of the synthesis method. The EDX result also showed a negligible amount of residual Cl in the final product.

na

Figure 1 Schematic illustration of synthesis process for IrO2@TiO2 catalyst. Surface charge manifesting in Zeta potential of TiO2 particle in ethanol solution containing H2O changes upon pH changing and attachment of [IrCl6]2- anions.

Jo

ur

TEM investigations revealed a full coverage of IrO2 layers around TiO2 particles as shown in Figure 2 c, d and e, when compared with the blank TiO2 particles in Figure 2a and b. The IrO2 layer exhibited typical rutile crystalline structures. The d-space of its crystallographic planes was determined to be 0.332 nm via Matlab analysis of a high resolution TEM image (Figure 2f).[34] Powder XRD pattern of the IrO2@TiO2 is shown in Figure 3a, revealing sharp typical peaks of rutile IrO2 [35], confirming the high crystallinity of the IrO2 shells. The result also revealed the majority of rutile TiO2 besides anatase TiO2 for the core. The mean crystal size of the IrO2 shell was calculated being  10 nm, using the Scherrer equation with the diffraction peak at 34.4 degree of the (101) planes of IrO2. This peak was used because of its highest intensity and isolation. In comparison to IrO2@TiO2 , IrO2/TiO2 (Umicore) exhibited XRD peaks with the same locations, but broader width due to its smaller crystal sizes. Further EDX elemental mapping for a single IrO2@TiO2 particle demonstrated a homogeneous distribution of Ir element across the entire surface of TiO2 particle as shown in Figure 4 b1, b2, b3, and b4. While the blank TiO2 particles tended to aggregate when dispersed in IPA solvent, after being 8

re

-p

ro of

covered with IrO2, the resulting IrO2@TiO2 particles were well dispersed in polar solvents such as IPA. As a result, well-distributed IrO2@TiO2 particles were deposited on a silicon substrate (Figure 4 a3 and a4 and Figure S4 and S5). In contrast, TiO2 particles were agglomerated when deposited on a silicon substrate using the same method (Figure 4 a1 and a2, and Figure S5). This facilitated the ink preparation for electrode fabrication using a spray-coating technique as later presented.

Jo

ur

na

lP

Figure 2 TEM images of a blank TiO2 particle at different scales (a) and (b); a IrO2@TiO2 particle (c), (d), and (e), and (f) an zoom-in TEM image of a IrO2 crystal from the marked spot in (e) with crystallographic plane distance analysis.

Figure 3 Powder X-ray diffraction patterns of IrO2@TiO2 (a) and IrO2/TiO2 Umicore (b)

9

ro of

Figure 4 SEM images of blank TiO2 particles (a1) and (a2), and of IrO2@TiO2 (a3) and (a4). EDX elemental mapping images for Ir (b2), Ti (b3), and O (b4), respectively for a single particle of IrO2@TiO2 in (b1) SEM image, showing homogenously covering of Ir over the TiO2 particle.

Jo

ur

na

lP

re

-p

3.1.2 Electrochemical activity and stability in half-cell tests Figure 5a and c show the electrochemical characterizations of IrO2@TiO2 in comparison to commercial unsupported IrO2 (Alfa Aesar) and IrO2/TiO2 (Umicore), using three electrode scanning flow cell (SFC) setup. IrO2@TiO2 exhibited an OER mass activity of 112 mA mgIr-1 at 1.55 V (RHE) in 0.1 M HClO4 solution, which is more than two times that of commercial IrO2/TiO2 (Umicore) (47 mA mgIr-1). However, the activity of IrO2@TiO2 is lower than pure IrO2 catalysts (Alfa Aesar) which exhibited 279 mA mgIr-1 at the same potential. As the activity of IrO2 catalysts is conversely correlated to their crystallinity, which decides on the stability, a compromise between the activity and stability should be considered for catalyst design.[27] As such we investigated the stability of IrO2@TiO2 in comparison with the two commercial catalysts by an in situ dissolution analysis using the scanning flow cell in combination with ICP-MS.[27] Figure 5b presents dissolution rates of Iridium in parallel with potential profile applied to each catalysts vs time. The potential profiles included a linear sweep of potential (1.1-1.65 V vs RHE at 5 mV s-1) and iridium-normalized constant current densities of 100 mA mg-1 for IrO2@TiO2 and IrO2 catalysts while the IrO2/TiO2 catalyst was measured at a lower current density of 20 mA mg-1. The lower current density for IrO2/TiO2 was necessary due to its higher crystallinity and therefore lower activity. The results showed a lower stability of IrO2@TiO2 as compared to IrO2 (Alfa Aesar). IrO2/TiO2 (Umicore) exhibited the highest stability. To evaluate catalyst utilization efficiency (activity and stability), Geiger et. al.[27] introduced a descriptor, S-number, which is the number of produced oxygen molecules (n(O2)) per number of dissolved iridium ions (n(Ir)) for OER catalysts. As shown in Figure 5d, IrO2@TiO2 yielded an S-number slightly lower than IrO2 (Alfa Aesar) and IrO2/TiO2 (Umicore). Compared with other benchmarking materials such as commercial hydrous iridium oxide (S-number of ca. 104) and RuO2 (S-number of ca. 103) nanoparticles,[27] IrO2@TiO2 provides a stability higher than RuO2 and similar to hydrous iridium oxide. Therefore, when considering both energy efficiency and catalyst utilization efficiency, IrO2@TiO2 might be the suitable choice for practical application. The intrinsic activity is in the order of IrO2 (Alfa Aesar) > IrO2@TiO2 > IrO2/TiO2 (Umicore), while the stability is in the order of IrO2/TiO2 (Umicore) > IrO2 (Alfa Aesar) > IrO2@TiO2. As a direct correlation between IrOx crystallinity and OER stability was established,[36] the stability of IrO2@TiO2 might be further tunable towards higher stability in future works. 10

Jo

ur

na

lP

re

-p

ro of

Previous developments mainly focused on using conductive supports for IrO2. Interestingly, in spite of low conductivity of TiO2, IrO2@TiO2 exhibited a superior mass activity to most other supports in half-cell characterization (Table S1). At 1.6 V vs RHE, IrO2@TiO2 showed a mass activity of 364 mA mg1 Ir, whereas, mesoporous Ir supported TiOx (sub-oxide) [37], thermal decomposed Ir/TiOx, [38] and nano IrO2/Nb-doped TiO2 [39] displayed that of 158.3, 6.2, and 157 mA mg-1 Ir, respectively. Our catalyst also exhibited a higher mass activity than IrO2 supported on TaC, TiC, NbC, NbO2, WC, ATO as reported by Karimi and Peppley [14], but slightly lower than IrO2@Ir supported TiN [19], and mixed oxide of Ir-Ti (Table S1). [22] We note that the design of IrO2@TiO2 core-shell microparticles conceptually differs from conventional core-shell catalyst particles, which often comprise of a single nanosized core coated with a continuous nano-thin crystalline shell. The conventional core-shell configurations are well established in Pt-based catalysts for fuel cells [40,41], but have been rarely reported for Ir-based OER catalysts [24–26] due to the synthesis challenges as mentioned before. In our design, the catalyst particles are multi-scale structures: IrO2 shells comprising of IrO2 nanoparticles; and microscale TiO2 cores with diameters from a few hundred nm to 1 µm (Figure 2 and 4). This particular multiscale catalyst is suitable for direct deposition on PTL current collectors, which expose large pores in the µm-range at the interface towards the catalyst layers. This might be the reason for a more efficient catalyst utilization when IrO2@TiO2 was implemented into the catalyst layer directly on Ti PTL as later described in more detail, which is the main objective of this report.

Figure 5 Electrochemical performances of IrO2@TiO2 in comparison with IrO2 reference catalysts from Alfa Aesar and Umicore. (a) Linear sweep voltammograms (LSVs); (b) Stability investigation based on detected iridium dissolution during OER: potential profiles (upper image) and detected Ir concentration in the electrolyte during ramping potential to 1.65 V vs. RHE and constant current at 100 mA mg-1 for IrO2@TiO2 and IrOx (Alfa Aesar) and 20 mA mg-1 for IrO2/TiO2 (Umicore). (c) CVs recorded before AST for all catalysts. (d) S-numbers of the studied catalysts. Catalyst loadings: 10 µg cm-2 for IrO2@TiO2 and IrOx ( Alfa Aesar) and 50 µg cm-2 for IrO2/TiO2 (Umicore), LSVs without iR drop correction. 11

It is noteworthy to mention that electro catalyst stability under real PEMWE conditions is expected to be higher than that obtained in SFC experiments,[27] so further SFC-ICP-MS studies regarding approaches to improve IrO2@TiO2 stability, as well as representative data from PEMWE electrocatalyst dissolution are still pending. 3.2 MEA preparation and characterization

Jo

ur

na

lP

re

-p

ro of

3.2.1 Catalyst inks: preparation and stability test As the stability of the catalyst ink is a crucial parameter for the application of spray coating and other MEA manufacturing techniques, we examined the stability of our ink formula for IrO2@TiO2 compared to an ink with mixed TiO2+IrO2 powders. Both inks contained 1 wt. % solids mixed with methanol and water (3:1 vol. ratio). The solid content was the sum of 1 wt. % Nafion and 99 wt. % metal powder. The stability was studied by measuring the effect of ink destabilization on backscattering signal intensity. In general, a drop of backscattering intensity indicates instability of the tested ink. The results in Figure 6a show that the backscattering intensity of IrO2@TiO2 ink dropped by only < 2 % during 2.5 h aging, which is ten-fold less than that of the ink with the same ink formula but using mixed IrO2+TiO2 particles (> 10 %). As such, the results demonstrated that the IrO2@TiO2 ink was stable, while (IrO2+TiO2) ink was unstable. This result was also corroborated in digital images (Figure 6b), in which the white colour of TiO2 was observed at the bottom of the test vial of the (IrO2+TiO2) ink, while it was absent for the IrO2@TiO2 ink. The stability of IrO2@TiO2 ink can be explained by surface covering of IrO2 over TiO2 particle, which changes the surface properties of the received IrO2@TiO2 core-shell particles. At the same conditions, the bare TiO2 particles were demonstrated to be unstable in suspension. Therefore, it was not practically feasible to use the catalyst ink with physically-mixed IrO2 and TiO2 to fabricate spray coated anodes. To our knowledge, the stability of the TiO2+IrO2 mix based inks was not investigated before but is an important factor for future catalyst development: an unstable ink with its corresponding catalyst will not make it into the device.

Figure 6 Ink stability study for IrO2@TiO2 in comparison with an ink made by physically mixed IrO2 + TiO2 in the same conditions. (a) Evolution of backscattering spectra of the IrO2@TiO2 ink and the reference IrO2 + TiO2 ink upon aging time. (b) Digital images of two compared inks, which were taken prior to stability test. IrO2@TiO2 ink remained homogenous, while the IrO2+TiO2 ink segregated into two layers due to the precipitation of TiO2 particles onto bottom of the vial. 12

Due to the insufficient stability for spray coating with the device used in this work, no MEAs were fabricated using the mixed IrO2+TiO2 catalyst ink. In former works from Rozain et al.[12] it was claimed that pure Ti particles were mechanically mixed with an IrO2 catalyst powder. As the Ti powder however is exposed to air, there will be at least a small oxide layer forming on the Ti particles and the oxide layer will dominate the surface chemistry and physics of the particles when mixed into the solution of a catalyst ink. In our investigation the catalyst ink with the mixed IrO2+TiO2 catalyst ink like shown in Figure 6 undergoes sedimentation. Consequently, when aiming towards preparing large area electrodes with sizes of up to 1 m², homogeneous deposition of the electrode and thus reproducibility cannot be ensured with IrO2+TiO2 type catalyst systems. Therefore, the following structural characterization and MEA testing was performed on PTEs prepared with the IrO2@TiO2 catalyst, the unsupported IrO2 and the supported IrO2/TiO2 catalyst.

ro of

3.2.1 Structural characterization of porous transport electrodes A structural analysis of the cross-section of the different types of porous transport electrodes is shown in Figure 7. The catalyst layers contained 1.2 mgIr cm-2 for the IrO2@TiO2 and unsupported IrO2 catalysts, and 1.4 mgIr cm-2 for IrO2/TiO2(Umicore) and 5 wt. % Nafion. By the decision to take the same Nafion content and spray parameters for all different catalyst layers, the performance of the different MEAs, with respect to e.g. changes of the microstructure, was directly comparable.

na

lP

re

-p

The reference PTE made of commercial unsupported IrO2 catalyst (Premion, Alfa Aesar) showed a good porosity and a homogeneous catalyst layer (Figure 7a). When using the micrometer scale IrO2@TiO2 catalyst particles, the structure of the catalyst layer was clearly altered (Figure 7b). The relatively large TiO2 cores in the range of 500 nm led to a different distribution of IrO2 catalyst material within the catalyst layer compared to the unsupported IrO2 catalyst. An EDX analysis of the IrO2@TiO2 catalyst layer (Figure S6 and S7) showed the consistency of the catalyst composition with the material characterization (Figure 4). The supported IrO2/TiO2 catalyst (Umicore) (Figure 7c) showed particle sizes in the size range of the unsupported IrO2 catalyst which was approximately 100 nm. Therefore, the TiO2 particles were much smaller compared to the TiO2 cores used to prepare the IrO2@TiO2 catalyst. (b) IrO2@TiO2

ur

(a) IrO2

TiO2-core Ti-fiber

500 nm

Ti-fiber

500 nm

Jo

500 nm

(c) IrO2/TiO2

Figure 7 FIB/SEM cross-sections of the anodic catalyst layers deposited on titanium based porous transport layers (PTLs) containing 5 wt. % Nafion and three different IrO2 based catalysts. (a) Unsupported IrO2 Premion powder from Alfa Aesar; (b) IrO2@TiO2 catalyst. (c) IrO2/TiO2 supported catalyst from Umicore. The Pt-pad was deposited for assisting FIB/SEM analysis. According to the fabrication procedure, more layers of the supported catalysts have to be sprayed to reach the same overall loading as when the unsupported IrO2 powder is used. Therefore, the average thickness of the catalyst layers increased with an increasing wt. % of TiO2 in the supported catalyst powder. The volume of the catalyst layer is additionally increasing with an increasing particle size of 13

the TiO2 support. The IrO2@TiO2 catalyst layer has the in comparison biggest particle size and highest wt. % of the TiO2 support and is therefore thicker, than the IrO2 and IrO2/TiO2 based catalyst layer. The increased thickness of the IrO2@TiO2 catalyst layer was supposed to have an influence on mass transport properties. Also, the TiO2 was supposed to improve the mechanical stability of the catalyst layer. Conventional catalyst layers with loading < 0.5 mg cm-2 are often too thin, leading to mechanical instability when assembling in the PEMWE cell and an inefficient electrical contact with current collector as pointed out by Bernt et al.[13] The average thickness of the catalyst layers fabricated in this work was however not trivial to analyze due to the rough surface of the PTEs (Figure S8). A difference of several µm in local catalyst layer thickness made a direct comparison between the three different PTEs via the method of cross-sectional FIB-SEM analysis inaccurate and time consuming.

1.9 1.8 1.7 1.2 mgIr cm-2_IrO2

1.6 1.5

1.7

1.6

1.5

lP

1.4 mgIr cm-2_IrO2/TiO2

(b)

re

40 ml min, 80 °C N212 membrane

2.0

HFR-free cell voltage / V

2.1 Cell voltage / V

1.8

(a)

1.2 mgIr cm-2_IrO2@TiO2

1.4

1.4

0

1000

2000

3000

4000

5000

0

(c)

ur

0.10

Jo

0.05

0.00

0

1000

2000

1.60 HFR-free cell voltage / V

0.15

na

Current density / mA cm-2

HFR / W cm2

-p

2.2

ro of

3.2.2 MEA testing The MEAs used for electrochemical testing were made of a Nafion 212 membrane sandwiched between a PTE anode and Pt/C based cathode. To assess the reproducibility, three samples were prepared and tested for every type of PTE. Each curve shown in Figure 8 is plotted using the average data of the three samples for each PTE type cell with the standard deviation included.

1000

2000

3000

4000

5000

Current density / mA cm-2

(d)

1.55 1.50 1.45 1.40 1.35 1.30

3000

4000

Current density / mA cm-2

5000

1

10

100

Current density / mA cm-2

Figure 8 Full-cell characterizations of the IrO2@TiO2 catalyst in comparison with commercial reference catalysts. (a) Polarization curves of PEMWE cells using IrO2@TiO2 catalysts, unsupported IrO2 (Alfa Aesar), and IrO2/TiO2 (Umicore). Catalyst loadings are 1.2 mgIr cm-2 for IrO2@TiO2 and unsupported IrO2 (Alfa Aesar) catalysts, and 1.4 mgIr cm-2 for IrO2/TiO2; Nafion 212 membranes with thicknesses of 50.8 µm were used for all cells. (b) HFR-free cell voltage. (c) HFRs of the different PTE. (d) Kinetic region of the HFR-free cell voltages. 14

ur

na

lP

re

-p

ro of

When analyzing the polarization curves (Figure 8a), all PTEs showed a very good reproducibility, as the standard deviation of all polarization curves was in the range of 0.2 %. The IrO2@TiO2 catalyst showed a superior performance compared to the unsupported IrO2 and supported IrO2/TiO2 catalyst, which even had a 0.2 mgIr cm-2 higher Ir-loading. The HFR-free cell voltage (Figure 8b) was similar for the IrO2 and IrO2@TiO2-based PTEs of ca. 1.6 V at 2 A cm-2, but a ca. 40 mV higher overpotential for the supported IrO2/TiO2 catalyst was measured at the same current density. Due to no significant differences in the slopes of the HFR-free curves, differences in mass transport were rather small despite different catalyst layer thicknesses when using supported catalysts compared to unsupported IrO2 powder. The HFRs at 2 A cm-2 of supported catalysts were almost 37 mΩ cm2 lower than that of the unsupported IrO2 catalyst (Figure 8c). This lower HFR is the main origin of the improved cell performances, as EIS measurements (Figure S9) showed a similar charge transfer resistance of all PTEs. We attribute the lower HFR of both supported catalysts to the possibly better electrical conductivity resulted from their higher crystallinity, as compared to unsupported IrO2. A significant larger capacitance of unsupported IrO2 compared to supported catalysts, shown in the CV curves (Figure 5), is a behavior of amorphous surface of unsupported IrO2 catalysts. [42] This explanation is in line with a recent report [43], which revealed an additional interfacial resistance as a result of interface between amorphous hydrous surface IrO2 and Ti-PTL. A magnification of the kinetic region of the HFR-free cell voltage (Figure 8d) revealed the reason for the worse HFR-free cell voltage of the IrO2/TiO2 catalyst despite a similar mass transport overpotential. The IrO2/TiO2 catalyst showed a significantly worse activity on the cell level compared to the other catalyst systems. The activity of the IrO2@TiO2 catalyst was slightly better than that of the unsupported IrO2 within current density below 50 mA cm-2, but almost the same for current density > 50 mA cm-2. This is a surprising result, since half-cell measurements (Figure 5a) showed the unsupported IrO2 catalyst being the most active for the OER. Further investigation is therefore encouraged to fully elucidate this behavior. The use of bigger TiO2 particles in the case of the IrO2@TiO2 (Figure 7b) compared to the smaller particles in the IrO2/TiO2 catalyst (Figure 7c) seemed to increase the accessible catalyst surface area and therefore the iridium catalyst utilization. The combination of both a low HFR and high activity was the reason for the superior performance of the IrO2@TiO2 based PTE configuration compared to the pure IrO2 and IrO2/TiO2 based MEAs. These results demonstrate that the intrinsic catalyst activity showed in half-cell measurement is not always consistent with its performance at full-cell level due to the catalyst morphology incompatible with device configurations. In this case we suggest the evaluation of novel catalyst materials for a first screening in half-cell tests but then in a full-cell tester to state the final usability for PEMWE application.

Jo

Due to the need of reducing the overall noble metal content used in PEMWE applications, the goal of electrode fabrication in this section was to test the applicability of the novel IrO2@TiO2 catalyst for low loadings. The aims were first, to fabricate a homogeneous catalyst layer and second to determine the reproducibility and sufficient performance of the MEAs when tested in a full-cell. Due to the worse overall performance of the unsupported IrO2 catalyst already at higher loadings, only the two supported catalyst materials were further investigated. The Ir loading was 0.4 mgIr cm-2 for the IrO2@TiO2 and 0.5 mgIr cm-2 for IrO2/TiO2 (Umicore). The Nafion content in the catalyst layers was again constant at 5 wt. %. The configurations with lower loadings were tested two times (Figure 9). The IrO2@TiO2 catalyst did also perform significantly better than the IrO2/TiO2 catalyst at low Irloadings (Figure 9a). When comparing the polarization behavior of the PTEs in this work, the higher 15

performance of the low loading 0.4 mgIr cm-2 IrO2@TiO2 PTE was obvious. The same parameters for spray coating were used for all three catalysts. At a current density of 1 A cm-2, a cell voltage of 1.67 V for the 0.4 mgIr cm-2 IrO2@TiO2 PTE, and 1.72 V for 0.5 mgIr cm-2 IrO2/TiO2 PTE, were measured, respectively. The unsupported IrO2 and IrO2/TiO2 based PTEs required a loading of 1.2 mgIr cm-2 and 1.4 mgIr cm-2, respectively, for achieving similar performances as the low loading IrO2@TiO2 PTE

1.9 1.8 1.7 1.6

0.5 mgIr cm-2_IrO2/TiO2

1.5

0.4 mgIr cm-2_IrO2@TiO2

1.4

1.7

1.6

1.5

1.4

1000

2000

3000

4000

5000

-2

0.15

1.60

(c)

na

0.10

0.00 0

ur

0.05

1000

2000

3000

2000

3000

4000

5000

-2

(d)

1.55 1.50 1.45 1.40 1.35 1.30

4000 -2

Current density / mA cm

Jo

1000

Current density / mA cm

lP

Current density / mA cm

0

HFR-free cell voltage / V

0

HFR / W cm2

(b)

re

40 ml min, 80 °C N212 membrane

2.0

HFR-free cell voltage / V

2.1 Cell voltage / V

1.8

(a)

-p

2.2

ro of

(Figure S10). The HFR-free cell voltage (Figure 9b) showed no significant differences in the slopes of the curves indicating as for the higher loadings (Figure 8b), a similar mass transport behavior. The IrO2@TiO2-based PTE showed an almost stable HFR of ca. 75 mΩ cm2 (Figure 9c). The HFR of the IrO2/TiO2-based PTE however decreased when approaching to higher current densities. Starting with an initially higher HFR than for the IrO2@TiO2-based PTE, the IrO2/TiO2 samples showed lower HFRs above 3 A cm-2. This trend of the decreasing HFR with current density was already present at higher loadings (Figure 8c) and might be the reason for the smaller difference in cell voltage when

5000

1

10

100 -2

Current density / mA cm

Figure 9 PEMWE performance at reduced catalyst loadings. (a) Polarization curves of PEMWE using IrO2@TiO2 catalysts in comparison to the IrO2/TiO2. Catalyst loadings were 0.4 mgIr cm-2 for IrO2@TiO2 and 0.5 mgIr cm-2 for IrO2/TiO2. A Nafion 212 membrane with a thickness of 50.8 µm was used. (b) HFR-free cell voltage. (c) HFR of the different PTEs (d) Kinetic region of the HFR-free cell voltage. approaching higher current densities. Comparing the HFR of the low loading PTEs with the higher loadings, the HFRs are in the same range. Therefore, the positive influence of improving the in-plane conductivity within the catalyst layer via adding TiO2 support was more pronounced at lower loadings. A zoom in the HFR-free cell voltage (Figure 9d) shows again the superior OER kinetics on the 16

Jo

ur

na

lP

re

-p

ro of

IrO2@TiO2 as already analyzed for the PTEs with higher Ir loadings (Figure 8d). The reproducibility of the IrO2/TiO2-based PTEs was slightly worse compared to PTEs prepared with the IrO2@TiO2 catalyst and could be related to an increased homogeneity due to an increased thickness as already discussed in the structural analysis section. As practical PEMWE applications often use thicker membranes to mitigate the gas crossover, we tested the performance of IrO2@TiO2 PTE with a thicker membrane (N115, 127 µm thickness). The cells’ performance reflexed the trend of the above N212 (50 µm thickness) based cells, manifesting in higher performance of IrO2@TiO2 PTEs compared to unsupported IrO2 PTEs (Figure S11). To evaluate the potential of IrO2 @TiO2 catalysts for practical applications, we performed a durability test in a single cell (0.5 mgIr cm-2, N212 membrane) based on monitoring the change of cell voltage upon a constant current hold at 2 A cm-2 for 150 h. We observed that the cell performance degraded during the first 135 h, but seemed to be stable after 135 h (see Figure S12 and S13). It is not possible to decipher the origin of cell degradation during the first 135 h, since the first 100-h cell operation is often considered the conditioning time. [8] Due to the course of this work, we could not continue the test to a proper duration e.g. > 1000 h current hold. It will be the subject of our future dedicated research on the stability of IrO2@TiO2 catalysts. Overall, IrO2@TiO2 catalyst demonstrated notably superior performance compared to unsupported IrO2 and supported IrO2/TiO2 commercial catalysts at full-cell level in our experimental systems. To compare the performance of IrO2@TiO2 PTE with state of the art, we summarized the cell performances and test conditions of published PEMWE using supported Ir-based catalysts as Table S2. From these data, it is not possible to directly compare the intrinsic activity of IrO2@TiO2 with other supported catalysts since the testing conditions such as membrane types and catalyst loading were different. However, in term of overall cell performance, the cells using IrO2@TiO2 PTE (with membrane thickness of either 50 µm or 127 µm) outperformed most other systems using supported Ir-based catalysts such as IrO2 supported Nb-doped TiO2 [39], IrO2 supported V-doped TiO2 [18], IrO2 supported Ti0.9Nb0.1O2-x [44], IrO2/Ti1-xTaxO2 [45], IrO2 supported TiO2 (P25) [23], and IrO2 supported TaC [17], and only underperformed IrOx supported Sb-doped SnO2 and Ir-Sn mixed oxide. [46] However, previous work revealed the limited stability of Sn-based catalyst supports against dissolution in acidic conditions. [16] Compared with the cells using Ir0.7Ru0.3Ox, which is considered the most active OER catalyst, reported by Siracusano et al. [8], with 0.34 mg (Ir + Ru) cm-2 and a 90µm-thick membrane, our IrO2@TiO2 based cells with 0.4 mg(Ir) cm-2 and a 50-µm-thick membrane, exhibited a 20-mV higher potential at 1 A cm-2. But compared to the cells using Ir0.7Ru0.3Ox, reported by Wang et al. [7], with 1.0 mg (oxide) cm-2 and a 50-µm-thick membrane, our cells with the same membrane but a lower loading (0.4 mgIr cm-2) exhibited a 30-mV lower potential at 1 A cm-2. We further used an Ir-specific power density metric, which focuses on catalyst utilization, excluding the influence of other testing factors.[13] The Ir-specific power density is calculated via dividing the noble metal loading by the cell power at a fixed cell voltage of 1.79 V, which corresponds to a target cell voltage efficiency of 70 % (based on lower heating value of hydrogen). For the PTEs with 0.4 mgIr cm-2 IrO2@TiO2 catalyst, the cell voltage of 1.79 V corresponded to a current density of ca. 2.1 A cm-2 and therefore a Ir-specific power density of ca. 0.106 gIr kW-1. This value was close to the minimum Ir-specific power density of ca. 0.08 gIr kW-1 of the CCM MEA with the lowest loading (0.2 mgIr cm-2) prepared by Bernt et al.[13] Rozain et al.[12] achieved the value of ca. 0.05 gIr kW-1 for CCM MEA using 0.12 mg cm-2 IrO2-loading of IrO2/Ti supported catalyst. However, the use of Ti metal particle support bears the stability and cost concern as mentioned previously. It should be noted, that in this work the same spray and ink parameters optimized for the unsupported IrO2 were the same for all three catalysts to have a comparable morphology. Neither the IrO2/TiO2 catalyst, nor the IrO2@TiO2 catalyst therefore were optimized in terms of the catalyst layers. Although, the Ir-specific 17

power density of IrO2@TiO2 PTE cell is still slightly inferior to the IrO2/TiO2 CCM cell, the IrO2@TiO2 PTE configuration has an advantage of a straightforward manufacturing process resulting in mechanically stable and homogeneous catalyst layers, and reproducible PTEs at low catalyst loadings. In PTEs, all catalyst particles being connected with each other and with the titanium PTL could lead to an interface connection stability, mitigating the passivation problem of PTL current collectors.[47] We note that our excellent results of 50-µm-thin membrane PEMWEs might not be directly applicable to practical applications, as high crossover of hydrogen might occur under pressurized operations. [48] Elsewhere, strategies to mitigate hydrogen crossover have been developed, such as integration of Pt-based catalysts into the anode [49] or membrane [50] for recombining hydrogen and oxygen into water. Since these strategies proved to effectively mitigate the hydrogen crossover even under pressurized conditions, their adoption might enable thin membrane PEMWEs e.g. in our work possible for practical applications.

ro of

4. Summary and Conclusion

lP

re

-p

A high catalyst dispersed electrode has been accomplished, allowing a high PEMWE performance at low Ir loading. This is achieved by the combination of a unique microstructural catalyst, IrO2 coated TiO2 core-shell microparticles, with new electrode configuration, catalyst coated porous transport layer (PTE). The core-shell catalyst (IrO2@TiO2) has been synthesized by a facile and scalable method. The method works on the basis of electrostatic interaction of [IrCl6]2- precursors to positively-charged surface of TiO2 microparticles to form H2IrCl6-coated TiO2 particles intermediate product, followed by pyrolysis to form IrO2@TiO2. TEM and SEM investigations revealed the core-shell like structures of IrO2@TiO2 microparticles with high coverage of IrO2 on the surface of TiO2 particles. High resolution TEM revealed high crystallinity of rutile IrO2 shell. Electrochemical testing by scanning flow cell (SFC) revealed that IrO2@TiO2 exhibits a two-fold higher mass activity than commercial IrO2/TiO2 (Umicore) towards oxygen evolution reaction (OER) in acidic solution.

Jo

ur

na

When implemented in PTE-configurations for PEMWEs, IrO2@TiO2 anode catalyst layers demonstrated a notably superior performance compared to two commercial benchmark catalysts: unsupported IrO2 (Alfar Aeasar) and IrO2/TiO2 (Umicore). The different catalysts layers were prepared with the same fabrication technique and compared at the same electrochemical testing parameters. IrO2@TiO2 PTEs with a loading as low as 0.4 mgIr cm-2 yielded current densities of 1 A cm-2 at 1.67 V, still outperforming the benchmark IrO2/TiO2 PTE with a 3-fold higher loading (1.4 mgIr cm-2). This high performance was mainly attributed to the combination of the best kinetics and low HFR. Conversely, the unsupported IrO2 catalyst PTE showed good kinetics but a higher HFR, and the IrO2/TiO2 PTE showed a low HFR but worst kinetics. The better catalyst utilization in the IrO2@TiO2 configuration was related to its catalyst morphology, as the high dispersion of IrO2 within porous structured catalyst layers was confirmed by FIB/SEM investigation. The addition of TiO2 particles reduced the ohmic resistance, which was related to an increased in-plane conductivity and electronic contact area between catalyst layer and titanium porous transport layer. It was found that the intrinsic catalyst activity showing in half-cell measurements is not always consistent with its performance in full-cell level due to the catalyst morphology incompatible with device configurations. Overall, in comparison to state-of-the-art PEMWE, the IrO2@TiO2 PTE exhibited amongst the best performances reported in the literature. The potential of our novel IrO2@TiO2 catalyst for industrial PEMWE applications was motivated in this work, given the facile synthesis route and high MEA performance at low Ir loadings. Further optimization of the IrO2@TiO2 catalyst layer in terms of 18

Nafion content and fabrication parameters could result in even further reduced Ir loading while maintaining high cell performance. For a better comparison to state-of-the-art PEMWE performance, catalyst coated membranes will be investigated. Still, the stability of IrO2@TiO2 in OER condition was not optimal, and still inferior to that of commercial IrO2/TiO2 (Umicore). We suggest that the stability could be improved by the optimization of thermal treatment process in the future.[27,51]

Credit author statement Dr. Chuyen Van Pham Conceptualization, Methodology, Validation, Formal analysis, Investigation, Writing - Original Draft, Writing - Review & Editing, Visualization.

ro of

Melanie Bühler Conceptualization, Methodology, Validation, Formal analysis, Investigation, Writing - Original Draft, Writing - Review & Editing, Visualization.

Julius Knöppel

-p

Methodology, Formal analysis, Investigation, Writing - Review & Editing, Visualization.

re

Markus Bierling

lP

Investigation, Formal analysis, Writing - Review & Editing.

Dominik Seeberger

na

Investigation, Formal analysis, Writing - Review & Editing.

Dr. Daniel Escalera-López

ur

Formal analysis, Writing - Review & Editing, Visualization.

Prof. Dr. Karl J. J. Mayrhofer

Jo

Conceptualization, Formal analysis, Resources, Writing - Review & Editing.

Dr. Serhiy Cherevko Conceptualization, Formal analysis, Resources, Writing - Review & Editing.

Prof. Dr. Simon Thiele Conceptualization, Formal analysis, Resources, Writing - Review & Editing, Supervision, Project administration, Funding acquisition. Declaration of interests 19

☒ The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

ro of

The authors gratefully acknowledge the financial support by the Federal Ministry of Education and Research of Germany in the framework of Power-MEE (03SF0536E, 03SF0536F). C.V.P acknowledges financial support from Deutsche Forschungsgemeinschaft (DFG) under project number 389154849. D.E.-L. and S.C. acknowledge financial support from Deutsche Forschungsgemeinschaft (DFG) under project number CH 1763/3-1. J.K, K.M. and S.C. acknowledge funding by the German Federal Ministry of Education and Research (BMBF) within the Kopernikus Project P2X and a further project (Kz: 033RC1101A). We thank Riko Moroni for Matlab analysis of a high resolution TEM image.

References

Jo

ur

na

lP

re

-p

[1] M. Carmo, D.L. Fritz, J. Mergel, D. Stolten, A comprehensive review on PEM water electrolysis, International Journal of Hydrogen Energy 38 (2013) 4901–4934. [2] C. Spöri, J.T.H. Kwan, A. Bonakdarpour, D.P. Wilkinson, P. Strasser, The Stability Challenges of Oxygen Evolving Catalysts: Towards a Common Fundamental Understanding and Mitigation of Catalyst Degradation, Angewandte Chemie (International ed. in English) 56 (2017) 5994–6021. [3] M. Blasco-Ahicart, J. Soriano-López, J.J. Carbó, J.M. Poblet, J.R. Galan-Mascaros, Polyoxometalate electrocatalysts based on earth-abundant metals for efficient water oxidation in acidic media, Nature Chemistry 10 (2018) 24–30. [4] J.S. Mondschein, K. Kumar, C.F. Holder, K. Seth, H. Kim, R.E. Schaak, Intermetallic Ni2Ta Electrocatalyst for the Oxygen Evolution Reaction in Highly Acidic Electrolytes, Inorganic Chemistry 57 (2018) 6010–6015. [5] J. Soriano-López, S. Goberna-Ferrón, L. Vigara, J.J. Carbó, J.M. Poblet, J.R. Galán-Mascarós, Cobalt polyoxometalates as heterogeneous water oxidation catalysts, Inorganic Chemistry 52 (2013) 4753–4755. [6] S. Siracusano, N. van Dijk, E. Payne-Johnson, V. Baglio, A.S. Aricò, Nanosized IrOx and IrRuOx electrocatalysts for the O2 evolution reaction in PEM water electrolysers, Applied Catalysis B: Environmental 164 (2015) 488–495. [7] L. Wang, V.A. Saveleva, S. Zafeiratos, E.R. Savinova, P. Lettenmeier, P. Gazdzicki, A.S. Gago, K.A. Friedrich, Highly active anode electrocatalysts derived from electrochemical leaching of Ru from metallic Ir 0.7 Ru 0.3 for proton exchange membrane electrolyzers, Nano Energy 34 (2017) 385– 391. [8] S. Siracusano, N. Hodnik, P. Jovanovic, F. Ruiz-Zepeda, M. Šala, V. Baglio, A.S. Aricò, New insights into the stability of a high performance nanostructured catalyst for sustainable water electrolysis, Nano Energy 40 (2017) 618–632. [9] S. Siracusano, V. Baglio, S.A. Grigoriev, L. Merlo, V.N. Fateev, A.S. Aricò, The influence of iridium chemical oxidation state on the performance and durability of oxygen evolution catalysts in PEM electrolysis, Journal of Power Sources 366 (2017) 105–114. 20

Jo

ur

na

lP

re

-p

ro of

[10] H. Su, V. Linkov, B.J. Bladergroen, Membrane electrode assemblies with low noble metal loadings for hydrogen production from solid polymer electrolyte water electrolysis, International Journal of Hydrogen Energy 38 (2013) 9601–9608. [11] O. Kasian, S. Geiger, P. Stock, G. Polymeros, B. Breitbach, A. Savan, A. Ludwig, S. Cherevko, Mayrhofer, Karl J. J., On the Origin of the Improved Ruthenium Stability in RuO2–IrO2 Mixed Oxides, Journal of The Electrochemical Society 163 (2016) F3099-F3104. [12] C. Rozain, E. Mayousse, N. Guillet, P. Millet, Influence of iridium oxide loadings on the performance of PEM water electrolysis cells, Applied Catalysis B: Environmental 182 (2016) 123–131. [13] M. Bernt, A. Siebel, H.A. Gasteiger, Analysis of Voltage Losses in PEM Water Electrolyzers with Low Platinum Group Metal Loadings, J. Electrochem. Soc. 165 (2018) F305-F314. [14] F. Karimi, B.A. Peppley, Metal Carbide and Oxide Supports for Iridium-Based Oxygen Evolution Reaction Electrocatalysts for Polymer-Electrolyte-Membrane Water Electrolysis, Electrochimica Acta 246 (2017) 654–670. [15] G. Liu, J. Xu, Y. Wang, X. Wang, An oxygen evolution catalyst on an antimony doped tin oxide nanowire structured support for proton exchange membrane liquid water electrolysis, J. Mater. Chem. A 3 (2015) 20791–20800. [16] S. Geiger, O. Kasian, A.M. Mingers, Mayrhofer, Karl J. J., S. Cherevko, Stability limits of tin-based electrocatalyst supports, Scientific reports 7 (2017) 4595. [17] J. Polonský, P. Mazúr, M. Paidar, E. Christensen, K. Bouzek, Performance of a PEM water electrolyser using a TaC-supported iridium oxide electrocatalyst, Int J Hydrogen Energ 39 (2014) 3072–3078. [18] C. Hao, H. Lv, Q. Zhao, B. Li, C. Zhang, C. Mi, Y. Song, J. Ma, Investigation of V-doped TiO 2 as an anodic catalyst support for SPE water electrolysis, International Journal of Hydrogen Energy 42 (2017) 9384–9395. [19] G. Li, K. Li, L. Yang, J. Chang, R. Ma, Z. Wu, J. Ge, C. Liu, W. Xing, Boosted Performance of Ir Species by Employing TiN as Support towards Oxygen Evolution Reaction, ACS Appl. Mater. Interfaces (2018). [20] M. Bernt, H.A. Gasteiger, Influence of Ionomer Content in IrO 2 /TiO 2 Electrodes on PEM Water Electrolyzer Performance, J. Electrochem. Soc. 163 (2016) F3179-F3189. [21] H.-S. Oh, H.N. Nong, T. Reier, A. Bergmann, M. Gliech, J. Ferreira de Araújo, E. Willinger, R. Schlögl, D. Teschner, P. Strasser, Electrochemical Catalyst-Support Effects and Their Stabilizing Role for IrOx Nanoparticle Catalysts during the Oxygen Evolution Reaction, Journal of the American Chemical Society 138 (2016) 12552–12563. [22] E. Oakton, D. Lebedev, M. Povia, D.F. Abbott, E. Fabbri, A. Fedorov, M. Nachtegaal, C. Copéret, T.J. Schmidt, IrO 2 -TiO 2, ACS Catal. 7 (2017) 2346–2352. [23] P. Mazúr, J. Polonský, M. Paidar, K. Bouzek, Non-conductive TiO2 as the anode catalyst support for PEM water electrolysis, International Journal of Hydrogen Energy 37 (2012) 12081–12088. [24] H.N. Nong, L. Gan, E. Willinger, D. Teschner, P. Strasser, IrOx core-shell nanocatalysts for costand energy-efficient electrochemical water splitting, Chem. Sci. 5 (2014) 2955–2963. [25] L.C. Seitz, C.F. Dickens, K. Nishio, Y. Hikita, J. Montoya, A. Doyle, C. Kirk, A. Vojvodic, H.Y. Hwang, J.K. Norskov, T.F. Jaramillo, A highly active and stable IrOx/SrIrO3 catalyst for the oxygen evolution reaction, Science 353 (2016) 1011–1014. [26] B.M. Tackett, W. Sheng, S. Kattel, S. Yao, B. Yan, K.A. Kuttiyiel, Q. Wu, J.G. Chen, Reducing Iridium Loading in Oxygen Evolution Reaction Electrocatalysts Using Core–Shell Particles with Nitride Cores, ACS Catal. 8 (2018) 2615–2621. 21

Jo

ur

na

lP

re

-p

ro of

[27] S. Geiger, O. Kasian, M. Ledendecker, E. Pizzutilo, A.M. Mingers, W.T. Fu, O. Diaz-Morales, Z. Li, T. Oellers, L. Fruchter, A. Ludwig, Mayrhofer, Karl J. J., M.T.M. Koper, S. Cherevko, The stability number as a metric for electrocatalyst stability benchmarking, Nat Catal 1 (2018) 508–515. [28] M. Ledendecker, S. Geiger, K. Hengge, J. Lim, S. Cherevko, A.M. Mingers, D. Göhl, G.V. Fortunato, D. Jalalpoor, F. Schüth, C. Scheu, Mayrhofer, Karl J. J., Towards maximized utilization of iridium for the acidic oxygen evolution reaction, Nano Res. 163 (2019) F3099. [29] S.O. Klemm, A.A. Topalov, C.A. Laska, K.J. Mayrhofer, Coupling of a high throughput microelectrochemical cell with online multielemental trace analysis by ICP-MS, Electrochemistry Communications 13 (2011) 1533–1535. [30] M. Bühler, F. Hegge, P. Holzapfel, M. Bierling, M. Suermann, S. Vierrath, S. Thiele, Optimization of anodic porous transport electrodes for proton exchange membrane water electrolyzers, J. Mater. Chem. A 165 (2019) F305. [31] M. Feng, R. Qu, Z. Wei, L. Wang, P. Sun, Z. Wang, Characterization of the thermolysis products of Nafion membrane: A potential source of perfluorinated compounds in the environment, Sci. Rep. 5 (2015) 9859. [32] M. Suermann, T.J. Schmidt, F.N. Büchi, Cell Performance Determining Parameters in High Pressure Water Electrolysis, Electrochim. Acta 211 (2016) 989–997. [33] K.C. Akratopulu, C. Kordulis, A. Lycourghiotis, Effect of temperature on the point of zero charge and surface charge of TiO2, J. Chem. Soc., Faraday Trans. 86 (1990) 3437–3440. [34] E. Willinger, C. Massué, R. Schlögl, M.G. Willinger, Identifying Key Structural Features of IrOx Water Splitting Catalysts, Journal of the American Chemical Society 139 (2017) 12093–12101. [35] G.C. da Silva, N. Perini, E.A. Ticianelli, Effect of temperature on the activities and stabilities of hydrothermally prepared IrO x nanocatalyst layers for the oxygen evolution reaction, Applied Catalysis B: Environmental 218 (2017) 287–297. [36] G.C. da Silva, N. Perini, E.A. Ticianelli, Effect of temperature on the activities and stabilities of hydrothermally prepared IrOx nanocatalyst layers for the oxygen evolution reaction, Applied Catalysis B: Environmental 218 (2017) 287–297. [37] D. Bernsmeier, M. Bernicke, R. Schmack, R. Sachse, B. Paul, A. Bergmann, P. Strasser, E. Ortel, R. Kraehnert, Oxygen Evolution Catalysts Based on Ir-Ti Mixed Oxides with Templated Mesopore Structure: Impact of Ir on Activity and Conductivity, ChemSusChem 11 (2018) 2367–2374. [38] K. Endo, Y. Katayama, T. Miura, T. Kishi, Composition dependence of the oxygen-evolution reaction rate on IrxTi1−xO2 mixed-oxide electrodes, Journal of Applied Electrochemistry 32 (2002) 173–178. [39] C. Hao, H. Lv, C. Mi, Y. Song, J. Ma, Investigation of Mesoporous Niobium-Doped TiO 2 as an Oxygen Evolution Catalyst Support in an SPE Water Electrolyzer, ACS Sustainable Chem. Eng. 4 (2015) 746–756. [40] X.X. Wang, M.T. Swihart, G. Wu, Achievements, challenges and perspectives on cathode catalysts in proton exchange membrane fuel cells for transportation, Nat Catal 2 (2019) 578– 589. [41] D. Göhl, A. Garg, P. Paciok, Mayrhofer, Karl J. J., M. Heggen, Y. Shao-Horn, R.E. Dunin-Borkowski, Y. Román-Leshkov, M. Ledendecker, Engineering stable electrocatalysts by synergistic stabilization between carbide cores and Pt shells, Nature materials (2019). [42] F. Karimi, A. Bazylak, B.A. Peppley, Effect of Calcination Temperature on the Morphological and Electrochemical Characteristics of Supported Iridium Hydroxyoxide Electrocatalysts for the PEM Electrolyzer Anode, Journal of The Electrochemical Society 164 (2017) F464-F474. 22

Jo

ur

na

lP

re

-p

ro of

[43] A. Weiß, A. Siebel, M. Bernt, T.-H. Shen, V. Tileli, H.A. Gasteiger, Impact of Intermittent Operation on Lifetime and Performance of a PEM Water Electrolyzer, J. Electrochem. Soc. 166 (2019) F487-F497. [44] H. Lv, S. Wang, C. Hao, W. Zhou, J. Li, M. Xue, C. Zhang, Oxygen‐deficient Ti0.9Nb0.1O2‐x as an Efficient Anodic Catalyst Support for PEM Water Electrolyzer, ChemCatChem (2019). [45] H. Lv, G. Zhang, C. Hao, C. Mi, W. Zhou, D. Yang, B. Li, C. Zhang, Activity of IrO 2 supported on tantalum-doped TiO 2 electrocatalyst for solid polymer electrolyte water electrolyzer, RSC Adv. 7 (2017) 40427–40436. [46] G. Li, H. Yu, D. Yang, J. Chi, X. Wang, S. Sun, Z. Shao, B. Yi, Iridium-Tin oxide solid-solution nanocatalysts with enhanced activity and stability for oxygen evolution, Journal of Power Sources 325 (2016) 15–24. [47] S. Choe, B.-S. Lee, M.K. Cho, H.-J. Kim, D. Henkensmeier, S.J. Yoo, J.Y. Kim, S.Y. Lee, H.S. Park, J.H. Jang, Electrodeposited IrO 2 /Ti electrodes as durable and cost-effective anodes in hightemperature polymer-membrane-electrolyte water electrolyzers, Applied Catalysis B: Environmental 226 (2018) 289–294. [48] P. Trinke, B. Bensmann, R. Hanke-Rauschenbach, Current density effect on hydrogen permeation in PEM water electrolyzers, International Journal of Hydrogen Energy 42 (2017) 14355–14366. [49] N. Briguglio, S. Siracusano, G. Bonura, D. Sebastián, A.S. Aricò, Flammability reduction in a pressurised water electrolyser based on a thin polymer electrolyte membrane through a Pt-alloy catalytic approach, Applied Catalysis B: Environmental 246 (2019) 254–265. [50] C. Klose, P. Trinke, T. Böhm, B. Bensmann, S. Vierrath, R. Hanke-Rauschenbach, S. Thiele, Membrane Interlayer with Pt Recombination Particles for Reduction of the Anodic Hydrogen Content in PEM Water Electrolysis, Journal of The Electrochemical Society 165 (2018) F1271F1277. [51] S. Geiger, O. Kasian, B.R. Shrestha, A.M. Mingers, Mayrhofer, Karl J. J., S. Cherevko, Activity and Stability of Electrochemically and Thermally Treated Iridium for the Oxygen Evolution Reaction, Journal of The Electrochemical Society 163 (2016) F3132-F3138.

23