Investigation of electrocatalysts for anion-exchange ...

HAL Id: tel-03463623https://tel.archives-ouvertes.fr/tel-03463623

Submitted on 2 Dec 2021

HAL is a multi-disciplinary open accessarchive for the deposit and dissemination of sci-entific research documents, whether they are pub-lished or not. The documents may come fromteaching and research institutions in France orabroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, estdestinée au dépôt et à la diffusion de documentsscientifiques de niveau recherche, publiés ou non,émanant des établissements d’enseignement et derecherche français ou étrangers, des laboratoirespublics ou privés.

Investigation of electrocatalysts for anion-exchangemembrane fuel cells

Pietro Giovanni Santori

To cite this version:Pietro Giovanni Santori. Investigation of electrocatalysts for anion-exchange membrane fuel cells.Material chemistry. Université Montpellier, 2019. English. �NNT : 2019MONTS129�. �tel-03463623�

I

THÈSE POUR OBTENIR LE GRADE DE DOCTEUR

DE L’UNIVERSITÉ DE MONTPELLIER

En Chimie des Matériaux

École doctorale Science Chimiques Balard ED459

Unité de recherche UMR 5253 ICGM AIME

Présentée par Pietro Giovanni Santori Le 29 Novembre 2019

Sous la direction du Dr. Frédéric Jaouen

Devant le jury composé de

M Frédéric JAOUEN, Chargé de Recherche, Université de Montpellier

Mme Elena SAVINOVA, Professeur, Université de Strasbourg

M Nicolas ALONSO-VANTE, Professeur, Université de Poitiers

M Serhiy CHEREVKO, Researcher, Helmholtz Institute Erlangen-Nürnberg

Mme Deborah JONES, Directeur de Recherche, Université de Montpellier

Directeur de thèse

Rapporteuse

Rapporteur

Examinateur

Présidente de Jury

Investigation d'électrocatalyseurs pour les pi les à

combustible à membrane anionique

Investigation of electrocatalysts for anion-exchange

membrane fuel cells

I

II

Table of Contents

Acknowledgements .................................................................................................................... 2

Chapter 1. Introduction ............................................................................................................ 4

Energy Situation .......................................................................................................... 4

PEMFC state of the art ................................................................................................ 7

AEMFC ....................................................................................................................... 10

Fe-N-C catalysts ......................................................................................................... 18

Hydrogen Peroxide Scavenger: Manganese Oxides.................................................. 30

Summary of the objectives of the different chapters ............................................... 31

References ............................................................................................................................ 34

Effect of Pyrolysis Atmosphere and Electrolyte pH on the Oxygen Reduction

Activity, Stability and Spectroscopic Signature of FeNx Moieties in Fe-N-C Catalysts

............................................................................................................................ 46

Abstract ..................................................................................................................... 47

Introduction............................................................................................................... 47

Experimental ............................................................................................................. 53

Results ....................................................................................................................... 56

Conclusions................................................................................................................ 68

References ............................................................................................................................ 70

Effect of Mn-oxides on the Oxygen and Peroxide Reduction Reactions for

MnOx/FeNC Composites in Alkaline Medium ....................................................... 72

Abstract ..................................................................................................................... 73

Introduction............................................................................................................... 73

Experimental ............................................................................................................. 78

Results and discussion ............................................................................................... 82

Conclusions................................................................................................................ 95

III

References ............................................................................................................................ 97

Mechanisms of Manganese Oxide Electrocatalysts Degradation during Oxygen

Reduction and Oxygen Evolution Reactions ....................................................... 104

Abstract ................................................................................................................... 105

Introduction............................................................................................................. 105

Experimental ........................................................................................................... 108

Results ..................................................................................................................... 111

Discussion ................................................................................................................ 120

Conclusions.............................................................................................................. 124

References .......................................................................................................................... 126

Critical importance of the ionomer on the electrochemical activity of platinum and

non-platinum catalysts in anion-exchange membrane fuel cells ....................... 130

Abstract ................................................................................................................... 131

Introduction............................................................................................................. 131

Experimental ........................................................................................................... 132

Results & Discussion ................................................................................................ 133

Conclusions.............................................................................................................. 138

References .......................................................................................................................... 140

Highly active and durable Fe0.5-NH3 cathode in AEMFC ..................................... 142

Introduction............................................................................................................. 143

Experimental ........................................................................................................... 144

Results ..................................................................................................................... 146

Conclusions.............................................................................................................. 155

References .......................................................................................................................... 156

Conclusions .......................................................................................................................... 158

Appendix .......................................................................................................................... 162

IV

Acronym .......................................................................................................................... 178

Résumé de la these ................................................................................................................ 181

1

2

Acknowledgements

I truly want to thank my supervisor Dr. Frédéric Jaouen for having guided me, taught me and

be patient, helping me to improve my skills and opening my mind for the analysis of the data,

as well as listening and respecting all my ideas, including the most unclear ones.

Special acknowledgement to the jury members Prof. Elena Savinova, Prof. Nicolas Alonso-

Vante, Dr. Serhiy Cherevko and Dr. Deborah Jones for their time spent on the reading and

correction of this manuscript.

I want to acknowledge all the members of the CREATE project, and in particular Dr. Serhiy

Cherevko and Florian Speck for having actively cooperate with us, studying the operando iron

and manganese dissolution with their unique setup.

I want also to acknowledge the synchrotron SOLEIL (Gif-sur-Yvette, France) for the provision

of synchrotron radiation facilities at beamline SAMBA and Dr. Andrea Zitolo, researcher at the

SAMBA beamline, for his help during the beamtime.

Thanks also to Prof. William Mustain and Dr. Xiong Peng from the University of South Carolina

(Columbia, SC, USA) for hosting me and helping me with the preparation and testing of MEAs

in fuel cells.

Special thanks to the past and current members of the research team, who helped me to

become the researcher I am. In chronological order, I want to acknowledge Dr. Sebastian

Brüller, Dr. Nastaran Ranjbar, Dr. Jingkun Li, Dr. Aaron Roy and Dr. Nicolas Bibent. You all have

helped me to carry out scientific investigations from both practical and theoretical aspects

through daily exchange, and helping me to watch the research from several points of view.

I also want to thank Dr. Nicolas Donzel for the great help during my first approach to the fuel

cell system, for all the BET measurement and more generally, for his great help for my

integration in the AIME team.

I want also to thank Dr. Marc Dupont and Dr. Frédéric Lecoeur for taking care of the lab.19

and for their help during the fuel cell tests.

Acknowledgments to Dr. Moulay Sougrati for the Mössbauer spectroscopy studies done on

my samples and for his time on analysing the results.

Special thanks to all the AIME members, for being always available for any question,

experiment or to share a moment or a coffee together.

3

Last, I would like to acknowledge the project CREATE (Critical Raw materials Elimination by a

top-down Approach To hydrogen and Electricity generation) funded from the European

Union’s Horizon 2020 research and innovation programme under grant agreement No

721065, that financed my PhD studentship.

4

Chapter 1. Introduction

Energy Situation

Over the last two centuries, the strong increase in world population combined with improved

living standards and industrialisation has resulted in a continuously growing demand for

energy.1 This demand has hitherto been mainly satisfied by the exploitation of different types

of fossil fuels, from coal to crude oil, natural gas and recently shale gas and shale oil.2,3

Unfortunately, the use of fossil energy sources poses two main problems in the long term: i)

fossil fuel reserves are limited and the rate at which they are naturally regenerated from dead

organisms is negligible compared to their current rate of consumption by human activities, ii)

the combustion of fossil fuels releases CO2 in the Earth atmosphere and other substances

(NOx, unburned particles, etc), which lead to local pollution in urban environment and in a

global increase in the CO2 concentration in the Earth’s atmosphere.4,5 These local and global

changes of the atmosphere affect the human health and the climate. In particular, the risk of

global climate changes induced by human activities has become of critical importance due to

dramatic prevision that forecast an increase of the average surface temperature on Earth by

up to 7°C in 2100.6

The expressed political goal, in some of the world’s regions and in particular Europe,7 of

cutting CO2 emissions related to the burning of fossil fuels challenges the industry and

scientific community to improve the efficiency of existing technologies and, in the medium or

longer term, to switch entirely to decarbonized energy solutions.8,9 Pathways to gradually

switch to a decarbonized energy production sector have been proposed across numerous

countries (Figure 1). This obviously implies the need for research & innovation in renewable

energy sources at large, from their production to storage and transformation into different

forms of energies that are useful for the end-users, such as electric power, heating power and

mechanical power. New energy converting systems and novel advanced materials will be

needed, in order to successfully replace the existing technologies based on fossil fuels, with

similar or improved performance, including cost, durability and efficiency. 10

5

Figure 1 Summary of the objective of 139 countries for the next 30 years. The graph represents the

present and projected energy power capacity for the main types of energy resources as a function of

years. The black area mainly represents fossil fuel energy. The aim is to gradually replace fossil and

nuclear energies with renewable energy, to reach 50 % renewable energy in 2025 and 100 % in 2050.

Beyond the necessary development and deployment of various renewable energy sources

such as solar, wind, hydro and tidal energies for example, the production of decarbonized

fuels, or green fuels, will be necessary for the transportation sector in particular, and also as

a buffer to store renewable electricity.11-14 The latter will be produced in large amounts but

in a non-continuous and non-predictable manner at daily and seasonal timescale by

windfarms and solar panels. A strong interest is growing in the direct production of hydrogen

from electric power, as a green fuel, and its efficient reconversion into electric power within

different types of fuel cells.15-17

The general concept of fuel cell operation is the direct conversion of the chemical energy

contained in the fuel into electric energy, without resorting to combustion.18-20 The fuel is

continuously fed at the anode side where it is electrochemically oxidized, while the reaction

at the cathode is, for terrestrial application, almost always the electroreduction of the

dioxygen from air. A single-cell fuel cell is thus composed of two electrodes (anode and

6

cathode) that are electrically separated within the device by a thin layer of electrolyte, which

can be a liquid, a solid or a polymer separator, depending on the fuel cell type and its

operating temperature. Outside the device, the anode and cathode are electrically connected

to a load, which can be any electric device that will consume the electric power generated by

the fuel cell.

Thus, in strong contrast to a battery, a fuel cell does not contain any chemical or electric

energy by itself, but converts the chemical energy of the fuel into electric power, as long as

the fuel is fed into the anode side. If the fuel is H2, the fuel cell system will thus comprise both

a fuel cell and a H2 tank, thereby decoupling the function of converting H2 into electric power

from that of storing H2.

While several types of fuel cell exist, defined by the type of electrolyte (solid-oxide fuel cell,

molten carbonate fuel cell, proton ceramic fuel cell, etc15) or by the nature of the fuel

(methanol fuel cell, ethanol fuel cell15), in this thesis we will focus on two fuel cell devices,

namely the Proton Exchange Membrane Fuel Cell (PEMFC) and the Anion Exchange

Membrane Fuel Cell (AEMFC). Those two systems have in common the polymeric nature of

the electrolyte, the operating temperature range (sub-zero to circa 80-100°C) and the need

for a humidified environment to reach a sufficient ionic conductivity of the polymer

membrane.21,22 They differ by the sign of the electric charge of the ions that the polymer

membrane conducts, i.e. cations for the PEMFC, and anions for the AEM, as can be easily

understood from their name. The operating principle of H2/air PEMFC and AEMFC is shown in

Figure 2, including the flow of electrons outside the device, the flow of ions through the

membrane, the incoming gases and the electrode outlet where the highest amount of

electrochemically produced water is expected.

7

Figure 2 Scheme of the operating principle of PEMFC (left) and AEMFC (right). Reproduced from Ref.

23.

Due to the relatively low operating temperature of PEMFC and AEMFC and the high stability

of the double bond in O2, the power and durability performance of such fuel cells entirely

depend on the nature and amount of catalysts, which must be appropriately chosen for each

reaction and each polymeric environment. The current status and challenges of the PEMFC

and AEMFC are shortly described in the following paragraphs.

PEMFC state of the art

R&D on PEMFC is ongoing since the 1960’s, and the advent of the highly conductive and stable

Nafion® membranes from Dupont was a first breakthrough for this technology.24,25 The drastic

reduction of the platinum content at the cathode from several mg·cm2 for platinum black

down to 0.4 mg·cm2 for advanced cathodes in the early 1990’s by Gottesfeld group at Los

Alamos National Laboratory was a second breakthrough toward improved power

performance and decreased cost.26 The key for this achievement was the nano-structuration

of Pt and its dispersion on high surface area carbon and, even more important, the

understanding of the need for providing a proton-conduction path inside the cathode active

layer.26,27 The triple-phase boundary (electron path, proton path and gas phase path, Figure

8

3a) has since been a cornerstone concept for designing high performance gas diffusion

electrodes and for achieving high catalyst utilization, even though O2 and H2 molecules can

also reach the catalyst surface via diffusion through a thin ionomer film (Figure 3b).

Figure 3a. Scheme defining the triple phase boundary (TPB). Reproduced from 28.

Figure 3b. Scheme for an active layer comprising macropores and catalyst agglomerates. These pore-

free agglomerates comprise primary catalytic particles embedded in a matrix of ionomer. Reproduced

from 29.

Further improvements in Pt-catalysts (Pt-M alloys, core-shell structures, etc,30-34) and all fuel

cell components (membranes, gas diffusion layers cell design) have then contributed to

significantly increase the power performance and durability of H2/air PEMFC, now reaching

typically 1.5 A cm-2 at 0.6 V,35 which has allowed their implementation in the first commercial

9

fuel cell vehicles by Hyundai and Toyota.36,37 While this technology now shows good

performance with clean H2 (free of sulfur and CO contaminants) and air, its large scale

commercialization is threatened by high raw materials costs. While the production cost of

almost all fuel cell components will decrease with increasing volumes of production according

to economies of scale (typically, cost divided by two for every x10 increase in production), the

cost of catalysts based on platinum group metals (PGM) will move in the opposite direction

due to the increasing demand and stagnating supply of such rare metals. As a result, the

scenarios identify that PGMs will account for the major fraction of the raw material’s cost of

an automotive fuel cell stack, even for relatively low volume production of 500,000 units per

year for the automobile industry (Figure 4).38-40

Figure 4 Scheme on the expected price of fuel cell components as a function of production volumes

of fuel cell stacks per year. The main fraction of the costs is due to the electrode catalysts, especially

at high production volumes.

While the loading of PGM and Pt in particular at the cathode, has been reduced from ca 0.4

mg·cm2 in the early 2000’s to circa 0.2 mg·cm2 nowadays,41-43 with concomitant increase in

PEMFC power performance, further decreasing the Pt loading at the cathode without

sacrificing power performance seems difficult. This will result in a lower-end minimum value

10

of platinum mass per kW rated electric power for PEMFC stack, which may be too high to

compete with the low cost of internal combustion engines. In the acidic environment of

PEMFCs, the anode side, where the hydrogen oxidation reaction (HOR) takes place, is not an

issue due to the almost reversible HOR on Pt in acidic medium. Pt loadings as low as 50-100

µgPt cm-2 can be used, with sufficient durability.44-46

In summary, Platinum is a rare and precious metal, currently used at both the anode and

cathode of PEMFCs, and with no viable alternative solutions currently. Intense R&D efforts

are ongoing in the field of PGM-free catalysis of the ORR in acid medium. While several

breakthroughs on power performance and understanding in PGM-free cathodes for PEMFCs

have been realized since 2009, the long-term durability of PGM-free catalysts in the acidic

environment is, at present time, a grand challenge for an industrial application.39,47-49 Beside

their poor durability during operation, PGM-free ORR catalysts have also not yet reached the

high activity needed for competing with Pt-based catalysts in air/H2 PEMFCs. Even with ten

times thicker electrodes, PGM-free Metal-N-C cathodes still deliver less power than state-of-

art Pt/C cathodes.

The cost reduction of automotive fuel cell stacks and the dependence on Pt for ORR catalysis

are thus remaining issues for the large-scale deployment of PEMFCs. In contrast, the high-pH

environment in high performance AEMFCs, as recently achieved, now offers the opportunity

to switch to PGM-free AEMFCs, due to the high stability of numerous PGM-free materials in

this environment. To really compete with state-of-art PGM-based catalysts, advances in both

cathode and anode catalysts and electrode structuration are however needed. The status and

challenges of AEMFC is described below.

AEMFC

While the research community now tackles the remaining challenges facing the application

of M-N-C catalysts in PEMFC cathodes (higher activity and, especially, higher stability), an

alternative route to apply M-N-C catalysts in low-temperature, high-performance fuel cells

recently surged, with the advent of highly conductive anion-exchange membranes, and fast

advances in the preparation of MEAs for AEMFC, having resulted in extraordinary increase in

11

the power performance of AEMFCs over the last 7 years.23 Figure 5 summarises the growth

in the number of papers published since 2000 on AEMFCs (right) and the reported increase in

the initial power performance of AEMFC (left).23 When in their OH-form (no carbonation from

air), the AEM correspond to a high-pH environment, leading to a paradigm shift in the

development of low-temperature polymer-electrolyte fuel cells, shifting away from the acidic

environment of PEMFCs. In contrast to PEMFC, the AEMFC leans on the conduction through

the AEM of OH- anions, from the cathode to the anode, where water is produced by the

combination of hydroxyl anion and H2 (Fig. 2).

Cathode : O2 + 2 H2O + 4 e- 4 OH- (1.1)

Anode: 4 OH- + 2 H2 4 H2O + 4 e- (1.2)

AEMFCs open the door to the potential application of numerous PGM-free catalysts, either at

the anode and/or cathode, due to the expected stability of many non-PGM metal and metal

oxides at such equivalent high-pH value. However, to date, most of the high-performance

AEMFC data has been reached with PGM-based catalysts at the cathode and anode, and

efforts focused primarily on the development of AEM and Anion exchange ionomer (AEI), as

well as on the MEA preparation and operating conditions leading to high performance.50-53

This was particularly true in early 2017 at the starting time of this PhD thesis.

Figure 5 The histogram on the left summarizes the number of papers published each year since 2000

on the topic of AEMFC (Source: Web of Science). The figure on the right reports the increase in the

AEMFC performance (H2/O2) since 2006. Reproduced from Ref. 23.

12

AEM and AEI conductivity and stability

Initially, the main challenge in the field of AEMFC was the lower ionic conductivity of AEMs

compared to PEMs, intrinsically related to the lower mobility of OH- compared to protons in

PEMFC.54,55 To improve the conductivity, several research groups have investigated many

different polymer backbones, onto which the cationic functional groups can be covalently

bound, either directly or indirectly via a side chain (an alkyl or aromatic types). The main

chemistries studied to prepare AEMs were summarized by Varcoe et al in a review in 2014.56.

The principal chemistries of the polymer backbone are poly-(phenylene oxides) (PPO);57-61

polybenzimidazole (PBI) AEMs including cationic functional groups directly bound to the

backbone;62-66 electrospun fibres;67,68 PTFE-reinforced polymers;69-71 ethylene (tetrafluoro

ethylene) (ETFE) based AEMs obtained by the radiation-grafting of vinyl groups on the

backbone and, more recently, high density polyethylene (HDPE) AEMs with improved

stability, both ETFE and HDPE having been developped in Varcoe’s group.51,72 In particular the

latter two AEMs, combined with the use of ETFE-based powder ionomer, resulted in great

performance in AEMFC, as highlighted by Mustain’s group in the last two years.73-76

The cationic functional group plays a key role in the conductivity and stability of the AEM and

AEI. The main cationic groups studied hitherto for AEM and AEI can be classified as: i)

quaternary ammonium (QA) such as benzyl-ammonium, alkyl-bound QAs and QAs based on

bicyclic ammonium systems; ii) heterocyclic systems including imidazolium,

benzimidazoliums, PBI systems and pyridinium types. 62-66 Recently the use of ETFE-based AEI

has been obtained in solid form and combined with the catalyst, obtaining great results in

operating AEMFC.53,74

The main drawback of these types of functional group is their stability, in particular in highly

alkaline environment, where OH- can attack the polymer backbone, the sidechain and,

particularly, the cationic group itself, by nucleophilic attack, or, in the case of the presence of

a β-Hs, by Hoffmann elimination (Figure 6).

13

Figure 6 Scheme of the nucleophilic attack of the OH- on benzyltrimethylammonium cationic group.

In the dashed-box, the Hoffman Elimination mechanism occurring on the alkyl-bound of the QA group,

which possess a β-H. Reproduced from Ref. 56.

In addition, the formation of HO· radical, which are extremely aggressive, can further

decrease the lifetime of the AEM and/or AEI in the operating device, since such radicals can

be produced in operando at the cathode (ORR by-products) or in the membrane (via cross-

over of H2 and O2). For these reasons, research efforts focused on the study of different

backbone and functional group chemistries in order to enhance the stability and conductivity

of AEM and AEI.

To improve membrane stability, suitable polymer chemistry of the backbone have to be

chosen. For example, non-fluorinated or fully fluorinated polyolefin backbone showed good

stability at high pH, while partially fluorinated polyolefin undergoes degradation due to

dehydrofluorination derived by the OH--attack, with the consequent chain scission.77-79

Further studies identified three possible approaches to enhance the lifetime of AEMs: the

removal of electro-withdrawing cationic groups from the proximity of the aryl-ether linkage

by the addition of a spacer in the sidechain; the removal of sulfur and fluor from the polymer

14

backbone chemistry in order to limit their electro-withdrawing effect; the synthesis of

polymers free of aryl-groups.80-83 An example was reported in 2017, where ETFE-AEM

performance was compared to LDPE-AEM (the backbone chemistries are summarized in

Fig.7), showing better performance and stability for the second backbone.

N+

CH3

CH3 CH3

CH3O

CCR R

H

H

H

m

n

N+

CH3

CH3 CH3

CH3

CC C

H

H

H

C

F

F

F

F

RR

x

mn

Polyethylene (PE)Ethylene tetrafluoroethylene (ETFE)

Figure 7 Representation of the ETFE (left) and of the PE (right). The side chains and the cationic groups

are the same, while the polymer backbones are different. The absence of fluoride groups in the

backbone increased the stability of the AEM.51,84

Carbonation

Compared to PEMFC, the risk of partial carbonation of AEM and AEI is a specific challenge,

which can result in drastic decrease of AEM and AEI conductivity, and thus of the AEMFC

performance. QA groups in OH- form are not stable in ambient conditions, due to their higher

affinity for CO2, which is converted into carbonate (CO32-) and bicarbonate (HCO3

-).23 Those

anions have a larger radius than the hydroxyl anion and therefore a lower mobility.85 This

consequently decreases the AEM conductivity drastically, with a 75% decrease of the

conductivity typically observed from 100% OH-form to a fully carbonated/bi-carbonated

membrane, and a 50% conductivity decrease when the feeding gas is switched from pure O2

to ambient air.86-88 For this reason, the performance of AEMFC is typically studied presently

in an environment free of CO2. However, for a practical application, it will be important to

develop membrane solutions to mitigate the effect of CO32- and HCO3

- on the polymer

15

conductivity, or to develop system solutions to decrease the CO2 concentration in the cathode

feed. One option is described by Krewer et al., evidencing with a mathematical model how an

AEMFC operated at high current density (> 1 A cm-2 in their model) can lead to the

electrochemical generation of OH- groups at the cathode that is sufficiently fast to mitigate

the formation of (bi) carbonate groups from airborne CO2.89-91

Electrocatalysis

1.3.3.1 HOR

In contrast to the case in PEMFC, the electrocatalysis of the HOR in the high-pH environment

of AEMFC is not reversible, even on PGM-based catalyst. It was reported in 2010 by

Gasteiger’s group that the HOR kinetics on Pt is circa 100 times lower in pH 13 than in pH 1.92

The fundamental reason seems related to the different requirements to catalyse HOR, as is

visible from the writing of the electrochemical HOR in each pH environment.

HOR in alkaline: O2 + 2 H2O + 4 e- 4 OH- (1.2)

HOR in acid: 2 H2 4 H+ + 4 e- (1.3)

The HOR in high-pH requires two different properties from a catalytic surface, namely the

ability to bind H2 and the ability to provide hydroxyl groups, often coined as oxophylic

property (Figure 8).93,94

16

Figure 8 Scheme of the mechanism for HOR on PtRu anode catalysts. The ruthenium oxide surface

binds the hydroxyl anions while the platinum surface binds H2, synergistically increasing the HOR

kinetics in alkaline medium. Re-adapted from Ref. 95.

For this reason, much higher HOR activity is observed in alkaline electrolyte in RDE setup with

bimetallic PtRu catalysts relative to Pt, where the ruthenium oxide was described to act as an

oxophylic agent.96 The identification of the HOR activity issue in alkaline electrolyte, and the

switch from Pt to PtRu anodes has resulted in greatly improved AEMFC performance.96

Alternative PGM catalysts, without platinum has been studied. One can mention in particular

palladium nanoparticles dispersed on ceria, and ceria in turn dispersed on carbon black. Such

catalysts have demonstrated an AEMFC power performance of 500 mW cm-2 at 0.8 V cell

voltage, when combined with a Pt/C cathode.97,98

The holy grail for AEMFC and for the full exploitation of their competitive advantage vs.

PEMFC is however clearly to move to PGM-free catalysts, at both the anode and cathode. At

present time, Ni-based catalysts have principally been studied for the HOR in alkaline

medium, but the performance has hitherto remained too low in AEMFC, and/or highly active

Ni-M alloys have shown strong instability.99,100

1.3.3.2 ORR

Well known Pt/C catalysts have been applied at the cathode side to obtain benchmark results

and to optimize the operating conditions for AEMFC. Several studies and from different

groups have now reported great performance with Pt/C cathodes and PtRu/C anodes, when

combined with ETFE/HDPE AEM and ETFE-based AEI. The recent reports describe not only

Pt

Ru

17

impressive initial performance, with peak power density above 2.5 W cm-2 and current density

close to 7 A cm-2, but also promising stability for more than 500 h of steady-state operation.72-

74

The high-pH environment of AEMFCs however holds the promise to replace PGM-based

catalysts with PGM-free ones, without sacrificing power performance and stability. Numerous

PGM-free ORR catalysts have been identified from RDE studies in liquid electrolyte, some

many decades ago.101-111

While many studies of non-PGMs cathode catalysts have been performed in RDE, only few

studies have been carried out in operating AEMFC hitherto. Tammeveski et al. reported

AEMFC results using metal nitrogen-doped carbon materials as cathode (loading of 1-5 mg-2)

and Pt/C as the anode (40 wt% Pt) with a loading of 0.3-0.5 mgPt cm-2. The membrane

electrode assembly was prepared with Tokuyama A201 series membrane and AS-4 ionomer.

The results showed a maximum peak power density of 125 mW·cm-2 and a current density of

400 mA·cm-2.112 Another work from Joo et al in 2016 combined a cathode based on carbon

nanotube (CNT) pyrolyzed with ironIII porphyrin chloride (cathode loading of 2 mg·cm-2), an

anode based on Pt/C (40 wt% Pt) with a loading of 0.5 mgPt cm-2 and in-house AEM and AEI.

The polarization curves described good performance, reaching ca. 0.4 W cm-2 at 0.4 V and a

maximum current density above 1.2 A cm-2 at 0.2 V.113

While many catalytic systems free of PGMs are good candidates for the cathode, reaching the

same activity per mass of catalyst than the state-of-art Pt/C materials (with up to 40-60 wt %

Pt as nanoparticles) can be challenging. For example, metal oxides are typically non-

conductive, and loading a high wt% of metal oxides on carbon could impair the cathode

conductivity. The interaction between metal-oxide or non-metallic catalysts with the AEI may

also be different from that of Pt/C, and the successful implementation of PGM-free catalysts

still needs to be demonstrated in the novel MEA designed for AEMFCs.

As an alternative to Pt-based cathode catalysts in AEMFC, silver-based catalysts have also

been applied, reporting good performance even in combination with non-PGM anodes.100

While Ag might be perceived as a precious metal, it is much less expensive than platinum. A

few other PGM-free ORR catalysts have also been investigated in AEMFC over the last years,

evidencing progress toward precious metal free cathodes.114-116 Even more recently, in 2019,

Peng et al. reported great performance obtained using cobalt-ferrite dispersed on Vulcan

18

carbon, with peak power density up to 1.8 W cm-2 in H2/O2 and up to 0.8 W cm-2 in H2/air,

respectively. 75

Figure 9. Summary of AEMFC state-of-art performance with non-PGM cathode catalysts: FU,117 Fe-N-

C nanotubes,118 N-C-CoOx,119 CNT/PC113 and CF-VC75. Reproduced from Ref. 75.

Fe-N-C catalysts

One of the main class of materials being investigated as an alternative to Pt-based catalysts

at the cathode side of a fuel cell is the so-called Metal-Nitrogen-Carbon (M-N-C) family. While

their preparation methods, type of sources for the M, N and C elements and heterogeneity of

active sites differ vastly among the different research groups, all those materials share the

simultaneous presence of carbon (typically > 85 at %), nitrogen (typically in the range 1-10 at

%) and the presence of naturally abundant and inexpensive transition metals from the 3d row

(e.g. Mn, Fe, Co, Ni, Cu),120-122 although recent reports have also explored M-N-C materials

prepared with heavier metals.123 The presence of nitrogen in the synthesis is critical to reach

high activity, since there is growing evidence that the most active sites for the oxygen

reduction reaction (ORR) (especially in acidic medium) are atomically-dispersed MNx

moieties.124-127 However, most catalysts prepared via a pyrolysis step also contain a small, or

even a large, fraction of metallic or metal-carbide particles as well.128-130 Depending on the

nature of the metal, such particles can form core-shell Metal@NC structures during pyrolysis,

with non-negligible ORR activity, especially in alkaline medium.131 For application in acidic

medium, the stabilization of such 3d metals by coordination with nitrogen is essential not only

for activity but also for stability aspect, since they would otherwise be unstable in their

metallic form, and would leach in the electrolyte as metal cations. In the case of Metal@NC

19

particles, the N-doped carbon shell can theoretically protect the metallic core from leaching,

but the shell thickness must be thin enough for ORR activity and thick enough for stability.132

The class of M-N-C materials has shown promising activity and power performance for

catalysing the ORR in a broad range of pH, particularly Fe-N-C materials.102,133-135

The first report on a M-N-C material dates back to 1964, with the bio-inspired work from

Jasinski who reported that a Co-phthalocyanine is active for ORR in alkaline medium.136 This

initial work triggered research on the investigation of the ORR activity of MN4 macrocycles

supported on carbon as inexpensive catalysts in the 1970-80’s, either in alkaline or acidic

medium. Several macrocycles based on a central transition metal cation (Fe, Mn, Cu etc)

coordinated by four nitrogen atoms were shown to be ORR active at low pH, but suffering of

low stability.48,137 Subsequent studies in the second half of the 1970’s addressed the stability

issue of such active sites by applying a heat treatment to MN4 macrocycles supported on

carbon (MN4/C), in a controlled atmosphere.138-140 The heat treatment was applied in the

range of temperature of 800-900°C in protective inert gas atmosphere, and it was claimed

that this kept the N4-coordination of the metal centre intact, but the remainder of the

macrocyclic structure was destroyed and new bonds between N-containing groups and the

carbon support were formed.141,142 After the heat treatment of cobalt complexes

(phthalocyanine, tetraphenylporphyrin, tetrabenzoporphyrin, tetra(p-metoxyphenyl-

porphyrin), it was reported by Bagotzky et al that the electrochemical activity was unchanged,

but the stability in acidic medium was improved.139

If the activity and stability of heat-treated MN4/C catalysts were intensively studied by several

groups, the structure of the active sites after the heat treatment was less investigated, due

to the technical challenge at that time for a structural characterization of the allegedly single

metal-ion centers after pyrolysis, especially due to their dilution (typically 0.5-2 wt%) in the

carbon matrix.141,143-145

For the first time in 1988, Yeager et al. synthetized Co and Fe based catalysts, mixing

polyacrylonitrile (PAN) with metal salts (metal acetate) and Vulcan XC-72. The precursors

were mixed in 100 mL of dimethylformamide (DMF) and the solvents removed by evaporation

under N2 environment. The catalyst precursor powder was then pyrolyzed at 800°C in argon.

The ORR activity was fairly high in both alkaline and acidic electrolytes, demonstrating for the

20

first time that the formation of Fe or Co-based active sites at high temperature does not

require the use of metal-N4 macrocycles as a starting precursor, opening the route to the

modern synthetic procedure of MNC catalysts.146

This breakthrough also opened the still ongoing debate on the nature of the active sites in

pyrolyzed MNC catalysts. Time-of-flight secondary-ion mass spectroscopy (TOF-SIMS) and X-

ray absorption spectroscopy (XAS) were the first techniques applied to support the existence

of atomically dispersed MNx moieties in such materials.147,148 For example, in 2002, Lefèvre et

al. characterized with TOF-SIMS and RDE three different series of Fe-N-C catalysts, which

differed by either the iron precursor used during the synthesis (iron porphyrin or iron

acetate), the Fe content (0.2 or 2.0 wt % Fe on carbon) and the pyrolysis temperature (400-

1000°C), to evidence activity – structure correlations. The study revealed the presence of two

different catalytic sites, one labelled FeN4, present in all their catalysts, and a second one,

labelled FeN2, mainly observed on Fe-N-C catalysts prepared from iron acetate precursor and

pyrolyzed at high temperature (700-900°C). The study revealed a positive correlation

between the TOF-SIMS signal intensity for FeN2 sites and the ORR electrochemical

activity.149,150 Due to the destructive TOF-SIMS method, this does not however imply that the

active site coordination is restricted to FeN2. The identification of clear correlations between

the synthetic path for pyrolyzed MNC materials and their resulting activity became of critical

importance in 2000-2010 in order to rationally improve the performance of Fe-N-C and Co-N-

C catalysts and to compete with the PGM-based catalysts for PEMFC application. In 2006,

Jaouen et al. revealed a correlation between the amount of micropores formed during

pyrolysis in flowing NH3 (used as sole nitrogen precursor) and the ORR activity of FeNC

materials prepared via impregnation of iron acetate on an initially non-microporous carbon

black.151 It was shown that ammonia pyrolysis leads to a weight loss of carbon, etching the

carbon support as volatile NC compounds (likely HCN) and simultaneously creating

micropores and introducing N-groups on the carbon surface. A positive correlation between

the microporous surface area and the ORR activity was evidenced, suggesting that the key

factor limiting the ORR activity is the micropore area, while the N-content was sufficient for

all catalysts. It was hypothesized that the FeNx sites, or at least the most active ones, are

located inside micropores and that micropores are needed to efficiently transform the Fe salt

into FeNx sites.

21

Since the ORR activity of MNC materials was shown to be strongly tied to the porous structure

of the carbon support either before or during the pyrolysis (e.g. via a chemical reaction with

ammonia), in 2008 efforts addressed the identification of the key parameters of the carbon

support for preparing highly active FeNC catalysts. Before 2008, three structural or

composition parameters were found to be fundamental: i) the amount of disordered carbon

in the carbon support,152 ii) the diameter of the particles, and iii) the average lateral size of

the graphene layers.153 Parameter i) is important since disordered carbon reacts faster with

ammonia than graphitic carbon, and this difference in etching rate in carbon black particles

with graphitic and amorphous domains leads to the formation of micropores via preferential

etching of the disordered phase, in turn promoting the formation of FeNx active sites.

Based on the results showing the importance of micropores as well as the 57Fe Mössbauer

spectroscopy and EXAFS results, the hypothetical structure and location of the FeNx active

sites was proposed (Figure 10).153,154

The iron cation was proposed to be coordinated by either two or four nitrogen atoms in the

first coordination sphere, proposing a combination of FeN4, FeN2 or FeN2+2 sites, in agreement

with the TOF-SIMS results. Such sites can be localized in-plane (FeN4), on the edge of a

graphite plane (FeN2) or bridging two different crystallites (FeN2+2).153 For the FeN2 site, the

authors stated that the coordination is probably incomplete, and the remainder of the

coordination was left undefined.

22

Figure 10 Proposed structures of catalytic sites in FeNC catalysts. FeN2/C: previously incomplete

model of FeN2+2/C. FeN4/C: possible structure of in-plane catalytic sites in FeNC catalysts. Reproduced

from Ref. 153.

In 2009, Science from Lefèvre et al.135 reported a new synthetic procedure to better use the

micropores present in carbon black, overcoming the limitations intrinsic to the impregnation

method. To do this, two pore fillers were selected (perylene-tetracarboxylic dianhydride or

1,10-phenantroline) and mixed to an iron salt and to the carbon support, then the mixture is

milled in planetary ball milling. During the mixing, the friction and impacts force the iron salts

and the pore filler inside the micropores of the carbon black. The mixed catalyst precursor is

then pyrolyzed in Ar and then in NH3 , or directly in NH3 in the case of the nitrogen-free pore

filler.135

Since a method to efficiently use the microporosity in the carbon support before pyrolysis

was established in 2009,135 leading to a breakthrough in in the initial activity and PEMFC

performance of FeNC catalysts, more efforts have subsequently addressed the use of

23

different microporous materials for preparing MNC materials, in particular the use of Metal-

Organic Frameworks (MOFs). MOFs are crystalline materials comprising metal ions

coordinated by organic ligands, of interest to the catalysis and gas separation community due

the presence of well-defined metal center coordinations, high porosity, high specific area (up

to 6200 m2g-1), well defined pore size, and, in some cases, their scalable and inexpensive

synthesis.155-159

The group of Liu at Argonne National Laboratory pioneered in 2010 the use of MOFs as a

precursor to synthetize MNC catalysts via pyrolysis, showing that the ZIF-67 MOF structure

(cobalt ions and methyl-imidazole ligands) was transformed into a microporous N-doped

carbon with cobalt active sites at temperatures above 600°C.160-162 The amount of cobalt in

ZIF-67 was however above the optimum amount of cobalt, leading to the formation of

metallic cobalt particles during pyrolysis.160,161 In parallel, the group of Dodelet at INRS

investigated the use of ZIF-8 (Zn(II) and 2-methylimidazole ligand, with same crystalline

structure as ZIF-67), with the advantage over ZIF-67 that the high amount of Zn can be mostly

removed as volatile products during pyrolysis, due to the low boiling point of Zn. With ZIF-8,

the amount of the ORR-active metal, Fe or Co, can thus be adjusted while it cannot with ZIF-

67. The first study on the preparation of FeNC by ZIF-8 came out in 2010 from Dodelet’s group,

and showed further improved activity and performance in PEMFC than in the Science paper

from 2009. The performance reported by Proietti et al. achieved a current density of 1.2 A

cm-2 at 0.6 V in PEMFC and a peak power density of 0.91 W cm-2.163 The synthesis involved

Basolite® Z1200 (commercial ZIF-8 from Sigma Aldrich) for its high microporous surface area

(1700 m2 g-1), mixing it with optimized amounts of 1,10-phenantroline and iron acetate.163

The catalyst precursor powder was mixed via planetary ball milling and then pyrolysed in

argon at 1050°C and then in ammonia at 950°C. The heat treatment transforms the MOF into

a highly microporous nitrogen-doped carbon structure (1000 m2 g-1), where the FeNx-sites are

allegedly hosted. The Fe speciation in those novel FeNC catalysts had not yet been studied at

that time.

In the following years, similarly prepared catalysts have been intensively studied, in particular

the Fe speciation and structure of their active sites, in order to better elucidate the active site

structure and to understand the ORR mechanism. In 2015, Zitolo et al. investigated the

structure of Fe-based sites in FeNC and the effect of the synthetic path on the formation of

24

FeNC catalysts free of inactive particles. To study the structure of the catalyst, 57Fe Mössbauer

spectroscopy and XAS was applied, evidencing the exclusive presence of atomically dispersed

FeN4 moieties. 57Fe Mössbauer spectra (Figure 11) shows that three FeNC catalysts prepared

via dry ballmilling pyrolysis and with only 0.5 wt% Fe before pyrolysis have the same spectra,

with two doublets that are characteristic for FeNx sites. The catalyst prepared with 1 wt% of

iron in the catalyst precursor and wet impregnation, followed by dry ballmilling, leads to

metallic and metal carbide particles due to excess iron during pyrolysis (Figure 11-a).124

Figure 11 Mössbauer spectra of: a Fe1.0; b Fe0.5; c Fe-900; d Fe-950. All the catalysts were obtained

mixing ZIF-8 (80 wt%), 1,10-phenantroline (20 wt%) and Fe(Ac)2 0.5 wt% (b, c, d) or 1 wt% (a). The

precursors were mixed together either directly by ballmilling or after a wet impregnation step, and

then pyrolyzed in Ar at 1050°C (a, b) and further pyrolyzed in NH3 at 900°C (c) or 950°C (d). Reproduced

from Ref. 124.

While Mössbauer spectroscopy shows that the three catalysts derived from 0.5 wt% Fe

contain only FeNx sites, the detailed structure of the active sites cannot be derived from

Mössbauer spectroscopy due to the lack of theoretical methods until this year to compare

the experimental isomer shift and quadrupole splitting to. To better understand the structure

of the active sites, EXAFS was used. The experimental FT-EXAFS spectra could be well fitted

with the assumption of FeN4 structures. The strongest peak around 1.2-1.4 Å that describe

25

the first coordination shell was fitted, evidencing the presence of 4 nitrogen atoms in-plane

with Fe, and 1 or 2 axial oxygen atoms, leading to an octahedral coordination (Figure 12).

Figure 12 Fe K-edge EXAFS analysis fitted using FeN4-moieties coordinated by one (left) or two (right)

oxygen atoms (red sphere). The Brown sphere indicate the iron and the blue spheres represent the

nitrogen. Reproduced from Ref. 124.

To better understand the structure around the X-ray absorbing Fe atoms, Fe K-edge XANES

spectroscopy was also studied, since it is more sensible to the local structure arrangement.

The fit of the spectra highlight the presence of four nitrogen coordinated to an iron centre in

the equatorial position and further coordinated in the axial position by O or N. This is

confirmed from the absence in the spectra of the pre-edge peak at 7,118 eV, which is

characteristic of a square planar structure, as observed e.g. in the XANES spectrum of Fe(II)

phthalocyanines. Several structures have been hypothesized and the calculated XANES

spectra compared to the experimental data. A good match was obtained with a structure

involving the coordination of iron by four nitrogen atoms at a distance of 2.00 Å (as previously

proposed) and the presence of dioxygen in the axial position at 1.87 Å as represented in Figure

13 a (two end-on adsorbed O2 molecules). A second structure was able to equally fit the

experimental data (Figure 13 b, side-on adsorbed O2), in agreement with the Mössbauer

spectra that evidenced two different active sites.

26

Figure 13. XANES experimental spectra (dotted line) compared with theoretical spectra (red line).

Reproduced from Ref. 124.

The investigation of the electrochemical behaviour showed that the ammonia pyrolysis does

not change the structure of the FeN4 active sites, but has a clear effect on the activity of the

catalyst, increasing the initial activity in RDE and PEMFC with respect to the catalyst pyrolyzed

only in argon. Figure 14 summarizes the results obtained in RDE, showing that the catalyst

obtained after the ammonia treatment at 950°C (red curves) is more than 10 times more

active compared to the Ar catalyst (blue curve), as can be seen in the Tafel slope (Figure 14 c

and d) at 0.8V.124

27

Figure 14. RRDE results of four FeNC catalysts obtained by: Ar-pyrolysis (Fe0.5); NH3-pyrolysis at 900°C

(Fe0.5-900); NH3-pyrolysis at 950°C (Fe0.5-950); 1.0 wt% of iron in the precursor and Ar-pyrolysis (Fe1.0).

a amount of H2O2 produced during ORR; b polarization curved measured in RRDE; c Tafel plot of RRDE

data; d Tafel plot of PEMFC polarization curves. Reproduced from Ref. 124.

The higher ORR activity of the ammonia treated catalyst is correlated to the higher basicity

(Figure 15 red square, red circle, orange circle). After NH3 pyrolysis, a positive shift of about

+4 units in the pKa of the N-groups is observed. This has been measured, dispersing Fe-N-C in

an aqueous solution of pH 6, saturated with N2 to avoid the acidification of the air. The final

pH of the solution after the dispersion of FeNC was measured once the pH value of the

solution became stable. The experiment evidenced that the basicity is positively correlated

with the electrochemical activity of the catalyst, as previously reported.164,165

28

Figure 15 Initial ORR activity as a function of the basicity of different Fe-N-C catalysts. Reproduced

from Ref. 124.

If the activity and the current density of the ammonia pyrolyzed iron nitrogen-doped carbon

in both RDE configuration and during PEMFC operation are promising, the long-term tests

evidenced a critical drop in the performance after few hours of operation, different from the

results obtained using argon treated FeNC catalysts.124,163

Figure 16 Comparison on the stability of Ar- (blue line) and NH3- (red line) treated catalysts in PEMFC

at 0.5V. Reproduce from Ref. 124.

29

Figure 16 highlights the rapid and important decay in performance of NH3-treated catalyst

(red curve) with respect to the Fe-N-C obtained by a single heat treatment in Ar. The main

difference between those two catalysts is the higher basicity and the higher microporosity

induced by the ammonia heat treatment. It was demonstrated in 2011 that a short immersion

in acidic medium of NH3 treated FeNC resulted in a decrease in the activity of a factor of ten,

partially recovered after a heat treatment at 400°C.164 This indicates that a fraction of the iron

was weakly bounded and dissolved during the first immersion in the electrolyte, while

another fraction was only reversibly deactivated. It has been proposed that the reversible

deactivation was related to the protonation of highly basic nitrogen groups and subsequent

anion adsorption. It was proposed that the protonation initially increases the Turnover

Frequency (TOF) of the FeN4 active sites (explaining the high initial activity of NH3-treated

catalysts), while further neutralization by a counter-anion present in the electrolyte reduced

the TOF (Figure 17). This mechanism is expected to be fast in liquid electrolyte, while in

PEMFC it should be slower since the electrolyte is a polymer and there are no free counter

anions, initially.164

Figure 17 Scheme of the deactivation mechanism due to the neutralization of a protonated nitrogen

near the active site. Reproduced from Ref. 164.

To better understand other degradation mechanisms, inert-gas-pyrolyzed Fe-N-C have been

considered, since they show higher durability and are not prone to this protonation effect

(Figure 16). Two main degradation mechanisms for Ar-pyrolyzed catalysts can be proposed:

i) demetallation of iron; ii) production of hydrogen peroxide, further converted into reactive

oxygen species (ROS) via Fenton reactions.166-168 The first degradation mechanism is mostly

triggered by carbon corrosion that takes place at high potential (startup/shutdown

30

accelerated stress test), leading to the destruction of FeNx sites and to iron dissolution, as

evaluated in Scanning Flow Cell combined with Inductively Coupled Plasma (SFC-ICP/MS).167

During load cycling conditions (0.6-1.0 V) in inert gas, FeNx sites are stable and the iron

dissolution can mainly be tracked to originate from Fe particles imperfectly surrounded by a

graphite shell.166,169,170 Regarding the second mechanism, the effect of exposing H2O2 to Fe-

N-C activity had been investigated first in 2003.150,171 H2O2 is an intermediate product of the

ORR, considered to be a possible reason for the low stability of Fe-N-C in operating fuel cell,

due to a direct and indirect attack.48 A direct attack of hydrogen peroxide was proposed to be

directed on N-functionalities present in the carbon support, some of them binding the metal

center, which would then lead to the leaching of iron from the active site.172 The indirect

attack of hydrogen peroxide can result from its conversion into ROS (OH·, HOO·) by Fenton

and electro-Fenton reaction that take place at low pH in presence of metal cations, in

particular Fe2+, degrading also the membrane.173,174

Hydrogen Peroxide Scavenger: Manganese Oxides

As above-mentioned, the strategy to switch to alkaline operating pH is mostly related to the

improvement of the lifetime of the Fe-N-C cathode catalyst, suppressing the main

degradation steps previously reported in acid medium. As reported in the literature, hydrogen

peroxide production usually increases at high pH, due to different ORR mechanisms and/or

due to the non-negligible activity but low selectivity of the carbon support itself (and N-doped

carbons),125,165,175,176 producing superoxide radicals (O2-·) by the conversion of the

peroxide.177,178 The latter is aggressive in particular for the membrane, and literature showed

that the hydrogen peroxide reduction reaction (HPRR) is sluggish on Fe-N-C based

catalysts.179,180

One strategy to reduce the amount of peroxide during ORR consists in the addition to the

catalyst of a co-catalyst that acts in situ to enhance the chemical and electrochemical

decomposition of HO2-. A secondary parameter for the selection of such a co-catalyst is its

activity towards the ORR, so as to avoid significant decrease of the overall ORR activity of

catalyst/co-catalyst composites. To achieve this purpose, manganese oxides have been

selected and studied in this work, supported either on Vulcan carbon or on Fe-N-C.

31

The main reason on the choice of this class of co-catalysts lies in the natural abundance of

Mn, the low cost and the relative simplicity in the preparation of the systems, in addition to

their reported chemical activity toward the decomposition of hydrogen peroxide.181-183

Manganese oxides have been widely studied for electrochemical ORR catalysis, resorting to

oxides with different structures and Mn/O stoichiometries.107,184-192 Interesting works have

been done by Ryabova et al. in 2015, deeply entering in the electrochemical mechanism of

the ORR. The study compared the efficiency of a set of different manganese oxide catalysts,

comparing their electrochemistry in both N2 and O2 saturated medium, reporting that their

activity is strongly related to the surface potential of the redox couple Mn(IV)/Mn(III) and

from the effects of the structure and the composition of the oxides.193 These electrochemical

investigations coupled with characterisation techniques and micro-kinetic models clarified

their electro-reduction activity for hydrogen peroxide in alkaline medium, highlighting the

good performance of Mn2O3 for both the ORR and HPRR reactions, highlighting it as the most

interesting system for combination with Fe-N-C.104,194-196

Summary of the objectives of the different chapters

This PhD thesis focuses on Fe-N-C and Mn-oxide/Fe-N-C composite catalysts, evaluating their

active sites, activity, selectivity and stability in alkaline medium and looking at both the ORR

and HPRR reactions. The performance of down-selected catalysts is finally investigated in

AEMFC devices, both in O2 and air.

In Chapter 2, the effect of the pyrolysis environment on the electrocatalytic properties of Fe-

N-C materials towards ORR in alkaline medium is investigated. A first catalyst was prepared

via a single heat treatment in Ar at 1050°C and a second one by subjecting the first catalyst to

a NH3 pyrolysis at 950°C. We will describe how the nature of the pyrolysis atmosphere affects

the activity and stability of Fe-N-C catalysts, comparing results obtained in alkaline electrolyte

with those observed in acidic electrolyte with rotating (ring) disc electrode. Those

electrochemical results were combined with online measurements of Fe leaching rates under

potential control, with online scanning flow Cell combined with Inductive Coupled Plasma

32

Mass Spectroscopy (SFC-ICP/MS). High activity and stability is observed for the NH3-treated

Fe-N-C catalyst in alkaline medium, while it has poor stability in acidic medium.

In Chapter 3, the aim is to identify the Mn-oxide with highest activity toward HPRR and

combine it with Fe-N-C in order to increase the selectivity toward four-electron reduction

during ORR. To do this, the ORR and HPRR activity of four manganese oxide polymorphs was

first studied by supporting them on Vulcan carbon black. The results are discussed in term of

apparent activity and activity after normalisation by the oxide surface area. Then, the HPRR

activity, ORR activity and selectivity of the four MnOx/FeNC composite catalysts was studied.

Mn2O3 is identified as the most active Mn-oxide for HPRR, and was successfully combined

with Fe-N-C, leading to improved selectivity during ORR.

In Chapter 4, the stability of the manganese oxides is investigated in 0.05 M NaOH, resorting

to the SFC-ICP/MS to evaluate the leaching of manganese in different conditions (Ar or O2

saturated electrolytes, ORR and OER potential ranges, presence or not of hydrogen peroxide).

It is found that the leaching of manganese in the ORR potential range is low in Ar-saturated

conditions, but high in O2-saturated conditions, and that the key reason for Mn-leaching is

the presence of HO2- in solution. In OER conditions, Mn leaching is observed for the four

oxides and is only related to the electrochemical potential, not to the presence of peroxide.

In Chapter 5, the effect of the nature of the anion exchange ionomer on the activity of two

ORR catalysts and two HOR catalysts was investigated. A set of anion exchange ionomers

prepared at Technion is studied in combination with two highly loaded PGMs (Pt/C, PtRu/C)

one non-PGM (FeNC) and one PGM-lean catalyst (Pd-CeO2/C). The study identifies significant

effects of the ionomer on the activity and diffusion-limited current densities, particularly

pronounced for catalysts with low wt% metal on carbon. The reasons behind these different

behaviours are discussed, such as interaction between the polymer and the catalysts, the

effect of the site density and the optimum ionomer/carbon ratio.

In Chapter 6, the performance in AEMFC is investigated with ammonia-pyrolyzed FeNC. The

MEAs were prepared according to the method reported by Omasta et al.74 The anode was

33

PtRu/C. Reference measurement with a Pt/C cathode (0.4 mg cm-2) was also recorded for

comparison. Similar activity at 0.85 V was obtained with FeNC (loading of ca 0.9 mgFeNC cm-2)

and Pt/C, demonstrating the high potential of this class of catalysts. At lower cell voltages, the

FeNC cathode reached ca 80% of the current density reached with the Pt/C cathode.

Compared to other recently studied PGM-free cathode catalysts, the results show superior

activity of FeNC compared to FeCo-, Co-oxides and Ag/C.75 The durability of the FeNC cathode

was also tested for 100 h in air, showing an initial cell voltage of 0.69 V at 0.6 A cm-2 and a

restricted activity decrease during 100 h (30%).

34

References

1 Pérez-Lombard, L., Ortiz, J. & Pout, C. A review on buildings energy consumption information. Energy and Buildings 40, 394-398, doi:https://doi.org/10.1016/j.enbuild.2007.03.007 (2008).

2 Mohr, S. H., Wang, J., Ellem, G., Ward, J. & Giurco, D. Projection of world fossil fuels by country. Fuel 141, 120-135, doi:https://doi.org/10.1016/j.fuel.2014.10.030 (2015).

3 Lu, S. et al. Classification and evaluation criteria of shale oil and gas resources: Discussion and application. Petroleum Exploration and Development 39, 268-276, doi:https://doi.org/10.1016/S1876-3804(12)60042-1 (2012).

4 Alam, M. M., Murad, M. W., Noman, A. H. M. & Ozturk, I. Relationships among carbon emissions, economic growth, energy consumption and population growth: Testing Environmental Kuznets Curve hypothesis for Brazil, China, India and Indonesia. Ecological

Indicators 70, 466-479, doi:https://doi.org/10.1016/j.ecolind.2016.06.043 (2016). 5 Höök, M. & Tang, X. Depletion of fossil fuels and anthropogenic climate change—A review.

Energy Policy 52, 797-809, doi:https://doi.org/10.1016/j.enpol.2012.10.046 (2013). 6 Garric, A. in Le Monde (2019). 7 COMMUNICATION FROM THE COMMISSION TO THE EUROPEAN PARLIAMENT, T. E. C., THE

COUNCIL, THE EUROPEAN ECONOMIC AND SOCIAL COMMITTEE, THE COMMITTEE OF THE REGIONS AND THE EUROPEAN INVESTMENT BANK. A Clean Planet for all. A European strategic

long-term vision for a prosperous, modern, competitive and climate neutral economy 2018 ). 8 La COP 21, 2019). 9 Dresselhaus, M. S. & Thomas, I. L. Alternative energy technologies. Nature 414, 332-337,

doi:10.1038/35104599 (2001). 10 Jacobson, M. Z. et al. 100% Clean and Renewable Wind, Water, and Sunlight All-Sector Energy

Roadmaps for 139 Countries of the World. Joule 1, 108-121, doi:https://doi.org/10.1016/j.joule.2017.07.005 (2017).

11 Dunn, S. Hydrogen futures: toward a sustainable energy system. International Journal of

Hydrogen Energy 27, 235-264, doi:https://doi.org/10.1016/S0360-3199(01)00131-8 (2002). 12 Schlapbach, L. & Züttel, A. in Materials for Sustainable Energy 265-270 (2010). 13 Edwards, P. P., Kuznetsov, V. L., David, W. I. F. & Brandon, N. P. Hydrogen and fuel cells:

Towards a sustainable energy future. Energy Policy 36, 4356-4362, doi:https://doi.org/10.1016/j.enpol.2008.09.036 (2008).

14 Momirlan, M. & Veziroglu, T. N. Current status of hydrogen energy. Renewable and

Sustainable Energy Reviews 6, 141-179, doi:https://doi.org/10.1016/S1364-0321(02)00004-7 (2002).

15 Ellis, M. W., Spakovsky, M. R. V. & Nelson, D. J. Fuel cell systems: efficient, flexible energy conversion for the 21st century. Proceedings of the IEEE 89, 1808-1818, doi:10.1109/5.975914 (2001).

16 Dicks, A. L. Hydrogen generation from natural gas for the fuel cell systems of tomorrow. Journal of Power Sources 61, 113-124, doi:https://doi.org/10.1016/S0378-7753(96)02347-6 (1996).

17 Steele, B. C. H. Material science and engineering: The enabling technology for the commercialisation of fuel cell systems. Journal of Materials Science 36, 1053-1068, doi:10.1023/a:1004853019349 (2001).

18 J. Larminie, A. Dicks & McDonald, M. S. Fuel cell system explained. (2003). 19 L. J. M. J. Blomen & Mugerwa., M. N. Fuel Cell Systems. (2013).

20 Haile, S. M. Fuel cell materials and components☆☆☆The Golden Jubilee Issue—Selected

topics in Materials Science and Engineering: Past, Present and Future, edited by S. Suresh. Acta Materialia 51, 5981-6000, doi:https://doi.org/10.1016/j.actamat.2003.08.004 (2003).

35

21 Wang, Y., Chen, K. S., Mishler, J., Cho, S. C. & Adroher, X. C. A review of polymer electrolyte membrane fuel cells: Technology, applications, and needs on fundamental research. Applied

Energy 88, 981-1007, doi:https://doi.org/10.1016/j.apenergy.2010.09.030 (2011). 22 Merle, G., Wessling, M. & Nijmeijer, K. Anion exchange membranes for alkaline fuel cells: A

review. Journal of Membrane Science 377, 1-35, doi:10.1016/j.memsci.2011.04.043 (2011). 23 Dekel, D. R. Review of cell performance in anion exchange membrane fuel cells. Journal of

Power Sources 375, 158-169, doi:10.1016/j.jpowsour.2017.07.117 (2018). 24 Stinson, S. C. Electrolytic cell membrane development surges. Chemical and Engineering

News, 22-25 (1982). 25 Banerjee, S. & Curtin, D. E. Nafion® perfluorinated membranes in fuel cells. Journal of Fluorine

Chemistry 125, 1211-1216, doi:https://doi.org/10.1016/j.jfluchem.2004.05.018 (2004). 26 Wilson, M. S. & Gottesfeld, S. Thin-film catalyst layers for polymer electrolyte fuel cell

electrodes. Journal of Applied Electrochemistry 22, 1-7, doi:10.1007/bf01093004 (1992). 27 Therdthianwong, A., Saenwiset, P. & Therdthianwong, S. Cathode catalyst layer design for

proton exchange membrane fuel cells. Fuel 91, 192-199, doi:https://doi.org/10.1016/j.fuel.2011.07.003 (2012).

28 O’Hayre, R., Barnett, D. M. & Prinz, F. B. The Triple Phase Boundary: A Mathematical Model and Experimental Investigations for Fuel Cells. Journal of The Electrochemical Society 152, A439-A444, doi:10.1149/1.1851054 (2005).

29 Nonoyama, N., Okazaki, S., Weber, A. Z., Ikogi, Y. & Yoshida, T. Analysis of oxygen-transport diffusion resistance in proton-exchange-membrane fuel cells. Journal of The Electrochemical

Society 158, B416-B423 (2011). 30 Paulus, U. A. et al. Oxygen Reduction on Carbon-Supported Pt−Ni and Pt−Co Alloy Catalysts.

The Journal of Physical Chemistry B 106, 4181-4191, doi:10.1021/jp013442l (2002). 31 Stamenkovic, V. R. et al. Trends in electrocatalysis on extended and nanoscale Pt-bimetallic

alloy surfaces. Nature materials 6, 241-247, doi:10.1038/nmat1840 (2007). 32 Stamenkovic, V. R. et al. Improved Oxygen Reduction Activity on Pt₃Ni(111) via

Increased Surface Site Availability. Science 315, 493-497, doi:10.1126/science.1135941 (2007).

33 Mukerjee, S., Srinivasan, S., Soriaga, M. P. & McBreen, J. Effect of Preparation Conditions of Pt Alloys on Their Electronic, Structural, and Electrocatalytic Activities for Oxygen Reduction - XRD, XAS, and Electrochemical Studies. The Journal of Physical Chemistry 99, 4577-4589, doi:10.1021/j100013a032 (1995).

34 Lim, B. et al. Pd-Pt Bimetallic Nanodendrites with High Activity for Oxygen Reduction. Science 324, 1302-1305, doi:10.1126/science.1170377 (2009).

35 Sasikumar, G., Ihm, J. W. & Ryu, H. Optimum Nafion content in PEM fuel cell electrodes. Electrochimica Acta 50, 601-605, doi:https://doi.org/10.1016/j.electacta.2004.01.126 (2004).

36 Ahn, B. K. & Lim, T. W. in 2006 International Forum on Strategic Technology. 199-201. 37 Aso, S., Kizaki, M. & Nonobe, Y. in 2007 Power Conversion Conference - Nagoya. 1606-1611. 38 Thompson, S. T. & Papageorgopoulos, D. Platinum group metal-free catalysts boost cost

competitiveness of fuel cell vehicles. Nature Catalysis 2, 558-561, doi:10.1038/s41929-019-0291-x (2019).

39 Litster, S. & McLean, G. PEM fuel cell electrodes. Journal of Power Sources 130, 61-76, doi:https://doi.org/10.1016/j.jpowsour.2003.12.055 (2004).

40 Gasteiger, H. A., Kocha, S. S., Sompalli, B. & Wagner, F. T. Activity benchmarks and requirements for Pt, Pt-alloy, and non-Pt oxygen reduction catalysts for PEMFCs. Applied

Catalysis B: Environmental 56, 9-35, doi:https://doi.org/10.1016/j.apcatb.2004.06.021 (2005).

41 Wilson, M. S., Garzon, F. H., Sickafus, K. E. & Gottesfeld, S. Surface Area Loss of Supported Platinum in Polymer Electrolyte Fuel Cells. Journal of The Electrochemical Society 140, 2872-2877, doi:10.1149/1.2220925 (1993).

36

42 Zawodzinski, C., Wilson, M. S. & Gottesféld, S. PEM Fuel Cell Stack Development Based on Membrane-Electrode Assemblies of Ultra-Low Platinum Loadings. ECS Proceedings Volumes 1995-23, 57-65, doi:10.1149/199523.0057pv (1995).

43 Wilson, M. S., Valerio, J. A. & Gottesfeld, S. Low platinum loading electrodes for polymer electrolyte fuel cells fabricated using thermoplastic ionomers. Electrochimica Acta 40, 355-363, doi:https://doi.org/10.1016/0013-4686(94)00272-3 (1995).

44 Gasteiger, H. A., Panels, J. E. & Yan, S. G. Dependence of PEM fuel cell performance on catalyst loading. Journal of Power Sources 127, 162-171, doi:https://doi.org/10.1016/j.jpowsour.2003.09.013 (2004).

45 Zeis, R., Mathur, A., Fritz, G., Lee, J. & Erlebacher, J. Platinum-plated nanoporous gold: An efficient, low Pt loading electrocatalyst for PEM fuel cells. Journal of Power Sources 165, 65-72, doi:https://doi.org/10.1016/j.jpowsour.2006.12.007 (2007).

46 Eom, K. et al. Effects of Pt loading in the anode on the durability of a membrane–electrode assembly for polymer electrolyte membrane fuel cells during startup/shutdown cycling. International Journal of Hydrogen Energy 37, 18455-18462, doi:https://doi.org/10.1016/j.ijhydene.2012.09.077 (2012).

47 Shao, Y., Dodelet, J. P., Wu, G. & Zelenay, P. PGM-Free Cathode Catalysts for PEM Fuel Cells: A Mini-Review on Stability Challenges. Advanced materials, e1807615, doi:10.1002/adma.201807615 (2019).

48 Banham, D. et al. A review of the stability and durability of non-precious metal catalysts for the oxygen reduction reaction in proton exchange membrane fuel cells. Journal of Power

Sources 285, 334-348, doi:10.1016/j.jpowsour.2015.03.047 (2015). 49 Chen, Z., Higgins, D., Yu, A., Zhang, L. & Zhang, J. A review on non-precious metal

electrocatalysts for PEM fuel cells. Energy & Environmental Science 4, 3167-3192, doi:10.1039/C0EE00558D (2011).

50 Britton, B. & Holdcroft, S. The Control and Effect of Pore Size Distribution in AEMFC Catalyst Layers. Journal of The Electrochemical Society 163, F353-F358, doi:10.1149/2.0421605jes (2016).

51 Wang, L. et al. An optimised synthesis of high performance radiation-grafted anion-exchange membranes. Green Chemistry 19, 831-843, doi:10.1039/c6gc02526a (2017).

52 Omasta, T. J., Peng, X., Lewis, C. A., Varcoe, J. & Mustain, W. E. Improving Performance in Alkaline Membrane Fuel Cells through Enhanced Water Management. ECS Transactions 75, 949-954, doi:10.1149/07514.0949ecst (2016).

53 Poynton, S. D. et al. Preparation of radiation-grafted powders for use as anion exchange ionomers in alkaline polymer electrolyte fuel cells. Journal of Materials Chemistry A 2, 5124-5130, doi:10.1039/C4TA00558A (2014).

54 Atkins, P. W., De Paula, J. & Keeler, J. Atkins' physical chemistry. (Oxford university press, 2018).

55 Hibbs, M. R. et al. Transport Properties of Hydroxide and Proton Conducting Membranes. Chemistry of Materials 20, 2566-2573, doi:10.1021/cm703263n (2008).

56 Varcoe, J. R. et al. Anion-exchange membranes in electrochemical energy systems. Energy &

Environmental Science 7, 3135-3191, doi:10.1039/C4EE01303D (2014). 57 Rebeck, N. T., Li, Y. & Knauss, D. M. Poly(phenylene oxide) copolymer anion exchange

membranes. Journal of Polymer Science Part B: Polymer Physics 51, 1770-1778, doi:10.1002/polb.23245 (2013).

58 Katzfuß, A., Gogel, V., Jörissen, L. & Kerres, J. The application of covalently cross-linked BrPPO as AEM in alkaline DMFC. Journal of Membrane Science 425-426, 131-140, doi:https://doi.org/10.1016/j.memsci.2012.09.022 (2013).

59 Lin, X., Wu, C., Wu, Y. & Xu, T. Free-standing hybrid anion-exchange membranes for application in fuel cells. Journal of Applied Polymer Science 123, 3644-3651, doi:10.1002/app.35030 (2012).

37

60 Gu, S., Skovgard, J. & Yan, Y. S. Engineering the Van der Waals Interaction in Cross-Linking-Free Hydroxide Exchange Membranes for Low Swelling and High Conductivity. ChemSusChem 5, 843-848, doi:10.1002/cssc.201200057 (2012).

61 Ong, A. L., Saad, S., Lan, R., Goodfellow, R. J. & Tao, S. Anionic membrane and ionomer based on poly(2,6-dimethyl-1,4-phenylene oxide) for alkaline membrane fuel cells. Journal of Power

Sources 196, 8272-8279, doi:https://doi.org/10.1016/j.jpowsour.2011.06.015 (2011). 62 Henkensmeier, D. et al. Anion conducting polymers based on ether linked polybenzimidazole

(PBI-OO). International Journal of Hydrogen Energy 39, 2842-2853, doi:https://doi.org/10.1016/j.ijhydene.2013.07.091 (2014).

63 Lee, H.-J. et al. Synthesis and characterization of poly(benzimidazolium) membranes for anion exchange membrane fuel cells. Polymer Bulletin 70, 2619-2631, doi:10.1007/s00289-013-0978-0 (2013).

64 An, L., Zeng, L. & Zhao, T. S. An alkaline direct ethylene glycol fuel cell with an alkali-doped polybenzimidazole membrane. International Journal of Hydrogen Energy 38, 10602-10606, doi:https://doi.org/10.1016/j.ijhydene.2013.06.042 (2013).

65 Thomas, O. D., Soo, K. J. W. Y., Peckham, T. J., Kulkarni, M. P. & Holdcroft, S. A Stable Hydroxide-Conducting Polymer. Journal of the American Chemical Society 134, 10753-10756, doi:10.1021/ja303067t (2012).

66 Xia, Z. et al. Polybenzimidazoles with pendant quaternary ammonium groups as potential anion exchange membranes for fuel cells. Journal of Membrane Science 390-391, 152-159, doi:https://doi.org/10.1016/j.memsci.2011.11.032 (2012).

67 Park, A. M., Turley, F. E., Wycisk, R. J. & Pintauro, P. N. Electrospun and Cross-Linked Nanofiber Composite Anion Exchange Membranes. Macromolecules 47, 227-235, doi:10.1021/ma401932h (2014).

68 Roddecha, S., Dong, Z., Wu, Y. & Anthamatten, M. Mechanical properties and ionic conductivity of electrospun quaternary ammonium ionomers. Journal of Membrane Science 389, 478-485, doi:https://doi.org/10.1016/j.memsci.2011.11.016 (2012).

69 Zhao, Y. et al. High-performance alkaline fuel cells using crosslinked composite anion exchange membrane. Journal of Power Sources 221, 247-251, doi:https://doi.org/10.1016/j.jpowsour.2012.08.053 (2013).

70 Li, G. et al. Ultrathin composite membrane of alkaline polymer electrolyte for fuel cell applications. Journal of Materials Chemistry A 1, 12497-12502, doi:10.1039/C3TA12626A (2013).

71 Zhao, Y. et al. Quaternary ammonia polysulfone-PTFE composite alkaline anion exchange membrane for fuel cells application. International Journal of Hydrogen Energy 38, 1983-1987, doi:https://doi.org/10.1016/j.ijhydene.2012.11.055 (2013).

72 Wang, L., Peng, X., Mustain, W. E. & Varcoe, J. R. Radiation-grafted anion-exchange membranes: the switch from low- to high-density polyethylene leads to remarkably enhanced fuel cell performance. Energy & Environmental Science 12, 1575-1579, doi:10.1039/c9ee00331b (2019).

73 Omasta, T. J. et al. Beyond catalysis and membranes: visualizing and solving the challenge of electrode water accumulation and flooding in AEMFCs. Energy & Environmental Science 11, 551-558, doi:10.1039/C8EE00122G (2018).

74 Omasta, T. J. et al. Beyond 1.0 W cm−2 Performance without Pla]num: The Beginning of a New Era in Anion Exchange Membrane Fuel Cells. Journal of The Electrochemical Society 165, J3039-J3044, doi:10.1149/2.0071815jes (2018).

75 Peng, X. et al. High-Performing PGM-Free AEMFC Cathodes from Carbon-Supported Cobalt Ferrite Nanoparticles. Catalysts 9, 264, doi:10.3390/catal9030264 (2019).

76 Gonçalves Biancolli, A. L. et al. ETFE-based anion-exchange membrane ionomer powders for alkaline membrane fuel cells: a first performance comparison of head-group chemistry. Journal of Materials Chemistry A 6, 24330-24341, doi:10.1039/c8ta08309f (2018).

38

77 Zhang, M. et al. Highly stable anion exchange membranes based on quaternized polypropylene. Journal of Materials Chemistry A 3, 12284-12296, doi:10.1039/C5TA01420D (2015).

78 Taguet, A., Sauguet, L., Ameduri, B. & Boutevin, B. Fluorinated cotelomers based on vinylidene fluoride (VDF) and hexafluoropropene (HFP): Synthesis, dehydrofluorination and grafting by amine containing an aromatic ring. Journal of Fluorine Chemistry 128, 619-630, doi:https://doi.org/10.1016/j.jfluchem.2007.02.005 (2007).

79 Ameduri, B. From Vinylidene Fluoride (VDF) to the Applications of VDF-Containing Polymers and Copolymers: Recent Developments and Future Trends. Chemical Reviews 109, 6632-6686, doi:10.1021/cr800187m (2009).

80 Dang, H.-S., Weiber, E. A. & Jannasch, P. Poly(phenylene oxide) functionalized with quaternary ammonium groups via flexible alkyl spacers for high-performance anion exchange membranes. Journal of Materials Chemistry A 3, 5280-5284, doi:10.1039/C5TA00350D (2015).

81 Yang, Z. et al. A strategy to construct alkali-stable anion exchange membranes bearing ammonium groups via flexible spacers. Journal of Materials Chemistry A 3, 15015-15019, doi:10.1039/C5TA02941D (2015).

82 Lin, B., Qiu, L., Qiu, B., Peng, Y. & Yan, F. A Soluble and Conductive Polyfluorene Ionomer with Pendant Imidazolium Groups for Alkaline Fuel Cell Applications. Macromolecules 44, 9642-9649, doi:10.1021/ma202159d (2011).

83 Zhuo, Y. Z. et al. Enhancement of hydroxide conductivity by grafting flexible pendant imidazolium groups into poly(arylene ether sulfone) as anion exchange membranes. Journal

of Materials Chemistry A 3, 18105-18114, doi:10.1039/C5TA04257G (2015). 84 Wang, L. et al. Non-fluorinated pre-irradiation-grafted (peroxidated) LDPE-based anion-

exchange membranes with high performance and stability. Energy & Environmental Science 10, 2154-2167, doi:10.1039/C7EE02053H (2017).

85 Kiss, A. M. et al. Carbonate and Bicarbonate Ion Transport in Alkaline Anion Exchange Membranes. Journal of The Electrochemical Society 160, F994-F999, doi:10.1149/2.037309jes (2013).

86 Suzuki, S., Muroyama, H., Matsui, T. & Eguchi, K. Influence of CO2 dissolution into anion exchange membrane on fuel cell performance. Electrochimica Acta 88, 552-558, doi:https://doi.org/10.1016/j.electacta.2012.10.105 (2013).

87 Unlu, M., Zhou, J. & Kohl, P. A. Anion Exchange Membrane Fuel Cells: Experimental Comparison of Hydroxide and Carbonate Conductive Ions. Electrochemical and Solid-State

Letters 12, B27-B30, doi:10.1149/1.3058999 (2009). 88 Li, G. et al. Carbonation effects on the performance of alkaline polymer electrolyte fuel cells.

International Journal of Hydrogen Energy 40, 6655-6660, doi:10.1016/j.ijhydene.2015.03.119 (2015).

89 Krewer, U., Weinzierl, C., Ziv, N. & Dekel, D. R. Impact of carbonation processes in anion exchange membrane fuel cells. Electrochimica Acta 263, 433-446, doi:10.1016/j.electacta.2017.12.093 (2018).

90 Ziv, N., Mustain, W. E. & Dekel, D. R. The Effect of Ambient Carbon Dioxide on Anion-Exchange Membrane Fuel Cells. ChemSusChem 11, 1136-1150, doi:10.1002/cssc.201702330 (2018).

91 Zheng, Y. et al. Quantifying and elucidating the effect of CO2 on the thermodynamics, kinetics and charge transport of AEMFCs. Energy & Environmental Science 12, 2806-2819, doi:10.1039/C9EE01334B (2019).

92 Sheng, W., Gasteiger, H. A. & Shao-Horn, Y. Hydrogen Oxidation and Evolution Reaction Kinetics on Platinum: Acid vs Alkaline Electrolytes. Journal of The Electrochemical Society 157, B1529-B1536, doi:10.1149/1.3483106 (2010).

93 Strmcnik, D. et al. Improving the hydrogen oxidation reaction rate by promotion of hydroxyl adsorption. Nature chemistry 5, 300, doi:10.1038/nchem.1574 https://www.nature.com/articles/nchem.1574#supplementary-information (2013).

39

94 Campos-Roldán, C. A. & Alonso-Vante, N. The Hydrogen Oxidation Reaction in Alkaline Medium: An Overview. Electrochemical Energy Reviews 2, 312-331, doi:10.1007/s41918-019-00034-6 (2019).

95 Davydova, E. S., Mukerjee, S., Jaouen, F. & Dekel, D. R. Electrocatalysts for Hydrogen Oxidation Reaction in Alkaline Electrolytes. ACS Catalysis 8, 6665-6690, doi:10.1021/acscatal.8b00689 (2018).

96 Wang, Y. et al. Pt–Ru catalyzed hydrogen oxidation in alkaline media: oxophilic effect or electronic effect? Energy & Environmental Science 8, 177-181, doi:10.1039/C4EE02564D (2015).

97 Miller, H. A. et al. A Pd/C-CeO2 Anode Catalyst for High-Performance Platinum-Free Anion Exchange Membrane Fuel Cells. Angewandte Chemie International Edition 55, 6004-6007, doi:10.1002/anie.201600647 (2016).

98 Miller, H. A. et al. Highly active nanostructured palladium-ceria electrocatalysts for the hydrogen oxidation reaction in alkaline medium. Nano Energy 33, 293-305, doi:https://doi.org/10.1016/j.nanoen.2017.01.051 (2017).

99 Wang, T. et al. Weakening hydrogen adsorption on nickel via interstitial nitrogen doping promotes bifunctional hydrogen electrocatalysis in alkaline solution. Energy & Environmental

Science, doi:10.1039/C9EE01743G (2019). 100 Lu, S., Pan, J., Huang, A., Zhuang, L. & Lu, J. Alkaline polymer electrolyte fuel cells completely

free from noble metal catalysts. Proceedings of the National Academy of Sciences 105, 20611-20614, doi:10.1073/pnas.0810041106 (2008).

101 Zhong, G., Wang, H., Yu, H. & Peng, F. Nitrogen doped carbon nanotubes with encapsulated ferric carbide as excellent electrocatalyst for oxygen reduction reaction in acid and alkaline media. Journal of Power Sources 286, 495-503, doi:10.1016/j.jpowsour.2015.04.021 (2015).

102 Tan, H. et al. Perfectly ordered mesoporous iron-nitrogen doped carbon as highly efficient catalyst for oxygen reduction reaction in both alkaline and acidic electrolytes. Nano Energy 36, 286-294, doi:10.1016/j.nanoen.2017.04.014 (2017).

103 Sa, Y. J. et al. A General Approach to Preferential Formation of Active Fe-Nx Sites in Fe-N/C Electrocatalysts for Efficient Oxygen Reduction Reaction. Journal of the American Chemical

Society 138, 15046-15056, doi:10.1021/jacs.6b09470 (2016). 104 Ryabova, A. S. et al. Further insights into the role of carbon in manganese oxide/carbon

composites in the oxygen reduction reaction in alkaline media. Electrochimica Acta 246, 643-653, doi:10.1016/j.electacta.2017.06.017 (2017).

105 Li, X., Liu, G. & Popov, B. N. Activity and stability of non-precious metal catalysts for oxygen reduction in acid and alkaline electrolytes. Journal of Power Sources 195, 6373-6378, doi:10.1016/j.jpowsour.2010.04.019 (2010).

106 Kruusenberg, I., Matisen, L., Shah, Q., Kannan, A. M. & Tammeveski, K. Non-platinum cathode catalysts for alkaline membrane fuel cells. International Journal of Hydrogen Energy 37, 4406-4412, doi:10.1016/j.ijhydene.2011.11.143 (2012).

107 Cheng, F., Su, Y., Liang, J., Tao, Z. & Chen, J. MnO2-Based Nanostructures as Catalysts for Electrochemical Oxygen Reduc]on in Alkaline Media†. Chemistry of Materials 22, 898-905, doi:10.1021/cm901698s (2010).

108 Cheng, F., Su, Y., Liang, J., Tao, Z. & Chen, J. MnO2-Based Nanostructures as Catalysts for Electrochemical Oxygen Reduction in Alkaline Media. Chemistry of Materials 22, 898-905, doi:10.1021/cm901698s (2010).

109 Meng, Y. et al. Structure–Property Relationship of Bifunctional MnO2 Nanostructures: Highly Efficient, Ultra-Stable Electrochemical Water Oxidation and Oxygen Reduction Reaction Catalysts Identified in Alkaline Media. Journal of the American Chemical Society 136, 11452-11464, doi:10.1021/ja505186m (2014).

110 Liang, Y. et al. Co3O4 nanocrystals on graphene as a synergistic catalyst for oxygen reduction reaction. Nature materials 10, 780-786, doi:10.1038/nmat3087 (2011).

40

111 Stariha, S. et al. PGM-free Fe-N-C catalysts for oxygen reduction reaction: Catalyst layer design. Journal of Power Sources 326, 43-49, doi:https://doi.org/10.1016/j.jpowsour.2016.06.098 (2016).

112 Sarapuu, A., Kibena-Põldsepp, E., Borghei, M. & Tammeveski, K. Electrocatalysis of oxygen reduction on heteroatom-doped nanocarbons and transition metal–nitrogen–carbon catalysts for alkaline membrane fuel cells. Journal of Materials Chemistry A 6, 776-804, doi:10.1039/C7TA08690C (2018).

113 Sa, Y. J. et al. A General Approach to Preferential Formation of Active Fe–Nx Sites in Fe–N/C Electrocatalysts for Efficient Oxygen Reduction Reaction. Journal of the American Chemical

Society 138, 15046-15056, doi:10.1021/jacs.6b09470 (2016). 114 Serov, A. et al. Highly stable precious metal-free cathode catalyst for fuel cell application.

Journal of Power Sources 327, 557-564, doi:https://doi.org/10.1016/j.jpowsour.2016.07.087 (2016).

115 Kishi, H. et al. Structure of Active Sites of Fe-N-C Nano-Catalysts for Alkaline Exchange Membrane Fuel Cells. Nanomaterials 8, 965 (2018).

116 Hossen, M. M., Artyushkova, K., Atanassov, P. & Serov, A. Synthesis and characterization of high performing Fe-N-C catalyst for oxygen reduction reaction (ORR) in Alkaline Exchange Membrane Fuel Cells. Journal of Power Sources 375, 214-221, doi:10.1016/j.jpowsour.2017.08.036 (2018).

117 Lu, Y. et al. Halloysite-derived nitrogen doped carbon electrocatalysts for anion exchange membrane fuel cells. Journal of Power Sources 372, 82-90 (2017).

118 Ren, H. et al. Fe/N/C nanotubes with atomic Fe sites: A highly active cathode catalyst for alkaline polymer electrolyte fuel cells. Acs Catalysis 7, 6485-6492 (2017).

119 Peng, X. et al. Nitrogen-doped Carbon–CoOx Nanohybrids: A Precious Metal Free Cathode that Exceeds 1.0 W cm−2 Peak Power and 100 h Life in Anion-Exchange Membrane Fuel Cells. Angewandte Chemie International Edition 58, 1046-1051, doi:10.1002/anie.201811099 (2019).

120 Kang, J., Wang, H., Ji, S., Key, J. & Wang, R. Synergy among manganese, nitrogen and carbon to improve the catalytic activity for oxygen reduction reaction. Journal of Power Sources 251, 363-369, doi:https://doi.org/10.1016/j.jpowsour.2013.11.045 (2014).

121 Hou, Y. et al. An Advanced Nitrogen-Doped Graphene/Cobalt-Embedded Porous Carbon Polyhedron Hybrid for Efficient Catalysis of Oxygen Reduction and Water Splitting. Advanced

Functional Materials 25, 872-882, doi:10.1002/adfm.201403657 (2015). 122 Volosskiy, B. et al. Tuning the Catalytic Activity of a Metal–Organic Framework Derived Copper

and Nitrogen Co-Doped Carbon Composite for Oxygen Reduction Reaction. ACS applied

materials & interfaces 8, 26769-26774, doi:10.1021/acsami.6b08320 (2016). 123 Wang, S. et al. Molybdenum-Carbide-Modified Nitrogen-Doped Carbon Vesicle Encapsulating

Nickel Nanoparticles: A Highly Efficient, Low-Cost Catalyst for Hydrogen Evolution Reaction. Journal of the American Chemical Society 137, 15753-15759, doi:10.1021/jacs.5b07924 (2015).

124 Zitolo, A. et al. Identification of catalytic sites for oxygen reduction in iron- and nitrogen-doped graphene materials. Nature materials 14, 937-942, doi:10.1038/nmat4367 (2015).

125 Artyushkova, K. et al. Correlations between Synthesis and Performance of Fe-Based PGM-Free Catalysts in Acidic and Alkaline Media: Evolution of Surface Chemistry and Morphology. ACS

Applied Energy Materials 2, 5406-5418, doi:10.1021/acsaem.9b00331 (2019). 126 Artyushkova, K., Serov, A., Rojas-Carbonell, S. & Atanassov, P. Chemistry of Multitudinous

Active Sites for Oxygen Reduction Reaction in Transition Metal–Nitrogen–Carbon Electrocatalysts. The Journal of Physical Chemistry C 119, 25917-25928, doi:10.1021/acs.jpcc.5b07653 (2015).

41

127 Zhu, Q.-L. et al. Atomically dispersed Fe/N-doped hierarchical carbon architectures derived from a metal–organic framework composite for extremely efficient electrocatalysis. ACS

Energy Letters 2, 504-511 (2017). 128 Meng, H. et al. Iron porphyrin-based cathode catalysts for polymer electrolyte membrane fuel

cells: Effect of NH3 and Ar mixtures as pyrolysis gases on catalytic activity and stability. Electrochimica Acta 55, 6450-6461, doi:10.1016/j.electacta.2010.06.039 (2010).

129 Lee, J. S., Park, G. S., Kim, S. T., Liu, M. & Cho, J. A highly efficient electrocatalyst for the oxygen reduction reaction: N-doped ketjenblack incorporated into Fe/Fe3C-functionalized melamine foam. Angewandte Chemie International Edition 52, 1026-1030 (2013).

130 Wen, Z. et al. Nitrogen-enriched core-shell structured Fe/Fe3C-C nanorods as advanced electrocatalysts for oxygen reduction reaction. Advanced materials 24, 1399-1404 (2012).

131 Ratso, S. et al. Synthesis of highly-active Fe–N–C catalysts for PEMFC with carbide-derived carbons. Journal of Materials Chemistry A 6, 14663-14674, doi:10.1039/c8ta02325e (2018).

132 Gu, L., Jiang, L., Jin, J., Liu, J. & Sun, G. Yolk–shell structured iron carbide/N-doped carbon composite as highly efficient and stable oxygen reduction reaction electrocatalyst. Carbon 82, 572-578, doi:https://doi.org/10.1016/j.carbon.2014.11.010 (2015).

133 Tylus, U. et al. Elucidating Oxygen Reduction Active Sites in Pyrolyzed Metal-Nitrogen Coordinated Non-Precious-Metal Electrocatalyst Systems. The journal of physical chemistry.

C, Nanomaterials and interfaces 118, 8999-9008, doi:10.1021/jp500781v (2014). 134 Yang, L. et al. Activity, Performance, and Durability for the Reduction of Oxygen in PEM Fuel

Cells, of Fe/N/C Electrocatalysts Obtained from the Pyrolysis of Metal-Organic-Framework and Iron Porphyrin Precursors. Electrochimica Acta 159, 184-197, doi:10.1016/j.electacta.2015.01.201 (2015).

135 Lefèvre, M., Proietti, E., Jaouen, F. & Dodelet, J.-P. Iron-Based Catalysts with Improved Oxygen Reduction Activity in Polymer Electrolyte Fuel Cells. Science 324, 71-74, doi:10.1126/science.1170051 (2009).

136 Jasinski, R. A New Fuel Cell Cathode Catalyst. Nature 201, 1212-1213, doi:10.1038/2011212a0 (1964).

137 Chen, M., He, Y., Spendelow, J. S. & Wu, G. Atomically Dispersed Metal Catalysts for Oxygen Reduction. ACS Energy Letters 4, 1619-1633, doi:10.1021/acsenergylett.9b00804 (2019).

138 Jahnke, H., Schönborn, M. & Zimmermann, G. Organic dyestuffs as catalysts for fuel cells. Topics in Current Chemistry 61, 133-180, doi:https://doi.org/10.1007/BFb0046059 (1976).

139 Bagotzky, V. S., Tarasevich, M. R., Radyushkina, K. A., Levina, O. A. & Andrusyova, S. I. Electrocatalysis of the oxygen reduction process on metal chelates in acid electrolyte. Journal

of Power Sources 2, 233-240, doi:https://doi.org/10.1016/0378-7753(78)85014-9 (1978). 140 van Veen, J. A. R. & Visser, C. Oxygen reduction on monomeric transition metal

phthalocyanines in acid electrolyte. Electrochimica Acta 24, 921-928, doi:https://doi.org/10.1016/0013-4686(79)87088-7 (1981).

141 van Veen, J. A. R., Colijn, H. A. & van Baar, J. F. On the effect of a heat treatment on the structure of carbon-supported metalloporphyrins and phthalocyanines. Electrochimica Acta 33, 801-804, doi:https://doi.org/10.1016/S0013-4686(98)80010-8 (1988).

142 van Wingerden, B., van Veen, J. A. R. & Mensch, C. T. J. An extended X-ray absorption fine structure study of heat-treated cobalt porphyrin catalysts supported on active carbon. Journal

of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases 84, 65-74, doi:10.1039/F19888400065 (1988).

143 Van Der Putten, A., Elzing, A., Visscher, W. & Barendrecht, E. Oxygen reduction on pyrolysed carbon-supported transition metal chelates. Journal of Electroanalytical Chemistry and

Interfacial Electrochemistry 205, 233-244, doi:https://doi.org/10.1016/0022-0728(86)90234-2 (1986).

42

144 Scherson, D. et al. Transition metal macrocycles supported on high area carbon: Pyrolysis—mass spectrometry studies. Electrochimica Acta 31, 1247-1258, doi:https://doi.org/10.1016/0013-4686(86)80144-X (1986).

145 van Veen, J. A. R., van Baar, J. F. & Kroese, K. J. Effect of heat treatment on the performance of carbon-supported transition-metal chelates in the electrochemical reduction of oxygen. Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed

Phases 77, 2827-2843, doi:10.1039/F19817702827 (1981). 146 Gupta, S., Tryk, D., Bae, I., Aldred, W. & Yeager, E. Heat-treated polyacrylonitrile-based

catalysts for oxygen electroreduction. Journal of Applied Electrochemistry 19, 19-27, doi:10.1007/bf01039385 (1989).

147 Lefèvre, M., Dodelet, J. P. & Bertrand, P. O2 Reduction in PEM Fuel Cells: Activity and Active Site Structural Information for Catalysts Obtained by the Pyrolysis at High Temperature of Fe Precursors. The Journal of Physical Chemistry B 104, 11238-11247, doi:10.1021/jp002444n (2000).

148 Bouwkamp-Wijnoltz, A. L. et al. On Active-Site Heterogeneity in Pyrolyzed Carbon-Supported Iron Porphyrin Catalysts for the Electrochemical Reduction of Oxygen:  An In Situ Mössbauer Study. The Journal of Physical Chemistry B 106, 12993-13001, doi:10.1021/jp0266087 (2002).

149 Lefèvre, M., Dodelet, J. P. & Bertrand, P. Molecular Oxygen Reduction in PEM Fuel Cells:  Evidence for the Simultaneous Presence of Two Active Sites in Fe-Based Catalysts. The Journal

of Physical Chemistry B 106, 8705-8713, doi:10.1021/jp020267f (2002). 150 Lefèvre, M. & Dodelet, J.-P. Fe-based catalysts for the reduction of oxygen in polymer

electrolyte membrane fuel cell conditions: determination of the amount of peroxide released during electroreduction and its influence on the stability of the catalysts. Electrochimica Acta 48, 2749-2760, doi:10.1016/s0013-4686(03)00393-1 (2003).

151 Jaouen, F., Lefèvre, M., Dodelet, J.-P. & Cai, M. Heat-Treated Fe/N/C Catalysts for O2 Electroreduction:  Are Active Sites Hosted in Micropores? The Journal of Physical Chemistry B 110, 5553-5558, doi:10.1021/jp057135h (2006).

152 Jaouen, F., Charreteur, F. & Dodelet, J. P. Fe-Based Catalysts for Oxygen Reduction in PEMFCs: Importance of the Disordered Phase of the Carbon Support. Journal of The Electrochemical

Society 153, A689-A698, doi:10.1149/1.2168418 (2006). 153 Charreteur, F., Jaouen, F., Ruggeri, S. & Dodelet, J.-P. Fe/N/C non-precious catalysts for PEM

fuel cells: Influence of the structural parameters of pristine commercial carbon blacks on their activity for oxygen reduction. Electrochimica Acta 53, 2925-2938, doi:10.1016/j.electacta.2007.11.002 (2008).

154 Kramm, U. I. et al. Structure of the catalytic sites in Fe/N/C-catalysts for O2-reduction in PEM fuel cells. Physical chemistry chemical physics : PCCP 14, 11673-11688, doi:10.1039/c2cp41957b (2012).

155 Hoskins, B. F. & Robson, R. Infinite polymeric frameworks consisting of three dimensionally linked rod-like segments. Journal of the American Chemical Society 111, 5962-5964, doi:10.1021/ja00197a079 (1989).

156 Hoskins, B. F. & Robson, R. Design and construction of a new class of scaffolding-like materials comprising infinite polymeric frameworks of 3D-linked molecular rods. A reappraisal of the zinc cyanide and cadmium cyanide structures and the synthesis and structure of the diamond-related frameworks [N(CH3)4][CuIZnII(CN)4] and CuI[4,4',4'',4'''-tetracyanotetraphenylmethane]BF4.xC6H5NO2. Journal of the American Chemical Society 112, 1546-1554, doi:10.1021/ja00160a038 (1990).

157 Abrahams, B. F., Hoskins, B. F., Michail, D. M. & Robson, R. Assembly of porphyrin building blocks into network structures with large channels. Nature 369, 727-729, doi:10.1038/369727a0 (1994).

158 Yaghi, O. M., Li, G. & Li, H. Selective binding and removal of guests in a microporous metal–organic framework. Nature 378, 703-706, doi:10.1038/378703a0 (1995).

43

159 Yaghi, O. M. & Li, H. Hydrothermal Synthesis of a Metal-Organic Framework Containing Large Rectangular Channels. Journal of the American Chemical Society 117, 10401-10402, doi:10.1021/ja00146a033 (1995).

160 Goenaga, G., Ma, S., Yuan, S. & Liu, D.-J. New Approaches to Non-PGM Electrocatalysts Using Porous Framework Materials. ECS Transactions 33, 579-586, doi:10.1149/1.3484554 (2010).

161 Ma, S., Goenaga, G. A., Call, A. V. & Liu, D.-J. Cobalt Imidazolate Framework as Precursor for Oxygen Reduction Reaction Electrocatalysts. Chemistry – A European Journal 17, 2063-2067, doi:10.1002/chem.201003080 (2011).

162 Liu, B. et al. Converting cobalt oxide subunits in cobalt metal-organic framework into agglomerated Co3O4 nanoparticles as an electrode material for lithium ion battery. Journal of

Power Sources 195, 857-861, doi:https://doi.org/10.1016/j.jpowsour.2009.08.058 (2010). 163 Proietti, E. et al. Iron-based cathode catalyst with enhanced power density in polymer

electrolyte membrane fuel cells. Nature communications 2, 416, doi:10.1038/ncomms1427 (2011).

164 Herranz, J. et al. Unveiling N-protonation and anion-binding effects on Fe/N/C-catalysts for O2 reduction in PEM fuel cells. The journal of physical chemistry. C, Nanomaterials and interfaces 115, doi:10.1021/jp2042526 (2011).

165 Ramaswamy, N., Tylus, U., Jia, Q. & Mukerjee, S. Activity descriptor identification for oxygen reduction on nonprecious electrocatalysts: linking surface science to coordination chemistry. Journal of the American Chemical Society 135, 15443-15449, doi:10.1021/ja405149m (2013).

166 Goellner, V. et al. Degradation of Fe/N/C catalysts upon high polarization in acid medium. Physical chemistry chemical physics : PCCP 16, 18454-18462, doi:10.1039/c4cp02882a (2014).

167 Choi, C. H. et al. Unraveling the Nature of Sites Active toward Hydrogen Peroxide Reduction in Fe-N-C Catalysts. Angewandte Chemie 56, 8809-8812, doi:10.1002/anie.201704356 (2017).

168 Nidheesh, P. V. Heterogeneous Fenton catalysts for the abatement of organic pollutants from aqueous solution: a review. RSC Advances 5, 40552-40577, doi:10.1039/c5ra02023a (2015).

169 Kumar, K. et al. Physical and Chemical Considerations for Improving Catalytic Activity and Stability of Non-Precious-Metal Oxygen Reduction Reaction Catalysts. ACS Catalysis 8, 11264-11276, doi:10.1021/acscatal.8b02934 (2018).

170 Choi, C. H. et al. Stability of Fe-N-C Catalysts in Acidic Medium Studied by Operando Spectroscopy. Angewandte Chemie 54, 12753-12757, doi:10.1002/anie.201504903 (2015).

171 Wiesener, K. N4-chelates as electrocatalyst for cathodic oxygen reduction. Electrochimica

Acta 31, 1073-1078 (1986). 172 Bron, M., Fiechter, S., Bogdanoff, P. & Tributsch, H. Thermogravimetry/Mass Spectrometry

Investigations on the Formation of Oxygen Reduction Catalysts for PEM Fuel Cells on the Basis of Heat-Treated Iron Phenanthroline Complexes. Fuel Cells 2, 137-142, doi:10.1002/fuce.200290012 (2002).

173 Gubler, L., Dockheer, S. M. & Koppenol, W. H. Radical (HO•, H• and HOO•) Formation and Ionomer Degradation in Polymer Electrolyte Fuel Cells. Journal of The Electrochemical Society 158, B755-B769, doi:10.1149/1.3581040 (2011).

174 Coms, F. D. The Chemistry of Fuel Cell Membrane Chemical Degradation. ECS Transactions 16, 235-255, doi:10.1149/1.2981859 (2008).

175 Ramaswamy, N. & Mukerjee, S. Influence of Inner- and Outer-Sphere Electron Transfer Mechanisms during Electrocatalysis of Oxygen Reduction in Alkaline Media. The Journal of

Physical Chemistry C 115, 18015-18026, doi:10.1021/jp204680p (2011). 176 Jiang, K. et al. Highly selective oxygen reduction to hydrogen peroxide on transition metal

single atom coordination. Nature communications 10, 3997, doi:10.1038/s41467-019-11992-2 (2019).

177 Parrondo, J., Wang, Z., Jung, M. S. & Ramani, V. Reactive oxygen species accelerate degradation of anion exchange membranes based on polyphenylene oxide in alkaline

44

environments. Physical chemistry chemical physics : PCCP 18, 19705-19712, doi:10.1039/c6cp01978a (2016).

178 Zhang, C., Fan, F.-R. F. & Bard, A. J. Electrochemistry of Oxygen in Concentrated NaOH Solutions: Solubility, Diffusion Coefficients, and Superoxide Formation. Journal of the

American Chemical Society 131, 177-181, doi:10.1021/ja8064254 (2009). 179 Song, Y., Zhang, X., Cui, X. & Shi, J. The ORR kinetics of ZIF-derived FeNC electrocatalysts.

Journal of Catalysis 372, 174-181, doi:https://doi.org/10.1016/j.jcat.2019.02.023 (2019). 180 Jaouen, F. & Dodelet, J.-P. O2 Reduction Mechanism on Non-Noble Metal Catalysts for PEM

Fuel Cells. Part I: Experimental Rates of O2 Electroreduction, H2O2 Electroreduction, and H2O2 Disproportionation. The Journal of Physical Chemistry C 113, 15422-15432, doi:10.1021/jp900837e (2009).

181 Zhang, W., Wang, H., Yang, Z. & Wang, F. Promotion of H2O2 decomposition activity over β-MnO2 nanorod catalysts. Colloids and Surfaces A: Physicochemical and Engineering Aspects 304, 60-66, doi:10.1016/j.colsurfa.2007.04.022 (2007).

182 Ghodbane, O., Pascal, J. L. & Favier, F. Microstructural effects on charge-storage properties in MnO2-based electrochemical supercapacitors. ACS applied materials & interfaces 1, 1130-1139, doi:10.1021/am900094e (2009).

183 Cao, X. et al. A novel non-enzymatic hydrogen peroxide biosensor based on ultralong manganite MnOOH nanowires. Sensors and Actuators B: Chemical 147, 730-734, doi:10.1016/j.snb.2010.03.087 (2010).

184 Cheng, F. et al. Rapid room-temperature synthesis of nanocrystalline spinels as oxygen reduction and evolution electrocatalysts. Nature chemistry 3, 79-84, doi:10.1038/nchem.931 (2011).

185 Gorlin, Y. et al. In situ X-ray absorption spectroscopy investigation of a bifunctional manganese oxide catalyst with high activity for electrochemical water oxidation and oxygen reduction. Journal of the American Chemical Society 135, 8525-8534, doi:10.1021/ja3104632 (2013).

186 Meng, Y. et al. Structure-property relationship of bifunctional MnO2 nanostructures: highly efficient, ultra-stable electrochemical water oxidation and oxygen reduction reaction catalysts identified in alkaline media. Journal of the American Chemical Society 136, 11452-11464, doi:10.1021/ja505186m (2014).

187 Selvakumar, K. et al. Physiochemical Investigation of Shape-Designed MnO2 Nanostructures and Their Influence on Oxygen Reduction Reaction Activity in Alkaline Solution. The Journal of

Physical Chemistry C 119, 6604-6618, doi:10.1021/jp5127915 (2015). 188 Kéranguéven, G., Royer, S. & Savinova, E. Synthesis of efficient Vulcan–LaMnO3 perovskite

nanocomposite for the oxygen reduction reaction. Electrochemistry Communications 50, 28-31, doi:10.1016/j.elecom.2014.10.019 (2015).

189 Lehtimäki, M. et al. Targeted design of α-MnO2 based catalysts for oxygen reduction. Electrochimica Acta 191, 452-461, doi:https://doi.org/10.1016/j.electacta.2016.01.070 (2016).

190 Arunchander, A., Vivekanantha, M., Peera, S. G. & Sahu, A. K. MnO–nitrogen doped graphene as a durable non-precious hybrid catalyst for the oxygen reduction reaction in anion exchange membrane fuel cells. RSC Advances 6, 95590-95600, doi:10.1039/c6ra20627a (2016).

191 Zuo, L. X., Jiang, L. P., Abdel-Halim, E. S. & Zhu, J. J. Sonochemical preparation of stable porous MnO2 and its application as an efficient electrocatalyst for oxygen reduction reaction. Ultrasonics sonochemistry 35, 219-225, doi:10.1016/j.ultsonch.2016.09.021 (2017).

192 Lambert, T. N. et al. Understanding the Effects of Cationic Dopants on α-MnO2 Oxygen Reduction Reaction Electrocatalysis. The Journal of Physical Chemistry C 121, 2789-2797, doi:10.1021/acs.jpcc.6b11252 (2017).

193 Ryabova, A. S. et al. Rationalizing the Influence of the Mn(IV)/Mn(III) Red-Ox Transition on the Electrocatalytic Activity of Manganese Oxides in the Oxygen Reduction Reaction. Electrochimica Acta 187, 161-172, doi:10.1016/j.electacta.2015.11.012 (2016).

45

194 Poux, T. et al. Electrocatalysis of hydrogen peroxide reactions on perovskite oxides: experiment versus kinetic modeling. Physical chemistry chemical physics : PCCP 16, 13595-13600, doi:10.1039/c4cp00341a (2014).

195 Ryabova, A. S. et al. Study of Hydrogen Peroxide Reactions on Manganese Oxides as a Tool To Decode the Oxygen Reduction Reaction Mechanism. ChemElectroChem 3, 1667-1677, doi:10.1002/celc.201600236 (2016).

196 Bonnefont, A. et al. Challenges in the understanding oxygen reduction electrocatalysis on transition metal oxides. Current Opinion in Electrochemistry 14, 23-31, doi:10.1016/j.coelec.2018.09.010 (2019).

46

Effect of Pyrolysis Atmosphere and Electrolyte pH on the

Oxygen Reduction Activity, Stability and Spectroscopic

Signature of FeNx Moieties in Fe-N-C Catalysts

Pietro Giovanni Santori,1 Florian Speck,2 Jingkun Li,1 Andrea Zitolo,3 Qingying Jia,4 Sanjeev

Mukerjee,4 Serhiy Cherevko,2 * Frédéric Jaouen1 *

1. Institut Charles Gerhardt Montpellier, UMR 5253, CNRS, Université Montpellier,

ENSCM, Place Eugène Bataillon, 34095 Montpellier cedex 5, France

2. Helmholtz-Institute Erlangen-Nürnberg for Renewable Energy (IEK-11),

Forschungszentrum Jülich, Egerlandstr. 3, 91058 Erlangen, Germany

3. Synchrotron SOLEIL, L’Orme des Merisiers, BP 48 Saint Aubin, 91192, Gif-sur-Yvette,

France

4. Department of Chemistry and Chemical Biology, Northeastern University, 364 Egan

Research Center, 360 Huntington avenue, Boston, MA 02115, USA

link: http://jes.ecsdl.org/content/166/7/F3311.short

47

Abstract

Two Fe-N-C catalysts comprising only atomically-dispersed FeNx moieties were prepared,

differing only in the fact that the second catalyst (Fe0.5-NH3) was obtained by subjecting the

first one (Fe0.5-Ar) to a short pyrolysis in ammonia. While the initial ORR activity in acid

medium in rotating disk electrode is similar for both catalysts, the activity in alkaline medium

is significantly higher for Fe0.5-NH3. Time-resolved Fe dissolution reveals a circa 10 times

enhanced Fe leaching rate in acidic electrolyte for Fe0.5-NH3 relative to Fe0.5-Ar. Furthermore,

for the former, the leaching rate is strongly enhanced when the electrochemical potential is

in the range 0.75-0.3 V vs. RHE. This may explain the reduced stability of ammonia-pyrolyzed

Fe-N-C catalysts in operating PEMFCs. In alkaline medium in contrast, Fe0.5-NH3 is more active

and more stable, with minimized Fe leaching during electrochemical operation in load-cycling

mode. Operando X-ray absorption spectroscopy measurements in alkaline electrolyte reveals

similar trends of the XANES and EXAFS spectra as a function of the electrochemical potential

for both catalysts, but the magnitude of change is much less for Fe0.5-NH3, as evidenced by a

Δµ analysis. This is interpreted as a lower average oxidation state of FeNx moieties in Fe0.5-

NH3 at open circuit potential.

Introduction

Due to growing concerns related to local and global impacts on the Earth’s atmosphere and

long-term sustainability of fossil fuels, new ways of producing and using renewable fuels are

being explored for both transportation and stationary applications. Hydrogen is a promising

fuel that can be produced from water and renewable energy via electrolysis or other means,

and back-converted to electric power on demand in high-efficiency fuel cells rather than in

low-efficiency combustion engines.1 While different types of fuel cell technologies exist, the

proton exchange membrane fuel cell (PEMFC) has key advantages such as low internal

resistance, fast start-up/shutdown and high electric power density2–5. This makes it suitable

for small and medium size applications ranging from portable devices to automotive

application and distributed electric power production. While high power densities at

reasonable energy efficiency can be reached with PEMFC, several drawbacks hold its large-

scale commercialization. In particular, the sluggish oxygen reduction reaction (ORR) at the

cathode requires the most active catalysts to overcome the kinetic barrier of splitting the

48

double-bond in O2. To date, all catalysts that meet the ORR activity and stability targets in the

acidic PEMFC environment are based on platinum.4,5 The low natural abundance in the Earth’s

crust and geographically-constrained resources of this element raises concerns of cost

competitiveness and long-term sustainability of the present PEMFC technology6.

Research on non-noble ORR catalysts for PEMFC started ca 50 years ago, and significantly

intensified in the past decade7–10. Among catalysts free of platinum group metals (PGMs),

metal-nitrogen-carbon catalysts have hitherto shown the highest initial activity toward the

ORR in acidic media, especially Fe-N-C catalysts7–12. Compared to Pt-based catalysts, the ORR

activity of pyrolyzed Fe-N-C materials is highly dependent on the synthetic path, due to

numerous Fe species that may form at high temperature, ranging from metallic, metal-

carbide and metal-nitride Fe particles (often embedded in carbon) to various FeNx moieties

featuring single iron-atoms covalently integrated in the N-doped carbon matrix.13–16 Among

the preparation methods, the sacrificial metal-organic-framework method has resulted in Fe-

N-C catalysts with state-of-the-art activity and power performance in a rotating disk electrode

(RDE) and PEMFC, respectively17–19. Characterization by 57Fe Mössbauer and X-ray Absorption

Spectroscopy (XAS) has revealed that the active sites in Fe-N-C catalysts featuring exclusively

atomically-dispersed iron have an FeN4 structure that is similar to the iron-porphyrin

core.18,20,21 While Fe-N-C catalysts featuring exclusively Fe-carbide particles embedded in N-

doped carbon matrix can also show interesting ORR activity at beginning-of-life (BoL), their

stability during load-cycling in acidic medium was recently shown to be insufficient.22–24 In

contrast, an Fe-N-C catalyst prepared via ramp pyrolysis in inert atmosphere and exclusively

comprising FeN4 moieties showed only 25% loss in activity after 30,000 load cycles at 80°C in

inert-gas saturated acidic electrolyte, well above the U.S. Department of Energy target of 40%

maximum loss in activity after such an accelerated stress test (AST).24 While such AST protocol

cannot capture all degradation mechanisms that will occur in a PEMFC, it is useful in

identifying which PGM-free catalysts meet the necessary (but not sufficient) criterion of

structural stability when cycled in the cathode potential range and pH environment expected

in PEMFC. In particular it is useful to detect demetallation that may occur even in the absence

of O2.

49

If the ORR activity and stability in acidic medium in RDE setup (stability to load-cycling AST in

inert-gas saturated electrolyte) of at least some atomically-dispersed Fe-N-C catalysts is

unquestionable, their stability in operating PEMFC has hitherto still been low.7,9,17,18,25 This is

particularly true for Fe-N-C catalysts prepared via pyrolysis in flowing NH3, resulting in circa

20-30 times higher BoL activity at 0.9 V in PEMFC than similarly prepared Fe-N-C catalysts but

pyrolyzed in inert atmosphere.17,18 After 15-20 h of operation in PEMFC, the current density

at 0.5 V obtained with NH3-pyrolyzed Fe-N-C is however typically halved, while that with Ar-

pyrolyzed Fe-N-C undergoes less than 10% decay.18,26 Advanced ex situ spectroscopic

comparison between an Ar- and a NH3-pyrolyzed Fe-N-C catalyst revealed no or small

difference of Fe coordination, both featuring FeN4 sites with nearly identical XAS and

Mössbauer spectroscopy fingerprints.18 The only key differences are i) the higher basicity of

the catalyst’s surface and ii) higher micropore volume for the NH3-pyrolyzed material. A

previous study on another NH3-pyrolyzed Fe-N-C catalyst revealed that a short immersion in

acidic medium (sulphuric acid or perchloric acid) decreased its activity by a factor ten

(activity/10), and that this deactivation was partially reversible after the acid-washed catalyst

had been subjected to a cleaning treatment at 400°C in inert gas (activity/10 x5).27 The

irreversible activity loss (activity/2) was assigned to metal leaching from weakly bound Fe

sites (ca 50% loss of metal upon first acid wash). Since that pristine catalyst contained not

only FeNx moieties but also a significant content of crystalline Fe particles, the metal leached

during first acid-wash may have however originated (at least partially) from metallic Fe

particles. The reversible deactivation phenomenon (between activity/10 to activity/2) was

explained as a protonation of highly basic N-groups (leading to a high turn-over frequency,

TOF, of FeNx moieties) and their charge-neutralization by the electrolyte’s counter-anion.27

The latter anion-adsorption event was proposed to decrease the FeNx site’s TOF. Strong

support for this reversible deactivation mechanism, not related to any Fe-leaching event, was

given by the possibility to recover most of the initial activity of the ammonia-treated Fe-N-C

catalyst by removing the adsorbed bisulfate anions (as proven by X-ray photoelectron

spectroscopy) during the 400°C treatment. The ORR activity increased from ca only 10% of

the initial activity after a short acid wash to ca 50% of the initial activity after removal of the

bisulfate anions.27

50

More recently, the reason for the reduced stability of NH3-pyrolyzed Fe-N-C catalysts in

PEMFC was re-investigated and debated in three follow-up papers.25,26,28 Dodelet's group first

proposed that the instability of NH3-pyrolyzed Fe-N-C catalysts in PEMFC is due to oxidation

of the carbon surface in micropores, increasing the hydrophilicity and leading to micropore

flooding.25 Iron was first claimed to play no role in the deactivation, although Fe coordination

and demetallation during operation or post mortem had not been characterized. Choi et al

then focused on this micropore-flooding hypothesis28 and could demonstrate that this

mechanism could not explain the rapid loss of performance in PEMFC observed with another

NH3-pyrolyzed Fe-N-C catalyst. From polarization curves recorded at different relative

humidity and cyclic voltammograms before/after stability test, they concluded that the

micropores were partially or completely filled by water already at BoL, and that the rapid

performance loss was mainly due to a decrease of the ORR kinetics, and not due to a decay

of the mass-transport performance of the cathode layer. While clearly ruling out the

micropore flooding hypothesis, their work could not conclusively point what is the mechanism

for the rapid decay of the ORR kinetics of NH3-pyrolyzed Fe-N-C catalysts. Following that work,

Dodelet’s group agreed that micropore flooding was not the cause for rapid ORR activity

decay of their NH3-pyrolyzed Fe-N-C catalyst as well and, after examination of the Fe content

and speciation with Mössbauer spectroscopy following various short-duration PEMFC tests

(in the range of 0-20 h), could establish similar trends between a) the current density at 0.6 V

vs. duration of operation and b) the relative fraction of Fe present as FeNx moieties in the

cathode as a function of duration of operation in PEMFC, both experiencing a relative decay

of ca 50% after 20 h operation.26 The novel hypothesis was that the fast activity decay of NH3-

pyrolyzed Fe-N-C catalyst is due to a demetallation of unstable FeNx moieties in this material,

such moieties being specifically located in micropores. Despite the above-mentioned trend,

it must be noted however that the current density at 0.6 V in PEMFC is inappropriate to track

changes in the ORR activity, and that the relative decay in ORR activity at 0.8 V after 20 h

operation was much higher than 50%, in fact ca 90% (Fig. 4 in26). The demetallation from

specific FeNx moieties existing in highly microporous Fe-N-C materials may thus explain part

of the initial ORR activity decay (e.g. from 1.0 to 0.5 in normalized activity, 1.0 corresponding

to BoL activity), but not necessarily all of the decay. It is also possible that the loss of 50% of

FeNx moieties in fact implied that ca 90% of the electrochemically-accessible FeNx moieties

51

had been leached. Since Mössbauer spectroscopy is a bulk technique, the fact that ca 50% of

FeNx moieties were still present after PEMFC operation does not necessarily imply that those

moieties did contribute to the initial ORR activity, if they were not located on the top surface

of the carbon matrix.

With no clear path at the moment on how highly-basic N-groups could be stabilized in acidic

medium, our scientific interest was first oriented on the degradation mechanisms of inert-

gas-pyrolyzed Fe-N-C catalysts with FeNx moieties as the main active sites. While significantly

more durable than NH3-pyrolyzed catalysts, they also suffer from a slow but steady linear

decline with duration of operation in PEMFC.26,29 Two main degradation mechanisms of such

materials have been identified: (1) irreversible iron leaching from Fe particles imperfectly

embedded in carbon or from FeN4 active sites30; and (2) reversible degradation induced by

the hydrogen peroxide by-product formed during ORR in acidic medium;31,32 Mechanism (1)

was investigated with online mass-spectrometry and revealed that Fe leaching resulted from

carbon corrosion during startup/shutdown AST, while Fe leaching during load-cycling AST

mainly originated from Fe particles imperfectly surrounded by a N-C layer. The FeN4 moieties

were mostly stable during load-cycling AST in inert-gas saturated acidic medium, in a broad

range of temperature and up to 10,000 cycles.24,30,33 The high stability in acidic medium during

load-cycling of FeN4 moieties stood in apparent paradox to their poor durability in operating

PEMFC. This conundrum was recently clarified by a study showing that the surface

modification of Fe-N-C by hydrogen peroxide does not leach FeN4 moieties but decreases their

TOF.34 This deactivation originates from the chemical reaction of minute amounts of H2O2

with the FeN4 moieties, leading to the formation of reactive oxygen species. These radicals

then react with the carbon surface, introducing a high number of oxygen functionalities on

the top-surface. From a combined experimental and theoretical investigation, we showed

that this decreases i) the electron density at the Fe centre and ii) the O2 binding energy of

FeN4 moieties, resulting in a much decreased single-site TOF.34 Interestingly, the controlled

ex situ deactivation of Fe-N-C by hydrogen peroxide revealed that this mechanism is highly

pH-dependent, the same protocol but applied in 0.1 M solution of KOH instead of HClO4

resulting in no deactivation of an Ar-pyrolyzed Fe-N-C catalyst.34 This makes it promising for

the application of Fe-N-C catalysts based on FeN4 moieties in anion exchange membrane fuel

52

cell (AEMFC).35–37 In particular, the AEMFC environment might allow combining the highest

ORR activity of NH3-pyrolyzed Fe-N-C catalysts with high durability. Expectation for high

durability is supported by i) the stability at high pH of highly-basic N-groups present in NH3-

pyrolyzed catalysts and ii) the lack of peroxide-induced deactivation at high pH on a specific

Ar-pyrolyzed Fe-N-C catalyst comprising exclusively FeN4 sites.

The aims of this study are to assess the site-structure, activity and stability during load-cycling

in alkaline electrolyte of a NH3-pyrolyzed catalyst showing an extremely high ORR activity and

exclusively comprising Fe as atomically-dispersed FeNx moieties. A second Fe-N-C catalyst

prepared similarly but obtained via a single pyrolysis in inert gas is also studied, allowing the

identification of the effect of pyrolysis atmosphere on the properties of Fe-N-C catalysts.

Activity and stability during load-cycling AST was also performed in an acidic electrolyte to

assess the effect of electrolyte pH. Activity and stability are measured with rotating disk

electrode. Fe coordination and Fe leaching were monitored as a function of the

electrochemical potential with operando XAS and a flow cell coupled to online mass-

spectrometry, respectively, in order to shed light on activity and stability properties.

This study demonstrates that FeNx moieties in a NH3-pyrolyzed Fe-N-C catalyst, while being

structurally very similar to those present in the Ar-pyrolyzed Fe-N-C material, show

exacerbated demetallation in acidic medium. Online measurement of Fe dissolution rate

shows the Fe demetallation from NH3-pyrolyzed Fe-N-C increases when scanning negatively

the potential from 1.0 V vs. RHE, to reach a peak of dissolution rate at ca 0.3 V vs. RHE. In

alkaline electrolyte, the metal-Nx moieties' stability of Ar- and NH3-pyrolyzed Fe-N-C are

comparable, and the very highly ORR-active NH3-pyrolyzed Fe-N-C material is stable in 0.1 M

KOH for several thousands of cycles. Operando XAS of the NH3-pyrolyzed Fe-N-C catalyst in

alkaline electrolyte shows similar trends as for the Ar-pyrolyzed Fe-N-C catalyst, but with a

reduced magnitude of the changes with electrochemical potential. This is interpreted as a

lower average oxidation state of the FeNx moieties at open circuit potential in the NH3-

pyrolyzed Fe-N-C material compared to the Ar-pyrolyzed Fe-N-C.

53

Experimental

Synthesis

Two Fe-N-C catalysts are prepared with the sacrificial metal-organic-framework method,

using ZIF-8 (Basolite® Z1200, Sigma Aldrich) as support, 1,10-phenantroline (≥99%, Sigma

Aldrich) as a secondary source of nitrogen and iron (II) acetate (≥99.99%, Sigma Aldrich) as

the source of metal. The catalyst precursor is prepared with a weight ratio of 4/1 for ZIF-

8/phenanthroline and 0.5 wt% of iron in the complete catalyst precursor. The three

precursors are initially mixed using low-energy ball milling at 400 rpm for 2 h 20 min, with 5

minutes pause every 30 minutes of milling. The obtained catalyst precursor is transferred into

a quartz boat and inserted in a quartz tube. The first pyrolysis is performed in flash-pyrolysis

mode, pre-equilibrating the quartz tube and oven at 1050 °C, then pushing the quartz boat

and catalyst precursor within 1 min in the heating zone of the furnace with an outer magnet.

The pyrolysis duration at 1050 °C in flowing Ar is exactly 1 h. The pyrolysis is terminated by

opening the split-hinge furnace, removing the quartz tube and letting it cool down at room

temperature for 20 minutes. The obtained catalyst is labelled Fe0.5-Ar. To prepare the NH3-

pyrolyzed catalyst, Fe0.5-Ar is re-pyrolysed with the same flash-pyrolysis mode, but in flowing

pure NH3 and for only 5 minutes at 950 °C. The obtained catalyst is labelled Fe0.5-NH3.

Electrochemical characterization

Activity and durability in acidic and alkaline electrolytes are obtained using a RDE set-up (Pine

instruments) and either 0.1 M KOH or 0.1 M H2SO4 electrolytes. The three-electrode

configuration involves a platinum wire immersed in a H2-saturated electrolyte compartment,

separated from the main compartment by a fritted glass, as a reversible hydrogen electrode

(RHE) reference; a graphite plate as a counter electrode and a glassy carbon (GC) rotating disk

(5 mm diameter, Pine Research) as a support for the active layer forming the working

electrode. The ink is prepared by adding in sequence 5 mg catalyst, 54 µL Nafion® (5%

perfluorinated resin solution), 744 µL ethanol, 92 µL ultrapure water and sonicating for 1 hour

in an ice bath. An aliquot of 7 µL of the ink is pipetted on the GC disk and dried at room

temperature, resulting in a catalyst loading of 200 µg cm-2. The Initial activity is measured in

O2-saturated electrolyte at a scan rate of 1 mV·s-1 (SP-300, BioLogic Potentiostat) and at

rotation rate of 1600 rpm. Due to the low scan rate and low catalyst loading, no correction

54

for the capacitive current is needed. To evaluate the durability of the catalysts in RDE set-up,

a load-cycling protocol is applied, comprising 5000 triangular cycles performed at a scan rate

of 100 mV·s-1 in the potential range of 0.6-1.0 V vs RHE and in N2-saturated electrolyte. The

ORR activity after the AST is measured after re-saturating the solution with O2.

The leaching of Fe during short electrochemical cycling, performed either before or after the

AST, is investigated in an on-line electrochemical scanning flow cell (SFC) directly connected

to an inductively-coupled plasma mass spectrometer (ICP–MS), previously developed by

us.189-191 It is however challenging to continuously measure Fe leaching over the length of the

AST, due to drift of the ICP-MS with time and need for constant recalibration. To measure

56Fe, the ICP–MS (Perkin Elmer, NexION 350) is operated in dynamic-reaction-cell mode, using

methane as the reaction gas. The cell is calibrated to both an acidic and alkaline standard

solution of iron to ensure maximized detection of 56Fe. Daily calibration of the ICP–MS is done

by a four-point calibration curve (0, 0.5, 1.0, 5.0 μg·L−1) of standard iron solutions prepared

from Merck Centripur® ICP standards (Fe(NO3)3, 1000 mg·L−1, in 2–3% HNO3). As an internal

standard, we use 58Co (Merck Centripur®, Co(NO3)2, 1000 mg·L−1, in 2–3% HNO3) diluted to 50

μg·L−1 in HNO3 (0.15 mol·L−1) to ensure full acidification of the electrolyte in a y- connector

before its introduction in the ICP–MS. The SFC consists of a three-electrode setup using a

Ag/AgCl (Metrohm, 3 M KCl) reference electrode, a graphite rod counter electrode and a GC

RDE as a working electrode, on which the catalyst is drop cast. A positioning stage (Physik

Instrumente, M-403.6 DG) is used to approach individual catalyst spots on the working

electrode. Stability measurements are conducted in alkaline (99.99%, Suprapur®, NaOH, 0.05

mol·L−1) as well as in acidic (Suprapur®, 0.05 mol·L−1 H2SO4) electrolyte. The potentiostat

(Gamry, Reference 600) as well as purging gases and the positioning stage is controlled by a

custom LabVIEW software. The catalyst ink is prepared from the Fe-N-C catalyst, Nafion® (5%

perfluorinated resin solution) and water, with a mass ratio of catalyst/dry-ionomer of 4 and a

catalyst concentration of 3.3 g·L−1 in the liquid ink. An aliquot of 2.75 μL is deposited on the

GC, resulting in a catalyst loading of 400 μg·cm−2. Such a high loading is necessary to reach a

sufficient signal-to-noise ratio in the ICP-MS measurements. This is due to the

aforementioned interference of the 40Ar16O dimers and high background noise of iron in

alkaline solution. For Fe-leaching measurement before and after the AST, the latter is

conducted in a separate Teflon RDE-cell containing 100 mL electrolyte, and the RDE tip is then

55

quickly transferred from the Teflon cell to the SFC set-up, with the catalyst still wetted by

electrolyte. The RDE cell used for the AST consists of four individual compartments, one each

for the three electrodes and for the purging tube. The counter and reference electrodes are

the same as in the SFC setup.

Physico-chemical characterization

The pristine catalysts are characterized with 57Fe Mössbauer spectroscopy at room

temperature. To this end, 57Fe-enriched catalysts are used, prepared identically as the ones

otherwise investigated in this study, except for the use of 57Fe acetate during their synthesis.

Mössbauer spectra are measured at room temperature with a 57Co:Rh source. The

measurements are carried out in triangular velocity waveform using NaI scintillation detector

for γ-rays. The velocity calibration is done with an α-Fe foil. A mass of 30 mg of 57Fe-enriched

Fe0.5-Ar and Fe0.5-NH3 powders are necessary for a proper signal-to-noise resolution.

The pristine catalysts are also characterized with XAS in both ex situ and operando conditions.

The XAS spectra are collected at SAMBA beamline (synchrotron SOLEIL) at the Fe K-edge using

a double crystal Si 220 monochromator and a Canberra 35-elements germanium detector for

operando acquisition in fluorescence mode. The catalyst ink (10 mg catalyst, 100 µL 5% Nafion

solution and 50 µL ultrapure H2O) is prepared via ultrasonication, and 50 µL is deposited on

circa 3 cm2 circular area of a larger conductive carbon foil, resulting in a catalyst loading of

circa 1 mg cm-2 41. The carbon foil is then inserted in a three-electrode cell, 0.1 M KOH

electrolyte is added and the three electrodes are connected, using Pt-wire counter electrode

and a Hg/HgO reference electrode. Note that all potentials are however reported in V vs. RHE

in this work. Air is continuously bubbled in the electrolyte during the measurements. The

operando XAS spectra are collected at open circuit potential (OCP), 0.2, 0.4, 0.6, 0.8 and 1.0

V vs. RHE. Ex situ spectra were collected in transmission geometry on pellets of 1 mm

diameter using Teflon powder as a binder.

To measure the specific surface area of carbon in the catalysts, sorption isotherms of N2 are

measured in liquid nitrogen (77 K) with a Micromeritics, ASAP 2020 instrument. The sample

is previously cleaned at 200°C for 5 h in flowing nitrogen. The specific surface area is

determined by the multipoint Brunauer-Emmett-Teller (BET) method.

56

Results

Initial ORR activity and ex situ spectroscopic characterization

The ORR activity before and after the AST is shown in Figure 1 in linear- and semi-logarithmic

scales, for Fe0.5-Ar (Fig. 1a-b) and Fe0.5-NH3 (Fig. 1c-d). In each sub-figure, the measurements

performed in alkaline electrolyte are shown as purple curves while those performed in acidic

electrolyte are plotted in grey color. The curves before and after AST are identified with filled

and open symbols, respectively. Before the AST, Fe0.5-Ar shows a similar activity in both

electrolytes (Fig. 1b, curves with filled symbols), with ORR mass activity at 0.9 VRHE of 0.35 and

0.25 A·g-1 in alkaline and acidic electrolytes, respectively. A slight difference is visible only at

low-potential, with a less defined diffusion-limited current density in acidic medium (Figure

1a, filled grey symbols).

0,0 0,2 0,4 0,6 0,8 1,0-6

-5

-4

-3

-2

-1

0

1 10 100

0,6

0,7

0,8

0,9

0,0 0,2 0,4 0,6 0,8 1,0-6

-5

-4

-3

-2

-1

0

1 10 100

0,6

0,7

0,8

0,9

dc

b Fe

0.5-Ar; KOH; before AST

Fe0.5

-Ar; KOH; after AST

Fe0.5

-Ar; H2SO

4; before AST

Fe0.5

-Ar; H2SO

4; after AST

Cu

rren

t d

ensity (

mA

cm

-2)

aP

ote

ntial (V

vs.

RH

E)

Fe0.5

-NH3; KOH; before AST

Fe0.5

-NH3; KOH; after AST

Fe0.5

-NH3; H

2SO

4; before AST

Fe0.5

-NH3; H

2SO

4; after AST

Cu

rre

nt

den

sity (

mA

cm

-2)

Potential (V vs. RHE)

Po

tential (V

vs.

RH

E)

Mass activity (A g-1)

Fig. 1. RDE determination of the ORR activity of Fe0.5-Ar (a, b) and Fe0.5-NH3 (c, d) in acidic

(grey curves) and alkaline (purple curves) electrolyte before (filled symbols) and after (empty

symbols) the load-cycling AST. Polarization curves were measured in O2-saturated electrolyte

57

at a scan rate of 1 mV s-1, a rotation rate of 1600 rpm and a catalyst loading of 200 µg·cm-2.

The curves are not corrected for Ohmic loss. The AST has been conducted in N2-saturated

electrolyte, in a potential range of 0.6-1.0 V vs RHE, with a scan rate of 100 mV s-1 for 5000

cycles. The semi-logarithmic Tafel plots on the left handside have been obtained from the

polarization curve by applying the Koutecky-Levich equation, taking the value of diffusion-

limited current density as the current density at 0.4 V vs. RHE.

Table 1. Mass activity at 0.9 V vs. RHE before and after the AST for Fe0.5-Ar and Fe0.5-NH3 in acidic and

alkaline electrolytes. The AST comprises 5000 cycles between 0.6 V and 1.0 V vs. RHE, in N2-saturated

0.1 M KOH at room temperature. The catalyst loading was 200 µg·cm-2.

Mass activity / A g-1

Acidic electrolyte Alkaline Electrolyte

Catalyst ↓ BET Surface / m2 g-1 Before AST After AST Before AST After AST

Fe0.5-Ar 635 0.25 0.25 (-0%) 0.35 0.55 (+ 57%)

Fe0.5-NH3 970 0.50 0.35 (-30%) 4.60 4.15 (-10%)

A distinct behavior appears for Fe0.5-NH3, characterized with a much higher initial activity in

alkaline vs. acidic electrolyte (Figures 1c-1d, filled purple vs. filled grey symbols), with ORR

mass activity at 0.9 V vs. RHE of 4.6 and 0.5 A·g-1, respectively. The former is among the highest

reported value of the ORR activity in alkaline medium for PGM-free catalysts.42–46 Its lower

initial activity in acidic than in alkaline electrolyte can mostly be assigned to a fast protonation

of highly-basic nitrogen groups followed by anion-adsorption on positively-charged [NH]+

groups, which had been previously shown, on other NH3-pyrolyzed Fe-N-C catalysts, to divide

the activity by a factor 5 to 10.27,45,47 High initial activity of NH3-pyrolyzed Fe-N-C catalysts in

acidic liquid-electrolyte could previously be achieved only with high catalyst loading and

optimized Nafion/catalyst ratio.47 It is anticipated that these two parameters allow a short

time protection of the highly basic N-groups from the liquid acidic electrolyte. With the

present experimental conditions (low scan rate and low catalyst loading), the high ORR

activity of Fe0.5-NH3 cannot be captured in acidic liquid-electrolyte because the highly basic

N-groups were likely protonated and charge-neutralized even before the first polarization

curve was recorded. The surface state of Fe0.5-NH3 in acidic medium is then similar to that of

58

Fe0.5-Ar, explaining similar initial ORR activities in acid (compare filled grey symbols in Fig. 1b

and Fig. 1d). The slightly higher initial ORR mass activity in acid of Fe0.5-NH3 vs. Fe0.5-Ar (0.50

vs 0.25 A·g-1 at 0.9 V vs. RHE, see Table 1) can be explained by its higher BET specific area, 970

vs. 635 m2g-1.

A second possibility to explain the lower initial activity of Fe0.5-NH3 in acid vs. alkaline

electrolyte is that (at least some) FeNx moieties after NH3 pyrolysis are intrinsically different

from those after Ar-pyrolysis and, while being more active for ORR (regardless of which pH),

would also be less stable in acidic medium. It might be that such highly-active FeNx moieties

were leached in acidic medium even before completing the acquisition of the first polarization

curve. Evidence for increased Fe leaching with the Fe0.5-NH3 catalyst in acidic medium is

reported in section 3.3. We also note that the contribution of N-groups (not binding Fe) to

the overall initial ORR activity of those two catalysts can be neglected. Two reference

materials prepared similarly as Fe0.5-Ar and Fe0.5-NH3 but without addition of Fe acetate were

studied in RDE and their initial ORR activity in 0.1 M KOH at 0.9 V vs. RHE is 0.03 and 0.57 A·g-

1, respectively. This is circa 10 times lower than the activity of corresponding Fe-NC materials,

0.35 and 4.60 A·g-1 (Table 1). The difference would be even larger in acidic medium.

In order to characterize the active-site structure in Fe0.5-NH3 and Fe0.5-Ar, we first resorted to

ex situ spectroscopic characterization. The ex situ 57Fe Mössbauer spectra could not reveal

any significant difference between the two catalysts (Figure 2a-b and supporting Table S1).

They were fitted with two doublets D1 and D2, each one having similar Mössbauer

parameters for both catalysts and being present in similar ratio. These doublets are assigned

to atomically-dispersed FeNx moieties.18 Similar Mössbauer spectra imply that the local Fe

coordination and site geometry, up to two coordination spheres, are similar in both catalysts.

Further identification of the active-site structure was performed using XAS. Figure 2c shows

the ex situ X-ray absorption near-edge structure (XANES) spectra of Fe0.5-NH3 (circle symbols)

and Fe0.5-Ar (solid curve), that are characteristic for Fe-N-C catalysts free of metallic

particles.18 The absence of a strong signal at ca 2.2 Å (Fe-Fe bond distance in metallic and

metal-nitride particles, uncorrected for phase shift) in the Fourier transform (FT) of the EXAFS

spectra of both Fe0.5-Ar and Fe0.5-NH3 indicates that both catalysts are free of Fe-based

particles, or that the amount of particles is below the XAS detection limit, which is

approximately 3% relative to the total amount of Fe in the catalysts (Figure 2d). The absence

59

in the XANES spectra of the pre-edge peak at 7118 eV, characteristic for unpyrolyzed Fe(II)

phthalocyanine (FeN4 square-planar structure with identical Fe-N bond distances), reveals a

broken D4h symmetry of the FeNx moieties. This may be due to structural disorder or existence

of ferric moieties with an axial ligand such as O2. The latter case is particularly possible when

recording ex situ spectra of catalysts in their resting state in air environment. The white line

intensity is slightly stronger for Fe0.5-Ar than Fe0.5-NH3. This may be interpreted as a higher

coordination number in the first coordination sphere surrounding Fe (either N, C or O atoms).

This hypothesis is supported by the stronger signal of the first peak (at ca 1.4 Å) in the FT-

EXAFS spectra for Fe0.5-Ar vs. Fe0.5-NH3 (Figure 2d). Such ex situ changes may be assigned to

a stronger FeN4-O2 interaction ex situ for Fe0.5-Ar, possibly due to a higher average oxidation

state of Fe in Fe0.5-Ar vs. Fe0.5-NH3. In the latter, a lower oxidation state of Fe may be expected

due to the presence of Lewis-base (highly basic) nitrogen groups, if some of those groups are

directly involved in Fe cations ligation.

-6 -4 -2 0 2 4 6

0,94

0,96

0,98

1,00

-6 -4 -2 0 2 4 60,92

0,94

0,96

0,98

1,00

7110 7120 7130 7140 7150 7160 71700,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

0 1 2 3 4 50,0

0,2

0,4

0,6

0,8

1,0

1,2

1,4

Exp

Fit

D1 - 62,0%

D2 - 38,0%

Tra

nsm

issio

n

Velocity (mm s-1)

a

Tra

nsm

issio

n

Velocity (mm s-1)

Exp

Fit

D1 - 65,7%

D2 - 34,3%

Fe0.5

-NH3

Fe0.5

-Ar

d

b

c

No

rma

lize

d X

AN

ES

sig

na

l

Energy (eV)

Fe0.5

-NH3

Fe0.5

-Ar

Fo

uri

er

tra

nsfo

rm / A

-3

R (Angstrom)

60

Figure 2. Ex situ spectroscopic characterization of Fe0.5-NH3 and Fe0.5-Ar. 57Fe Mössbauer

transmission spectra (top) for a) Fe0.5-Ar and b) Fe0.5-NH3. XAS spectroscopic characterization

at the Fe K-edge (bottom) by c) XANES and d) EXAFS. The spectra were recorded in air at room

temperature. The distance in the FT-EXAFS spectra is not corrected for phase-shift.

Operando spectroscopic characterization

In order to investigate whether the small differences observed ex situ with XANES and FT-

EXAFS spectra of both catalysts remained, disappeared or were exacerbated during the ORR

in alkaline medium, we performed operando XAS. The XANES and EXAFS spectra were

recorded in alkaline electrolyte between 0.2 and 1.0 V vs. RHE, covering all regions of the RDE

polarization curves (Supporting Figure S1). Figure 3 shows the operando XANES spectra (left-

handside) and FT-EXAFS spectra (right-handside) for Fe0.5-Ar (top) and Fe0.5-NH3 (bottom). For

Fe0.5-Ar, both the operando XANES and FT-EXAFS spectra reveal large changes with the

electrochemical potential (Figure 3a-b). The change of the XANES spectra with potential

observed here for Fe0.5-Ar in alkaline medium is similar to that reported by us for the same

catalyst in acidic medium.41 The magnitude of change is however ca twice smaller in alkaline

vs. acidic medium for Fe0.5-Ar, as can be seen by comparing the Δµ spectra (compare the inset

of Fig. 3a in this work with the inset of Fig. 4b in Ref. 41). In the ORR potential region, the FeN4-

moieties in Fe0.5-Ar undergo a structural change, as revealed by the complex change and

existence of isobestic points at 7130.5 and 7156.4 eV in the set of XANES spectra.

61

7110 7120 7130 7140 7150 7160 7170

0,0

0,2

0,4

0,6

0,8

1,0

1,2

0 1 2 3 4 5

0,0

0,2

0,4

0,6

0,8

1,0

7110 7120 7130 7140 7150 7160 7170

0,0

0,2

0,4

0,6

0,8

1,0

1,2

0 1 2 3 4 5

0,0

0,2

0,4

0,6

0,8

1,0

7120 7140 7160

-0,2

-0,1

0,0

0,1

7120 7140 7160

-0,2

-0,1

0,0

0,1

0.2

0.4

0.6

0.8

1.0

Norm

aliz

ed X

AN

ES

sig

nal

Norm

aliz

ed X

AN

ES

sig

nal

Energy (eV)

1.0

0.8

0.6

0.4

0.2

a b

R(Angstrom)

1.0

0.8

0.6

0.4

0.2

c

Norm

aliz

ed µ

(E)

Energy (eV)

0.2

0.4

0.6

0.8

1.0

Fourier

transfo

rm / A

-3

d

Fourier

transfo

rm / A

-3

R(Angstrom)

1.0

0.8

0.6

0.4

0.2

Figure 3. Operando XAS characterization of Fe0.5-Ar (a-b) and Fe0.5-NH3 (c-d) in O2-

saturated 0.1 M KOH electrolyte. XANES (left-handside) and FT-EXAFS (right-handside) was

recorded between 0.2 and 1.0 V vs. RHE. The legend indicates the potential (in V vs RHE) at

which each spectrum was recorded. The insets in subfigures a and c show the Δµ spectra,

obtained by subtracting the normalized XANES spectrum at a given potential to the spectrum

recorded at 0.2 V vs. RHE.

These isobestic points support the existence of at least two Fe-site geometries, whose fraction

switches gradually from 0 to 1 across the potential range of 0.2 to 1.0 V vs. RHE. This

observation has also been reported, but only in acidic electrolyte hitherto, for other Fe-N-C

materials19,48,49 and interpreted as a change from in-plane FeN4 to out-of-plane FeN4

configuration as a function of potential. The operando FT-EXAFS (Figure 3b) also support a

structural change, with a decreasing signal intensity at 1.4 Å with decreasing potential. This

can be interpreted as the presence of oxygen adsorbates (O2, OH and H2O) strongly adsorbed

on FeN4 sites at high potential, and their absence or elongated Fe-O bond distance at low

potential. Those spectroscopic changes were reversible, similar spectra being recorded at a

62

given potential, when scanning down and then up the potential from 1.0 to 0.2 V and then

back up to 1.0 V.

The operando XANES spectra of Fe0.5-NH3 in alkaline electrolyte reveal much smaller changes

with electrochemical potential, as compared to changes observed for Fe0.5-Ar in the same

electrolyte (compare Figure 3c vs. Figure 3a). The Δµ spectra (inset of Fig. 3c) reveals a trend

that is however comparable to the Δµ signals observed with Fe0.5-Ar in alkaline electrolyte

(inset of Fig. 3a), but with ca twice lower magnitude of change. The smaller spectral changes

with electrochemical potential observed for Fe0.5-NH3 vs. Fe0.5-Ar may appear at first

counterintuitive, since the former has a higher BET area and expectedly a higher exposure of

Fe-sites to the top surface and to the electrolyte. The smaller spectral changes with potential

observed for Fe0.5-NH3 must therefore be assigned to a distinct environment of Fe-sites in

Fe0.5-NH3 compared to Fe0.5-Ar, in line with their much different ORR activities in alkaline

electrolyte (Figure 1).

In line with the operando XANES spectra, the corresponding FT-EXAFS spectra of Fe0.5-NH3 are

almost unchanged with potential (Fig. 3d). The trend of the signal intensity at 1.4 Å with

electrochemical potential is the same as for Fe0.5-Ar (decreasing signal with decreasing

potential), but the magnitude of the change is also much smaller. It is in fact only significant

at the lowest potential studied, namely 0.2 V vs. RHE.

Coming back to the initial question whether the small differences observed ex situ with XANES

and FT-EXAFS spectra of both catalysts remained or disappeared in operando, the direct

comparison of the XANES and EXAFS spectra of both catalysts at low potential (0.2 V vs. RHE,

Figure S2) reveals that the spectroscopic signatures are almost identical. Thus, the two

catalysts differ in ex situ conditions or in situ at high electrochemical potential, but the

differences become smaller as the potential is lowered, and become negligible at 0.2 V vs.

RHE. This supports the hypothesis that, in ex situ conditions, Fe is in a higher oxidation state

in Fe0.5-Ar than in Fe0.5-NH3, and binds more O2 or oxygen adsorbates. In operando conditions,

the difference progressively vanishes as the electrochemical potential is reduced (ferric

moieties turning into ferrous moieties in Fe0.5-Ar, becoming then in a similar state as Fe0.5-

NH3).

63

Electrochemical stability and operando iron leaching

The stability of Fe0.5-Ar and Fe0.5-NH3 catalysts was studied in both acidic and alkaline

electrolytes. Table 1 summarizes the activity observed before and after the 5000 load-cycle

AST. Starting with Fe0.5-Ar in alkaline medium, an activity increase from 0.07 to 0.11 mA cm-2

at 0.9 V vs. RHE can be seen after 5000 load-cycles in N2-saturated solution (purple curves in

Fig. 1a-b). Similarly, also Fe0.5-NH3 shows high stability in alkaline electrolyte, with very slight

decay in activity at 0.9 V vs. RHE (from 0.92 to 0.83 mA·cm-2, see purple curves in Fig. 1c-d).

In acidic medium, no activity loss was observed for Fe0.5-Ar, but a significant loss observed for

Fe0.5-NH3, with a relative decrease of about 24% (Table 1).

We attribute this reduced activity following AST in N2-saturated electrolyte to a loss of a

fraction of the active centers, i.e. FeN4 or only Fe cations from FeN4 moieties, from the

nitrogen-doped carbon network. Since the AST was performed in N2-saturated electrolyte, no

ORR occurred during the AST and we can exclude a degradation or deactivation due to

hydrogen peroxide or reactive oxygen species formed from peroxide and FeNx sites, which

was recently demonstrated to cause a main deactivation of Fe-N-C catalysts in acidic

medium.34 To show this loss of iron in situ, we conducted SFC-ICP–MS measurements. Figure

4 summarizes the operando Fe leaching measurements where we applied the same potential-

scan protocol (upper plot) to Fe0.5-Ar and Fe0.5-NH3 in oxygen-saturated electrolyte. The

electrode was first contacted by the SFC (corresponding time marked with * in the graph, t ~

250 s) at open circuit potential (OCP). The applied electrochemical potential protocol is then

20 cyclic voltammograms (CVs) in the range 1.0 to 0.6 V vs. RHE at a scan rate of 100 mV·s−1,

a potential range typically occurring during the ORR in fuel cell devices. Another

chronoamperometry at 1.0 V was recorded for 200 s, before applying one CV at a low scan

rate of 2 mV·s−1 from 1.0 V down to 0.0 V vs. RHE, and back up to 1.0 V vs. RHE to identify in

more detail the potential-dependence of Fe dissolution

For each catalyst considered separately, Figure 4 clearly identifies higher Fe dissolution rates

in acid than in alkaline electrolyte, especially during the slow CV. We first discuss the transient

release of Fe occurring when the SFC contacts the electrode with the electrolyte (time marked

with an asterisk), leading to an initial loss of iron. For Fe0.5-Ar, this initial loss is similar in both

electrolytes (Fig. 4a, middle panel), while for Fe0.5-NH3 the dissolution rate in acid medium is

64

more than double that in alkaline electrolyte (Fig. 4b, middle panel). Comparing the contact

dissolution in acidic medium, the peak dissolution rates are ca 1.0 and 0.4 ngFe·cm-2·s-1 for

Fe0.5-NH3 and Fe0.5-Ar, respectively. After 400 s at OCP in acid medium, the Fe dissolution rate

became < 0.15 ngFe·cm-2·s-1 for Fe0.5-Ar but remained significant (~ 1.0 ngFe·cm-2·s-1) and quite

constant for Fe0.5-NH3. The cumulative Fe dissolution of Fe0.5-NH3 at that stage is however

restricted to ca 0.5 µgFe·cm-2 (Fig. 4b, lower panel), much lower than the total amount of Fe

in the catalyst layer (ca 2.5 wt% Fe in Fe0.5-NH3, leading to ca 10 µgFe·cm-2). It is therefore

unlikely that the lower activity measured for Fe0.5-NH3 in acid vs. alkaline originates from a

very fast dissolution of iron. In alkaline medium, the curves of Fe dissolution rate vs. time of

both catalysts are nearly superimposed (from immersion to OCP hold, time 250 to 850 s on x-

axis), with a peak value of Fe dissolution rate of ca 0.5 ngFe·cm-2·s-1, quickly decreasing (only

ca 0.1 ngFe·cm-2·s-1 before starting the fast CVs). Therefore, a trend is observed that Fe0.5-Ar is

more stable than Fe0.5-NH3, and that alkaline environment leads to lower dissolution rate than

acidic medium.

Figure 4. Time-resolved potential (top) dissolution rates (middle) and totally dissolved

amount (bottom) of iron from Fe0.5-Ar (a) and Fe0.5-NH3 (b) according to the colour scheme of

Figure 1. Note the different range of dissolution values in the Y-axis for a) and b). The contact

dissolution peak is marked with an asterisk in the upper panels. Dissolution rates after AST

have been omitted for clarity in the middle panels since they are negligible.

0,0

0,5

1,0

0

3

6

9

0 500 1000 1500 2000

0

1

2

E (

V v

s. R

HE

)d

(Fe)/

dt*

S (

ng c

m-2 s

-1)

Fe

Dis

s (µ

g c

m-2)

t (s)

NaOH ; before AST

NaOH ; after AST

H2SO

4 ; before AST

H2SO

4 ; after AST

*

b

0,0

0,5

1,0

0,0

0,5

1,0

1,5

0 500 1000 1500 2000

0,0

0,5

E (

V v

s. R

HE

)d

(Fe)/

dt*

S (

ng c

m-2 s

-1)

a

Fe

Dis

s (

µg c

m-2)

t (s)

NaOH ; before AST

NaOH ; after AST

H2SO

4 ; before AST

H2SO

4 ; after AST

*

65

The subsequent fast 20 CV scans increased the Fe release rate in acidic medium, especially

for Fe0.5-NH3 (increasing to 1.5 ngFe·cm-2·s-1), while it had no impact on the Fe dissolution rate

in alkaline medium. Note that, due to the high scan rate used during the 20 CVs, the effect of

scanning up or down the potential cannot be distinguished, and only a lump Fe dissolution

rate is observed.

The time-resolved Fe dissolution rate during the subsequent slow potential scan then allows

us identifying at which potential Fe is dissolved. In sulfuric acid, the onset of Fe dissolution

while scanning the potential from 1 V down to 0 V occurs at ca 0.75 V vs. RHE, and the peak

of dissolution rate occurs at ca 0.2-0.3 V vs. RHE, for both catalysts. The intensity of the peak

of Fe dissolution is however 10 times higher for Fe0.5-NH3 vs. Fe0.5-Ar. At potentials E < 0.2 V

vs. RHE, the Fe dissolution rate decreases for both catalysts, and remains very low also during

the positive-going scan from 0.0 V to 1.0 V vs. RHE. For Fe0.5-NH3, the cumulative Fe amount

leached after the slow CV reaches ca 2 µg·cm-2, representing about 50% of the total Fe content

initially present (Fig. 4b, lower panel). Regarding the Fe release rate during the slow CV in

alkaline electrolyte, there is no significant effect of the electrochemical potential in the

negative-going branch of the scan, while reverting the scan direction from 0.0 V and upwards

resulted in increased Fe dissolution rate for Fe0.5-NH3 but unmodified Fe dissolution rate for

Fe0.5-Ar. These experiments were repeated multiple times, and showed reproducible trends.

While Figure 4 informs on the electrochemical conditions in which Fe is dissolved, Figure 5

quantitatively shows how much Fe from the catalysts was dissolved as a function of time in

the SFC-ICP-MS protocol before the AST. The y-axis shows the %Fe remaining in the catalyst

relative to the initial Fe content. The cumulative dissolved Fe content was obtained from the

integral of the curves shown in the lower panels of Figure 4 while the total Fe content in each

electrode was derived from i) the fixed Fe-N-C catalyst loading value and the exact geometric

area investigated by SFC-ICP-MS (verified each time by a microscope) and ii) the knowledge

of the initial Fe content in each catalyst. The latter were measured by ICP-MS on the catalyst

powders to be 1.45 wt% for Fe0.5-Ar and 1.57 wt% for Fe0.5-NH3. Figure 5 shows that the

absolute Fe dissolution is restricted for Fe0.5-Ar (at both pH) and for Fe0.5-NH3 at high pH (5 to

10% relative Fe content is dissolved after 20 fast CVs and a slow scan) while for Fe0.5-NH3 at

acidic pH, more than 50% of the initial Fe content present in the active layer was dissolved

after the same time.

66

0 500 1000 1500 20000

10

20

30

40

50

60

70

80

90

100

Re

lative

Fe c

onte

nt

rem

ain

ing

in t

he

cata

lyst

(%)

t (s)

Fe0.5

- Ar, H2SO

4

Fe0.5

- Ar, NaOH

Fe0.5

- NH3, H

2SO

4

Fe0.5

- NH3, NaOH

*

0,0

0,2

0,4

0,6

0,8

1,0

1,2

E (

V v

s. R

HE

)

Figure 5. Percentage of initial iron remaining in the catalyst as a function of time. The

electrochemical potential applied as a function of time is the same as that shown in Figure 6.

These time-resolved Fe dissolution data reveal that the Fe-based sites in Fe0.5-NH3 are less

stable in acidic medium than those present in Fe0.5-Ar, while in alkaline medium the stability

of Fe0.5-NH3 is as good, or even better, than that of Fe0.5-Ar. While the data might be

interpreted by assuming that a much higher fraction of all FeNx sites are exposed to the

electrolyte in Fe0.5-NH3 than in Fe0.5-Ar, this assumption should have resulted in a slightly

increased Fe dissolution for Fe0.5-NH3 in alkaline electrolyte compared to that for Fe0.5-Ar in

the same electrolyte. This is however not observed. The operando XAS data are also not in

support of an increased fraction of FeNx sites being exposed to the electrolyte in Fe0.5-NH3

(smaller magnitude of change for the XANES and EXAFS spectra with potential than for Fe0.5-

Ar). Thus, the electrolyte-exposed FeNx sites in Fe0.5-NH3 seem to be intrinsically less stable in

acidic medium than those in Fe0.5-Ar.

Discussion

The operando XANES and EXAFS data in alkaline electrolyte reveal that the catalyst Fe0.5-NH3

experiences less change of its site geometry and Fe oxidation state as a function of the

electrochemical potential, as compared to Fe0.5-Ar. This is assigned to a lower average

67

oxidation state of Fe cations in FeNx moieties in the resting state for Fe0.5-NH3 than for Fe0.5-

Ar. These fine differences between FeN4 sites in Ar-pyrolyzed and NH3-pyrolyzed catalysts are

revealed here for the first time by operando XAS, and can explain the higher TOF at high

potential for ORR of Fe0.5-NH3 relative to Fe0.5-Ar. The lower average oxidation state of Fe in

NH3-pyrolysed catalysts may be a consequence of the presence of nitrogen groups with Lewis

basicity. It can be reasonably proposed that the involvement of highly basic nitrogen groups

in Fe ligation in Fe0.5-NH3 results in increased electron density at the Fe centers, increased O2

binding and also introduces the possibility to immobilize protons near the Fe centers, which

could reduce the energy barrier during the rate determining step of the ORR. However, if

highly basic nitrogen groups are directly involved in the coordination of all or some FeNx

moieties, it can be expected that such moieties will be stable only in alkaline electrolyte, and

not in acidic medium. The operando Fe leaching measurements support this hypothesis, with

increased Fe leaching specifically observed for Fe0.5-NH3 in acidic medium. The instability in

acidic medium of some FeNx moieties present in Fe0.5-NH3 may thus be assigned to the higher

basicity of N-groups that ligate some of the iron cations. Upon their protonation in acidic

medium, the covalent bond that previously existed between such Fe cations and nitrogen is

broken or weakened, and the iron cations are dissolved in the electrolyte.

It is however unresolved from the dissolution data whether such unstable FeNx moieties in

acidic medium account for the vast majority of the ORR activity of pristine Fe0.5-NH3, or both

stable and unstable FeNx moieties co-exist in comparable amount. The latter hypothesis is

more likely. Due to the disorder of the system formed during high-temperature pyrolysis in

NH3, one might expect that two types of moieties coexist, i) FeNx moieties with Fe ligated by

at least one highly-basic nitrogen group, and ii) FeNx moieties with Fe ligated only by nitrogen

groups with low pKa value (non-protonating in pH 1). The existence of this mixed system of

FeNx moieties would explain the irreversible loss of ORR activity experienced by NH3-

pyrolyzed Fe-N-C catalysts after an acid-wash but also the fact that the very low ORR activity

after acid-wash (activity / initial activity = 0.1) can be recovered to about 0.5 of the initial

activity after a mild re-heat treatment at 300 °C (that removes anions and restores the N-

groups in a non-protonated state).27 Bringing further complexity, the online Fe dissolution

data reveals here that the Fe leaching from Fe0.5-NH3 in acid medium is significantly enhanced

68

when the electrochemical potential is < 0.75 V vs. RHE, and almost peaks at 0.5 V vs. RHE, a

potential close to the one often chosen during stability testing of PGM-free cathode catalysts

in PEMFC.

Thus, while there is no doubt that the nitrogen protonation and anion-binding phenomenon

reduces the high activity of NH3-pyrolysed Fe-N-C catalysts in liquid acid electrolyte in RDE

set-up, it is unclear whether this effect is responsible for the fast decay of NH3-pyrolysed Fe-

N-C catalysts during the first 10-15 h of operation in PEMFC. The online Fe dissolution data

presented here suggest that the Fe dissolution rate of NH3-pyrolysed Fe-N-C catalysts in acid

medium may be very fast at cathode potentials of 0.3-0.6 V vs. RHE. The circa 10 x faster Fe

leaching rate from Fe0.5-NH3 than from Fe0.5-Ar in liquid acid medium in this potential range

is in line with the relative degradation rate of Fe0.5-NH3 vs Fe0.5-Ar in PEMFCs. Further study

exploring the potential-dependence and atmosphere-dependence (O2, air or simply N2) of the

performance degradation of Fe0.5-NH3 during potentiostatic control of PEMFC cathodes,

combined with Fe dissolution measurements may strengthen this hypothesis.

The antagonism between ORR activity and stability of Fe0.5-NH3 revealed here in acid medium

does not exist in alkaline electrolyte, where high activity and high stability are simultaneously

met. This supports the idea that highly-basic N-groups are at the root of the high ORR activity

of FeNx moieties in ammonia-pyrolyzed Fe-N-C catalysts. Such catalysts are therefore proper

candidates for replacing Pt-based catalysts in AEMFCs.

Conclusions

Two Fe-N-C catalysts comprising only atomically-dispersed FeNx moieties were prepared,

differing only in the fact that the second catalyst (Fe0.5-NH3) was obtained by subjecting the

first one (Fe0.5-Ar) to a short pyrolysis in ammonia. While the initial ORR activity in acid

medium in RDE setup is similar for both catalysts, the activity in alkaline medium is

significantly higher for Fe0.5-NH3. Operando XAS measurements in alkaline electrolyte reveals

similar trends of the spectra as a function of the electrochemical potential for both catalysts,

but the magnitude of change is much less for Fe0.5-NH3, as evidenced by a Δµ analysis.

Accelerated stress tests in alkaline and acidic electrolyte revealed that the ORR activity of

both catalysts was very stable in alkaline electrolyte, while some activity decay is observed

for both catalysts in acidic electrolyte after 5000 cycles. Time-resolved Fe dissolution

69

combined with previous literature studies point that the lower ORR activity of Fe0.5-NH3 in

acid vs. alkaline liquid electrolyte is the outcome of two phenomena, i) the leaching of a

fraction of acid-unstable FeNx moieties, and ii) the protonation and charge-neutralization by

counter-anions of the electrolyte of highly-basic N-groups. Overall, ammonia pyrolysis of Fe-

N-C catalysts is shown to result, in alkaline medium, in high ORR activity of atomically-

dispersed FeNx moieties, high ORR durability and minimized Fe leaching during

electrochemical operation in load-cycling mode. In acid electrolyte, the ammonia pyrolysis of

Fe-N-C catalysts results in circa 10 times enhanced Fe leaching relative to the reference inert-

gas pyrolyzed catalyst, with a Fe leaching rate that is strongly enhanced when an

electrochemical potential in the range 0.75 to 0.3 V vs. RHE is applied. This may explain the

recognized reduced stability of ammonia-pyrolyzed Fe-N-C catalysts in operating PEMFCs.

70

References

1. P. P. Edwards, V. L. Kuznetsov, W. I. F. David, and N. P. Brandon, Energy Policy, 36, 4356–4362 (2008).

2. V. Mehta and J. S. Cooper, J. Power Sources, 114, 32–53 (2003). 3. J. Wu, X. Yuan, J. Martin, H. Wang, J. Zhang, J. Shen, S. Wu and W. Merida, J. Power Sources,

184, 104–119 (2008). 4. H. A. Gasteiger, S. S. Kocha, B. Sompalli, and F. T. Wagner, Appl. Catal. B Environ., 56, 9–35

(2005). 5. K. C. Neyerlin, W. Gu, J. Jorne, and H. A. Gasteiger, J. Electrochem. Soc., 153, A1955 (2006) 6. Y. Wang, K. S. Chen, J. Mishler, S. C. Cho, and X. C. Adroher, Appl. Energy, 88, 981–1007 (2011) 7. D. Banham, S. Ye, K. Pei, J. Ozaki, and T. Kishimoto, J. Power Sources, 285, 334–348 (2015) 8. Z. Chen, D. Higgins, A. Yu, L. Zhang, and J. Zhang, Energy Environ. Sci., 4, 3167–3192 (2011). 9. M. Shao, Q. Chang, J.-P. Dodelet, and R. Chenitz, Chem. Rev., 116, 3594–3657 (2016) 10. D. Banham and S. Ye, ACS Energy Lett., 2, 629–638 (2017). 11. A. Bonakdarpour, M. Lefevre, R. Yang, F. Jaouen, T.Dahn, J. Dodelet and J.R. Dahn,

Electrochem. Solid-State Lett., 11, B105 (2008) 12. F. Jaouen J. Herranz, M. Lefévre, J. Dodelet, U.I. Kramm, I. Herrmann, P. Bogdanoff, J.

Maruyama, T. Nagaoka, A. Garsuch, J.R. Dahn, T. Olson, S. Pylypenko, P. Atanassov and E.A. Ustinov, ACS Appl. Mater. Interfaces, 1, 1623–1639 (2009).

13. H. Tan, H. Tan, Y. Li, X. Jiang, J. Tang, Z. Wang, H. Qian, P. Mei, V. Malgras, Y. Bando and Y. Yamauchi, Nano Energy, 36, 286–294 (2017)

14. M. Hossen, K. Artyushkova, P. Atanassov, and A. Serov, J. Power Sources, 375, 214–221 (2017) 15. J. Tian, A. Morozan, M.T. Sougrati, M. Lefèvre, R. Chenitz, J. Dodelet, D. Jones and F. Jaouen,

Angew. Chemie - Int. Ed., 52, 6867–6870 (2013). 16. S. Ratso, I. Kruusenberg, M. Käärik, M. Kook, R. Saar, P. Kanninen, T. Kallio, J. Leis and K.

Tammeveski, Appl. Catal. B Environ., 219, 276–286 (2017) 17. E. Proietti, F. Jaouen, M. Lefevre, N. Larouche, J. Tian, J. Herranz and J. Dodelet, Nat. Commun.,

2, 1–6 (2011). 18. . Zitolo, V. Goellner, V. Armel, M.T. Sougrati, T. Mineva, L. Stievano, E. Fonda and F. Jaouen,

Nat. Mater., 14, 937–942 (2015) 19. J. Li, S. Ghoshal, W. Liang, M.T. Sougrati, F. Jaouen, B. Halevi, S. McKinney, G. McCool, C. Ma,

X. Yuan, Z. Ma, S. Mukerjee and Q. Jia, Energy Environ. Sci., 9, 2418–2432 (2016). 20. U. I. Kramm, J. Herranz, N. Larouche, T.M. Arruda, M. Lefèvre, F. Jaouen, P. Bogdanoff, S.

Fiechter, I. Abs-Wurmbach, S. Mukerjee and J. Dodelet, Phys. Chem. Chem. Phys., 14, 11673 (2012)

21. J. Li and F. Jaouen, Curr. Opin. Electrochem., 9, 198–206 (2018) 22. M. Xiao, J. Zhu, L. Feng, C. Liu, and W. Xing, Adv. Mater., 27, 2521–2527 (2015). 23. G. Zhong, H. Wang, H. Yu, and F. Peng, J. Power Sources, 286, 495–503 (2015) 24. K. Kumar, P. Gairola, M. Lions, N. Ranjbar-Sahraie, M. Mermoux, L. Dubau, A. Zitolo, F. Jaouen

and F. Maillard, ACS Catal., 8, 11264–11276 (2018) 25. G. Zhang, R. Chenitz, M. Lefèvre, S. Sun, and J. P. Dodelet, Nano Energy, 29, 111–125 (2016) 26. R. Chenitz, U.I. Kramm, M. Lefevre, V. Glibin, G. Zhang, S. Sun and J. Dodelet, Energy Environ.

Sci., 11, 365– 382 (2018). 27. J. Herranz, F. Jaouen, M. Lefèvre, U.I. Kramm, E. Proietti, J. Dodelet, P. Bogdanoff, S. Fiechter,

I. Abs-Wurmbach, P. Bertrand, T.M. Arruda and S. Mukerjee, J. Phys. Chem. C, 115, 16087–16097 (2011).

28. J. Choi, L. Yang, T. Kishimoto, X. Fu, S. Ye, Z. Chen and D. Banham, Energy Environ. Sci., 10, 296–305 (2017)

29. A. K. Mechler, N. Ranjbar-Sahraie, V. Armel, A. Zitolo, M.T. Sougrati, J.N. Schwämmlein, D. Jones and F. Jaouen, J. Electrochem. Soc., 165, F1084–F1091 (2018)

71

30. C. H. Choi, C. Baldizzone, J.P. Grote, A.K. Schuppert, F. Jaouen and K.J. Mayrhofer, Angew.

Chemie - Int. Ed., 54, 12753–12757 (2015). 31. V. Goellner, V. Armel, A. Zitolo, E. Fonda, and F. Jaouen, J. Electrochem. Soc., 162, 403–414

(2015). 32. C. H. Choi, W.S. Choi, O. Kasian, A.K. Mechler, M.T. Sougrati, S. Brüller, K. Strickland, Q. Jia, S.

Mukerjee, K.J. Mayrhofer and F. Jaouen, Angew. Chemie - Int. Ed., 56, 8809–8812 (2017) 33. V. Goellner, C. Baldizzone, A. Schuppert, and T. Sougrati, Phys. Chem. Chem. Phys., 16, 18454–

18462 (2014) 34. C. H. Choi, H. Lim, M.W. Chung, G. Chon, N. Ranjabar-Sahraie, A. Altin, M.T. Sougrati, L.

Stievano, H.S. Oh, E. Park, F. Luo, P. Strasser, G. Dražić, K. Mayrhofer, H. Kim and F. Jaouen, Energy Environ. Sci., 11, 3176– 3182 (2018)

35. D. R. Dekel, J. Power Sources, 375, 158–169 (2018) 36. S. Lu, J. Pan, A. Huang, L. Zhuang, and J. Lu, Proc. Natl. Acad. Sci., 105, 20611 LP-20614 (2008). 37. T. J. Omasta, X. Peng, H.A. Miller, F. Vizza, L. Wang, J.R. Varcoe, D.R. Dekel and W.E. Mustain,

J. Electrochem. Soc., 165, J3039–J3044 (2018) 38. S. O. Klemm, A. A. Topalov, C. A. Laska, and K. J. J. Mayrhofer, Electrochem. commun., 13,

1533–1535 (2011) 39. N. Kulyk, S. Cherevko, M. Auinger, C. Laska, and K. J. J. Mayrhofer, J. Electrochem. Soc., 162,

H860–H866 (2015) 40. A. K. Schuppert, A. A. Topalov, I. Katsounaros, S. O. Klemm, and K. J. J. Mayrhofer, J.

Electrochem. Soc., 159, F670–F675 (2012) 41. A. Zitolo, N. Ranjbar-Sahraie, T. Mineva, J. Li, Q. Jia, S. Stamatin, G.F. Harrington, S.M. Lyth, P.

Krtil, S. Mukerjee, E. Fonda, and F. Jaouen, Nat. Commun., 8, 1–10 (2017) 42. H. Meng, F. Jaouen, E. Proietti, M. Lefèvre, and J. P. Dodelet, Electrochem. commun., 11, 1986–

1989 (2009) 43. M. Piana, S. Catanorchi, and H. A. Gasteiger, Trans. E C S Soc. Electrochem., 16, 2045–2055

(2008). 44. H. Su, S. Zhou, X. Zhang, H. Sun, H. Zhang, Y. Xiao, K. Yu, Z. Dong, X. Dai, and X. Huang, Dalt.

Trans., 47, 16567–16577 (2018) 45. M. Rauf, Y. Zhao, Y. Wang, Y. Zheng, C. Chen, X. Yang, Z. Zhou, and S. Sun, Electrochem.

commun., 73, 71–74 (2016) 46. H. T. Chung, J. H. Won, and P. Zelenay, Nat. Commun., 4, 1922 (2013) 47. F. Jaouen, V. Goellner, M. Lefèvre, J. Herranz, E. Proietti, and J. Dodelet,Electrochim. Acta, 87,

619–628 (2013) 48. Q. Jia, N. Ramaswamy, H. Hafiz, U. Tylus, K. Strickland, G. Wu, B. Barbiellini, A. Bansil, E.F.

Holby, P. Zelenay, and S. Mukerjee, ACS Nano, 9, 12496–12505 (2015) 49. M. Xiao, J. Zhu, L. Ma, Z. Jin, J. Ge, X. Deng, Y. Hou, Q. He, J. Li, Q. Jia, S. Mukerjee, R. Yang, Z.

Jiang, D. Su, C. Liu, and W. Xing ACS Catal., 8, 2824–2832 (2018)

72

Effect of Mn-oxides on the Oxygen and Peroxide

Reduction Reactions for MnOx/FeNC Composites in

Alkaline Medium

Pietro Giovanni Santori,1 Florian Speck,2 Serhiy Cherevko,2 Xiong Peng,3 William E. Mustain,3

Frédéric Jaouen1

1. Institut Charles Gerhardt Montpellier, UMR 5253, CNRS, Université Montpellier, ENSCM,

Place Eugène Bataillon, 34095 Montpellier cedex 5, France

2. Helmholtz-Institute Erlangen-Nürnberg for Renewable Energy (IEK-11),

Forschungszentrum Jülich, Egerlandstr. 3, 91058 Erlangen, Germany

3. Department of Chemical Engineering, University of South Carolina, Columbia, SC 29208,

USA

73

Abstract

The Anion Exchange Membrane Fuel Cell is gaining interest due to the high power

performance recently achieved, even though with precious metal catalysts. Promising results

have recently been obtained with non-noble cathode catalysts, an important step toward the

quest of precious-metal-free AEMFCs. FeNC catalysts have shown high ORR activity as well as

high stability in alkaline electrolyte, but the production of even minute amount of hydrogen

peroxide during ORR may lead to premature catalyst or ionomer degradation during

operation. To increase the activity of FeNC cathodes towards the electroreduction of

hydrogen peroxide, four manganese oxides (α-MnO2, β-MnO2, δ-MnO2 and α-Mn2O3) were

investigated as co-catalysts. The HPRR activity of the four Mn-oxides supported on Vulcan

carbon black was evaluated showing comparably high apparent activity, but significantly

higher surface-specific activity for Mn2O3. Combined with a FeNC catalyst comprising mainly

FeNx sites and that was prepared via pyrolysis in ammonia, the composites 20%-MnOx/80%-

FeNC show a circa 20 % lower initial ORR activity, only due to a dilution effect of FeNC in the

composites. Higher HPRR activity is observed for all four composites relative to FeNC, in

particular for Mn2O3/Fe-N-C. Higher selectivity during ORR is also observed for the

composites. The stability of the Mn2O3/Fe0.5-NH3 composite catalyst was studied in operando

SFC-ICP/MS to evaluate the online leaching of Mn and Fe and to understand if there is a

stabilization effect of either Mn or Fe in the Mn2O3/Fe0.5-NH3 composite relative to either

Fe0.5-NH3 or to Mn2O3/C. In a final step, Fe0.5-NH3 and the composite catalyst Mn2O3/Fe0.5-NH3

were applied at the cathode of an AEMFC, achieving similar current density at 0.9 V than a

commercial Pt/C cathode with 0.4 mgPt cm-2.

Introduction

The energy-environment nexus calls for the development of renewable energy sources and

more efficient machines. Renewable electric power, produced from solar or wind energies for

example, can be efficiently converted to hydrogen via water electrolysis, and hydrogen used

as a green fuel, on demand.1 To this end, fuel cell devices can efficiently re-convert H2 and O2

to electric power with high efficiency. This technology is well positioned to compete, among

other applications, with internal combustion engines for automotive application.2 Among

different types of fuel cells, Proton Exchange Membrane Fuel Cells (PEMFCs) are currently the

74

ones envisioned and developed by the automotive industry for transportation application,

due to their key advantages in power density, internal resistance, maintenance, durability and

fast start-up.3,4 In the long term however, the main drawback of PEMFCs is tied to the kinetics

of the Oxygen Reduction Reaction (ORR) that occurs at the cathode side. The sluggishness of

the ORR in the acidic medium of PEMFC forces the use of the most possibly active catalysts,

namely platinum-based ones, in order to mitigate the energy efficiency and power density

losses in PEMFC that are associated with the ORR overpotential.5,6 Platinum is a rare and

expensive metal that increases the production cost of the device, threatening the large-scale

commercialization of PEMFCs.5-8 The need to reduce the amount of platinum (at large,

platinum group metals (PGMs), Pt and Pd being the most abundant among the PGMs) in high-

performance PEMFCs has led the scientific community involved in electrocatalysis to

investigate two separate approaches, one leaning on the improvement of PGM catalysts and

cathode layers to utilize more efficiently such metals,9-14 and one exploring the synthesis of

PGM-free catalysts with activity and durability that must, ultimately, become comparable to

those of Pt-based catalysts.

Among the PGM-free catalysts investigated since the early 2000’s, the class of MNC catalysts

with 3d transition metals (Fe and Co, in particular) integrated in a N-doped carbon matrix has

hitherto shown the most promising activity and stability results in acidic medium.15-21 While

oxynitrides of transition metals of group 4 and 5 in the periodic table have also been reported

with promising activity.22-24 The nature of the most active catalytic sites in MNC catalysts

obtained via pyrolysis has been controversial since their emergence. Recent studies from

several groups are now however converging to a same conclusion that, in acidic medium, the

most active and stable sites in MNC catalysts are atomically-dispersed MNx moieties.25-34 This

conclusion could be reached with the preparation of model MNC materials, exclusively

comprising single-metal-atom structures and no metal clusters (to the detection limit of X-ray

absorption and 57Fe Mössbauer spectroscopies). Combined theoretical and experimental

interpretation of the spectroscopic response of such model MNC catalysts prepared from

different precursors are also converging to the picture of FeNx and CoNx moieties (x ≤ 4)

embedded in the carbon matrix as porphyrinic MN4C12 or pyridinic MN4C10 structures, in +II

or +III oxidation state and in different spin states.17,26,27,35-38 Binuclear Fe2N5 sites have also

75

been proposed and could be present as a minority site, but possibly with higher turnover

frequency, in FeNC catalysts.39

Besides the need to further improve the activity of MNC catalysts for matching the high

PEMFC performance currently obtained with Pt-based catalysts, the main bottleneck for the

industrial application of MNC catalysts in PEMFC is however, presently, their limited durability

during PEMFC operation. Their degradation mechanisms have been intensively explored in

the last few years, in rotating disk electrode in acid electrolyte, in PEMFC and also when

subjected to ex situ Fenton reaction tests.19,34,40-50 While some MNC catalysts show high

stability when load-cycled in acidic medium in deaerated conditions,34,42 all MNC catalysts can

be expected to suffer from severe carbon corrosion during startup/shutdown accelerated

stress tests, leading to the destruction of MNx sites and the leaching of the metal cations.34,51

In addition, highly microporous FeNC catalysts, often prepared via NH3 pyrolysis, are the most

active MNC catalysts in acidic medium but also the least stable ones in PEMFC.16,18,40,52,53

Recent studies by Dodelet’s group and also our group have shown that NH3-pyrolyzed

microporous FeNC materials show distinctly faster Fe leaching rates in acidic medium than

FeNC catalysts prepared similarly but pyrolyzed in inert gas.43,54 Our study reported that the

leaching rate of iron in O2-saturated acid electrolyte during a linear scan between 1 and 0 V

vs. RHE was circa 10 times higher for a microporous FeNC catalyst (prepared via a first

pyrolysis in Ar and a second pyrolysis in ammonia, labelled Fe0.5-NH3) relative to the FeNC

catalyst prepared identically except that it was not subjected to the second pyrolysis in

ammonia (labelled Fe0.5-Ar).43 Furthermore, the Fe leaching rate was measured online and in

operando, identifying that the Fe leaching rate increased when the potential was decreased

from 1 to 0, with an onset potential for Fe leaching at ca 0.75 V vs. RHE and a peak of

dissolution rate at 0.2-0.3 V vs. RHE. It was also shown that both the Fe0.5-NH3 and Fe0.5-Ar

catalysts exclusively comprised FeNx sites. Thus, the distinct lower stability of FeNx sites from

Fe0.5-NH3 in acidic electrolyte could be assigned to either the higher microporosity in Fe0.5-

NH3, or to weaker Fe-N bonds in acidic medium. These new insights must also be connected

with previous studies from Dodelet’s group where it was shown that NH3 pyrolysis results in

nitrogen-groups with high pKa values, and the high initial activity in acid medium was

proposed to be due to the protonation of such N-groups in acidic medium, and their fast

deactivation due to subsequent anion adsorption on positively-charged NH groups.55

76

Combined together, those results can be interpreted as the coordination of Fe cations with

highly basic N-ligands in Fe0.5-NH3, such a Metal-N coordination leading to high ORR activity

but low stability in acidic medium, due to the competition between Fe cations and protons

for such highly basic N-groups. Strongly supporting this hypothesis, Fe0.5-NH3 was, in alkaline

electrolyte, as stable as Fe0.5-Ar, both from the perspective of the Fe leaching rate measured

in operando during a slow scan in O2-saturated electrolyte and from the measured ORR

activity, retained after 5,000 load cycles. The mass activity of Fe0.5-NH3 at 0.9 V vs. RHE in 0.1

M KOH was 4.6 A·g-1, among the highest reported mass activities for MNC catalysts thus far.56-

59

The high activity combined with high stability in alkaline medium of the most active FeNC

catalysts prepared via ammonia pyrolysis is thus promising for their application in Anion

Exchange Membrane Fuel Cell (AEMFC). However, the production of hydrogen peroxide (H2O2

in acid, HO2- in alkaline) on FeN4 sites or on the N-doped carbon surface during ORR might

limit the lifetime of Membrane Electrode Assemblies (MEAs) when operated at high current

density in AEMFC devices.60,61 Hydrogen peroxide, a reaction intermediate produced by the

2+2e- ORR pathway, is a strong oxidizer, which, at low pH and, in the presence of 3d transition

metal cations, undergoes Fenton’s reaction, producing reactive oxygen species (ROS) such as

·OH and ·OOH.45,62 In high-pH environment, rather than ·OH and ·OOH, it is the superoxide

radical (O2·-) that is expected from Fenton reactions between hydrogen peroxide and 3d metal

cations, or as an ORR intermediate. The superoxide radical has recently been shown to be

positively correlated with the decrease in ion-exchange capacity and conductivity of AEMs in

an operating AEMFC with in situ fluorescence spectroscopy. 16,17 In contrast, Fenton reactions

between FeNC catalysts and hydrogen peroxide have recently been shown to lead to strong

ORR deactivation when occurring in acid electrolyte, but to no ORR deactivation when

occurring in alkaline electrolyte.45 Nevertheless, HO2- might strongly bind on the Fe-based

active sites and, if not electro-reduced to water sufficiently fast, might poison the FeN4 active

sites via increased concentration of HO2- in the electrodes with operation time, especially in

the confined environment of micropores, where access of O2 to the active site is limited and,

vice versa, the removal of formed peroxide is limited .45

For those reasons, the aim of this work is to identify a stable and active PGM-free

electrocatalyst for the 2e- reduction of HO2- to OH- in alkaline electrolyte, and to study such a

77

catalyst in combination with the highly active and stable Fe0.5-NH3 catalyst mentioned

previously. With such a composite catalyst, we aim to decrease the amount of peroxide

produced during ORR in alkaline electrolyte below the level observed on Fe0.5-NH3. It is

important to remind here that, in contrast to the case in acidic medium, the ORR activity of

metal-free N-doped carbons is not negligible.63-65 Thus, metal-free nitrogen functional groups

can significantly contribute to the ORR in parallel with FeNx sites, in any FeNC catalyst. Such

metal-free N-groups typically catalyse the 2e- ORR to HO2-. The role of a co-catalyst that is

selected for its high activity toward hydrogen peroxide reduction reaction (HPRR) is thus to

scavenge hydrogen peroxide produced either at FeNx sites and, mostly, at metal-free N-sites,

in alkaline electrolyte. Manganese oxides are well known since decades for their high

heterogeneous catalytic activity toward H2O2 disproportionation, which has been

investigated in particular for propulsion of space and undersea vehicles.66,67 Manganese

oxides are also identified catalysts for ORR and for the oxygen evolution reaction (OER) in

alkaline electrolyte. The specific activity, stability and in operando structural changes of

various MnOx polymorphs have been investigated by numerous groups, and this research field

is still highly active.68-73 Recent reports however have revealed the restricted stability of Mn-

oxides for electrochemical energy-conversion applications, due to deep structural changes of

such oxides when moving the electrochemical potential from equilibrium (zero-current) in an

oxygen environment to either the OER or ORR regions.69,74,75 Surprisingly however, while

various transition-metal-oxides have previously been shown to work in tandem with carbon

catalysts (carbon catalysing O2 electro-reduction to HO2- and metal-oxides catalysing the

HPRR),76-78 the investigation of the specific HPRR activity of different Mn-oxide polymorphs

has only recently been investigated for itself by Ryabova et al.79,80 Among the different Mn-

oxide polymorphs that were investigated (α-Mn2O3, LaMnO3, β-MnO2, α-MnOOH and Mn3O4),

α-Mn2O3 was identified as the initially most active one, followed by β-MnO2. Structure-

Activity correlations were established, identifying the MnIV/MnIII redox potential as a key

descriptor to predict the onset potential for HPRR.

From these previous works, we selected four Mn-oxide polymorphs, i.e. three allotropes of

MnO2, namely α-MnO2, β-MnO2 and δ-MnO2, and one Mn-oxide with a different

stoichiometry, i.e. Mn2O3, which showed to be the most active for HPRR in Ref.79 The stability

of the same four MnOx materials was recently investigated by us in a parallel work using a

78

scanning flow cell coupled to inductively coupled mass spectroscopy (SFC/ICP-MS).75 That

work evidenced higher leaching rates of manganese during electrochemical load-cycling

conditions in O2-saturated 0.1 M NaOH electrolyte than in N2-saturated electrolyte. By

following the Mn leaching rate upon controlled addition of hydrogen peroxide in the

electrolyte, we could demonstrate that the increased Mn leaching during ORR is triggered by

hydrogen peroxide produced during the ORR. We also evidenced that Mn2O3 was the most

stable Mn-oxide in this series in the ORR potential range, with lowest dissolution rate of

manganese.

The objective of this work is to investigate the HPRR activity in alkaline electrolyte of α-MnO2,

β-MnO2, δ-MnO2, Mn2O3, and to investigate the effect of the Mn-oxides on the ORR activity,

selectivity and HPRR activity of composite catalysts prepared from a mix of Fe0.5-NH3 and each

one of these oxides. The HPRR activity was studied with a Rotating Disk Electrode (RDE) while

the ORR activity and selectivity were measured with a Rotating Ring Disk Electrode (RRDE).

The activity and performance was investigated in AEMFC for Fe0.5-NH3 and Mn2O3/ Fe0.5-NH3

cathodes.

Experimental

Synthesis of manganese oxides

The synthesis of pure α‒, β‒, δ‒MnO2 and α‒Mn2O3 phases was performed according to

previous literature reports, and as detailed in our recent study.75 A brief description is given

here for each phase. The α‒MnO2 phase was obtained by reducing KMnO4 in water and

fumaric acid at room temperature. The resulting gel was filtered, washed with ultrapure

water and dried.25 δ‒MnO2 was obtained via the reduction of KMnO4 in a mixture of water,

H2SO4 and ethanol, also at room temperature.81 To obtain β‒MnO2, the dried δ‒MnO2

powder was calcined at 450 °C. For preparing α‒Mn2O3, γ-MnOOH was prepared in a first

step, and then calcined at 550°C for 12 h.74 For the first step, γ-MnOOH was synthesized

dissolving Mn(CH3COO)2·4H2O and KMnO4 in deionized water, the solution refluxed for 12 h

and then washed with water and dried.82

79

Synthesis of Fe0.5-NH3

Fe0.5-NH3 was prepared via the sacrificial metal-organic framework method, as reported e.g.

in Ref.43 Commercial ZIF-8 (Basolite® Z1200, Sigma Aldrich), 1,10-phenantroline (≥ 99%, Sigma

Aldrich) and iron (II) acetate (≥99.99%, Sigma Aldrich) were homogeneously mixed in weight

ratios of 4/1 for ZIF-8/phenanthroline and 0.5 wt% of iron in the complete catalyst precursor, using

planetary ball milling (400 rpm, four cycles of 30 minutes milling with 5 minutes of pause

between each cycle). The catalyst precursor powder was then collected from the jar,

transferred in a quartz boat and inserted in a quartz tube. For the first pyrolysis, the oven was

pre-equilibrated at 1050°C for 2 h while a continuous flow of argon passed through the tube,

with the quartz boat and catalyst precursor inside the tube but outside the heating zone. After

2 h, the quartz boat was pushed in the heating zone in three steps of 30 seconds (“flash

pyrolysis”), with the help of an outer magnet and a quartz rod with magnet attached at one

end, located inside the tube. The catalyst precursor was pyrolysed in Ar at 1050°C for exactly

1 h, then the split-hinge oven opened and the tube removed and let to cooldown for 20 min,

still under Ar flow. A second pyrolysis was then performed, following the above-mentioned

procedure, but using pure NH3 instead of Ar, an oven temperature of 950°C and a pyrolysis

duration of only 5 min. The final product is collected directly from the quartz boat (Fe0.5-NH3).

Preparation of MnOx/Vulcan and MnOx/Fe0.5-NH3 composites

Each Mn-oxide phase was mixed with Vulcan carbon black in a weight ratio of 1/4 by manual

grinding. The Mn-oxides were mixed with Fe0.5-NH3 with the same procedure and weight

ratio.

Physical and morphological characterization

The crystalline structure of the four manganese oxides was verified using X-ray Powder

Diffraction (XRPD) with PANAlytical X’pert diffractometer in Bragg-Brentano configuration,

using a CuKα source (λ=1.5406 Å) in a 2θ range of 5-80° with a step size of 0.035°. The most

intense peaks were compared to the literature using PANalytical X’Pert Highscore Plus

(version 3.0e). To analyse the oxide morphology, scanning electron microscopy (SEM) was

applied (Hitachi S4800). Specific surface area of the four Mn-oxides was determined using the

Brunauer–Emmett–Teller (BET) method, cleaning previously the sample with flowing nitrogen

80

at 200°C for 5 h and then measuring the sorption of nitrogen at 77 K (Micromeritics ASAP

2020).

Electrochemical characterization

The ORR activity and HO2- production during ORR were measured with a RRDE setup (Pine

Instruments) in 0.1 M KOH, using a three-electrode configuration. The reference electrode

was a Pt-wire immersed in a H2-saturated electrolyte separated from the main compartment

by a fritted glass, acting as a reversible hydrogen electrode (RHE). The counter electrode was

a graphite rod immersed in the electrolyte, at a fixed distance from the working electrode.

The latter was a glassy carbon tip (5.6 mm of diameter) where the catalyst ink is deposited.

The peroxide produced at the working electrode was detected at the Pt ring. The HPRR activity

was measured in a RDE setup (Pine instruments) in the same conditions (glassy carbon tip of

5 mm diameter).

The inks were prepared adding in sequence 54 µL of Nafion® (5% perfluorinated resin

solution), 744 µL of ethanol and 92 µL of ultrapure water to 5 mg of FeNC, 20%-MnOx/C or

20%-MnOx/FeNC. The mixture was then sonicated for 1 h for homogenization. Then, 8.8 µL of

the dispersion was deposited on the RRDE-GC tip, while 7 µL are drop cast on RDE-GC tip, and

dried in air at room temperature, to obtain a catalyst loading of 0.2 mg cm-2 (total mass of

catalyst, supported catalyst or composite). For measuring the ORR activity of the Vulcan

carbon black support used for Mn-oxides, an aliquot of 7.0 µL was drop cast on the RRDE-tip

and 5.6 µL on the RDE-tip, in order to reach a loading of 0.16 mg cm-2. This corresponds to the

loading of Vulcan in the active layers of 20%-MnOx/C.

The electrochemical surface area was measured in N2-saturated electrolyte in the potential

range from 0 to 1 V vs. RHE (SP-300, BioLogic Potentiostat) for Fe0.5-NH3 while, for manganese

oxides, the lower potential limit was set to 0.45 V vs. RHE to minimize Mn leaching during the

experiments. The scan rate of 10 mV s-1 and rotation rate of 1600 rpm were applied for

electrochemical surface area assessment. To measure the ORR activity and selectivity, the

solution was saturated with O2 and the potential scanned in the potential range 0 - 1 V vs.

RHE with a scan rate of 1 mV s-1, while the potential of 1.2 V vs. RHE was applied to the Pt

ring.

81

To measure the activity toward HPRR, a RDE configuration was used, and the measurement

performed in N2-saturated 0.1 M KOH with 2mM H2O2, scanning the potential at 1 mV s-1 in

the range 0.45 – 1.0 V vs. RHE. During ORR and HPRR measurements, the scan rate of 1 mV s-

1 was found to be sufficiently low to neglect non-Faradaic currents.

Operando Scanning Flow Cell – Inductively Coupled Plasma Mass Spectroscopy

To evaluate the stability of the Mn2O3/Fe0.5-NH3 catalyst, a custom made polycarbonate SFC

coupled with ICP-MS is used to measure in situ the dissolution of manganese and iron during

electrochemical protocols. Electrochemical measurements have been carried out in a three

electrode configuration using a graphite counter electrode, Ag/AgCl reference electrode

(Metrohm) and glassy carbon as working electrode. Electrochemistry is applied with a

Potentiostat (Gamry, Reference 600) in 0.05 M NaOH (Merck, Certipur), reaching a controlled

pH (Mettler Toledo, SevenExcellence) of 11, used to convert the potentials of the Ag/AgCl to

the RHE.

AEMFC testing

The catalytic inks were prepared following the procedure described in Omasta et al., manually

grinding the catalyst (anode or cathode catalyst) and ETFE (ethylene tetrafluoroethylene)

powder ionomer (20 wt % with respect to the carbon contained in the catalyst)83 with 1 mL

of H2O and 9 mL of 1-propanol. The dispersion was then sonicated in an ice bath for 1 h and

then sprayed on a gas diffusion layer (Toray 60, 5 wt % PTFE wet-proofing) using an airbrush

(Iwata Eclipse HP CS). The obtained gas diffusion electrodes and ETFE membrane were then

soaked for 20 min in 1 M KOH, and this was repeated two times. The MEA was then assembled

in the single-cell fuel cell hardware using Teflon gaskets, with gasket thickness chosen to reach

25% compression. The AEMFC was operated using a Scribner 850e Fuel cell test system,

flowing H2/O2 at 1.0 L min-1 with a cell temperature of 60°C. The break-in was performed in

potentiostatic mode at 0.5 V, adjusting the relative humidity (RH) at both electrodes.34,35

AEMFC tests have been carried with three cathodes: a) Pt/C (40 wt% Pt, Johnson Matthey),

with a loading of 0.45 mgPt cm-2, b) Fe0.5-NH3 and c) Mn2O3/Fe0.5-NH3 (20 wt % Mn2O3) both

with a total loading of 1.5 mg cm-2. All AEMFC experiments have been carried out using PtRu/C

82

at the anode side (25 wt% of Pt and 15 wt% Ru, Johnson Matthey) with a total loading of 0.9

mgPGM cm-2.

Results and discussion

Physical and morphological characterization of Mn-oxides

The XRPD patterns shown in Figure 1a demonstrate the formation of pure or almost pure

manganese oxide phases. All diffraction peaks in each material could be assigned to the

targeted oxide phase, except for α-Mn2O3 pattern which shows minor peaks that can be

assigned to α-MnO2 (5% of α-MnO2 was estimated,75). Figure 1a also underlines the higher

crystallite dimension of α-Mn2O3 compared to MnO2 structures in general. The δ-MnO2 phase,

especially, shows only broad peaks. The high crystallinity of the present α-Mn2O3 material

may be related to the relatively high temperature of 550°C at which it was formed (see

Experimental section), while the higher crystallinity of β-MnO2 relative to δ-MnO2 may also

be explained from the viewpoint of the synthesis temperature, namely room temperature for

δ-MnO2 and then its calcination at 450°C to obtain β-MnO2.

83

Figure 1 Physical characterization of the four Mn-oxides with XRPD (a) and N2 physisorption

isotherms (b).

Figure 1b shows the N2 sorption isotherms of the pure Mn-oxide materials, revealing micro

and mesopores. The isotherm of δ-MnO2 clearly shows the higher amount of N2 adsorbed,

especially at low P/P0 and also shows a strong hysteresis closing at P/P0 of circa 0.4. The shape

of the isotherms of β-MnO2 and α-MnO2 are similar up to P/P0 = 0.7, with the one for β-MnO2

vertically translated by ca +10 cm3/g due to higher amount of N2 adsorbed at low P/P0. The

sorption isotherm for Mn2O3 clearly shows a much lower amount of N2 adsorbed relative to

all others, and identifies a non-porous material. The specific surface area of the oxides was

obtained with the BET method. The values are in increasing order 14 (α-Mn2O3), 98 (α-MnO2),

136 (β-MnO2) and 214 m2 g-1 (δ-MnO2). In line with the very different BET areas, SEM

micrographs evidence strong morphological differences among the different oxides.

Macroscopically, δ-MnO2 and β-MnO2 share a common particle morphology, with a size in

the range of 300-600 nm (low resolution SEM images, left column in Figure 2a, c), in line with

the fact that β-MnO2 was simply derived from δ-MnO2 via calcination. The high resolution

SEM micrographs show however that the δ-MnO2 surface has a higher roughness, with a sand-

10 20 30 40 50 60 70 80

Inte

nsity (

arb

. unit)

Scattering Angle, 2θ

α-Mn2O

3

α-MnO2

β-MnO2

δ-MnO2

0,0 0,2 0,4 0,6 0,8 1,0

0

25

50

75

100

125

150

175

200

225

250

Quantity

Adsorb

ed (

cm

³/g S

TP

)

Relative Pressure (p/p°)

δ-MnO2

β-MnO2

α-MnO2

α-Mn2O

3

84

rose morphology and chambers of 30-60 nm, which partially collapsed on the surface of the

β-MnO2 particles (Figure 2 b, d). The macroscopic morphology of α-MnO2 is different, with

large chunks of several µm in size, composed of numerous well-defined rod-shaped sub-

structures with ~130-300 nm length and 30-60 nm width, and also of even smaller objects

with less-defined particle-like morphology (Figure 2 e, f). Last, α-Mn2O3 has an interconnected

particle morphology, molten-like, with average particle dimension of ~150 nm with a smooth

surface (Figure 2 g, h), in line with the higher temperature of 550°c used for its synthesis.

The BET area ranking can be reasonably well explained from the surface roughness as

identified by the higher magnification SEM images. The highest BET area was measured for δ-

MnO2 (214 m2·g-1) with highest surface roughness (Figure 2b), then slightly decreased BET

area of 136 m2·g-1 for β-MnO2 due to partial collapse of surface tortuosity (Figure 2d), still a

high surface area of 98 m2·g-1 for α-MnO2 with smooth but small rod substructures (Figure 2f)

and last, drastically lowered BET area of 14 m2·g-1 due to interconnected and molten particles

morphology for Mn2O3 (Figure 2h).

85

a-b) δ-MnO2

c-d) β-MnO2

e-f) α-MnO2

g-h) α-Mn2O3

Figure 2. SEM micrographs of the different Mn-oxide polymorphs, in two different

magnifications (scale bar is 1.20 µm on the left, 300 nm on the right).

86

To further verify the apparent correlation between surface roughness and BET areas, we

calculated the expected size of Mn-oxide nanostructures from their measured BET areas and

assuming either a slab, rod, or spherical geometry (depending on samples) for the non-porous

oxide substructures assumed to contribute mainly to the surface area of the Mn-oxides (Table

1). For δ-MnO2 and β-MnO2, the assumed slab geometry results in estimated slab thickness

of ca 3 nm, in line with the very thin walls of the sand rose morphology observed in SEM. For

α-MnO2, the assumed rod morphology results in ca 9 nm width (diameter) for the rods, which

is significantly smaller than observed by SEM. Therefore, the surface of α-MnO2 seems to

result from a contribution of the observed rods but also from the smaller and less defined

objects seen in SEM (Figure 2f). Last, for α-Mn2O3, the estimated particle size of ca 95 nm is

in line with the smooth particles observed in SEM images (Figure 2h). Due to these large

morphological and BET differences, the apparent ORR and HPRR activity that will be measured

for a given loading of Mn-Oxides need to be normalized to the oxide area. This will be done

with the BET area values.

Oxide BET (m2 g-1) Density (g cm-3) Theoretical

specific surface area

Calculated

object size / nm

δ-MnO2 214 3.00 2/ (ρ·t ) (slab) 3.1

β-MnO2 136 5.03 2/ (ρ·t ) (slab) 2.9

α-MnO2 98 4.36 4/ (ρ·d) (rod) 9.4

α-Mn2O3 14 4.50 6/(ρ·d) (sphere) 95.2

Table 1. Measured BET areas of the Mn-oxides and estimated object size of the oxide

nanostructures forming the BET area. A slab geometry was assumed for δ-MnO2 and β-MnO2

and spherical geometry for α-Mn2O3.

Electrochemical characterization of Mn-oxides

The HPRR activity was then investigated in a RDE setup in 0.1 M KOH with 2 mM HO2-, for the

four Mn-oxide polymorphs supported either on Vulcan carbon black (Figure 3a) or on Fe0.5-

NH3 (Figure 3b).

87

0,5 0,6 0,7 0,8 0,9 1,0

-1,0

-0,5

0,0

0,5

1,0

0,5 0,6 0,7 0,8 0,9 1,0

-1,0

-0,5

0,0

0,5

1,0

0,5 0,6 0,7 0,8 0,9 1,0

-0,06

-0,04

-0,02

0,00

0,02

0,04

0,06

Cu

rre

nt

(mA

cm

-2)

Potential (V vs RHE)

δ-MnO2

β-MnO2

α-MnO2

Mn2O

3

Vulcan

δ-MnO2 β-MnO

2 α-MnO

2 α-Mn

2O

3

Cu

rre

nt

(mA

cm

-2)

Potential (V vs RHE)

Cu

rre

nt

de

nsit

y / m

A p

er

cm

2 o

xid

e

Potential (V vs RHE)

0,0

0,1

0,2

0,3

0,4

0,5

0,6

0,7

Slo

pe

of

no

rma

lize

d c

urr

en

t /

mA

pe

r c

m2 o

xid

e V

-1

a. b.

c. d.

Figure 3. Activity toward HPRR. a. Polarization curves of the four Mn-oxides supported on

Vulcan (20 wt% MnOx/C), normalizing the current with respect to the geometrical area of the

electrode; b. Polarization curves of the four Mn-oxides supported on Fe0.5-NH3 (20 wt%

MnOx) and for Fe0.5-NH3 alone; c. Polarisation curves of MnOx/C normalized by the BET

specific surface of the Mn-oxides; d. Histograms reporting the slopes near j = 0 obtained for

the BET-normalized polarization curves in c. All experiments have been carried out in 2mM

H2O2 + 0.1 M KOH electrolyte, with a three-electrode RDE configuration at a rotation rate of

1600 rpm, scan rate of 1 mV s-1 and a total catalyst loading of 0.2 mg cm-2.

The polarization curves seen in Figure 3a-b show a cathodic branch that can mainly be

assigned to HPRR (Eq. 1), and an anodic branch that can be assigned to the hydrogen peroxide

oxidation reaction (HPOR) (Eq. 2). In addition, a minor fraction of the cathodic current might

also be related to ORR, with O2 being produced in situ via non-electrochemical

disproportionation of peroxide on the Mn-oxide surface (Eq. 3).

88

HO2- + H2O + 2 e- → 3 OH- (1)

HO2- + OH- → O2 + H2O + 2e- (2)

2 HO2- → 2 OH- + O2 (3)

Considering the electrochemical reactions, it is important to note that reaction 2 is not the

reverse of reaction 1, and the HPRR and HPOR have therefore very different standard

equilibrium potentials, ca 1.73 V vs. RHE at pH 13 for HPRR, and ca 0.72 V vs. RHE at pH 13 for

HPOR.84 This expresses that, in the entire potential region between 0.72 and 1.73 V vs. RHE,

HO2- should be highly unstable and could, theoretically, be electro-oxidized to O2 and electro-

reduced to OH- at the same time. In practice however, a net oxidation current is generally

observed at potentials above 0.8-0.9 V, depending on the electrocatalyst investigated, and

assigned to HPOR.79,85,86 The oxidation reaction 2 may be kinetically much faster than the

reduction reaction 1, explaining why a net positive current is observed on all types of catalysts

in that potential region. A net reduction current is observed for MnOx/Vulcan only when the

HPOR becomes thermodynamically unfavored, i.e. at potentials lower than 0.83-0.85 V vs.

RHE (Figure 3a). In those conditions, peroxide anions at the electrode surface can no longer

be oxidized, and become available for HPRR. No clear redox process associated with Mn-

oxides is observed near the open circuit potential of 0.83-0.85 V vs. RHE, except for δ-MnO2,

as shown by cyclic voltammetry in a N2-saturated electrolyte free of peroxide (Figure S1).

Therefore, the experimental OCP values should be close to the equilibrium potential of

reaction (2). With 2 mM HO2- in the electrolyte and assuming an activity of 1 for O2 (i.e.

assuming that reaction 3 can saturate in O2 a very thin layer of electrolyte near the catalyst

surface), one can calculate from a Nernst equation that the equilibrium potential for reaction

(2) should be 0.801 V vs. RHE, assuming a pH = 13. This is close to the experimentally observed

values of OCP, ranging from 0.83 to 0.85 V vs. RHE for δ-MnO2, β-MnO2, α-MnO2 and Mn2O3,

in the order of increasing OCP values. As discussed below but already obvious in Fig. 3b, the

HPOR activity after normalization by the Mn-oxide area is much higher for Mn2O3 compared

to the other Mn-oxides, and this could explain why its OCP value is even closer to the

calculated Nernst equilibrium potential (calculated assuming O2 saturation at the catalyst

surface) than the other Mn-oxides.

89

Looking at the cathodic branch, assigned mainly to HPRR, the current density reaches

different values of plateau depending on the Mn-oxide considered, which is in contradiction

with a same diffusion-limited current density expected for the two-electron HPRR. The

plateau of current density is highest for Mn2O3, followed by β-MnO2 and δ-MnO2, and lowest

for α-MnO2. It was shown by Ryabova et al that this limiting current density can be assigned

entirely to a rate-determining chemical step in the HPRR on Mn3O4 and MnOOH (with the

limiting current density showing no variation with the rotation rate), while β-MnO2 was in an

intermediate situation in which the limiting current density slightly varied with rotation rate,

but much less than expected by Levich equation.79 In contrast, the limiting current density of

HPRR on Mn2O3 obeyed Levich equation and increased proportionally with the square root of

the electrode rotation rate. The authors concluded that for MnOOH, Mn3O4 (and β-MnO2,

partially) the rate of HPRR at low potential is not controlled by the diffusion of HO2- species

toward the electrode, but by a slow chemical step. The latter was proposed to be the chemical

dissociation of the O-O bond of an adsorbed HO2- on a MnIII site. Our experimental results are

in line with this, showing that other MnO2 allotropes behave similarly as β-MnO2 (Fig. 3a). In

the anodic branch in contrast (HPOR), no such effect was reported by Ryabova et al, and the

HPOR was limited by peroxide diffusion at potentials above 1 V vs. RHE for both Mn2O3 and

β-MnO2.79 Similar observation is made here on the present set of Mn-oxides, with similar

oxidation current densities at 1 V vs. RHE. The anodic branches at this potential have however

not yet reached the plateau. This may be due to the lower OH- concentration used in our work

compared to Ref.79 , 0.1 and 1.0 M respectively. As one can see in Eq. 2, the HPOR involves

OH- as a reactant, and this could therefore impact the HPOR reaction rate.

As discussed in the previous section, the different Mn-oxides possess very different BET areas,

α-Mn2O3 in particular having a lower BET area than the other materials. To assess their

intrinsic activity for HPRR/HPOR, we normalized the overall geometric current density by the

BET areas of the oxides. Since Vulcan carbon black has negligible activity for HPRR/HPOR

(Figure 3a), this normalization should correctly represent the intrinsic activity of the Mn-

oxides. Figure 3c highlights the superior intrinsic activity of Mn2O3 compared to the three

MnO2 allotropes. In this graph, Mn2O3 not only reaches higher reduction currents at low

potential (due to non-limitation by a slow chemical step) but also higher oxidation and

reduction currents near OCP, i.e. faster intrinsic electrochemical kinetics. To quantify the

90

latter, we fitted the polarisation curves in the region close to OCP (-30/+50 mV) with a straight

line and report the derivative dj/dE (units of mA cmoxide-2 V-1) in Figure 3d. The figure shows a

seven-fold higher intrinsic activity for Mn2O3 compared to β- and α-MnO2, and fourteen-fold

higher intrinsic activity compared to δ-MnO2. These results are in line with the five-fold higher

kinetic rate constants k2 and k3 for Mn2O3 vs. β-MnO2 reported in Ref.79, and describing the

kinetics of reaction (2).

In conclusion, Mn2O3 is shown to be intrinsically more active toward HPRR and HPOR than the

other Mn-oxides. Furthermore, our recent study showed that Mn2O3 is also more stable than

the other Mn-oxides in the ORR region, even after normalization for the oxide area.75

Therefore, both from activity and stability viewpoints, Mn2O3 is clearly identified as the most

promising Mn-oxide material for scavenging peroxide species produced during ORR. The full

potential of Mn2O3 vs. other Mn-oxides is however challenged by its generally lower BET area,

related to the high temperature calcination usually required to form this phase.87-89

In the next step, the Mn-oxides were dispersed on Fe0.5-NH3, in the same weight ratio as was

used for Vulcan, i.e. 20 wt% Mn-oxide on FeNC. The HPRR and HPOR polarisation curves of

Fe0.5-NH3 and of the four composite materials are shown in Figure 3b, confirming the

previously discussed feature of cathode limiting current density but also revealing a new

feature with respect to trends in the OCP values. The Fe0.5-NH3 catalyst alone, while showing

a much higher OCP value (ca 0.94 V vs. RHE) compared to all MnOx/Vulcan, has poor HPRR

and HPOR kinetics, as visible from its lower derivative dj/dE near OCP and lower cathodic

limiting current density at low potentials (Figs. 3a-b). The poor HPRR and HPOR activity of

Fe0.5-NH3 in alkaline medium, with clearly lower limiting cathodic current than expected for a

diffusion-limited HPRR at low potential, is a general phenomenon observed for FeNC catalysts

in both acid and alkaline electrolytes.90-96

The maximum HPOR and HPRR current densities for Fe0.5-NH3 in the potential range 0.5 – 1.0

V vs. RHE are ca 0.2 and 0.4 mA cm-2, respectively (Figure 3b). These values are significantly

increased for the MnOx/Fe0.5-NH3 composites. In the cathodic region, the limiting current

density increases in the same order than for the MnOx/Vulcan catalysts, i.e. α-MnO2 < δ-

MnO2 = β-MnO2 < Mn2O3, demonstrating the role of Mn-oxides in increasing the HPRR

activity of MnOx/Fe0.5-NH3 composites. In the anodic region of 0.95 – 1.0 V vs. RHE, the HPOR

current density is also higher for all MnOx/Fe0.5-NH3 composites relative to Fe0.5-NH3 (Figure

91

3b). The composites MnOx/Fe0.5-NH3 composites also show higher OCP values than the

corresponding MnOx/Vulcan layers, with OCP values ranging from 0.86 V vs. RHE to 0.91 V vs.

RHE for α-MnO2/Fe0.5-NH3 and Mn2O3/Fe0.5-NH3, respectively. This shift of OCP toward higher

values seems driven by the even higher OCP of Fe0.5-NH3 (0.94 V vs. RHE). In fact, due to the

negligible activity of Vulcan in MnOx/Vulcan layers, and the comparable loading of Fe0.5-NH3

in the pure Fe0.5-NH3 layer and in the composite MnOx/Fe0.5-NH3 layers (0.20 and 0.16 mg cm-

2, respectively), one may expect the polarisation curves of MnOx/Fe0.5-NH3 layers to behave,

in first approximation as the linear superimposition of the polarisation curves of MnOx/Vulcan

and Fe0.5-NH3, for a given MnOx material.

Overall, the results confirm the higher HPRR activity of Mn2O3/Fe0.5-NH3 relative to Fe0.5-NH3

itself, at potentials of 0.50 - 0.85 V vs. RHE. At the cathode of an AEMFC device, large amount

of peroxide will be produced at low potentials due to i) increased current density and ii) lower

selectivity (due to N-doped sites contribution to ORR at low potential, e.g. at 0.5 V, while such

sites do not contribute to the ORR at high potentials, e.g. at > 0.8 V).

The demonstration of the concept of scavenging HO2- formed during ORR with the addition

of Mn-oxides to Fe0.5-NH3 requires RRDE measurements in O2-saturated electrolyte. Figure

4a-b shows, for reference, the ORR activity, selectivity and Tafel plots of MnOx/Vulcan layers.

The ORR onset of MnOx/Vulcan layers is ca 0.9 V vs. RHE for Mn2O3, α- and β-MnO2 and ca 20

mV lower for δ-MnO2. The ORR polarisation curves show otherwise a similar shape and reach

comparable limiting current densities. The latter are close to the value expected for an overall

apparent four-electron ORR, and this is correlated by relatively low % HO2- detected at the

ring (Figure 4a, top). A closer detail at the % HO2- (or limiting current density) reveal however

fine differences, with decreasing % HO2- (or increasing absolute value of limiting current

density) in the order α-MnO2, δ-MnO2, β-MnO2, Mn2O3. This is in the exact same order as the

absolute value of limiting current density observed on MnOx/C layers in the HPRR

measurement (Figure 3a). This suggests that ORR on these MnOx proceeds via a 2e x 2e

pathway (subsequent 2 electron reduction on same MnOx sites) and/or via 2 electron

pathway to peroxide followed by catalytic HO2- decomposition (reaction 3), which could also

lead to apparent overall near four-electron ORR.

92

0.0 0.2 0.4 0.6 0.8 1.0-6

-5

-4

-3

-2

-1

0

α-MnO2

β-MnO2

δ-MnO2

α-Mn2O

3

Vulcan

Cu

rre

nt

De

ns

ity(m

A c

m-2)

Potential (V vs RHE)

0

10

20

30

40

50

% H

O2

-

0.01 0.1 1 10

0.70

0.75

0.80

0.85

0.90

0.95

Po

ten

tia

l (V

vs

RH

E)

Jk (mA cm

-2)

a. b.

0,1 1 10

0,70

0,75

0,80

0,85

0,90

0,95

1,00

0,0 0,2 0,4 0,6 0,8 1,0-6

-5

-4

-3

-2

-1

0

Fe0,5

-NH3 / δ-MnO

2

Fe0,5

-NH3 / β-MnO

2

Fe0,5

-NH3 / α-MnO

2

Fe0,5

-NH3 / α-Mn

2O

3

Fe0,5

-NH3

Cu

rre

nt

De

ns

ity

(m

A c

m-2)

Potential (V vs RHE)

0

5

10

15

20

% H

O2

-

c. d.

Po

ten

tial (V

vs R

HE

)

Jk (mA cm-2)

Figure 4 ORR activity and selectivity. RRDE measurements and ORR Tafel plots for the four

Mn-oxides supported on Vulcan carbon (20 wt% MnOx) (a, b). RRDE measurements and ORR

Tafel plots for the four Mn-oxides supported on Fe0.5-NH3 (20 wt% MnOx) and for Fe0.5-NH3

alone; (c, d). The total catalyst loading is 0.2 mg cm-2, except for Vulcan alone, with a catalyst

loading of 0.16 mg cm-2 mimicking the amount of Vulcan deposited for the MnOx/Vulcan

active layers. All experiments have been carried out in O2 saturated 0.1 M KOH electrolyte,

with a three-electrode RRDE configuration at a rotation rate of 1600 rpm, the scan rate is 1mV

s-1. No correction was done with respect to the ohmic drop.

93

Then, the ORR activity, selectivity and Tafel plots were recorded for Fe0.5-NH3 and the four

MnOx/Fe0.5-NH3 composite layers (Figure 4c-d). Fe0.5-NH3 alone shows high ORR activity of ca

1 mA cm-2, i.e. a mass activity of ca 5 A·g-1, similar to our recent study on Fe0.5-NH3 in alkaline

electrolyte.43 The ORR activity of the four MnOx/Fe0.5-NH3 composite layers is ca 30 % lower,

which is as expected from the dilution effect (such layers having only 0.16 mgFeNC cm-2, while

the reference FeNC layer has 0.2 mgFeNC cm-2) coupled to the presence of a more resistive

material. In the potential range of 0.7-0.8 V vs. RHE (mixed kinetic and O2 diffusion control),

the composite layers with α- and β-MnO2 show a distinctly higher mass-transport issues than

Fe0.5-NH3. This could be due to less homogeneous mix of these oxides and the FeNC, or to low

electronic conductivity of those oxide phases. In contrast, the composite layers with δ-MnO2

and Mn2O3 do not show this additional mass-transport issue (Figure 4c).97,98 Last but not least,

Figure 4c (top) demonstrates higher selectivity for all MnOx/Fe0.5-NH3 composite layers

relative to Fe0.5-NH3.

Then, the stability of the composite catalyst Mn2O3/Fe0.5-NH3 was evaluated in operando SFC-

ICP/MS. That particular Mn-oxide was selected for the composite catalyst due its higher

surface-normalized activity, as shown in this study and previously reported also, as well as its

slightly higher stability among this set of four oxides, as recently reported by Speck et al.75

Figure 5a (middle panel) in the region from 0 to ca 500 s shows similar low Mn dissolution

rate in Ar- and O2-saturated electrolytes when the potential is cycled for the first time

between 1 and 0.6 V, below ca 0.1 ng cm-2 s-1. This is due to the relative thermodynamic

stability of Mn2O3 in that restricted potential range. In contrast, when the potential is cycled

between 1 and 0 V (time between 500 and 1500 s in Figure 5a middle panel), the Mn leaching

rates increased when compared to the 1-0.6 V cycling and a higher Mn leaching rate is

observed in O2-saturated electrolyte vs. Ar-saturated electrolyte. This trend is similar to that

observed for Mn2O3/C.75 An AST protocol showed that the overall dissolution of manganese

in oxygen saturated electrolyte is higher than in Ar-saturated electrolyte (Figure 5b middle

panel), which is explained by the presence of hydrogen peroxide produced in O2-saturated

solution, as already reported by us.75 Concerning Fe dissolution, the catalyst does not show

evident dissolution peaks during the first slow scan between 1 and 0 V in oxygen saturated

electrolyte (Figure 5b, bottom panel). The dissolution rate of Fe is 0.05-0.08 ng cm-2 s-1,

comparable to the Fe leaching rate observed for Fe0.5-NH3 and previously reported by us.43

94

Figure 5b (middle panel) shows that the Mn dissolution rate from Mn2O3/Fe0.5-NH3 is

comparable to the dissolution rate previously reported for the same Mn-oxide supported on

Vulcan,75 with maximum leaching rates of 0.15-0.30 ng cm-2 s-1 in O2-saturated electrolyte

during the low-scan from 1 to 0 V vs. RHE.

Figure 5 Operando SFC-ICP/MS of the composite catalyst 20%-Mn2O3/Fe0.5-NH3 in alkaline

electrolyte saturated in Ar (blue curve) or O2 (red curve). a.) Durability test comprising an

initial high load CV followed by fast CVs and a final high load CV to evaluate the leaching of

manganese (middle plot) and iron (bottom plot). b) Relation between the potential and the

dissolution of manganese (middle graph) or iron (bottom graph) during low-load CVs (1-0.6 V

vs RHE) and high-load CVs (1-0 V vs RHE) in either Ar- (blue curve) or O2-saturated (red curve)

electrolyte.

Finally, the performance in AEMFC was evaluated and compared for Fe0.5-NH3, Mn2O3/Fe0.5-

NH3 and Pt/C cathodes. The initial activity of the non-PGM cathodes is comparable with that

of Pt/C. Fe0.5-NH3 shows a higher current density at 0.9 V (ca. 65 mA cm-2) with respect to the

composite catalyst, which shows ca. 52 mA cm-2, identical to Pt/C (value of 55 mA cm-2)

(Figure 6). At lower cell voltages, the non-PGM cathodes do not perform as well as the Pt/C

cathode, due to mass transport issues. The latter may be assigned to a more difficult path for

the oxygen to reach the active centres in such cathodes (possibly due to the microporous

nature of the carbon in FeNC), but also to local O2 starvation due to the much lower site

density in FeNC (ca 1.57 wt % Fe)43 compared to Pt/C (40 wt% Pt) due to the micropores.

Nevertheless, the absolute performance obtained with these non-PGM cathodes is promising,

0,0

0,1

0,2

0,3

0 500 1000 1500 2000

0,0

0,4

0,8

0,0

0,5

1,0

1,5

Mn

Dis

s (n

g c

m-2 s

-1)

O2 purged

Ar purged

Fe

Dis

s (n

g c

m-2 s

-1)

t (s)

E (

V v

s. R

HE

)b.

0.0

0.1

0.2

0.3

0 500 1000 1500 2000 2500 3000 3500 4000 4500 5000

0.03

0.04

0.050.060.070.080.09

0.0

0.5

1.0

1.5

Mn

Dis

s (n

g c

m-2 s

-1)

t (s)

Fe

Dis

s (n

g c

m-2 s

-1)

a.

E (

V v

s. R

HE

)a b

95

reaching at 0.6 V a current density of 1.43 A cm-2 and 1.25 A cm-2 for the Fe0.5-NH3 cathode

and the composite cathode, respectively. This is more than half the current density at 0.6 V

obtained with 0.4 mgPt cm-2 (2.41 A cm-2). The peak power density is 1.04 W cm-2 for Fe0.5-

NH3 and 0.98 W cm-2 for the Mn2O3/Fe0.5-NH3 composite, compared to 1.53 W cm-2 for Pt/C

(Figure 6b).

0 1000 2000 3000 40000

500

1000

1500

Pow

er

(mW

/cm

²)

I (mA/cm²)

0 1000 2000 3000 40000,0

0,2

0,4

0,6

0,8

1,0

I (mA/cm²)

Pt/C (CLPt

=0,45mg cm-2)

FeNC (CLFeNC

= 1,5mg cm-2)

FeNC/Mn2O

3 (CL

FeNC= 1,5mg cm

-2)

Pote

ntial (V

)

a. b.

Figure 6. AEMFC polarization curves measured for Pt/C, Fe0.5-NH3 and Mn2O3/Fe0.5-NH3

cathodes (red and blue curves respectively) using ETFE-based membrane and ionomer at a

cell temperature of 60°C and using gas flows of 1 L min-1. The anode was identical in all cases,

PtRu/C. The polarization curve was acquired at a scan rate of 10 mV s-1.

The higher current density at low potential (below 0.4 V) obtained using the composite non-

PGM cathode relative to Fe0.5-NH3 cathode can be attributed to the effect of the Mn-oxide

(helping in reducing peroxide into water at those potentials), or to the slightly lower loading

of FeNC in the layer (80% the loading In the pure FeNC layer), helping the mass-transport.

Conclusions

This study demonstrates that the addition of Mn-oxides to FeNC can help in scavenging

chemically and electrochemically hydrogen peroxide during ORR. The lower yields of

hydrogen peroxide observed with the composite MnOx/Fe0.5-NH3 catalysts compared to Fe0.5-

NH3 alone derives from the high HPRR kinetics of the four Mn-oxides, as determined by RDE

in 2 mM H2O2 alkaline electrolyte. Mn2O3 was shown to have higher HPRR activity than the

other three Mn-oxide polymorphs, once normalized per oxide surface area. The present

96

Mn2O3 material has however lower BET area than the other Mn-oxides, and further progress

could thus be made by synthesizing Mn2O3 powders with higher BET area. Mn2O3 was selected

for further stability study in operando SFC-ICP/MS, when combined with Fe0.5-NH3. The results

show similar dissolution rates for Mn and Fe from the composite 20wt% Mn2O3/Fe0.5-NH3

when compared to 20wt% Mn2O3/C and Fe0.5-NH3, respectively. Thus, no synergy effect is

observed regarding the metal leaching rates. The results also show similar trends as for the

separate materials with respect to the saturating gas, with higher Mn dissolution rates in O2

vs. Ar-saturated electrolytes, due the production of hydrogen peroxide that triggers the Mn

leaching. Finally the composite catalyst Mn2O3/Fe0.5-NH3 was tested at the cathode of an

AEMFC, showing comparable activity at 0.9 V to the Fe0.5-NH3 cathode, and also to the Pt/C

cathode. At high current density, lower mass transport is observed for the two non-PGM

cathodes with respect to the Pt/C cathode.

97

References

1 Staffell, I. et al. The role of hydrogen and fuel cells in the global energy system. Energy &

Environmental Science 12, 463-491, doi:10.1039/C8EE01157E (2019). 2 Stephens, I. E. L., Rossmeisl, J. & Chorkendorff, I. Toward sustainable fuel cells. Science 354,

1378-1379 (2016). 3 Cano, Z. P. et al. Batteries and fuel cells for emerging electric vehicle markets. Nature Energy

3, 279-289, doi:10.1038/s41560-018-0108-1 (2018). 4 Kongkanand, A. & Mathias, M. F. The Priority and Challenge of High-Power Performance of

Low-Platinum Proton-Exchange Membrane Fuel Cells. The Journal of Physical Chemistry

Letters 7, 1127-1137, doi:10.1021/acs.jpclett.6b00216 (2016). 5 Wu, J. et al. A review of PEM fuel cell durability: Degradation mechanisms and mitigation

strategies. Journal of Power Sources 184, 104-119, doi:10.1016/j.jpowsour.2008.06.006 (2008).

6 Gasteiger, H. A., Kocha, S. S., Sompalli, B. & Wagner, F. T. Activity benchmarks and requirements for Pt, Pt-alloy, and non-Pt oxygen reduction catalysts for PEMFCs. Applied

Catalysis B: Environmental 56, 9-35, doi:10.1016/j.apcatb.2004.06.021 (2005). 7 Majlan, E. H., Rohendi, D., Daud, W. R. W., Husaini, T. & Haque, M. A. Electrode for proton

exchange membrane fuel cells: A review. Renewable and Sustainable Energy Reviews 89, 117-134, doi:10.1016/j.rser.2018.03.007 (2018).

8 Thompson, S. T. et al. Direct hydrogen fuel cell electric vehicle cost analysis: System and high-volume manufacturing description, validation, and outlook. Journal of Power Sources 399, 304-313, doi:https://doi.org/10.1016/j.jpowsour.2018.07.100 (2018).

9 Henning, S. et al. Unsupported Pt-Ni Aerogels with Enhanced High Current Performance and Durability in Fuel Cell Cathodes. Angewandte Chemie 129, 10847-10850, doi:10.1002/ange.201704253 (2017).

10 Chung, D. Y., Yoo, J. M. & Sung, Y.-E. Highly Durable and Active Pt-Based Nanoscale Design for Fuel-Cell Oxygen-Reduction Electrocatalysts. Advanced materials 30, 1704123, doi:10.1002/adma.201704123 (2018).

11 Ding, J. et al. Nano-engineered intrapores in nanoparticles of PtNi networks for increased oxygen reduction reaction activity. Journal of Power Sources 374, 48-54, doi:https://doi.org/10.1016/j.jpowsour.2017.11.002 (2018).

12 Asset, T. et al. A Review on Recent Developments and Prospects for the Oxygen Reduction Reaction on Hollow Pt-alloy Nanoparticles. ChemPhysChem 19, 1552-1567, doi:10.1002/cphc.201800153 (2018).

13 Ishikawa, H. et al. Tomographic analysis and modeling of polymer electrolyte fuel cell unsupported catalyst layers. Journal of The Electrochemical Society 165, F7-F16 (2018).

14 Chattot, R. et al. Surface distortion as a unifying concept and descriptor in oxygen reduction reaction electrocatalysis. Nature materials 17, 827 (2018).

15 Banham, D. & Ye, S. Current Status and Future Development of Catalyst Materials and Catalyst Layers for Proton Exchange Membrane Fuel Cells: An Industrial Perspective. ACS Energy

Letters 2, 629-638, doi:10.1021/acsenergylett.6b00644 (2017). 16 Lefèvre, M., Proietti, E., Jaouen, F. & Dodelet, J.-P. Iron-Based Catalysts with Improved Oxygen

Reduction Activity in Polymer Electrolyte Fuel Cells. Science 324, 71-74, doi:10.1126/science.1170051 (2009).

17 Li, J. & Jaouen, F. Structure and activity of metal-centered coordination sites in pyrolyzed metal–nitrogen–carbon catalysts for the electrochemical reduction of O2. Current Opinion in

Electrochemistry 9, 198-206, doi:10.1016/j.coelec.2018.03.039 (2018). 18 Shao, M., Chang, Q., Dodelet, J.-P. & Chenitz, R. Recent Advances in Electrocatalysts for

Oxygen Reduction Reaction. Chemical Reviews 116, 3594-3657, doi:10.1021/acs.chemrev.5b00462 (2016).

98

19 Shao, Y., Dodelet, J.-P., Wu, G. & Zelenay, P. PGM-Free Cathode Catalysts for PEM Fuel Cells: A Mini-Review on Stability Challenges. Advanced materials 31, 1807615, doi:10.1002/adma.201807615 (2019).

20 Wang, X. X., Swihart, M. T. & Wu, G. Achievements, challenges and perspectives on cathode catalysts in proton exchange membrane fuel cells for transportation. Nature Catalysis 2, 578 (2019).

21 Martinez, U. et al. Progress in the Development of Fe-Based PGM-Free Electrocatalysts for the Oxygen Reduction Reaction. Advanced materials 31, 1806545, doi:10.1002/adma.201806545 (2019).

22 Chisaka, M., Ishihara, A., Ota, K.-i. & Muramoto, H. Synthesis of carbon-supported titanium oxynitride nanoparticles as cathode catalyst for polymer electrolyte fuel cells. Electrochimica

Acta 113, 735-740 (2013). 23 Chisaka, M., Ando, Y. & Muramoto, H. Facile Combustion Synthesis of Carbon-Supported

Titanium Oxynitride to Catalyse Oxygen Reduction Reaction in Acidic Media. Electrochimica

Acta 183, 100-106, doi:https://doi.org/10.1016/j.electacta.2015.03.211 (2015). 24 Ishihara, A. et al. Emergence of Oxygen Reduction Activity in Zirconium Oxide-Based

Compounds in Acidic Media: Creation of Active Sites for the Oxygen Reduction Reaction. The

Journal of Physical Chemistry C 123, 18150-18159, doi:10.1021/acs.jpcc.9b02393 (2019). 25 Zhang, H. et al. High-performance fuel cell cathodes exclusively containing atomically

dispersed iron active sites. Energy & Environmental Science 12, 2548-2558, doi:10.1039/C9EE00877B (2019).

26 Wagner, S. et al. Elucidating the Structural Composition of an Fe–N–C Catalyst by Nuclear- and Electron-Resonance Techniques. Angewandte Chemie 131, 10596-10602, doi:10.1002/ange.201903753 (2019).

27 Mineva, T. et al. Understanding Active Sites in Pyrolyzed Fe–N–C Catalysts for Fuel Cell Cathodes by Bridging Density Functional Theory Calculations and 57Fe Mössbauer Spectroscopy. ACS Catalysis, 9359-9371, doi:10.1021/acscatal.9b02586 (2019).

28 Kneebone, J. L. et al. A Combined Probe-Molecule, Mössbauer, Nuclear Resonance Vibrational Spectroscopy, and Density Functional Theory Approach for Evaluation of Potential Iron Active Sites in an Oxygen Reduction Reaction Catalyst. The Journal of Physical Chemistry C 121, 16283-16290, doi:10.1021/acs.jpcc.7b03779 (2017).

29 Kramm, U. I. et al. On an Easy Way To Prepare Metal–Nitrogen Doped Carbon with Exclusive Presence of MeN4-type Sites Active for the ORR. Journal of the American Chemical Society 138, 635-640, doi:10.1021/jacs.5b11015 (2016).

30 Jiang, R. et al. Edge-Site Engineering of Atomically Dispersed Fe–N4 by Selective C–N Bond Cleavage for Enhanced Oxygen Reduction Reaction Activities. Journal of the American

Chemical Society 140, 11594-11598, doi:10.1021/jacs.8b07294 (2018). 31 Jia, Q. et al. Spectroscopic insights into the nature of active sites in iron–nitrogen–carbon

electrocatalysts for oxygen reduction in acid. Nano Energy 29, 65-82, doi:https://doi.org/10.1016/j.nanoen.2016.03.025 (2016).

32 Lai, Q. et al. Metal–Organic-Framework-Derived Fe-N/C Electrocatalyst with Five-Coordinated Fe-Nx Sites for Advanced Oxygen Reduction in Acid Media. ACS Catalysis 7, 1655-1663, doi:10.1021/acscatal.6b02966 (2017).

33 Zhang, H. et al. Single Atomic Iron Catalysts for Oxygen Reduction in Acidic Media: Particle Size Control and Thermal Activation. Journal of the American Chemical Society 139, 14143-14149, doi:10.1021/jacs.7b06514 (2017).

34 Kumar, K. et al. Physical and Chemical Considerations for Improving Catalytic Activity and Stability of Non-Precious-Metal Oxygen Reduction Reaction Catalysts. ACS Catalysis 8, 11264-11276, doi:10.1021/acscatal.8b02934 (2018).

35 Zitolo, A. et al. Identification of catalytic sites for oxygen reduction in iron- and nitrogen-doped graphene materials. Nature materials 14, 937-942, doi:10.1038/nmat4367 (2015).

99

36 Zitolo, A. et al. Identification of catalytic sites in cobalt-nitrogen-carbon materials for the oxygen reduction reaction. Nature communications 8, 957, doi:10.1038/s41467-017-01100-7 (2017).

37 Kattel, S., Atanassov, P. & Kiefer, B. Stability, Electronic and Magnetic Properties of In-Plane Defects in Graphene: A First-Principles Study. The Journal of Physical Chemistry C 116, 8161-8166, doi:10.1021/jp2121609 (2012).

38 Yang, L. et al. Unveiling the high-activity origin of single-atom iron catalysts for oxygen reduction reaction. Proceedings of the National Academy of Sciences 115, 6626-6631, doi:10.1073/pnas.1800771115 (2018).

39 Holby, E. F., Wu, G., Zelenay, P. & Taylor, C. D. Structure of Fe–Nx–C Defects in Oxygen Reduction Reaction Catalysts from First-Principles Modeling. The Journal of Physical Chemistry

C 118, 14388-14393, doi:10.1021/jp503266h (2014). 40 Banham, D. et al. A review of the stability and durability of non-precious metal catalysts for

the oxygen reduction reaction in proton exchange membrane fuel cells. Journal of Power

Sources 285, 334-348, doi:10.1016/j.jpowsour.2015.03.047 (2015). 41 Choi, C. H. et al. Minimizing Operando Demetallation of Fe-N-C Electrocatalysts in Acidic

Medium. ACS Catalysis 6, 3136-3146, doi:10.1021/acscatal.6b00643 (2016). 42 Choi, C. H. et al. Stability of Fe-N-C Catalysts in Acidic Medium Studied by Operando

Spectroscopy. Angewandte Chemie 54, 12753-12757, doi:10.1002/anie.201504903 (2015). 43 Santori, P. G. et al. Effect of Pyrolysis Atmosphere and Electrolyte pH on the Oxygen Reduction

Activity, Stability and Spectroscopic Signature of FeNx Moieties in Fe-N-C Catalysts. Journal of

The Electrochemical Society 166, F3311-F3320, doi:10.1149/2.0371907jes (2019). 44 Goellner, V., Armel, V., Zitolo, A., Fonda, E. & Jaouen, F. Degradation by Hydrogen Peroxide of

Metal-Nitrogen-Carbon Catalysts for Oxygen Reduction. Journal of The Electrochemical

Society 162, H403-H414, doi:10.1149/2.1091506jes (2015). 45 Choi, C. H. et al. The Achilles' heel of iron-based catalysts during oxygen reduction in an acidic

medium. Energy & Environmental Science 11, 3176-3182, doi:10.1039/c8ee01855c (2018). 46 Wang, X. X., Prabhakaran, V., He, Y., Shao, Y. & Wu, G. Iron-Free Cathode Catalysts for Proton-

Exchange-Membrane Fuel Cells: Cobalt Catalysts and the Peroxide Mitigation Approach. Advanced materials 31, 1805126, doi:10.1002/adma.201805126 (2019).

47 Ferrandon, M. et al. Stability of iron species in heat-treated polyaniline–iron–carbon polymer electrolyte fuel cell cathode catalysts. Electrochimica Acta 110, 282-291, doi:https://doi.org/10.1016/j.electacta.2013.03.183 (2013).

48 Singh, H., Zhuang, S., Ingis, B., Nunna, B. B. & Lee, E. S. Carbon-based catalysts for oxygen reduction reaction: A review on degradation mechanisms. Carbon 151, 160-174, doi:https://doi.org/10.1016/j.carbon.2019.05.075 (2019).

49 Martinez, U., Komini Babu, S., Holby, E. F. & Zelenay, P. Durability challenges and perspective in the development of PGM-free electrocatalysts for the oxygen reduction reaction. Current

Opinion in Electrochemistry 9, 224-232, doi:https://doi.org/10.1016/j.coelec.2018.04.010 (2018).

50 Serov, A. et al. Nano-structured non-platinum catalysts for automotive fuel cell application. Nano Energy 16, 293-300 (2015).

51 Goellner, V. et al. Degradation of Fe/N/C catalysts upon high polarization in acid medium. Physical chemistry chemical physics : PCCP 16, 18454-18462, doi:10.1039/c4cp02882a (2014).

52 Proietti, E. et al. Iron-based cathode catalyst with enhanced power density in polymer electrolyte membrane fuel cells. Nature communications 2, 416, doi:10.1038/ncomms1427 (2011).

53 Yang, L. et al. Activity, Performance, and Durability for the Reduction of Oxygen in PEM Fuel Cells, of Fe/N/C Electrocatalysts Obtained from the Pyrolysis of Metal-Organic-Framework and Iron Porphyrin Precursors. Electrochimica Acta 159, 184-197, doi:https://doi.org/10.1016/j.electacta.2015.01.201 (2015).

100

54 Chenitz, R. et al. A specific demetalation of Fe–N4 catalytic sites in the micropores of NC_Ar + NH3 is at the origin of the initial activity loss of the highly active Fe/N/C catalyst used for the reduction of oxygen in PEM fuel cells. Energy & Environmental Science 11, 365-382, doi:10.1039/c7ee02302b (2018).

55 Herranz, J. et al. Unveiling N-protonation and anion-binding effects on Fe/N/C-catalysts for O2 reduction in PEM fuel cells. The journal of physical chemistry. C, Nanomaterials and interfaces 115, doi:10.1021/jp2042526 (2011).

56 Meng, H., Jaouen, F., Proietti, E., Lefèvre, M. & Dodelet, J.-P. pH-effect on oxygen reduction activity of Fe-based electro-catalysts. Electrochemistry Communications 11, 1986-1989, doi:10.1016/j.elecom.2009.08.035 (2009).

57 Yasuda, S., Furuya, A., Uchibori, Y., Kim, J. & Murakoshi, K. Iron–Nitrogen-Doped Vertically Aligned Carbon Nanotube Electrocatalyst for the Oxygen Reduction Reaction. Advanced

Functional Materials 26, 738-744, doi:10.1002/adfm.201503613 (2016). 58 Sarapuu, A. et al. Electrocatalysis of oxygen reduction by iron-containing nitrogen-doped

carbon aerogels in alkaline solution. Electrochimica Acta 230, 81-88, doi:https://doi.org/10.1016/j.electacta.2017.01.157 (2017).

59 Ratso, S. et al. Enhanced oxygen reduction reaction activity of iron-containing nitrogen-doped carbon nanotubes for alkaline direct methanol fuel cell application. Journal of Power Sources 332, 129-138, doi:https://doi.org/10.1016/j.jpowsour.2016.09.069 (2016).

60 Zhang, Y., Parrondo, J., Sankarasubramanian, S. & Ramani, V. Detection of Reactive Oxygen Species in Anion Exchange Membrane Fuel Cells using In Situ Fluorescence Spectroscopy. ChemSusChem 10, 3056-3062, doi:10.1002/cssc.201700760 (2017).

61 Parrondo, J., Wang, Z., Jung, M. S. & Ramani, V. Reactive oxygen species accelerate degradation of anion exchange membranes based on polyphenylene oxide in alkaline environments. Physical chemistry chemical physics : PCCP 18, 19705-19712, doi:10.1039/c6cp01978a (2016).

62 Nidheesh, P. V. Heterogeneous Fenton catalysts for the abatement of organic pollutants from aqueous solution: a review. RSC Advances 5, 40552-40577, doi:10.1039/c5ra02023a (2015).

63 Masa, J., Zhao, A., Xia, W., Muhler, M. & Schuhmann, W. Metal-free catalysts for oxygen reduction in alkaline electrolytes: Influence of the presence of Co, Fe, Mn and Ni inclusions. Electrochimica Acta 128, 271-278, doi:https://doi.org/10.1016/j.electacta.2013.11.026 (2014).

64 Yang, M., Liu, Y., Chen, H., Yang, D. & Li, H. Porous N-Doped Carbon Prepared from Triazine-Based Polypyrrole Network: A Highly Efficient Metal-Free Catalyst for Oxygen Reduction Reaction in Alkaline Electrolytes. ACS applied materials & interfaces 8, 28615-28623, doi:10.1021/acsami.6b09811 (2016).

65 Yang, W., Fellinger, T.-P. & Antonietti, M. Efficient Metal-Free Oxygen Reduction in Alkaline Medium on High-Surface-Area Mesoporous Nitrogen-Doped Carbons Made from Ionic Liquids and Nucleobases. Journal of the American Chemical Society 133, 206-209, doi:10.1021/ja108039j (2011).

66 Broughton, D. B., Wentworth, R. L. & Laing, M. E. Mechanism of Decomposition of Hydrogen Peroxide Solutions with Manganese Dioxide. II. Journal of the American Chemical Society 69, 744-747, doi:10.1021/ja01196a004 (1947).

67 Rophael, M. W., Petro, N. S. & Khalil, L. B. II — kinetics of the catalytic decomposition of hydrogen peroxide solution by manganese dioxide samples. Journal of Power Sources 22, 149-161, doi:https://doi.org/10.1016/0378-7753(88)87004-6 (1988).

68 Risch, M. et al. Redox Processes of Manganese Oxide in Catalyzing Oxygen Evolution and Reduction: An in Situ Soft X-ray Absorption Spectroscopy Study. The Journal of Physical

Chemistry C 121, 17682-17692, doi:10.1021/acs.jpcc.7b05592 (2017).

101

69 Rabe, M. et al. Alkaline manganese electrochemistry studied by in situ and operando spectroscopic methods – metal dissolution, oxide formation and oxygen evolution. Physical

Chemistry Chemical Physics 21, 10457-10469, doi:10.1039/C9CP00911F (2019). 70 Lima, F. H. B., Calegaro, M. L. & Ticianelli, E. A. Investigations of the catalytic properties of

manganese oxides for the oxygen reduction reaction in alkaline media. Journal of

Electroanalytical Chemistry 590, 152-160, doi:https://doi.org/10.1016/j.jelechem.2006.02.029 (2006).

71 Roche, I. & Scott, K. Effect of pH and temperature on carbon-supported manganese oxide oxygen reduction electrocatalysts. Journal of Electroanalytical Chemistry 638, 280-286, doi:https://doi.org/10.1016/j.jelechem.2009.10.030 (2010).

72 Lu, M., Kharkwal, S., Ng, H. Y. & Li, S. F. Y. Carbon nanotube supported MnO2 catalysts for oxygen reduction reaction and their applications in microbial fuel cells. Biosensors and

Bioelectronics 26, 4728-4732, doi:https://doi.org/10.1016/j.bios.2011.05.036 (2011). 73 Lehtimäki, M. et al. Targeted design of α-MnO2 based catalysts for oxygen reduction.

Electrochimica Acta 191, 452-461, doi:https://doi.org/10.1016/j.electacta.2016.01.070 (2016).

74 Ryabova, A. S. et al. Rationalizing the Influence of the Mn(IV)/Mn(III) Red-Ox Transition on the Electrocatalytic Activity of Manganese Oxides in the Oxygen Reduction Reaction. Electrochimica Acta 187, 161-172, doi:10.1016/j.electacta.2015.11.012 (2016).

75 Speck, F. D., Santori, P. G., Jaouen, F. & Cherevko, S. Mechanisms of Manganese Oxide Electrocatalysts Degradation During Oxygen Reduction and Oxygen Evolution Reactions. The

Journal of Physical Chemistry C, doi:10.1021/acs.jpcc.9b07751 (2019). 76 Malkhandi, S. et al. Electrocatalytic Activity of Transition Metal Oxide-Carbon Composites for

Oxygen Reduction in Alkaline Batteries and Fuel Cells. Journal of The Electrochemical Society 160, F943-F952, doi:10.1149/2.109308jes (2013).

77 Poux, T. et al. Dual role of carbon in the catalytic layers of perovskite/carbon composites for the electrocatalytic oxygen reduction reaction. Catalysis Today 189, 83-92, doi:https://doi.org/10.1016/j.cattod.2012.04.046 (2012).

78 Kéranguéven, G., Royer, S. & Savinova, E. Synthesis of efficient Vulcan–LaMnO3 perovskite nanocomposite for the oxygen reduction reaction. Electrochemistry Communications 50, 28-31, doi:https://doi.org/10.1016/j.elecom.2014.10.019 (2015).

79 Ryabova, A. S. et al. Study of Hydrogen Peroxide Reactions on Manganese Oxides as a Tool To Decode the Oxygen Reduction Reaction Mechanism. ChemElectroChem 3, 1667-1677, doi:10.1002/celc.201600236 (2016).

80 Bonnefont, A. et al. Challenges in the understanding oxygen reduction electrocatalysis on transition metal oxides. Current Opinion in Electrochemistry 14, 23-31, doi:10.1016/j.coelec.2018.09.010 (2019).

81 Quan, W. et al. New nanocomposite material as supercapacitor electrode prepared via restacking of Ni-Mn LDH and MnO2 nanosheets. Electrochimica Acta 247, 1072-1079, doi:10.1016/j.electacta.2017.07.010 (2017).

82 Cao, X. et al. A novel non-enzymatic hydrogen peroxide biosensor based on ultralong manganite MnOOH nanowires. Sensors and Actuators B: Chemical 147, 730-734, doi:10.1016/j.snb.2010.03.087 (2010).

83 Gonçalves Biancolli, A. L. et al. ETFE-based anion-exchange membrane ionomer powders for alkaline membrane fuel cells: a first performance comparison of head-group chemistry. Journal of Materials Chemistry A 6, 24330-24341, doi:10.1039/c8ta08309f (2018).

84 Pourbaix, M. Atlas of Electrochemical Equilibria in Aqueous Solutions NACE International (1974).

85 Poux, T. et al. Electrocatalysis of hydrogen peroxide reactions on perovskite oxides: experiment versus kinetic modeling. Physical Chemistry Chemical Physics 16, 13595-13600, doi:10.1039/C4CP00341A (2014).

102

86 Poux, T. et al. Electrocatalysis of hydrogen peroxide reactions on perovskite oxides: experiment versus kinetic modeling. Physical chemistry chemical physics : PCCP 16, 13595-13600, doi:10.1039/c4cp00341a (2014).

87 Zhang, X. et al. Synthesis of highly efficient Mn2O3 catalysts for CO oxidation derived from Mn-MIL-100. Applied Surface Science 411, 27-33, doi:https://doi.org/10.1016/j.apsusc.2017.03.179 (2017).

88 Saputra, E. et al. Shape-controlled activation of peroxymonosulfate by single crystal α-Mn2O3 for catalytic phenol degradation in aqueous solution. Applied Catalysis B: Environmental 154-

155, 246-251, doi:https://doi.org/10.1016/j.apcatb.2014.02.026 (2014). 89 Qiu, Y. et al. Morphology-conserved transformation: synthesis of hierarchical mesoporous

nanostructures of Mn2O3 and the nanostructural effects on Li-ion insertion/deinsertion properties. Journal of Materials Chemistry 21, 6346-6353, doi:10.1039/C1JM00011J (2011).

90 Jaouen, F. & Dodelet, J.-P. O2 Reduction Mechanism on Non-Noble Metal Catalysts for PEM Fuel Cells. Part I: Experimental Rates of O2 Electroreduction, H2O2 Electroreduction, and H2O2 Disproportionation. The Journal of Physical Chemistry C 113, 15422-15432, doi:10.1021/jp900837e (2009).

91 Muthukrishnan, A., Nabae, Y., Okajima, T. & Ohsaka, T. Kinetic Approach to Investigate the Mechanistic Pathways of Oxygen Reduction Reaction on Fe-Containing N-Doped Carbon Catalysts. ACS Catalysis 5, 5194-5202, doi:10.1021/acscatal.5b00397 (2015).

92 Choi, C. H. et al. Unraveling the Nature of Sites Active toward Hydrogen Peroxide Reduction in Fe-N-C Catalysts. Angewandte Chemie 56, 8809-8812, doi:10.1002/anie.201704356 (2017).

93 Shah, S. S. A. et al. Exploring Fe-Nx for Peroxide Reduction: Template-Free Synthesis of Fe-Nx Traumatized Mesoporous Carbon Nanotubes as an ORR Catalyst in Acidic and Alkaline Solutions. Chemistry – A European Journal 24, 10630-10635, doi:10.1002/chem.201802453 (2018).

94 Ramaswamy, N. & Mukerjee, S. Fundamental Mechanistic Understanding of Electrocatalysis of Oxygen Reduction on Pt and Non-Pt Surfaces: Acid versus Alkaline Media. Advances in

Physical Chemistry 2012, 17, doi:10.1155/2012/491604 (2012). 95 Osmieri, L., Escudero-Cid, R., Monteverde Videla, A. H. A., Ocón, P. & Specchia, S. Application

of a non-noble Fe-N-C catalyst for oxygen reduction reaction in an alkaline direct ethanol fuel cell. Renewable Energy 115, 226-237, doi:https://doi.org/10.1016/j.renene.2017.08.062 (2018).

96 Ramaswamy, N. & Mukerjee, S. Influence of Inner- and Outer-Sphere Electron Transfer Mechanisms during Electrocatalysis of Oxygen Reduction in Alkaline Media. The Journal of

Physical Chemistry C 115, 18015-18026, doi:10.1021/jp204680p (2011). 97 Bhide, V. G. & Dani, R. H. Electrical conductivity in oxides of manganese and related

compounds. Physica 27, 821-826, doi:https://doi.org/10.1016/0031-8914(61)90079-9 (1961). 98 Klose, P. H. Electrical Properties of Manganese Dioxide and Manganese Sesquioxide. Journal

of The Electrochemical Society 117, 854-858, doi:10.1149/1.2407656 (1970).

103

104

Mechanisms of Manganese Oxide Electrocatalysts

Degradation during Oxygen Reduction and Oxygen

Evolution Reactions

Florian D. Speck1,2,*, Pietro G. Santori3, Frédéric Jaouen3,*, Serhiy Cherevko1,*

1. Helmholtz‒Institute Erlangen‒Nürnberg for Renewable Energy (IEK‒11),

Forschungszentrum Jülich, Egerlandstr. 3, 91058 Erlangen, Germany

2. Department of Chemical and Biological Engineering, Friedrich‒Alexander‒Universität

Erlangen‒Nürnberg, Egerlandstr. 3, 91058 Erlangen, Germany

3. Institut Charles Gerhardt Montpellier, UMR 5253, CNRS, Université Montpellier,

ENSCM, Place Eugène Bataillon, 34095 Montpellier cedex 5, France

Link: https://pubs.acs.org/doi/abs/10.1021/acs.jpcc.9b07751

105

Abstract

Anion exchange membrane fuel cells and electrolyzers offer a unique opportunity of using

non-noble metal electrocatalysts for catalyzing the sluggish oxygen reduction and oxygen

evolution reactions (ORR, OER). In recent years, various Mn-based oxides were identified as

promising catalysts for both reactions. While electrocatalytic activity of such oxides is well

addressed, their stability is still to be proven. Herein, we investigate the stability of four main

manganese oxide allotropes by following their Mn dissolution rate in operando ORR and OER

conditions. Using an electrochemical on-line inductively coupled plasma mass spectrometer

(on-line ICP-MS), we uncover unexpected instability of this class of catalysts, with different

degradation mechanisms identified under OER and ORR conditions. The reason for their

degradation is shown to be related to the production of hydrogen peroxide species on

manganese oxides during ORR. Furthermore, we discuss how limits in thermodynamically

stable windows of each Mn oxidation state leads to increased dissolution during applications

with high potential perturbations, i.e. change in load, start/stop conditions and especially

bifunctional application. Therefore we recommend clear guidelines for future development

of platinum group metal free electrocatalysts for affordable alkaline energy conversion

technologies.

Introduction

Electrochemical energy conversion devices such as water electrolyzers (EL) and hydrogen fuel

cells (FC) hold the promise for a closed electrocatalytic water cycle with high round-trip

efficiency, allowing the storage of renewable energy and electricity generation, respectively.

The state of the art in R&D and commercialization are proton exchange membrane fuel cells

(PEM-FC) and electrolyzers (PEM-EL).1-6 In both of these technologies, the electrochemical

reactions at the positive electrode have been identified as the most challenging ones for

electrocatalysis.7-10 In PEM-FCs, the oxygen reduction reaction (ORR) is the main limiting

factor, both in overcoming ORR kinetic restrictions of the four electron process, as well as

upscaling the technology due to the high price of Pt.7,11,12 To overcome the latter, anion

exchange membrane fuel cells (AEM-FCs) have been investigated with increasing attention in

the last years, as they offer a more favorable environment for platinum group metal (PGM)

106

free catalysts, both in terms of thermodynamic stability and ORR activity.13,14 Several recent

research papers have also reported promising replacement of PGM- by PGM-free cathodes

reaching peak powers > 1 W cm−2.15,16 Furthermore, continuous progress in membrane

development has drastically increased the initial power performance of AEM-FC, now at par

or even higher than that reached with PEM-FCs, at a comparable total PGM loading in the

entire membrane-electrode-assemblies.17,18 This advantage could lead to a significant

reduction in the cost of FC technology if PGM loading at the anode of AEM-FC can be

significantly reduced or zeroed in the future. For electrolyzers, state of the art PEM-EL faces

the same challenges as encountered at the positive electrode of PEM-FC, namely high

overpotential to drive the OER, even on the best PGM-based catalysts, and significant cost

associated with the use of Ir oxide (IrOx) and IrRuOx catalysts as well as titanium bipolar plates.

Here the price of the desired H2 production is mainly governed by the scarcity of relatively

stable and active IrOx catalysts, needed to drive the anodic side reaction.19,20 Similar to the

FC, switching to alkaline or high pH environment opens a broader selection of catalysts for

the oxygen evolution reaction (OER), including numerous PGM-free and Earth-abundant

metal oxides. While alkaline electrolyzers are already commercialized and compete for the

market of low-temperature electrolyzers with PEM-EL, they suffer from higher internal cell

resistance, less pure H2 produced and difficulty to electrochemically pressurize the hydrogen.

For those reasons, AEM-EL are perceived as a promising technology that could unify the

technical advantages of PEM-EL with the cost advantages of AEM-EL.

While the catalyst cost seems to not be an issue for AEM technology, the aforementioned

sluggishness of the 4-e− ORR and OER processes still needs urgent optimization to compete

with PEM technology. To achieve this, research efforts are focused on a better understanding

of electrocatalysis during ORR and OER on PGM-free catalysts in alkaline media. Numerous

PGM-free catalysts have been investigated for catalyzing the ORR or OER in alkaline

electrolytes. Ranging from metal oxides to perovskites for OER and ORR,21-23 all the way to

advanced single atom M-N-C catalysts for ORR.21 Manganese is considered as an interesting

candidate since in aqueous media at pH 13, thermodynamically stable oxidation states of the

solid phase in equilibrium include 0, +II, +II/+III, +III, +IV. In solution, stable ions reaching

oxidation states of +VI and +VII are present.13,24 These thermodynamic considerations,

however, neglect adsorbates and intermediates during electrocatalysis, as well as transient

107

changes of the surface during potential cycling. Nevertheless, due to its abundance and

associated low cost, rich redox chemistry and inspired by the key role of this element in

Nature’s photocatalytic water splitting systems, Mn-based materials have attracted a special

attention.20,25-28 Such materials have also been considered for bifunctional application, i.e. as

a single device capable of both FC and EL operation, and for rechargeable metal-air

batteries.28-30

Surprisingly, while Mn-oxides have been investigated in numerous studies for their activity in

alkaline electrolyte toward ORR, OER, bifunctional ORR/OER,31,32 as well as toward hydrogen

peroxide reduction or decomposition (e.g as tandem ORR catalysts with N-C or Fe-N-C

catalysts)33,34, their stability has been hitherto relatively poorly investigated. While there are

few reports, investigations on dissolution stability of Mn are rare, with bulk dissolution

typically measured ex situ.35 For OER in acidic environment, a stable potential window for

MnOx was recently derived from UV-vis spectroscopy measurements,36 with faradaic

efficiencies for OER measured to be near unity in that potential window, pointing towards

minimal corrosion current.23 In contrast, intense Mn dissolution starting at ca. 1.20 VRHE was

observed for OER in alkaline electrolyte, and the onset of dissolution was related to the

thermodynamic potential of MnO4−

(aq) formation from MnO2.37 Adverse studies reported the

stabilization of Mn in perovskite materials by optimizing lattice parameters and excluded the

possibility of Mn dissolution in alkaline electrolyte during OER in rotating ring disk electrode

(RRDE) experiments.20,38 However, most stability investigations usually rely on long term

activity assessments and ex-situ physical methods. As J. Kibsgaard and I. Chorkendorff

recently commented, apparent stable OER activity of metal-oxides is not a proof of material

stability, since the continuous leaching of metal cations from an oxide surface may lead to a

constantly changing oxide surface or even increased surface area, the leached metal ions

exposing a new surface.6 Therefore, before moving towards application of Mn

electrocatalysts in FC and EL, we find it crucial to gain improved understanding about the

thermodynamic implications of Mn redox reactions in operando, as well as the effect of the

catalysis reaction itself, which happens outside of thermodynamic equilibrium and can

involve intermediate species such as HO2− not accounted for in classical thermodynamic

considerations.

108

Herein, to uncover possible degradation mechanisms of Mn oxides in AEM-FC, AEM-EL and

bifunctional FC/EL applications, we synthesized four manganese oxide allotropes commonly

used in the literature, viz. α-Mn2O3, α-MnO2, β-MnO2, and δ-MnO2, deposited them on

carbon support, and investigated their dissolution rates during accelerated stress tests (ASTs)

designed either for FC or EL operation modes in an alkaline medium. We also specifically

investigated the effect of O2 or hydrogen peroxide present in the electrolyte on Mn

dissolution rates and compared it to the dissolution rates acquired in an inert-gas saturated

electrolyte. During ORR, we reveal that the in operando formation of HO2− on MnOx is the

main driver of Mn dissolution, while during OER, the main driver is the MnO2/MnO4−

(aq)

thermodynamic redox potential stability limitation. Lastly, apparent differences in the

dissolution behavior between the four allotropes were ruled out by normalizing to the

catalysts’ specific surface areas. We conclude that none of the MnOx allotropes is

satisfactorily stable, neither in ORR nor in OER operation modes in alkaline media and that

strict guidelines to verify material stability should be especially followed for PGM free

materials.

Experimental

Catalyst Synthesis and Preparation

The preparation of four different, phase-pure, manganese oxides was targeted, namely the

α‒, β‒, δ‒MnO2 and α‒Mn2O3 phases. Their synthesis was performed according to the

literature. In short, α‒MnO2 was obtained by reducing KMnO4 (3.16 g) in a mixture of water

(200 mL) and fumaric acid (0.78 g), kept under stirring at room temperature for 30 minutes.

The resultant gel is settled for 1 h and then filtered, washed with ultrapure water and dried

to yield α‒MnO2.39 The reduction of KMnO4 (0.395 g) to δ‒MnO2 was carried out under

stirring over 3 h in a mixture of water (80 mL), H2SO4 (96%, 70.2 µL) and ethanol (3 mL).40 The

dried powder was calcined at 450 °C to yield β‒MnO2.41 α‒Mn2O3 was obtained by calcining

γ‒MnOOH at 550 °C.14 The latter was obtained via KMnO4 (0.2 g) reduction together with

Mn(CH3COO)2·4H2O (1.2 g) in water (150 mL) over 12 h, keeping the mixture under stirring

and refluxed.42 Subsequently, the four manganese oxides were dispersed on carbon black

(CABOT, Vulcan® XC72R), with a ratio of 1:5, placed in zirconia jars with 100 balls (0.4 g per

ball) and mixed at low energy ball milling (Fritsch, Pulverisette 7) for 10 minutes at 200 rpm.

109

For scanning flow cell (SFC) measurements, the catalyst materials were dropcast onto a glassy

cacrrbon (GC) plate (HTW, Sigradur®). Therefore 5 mg of catalyst were dispersed firstly in

water (Merck, Milli-IQ® IQ 7000, 18.2 MΩ, 1176 μl) using a sonication horn (Branson, SFX 150)

for 10 minutes at 4 second intervals of 40% amplitude while cooling in an ice bath.

Afterwards, Nafion® perfluorinated resin solution (Sigma Aldrich, 5 wt.%, 28.6 μL) with

isopropanol (294 μL Merck, Emsure®) was added and sonicated again. This concentrated ink

was further diluted by half using equal parts of the ink and water and subsequently the pH

was adjusted to 11 using NaOH (Merck, Suprapur®, 0.05 M). After a final sonication step,

catalyst spots were dropcast onto a GC plate from 0.2 μL of the ink (c = 3.34 g L−1,

mcat/mion = 4) yielded spots with a radius of 0.65 mm ± 0.03 mm (≈ 50 μg cm−2), which was

measured for each spot using a laser scanning microscope (Keyence, VK-X250) and used for

normalization of all electrochemical and dissolution data towards geometric surface area.

On-line ICP-MS

For all stability investigations we used an inductively coupled plasma mass spectrometer

(ICP-MS) (Perkin Elmer, NexION 350x) for the in situ detection of dissolved manganese

species, by coupling it to the electrolyte outlet of a custom made polycarbonate SFC. On the

electrolyte inlet (angled 60 ° to the outlet) the SFC was connected to a graphite counter

electrode (Sigma Aldrich, 99.995% trace metal basis). A third capillary channel connected the

reference electrode (Metrohm, Ag/AgCl) closely to the working electrode surface. Contact

with the working electrode was made with a translation stage (Physik Instrumente, M-403).

A Potentiostat (Gamry, Reference 600) was used to employ all electrochemical protocols

during dissolution measurements. The purged electrolytes flow rate was controlled and

regularly calibrated by the peristaltic pump of the ICP-MS (Elemental Scientific, MP2 Pump).

The sensitivity of the ICP-MS towards dissolved Mn ions was calibrated daily from Mn

calibration standards (Merck, Centripur®) and dissolution rates are normalized to the

geometric surface area of each catalyst spot. For further information on the experimental SFC

ICP-MS setup, please refer to our previous reports.43

As an electrolyte NaOH (Merck, Suprapur) was dissolved in water to 0.05 M. For individual

experiments, an addition of small amounts of 30wt.% H2O2 (Merck) was achieved by multiple

dilution steps. The pH was controlled (Mettler Toledo, SevenExcellence) and used to convert

110

the potentials of the Ag/AgCl to the reversible hydrogen electrode (RHE) using

equation (Eq. 1).

ERHE =Eapplied + EAg/AgCl + 0.0591 × pH (Eq. 1)

RRDE Experiments

Rotating ring disk electrode (RRDE) analysis was performed in 0.05 M NaOH employing a

conventional three-electrode setup, where all electrodes were connected to a Potentiostat

(BioLogic, SP-300). Using a RHE, based on a Pt-wire immersed in H2-satured electrolyte inside

a fritted glass, as the reference electrode and a graphite rod as the counter electrode.

The ink is prepared by mixing 5 mg of the catalyst with 54 µL of Nafion®, 744 µL ethanol and

92 µL ultrapure water and subsequent sonication for 1 h in an ice bath. Drop casting of 8.8 µL

dispersion onto the glassy carbon RRDE yielded a catalyst loading of 0.25 mg cm−2, after

drying the droplet at room temperature. The active surface area is determined applying cyclic

voltammograms (CVs) in N2-saturated electrolyte in a range potential of 1 – 0.45 V (Figure S5),

with a scan rate of 5 mV s–1 and a rotation rate of 1600 rpm (identical for all the experiments).

Afterwards, activity is evaluated in O2-saturated electrolyte in a potential range of 1 – 0 V, at

a scan rate of 5 mV s−1. Lastly, by applying 1.2 VRHE to the Pt ring, the amount of HO2– produced

during the oxygen reduction reaction (ORR) could be obtained.

Physical Analysis

X-ray Photoelectron Spectroscopy (XPS) was conducted using a PHI Quantera II scanning X-ray

microprobe. Spectra of catalyst spots (90 μg cm−2) on GC were acquired using Al‒Kα

irradiation of a 200 μm diameter spot at 50 W and 15 kV. Survey scans at a step size of 1 eV

and 280 eV pass energy as well as high resolution narrow scans in 0.125 eV steps at 140 eV

pass energy were collected for 500 ms dwell time per step. Data was analyzed in CasaXPS

(V.2.3.18) using instrument specific relative sensitivity factors, Shirley type backgrounds and

a binding energy scale calibrated to the adventitious carbon peak at 284.8 eV.

X-ray Diffraction (XRD) patterns were measured on a PANAlytical X’pert diffractometer in

Bragg-Brentano configuration, using CuKα source (λ = 1.5406 Å) in a 2θ range of 5 – 80 ° in a

step size of 0.035 °. The most intense reflexes are compared in PANalytical X’Pert Highscore

Plus (version 3.0e) to literature data taken from α-MnO2,44 β-MnO2,45 δ-MnO2,46 Mn2O347.

The same literature data was also taken to simulate the ideal XRD patterns (FWHM = 0.1 °2ϴ)

111

in Figure 2a, using Mercury 4.0.0. Crystallite sizes D were estimated using the transformed

Scherrer Equation (Eq. 2) using a shape factor K = 0.9.

� �

��

Δ2Θ���� Θ

(Eq. 2)

Here, Δ2ϴFWHM is the main reflex’ width at 0.5 normalized intensity in radiants and ϴ its

position.

A Hitachi S4800 SEM was used to acquire images of the materials morphology.

The catalysts porosity was determined from N2 sorption isotherms at 77 K using the BET

Method. Isotherms were recorded on all catalysts using a Micromeritics ASAP 2020

instrument.

Results

Catalyst characterization

Figure 1. Structural and compositional analysis of the MnOx materials. (a) XRD patterns of the

as synthesized α‒Mn2O3, α‒MnO2, β‒MnO2 and δ‒MnO2 materials compared to calculated

positions of the reflexes for the targeted allotropes, using Mercury (4.0.0). (the assigned

reflexes can be found in Figure S1) (b) XPS analysis of the dropcasted inks, containing Vulcan®

Carbon and Nafion® is shown in form of the Mn 3s, Mn 2p regions as well as partial C 1s and

O 1s regions, selected to highlight C-F and O-Mn interactions, respectively.

112

The synthesized MnOx materials were investigated using X-ray diffraction (XRD) and X-ray

photoelectron spectroscopy (XPS) to confirm their crystal structure and oxidation state.

Figure 1 shows a summary of the physical catalyst analysis.

For the studied MnOx materials, the position of the main XRD reflexes in the recorded

patterns (Figure 1a) is in line with literature data,44,45,48,49 confirming successful synthesis of

the targeted phases. As can be seen in Fig. 1a, some minor reflexes in the XRD pattern of the

sample labelled as α-Mn2O3 can be assigned to α-MnO2. Using MAUD software, we quantified

that 5% α-MnO2 co-exists within the sample labelled as α-Mn2O3. Further, the crystallinity,

estimated from the broadness at full width half maximum (FWHM) of the most intense

diffraction reflexes, decreases in the order α-Mn2O3 > α-MnO2 > β-MnO2 >> δ-MnO2. Using

the Scherrer equation we estimate crystal sizes (D) of 34.9 > 31.3 > 20.1 >> 1.4 nm.

XPS survey scans of the MnOx/C catalyst ink deposited on glassy carbon (GC), presented in

Figure S2, show the presence of C, F, O, Mn and S at the catalyst surface while only C and O

are detected at the GC reference surface. The F and S signals seen on the catalyst surface

originate from the Nafion® ionomer in the dried ink while the C signal originates from the

Vulcan® (XC72R) support as well as from the underlying GC plate. Therefore, Figure 1b focuses

on narrow, high resolution scans of the Mn2p, Mn3s as well as partial C1s and O1s regions to

identify the surface composition and oxidation state of MnOx. Both Mn regions of the Mn2O3

sample clearly show a higher Mn atomic concentration. Further, a more defined peak in the

O1s signal at 529.8 eV compared to the other three MnO2 materials, is assigned to O—Mn

interaction as a feature next to the adventitious O1s peak that is also observed on the pure

GC background spectra (GC bg)50. Most indicative of the lower average surface oxidation state

of Mn for the Mn2O3 sample is the increased multiplet splitting (5.49 eV) of the Mn3s scan

and the shift of the Mn2p1/2 and Mn2p3/2 peaks to lower binding energies (641.7 eV and

653.3 eV, respectively) compared to the MnIVOx. Supporting the XRD findings (co-existence of

5% α-MnO2 with 95% Mn2O3 in the material labelled Mn2O3), we find 25% MnIV and 75% MnIII

on the catalyst surface, using CasaXPS peak fitting capabilities employing instrument specific

relative sensitivity factors. This increased MnIV content as compared to XRD is to be expected

if the higher oxidation state occurs mainly at the surface due to post synthesis effects, since

XRD only gives us the bulk composition. In contrast, the three MnIV‒oxide materials show

superimposed Mn3s and 2p regions, with a 3s multiplet splitting of 4.96 eV and Mn2p3/2 peak

113

at 642.3 eV.24 The O1s peak of each MnIV-oxide at 529.8 eV appears only as a shoulder to the

adventitious O1s peak. The C1s region displays a specific feature at 291.3 eV in all catalyst

spots, which corresponds to Nafion® (i.e. F—C)51 carbon which is not present on the pure GC

scan.

The morphology of the catalysts was investigated using scanning electron microscope (SEM)

analysis, shown in Figure S3. The various synthesis routes selected to prepare the phase-pure

MnOx materials resulted in different morphologies. All materials seem to consist of sintered

particles, however, the surface roughness increases from compact material (α-Mn2O3) to

more platelet looking sheets (δ-MnO2) which correlates with the crystal size estimated by

XRD. Using the Brunauer Emmett Teller (BET) equation applied to N2 sorption isotherms

gathered of the unsupported MnOx materials, we confirm that the surface area is inversely

correlated with the average crystal size estimated from the XRD patterns.

(α-Mn2O3; 14 m2 g −1 > α-MnO2; 98 m2 g −1 > β-MnO2; 136 m2 g −1 > δ-MnO2; 214 m2 g −1), see

Figure S4. Since the BET area is not necessarily a direct descriptor of the electrochemical

surface area (ECSA), cyclic voltammetry (CV) in N2-saturated electrolyte was performed on

the MnOx/C materials, with a RRDE setup, with data presented in Figures S5 and S6. Most

notably, Figure S5 shows the initial changes occurring during the first six CVs performed after

immersion in the electrolyte. Since the cathodic current is usually larger than the anodic one

in the scanned potential range, we anticipate that the surface of all polymorphs slowly

converts to the mixed MnII/III valence state during potential cycling. Furthermore, the intensity

of the redox couple centered at around 0.75 VRHE in the first CV follows the same trend as the

BET area measured for unsupported MnOx materials, or as D−1. As directly compared in

Figure S5, the charge passed could therefore reflect the trend of ECSA and therefore confirms

good correlation with the BET area. Lastly Figure S6 provides linear sweep voltammograms

(LSV) in O2 purged environment to compare ORR activity with the HO2− yields of the various

MnOx catalysts. The LSVs in Figure S6b show similar apparent activity for all MnIV-oxides, while

the least porous MnIII-oxide exhibits a higher overpotential towards ORR. All MnOx catalysts

supported on carbon, however, produce significant amounts of HO2− as detected by the ring

current, starting at 0.71 VRHE, and reaching yields of up to 16% at lower potential.

114

Degradation during the oxygen reduction reaction

To understand the degradation of Mn oxide-based catalysts during FC operation, dissolution

of Mn is studied first in the ORR potential region. As a representative example, the dissolution

of Mn from the δ-MnO2 catalyst is shown in Figure 2. The on-line dissolution rate of Mn, ORR

current density and HO2− production yield during negative-going LSV are presented in

Figure 2a, 2b and 2c, respectively, as a function of potential.

Figure 2. Dissolution rate (a) and current density (b) in the SFC as well as HO2− yields in a RRDE

setup (c) measured for δ‒MnO2 in argon- (blue) and oxygen-purged (red) 0.05 M NaOH

electrolyte. LSV was performed by scanning from 1 to 0 VRHE, at a scan rate of 5 mV s−1. The

flow rate in the SFC was 3.65 μL s−1.

Starting at 1.0 – 0.9 VRHE, neither dissolution (Figure 2a) nor any significant current (Figure 2b)

are detected in SFC, both in Ar and O2 purged electrolytes. While the negligible ring currents

suggest that no HO2− is being produced at the disk (Figure 2c). At a slightly lower potential,

the onset of Mn dissolution is detected at 0.88 VRHE, independent of the nature of the purging

gas. The onset of dissolution is defined as the potential at which the ICP-MS signal is three

times higher than the standard deviation of the baseline signal. This dissolution onset

potential value corresponds well to the theoretical (th) potential of MnIV reduction to

115

MnIII(Eth = 0.95 V).13 We correlate the onset potential of Mn dissolution to thermodynamic

limits of stability, since a change in surface oxidation state usually corresponds to

rearrangement of the structure leading to increased dissolution.52,53

Interestingly, additional to dissolution triggered by MnIV reduction, online ICP-MS data reveals

an increasing Mn dissolution rate in O2- vs Ar-purged environment at potentials at which the

ORR proceeds at significant rate at 0.68 VRHE. A more detailed inspection of the results in

Figure 2 shows that the additional Mn dissolution at low potential is closely related to the

HO2− species detected at the ring than with the ORR current density itself. In particular, in the

restricted potential range of 0.7-0.8 VRHE, significant ORR current is measured but without

production of peroxide and without higher Mn dissolution rate than what was measured in

Ar-saturated electrolyte (Figure 2a). As soon as peroxide is detected at the ring, however, the

Mn dissolution rate becomes higher than what was measured in Ar-purged electrolyte

(Figure 2a). In the range 0.7 to 0.0 VRHE, a correlation between %HO2− and Mn dissolution rate

in O2 is observed, suggesting that peroxide products play an active role in enhancing the Mn

dissolution rate.

To confirm the detrimental effect of ORR (HO2− produced during ORR) on the stability of other

MnOx polymorphs, we applied CVs in O2-purged electrolyte with three different potential

ranges that have been typically used in the FC literature as accelerated stress tests (ASTs), to

the four MnOx allotropes previously described.54 These three different CV ranges are herein

labelled as load, high load and start/stop cycles with corresponding potential ranges of

1.0 – 0.6, 1.0 – 0 and 1.0 – 1.5 VRHE, respectively. Figure 3a compares the Mn dissolution rates

of all four catalysts in O2 (red) and Ar (blue) saturated electrolyte. The nature of the purging

gas does not change the position of the MnIV oxide dissolution at 0.88 VRHE. This dissolution

onset is assigned to MnIV to MnIII surface reduction and reconstruction, and is expected to be

the same for all MnO2 oxides since oxides and metal ions in electrolyte follow the same

thermodynamic standard potentials. For the MnIII oxide (Mn2O3), the onset of Mn dissolution

is shifted to lower potential (0.63 VRHE) and corresponds to the reduction to a mixed valence

MnII/III oxide (0.69 Vth).13 As reported above for δ-MnO2, the presence of oxygen in the

electrolyte, has also a strong impact on the Mn dissolution rate (compare red and blue curves

in Figure 3a).

116

Figure 3. Mn dissolution rate from four MnOx within commonly used potential ranges at

5 mV s−1 (load cycles: 1 – 0.6 VRHE, high load cycles: 1 – 0 VRHE, start/stop cycles: 1 – 1.5 VRHE).

(a) Data gathered in both oxygen (red) and argon (blue) purged 0.05 M NaOH. The data shown

here is smoothed using a 20 point FFT Filter, while the raw data is shown as a pale grey line.

(b) Catalyst dissolution in Ar purged 0.05 M NaOH for various H2O2 concentrations

(0, 1, 5, 10 μM).

Similar to the results in Figure 2, one can clearly see that the ORR induces significantly

increased Mn dissolution on all studied materials in the load cycles. On the other hand, during

start/stop cycles (no ORR taking place, even in O2-purged electrolyte), the Mn dissolution

rates are similar for a given MnOx structure in Ar- and O2-purged electrolyte. This

independently confirms that ORR plays a role (direct or indirect) in enhanced Mn dissolution.

As extrapolated from the RRDE results in Figure 2 for δ–MnO2, reactive oxygen species (ROS)

formed during the ORR are expected to also trigger enhanced Mn dissolution of the other

MnOx allotropes (Figure S6a). Therefore, experiments in Ar-purged electrolyte but with

various amounts of intentionally added H2O2 were performed. Figure 3b shows the impact of

HO2− concentration on Mn dissolution during the different LSV protocols. The figure reveals

that Mn dissolution increases with increased H2O2 concentration in the electrolyte. In contrast

to Figure 3a, where increased Mn dissolution is observed only at potentials at which the ORR

117

occurs (i.e. potentials at which peroxide is formed in situ), in Figure 3b increased Mn

dissolution is observed at any potential when comparing peroxide-containing solution to

peroxide-free solution. This unambiguously demonstrates that it is peroxide, and not ORR or

presence of O2, which triggers the enhanced Mn dissolution.

To obtain quantitative information on the absolute amount of dissolved Mn, the dissolution

profiles were integrated over the time scale. Figure 4a shows the total dissolved amount

(TDA) of Mn during the first two load cycles, for all studied materials. In Ar-purged electrolyte,

the TDA trend is as follows, δ–MnO2 ≈ β–MnO2 ≈ α–MnO2 >> α–Mn2O3. In O2-purged

electrolyte, the TDA trend is δ–MnO2 > β–MnO2 ≈ α–MnO2 >> α–Mn2O3. To quantify the

impact of purge gas, the TDA in oxygen was divided by the TDA in argon with results presented

on the right y-axis in Figure 4a. Here, α‒Mn2O3, α‒MnO2 and β‒MnO2 dissolve twice as much

in oxygen, while still following the same stability trend. As for δ‒MnO2, it dissolves three times

as much in O2-purged electrolyte. The total loss of Mn during load and high load cycles in

varying H2O2 concentrations is shown in Figure 4b, revealing the same stability trends as

observed in Figure 4a between crystal structure and TDA, where δ–MnO2 is the least stable

structure, followed by β–MnO2 ≈ α–MnO2 and α–Mn2O3 is apparently more stable, in the

presence of H2O2. Also, the increase of TDA with increasing peroxide concentration is quite

restricted for α–Mn2O3 vs. the others (compare TDA at 0 and 10 μM for each MnOx).

118

Figure 4. (a) The integral of Mn online dissolution (left y-axis, Mndiss) during two load cycles in

Ar (blue) and O2(red) saturated electrolyte. The factor of dissolution increase during ORR is

shown on the right y-axis (green). (b) Dependence of Mn dissolution on H2O2 concentration

during load and high load cycle experiments in argon purged 0.05 M NaOH.

In previous dissolution studies we recognized that there are typically two main dissolution

processes, namely transient and steady state dissolution, characterized by short lived

dissolution during potentiodynamic scans and constant dissolution during potentials holds,

respectively.19,55,56 Therefore, we explored the effect of oxygen on Mn stability further by

performing chronoamperometric (CA) ORR experiments at 0.6 VRHE in oxygenated electrolyte

to measure the steady state dissolution. Figures 5a shows current vs. time profiles of ORR on

the MnOx and Figure 5b the corresponding steady state Mn dissolution rate vs. time curves.

At 0.6 VRHE, MnIV is reduced to MnIII, as follows from Figure S5 and additionally confirmed by

XPS spectra presented in Figure 5c. When we compare the photoelectron spectra of the Mn3s

region before ORR (Figure 1b) with the ones presented here after the experiment (Figure 5c),

it becomes clear that the surface of all catalysts fully assimilated the thermodynamically

favorable oxidation state at the imposed potential of 0.6 VRHE, which is the MnIII oxide. This is

most obvious in the Mn 3s region, where the multiplet splitting parameters of all MnIV-oxides

119

shift from 4.96 to 5.49 eV (compare Figure 2b and Figure 5c). In contrast, the Mn 3s spectrum

of Mn2O3 remains almost unchanged after 10 minutes of ORR, coherent with thermodynamic

expectations.

Figure 5. Chronoamperometric (CA) curves recorded on all catalysts over 10 min at 0.6 VRHE

in O2 purged 0.05 M NaOH (a) and Mn dissolution rate vs time (b). XPS Mn 3s region of the

same catalysts after the CA experiment with indicated peak splitting (c).

Degradation during the oxygen evolution reaction

To fully understand the degradation of MnOx catalysts and to probe their applicability for OER,

we also investigate Mn oxides stability at higher anodic potentials. For this, we use a

previously reported method of correlating the dissolved amount of active species with the

anodic current passed.19 Figure 6 shows the dissolution rate during a LSV from 1 VRHE to a

potential which corresponds to current density of 1 mA cm−2, from which the stability

numbers (S-numbers) were calculated.19 While all catalysts show a similar onset potential of

OER, the extent of dissolution varies between oxides. Most noticeably it is seen in case of

δ-MnO2, which also exhibits much slower OER kinetics. Nevertheless, it produces the highest

amount of oxygen per atom dissolved with an S-number of 160. Other structures

α-Mn2O3 (90), α-MnO2 (140) and β-MnO2 (60) are within the same order of magnitude.

120

Considering that state of the art commercial IrOx catalyst reach S-numbers of 5×105 in acidic

media, this is a grim outcome even for cheap, earth abundant materials. Seemingly, the OER

stability of Mn might be more promising in the PEM environment,23,36 where stability was

even improved by the addition of Mn/TiOx surface layers.57 This advantage in acid is also due

to the pH dependent character of the MnO2/MnO4−

(aq) couple on the RHE scale allowing

higher overpotentials to the H2O/O2 couple in acidic electrolytes.13

Figure 6. Dissolution of the MnOx in the potential range corresponding to oxygen evolution

reaction. (a) LSV recorded from 1 V to a potential corresponding to 1 mA cm−2. Scan rate

10 mV s−1. (b) Corresponding Mn dissolution profiles. (c) Comparison of S-numbers for Mn

oxides with literature values for Ir-based oxide in acid.

Discussion

Generally, electrocatalyst degradation has been shown to be influenced by various factors.

The most studied example is carbon supported platinum for which degradation is typically

narrowed down to mechanisms like Ostwald ripening, agglomeration and particle

detachment, as factors of morphology and synthesis method, as well as support and finally

catalyst corrosion as an intrinsic consequence of the materials thermodynamic

121

tendency.56,58,59 As this study shows, in the experiments conducted during ORR, an additional

cause of corrosion can be the catalytic reaction itself by forming unfavorable side products

due to slow kinetics on non-noble metals and in this case Mn oxides. In the following we

differentiate between transient-, ORR-, and OER-dissolution.

As a basis for discussion we first focus on the thermodynamics of Mn in aqueous solutions

without the influence of oxygen and its possible reduction side products during ORR.

Therefore, Table 1 summarizes reported thermodynamic potentials13 of redox reactions that

can play a role in Mn dissolution during potentiodynamic operation. These transitions are the

major reason for transient-dissolution on MnOx in an argon purged environment as seen in

Figure 2a. While the reactions in Table 1 are more complex and plentiful as in a comparable

table for state of the art PEM Pt catalysts, intensively investigated over the last years,55 we

still observe similar behavior of MnOx. Thus, some parallels can be drawn. In essence, if at a

given potential the thermodynamically stable Mn oxidation state is not the one of the bulk

oxide, surface oxides start to change their oxidation state and therefore their coordination by

neighboring atoms. During this transition, some of the Mnn+ cations can be hydrated by the

aqueous electrolyte and diffuse through the electrochemical double layer into the bulk

solution. These Mn ions in the bulk electrolyte then represent transient dissolution in the

on-line ICP-MS measurement.

Table 1. Redox transitions of Mn and H2O relevant to the surface processes at p 12.7.

# Redox Transitionsa E0 (V vs. SHE)

Manganese

1(s/s) Mn + 2 OH− ⇌ Mn(OH)2 + 2 e− −0.727 − 0.0591 × pH

2(s/s) 3 Mn(OH)2 + 2 OH− ⇌ Mn3O4 + 4 H2O + 2 0.462 − 0.0591 × pH

3(s/s) 2 Mn3O4 + 2 OH− ⇌ 3 Mn2O3 + H2O + 2 e 0.689 − 0.0591 × pH

4(s/s) Mn2O3 + 2 OH− ⇌ 2 MnO2 + H2O + 2 e− 1.014 − 0.0591 × pH

5(s/aq) MnO2 + 4 OH− ⇌ MnO4− + 2 H2O + 3 e− 1.692 − 0.0788 × pH + 0.0197 log(MnO4

−

Hydrogen Peroxide

6(aq/aq 3 OH− ⇌ HO2− + H2O + 2 e− 2.119 − 0.0886 × pH + 0.0295 log(HO2

−)

7(aq/g) OH− + HO2− ⇌ O2 + H2O + 2 e− 0.338− 0.0295 × pH + 0.0295 log(pO2/H

8(aq/g) 2 HO2− ⇌ O2 + 2 OH−

s: solid; aq: aqueous; g: gaseous; aAdapted from Ref.13 to show the more likely alkaline

122

Obviously however, Figure 2a shows a significant increase in dissolution when the electrolyte

is oxygen purged. Interestingly, the potential range of increased dissolution coincides with

the increased current density during ORR (Figure 2b) as well as with the detection of HO2− in

RRDE experiments (Figure 2c). As it is numerously reported that the production of ROS during

ORR can be traced back to poor kinetics on PGM-free electrocatalysts,26,34,35 degradation of

MnOx electrocatalysts proceeds to a large extent through surface redox processes initiated

by ROS. Therefore, we will further refer to this degradation mechanism as ORR-dissolution.

To show that the extent of ORR-dissolution is a function of the ORR, four MnOx with various

LSV characteristics (Figure S6b) underwent identical ORR stability tests in Ar and O2 purged

conditions. On-line dissolution rates in Figure 3a demonstrate that the ORR-dissolution is

unavoidable on all investigated MnOx catalysts. The apparently most stable oxide is the one

comprised of MnIII, it’s reduction according to reaction #3 leads to dissolution at 0.63 VRHE

(0.69 Vth) starting at lower potentials than for the other MnIV oxides. All MnIV oxides dissolve

transiently at a potential of 0.88 VRHE upon their reduction following reaction #4 (1.01 Vth). As

a more representative characteristic of ORR dissolution Figure 4a summarizes the TDA of Mn

during the load cycle experiments normalized to the geometric surface area. Here, the extent

of Mn transient-dissolution follows α-Mn2O3 << α-MnO2 ≈ β-MnO2 < δ-MnO2. However,

keeping the results from XRD, SEM, BET and N2-purged CVs in mind, it is obvious that there

are important morphological differences between the different oxides. Therefore, we

normalize the same data from Figure 4a to the BET surface area in Figure 7.

Figure 7. The TDA of Mn on-line dissolution from Figure 4a normalized to the BET surface

area, in Ar (blue) and O2(red) saturated electrolyte.

123

Here, no significant impact of crystal structure on the dissolution rate in Ar is observed which

relates to the fact that the apparently less stable δ-MnO2 simply had a much higher surface

area from which it can dissolve. More importantly, however, the increased ORR-dissolution is

still obvious with a ratio of ORR-dissolution to transient-dissolution of ca. 2

(α-Mn2O3, α-MnO2, β-MnO2) and even up to 3 for δ-MnO2. The scaling of O2 to Ar dissolution

ratio could suggest, that the leaching of Mnn+ during transient-dissolution is increased by the

presence of HO2–. To strengthen this hypothesis, we further investigated the effect of an

intentional addition of H2O2 to the alkaline electrolyte, in which it undergoes a transition to

HO2− which leads to a slight acidification. However, the measured pH only changed marginally

from 12.7 to 12.5 after the highest H2O2 addition of 10 μM, ruling out a pH effect. In Figure 3b

the influence of HO2− concentration on Mn dissolution is shown for all catalysts during the

same potentiodynamic protocol. Here the increase of transient-dissolution with higher HO2−

addition, better observed in Figure 4b, supports the above statement that ROS increase

degradation of the catalyst during ORR. To explain this, one has to address the

thermodynamics of H2O2 first. HO2− can be considered as redox amphoteric, since it can react

as both reducing agent at potentials according to reaction #7 and oxidizing agent according

to reaction #6. There is only a small window between #6 and #7 were it simply

disproportionates into O2 and H2O (Reaction #8) without involving other species for electron

transfers. In ORR conditions however, where it can be formed according to reaction #7, it

mostly acts as an oxidizing agent towards manganese. This oxidized Mnn+ species in return is

thermodynamically not favored at ORR potentials as discussed earlier (transient-dissolution)

and is prone to dissolution. Thereby HO2− can force stable MnOx into an oxidation state where

it dissolves. Secondly, ORR-dissolution can also depend on transient radical species formed

during the partial reduction of O2, which recombine with the manganese oxide to an unstable

Mn surface state. Lastly, harmful redox process on the catalyst surface could be induced by

the catalytic mechanism of ORR as suggested by Ryabova et al.14 This is however limited to

the MnIII/IV transition and does not account for increased dissolution at low potentials were

MnIII and MnII are the present surface spezies. Independent on the operative mechanism

however, we demonstrate that ORR and accompanied HO2− production can be harmful for

catalyst. Therefore, we urge the PGM-free FC research community to use oxygen in all AST

protocols to fully account for all possible degradation mechanisms.

124

Next to transient- and ORR-dissolution, we detect Mn leaching during start stop cycles with

an onset at 1.41 VRHE which can be correlated to reaction #5 and the oxidation of surface MnIII

oxides following reaction #4. Presumably, as presented in section 3.2, during ORR cycling all

MnOx form a reduced oxide surface film. This condition leads us to discussing a third

degradation pathway, labelled OER-dissolution which becomes important when we move on

to EL or even possible bifunctional applications. Regardless of which oxidation state the

original catalyst is in, the surface will always adopt the thermodynamically favored oxidation

state. Therefore, when switching the mode of a bifunctional assembly, MnOx will always need

to cross at least one redox reaction (Table1) leading to transient-dissolution (similar results

were recently shown by da Silva et al. for acidic bifunctional ORR/OER Pt/IrOx catalysts).60

Furthermore, ORR- and OER-dissolution need to be accounted for since they occur intrinsically

during an applied potential relevant to OER and ORR as an outcome of intermediate species.

In case of OER Mn dissolution can come from the thermodynamically favored soluble MnO42−

species or from constant surface oxidation state changes during a single catalytic

cycle.13,19,37,38,61 In case of ORR we contend that the often observed ROS formation during a

favored 2-e− reduction step on PGM-free electrocatalysts significantly increase

transient-dissolution. Therefore, we currently cannot confirm long term stability of any

investigated crystal structure during ORR, OER let alone bifunctional application.

Nevertheless, equilibrium conditions might play an important role in for example metal air

batteries, where the dissolved material cannot be diluted by an electrolyte flow, and an

equilibrium between dissolution and redeposition can be reached.

Conclusions

In an effort to understand the stability of Mn-oxides for both the ORR as well as OER, we

uncovered an imperative drawback of their highly versatile redox chemistry. First of all,

similar to transient dissolution during red/ox transitions in noble metals, we observed the

same for Mn oxides in alkaline environment. Three MnIV and one MnIII oxides showed good

correlation between dissolution onset potentials and thermodynamic potentials of a

transition in oxidatisecion state. Especially in a bifunctional device this can lead to constant

degradation, since the catalysts’ surface will always rearrange to the thermodynamically

favored oxidation state in combination with transient dissolution. The surface transition of

125

four MnOx was shown during a CA ORR on-line SFC-ICP-MS experiment, showing constant

dissolution while XPS before and after revealed the full transition of the surface to the

reduced state. Additionally, to these established degradation mechanisms we observe an

increase of up to 300% TDAMn during ORR. By the means of SFC-ICP-MS and RRDE

experiments, we find a good correlation between dissolution rate, ORR currents and HO2−

yields suggesting ROS as the main destructive participator. This hypothesis was confirmed by

intentional H2O2 addition revealing a clear dependence of transient dissolution on the H2O2

concentration. Under OER conditions all investigated MnOx revealed comparable stability.

Furthermore, with S-Numbers orders of magnitude lower than state of the art PEM

electrocatalysts,19 MnOx do not represent a stable catalyst. This is attributed to their

thermodynamic transition to MnO4−

(aq) (according to reaction #5, Table 1). Lastly our results

were discussed in regard to bifunctional ORR/OER applications with the pivotal observation,

that such potentiodynamic applications lead to constant surface changes and accompanied

dissolution. Especially Mn with such a highly versatile redox chemistry cannot be fully stable

in a broad potential window. In short, the three most noticeable degradation mechanisms in

such a device would include:

Transient dissolution during oxidation state transitions of the surface.

Degradation during ORR due to ROS.

Dissolution during OER due to constant surface oxidation state changes and a thermodynamic

window of corrosion for Mn.

Therefore, we content a critical reassessment of bifunctional devices with materials exhibiting

transitions in the designated potential operation range. Further we urge researchers to

redefine testing procedures, e.g. AST protocols should include O2 so that in operando

degradation mechanisms can occur and a clear before and after physical analysis can give

insights on degradation even without on-line analysis methods.

126

References

1 Yoshida, T. & Kojima, K. Toyota MIRAI Fuel Cell Vehicle and Progress Toward a Future Hydrogen Society. Electrochem. Soc. Interface 24, 45-49, doi:10.1149/2.F03152if (2015).

2 Cherevko, S., Keeley, G. P., Kulyk, N. & Mayrhofer, K. J. J. Pt Sub-Monolayer on Au: System Stability and Insights into Platinum Electrochemical Dissolution. J. Electrochem. Soc. 163, I1228-I1233, doi:10.1149/2.0981603jes (2016).

3 Zhang, T., Wang, P. Q., Chen, H. C. & Pei, P. C. A review of automotive proton exchange membrane fuel cell degradation under start-stop operating condition. Appl Energ 223, 249-262, doi:10.1016/j.apenergy.2018.04.049 (2018).

4 Shao, M., Chang, Q., Dodelet, J. P. & Chenitz, R. Recent Advances in Electrocatalysts for Oxygen Reduction Reaction. Chem. Rev. 116, 3594-3657, doi:10.1021/acs.chemrev.5b00462 (2016).

5 Carmo, M., Fritz, D. L., Merge, J. & Stolten, D. A comprehensive review on PEM water electrolysis. Int J Hydrogen Energ 38, 4901-4934, doi:10.1016/j.ijhydene.2013.01.151 (2013).

6 Kibsgaard, J. & Chorkendorff, I. Considerations for the scaling-up of water splitting catalysts. Nat. Energy 4, 430-433, doi:10.1038/s41560-019-0407-1 (2019).

7 Greeley, J. et al. Alloys of platinum and early transition metals as oxygen reduction electrocatalysts. Nat. Chem. 1, 552-556, doi:10.1038/nchem.367 (2009).

8 Speck, F. D. et al. On the Electrolytic Stability of Iron-Nickel Oxides. Chem 2, 590-597, doi:10.1016/j.chempr.2017.03.006 (2017).

9 Kumar, K. et al. Physical and Chemical Considerations for Improving Catalytic Activity and Stability of Non-Precious-Metal Oxygen Reduction Reaction Catalysts. ACS Catal. 8, 11264-11276, doi:10.1021/acscatal.8b02934 (2018).

10 Choi, C. H. et al. Stability of Fe-N-C Catalysts in Acidic Medium Studied by Operando Spectroscopy. Angew. Chem. Int. Ed. Engl. 54, 12753-12757, doi:10.1002/anie.201504903 (2015).

11 Keeley, G. P., Cherevko, S. & Mayrhofer, K. J. The Stability Challenge on the Pathway to Low and Ultra-Low Platinum Loading for Oxygen Reduction in Fuel Cells. ChemElectroChem 3, 51-54, doi:10.1002/celc.201500425 (2016).

12 Manberger, A. & Stenqvist, B. Global metal flows in the renewable energy transition: Exploring the effects of substitutes, technological mix and development. Energy Policy 119, 226-241, doi:10.1016/j.enpol.2018.04.056 (2018).

13 Pourbaix, M. Atlas of Electrochemical Equilibria in Aqueous Solutions. ( NACE International, 1974).

14 Ryabova, A. S. et al. Rationalizing the Influence of the Mn(IV)/Mn(III) Red-Ox Transition on the Electrocatalytic Activity of Manganese Oxides in the Oxygen Reduction Reaction. Electrochim.

Acta 187, 161-172, doi:10.1016/j.electacta.2015.11.012 (2016). 15 Wang, L., Brink, J. J. & Varcoe, J. R. The first anion-exchange membrane fuel cell to exceed 1

W cm(-2) at 70 degrees C with a non-Pt-group (O2) cathode. Chem. Commun. 53, 11771-11773, doi:10.1039/c7cc06392j (2017).

16 Peng, X. et al. Nitrogen-doped Carbon-CoOx Nanohybrids: A Precious Metal Free Cathode that Exceeds 1.0 W cm(-2) Peak Power and 100 h Life in Anion-Exchange Membrane Fuel Cells. Angew. Chem. Int. Ed. Engl. 58, 1046-1051, doi:10.1002/anie.201811099 (2019).

17 Dekel, D. R. Review of cell performance in anion exchange membrane fuel cells. J. Power

Sources 375, 158-169, doi:10.1016/j.jpowsour.2017.07.117 (2018). 18 Varcoe, J. R. et al. Anion-exchange membranes in electrochemical energy systems. Energ

Environ Sci 7, 3135-3191, doi:10.1039/c4ee01303d (2014). 19 Geiger, S. et al. The stability number as a metric for electrocatalyst stability benchmarking.

Nat. Catal. 1, 508-515, doi:10.1038/s41929-018-0085-6 (2018). 20 Scholz, J. et al. Tailoring the Oxygen Evolution Activity and Stability Using Defect Chemistry.

Catalysts 7, doi:ARTN 139

127

10.3390/catal7050139 (2017). 21 Santori, P. G. et al. Effect of Pyrolysis Atmosphere and Electrolyte pH on the Oxygen Reduction

Activity, Stability and Spectroscopic Signature of FeNx Moieties in Fe-N-C Catalysts. J.

Electrochem. Soc. 166, F3311-F3320, doi:UNSP F3311 10.1149/2.0371907jes (2019).

22 Suntivich, J., Perry, E. E., Gasteiger, H. A. & Shao-Horn, Y. The Influence of the Cation on the Oxygen Reduction and Evolution Activities of Oxide Surfaces in Alkaline Electrolyte. Electrocatalysis 4, 49-55, doi:10.1007/s12678-012-0118-x (2012).

23 Huynh, M., Shi, C., Billinge, S. J. & Nocera, D. G. Nature of Activated Manganese Oxide for Oxygen Evolution. J. Am. Chem. Soc. 137, 14887-14904, doi:10.1021/jacs.5b06382 (2015).

24 Biesinger, M. C. et al. Resolving surface chemical states in XPS analysis of first row transition metals, oxides and hydroxides: Cr, Mn, Fe, Co and Ni. Appl. Surf. Sci. 257, 2717-2730, doi:10.1016/j.apsusc.2010.10.051 (2011).

25 Najafpour, M. M. et al. Manganese Compounds as Water-Oxidizing Catalysts: From the Natural Water-Oxidizing Complex to Nanosized Manganese Oxide Structures. Chem. Rev. 116, 2886-2936, doi:10.1021/acs.chemrev.5b00340 (2016).

26 Stoerzinger, K. A., Risch, M., Han, B. H. & Shao-Horn, Y. Recent Insights into Manganese Oxides in Catalyzing Oxygen Reduction Kinetics. ACS Catal. 5, 6021-6031, doi:10.1021/acscatal.5b01444 (2015).

27 Antoni, H. et al. Oxidative Deposition of Manganese Oxide Nanosheets on Nitrogen-Functionalized Carbon Nanotubes Applied in the Alkaline Oxygen Evolution Reaction. ACS

Omega 3, 11216-11226, doi:10.1021/acsomega.8b01433 (2018). 28 Risch, M. et al. Redox Processes of Manganese Oxide in Catalyzing Oxygen Evolution and

Reduction: An in Situ Soft X-ray Absorption Spectroscopy Study. J. Phys. Chem. C 121, 17682-17692, doi:10.1021/acs.jpcc.7b05592 (2017).

29 Cheng, F. & Chen, J. Metal-air batteries: from oxygen reduction electrochemistry to cathode catalysts. Chem. Soc. Rev. 41, 2172-2192, doi:10.1039/c1cs15228a (2012).

30 Calegaro, M. L., Lima, F. H. B. & Ticianelli, E. A. Oxygen reduction reaction on nanosized manganese oxide particles dispersed on carbon in alkaline solutions. J. Power Sources 158, 735-739, doi:10.1016/j.jpowsour.2005.08.048 (2006).

31 Gorlin, Y. et al. In situ X-ray absorption spectroscopy investigation of a bifunctional manganese oxide catalyst with high activity for electrochemical water oxidation and oxygen reduction. J. Am. Chem. Soc. 135, 8525-8534, doi:10.1021/ja3104632 (2013).

32 Gorlin, Y. & Jaramillo, T. F. A bifunctional nonprecious metal catalyst for oxygen reduction and water oxidation. J. Am. Chem. Soc. 132, 13612-13614, doi:10.1021/ja104587v (2010).

33 Bonnefont, A. et al. Challenges in the understanding oxygen reduction electrocatalysis on transition metal oxides. Curr. Opin. Electrochem. 14, 23-31, doi:10.1016/j.coelec.2018.09.010 (2019).

34 Ryabova, A. S. et al. Study of Hydrogen Peroxide Reactions on Manganese Oxides as a Tool To Decode the Oxygen Reduction Reaction Mechanism. ChemElectroChem 3, 1667-1677, doi:10.1002/celc.201600236 (2016).

35 Roche, I., Chainet, E., Vondrak, J. & Chatenet, M. Durability of carbon-supported manganese oxide nanoparticles for the oxygen reduction reaction (ORR) in alkaline medium. J. Appl.

Electrochem. 38, 1195-1201, doi:10.1007/s10800-008-9537-z (2008). 36 Li, A. et al. Stable Potential Windows for Long-Term Electrocatalysis by Manganese Oxides

Under Acidic Conditions. Angew. Chem. Int. Ed. Engl. 58, 5054-5058, doi:10.1002/anie.201813361 (2019).

37 Rabe, M. et al. Alkaline manganese electrochemistry studied by in situ and operando spectroscopic methods - metal dissolution, oxide formation and oxygen evolution. Phys.

Chem. Chem. Phys. 21, 10457-10469, doi:10.1039/c9cp00911f (2019).

128

38 Scholz, J. et al. Rotating Ring–Disk Electrode Study of Oxygen Evolution at a Perovskite Surface: Correlating Activity to Manganese Concentration. J. Phys. Chem. C 120, 27746-27756, doi:10.1021/acs.jpcc.6b07654 (2016).

39 Ching, S., Roark, J. L., Duan, N. & Suib, S. L. Sol−Gel Route to the Tunneled Manganese Oxide Cryptomelane. Chem. Mater. 9, 750-754, doi:10.1021/cm960460k (1997).

40 Quan, W. et al. New nanocomposite material as supercapacitor electrode prepared via restacking of Ni-Mn LDH and MnO2 nanosheets. Electrochim. Acta 247, 1072-1079, doi:10.1016/j.electacta.2017.07.010 (2017).

41 Ghodbane, O., Pascal, J. L. & Favier, F. Microstructural effects on charge-storage properties in MnO2-based electrochemical supercapacitors. ACS Appl. Mater. Interfaces 1, 1130-1139, doi:10.1021/am900094e (2009).

42 Cao, X. et al. A novel non-enzymatic hydrogen peroxide biosensor based on ultralong manganite MnOOH nanowires. Sensor Actuat B-Chem 147, 730-734, doi:10.1016/j.snb.2010.03.087 (2010).

43 Klemm, S. O., Schauer, J. C., Schuhmacher, B. & Hassel, A. W. A microelectrochemical scanning flow cell with downstream analytics. Electrochim. Acta 56, 4315-4321, doi:10.1016/j.electacta.2011.01.052 (2011).

44 Kijima, N., Ikeda, T., Oikawa, K., Izumi, F. & Yoshimura, Y. Crystal structure of an open-tunnel oxide α-MnO2 analyzed by Rietveld refinements and MEM-based pattern fitting. J. Solid State

Chem. 177, 1258-1267, doi:10.1016/j.jssc.2003.10.038 (2004). 45 Bolzan, A. A., Fong, C., Kennedy, B. J. & Howard, C. J. Powder Neutron Diffraction Study of

Pyrolusite, β-MnO2. Aust. J. Chem. 46, doi:10.1071/ch9930939 (1993). 46 Holland, K. L. & Walker, J. R. Crystal structure modeling of a highly disordered potassium

birnessite. Clays Clay Miner. 44, 744-748, doi:Doi 10.1346/Ccmn.1996.0440604 (1996). 47 Geller, S. Structure of α-Mn2O3, (Mn0.983Fe0.017)2O3 and (Mn0.37Fe0.63)2O3 and relation

to magnetic ordering. Acta. Crystallogr., Sect. B: Struct. Sci. 27, 821-828, doi:10.1107/s0567740871002966 (1971).

48 Norrestam, R., Ingri, N., Östlund, E., Bloom, G. & Hagen, G. alpha-Manganese(III) Oxide --- a C-Type Sesquioxide of Orthorhombic Symmetry. Acta. Chem. Scand. 21, 2871-2884, doi:10.3891/acta.chem.scand.21-2871 (1967).

49 Post, J. E. & Veblen, D. R. Crystal-Structure Determinations of Synthetic Sodium, Magnesium, and Potassium Birnessite Using Tem and the Rietveld Method. Am. Mineral. 75, 477-489 (1990).

50 Oku, M., Hirokawa, K. & Ikeda, S. X-ray photoelectron spectroscopy of manganese—oxygen systems. J. Electron. Spectrosc. Relat. Phenom. 7, 465-473, doi:10.1016/0368-2048(75)85010-9 (1975).

51 Gelius, U. et al. Molecular Spectroscopy by Means of ESCA III.Carbon compounds. Phys. Scr. 2, 70-80, doi:10.1088/0031-8949/2/1-2/014 (1970).

52 Schalenbach, M. et al. The Electrochemical Dissolution of Noble Metals in Alkaline Media. Electrocatalysis 9, 153-161, doi:10.1007/s12678-017-0438-y (2017).

53 Cherevko, S. et al. Dissolution of Noble Metals during Oxygen Evolution in Acidic Media. ChemCatChem 6, 2219-2223, doi:10.1002/cctc.201402194 (2014).

54 Pizzutilo, E. et al. On the Need of Improved Accelerated Degradation Protocols (ADPs): Examination of Platinum Dissolution and Carbon Corrosion in Half-Cell Tests. J. Electrochem.

Soc. 163, F1510-F1514, doi:10.1149/2.0731614jes (2016). 55 Kasian, O., Geiger, S., Mayrhofer, K. J. J. & Cherevko, S. Electrochemical On-line ICP-MS in

Electrocatalysis Research. Chem. Rec., doi:10.1002/tcr.201800162 (2018). 56 Cherevko, S., Kulyk, N. & Mayrhofer, K. J. J. Durability of platinum-based fuel cell

electrocatalysts: Dissolution of bulk and nanoscale platinum. Nano Energy 29, 275-298, doi:10.1016/j.nanoen.2016.03.005 (2016).

129

57 Frydendal, R., Paoli, E. A., Chorkendorff, I., Rossmeisl, J. & Stephens, I. E. L. Toward an Active and Stable Catalyst for Oxygen Evolution in Acidic Media: Ti-Stabilized MnO2. Adv. Energy

Mater. 5, doi:ARTN 1500991 10.1002/aenm.201500991 (2015).

58 Katsounaros, I., Cherevko, S., Zeradjanin, A. R. & Mayrhofer, K. J. Oxygen electrochemistry as a cornerstone for sustainable energy conversion. Angew. Chem. Int. Ed. Engl. 53, 102-121, doi:10.1002/anie.201306588 (2014).

59 Meier, J. C. et al. Design criteria for stable Pt/C fuel cell catalysts. Beilstein J. Nanotechnol. 5, 44-67, doi:10.3762/bjnano.5.5 (2014).

60 da Silva, G. C., Mayrhofer, K. J. J., Ticianelli, E. A. & Cherevko, S. The degradation of Pt/IrOx oxygen bifunctional catalysts. Electrochim. Acta 308, 400-409, doi:10.1016/j.electacta.2019.04.017 (2019).

61 Man, I. C. et al. Universality in Oxygen Evolution Electrocatalysis on Oxide Surfaces. ChemCatChem 3, 1159-1165, doi:10.1002/cctc.201000397 (2011).

130

Critical importance of the ionomer on the electrochemical

activity of platinum and non-platinum catalysts in anion-

exchange membrane fuel cells

Pietro G. Santori,a Abhishek N. Mondal,b Dario R. Dekel,b,c * and Frédéric Jaouena *

a Institut Charles Gerhardt Montpellier, UMR 5253, CNRS, Université Montpellier, ENSCM,

Place Eugène Bataillon, 34095 Montpellier cedex 5, France

b The Wolfson Department of Chemical Engineering, Technion – Israel Institute of Technology,

Haifa 3200003, Israel

c The Nancy & Stephan Grand Technion Energy Program (GTEP), Technion – Israel Institute of

Technology, Haifa 3200003, Israel

131

Abstract

Anion-exchange membrane fuel cells (AEMFCs) show remarkable and rapid progress in

performance, significantly increasing the relevance for research on electrocatalysis of the

oxygen reduction reaction (ORR) and hydrogen oxidation reaction (HOR) for this technology.

Since much of the recent progress in AEMFC performance can be tied to the improved

interface between anion-exchange ionomers (AEI) and catalysts, this topic deserves a specific

attention.

This work reports the ORR and HOR activity measured in rotating disk electrode for several

ionomer-catalyst combinations involving five different AEIs and Nafion®, and four ORR and

HOR catalysts selected from the best in-class PGM-based and PGM-free catalysts. The results

show little impact of the ionomers on the ORR and HOR activity of Pt/C and PtRu/C catalysts,

respectively; however, the choice of the AEI has a critical importance on the ORR activity of

Fe-N-C and significant effect on the HOR activity of Pd-CeO2/C.

Introduction

The latest development of anion-exchange ionomers (AEIs) and membranes (AEMs) with high

hydroxide conductivity have resulted in tremendous progress in the performance of AEM fuel

cells (AEMFCs).1-3 As a consequence, materials research for AEMFCs is a blooming

multidisciplinary field mainly involving polymer-, solid-state science and electrocatalysis. For

the oxygen reduction reaction (ORR), while high activity and durability is presently reached

with PGM-based catalysts,4-7 the current focus is to replace them with PGM-free catalysts8-14

or even with metal-free carbon-based catalysts.12, 15-16 For the hydrogen oxidation reaction

(HOR), sluggish kinetics of HOR in the alkaline medium seems very challenging even for PGM-

based catalysts.17-21 For the AEIs and AEMs the current focus of research is to increase their

chemical stability under AEMFC operation conditions.22

While separate development for each class of material is necessary, mutual interaction

between catalytic and ionomeric materials at each electrode must be taken into account to

achieve optimized AEMFC performance. However, catalyst-AEI interaction is an unexplored

field, with only a very scarce number of studies have focused on the interaction effect

between ORR and HOR electrocatalysts and AEI in the alkaline medium of AEMFCs. Jervis et

al. showed that coupling Pt/C with acidic proton-exchange ionomer underestimated its HOR

132

activity in alkaline electrolyte.23 The authors claimed that Nafion acts as an insulator of OH–,

affecting hydroxide transport towards the catalytic surface and through the Nafion thin film.

They emphasized the imperative need to couple catalysts with AEIs while evaluating HOR

activity in alkaline medium. Kim et al. revealed that the adsorption of some AEI functional

groups onto the surface of Pt and Pt-M bimetallic catalysts negatively impacted their HOR

activity in alkaline electrolyte,24 and in turn, the AEMFC performance.25 The authors

demonstrated that Pd-based HOR catalysts deactivated following adsorption and

hydrogenation of phenyl group,26 indicating that co-adsorption of cation-hydroxide-water can

occur on the surface of PGM catalysts, limiting the hydrogen diffusion on their surface, and

therefore affecting catalyst activity towards HOR in alkaline medium.27 All the above studies

focused on PGM-based catalysts for HOR. To the best of our knowledge, there are no studies

on AEI-catalyst interactions either for ORR catalysis or for PGM-free materials.

In this study, we investigated the catalytic activity of different AEI-catalyst inks in alkaline

electrolyte using the RDE method. The materials of this work comprises different AEIs, HOR

and ORR catalysts, selected from the best-in-class Pt-based, low-PGM and PGM-free catalysts.

Experimental

The ionomers and catalysts used in this study are summarized in Table 1. To synthesize the

polyphenylene oxide (PPO)-based AEIs, we prepared brominated PPO (Br-PPO, 25%

bromination degree) as reported elsewhere.28 About 0.5 g Br-PPO was dissolved into 5 wt%

N-Methyl-2-pyrrolidone solution in a 25 mL round bottom flask and functionalized with excess

of trimethylamine (TMA), triethylamine (TEA), 1-methylpyrrolidine (MPy) or N,N-

dimethylbenzylamine (DMBA). Once the reaction was completed, the polymer solution was

cast and dried for 24 h into AEI films.

The Fe-N-C catalyst with atomically-dispersed Fe sites (labelled Fe0.5-950) was prepared as in

29 with ORR alkaline activity and stability using Nafion, reported in 30. The Pd-CeO2/C catalyst

was prepared as reported in 1. We selected a loading of 10 wt% Pd, as it was found to achieve

the highest HOR activity.1 Baseline inks were prepared adding 54 µL of either AEI (5 wt% in

dimethylformamide) or Nafion, and 836 µL of ethanol to 5 mg catalyst, resulting in

ionomer/catalyst mass ratio of 0.51. For FAA3 an additional ink formulation was also used,

with half amount of ionomer (named FAA3 halved).

133

The choice of the 0.51 ratio is derived from a previous study by us on the same Fe0.5-950

catalyst (> 95% carbon) in alkaline electrolyte using Nafion ionomer30 and also comparable to

ionomer/carbon ratios of 0.42 – 0.62 used by Omasta et al. for PtRu/C catalysts in

combination with an AEI in optimized AEMFCs.31 For the PGM-based catalysts, we used the

same baseline ink formulation as for Fe0.5-950 for consistency. With halved FAA-3/catalyst

ratio vs. baseline value (also studied in this work for all catalysts with FAA-3), the ionomer-to-

carbon ratio of the Pt/C and PtRu/C catalysts investigated in this study falls in the range

employed by Omasta et al in AEMFC.31 After sonicating the inks for 1 h, 7 µL of it was applied

onto glassy carbon tip (Pine Research), dried overnight under vacuum at room temperature,

to achieve working electrodes with 200 µgcat cm-2. All AEIs were in Br-form during ink

preparation, and exchanged in situ in RDE to OH-form.

The RDE studies were carried out in a three-electrode set-up with a reversible hydrogen

electrode (RHE) reference and a graphite plate counter electrode. Cyclic-voltammetry (CV)

was applied in N2-saturated electrolyte (SP-300, BioLogic Potentiostat). The ORR and HOR

activities were measured in the potential range 0 – 1 V and 0 – 0.4 V vs. RHE, respectively.

Results & Discussion

Figure 1 shows the ORR and HOR polarisation curves for all the catalysts, recorded for each

ionomer with baseline ink formulation and with a halved FAA3/catalyst ratio.

134

Figure 1. Polarisation curves for ORR on a) Pt/C and b) Fe0.5-950; and HOR on c) PtRu/C and

d) Pd-CeO2/C. 0.1 M KOH electrolyte saturated with O2 or H2, 1600 rpm, scan rate 1 mV s-1,

catalyst loading 200 µg cm-2.

Effect of ionomer for ORR

We discuss ORR activity on the basis of Faradaic current density at 0.9 V vs. RHE as read from

Figure 1. For Pt/C, the ORR activity is independent of the AEI (-1.6 to -1.8 mA cm-2 at 0.9 VRHE)

and similar to that obtained with Nafion (-1.8 mA cm-2 at 0.9 VRHE). Only for FAA3 the activity

is slightly lower (filled yellow circles, -1.2 mA cm-2 at 0.9 VRHE) but this effect disappears with

halved FAA3/catalyst mass ratio (open yellow circles). This might be interpreted as excess

FAA3 with the baseline ink formulation. The similar ORR activity for Pt/C with all AEIs is

correlated with similar CVs (Fig. 2a). While activities at 0.9 VRHE are similar, the transition from

the kinetically-controlled to the diffusion-limited region of polarisation curves is less sharp for

Pt/C with baseline FAA3 content (filled yellow circles, Fig. 1a). Also, it is noted that the

diffusion-limited current density (Jlim) is slightly lower for Pt/C with the synthesized PPO-based

AEIs as compared to the case of Pt/C-Nafion, at baseline ionomer content. This may indicate

135

either a less-selective ORR on Pt/C interfaced with such AEIs or a lower O2 permeability in

PPO-based AEIs compared to Nafion. A decreased permeability could lead to significant

diffusion barrier through the AEI thin film covering catalytic particles. AEIs and AEMs are

known to possess one order of magnitude lower O2 and H2 permeability than Nafion.32

Figure 2. CVs for a) Pt/C, b) Fe0.5-950, c) PtRu/C and d) Pd-CeO2/C. 0.1 M KOH electrolyte

saturated with N2, scan rate 10 mV s-1, catalyst loading 200 µg cm-2.

More significant changes of the ORR activity with AEI is observed for Fe0.5-950, especially with

FAA3 (Fig. 1b). While the ORR activity at 0.9 VRHE is similar for Fe0.5-950 interfaced with any of

the synthesized AEIs and comparable to that obtained with the Nafion, in the range of -0.37

to -0.47 mA cm-2, no ORR activity is observed for Fe0.5-950 interfaced with FAA3 at baseline

or halved FAA3/catalyst ratio (yellow or filled circles, Fig. 1b). The corresponding polarisation

curves can in fact be assigned to the ORR activity of glassy carbon (dashed curve). This is

correlated by the lack of signal in the CVs (yellow symbols, Fig. 2b). In contrast, the CVs of

Fe0.5-950 coupled with any other AEI presently investigated are comparable, and comparable

136

to that obtained for Fe0.5-950 with the Nafion (Fig. 2b). The strong effects of the ionomer

shown in these results obtained with Fe0.5-950-FAA3 may be interpreted by (i) a film of FAA3

formed on top of glassy-carbon, thereby electrically insulating Fe0.5-950 from the current

collector, or (ii) a film of FAA3 formed on the surface of all catalytic particles, electrically

insulating each of them. Hypothesis (ii) seems, however, more likely to explain the results

observed. We also note that for the other AEIs, the fine trends observed for Fe0.5-950 are

similar to those observed for Pt/C, with a lower Jlim-value reached with the AEIs compared to

that observed with Nafion. The decrease of Jlim when switching from Nafion to AEI is more

exacerbated here for Fe0.5-950 than for Pt/C. Also, the transition from the kinetically-

controlled region to the diffusion-limited region of the polarisation curves for Fe0.5-950 is less

sharp with PPO-based AEIs than with Nafion, possibly indicating an additional diffusion barrier

due to the AEI thin-film, in line with the generally lower gas permeability of such AEIs

compared to Nafion. This diffusion barrier would expectedly play an important role as the

density of active sites decreases (from 60 to 2 wt% metal from Pt/C to Fe0.5-950), similar to

what has been reported for Nafion-Pt/C, where decreased Pt content resulted in increased

diffusion barrier.33-34

Effect of ionomer for HOR

Similar to the case for ORR on Pt/C, the HOR activity on PtRu/C is invariable with the

synthesized AEI, as can be seen from nearly superimposed curves in the range 0 – 2 mA cm-2,

and similar to that obtained with the Nafion (Fig. 1c). However, in the case of FAA3 at baseline

content, the polarisation curve strongly deviates from the others already at 0.5 mA cm-2, and

has an apparent Jlim-value that is 2.5 times lower than the theoretical one (filled yellow circles,

Fig. 1c). This effect disappears when the FAA3 content is halved (open yellow circles), the

apparent Jlim-value now even slightly exceeding that reached with PtRu/C coupled with Nafion

(open yellow vs. filled black circles, Fig. 1c).

The identical HOR activity of PtRu/C obtained with the synthesized AEIs is correlated by similar

CVs (Fig. 2c), nearly superimposed with the CV for PtRu/C with Nafion. In addition, the

significantly lower Jlim-value of the layer with FAA3 at baseline content is obviously correlated

with a much supressed CV.

137

Similar to the case for ORR on Pt/C and on Fe0.5-950, slightly lower Jlim values are observed for

HOR on PtRu/C with the AEIs than for PtRu/C with Nafion. Because the HOR can only be a

two-electron reaction, this observation convincingly suggests that the AEI thin-film indeed

results in an additional diffusion barrier. As for Pt/C, the effect is restricted for PtRu/C due to

the high density of active sites.

With all the synthesized AEIs and Nafion, the HOR polarisation curves on Pd-CeO2/C show a

linear shape up to 0.3 V vs. RHE (Fig. 1d), consistent with previously reported results for the

same catalyst but with Nafion.2 The apparent smaller HOR activity of Pd-CeO2/C vs PtRu/C is

partly due to the lower PGM loading (20 µgPd cm-2 and 120 µgPtRu cm-2). The intrinsic activity

of Pd-CeO2/C is therefore also high, and peak power density > 1 W cm-2 has been achieved

with it in AEMFC.2, 35 The smaller slope of the polarisation curves observed on Pd-CeO2/C with

the synthesized AEIs in the 0 – 0.3 V region identifies a lower HOR activity and/or hydrogen

access to the catalytic surface as for Pd-CeO2/C coupled with the Nafion. With FAA3, no HOR

activity is observed at baseline AEI content (filled yellow circles, Fig. 1d), while the HOR

activity becomes comparable with that obtained with the other AEI when the FAA3 content

is halved (open yellow circles, Fig. 1d). The HOR results are well correlated by the CVs seen in

Fig. 2d. At baseline FAA3 content, there is no response in the CV (filled yellow circles), and

this can be paralleled to the observation made for Fe0.5-950 at baseline FAA3 content. For the

other cases, all CVs have a similar shape, differing only slightly in the overall signal intensity

(Fig. 2d). It is stressed that the impact of switching from Nafion to the AEI on the

electrocatalytic properties is stronger for Pd-CeO2/C than for Pt/C and PtRu/C, and this ties

with the lower density of active sites of Pd-CeO2/C (only 10 wt% Pd). The negative impact

when switching from Nafion to the AEI seems therefore to increase with a decreasing density

of active sites in catalysts: Pt/C ~ PtRu/C > Pd-CeO2/C > Fe0.5-950. This can be understood by

the increased resistance for gas (O2 or H2) permeation through an ionomer thin film towards

the lower number of catalytic sites of these catalysts. On top of this, a secondary parameter

playing a role might be the location of such active sites, with Fe-based sites being at least in

part located in micropores in Fe0.5-950.29-30 Figure 3 schematically illustrates this. For a fixed

electrode current density, the catalyst active sites have a higher ORR (or HOR) turnover

frequency in an electrode with less number of active sites, leading in turn to higher local flux

of O2 (or H2).

138

Figure 3. Scheme representing increasing O2 (or H2) local flux with decreasing catalyst active

site density, for a same total O2 (or H2) consumption rate.

Conclusions

In this study, we investigated the catalytic activity of different AEI-catalyst inks in alkaline

electrolyte using the RDE method. The materials included different ionomers (Nafion, FAA3

and four PPO-based AEIs functionalized with different functional groups) and ORR and HOR

catalysts (Pt/C, Fe0.5-950; PtRu/C and Pd-CeO2/C). Pt/C and PtRu/C lead to high ORR and HOR

activities, respectively, when coupled with any of the five AEIs. Compared to the coupling with

Nafion, the activities are comparable but lower diffusion-limited current densities are

observed for both for ORR and HOR. This may be assigned to a lower gas permeability in AEIs

compared to Nafion®. Coupling Fe0.5-950 catalyst and AEIs was found more challenging,

leading to a complete loss of ORR activity with FAA3, to high activity with the four synthesized

AEIs. In the latter case, the diffusion-limited current density was however also lower than that

reached with Nafion. Regarding HOR on Pd-CeO2/C, all AEIs lead to comparable activity at low

overpotential of 0-0.1 V, but significantly lower than the HOR activity observed with Nafion®.

This difference expanded at higher current density (higher potential).

The challenge of coupling AEI and catalysts to form active layers seems to be more critical for

the catalysts with low density of active sites (2 wt% Fe and 10 wt% Pd for Fe0.5-950 and Pd-

CeO2/C, respectively) than catalysts with high active sites density (60 wt% metal for Pt/C and

PtRu/C). This general observation also is in line with the hypothesis that a key issue with AEI

may be their low gas permeability. Catalysts with low site density would expectedly

exacerbate this effect due to enhanced reactant consumption per active site, at a given

geometric current density. This may play a critical role in AEMFC, where current densities are

139

orders of magnitude higher than those in RDE. Overall, this study shows the fundamental and

practical importance of ionomer-catalyst interfaces for different classes of catalysts, with

dramatic impact on apparent activity. Improved fundamental understanding of how ionomers

interact at the microscale with different classes of catalysts and different types of active sites

(atomically dispersed metal-sites or metal-free sites vs. metallic or metal oxide particles in

particular) is needed for the rational design of membrane-electrode assemblies and rational

choices of catalyst and catalyst-compatible ionomers.

140

References

1. H.A. Miller, F. Vizza, M. Marelli, A. Zadick, L. Dubau, M. Chatenet, S. Geiger, S. Cherevko, H. Doan, R.K. Pavlicek, S. Mukerjee, D.R. Dekel, Nano Energy 33 (2017) 293-305.

2. T.J. Omasta, X. Peng, H.A. Miller, F. Vizza, L. Wang, J.R. Varcoe, D.R. Dekel, W.E. Mustain, J. Electrochem. Soc. 165 (2018) J3039-J3044

3. D.R. Dekel, J. Power Sources 375 (2018) 158-169. 4. T.J. Omasta, L. Wang, X. Peng, C.A. Lewis, J.R. Varcoe, W.E. Mustain, J. Power Sources 375

(2018) 205-213. 5. L. Wang, E. Magliocca, E.L. Cunningham, W.E. Mustain, S.D. Poynton, R. Escudero-Cid, M.M.

Nasef, J. Ponce-Gonzalez, R. Bance-Souahli, R.C.T. Slade, D.K. Whelligan, J.R. Varcoe, Green Chem. 19 (2017) 831-843.

6. L. Wang, M. Bellini, H.A. Miller, J.R. Varcoe, J. Mater. Chem. A 6 (2018) 15404-15412. 7. Y. Wang, G. Wang, G. Li, B. Huang, J. Pan, Q. Liu, J. Han, L. Xiao, J. Lu, L. Zhuang, Energy

Environ. Sci. 8 (2015) 177-181. 8. X. Peng, V. Kashyap, B. Ng, S. Kurungot, L. Wang, W.E. Mustain, Catalysts 9 (2019) 264. 9. H. Ren, Y. Wang, Y. Yang, X. Tang, Y. Peng, H. Peng, L. Xiao, J. Lu, H.D. Abruna, L. Zhuang, ACS

Catal. 7 (2017) 6485-6492. 10. X. Peng, T.J. Omasta, E. Magliocca, L. Wang, J.R. Varcoe, W.E. Mustain, Angew. Chem. Int.

Ed. 58 (2019) 1046–1051. 11. M.M. Hossen, K. Artyushkova, P. Atanassov, A. Serov, J. Power Sources 375 (2018) 214-221. 12. A. Sarapuu, E. Kibena-Pöldsepp, M. Borghei, K. Tammeveski, J. Mater. Chem. A 6 (2018) 776-

804. 13 S. Ratso, I. Kruusenberg, M. Käärik, M. Kook, L. Puust, R. Saar, J. Leis, K. Tammeveski, J. Power

Sources 375 (2018) 233-243. 14. S. Ratso, M. Käärik, M. Kook, P. Paiste, V. Kisand, S. Vlassov, J. Leis, K. Tammeveski,

ChemElectroChem. 5 (2018) 1827-1836. 15. Y. Lu, L. Wang, K. Preuss, M. Qiao, M-M. Titirici, J.R. Varcoe, Q. Cai, J. Power Sources 372

(2017) 82-90. 16. M.P. Kumar, M.M. Raju, A. Arunchander, S. Selvaraj, G. Kalita, T.N. Narayanan, A.K. Sahu,

D.K. Pattanayak, J. Electrochem. Soc. 163 (2016) F848-F855. 17. E. Davydova, S. Mukerjee, F. Jaouen, D.R. Dekel, ACS Catalysis 8 (2018) 6665-6690. 18. D. Strmcnik, M. Uchimura, C. Wang, R. Subbamaran, N. Danilovic, D. van der Vliet, A.P.

Paulikas, V.R. Stamenkovic, N.M. Markovic, Nat. Chem. 5 (2013) 300–306. 19. D.R. Dekel, Current Opinion in Electrochem. 12 (2018) 182-188. 20. E.S. Davydova, J. Zaffran, K. Dhaka, M.C. Toroker, D.R. Dekel, Catalysts 8 (2018) 454. 21. A. Roy, M.R. Talarposhti, S.J. Normile, I.V. Zenyuk, V. De Andrade, K. Artyushkova, A. Serov,

P Atanassov, Sustain. Energy Fuels 2 (2018) 2268-2275. 22. C.E. Diesendruck, D.R. Dekel, Current Opinion in Electrochem. 9 (2018) 173–178. 23. R. Jervis, N. Mansor, A. J. Sobrido, S. Jones, C. Gibbs, T.P. Neville, J. Millichamp, P.R. Shearing,

D.J.L. Brett, J. Electrochem. Soc. 164 (2017) F1551-F1555. 24. I. Matanovic, H.T. Chung, Y.S. Kim, J. Phys. Chem. Lett. 8 (2017) 4918−4924. 25. S. Maurya, C.H. Fujimoto, M.R. Hibbs, C. Narvaez Villarrubia, Y.S. Kim, Chem. Mater. 30

(2018) 2188−2192. 26. S Maurya, J.H. Dumont, C. Narvaez Villarrubia, I. Matanovic, D. Li, Y.S. Kim, S. Noh, J. Han, C.

Bae, H.A. Miller, C.H. Fujimoto, D.R. Dekel, ACS Catalysis 8 (2018) 9429–9439. 27. D. Li, H.T. Chung, S. Maurya, I. Matanovic, Y.S. Kim, Current Opinion in Electrochemistry 12

(2018) 189–195. 28. S. Willdorf-Cohen, A.N. Mondal, D.R. Dekel, C.E. Diesendruck, J. Mater. Chem. A 6 (2018)

22234-22239.

141

29. A. Zitolo, V. Goellner, V. Armel, M-T. Sougrati, T. Mineva, L. Stievano, E. Fonda, F. Jaouen, Nature Mater. 14 (2015) 937-942.

30. P.G. Santori, F.D. Speck, J. Li, A. Zitolo, Q. Jia, S. Mukerjee, S. Cherevko, F. Jaouen, J. Electrochem. Soc. 166 (2019) F3311-F3320.

31. T.J. Omasta, A.M. Park, J.M. LaManna, Y. Zhang, X. Peng, L. Wang, D.L. Jacobson, J.R. Varcoe, B.S. Pivovar, W.E. Mustain, Energy Environ. Sci. 11 (2018) 551-558.

32. H. Ono, J. Miyake, S. Shimada, M. Uchida, K. Miyatake, J. Mater. Chem. A 3(2015) 21779-21788.

33. S. Jomori, N. Nonoyama, T. Yoshida, J. Power Sources 215 (2012) 18-27. 34. S. Kinoshita, T. Tanuma, K. Yamada, S. Hommura, A. Watakabe, S. Saito, T. Shimohira, ECS

Transactions 64 (2014) 371-375. 35. M. Bellini, M. Pagliaro, A. Lenarda, P. Fornasiero, M. Marelli, C. Evangelisti, M. Innocenti, Q.

Jia, S. Mukerjee, J. Jankovic, Lianqin Wang, John R. Varcoe, Chethana Krishnamurthy, Ilya Grinberg, Elena Davydova, D. R. Dekel, H. A. Miller, F. Vizza, ACS Applied Energy Materials, DOI: 10.1021/acsaem.9b00657

142

Highly active and durable Fe0.5-NH3 cathode in AEMFC

Pietro Giovanni Santori,1 Xiong Peng,2 William E. Mustain,2 Frédéric Jaouen1

1. Institut Charles Gerhardt Montpellier, UMR 5253, CNRS, Université Montpellier,

ENSCM, Place Eugène Bataillon, 34095 Montpellier cedex 5, France

2. Department of Chemical Engineering, University of South Carolina, Columbia, SC

29208, USA

143

Introduction

Non-noble catalysts based on metal nitrogen and carbon (M-N-C) were early shown to be

highly active towards ORR in rotating disk electrode (RDE) setup in alkaline electrolyte.1,2 The

pioneering work of Varcoe’s group on radiation-grafted polyethylene membrane and ETFE-

powder ionomer opened the route a few years ago to the design of high performance MEAs

for AEMFCs.3-5 Several laboratories worldwide have now reported excellent power

performance with different AEMS and AEIs and usually, Pt/C cathodes and PtRu/C or

Pd/CeO2/C anodes.6-8 Since this breakthrough occurred recently, the vast majority of studies

reporting AEMFC performance have hitherto relied on Pt/C catalysts at the cathode.9,10 Only

few studies have investigated non-PGM cathode catalysts in AEMFC,11,12 and even less have

investigated M-N-C cathodes in such a device.13-15 The investigation of PGM-free cathode

catalysts in AEMFC is therefore timely. Recently, Peng et al. reported high AEMFC

performance using a cathode based on carbon-support cobalt ferrite (CoFe2O4/C)

nanoparticles.16 While this catalyst showed poor ORR activity in RDE setup, with half-wave

potential of only 0.78 V vs RHE, the peak power density with a PtRu/C anode (0.7 mgPtRu cm-

2) and a cathode metal loading (Co + Fe) of 2.4 mg cm-2 reached 1.35 W cm-2 in H2/O2 AEMFC

and 0.67 W cm-2 in H2/air AEMFC. These results can be considered the state of art in term of

AEMFC performance with PGM-free cathode, surpassing the AEMFC performance reported

with an Ag/C cathode by Varcoe et al.15

The specific aim of this chapter is to investigate in AEMFC the application of the highly active

Fe-N-C catalysts discussed in chapter 2. As previously mentioned, Fe0.5-NH3 has been

intensively studied in acidic medium, reporting great initial performance in PEMFC, but

suffering of a fast performance decay, especially during the first 10 hours. Chapter 2

evidenced that the main deactivation mechanism of Fe0.5-NH3 in acid is the leaching of Fe

cations from the actives sites, triggered by the combination of low electrochemical potential

and the presence of O2. The ten times faster demetallation rate in acid medium for Fe0.5-NH3

compared to Fe0.5-Ar also supports the hypothesis that highly-basic nitrogen groups in Fe0.5-

NH3 are coordinating the Fe cations, resulting in more active but intrinsically less stable active

sites in acid medium. In contrast, in alkaline medium, it was shown that both the ORR activity

and stability of Fe0.5-NH3 are promising. Therefore, Fe0.5-NH3 is evaluated in AEMFC device in

144

this chapter, looking at its ORR activity at high potential, power density reached with pure O2

and synthetic air, and finally, exploring its durability during operation for 100 h.

Experimental

Fe0.5-NH3 was synthetized using the same procedure previously reported,17 but using iron

acetate (57Fe isotope) in order to carry out Mössbauer characterisation of the cathode at

Beginning of Test and End of Test (BoT and EoT).

The catalyst ink for RDE measurements was prepared mixing 5 mg of catalyst with 54 µL of

Nafion (5 wt% solution, Sigma Aldrich), 744 µL of ethanol and 92 µL of ultrapure water (18

MΩ). The dispersion is sonicated for 1 h in ice-bath and the ink is drop cast in the glassy carbon

tip, to reach a catalyst loading of 0.2 mg cm-2 and then dried at room temperature. A three-

electrode configuration is used for the electrochemical studies, using the catalysed GC as a

working electrode, a graphite rod as counter electrode and a platinum wire place in a fritted

glass compartment filled with H2-saturated electrolyte as reversible hydrogen electrode (RHE)

reference.

Initially, the electrolyte (0.1 M KOH) is saturated with nitrogen, and cycle-voltammetry (CV)

is applied in the potential range between 0.0 and 1.0 V vs. RHE, at a scan rate of 10 mV s-1

and a rotation rate of 1600 rpm, in order to evaluate the double-layer capacitance of the

catalyst. Then the electrolyte is saturated with oxygen to evaluate the ORR activity, applying

CVs in the same range of potential and at a same rotation rate, but lowering the scan rate to

1 mV s-1 in order to neglect capacitive currents.

For AEMFC testing, the catalytic inks were prepared following the procedure described in

Omasta et al.,10 manually grinding the catalyst (PtRu/C anode catalyst or 57Fe0.5-NH3 cathode

catalyst) and ETFE (ethylene tetrafluoroethylene) powder ionomer with 1 mL of H2O and 9 mL

of 1-propanol. The ETFE content was different at the anode and cathode, corresponding to

20 wt % with respect to 57Fe0.5-NH3 and 40 wt % with respect to the carbon content in the

PtRu/C anode catalyst. The anode catalyst was prepared by mixing 40wt%Pt-20wt%Ru/C

(Johnson Matthey) with Vulcan carbon black, to reach a total Pt+Ru content of 40 wt% on

carbon. The dispersion was then sonicated in an ice bath for 1 h and then sprayed on a gas

diffusion layer (Toray 60, 5 wt % PTFE wet-proofing) using an airbrush (Iwata Eclipse HP CS).

145

The obtained gas diffusion electrodes (GDE) and High Density polyethylene (HDPE)

membrane18 were then soaked for 20 min in 1 M KOH, and this was repeated two times. The

MEA was then assembled in the single-cell fuel cell hardware using Teflon gaskets, with gasket

thickness chosen to reach 25% compression. The AEMFC was operated using a Scribner 850e

Fuel cell test system, flowing H2/O2 at 1.0 L min-1 with a cell temperature of either 65°C or

80°C. The corresponding dew points were either 60 or 76°C at the cathode, and either 55 or

69°C at the anode. No back pressure (BP) has been applied for at 65°C (for both O2 and Air

fed cathode), while at 80°C a BP of 0.5 bar was applied to the anode and 1 bar to the cathode.

The choice of the dew points was made to reach the best power performance, while avoiding

flooding at the catalyst layers. The break-in was performed in potentiostatic mode at 0.5 V,

adjusting the relative humidity (RH) at both electrodes and decreasing the potential down to

0.11 V to reduce the quantity of carbonate formed during the exchange procedure done in

air.34,35

All AEMFC experiments have been carried out using a 57Fe0.5-NH3 loading of 0.91 mg cm-2 at

the cathode and a loading of 0.6 mgPt+Ru cm-2 at the anode. The initial activity and performance

of the MEA was evaluated in O2/H2 AEMFC at 65 and 80°C. Stability of the 57Fe0.5-NH3 cathode

was evaluated in air/H2 AEMFC at 65°C, applying chronopotentiometry (CP) at 600 mA cm-2

for a total duration of 100 h, during which the high-frequency resistance (HFR) was also

constantly measured by impedance spectroscopy. The CP experiment was regularly

interrupted to acquire polarisation curves during 100 h of test. The polarisation curves were

recorded by scanning the cell voltage from OCV to 0.1 V at a scan rate of 10 mV s-1.

57Fe Mössbauer spectra were measured at low temperature (5 K) on 57Fe0.5-NH3 cathodes

before and after H2/air AEMFC operation, with a 57Co:Rh source. The measurements were

carried out with a triangular velocity waveform, using NaI scintillation detector for detecting

the γ-rays. The velocity calibration was performed with an α-Fe foil. Two GDEs have been

studied with ex situ Mössbauer spectroscopy, one after break-in procedure and the other

after the break-in procedure and the stability test (including a cumulative 100 h of operation

in air at 600 mA cm-2 and all polarisation curves recorded during this test).

146

Results

57Fe0.5-NH3 was first electrochemically characterized in RDE configuration to evaluate its

activity towards ORR (Figure 1). The initial mass activity evaluated from the Tafel plot (Figure

1b) was ca 9 A g-1 at 0.9 V vs. RHE, similar to the high mass activity previously reported for

other ammonia-treated Fe-N-C catalysts in alkaline electrolyte.17 The polarization curve in

Figure 1a evidences fast kinetics for ORR, with the diffusion-limited current density of ca. 5.2

mA cm-2 around 0.6 V vs. RHE, indicating a near 4 e- pathway for ORR.

0.0 0.2 0.4 0.6 0.8 1.0

-6

-5

-4

-3

-2

-1

0

Curr

ent D

ensity / m

A c

m-2

Potential / V vs RHE

1 10 100

0.70

0.75

0.80

0.85

0.90

0.95

Pote

ntial / V

vs R

HE

Mass Activity / A g-1

a. b.

Figure 1. a) RDE polarisation curves of 57Fe0.5-NH3 in O2-satured 0.1 M KOH. The scan rate was

1 mV s-1, rotation rate 1600 rpm and the catalyst loading 0.2 mg cm-2. The curves are not

corrected for Ohmic losses. b) The semi-logarithmic Tafel plot obtained from the polarization

curve (left hand side) applying the Koutecky-Levich equation and using the value of diffusion-

limited current density observed at 0.4 V vs. RHE.

AEMFC polarization curves in O2/H2 (Figure 2a-2b) were obtained at 65°C and 80°C cell

temperature on the MEA combining 57Fe0.5-NH3 cathode, PtRu/C anode and HDPE-based

AEM. The latter was recently reported for its high conductivity and improved stability.18 At

65°C, the peak power density was 1.4 W cm-2 reached at ca 0.4 V, while the initial activity at

0.9 V was ca 75 mA cm-2 (see inset of Figure 2a), comparable to the activity in AEMFC seen in

chapter 3 for the natural-iron Fe0.5-NH3 cathode. Below 0.2 V, the polarization curve shows a

147

peculiar shape similar to a hook, characteristic for water management issues in the MEA. At

80°C, the activity at 0.9 V is unmodified, while the mass-transport properties improved,

leading to higher performance at all cell voltages below 0.85 V. The peak power density is

now 1.7 W cm-2 reached at ca 0.45 V.

0 1 2 3 4 5 6

0,0

0,5

1,0

1,5

2,0

Pow

er

density / W

cm

-2

Current Density / A cm-2

0 1 2 3 4 5 6

0,0

0,2

0,4

0,6

0,8

1,0

Tcell

= 65°C

Tcell

= 80°C

Cell

Voltage / V

Current Density / A cm-2

0,0 0,1 0,2 0,3 0,40,80

0,85

0,90

0,95

Cell V

olt

ag

e / V

Current Density / A cm-2

a.

b.

Figure 2. Effect of cell temperature on H2/O2 AEMFC polarization (a) and power density curves

(b), using 57Fe0.5-NH3 cathode (catalyst loading 0.91 mg cm-2, BP 1 bar), PtRu/C anode (Pt + Ru

loading 0.6mgPGM cm-2, BP 0.5 bar) and HDPE AEM. Both catalysts are combined with ETFE

powder AEI (ionomer/carbon ratio of 0.4 for PtRu/C, and 0.2 for 57Fe0.5-NH3). The curves are

measured with a scan rate of 10 mV s-1. The cathode and anode dew points are 60°C and 55°C

at Tcell=65°C; and 76°C and 69°C at Tcell=80°C. The flow rate are 1 L min-1 on both sides.

148

Furthermore, the hook shape at high current density is no longer seen at 80°C, probably due

to improved water management allowed by the higher possible partial pressures of water

vapour at higher temperature.

We now compare the initial activity and performance seen in Figure 2 to state-of-art activities

and performances reported for similar operating conditions (80°C), MEA preparation method

(ETFE ionomer) and same anode catalyst, but for different cathode catalysts.16,18,19 Figure 3

reports the current density at either 0.85 V or 0.7 V cell voltage. The voltage of 0.85 V was

selected to be representative of ORR activity, and also because 0.9 V leads to high noise in

the current density values due to the use of an electronic load in the fuel cell test system. In

contrast, the cell voltage of 0.7 V was chosen to represent the cell performance, when the

cell delivers high electric power but at an acceptable energy efficiency. Figure 3 highlights that

the 57Fe0.5-NH3 cathode (0.91 mg cm-2) has an activity at 0.85 V comparable to that of 0.4 mgPt

cm-2 cathode and much higher than the activity of CoFe2O4/C and CoOx/C cathodes (metal

loadings of 2.4 mg cm-2) and of an Ag/C cathode (0.85 mgAg cm-2). 16,18,19 At 0.7 V, the 57Fe0.5-

NH3 cathode also results in the highest current density among the state of art PGM-free

cathodes, but the difference is less strong than when the activities at 0.85 V are compared.

This may be due to increased mass-transport limitations for 57Fe0.5-NH3, which seems

reasonable due to the much lower site density in this catalyst (only ca 1.5-2.0 wt % Fe, i.e.

13.5-18.0 µgFe·cm-2 in the presently investigated cathode) compared to the other PGM-free

metal-based cathodes and also compared to the Pt/C cathode. The microporous nature of

carbon in 57Fe0.5-NH3 may also contribute to mass-transport issues that are specific to this

catalyst. The comparison therefore underlines the high ORR activity in AEM environment of

FeNx active sites after ammonia pyrolysis.

149

0

500

1000

1500

2000

2500

LDPE

65°C

CoOx/NC

19

Cu

rre

nt

De

nsity /

mA

cm

-2

J at 0,85V

J at 0,7V

HDPE

70°C

CoFe2O

4/C

16

HDPE

80°C

Pt/C 18

HDPE

80°C

Ag/C 18

HDPE

80°C

Fe0.5

-NH3

Figure 3 Comparison of AEMFC results obtained with previous state of art PGM and PGM-free

cathodes and with the present Fe0.5-NH3 cathode, in similar conditions of cell temperature,

anode loading and MEA preparation procedure and AEM or AEI materials. From left to right:

75wt% CoOx/NC cathode (metal loading 2.4 mg cm-2, BP 0.8 bar), AEM LDPE, anode PtRu/C

(loading in Pt+Ru 0.7 mg cm-2, BP 1.2 bar), cell temperature of 65°C;19 50wt% CoFe2O4/C

(metal loading 2.4 mg cm-2, no BP), anode PtRu/C (loading in Pt+Ru 0.7 mg cm-2, no BP), AEM

HDPE, cell temperature 70°C;16 40wt% Pt/C (loading in Pt 0.4 mg cm-2), AEM HDPE, anode

PtRu/C (loading in Pt 0.4 mg cm-2), cell temperature 80°C;18 40wt% Ag/C (Ag loading 0.85 mg

cm-2), AEM HDPE, anode PtRu/C (loading in Pt 0.4 mg cm-2), cell temperature 80°C;18 57Fe0.5-

NH3 (loading 0.91 mg cm-2, BP 1 bar) AEM HDPE, anode PtRu/C (loading in Pt+Ru 0.6 mg cm-

2, BP 0.5 bar), cell temperature 80°C.

After the determination of the activity and initial performance of the Fe-0.5-NH3 cathode in

both RDE and AEMFC fed with pure oxygen, further testing was performed with synthetic air,

free of CO2. Figure 4a shows the air/H2 polarization curves obtained at 65°C cell temperature,

evidencing a promising initial activity of 20 mA cm-2 at 0.9 V, ca 20 % the value reached in

pure oxygen, which is in line with the reduced concentration of O2 in air (21 %). The

polarization curve has a maximum current density of ca 2.4 A cm-2 and a peak power density

150

of 0.7 W cm-2, underlying the great performance of the 57Fe0.5-NH3 cathode in AEMFC fed with

air.

The MEA durability was then determined applying 100 h chronopotentiometry at 600 mA cm-

2 at a cell temperature of 65°C. Figure 4b shows a restricted decrease in the cell voltage from

0.69 V to 0.59 V over 100 h at 0.6 A·cm-2, which is accompanied by a continuous increase in

the HFR with time. The magnitude of the HFR increase (ca 7 mΩ·cm2) can however not explain

the magnitude of the drop in cell voltage with time at 0.6 A·cm-2. One would expect only ca 4

mV drop in cell voltage expected from increased HFR, while a drop of ca 100 mV is observed.

0,0 0,5 1,0 1,5 2,0 2,50,0

0,2

0,4

0,6

0,8

1,0

Current Density / A cm-2

Ce

ll V

olta

ge

/ V

0,0

0,2

0,4

0,6

0,8

Po

we

r D

en

sity /

W c

m-2

0 10 20 30 40 50 60 70 80 90 1000,0

0,2

0,4

0,6

0,8

1,0

Cell voltage at 600 mA cm-2

HFR

Time / h

Ce

ll V

olta

ge

/ V

25

30

35

HF

R /

mO

hm

cm

2

Figure 4. Durability measurement in H2/Air AEMFC for an MEA comprising a 57Fe0.5-NH3

cathode (catalyst loading 0.91 mg cm-2), PtRu/C anode (40 wt% PGM, PtRu loading 0.6 mgPGM

cm-2) and HDPE AEM. Both catalysts are combined with ETFE powder AEI (ionomer/carbon

ratio 0.41 for PtRu/C anode catalyst, 0.2 for 57Fe0.5-NH3 cathode catalyst). a) Initial polarization

and power density curves measured at a cell temperature of 65°C, dew points of 63°C and

58°C for cathode and anode respectively. The curves are measured at a scan rate of 10 mV s-

1. b) Chronopotentiometry (CP) at 600 mA cm-2 at 65°C cell temperature.

151

The origin of cell voltage loss can thus not be determined from the voltage- and HFR-vs-time

curves. To better understand the reason of the drop in the cell voltage with time, several

polarization curves has been collected to follow the changes in the performance at low and

high current density and the variation of the HFR. The main changes take place during the

first 50-60 h of operation, showing the decrease in current density (Fig. 5a, c and d) and the

parallel increase in the HFR (Fig. 5b and e). The decrease in the current density at high

potential (0.8 V) is 36 % between the initial and final polarisation curves (Fig. 5a and c,

compare blue and red curves/bars) can be mainly related to the cathode catalyst, as will be

later demonstrated by Mössbauer spectroscopy. A decrease in ORR activity of only 36%

should however, according to Tafel law with Tafel slope of 75 mV/dec, only lead to a negative

shift of ca 15 mV of the polarisation curve, at any given current density. This is not in line with

what is experimentally observed. We can thus conclude that the cell voltage loss is, for ca 15

mV, coming from decreased ORR activity, but the main voltage loss observed over 100 h must

be assigned to other factors, such as decreased OH-conductivity inside the electrodes, and/or

to decreased mass-transport properties. One previously saw that the decrease of cell voltage

with time is paralleled with increased HFR. The latter is related to the decrease of the OH-

conductivity in the membrane, but the OH-conductivity inside electrodes may have

experienced stronger decrease than the AEM, due to different local conditions during

operation. Further degradation mechanisms, related to the water management, have to be

taken into account, e.g. the dry-out of the cathode during operation, which could lead to

much enhanced ionomer degradation. The strong effect of water content on ionomer

chemical degradation was reported by Dekel’s group, lower water content leading to faster

degradation of quaternary ammonium groups.20

152

Figure 5. Evolution of key characteristics of the MEA at different times during the 100 h

durability test. a. Polarization curve collected at different times; b. HFR vs. current density

curves, as measured during each of the polarisation curves shown in a); c. bar plot of current

density at 0.8 V from the polarisation curves shown in a) ; d. bar plot of current density at 0.6

V from the polarisation curves shown in a) ; e. bar plot of the HFR collected at 0.6 A cm-2

corresponding to each of the polarisation curves shown in a).

The durability for the Fe-0.5-NH3 catalyst in AEMFC is thus good, losing only ca 30% at 0.85 V.

This strongly contrasts with the observations in PEMFC for the same catalyst and that

evidenced a very fast decay. The current density at 0.8 V was divided by ca a factor 10 during

the first 25 h of operation in PEMFC (Figure 3a in Ref. 21). The OH-rich environment of AEMFC

clearly leads to a better stability and durability of the Fe-based active sites present in

ammonia-pyrolyzed Fe-N-C catalysts.17

a.

b.

c.

d.

e.

153

To investigate possible structural changes in the cathode catalyst during the durability test,

the Fe species were analysed with Mössbauer spectroscopy, recorded before and after the

100 h durability test. The spectra were collected at low temperature (5 K) to be able to

separate the contributions of iron nano-oxides (leading to a spectral double component with

parameters very similar to D1 at room temperature, but leading to a sextet spectral

component at 5 K) from those of high-spin Fe(III)Nx species (doublet D1 at 5 K, which

represent surface-exposed sites that bind O2 or OH at OCP).22 The presence of D1 and D2

components is expected for pyrolyzed Fe-N-C materials. 17,23-25 The presence of iron oxide in

the BoT MEA measurement (Figure 6a) but not in the as-prepared catalyst (Figure 2 in Chapter

2) is due either to i) MEA preparation, or ii) MEA break-in in AEMFC. After the 100 h durability

test, the Mössbauer spectra shows a decrease in the absolute intensity of the D1, while the

absolute intensity of D2 and of the iron oxide stayed unmodified (Figure 6b-c). Both D1 and

D2 may contribute to the initial ORR activity, and are believed to be more active than Fe2O3.

Results in figure 6 can be combined to the results summarized in figure 5: during the CP at

600 mA cm-2, one can attribute the loss of performance at 0.8 V to the partial loss of D1-

species. If D1 was the only species that contributes to the activity of the catalyst, then the

relative decrease in the absolute area below D1 from BoT to EoT (around -53 %) should

translate in a corresponding relative decrease of the current density at 0.8 V. However, this is

not the case, the current density decrease being less pronounced (about -34 % at 0.8 V from

BoT to EoT) . In contrast, considering both the D1 and D2 species as active toward ORR, the

relative decrease of D1+D2 absolute area is -36 % from BoT to EoT (D2 absolute intensity

remaining unchanged), corresponding well with the relative loss in the current density at 0.8

V (-34%). This supports the idea that both D1 and D2 are active FeNx sites for ORR in AEMFC,

and that D2 is more stable than D1, resulting in nearly unchanged ORR activity between 50

and 100 h operation.

154

-6 -4 -2 0 2 4 60,975

0,980

0,985

0,990

0,995

1,000

-8 -4 0 4 8

0,98

0,99

1,00

c

Exp.

Total fit D1 (36%) D2 (37%) Oxide (27%)

Tra

nsm

issio

n

Velocity / mm s-1

0,98

0,99

1,00

Exp. Total fit

D1 (51%) D2 (27%) Oxide (22%)

Tra

nsm

issio

n

b

Tra

nsm

issio

n

Velocity / mm s-1

a

Figure 6. Comparison between the 57Fe Mössbauer spectra of the 57Fe0.5-NH3 cathode before

(a) and after (b) 100 h operation in AEMFC. Ex situ Mössbauer spectra on the cathode, either

right after the breaking procedure in AEMFC (a) and after the 100 h durability test in AEMFC

(b). The spectra were collected at 5 K. The green doublet indicates the doublet D1 assigned

to high spin Fe(III)-Nx sites, the blue curve the doublet D2 assigned to low or medium spin

Fe(II)Nx sites,22 and the sextet in magenta is assigned to iron oxide (Fe2O3). In (c) the spectra

before and after 100 h test are superimposed, evidencing that the absolute intensity of D1

decreased, while that of D2 and the sextet assigned to Fe2O3 were unchanged.

155

Conclusions

Fe0.5-NH3 showed high activity in RDE, with a mass activity of around 9 A g-1 at 0.9 V. This

catalyst has been tested in H2/O2 AEMFC, which results in high peak power density (ca. 1.7

reached at 0.4 V) and current density (maximum of 5.5 A cm-2) at a cell temperature of 80°C,

while slightly lower performance at 65°C was observed (1.4 W cm-2; maximum of 4.7 A cm-2).

These results are above the previously best reported results for a non-PGM cathode catalyst

in AEMFC.

In H2/Air AEMFC at 65°C, the MEA with the Fe0.5-NH3 cathode reached a peak power density

of 0.7 W cm-2 at 0.4 V and a maximum current density around 2.4 A cm-2. Stability test at 600

mA cm-2 for 100 h showed a 100 mV decrease in the cell voltage, underlying the promising

stability of the MEA. In addition, the polarization curves obtained before and after the stability

test have been compared showing a small decrease in the ORR activity at 0.8 V (about -34 %),

which underline the great stability of the Fe0.5-NH3 cathode catalyst. To better analyse the

reason of the loss in the current density at high voltage, Mössbauer spectra before and after

stability tests have been compared, showing a decrease in the intensity in the signal of the

quadrupole doublet component 1 while the quadrupole doublet 2 remained unmodified.

Comparison of the relative loss in D1 signal or D1+D2 signal intensities and the relative loss in

current density at 0.8 V, we can infer that both the FeNx sites corresponding to D1 and D2

contribute to the ORR activity, and that D2 is highly stable.

At lower cell voltage, the loss in cell performance observed with time can however not only

be assigned to loss in ORR activity, and can also not be assigned to increased AEM resistance.

We hypothesize that the main reasons for cell performance decrease over time is due to

either reduced mass-transport properties inside active layers and/or to strongly reduced OH-

conductivity inside the active layers, and in particular inside the cathode layer due to

expectedly dryer environment, leading to a more facile nucleophilic attack of the ETFE AEI

than in the AEM and in the anode layer.

156

References

1 Piana, M., Catanorchi, S. & Gasteiger, H. A. Kinetics of Non-Platinum Group Metal Catalysts for

the Oxygen Reduction Reaction in Alkaline Medium. ECS Transactions 16, 2045-2055,

doi:10.1149/1.2982044 (2008).

2 Meng, H., Jaouen, F., Proietti, E., Lefèvre, M. & Dodelet, J.-P. pH-effect on oxygen reduction

activity of Fe-based electro-catalysts. Electrochemistry Communications 11, 1986-1989,

doi:10.1016/j.elecom.2009.08.035 (2009).

3 Wang, L. et al. An optimised synthesis of high performance radiation-grafted anion-exchange

membranes. Green Chemistry 19, 831-843, doi:10.1039/c6gc02526a (2017).

4 Poynton, S. D. et al. Preparation of radiation-grafted powders for use as anion exchange

ionomers in alkaline polymer electrolyte fuel cells. Journal of Materials Chemistry A 2, 5124-

5130, doi:10.1039/C4TA00558A (2014).

5 Gonçalves Biancolli, A. L. et al. ETFE-based anion-exchange membrane ionomer powders for

alkaline membrane fuel cells: a first performance comparison of head-group chemistry. Journal

of Materials Chemistry A 6, 24330-24341, doi:10.1039/c8ta08309f (2018).

6 Britton, B. & Holdcroft, S. The Control and Effect of Pore Size Distribution in AEMFC Catalyst

Layers. Journal of The Electrochemical Society 163, F353-F358, doi:10.1149/2.0421605jes

(2016).

7 Bellini, M. et al. Palladium-ceria catalysts with enhanced alkaline hydrogen oxidation activity

for anion exchange membrane fuel cells. ACS Applied Energy Materials (2019).

8 Wang, Y. et al. Pt–Ru catalyzed hydrogen oxidation in alkaline media: oxophilic effect or

electronic effect? Energy & Environmental Science 8, 177-181, doi:10.1039/c4ee02564d

(2015).

9 Mamlouk, M., Scott, K., Horsfall, J. A. & Williams, C. The effect of electrode parameters on the

performance of anion exchange polymer membrane fuel cells. International Journal of

Hydrogen Energy 36, 7191-7198, doi:10.1016/j.ijhydene.2011.03.074 (2011).

10 Omasta, T. J. et al. Beyond 1.0 W cm−2 Performance without PlaGnum: The Beginning of a New

Era in Anion Exchange Membrane Fuel Cells. Journal of The Electrochemical Society 165, J3039-

J3044, doi:10.1149/2.0071815jes (2018).

11 Pham, C. V., Britton, B., Böhm, T., Holdcroft, S. & Thiele, S. Doped, Defect-Enriched Carbon

Nanotubes as an Efficient Oxygen Reduction Catalyst for Anion Exchange Membrane Fuel Cells.

Advanced Materials Interfaces 5, 1800184, doi:10.1002/admi.201800184 (2018).

12 Arunchander, A., Vivekanantha, M., Peera, S. G. & Sahu, A. K. MnO–nitrogen doped graphene

as a durable non-precious hybrid catalyst for the oxygen reduction reaction in anion exchange

membrane fuel cells. RSC Advances 6, 95590-95600, doi:10.1039/c6ra20627a (2016).

13 Serov, A. et al. Development of Electrocatalysts for Anion Exchange Membrane Fuel Cells.

Meeting Abstracts MA2018-01, 1734 (2018).

14 Hossen, M. M., Artyushkova, K., Atanassov, P. & Serov, A. Synthesis and characterization of high

performing Fe-N-C catalyst for oxygen reduction reaction (ORR) in Alkaline Exchange

Membrane Fuel Cells. Journal of Power Sources 375, 214-221,

doi:10.1016/j.jpowsour.2017.08.036 (2018).

15 Wang, L., Brink, J. J. & Varcoe, J. R. The first anion-exchange membrane fuel cell to exceed 1 W

cm−2 at 70 °C with a non-Pt-group (O2) cathode. Chemical Communications 53, 11771-11773,

doi:10.1039/C7CC06392J (2017).

16 Peng, X. et al. High-Performing PGM-Free AEMFC Cathodes from Carbon-Supported Cobalt

Ferrite Nanoparticles. Catalysts 9, 264, doi:10.3390/catal9030264 (2019).

17 Santori, P. G. et al. Effect of Pyrolysis Atmosphere and Electrolyte pH on the Oxygen Reduction

Activity, Stability and Spectroscopic Signature of FeNx Moieties in Fe-N-C Catalysts. Journal of

The Electrochemical Society 166, F3311-F3320, doi:10.1149/2.0371907jes (2019).

157

18 Wang, L., Peng, X., Mustain, W. E. & Varcoe, J. R. Radiation-grafted anion-exchange

membranes: the switch from low- to high-density polyethylene leads to remarkably enhanced

fuel cell performance. Energy & Environmental Science 12, 1575-1579,

doi:10.1039/c9ee00331b (2019).

19 Peng, X. et al. Nitrogen-doped Carbon–CoOx Nanohybrids: A Precious Metal Free Cathode that

Exceeds 1.0 W cm−2 Peak Power and 100 h Life in Anion-Exchange Membrane Fuel Cells.

Angewandte Chemie International Edition 58, 1046-1051, doi:10.1002/anie.201811099 (2019).

20 Dekel, D. R. Review of cell performance in anion exchange membrane fuel cells. Journal of

Power Sources 375, 158-169, doi:10.1016/j.jpowsour.2017.07.117 (2018).

21 Chenitz, R. et al. A specific demetalation of Fe–N4 catalytic sites in the micropores of NC_Ar +

NH3 is at the origin of the initial activity loss of the highly active Fe/N/C catalyst used for the

reduction of oxygen in PEM fuel cells. Energy & Environmental Science 11, 365-382,

doi:10.1039/c7ee02302b (2018).

22 Mineva, T. et al. Understanding Active Sites in Pyrolyzed Fe–N–C Catalysts for Fuel Cell

Cathodes by Bridging Density Functional Theory Calculations and 57Fe Mössbauer

Spectroscopy. ACS Catalysis, 9359-9371, doi:10.1021/acscatal.9b02586 (2019).

23 Zitolo, A. et al. Identification of catalytic sites for oxygen reduction in iron- and nitrogen-doped

graphene materials. Nature materials 14, 937-942, doi:10.1038/nmat4367 (2015).

24 Kramm, U. I. et al. Structure of the catalytic sites in Fe/N/C-catalysts for O2-reduction in PEM

fuel cells. Physical chemistry chemical physics : PCCP 14, 11673-11688,

doi:10.1039/c2cp41957b (2012).

25 Li, J. & Jaouen, F. Structure and activity of metal-centered coordination sites in pyrolyzed

metal–nitrogen–carbon catalysts for the electrochemical reduction of O2. Current Opinion in

Electrochemistry 9, 198-206, doi:10.1016/j.coelec.2018.03.039 (2018).

158

Conclusions

The present thesis focused on the investigation of electrocatalytic materials for oxygen

reduction reaction and hydrogen peroxide reduction reaction in alkaline medium, and in

particular the performance of an ammonia treated Fe-N-C catalyst from electrochemical

liquid electrolyte studies in alkaline medium and up to is activity and stability in AEMFC.

It has been demonstrated that the pyrolysis environment plays a critical role on the oxygen

reduction reaction activity and/or stability of Fe-N-C catalysts in acid and alkaline media. The

Fe-N-C catalyst pyrolyzed in argon atmosphere, Fe0.5-Ar, demonstrated similar activity at pH

1 and 13, while the stress test evidenced higher stability in alkaline medium. The Fe-N-C

catalyst pyrolyzed in ammonia, Fe0.5-NH3, shows a much higher activity at pH 13 than at pH 1,

and is also more active than Fe0.5-Ar, especially at pH 13. The increased activity of Fe0.5-NH3 is

due in part to its higher BET area, but mostly due to a slightly different coordination

environment, with lower coordination number evaluated at high potential with in situ XAS.

The Fe centers in the catalyst Fe0.5-NH3 are binding O2 less strongly than those in Fe0.5-Ar,

leading to a higher turnover frequency per active site. This higher turnover frequency is

durable in alkaline electrolyte, but not in acidic electrolyte. Operando dissolution experiments

revealed a 10 times higher iron leaching in acid medium for Fe0.5-NH3 vs. Fe0.5-Ar, while both

catalysts leached a very small amount of Fe in alkaline medium.

One drawback of alkaline medium is the higher production of hydrogen peroxide during ORR,

due to the non-negligible activity of carbon and nitrogen-doped carbon towards two-

electrons ORR. A series of four manganese oxides has been investigated as a peroxide

scavenger co-catalysts, with the goal to reduce the amount of HO2- produced during ORR on

Fe-N-C catalysts. Those Mn-oxide materials showed good activity for both the HPRR and ORR,

improving the selectivity of MnOx/Fe-N-C composites relative to Fe-N-C alone, and without

negatively affecting the ORR activity. In addition, operando dissolution measurements

showed that the addition of Mn2O3 further improved the stability of Fe0.5-NH3 in RDE

configuration. However, the main issue of this class of Mn-oxide materials derives from their

poor intrinsic stability, which has been shown to be higher in O2- than in Ar-saturated alkaline

electrolyte. The presence of HO2- was identified as the parameter triggering the increased

dissolution of manganese from Mn-oxides in alkaline. Mn2O3 was shown to have a lower Mn

159

leaching rate (even after normalization by BET area) compared to the others three Mn-oxides,

and was thus selected for testing a composite Mn-oxide/Fe-N-C catalyst in AEMFC.

In a separate study, the effect of the anion exchange ionomer on the activity of different

classes of ORR and HOR catalysts in RDE was investigated, also using Nafion as reference

proton conducting ionomer. It has been demonstrated that the choice of the ionomer can be

critical, in particular for catalysts with a low density of metal-based active sites, such as Fe-N-

C and the Pd-CeO2/C catalyst with low wt% Pd. The study underlined the necessity of a

screening of the compatibility between novel catalysts and ionomers, before such

combinations are investigated in AEMFC device.

The last part of the work evaluated the activity, power and durability performance of Fe-N-C

and Mn2O3/Fe-N-C cathodes in an operating AEMFC. The results highlighted the excellent

initial activity and power performance of Fe0.5-NH3 cathode fed with oxygen or air. The

durability is also very promising, with only circa 30% loss of electrochemical activity at 0.85 V

after 100 h of operation at 0.6 A·cm-2 in air. While this activity loss should have resulted in

only ca 15 mV loss of cell voltage at fixed current density, a drop of circa 100 mV was

observed. A trend of restricted increased high frequency resistance with time was seen for

the cell, but not in line with the magnitude of the cell voltage loss. It is hypothesized that

either the mass-transport properties of electrodes changed, or the local conductivity in

electrodes decreased (especially at the cathode, due to dryer operando conditions). The

durability of Fe0.5-NH3 is nevertheless very promising in AEMFC and much higher than in

PEMFC.

In conclusion, the experiments performed open new routes for the application of Fe-N-C

based catalysts in AEMFC. Further studies are needed to evaluate the durability after longer

periods of time in AEMFC and in different steady-state and dynamic conditions mimicking a

driving cycle. Optimization of the layer preparation and choice of ionomer is also still needed,

to match the state of art performance obtained with Pt/C catalysts in AEMFC. While the best

Fe-N-C catalyst in this study matched the activity of a 0.4 mgPt cm-2 cathode at high potential,

the performance at lower cell voltage was still ca 50 % lower, possibly due to mass-transport

160

issues. Further durability studies in AEMFC are also needed on composite cathode catalysts

based on Mn-oxides and Fe-N-C, or involving other metal-oxides that are more durable than

Mn-oxides. Finally, research on non-PGM anode catalysts will be critical for PGM-free

AEMFCs.

161

162

Appendix

Experimental Procedure

Introduction

The following Appendix summarize all the experimental procedure presented in the previous

chapter of the thesis.

Synthetic procedure

Fe-N-C

During the PhD thesis, two types of Fe-N-C catalysts has been prepared mixing together the

precursor composed of Zif-8 (Basolite Z1200, Sigma Aldrich) as sacrificial support, 1,10-

phenantroline (≥99%, Sigma Aldrich) as a secondary source of nitrogen and iron (II) acetate

(≥99.99%, Sigma Aldrich) as the source of metal keeping the same ratio of 80:20:0.5

respectively. The catalyst precursors are initially placed in zirconia jars with 100 balls

(diameter ca. 5mm, 0.4g per ball) and mixed together using low energy ball milling at 400 rpm

for 2 h 20 min (30 min of milling and 5 minute of pause). As obtained precursor powder is

then placed into quartz boat and inserted in a quartz tube, then saturated with argon. The

first pyrolysis in carried at 1050°C for 1h, pushing the quartz boat containing the catalyst

precursor powder within the heating zone using an outer magnet. The pyrolysis is terminated

by opening the split-hinge furnace, removing the quartz tube and letting it cool down at room

temperature for 20 minutes. As obtained catalyst is labelled Fe0.5-Ar. To prepare the NH3-

pyrolyzed catalyst, Fe0.5-Ar is re-pyrolysed with the same flash-pyrolysis mode, but in flowing

pure NH3 and for only 5 minutes at 950 °C. As obtained catalyst is labelled Fe0.5-NH3.

Manganese Oxides

The preparation of four different, phase-pure, manganese oxides was targeted, namely the

α‒, β‒, δ‒MnO2 and α‒Mn2O3 phases. Their synthesis was performed according to the

literature. In short, α‒MnO2 was obtained by reducing KMnO4 (3.16 g) in a mixture of water

(200 mL) and fumaric acid (0.78 g), kept under stirring at room temperature for 30 minutes.

163

The resultant gel is settled for 1 h and then filtered, washed with ultrapure water and dried

to yield α‒MnO2.39 The reduction of KMnO4 (0.395 g) to δ‒MnO2 was carried out under

stirring over 3 h in a mixture of water (80 mL), H2SO4 (96%, 70.2 µL) and ethanol (3 mL).40 The

dried powder was calcined at 450 °C to yield β‒MnO2.41 α‒Mn2O3 was obtained by calcining

γ‒MnOOH at 550 °C.14 The latter was obtained via KMnO4 (0.2 g) reduction together with

Mn(CH3COO)2·4H2O (1.2 g) in water (150 mL) over 12 h, keeping the mixture under stirring

and refluxed.42 Subsequently, the four manganese oxides were dispersed on carbon black

(CABOT, Vulcan® XC72R), with a ratio of 1:5, placed in zirconia jars with 100 balls (0.4 g per

ball) and mixed at low energy ball milling (Fritsch, Pulverisette 7) for 10 minutes at 200 rpm.

Composite MnOx/Fe0.5-NH3

The manganese oxides are combined to Fe0.5-NH3 in a weight ration of 1/4 by manual grinding

using agate mortar and pestle until the powder is uniformly dispersed.

Characterization methods

Rotating Disk Electrode (RDE) Rotating Ring Disk Electrode (RRDE)

Electrochemical measurements were carried out using a Bio-Logic® SP-300 dual channel

potentiostat connected to the three-electrodes electrochemical cell. The reference electrode

was Reversible Hydrogen Electrode (RHE) with hydrogen gas flowed through the electrolyte

contained in a glass frit; as counter electrode was used a graphite rod; as working electrodes

either rotating disk electrode (RDE) with a 5 mm diameter glassy carbon tip and rotating ring

disk electrode (RRDE) with a 5.6 mm diameter glassy carbon tip surrounded by a platinum

ring (used to measure the amount of hydrogen peroxide released in solution during the ORR).

The ink is prepared by adding in sequence 5 mg catalyst, 54 µL Nafion® (5% perfluorinated

resin solution), 744 µL ethanol, 92 µL ultrapure water and sonicating for 1 hour in an ice bath.

An aliquot of 7 µL of the ink is pipetted on the GC disk and dried at room temperature,

resulting in a catalyst loading of 200 µg cm-2. The Initial activity is measured in O2-saturated

electrolyte at a scan rate of 1 mV·s-1 (SP-300, BioLogic Potentiostat) and at rotation rate of

1600 rpm. Due to the low scan rate and low catalyst loading, no correction for the capacitive

current is needed.

164

Figure S1. Scheme of the R(R)DE configuration.

The ink is prepared by adding in sequence 5 mg catalyst, 54 µL Nafion® (5% perfluorinated

resin solution), 744 µL ethanol, 92 µL ultrapure water and sonicating for 1 hour in an ice bath.

An aliquot of 7 µL of the ink is pipetted on the GC disk and dried at room temperature,

resulting in a catalyst loading of 200 µg cm-2. The Initial activity is measured in O2-saturated

electrolyte at a scan rate of 1 mV·s-1 (SP-300, BioLogic Potentiostat) and at rotation rate of

1600 rpm. Due to the low scan rate and low catalyst loading, no correction for the capacitive

current is needed. To evaluate the durability of the catalysts in RDE set-up, a load-cycling

protocol is applied, comprising 5000 triangular cycles performed at a scan rate of 100 mV·s-1

in the potential range of 0.6-1.0 V vs RHE and in N2-saturated electrolyte. The ORR activity

after the AST is measured after re-saturating the solution with O2.

165

Scanning Flow Cell – Inductively Coupled Plasma Mass Spectroscopy (SFC-ICP/MS)

6.7.2.1 Fe leaching

Figure S2. Scheme of a SFC-ICP/MS

The leaching of Fe during short electrochemical cycling, performed either before or after the

AST, is investigated in an on-line electrochemical scanning flow cell (SFC) directly connected

to an inductively-coupled plasma mass spectrometer (ICP–MS), previously developed by us.

It is however challenging to continuously measure Fe leaching over the length of the AST, due

to drift of the ICP-MS with time and need for constant recalibration. To measure 56Fe, the

ICP–MS (Perkin Elmer, NexION 350) is operated in dynamic-reaction-cell mode, using

methane as the reaction gas. The cell is calibrated to both an acidic and alkaline standard

solution of iron to ensure maximized detection of 56Fe. Daily calibration of the ICP–MS is done

by a four-point calibration curve (0, 0.5, 1.0, 5.0 μg·L−1) of standard iron solutions prepared

from Merck Centripur® ICP standards (Fe(NO3)3, 1000 mg·L−1, in 2–3% HNO3). As an internal

standard, we use 58Co (Merck Centripur®, Co(NO3)2, 1000 mg·L−1, in 2–3% HNO3) diluted to 50

μg·L−1 in HNO3 (0.15 mol·L−1) to ensure full acidification of the electrolyte in a y- connector

before its introduction in the ICP–MS. The SFC consists of a three-electrode setup using a

Ag/AgCl (Metrohm, 3 M KCl) reference electrode, a graphite rod counter electrode and a GC

RDE as a working electrode, on which the catalyst is drop cast. A positioning stage (Physik

Instrumente, M-403.6 DG) is used to approach individual catalyst spots on the working

Electrolyte

Scanning Flow Cell

ICP/MS

166

electrode. Stability measurements are conducted in alkaline (99.99%, Suprapur®, NaOH, 0.05

mol·L−1) as well as in acidic (Suprapur®, 0.05 mol·L−1 H2SO4) electrolyte. The potentiostat

(Gamry, Reference 600) as well as purging gases and the positioning stage is controlled by a

custom LabVIEW software. The catalyst ink is prepared from the Fe-N-C catalyst, Nafion® (5%

perfluorinated resin solution) and water, with a mass ratio of catalyst/dry-ionomer of 4 and a

catalyst concentration of 3.3 g·L−1 in the liquid ink. An aliquot of 2.75 μL is deposited on the

GC, resulting in a catalyst loading of 400 μg·cm−2. Such a high loading is necessary to reach a

sufficient signal-to-noise ratio in the ICP-MS measurements. This is due to the

aforementioned interference of the 40Ar16O dimers and high background noise of iron in

alkaline solution. For Fe-leaching measurement before and after the AST, the latter is

conducted in a separate Teflon RDE-cell containing 100 mL electrolyte, and the RDE tip is then

quickly transferred from the Teflon cell to the SFC set-up, with the catalyst still wetted by

electrolyte. The RDE cell used for the AST consists of four individual compartments, one each

for the three electrodes and for the purging tube. The counter and reference electrodes are

the same as in the SFC setup.

6.7.2.2 Mn leaching

For all stability investigations we used an inductively coupled plasma mass spectrometer

(ICP-MS) (Perkin Elmer, NexION 350x) for the in situ detection of dissolved manganese

species, by coupling it to the electrolyte outlet of a custom made polycarbonate SFC. On the

electrolyte inlet (angled 60 ° to the outlet) the SFC was connected to a graphite counter

electrode (Sigma Aldrich, 99.995% trace metal basis). A third capillary channel connected the

reference electrode (Metrohm, Ag/AgCl) closely to the working electrode surface. Contact

with the working electrode was made with a translation stage (Physik Instrumente, M-403).

A Potentiostat (Gamry, Reference 600) was used to employ all electrochemical protocols

during dissolution measurements. The purged electrolytes flow rate was controlled and

regularly calibrated by the peristaltic pump of the ICP-MS (Elemental Scientific, MP2 Pump).

The sensitivity of the ICP-MS towards dissolved Mn ions was calibrated daily from Mn

calibration standards (Merck, Centripur®) and dissolution rates are normalized to the

geometric surface area of each catalyst spot. For further information on the experimental SFC

ICP-MS setup, please refer to our previous reports.

167

As an electrolyte NaOH (Merck, Suprapur) was dissolved in water to 0.05 M. For individual

experiments, an addition of small amounts of 30wt.% H2O2 (Merck) was achieved by multiple

dilution steps. The pH was controlled (Mettler Toledo, SevenExcellence) and used to convert

the potentials of the Ag/AgCl to the reversible hydrogen electrode (RHE) using

equation (Eq. 1).

ERHE =Eapplied + EAg/AgCl + 0.0591 × pH (Eq. 1)

Anion Exchange Membrane Fuel Cell Operation

For AEMFC testing, the catalytic inks were prepared manually grinding the catalyst (PtRu/C

anode catalyst or 57Fe0.5-NH3 cathode catalyst) and ETFE (ethylene tetrafluoroethylene)

powder ionomer with 1 mL of H2O and 9 mL of 1-propanol. The ETFE content was different at

the anode and cathode, corresponding to 20 wt % with respect to 57Fe0.5-NH3 and 40 wt %

with respect to the carbon content in the PtRu/C anode catalyst. The anode catalyst was

prepared by mixing 40wt%Pt-20wt%Ru/C (Johnson Matthey) with Vulcan carbon black, to

reach a total Pt+Ru content of 40 wt% on carbon. The dispersion was then sonicated in an ice

bath for 1 h and then sprayed on a gas diffusion layer (Toray 60, 5 wt % PTFE wet-proofing)

using an airbrush (Iwata Eclipse HP CS).

The obtained gas diffusion electrodes (GDE) and High Density polyethylene (HDPE) membrane

were then soaked for 20 min in 1 M KOH, and this was repeated two times. The MEA was then

assembled in the single-cell fuel cell hardware using Teflon gaskets, with gasket thickness

chosen to reach 25% compression. The AEMFC was operated using a Scribner 850e Fuel cell

test system, flowing H2/O2 at 1.0 L min-1 with a cell temperature of either 65°C or 80°C. The

corresponding dew points were either 60 or 76°C at the cathode, and either 55 or 69°C at the

anode. No back pressure (BP) has been applied for at 65°C (for both O2 and Air fed cathode),

while at 80°C a BP of 0.5 bar was applied to the anode and 1 bar to the cathode. The choice

of the dew points was made to reach the best power performance, while avoiding flooding at

the catalyst layers. The break-in was performed in potentiostatic mode at 0.5 V, adjusting the

relative humidity (RH) at both electrodes and decreasing the potential down to 0.11 V to

reduce the quantity of carbonate formed during the exchange procedure done in air.

All AEMFC experiments have been carried out using a 57Fe0.5-NH3 loading of 0.91 mg cm-2 at

the cathode and a loading of 0.6 mgPt+Ru cm-2 at the anode. The initial activity and performance

168

of the MEA was evaluated in O2/H2 AEMFC at 65 and 80°C. Stability of the 57Fe0.5-NH3 cathode

was evaluated in air/H2 AEMFC at 65°C, applying chronopotentiometry (CP) at 600 mA cm-2

for a total duration of 100 h, during which the high-frequency resistance (HFR) was also

constantly measured by impedance spectroscopy. The CP experiment was regularly

interrupted to acquire polarisation curves during 100 h of test. The polarisation curves were

recorded by scanning the cell voltage from OCV to 0.1 V at a scan rate of 10 mV s-1.

X-Ray Powder Diffraction (XRPD)

The crystalline structure and the purity of the four manganese oxides was verified using X-ray

Powder Diffraction (XRPD) with PANAlytical X’pert diffractometer in Bragg-Brentano

configuration, using a CuKα source (λ=1.5406 Å) in a 2θ range of 5-80° with a step size of

0.035°. The most intense peaks were compared to the literature using PANalytical X’Pert

Highscore Plus (version 3.0e).

X-Ray Absorption Spectroscopy (XAS)

The XAS spectra are collected at SAMBA beamline (synchrotron SOLEIL, Gif-sur-Yvette,

France) at the Fe K-edge using a double crystal Si 220 monochromator and a Canberra 35-

elements germanium detector for operando acquisition in fluorescence mode. The catalyst

ink (10 mg catalyst, 100 µL 5% Nafion solution and 50 µL ultrapure H2O) is prepared via

ultrasonication, and 50 µL is deposited on circa 3 cm2 circular area of a larger conductive

carbon foil, resulting in a catalyst loading of circa 1 mg cm-2. The carbon foil is then inserted

in a three-electrode cell, 0.1 M KOH electrolyte is added and the three electrodes are

connected, using Pt-wire counter electrode and a Hg/HgO reference electrode. Note that all

potentials are however reported in V vs. RHE in this work. Air is continuously bubbled in the

electrolyte during the measurements. The operando XAS spectra are collected at open circuit

potential (OCP), 0.2, 0.4, 0.6, 0.8 and 1.0 V vs. RHE. Ex situ spectra were collected in

transmission geometry on pellets of 1 mm diameter using Teflon powder as a binder.

Mössbauer Spectroscopy

The pristine catalysts are characterized with 57Fe Mössbauer spectroscopy at room

temperature. To this end, 57Fe-enriched catalysts are used, prepared identically as the ones

169

otherwise investigated in this study, except for the use of 57Fe acetate during their synthesis.

Mössbauer spectra are measured at room temperature with a 57Co:Rh source. The

measurements are carried out in triangular velocity waveform using NaI scintillation detector

for γ-rays. The velocity calibration is done with an α-Fe foil. A mass of 30 mg of 57Fe-enriched

Fe0.5-Ar and Fe0.5-NH3 powders are necessary for a proper signal-to-noise resolution.

57Fe Mössbauer spectra were measured at low temperature (5 K) on 57Fe0.5-NH3 cathodes

before and after H2/air AEMFC operation, with a 57Co:Rh source. The measurements were

carried out with a triangular velocity waveform, using NaI scintillation detector for detecting

the γ-rays. The velocity calibration was performed with an α-Fe foil. Two GDEs have been

studied with ex situ Mössbauer spectroscopy, one after break-in procedure and the other

after the break-in procedure and the stability test (including a cumulative 100 h of operation

in air at 600 mA cm-2 and all polarisation curves recorded during this test).

BET measurements

To measure the specific surface area of carbon in the catalysts, sorption isotherms of N2 are

measured in liquid nitrogen (77 K) with a Micromeritics, ASAP 2020 instrument. The sample

is previously cleaned at 200°C for 5 h in flowing nitrogen. The specific surface area is

determined by the multipoint Brunauer-Emmett-Teller (BET) method.

170

Chapter 2

Table S1 – Mössbauer spectral parameters obtained from the fitting analysis with two

doublets.

The isomer shift (IS), quadrupole splitting (QS) and line width (LW) are indicated as well as the

relative area % of each doublet.

Spectral component

IS (mm s-1)

QS (mm s-1)

LW (mm s-1)

Area %

Fe0.5-Ar

D1 0.36 0.97 0.66 62.0 D2 0.41 2.54 1.33 38.0

Fe0.5-NH3

D1 0.36 0.89 0.66 65.7 D2 0.40 2.51 1.35 34.3

0,0 0,2 0,4 0,6 0,8 1,0

-5

-4

-3

-2

-1

0

Fe0.5

- NH3

Fe0.5

- Ar

Cu

rre

nt D

en

sity (

mA

cm

-2)

Potential (V vs RHE)

Figure S1. Scheme showing the electrochemical potentials at which operando XANES and

EXAFS spectra were recorded in alkaline electrolyte, and where they fall in the ORR

polarisation curves of the Fe0.5-Ar and Fe0.5-NH3 catalysts as measured in RDE set-up in O2-

saturated 0.1 M KOH.

171

7110 7120 7130 7140 7150 7160 71700,0

0,2

0,4

0,6

0,8

1,0

1,2

0 1 2 3 4 50,0

0,2

0,4

0,6

0,8

1,0

Fe0.5

-Ar 0.2 V

Fe0.5

-NH3

0.2 V

Fe0.5

-NH3

1.0 V

Fe0.5

-Ar 1.0 V

bN

orm

aliz

ed

XA

NE

S s

ign

al

Energy (eV)

a

Fo

uri

er

tra

nsfo

rm / A

-3

R (Angstrom)

Figure S2. Comparison of a) XANES spectra and b) the FT-EXAFS spectra between Fe0.5-Ar and

Fe0.5-NH3 at 1.0 and 0.2 V vs. RHE in O2-saturated 0.1 M KOH. The spectra recorded at high

potential are in purple while those recorded at low potential are in red. Fe0.5-Ar is identified

by the solid curves and Fe0.5-NH3 by the filled circles.

172

Chapter 3

0,5 0,6 0,7 0,8 0,9 1,0

-0,4

-0,3

-0,2

-0,1

0,0

0,1

0,2

0,3

0,4

0,5

Cu

rre

nt

Den

sity (

mA

cm

-2)

Potential (V vs RHE)

α-MnO2

β-MnO2

δ-MnO2

α-Mn2O

3

Vulcan

Figure S1 CVs collected in N2-saturated electrolyte using the manganese oxides dispersed in

Vulcan. The scan rate used to do the experiment is 10 mV s-1, the rotation rate is 1600 rpm

and catalyst loading is 0.2 mg cm-2 for the four MnOx/Vulcan and 0.16 mg cm-2 for the Vulcan.

delta-MnO2 beta-MnO2 alfa-MnO2 Mn2O3

0

100

200

300

400

500

Capacitance

Oxide

Vulcan

Figure S2 Summarize of the capacitance of the four composite catalysts

173

Chapter 4

Figure S1. XRD Patterns from Figure 1 with reflexes assigned to literature values.

174

Figure S2. XPS survey scans of the four MnOx, dropcasted from a MnOx/C ink containing

Nafion, with compositional analysis of the elements present in the analytical volume, as well

as a reference spectra of the GC plate.

175

Figure S3. SEM images of the as synthesized metal oxides; bottom to top: α-Mn2O3, α-MnO2,

β-MnO2, δ-MnO2.

176

Figure S4. Inverse correlation between the crystallite size D calculated using the Scherrer

equation and the BET surface area obtained from N2 sorption isotherms of the unsupported

MnOx.

Figure S5. Six consecutive CVs on all four catalysts supported on Vulcan® in N2 purged 0.05 M

NaOH at 5 mV s−1 and 1600 rpm. (a) α‒Mn2O3 (b) α‒MnO2 (c) β‒MnO2 (d) δ‒MnO2.

177

Figure S6. Electrochemical characterization of all four catalysts supported on Vulcan® in

0.05 M NaOH. Bottom (b): LSV in O2, at 5 mV s−1 and 1600 rpm, ranging from 1 – 0 VRHE. For

clarity only the second cycle is shown. Top (a): corresponding Pt Ring currents at Ering = 1.2 V.

178

Acronym

AST Accelerated Stress Test

AEI Anion Exchange Ionomer

AEM Anion Exchange Membrane

AEMFC Anion Exchange Membrane Fuel Cell

BET Brunauer Emmett Teller

BoL Beginning of Life

BoT Beginning of Test

BP Back Pressure

CA Chronoamperometry

CNT Carbon Nanotube

CP Chronopotentiometry

CV Cyclo Voltammogram

DMF Dimethylformamide

DMBA N,N-dimethylbenzylamine

E Potential

EL Electrolyser

EoT End of Test

ETFE Ethlylene Tetrafluoroethylene

EXAFS Extended X-ray Absorption Fine Structure

FC Fuel Cell

FT Fourier Transform

FWHM Full Width Half Maximum

GC Glassy Carbon

GDE Gas Diffusion Electrode

GDL Gas Diffusion Layer

HDPE High Density Poly Ethylene

HFR High Frequency Resistance

HOR Hydrogen Oxidation Reaction

HPRR Hydrogen Peroxide Reduction Reaction

IPA Isopropanol

IS Isomer Shift

J lim Diffusion-limited currend density

LDPE Low Density Poly Ethylene

179

LSV Linear Sweep Voltammogram

LW Line Width

MEA Membrane Electrode Assembly

M-N-C Metal Nitrogen Carbon

MOF Metal-Organic Framework

Mpy 1-methylpyrrolidine

OCP Open Circuit Potential

OER Oxygen Evolution Reaction

ORR Oxygen Reduction Reaction

PAN Polyacrilonitrile

PBI Polybenzimidazole

PEM Proton Exchange Membrane

PEMFC Proton Exchange Membrane Fuel Cell

PGM Platinum Group Metals

PPO Poly-(Phenylene Oxides)

PTFE Poly-Tetrafluoroethylene

QA Quaternary Ammonium

QS Quadruple Splitting

RDE Rotating Disk Electrode

R&D Research and Development

RHE Reversible Hydrogen Electrode

ROS Reactive Oxygen Species

RRDE Rotating Ring Disk Electrode

SEM Scanning Electron Microscopy

SFC/ICP-MS Scanning Flow Cell Inductively Coupled Plasma Mass Spectroscopy

SHE Standard Hydrogen Electrode

SoA State of Art

TDA Total Dissolved Amount

TEA Triethylamine

TMA Trimethylamine

ToF Turnover Frequency

TOF-SIMS Time-of-Flight Secondary-Ion Mass Spectroscopy

TPB Triple Phase Boundary

UV-vis Ultra Violet-Visible Spectroscopy

XAS X-ray Absorption Spectroscopy

XANES X-ray Absorption Near Edge Structure

180

XPS X-ray Photoelectron Spectroscopy

XRD X-ray Diffraction

ZIF Zeolitic Imidazolate Framework

181

Résumé de la these

Français

Cette thèse de doctorat étudie la synthèse, caractérisation structurale et activité pour la réaction de réduction de O2 (ORR) de catalyseurs Fe-

N-C et de composites d’oxydes de manganèse supporté sur Fe-N-C, ainsi que leur utilisation en pile à combustible à membrane échangeuse

d’anions (AEMFC). Tandis que les piles à membrane échangeuse de protons (PEMFC) requièrent aujourd’hui du platine dans ses catalyseurs

pour atteindre des hautes performances, les piles AEMFC peuvent ouvrir la voie vers des piles sans métaux précieux. Si les catalyseurs Fe-N-

C sont actuellement étudiés comme alternative au platine à la cathode des PEMFC, ils souffrent d’une faible activité et d’une durabilité limitée

dans ce milieu. En revanche, on peut espérer que l’activité et la durabilité des catalyseurs Fe-N-C soient améliorées dans les AEMFC.

Ce travail démontre la haute activité, stabilité et durabilité en milieu alcalin de catalyseurs Fe-N-C comprenant des sites FeNx à un atome de

fer. Ils ont été préparés à partir de ZIF-8 et de sel de fer, pyrolysé sous Ar (Fe0.5-Ar) puis sous NH3 (Fe0.5-NH3). Leur activité a été mesurée en

électrode à disque tournant (RDE) et en AEMFC, tandis que la stabilité a été mesurée en RDE et, operando, avec un spectromètre de masse

(ICP-MS) en aval d’une cellule à flux (SFC), en électrolyte acide et alcalin. Le dispositif ICP-MS/SFC a été utilisé pour mesurer in operando la

dissolution du fer. En électrolyte acide oxygéné, la vitesse de dissolution du fer du catalyseur le plus actif (Fe0.5-NH3) est 10 fois plus rapide

que celle du catalyseur moins actif, Fe0.5-Ar. Ceci explique la faible stabilité des catalyseurs Fe-N-C pyrolysés sous NH3 en PEMFC. En revanche,

en électrolyte alcalin, les vitesses de dissolution du fer sont faibles, même pour Fe0.5-NH3. Ces résultats vont de pair avec l’absence de

changement d’activité en RDE après un test de dégradation accélérée. La nature des sites actifs a de plus été étudiée par spectroscopie

d’absorption de rayons X en mode operando.

Afin de réduire la quantité de peroxyde d’hydrogène sur Fe-N-C pendant l’ORR, plusieurs oxydes de manganèse ont été synthétisés et leur

activité pour l’ORR et la réaction de réduction du peroxyde d’hydrogène (HPRR) évaluée. Il a été démontré par ICP-MS/SFC que même l’oxyde

de manganèse le plus stable, Mn2O3, peut dissoudre une quantité importante de Mn pendant l’ORR en milieu alcalin. De plus, cette dissolution

est due au peroxyde d’hydrogène produit pendant l’ORR. Des composites MnOx/Fe0.5-NH3 ont été étudiés pour les réactions ORR et HPRR.

Tous ont montré une meilleure sélectivité pendant l’ORR que Fe0.5-NH3 seul, et l’effet le plus important fut avec Mn2O3.

Avant d’étudier ces catalyseurs en AEMFC, une étude a été faite sur la compatibilité entre différents catalyseurs de l’ORR et/ou de l’oxydation

de H2 (Pt/C, Fe0.5-NH3, PtRu/C, Pd-CeO2/C) et des ionomères échangeurs d’anion, en RDE dans 0.1 M KOH. Ceci a permis d’identifier certains

problèmes entre les ionomères étudiés et les catalyseurs comprenant une faible quantité de métal (Fe0.5-NH3, Pd-CeO2/C).

Les catalyseurs Fe0.5-NH3 et Mn2O3/Fe0.5-NH3 ont alors été étudiés en AEMFC avec un ionomère à base d’éthylène-tetrafluoroéthylène. Les

deux catalyseurs atteignent une densité de courant de 80 mA cm-2 à 0.9 V, avec un chargement de 1.0-1.5 mg cm-2. Le pic de puissance sous

H2/O2 est de 1 W cm-2 à 60°C, avec une AEM à base de polyéthylène basse densité, et de 1.4 W cm-2 à 65°C avec une AEM en polyéthylène

haute densité. En comparaison, une densité de courant de 70 mA cm-2 à 0.9 V et un pic de puissance de 1.5 W cm-2 ont été obtenus avec 0.45

mgPt cm-2 à la cathode (40 wt% Pt/C) à 60°C, avec l’AEM en polyéthylène basse densité. Un test de durabilité de 100 h à 0.6 A cm-2 sous air a

montré une bonne stabilité de Fe0.5-NH3.

En conclusion, ce travail met en exergue l’application prometteuse des catalyseurs Fe-N-C à la cathode de piles AEMFC, afin de s’affranchir

des catalyseurs à base de métaux précieux.

English

This PhD work demonstrates the high activity, stability and durability in alkaline medium of Fe-N-C catalysts with atomically-dispersed FeNx

sites. They were prepared from a mix of ZIF-8 and iron salt, pyrolyzed in argon (Fe0.5-Ar) and then ammonia (Fe0.5-NH3). The activity was

measured in a rotating disk electrode (RDE) and in AEMFC, while the stability was measured in RDE and in operando with mass spectroscopy

(ICP-MS) coupled with a scanning flow cell, in both acid and alkaline media. The latter setup was used to measure Fe dissolution in operando.

It was evidenced that, in oxygenated acid electrolyte, the iron leaching rate of the most active Fe-N-C catalyst (Fe0.5-NH3) is 10 times faster

compared to the less active Fe0.5-Ar. This explains the reduced stability of ammonia-treated Fe-N-C catalysts in operating PEMFC. In contrast,

in alkaline medium, very little demetallation was observed even for Fe0.5-NH3. This was correlated with almost unchanged activity after load

cycling in RDE. The nature of the active sites was investigated with X-ray absorption spectroscopy, including in operando measurements.

Then, to minimize the amount of peroxide species during ORR on Fe-N-C, different manganese oxides were synthesized and their activity for

ORR and hydrogen peroxide reduction reaction (HPRR) were evaluated, while operando manganese dissolution was investigated with ICP-MS.

It was found that even the most stable Mn-oxide, Mn2O3, leached a significant amount of Mn during ORR in alkaline medium. It was further

demonstrated that the Mn leaching is associated with hydrogen peroxide produced during ORR. Composites of Fe0.5-NH3 and Mn-oxides were

then investigated for ORR and HPRR. Improved selectivity during ORR was observed for all composites relative to Fe0.5-NH3 alone, but the

effect was strongest for Mn2O3.

Before investigating such catalysts in AEMFC, a study on the compatibility between different ORR and/or hydrogen oxidation reaction catalysts

(Pt/C, Fe0.5-NH3, PtRu/C, Pd-CeO2/C) and anion exchange ionomers was performed in RDE in 0.1 M KOH. The study identified issues between

the investigated ionomers and catalysts having low metal contents on the carbon support (Fe0.5-NH3, Pd-CeO2/C).

The catalyst Fe0.5-NH3 and its composite with Mn2O3 were then investigated in AEMFC with an ethylene-tetrafluoroethylene ionomer. Both

cathode catalysts reached a current density of ca 80 mA cm-2 at 0.9 V, with relatively low loading of 1.0-1.5 mg catalyst·cm-2. The peak power

density with H2/O2 reached 1 W cm-2 at 60°C with a low density polyethylene AEM and 1.4 W cm-2 with high density polyethylene AEM at 65°C.

By comparison, a current density of ca 70 mA cm-2 at 0.9 V and peak power density of 1.5 W cm-2 was reached with 0.45 mgPt cm-2 at the

cathode (40 wt% Pt/C) with low density polyethylene AEM at 60°C. A durability test of 100 h at 0.6 A cm-2 in air showed good stability of the

Fe0.5-NH3 catalyst.

In conclusion, this work highlights the promising application of Fe-N-C catalysts at the cathode of AEMFCs for replacing precious metal

catalysts.