Energy and Environment Research in China

Junliang Zhang
Shuiyun Shen

Low
Platinum
Fuel Cell
Technologies
Energy and Environment Research in China
More information about this series at http://www.springer.com/series/11888
Junliang Zhang Shuiyun Shen
•

Low Platinum Fuel Cell

Technologies

123
Junliang Zhang Shuiyun Shen
School of Mechanical Engineering, School of Mechanical Engineering,
Institute of Fuel Cells Institute of Fuel Cells
Shanghai Jiao Tong University Shanghai Jiao Tong University
Shanghai, China Shanghai, China

ISSN 2197-0238 ISSN 2197-0246 (electronic)

Energy and Environment Research in China
ISBN 978-3-662-56068-6 ISBN 978-3-662-56070-9 (eBook)
https://doi.org/10.1007/978-3-662-56070-9
Jointly published with Shanghai Jiao Tong University Press
The print edition is not for sale in China (Mainland). Customers from China (Mainland) please order the
print book from: Shanghai Jiao Tong University Press.

© Shanghai Jiao Tong University Press and Springer-Verlag GmbH Germany 2021
This work is subject to copyright. All rights are reserved by the Publishers, whether the whole or part
of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations,
recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission
or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar
methodology now known or hereafter developed.
The use of general descriptive names, registered names, trademarks, service marks, etc. in this
publication does not imply, even in the absence of a specific statement, that such names are exempt from
the relevant protective laws and regulations and therefore free for general use.
The publishers, the authors, and the editors are safe to assume that the advice and information in this
book are believed to be true and accurate at the date of publication. Neither the publishers nor the
authors or the editors give a warranty, express or implied, with respect to the material contained herein or
for any errors or omissions that may have been made. The publishers remain neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

This Springer imprint is published by the registered company Springer-Verlag GmbH, DE part of
Springer Nature.
The registered company address is: Heidelberger Platz 3, 14197 Berlin, Germany
Preface

Membrane electrode assembly (MEA) is the core component of proton exchange

membrane fuel cell, which is composed of proton exchange membrane, cathode and
anode catalytic layers and gas diffusion layers. The cost of MEA accounts for more
than 60% of that of the total system, and particularly, the cost of platinum
(Pt) catalysts accounts for nearly 70% of MEA cost. It has been well recognized that
the high cost caused by high Pt loading in MEA is one of the key issues that hinder
the commercialization of fuel cells, and thus the most direct way to reduce the cost
of fuel cells is to reduce the amount of Pt in the MEA. However, a continuous
decrease in the Pt loading in MEA will cause more serious activation overpotential
and oxygen transport problems. In order to solve the above difficulties, on one
hand, low Pt electrocatalysts, such as Pt alloy catalysts, Pt core–shell catalysts and
shape-controlled Pt-based nanocrystals, have been proposed to improve the cat-
alytic activities in the MEA, and on the other hand, accurately measuring the mass
transfer resistance in catalytic layers is conducive to the development of low Pt
membrane electrode technology. In Chap. 1, the introduction of PEMFCs and
components of MEA including proton exchange membrane (PEM), catalytic layers
and gas diffusion layers (GDLs) are then discussed. In addition, low Pt electro-
catalysts are our focus.
The high cost and poor durability are the pressing issues that need to be
addressed before pushing the polymer electrolyte membrane fuel cells (PEMFCs) to
the market. The primary effort is made on the electrocatalysis by lowering the
loading of Pt-group which are scarce and expensive but remain as the most efficient
material until now. State-of-the-art Pt and Pt-based electrocatalysts usually offer an
exceptional electrocatalytic activity as well as durability toward the oxygen
reduction reaction (ORR), but the trick lies in their well-controlled surface structure
and composition. Theoretical discussion based on the first principle about the
cathodic ORR in Chap. 2 will provide new insights into material choice, surface
element distribution and the closed crystallographic orientation design of electro-
catalysts. Therein, alloying Pt with 3d transition metals, Pt shells and {111}-facets
are regarded as excellent characteristics for improving the ORR performance. The
experimental strategies including chemical routes, electrodeposition and thermal

v
vi Preface

annealing for achieving those well-designed structures are then concluded. Besides,
we also shed light on the hinder for practical application of those electrocatalysts in
the PEMFCs.
Fuel cells are expected to be one of the major clean and sustainable energy
sources in the near future. However, the sluggish kinetics of ORR and the high Pt
loading in the cathode are the urgent issues to be addressed, since they determine
the efficiency and the cost of fuel cells. In Chap. 3, an approach is developed for
designing the electrocatalyst for ORR in fuel cells. These electrocatalysts consist of
only Pt monolayer or mixed transition-metal-Pt monolayer shells (MS) on suitable
carbon-supported metal or alloy nanoparticles. The synthesis involves depositing a
Cu shell on a suitable transition metal or metal alloy surface at underpotentials,
followed by the galvanic replacement of the Cu MS with Pt or mixed metal-Pt. It is
found that the electronic properties of Pt MS can be fine-tuned by electronic and
geometric effects introduced by the substrate metal (or alloy) and lateral effects
of the neighboring metal atoms. The role of substrates is found reflected in a
“volcano” plot of the MS activity for ORR as a function of their calculated d-band
centers. The Pt mass-specific activity of the Pt-MS electrocatalysts is up to twenty
times higher than state-of-the-art commercial Pt/C. The enhancement of the activity
is caused mainly by the decreased formation of Pt–OH (the blocking species for
ORR) and to a lesser degree by the electronic effects. Fuel cell tests show a very
good long-term stability of the electrocatalysts. Results demonstrate a viable way in
designing the electrocatalysts which can successfully alleviate two issues regarding
commercialization of fuel cells: the costs of electrocatalysts and their efficiency.
With a continuous decrease in Pt loading, mass transport resistances within
MEA are severely aggravated, especially that the oxygen transport resistance is
becoming an obstacle for commercialization of fuel cells. In Chap. 4, the proton
transport in the ultrathin perfluorosulfonic acid ionomer film covered on catalyst
surface is investigated first. Next, the oxygen transport behavior including gas
diffusion in porous electrode and gas permeation in ultrathin perfluorosulfonic acid
ionomer film is probed; thus, the underlying transport mechanism is discussed
detailedly. In addition, the application of numerous novel electrocatalysts leads to
high ORR performances but also cation contamination, which is also assessed in the
last part of this section. Herein, we shed light on transport and reaction mechanism
inside the MEA, as well as provide key insights for optimization design for
appropriate electrode structures.
Degradation of Pt-based catalysts is a major cause for the performance loss of
PEMFCs during normal operational conditions. To improve the durability of
PEMFCs, it is critical to fully understand the degradation process and mechanisms
of Pt-based catalysts. In Chap. 5, several degradation mechanisms for Pt catalysts
are investigated in detail, including particle growth due to Ostwald ripening and
particle agglomeration, Pt mass loss due to migration and particle detachment, and
impurities contamination. While for Pt-based alloy catalysts, the leaching of tran-
sitional metal out of catalysts is discussed emphatically, which is another vital
degradation mechanism resulting in the loss of catalytic activity here. Furthermore,
in-situ experimental methods for observation of these processes and the recent
Preface vii

mathematical methods for modeling the degradation behaviors of Pt-based catalysts

are concluded for further understanding the degradation processes. In addition,
several significant strategies are proposed for mitigating the above mentioned
degradation processes of Pt-based catalysts, such as Pt catalyst alloyed or modified
by proper metals, gradient design of cathode catalyst layer, and optimizing the
morphology, size and structure of Pt-based alloy catalysts.

Shanghai, China Junliang Zhang

Shuiyun Shen
Contents

1 Proton Exchange Membrane Fuel Cells (PEMFCs) . . . . . . . . . . . . . 1

1.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 PEMFC . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.2.1 PEMFC Components . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Membrane Electrode Assembly (MEA) . . . . . . . . . . . . . . . . . . . . 3
1.3.1 MEA Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.3.2 Proton Exchange Membrane (PEM) . . . . . . . . . . . . . . . . . 5
1.3.3 Catalyst Layer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
1.3.4 Gas Diffusion Layer (GDL) . . . . . . . . . . . . . . . . . . . . . . . 10
1.4 Electrocatalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
1.4.1 Pt/C Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
1.4.2 Pt Alloy/C Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14
1.4.3 Core–Shell Catalysts . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
1.4.4 Shape-Controlled Pt-Based Nanocrystals . . . . . . . . . . . . . . 19
1.4.5 Nonprecious Metal Catalysts (NPMC) . . . . . . . . . . . . . . . 20
1.5 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21
2 The Electrocatalysis of Oxygen Reduction on Platinum
and Its Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
2.1.1 Cathode Reaction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.1.2 Platinum and Platinum-Based Catalysts . . . . . . . . . . . . . . . 26
2.2 Reaction Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
2.2.1 The Dissociation Mechanism and the Association
Mechanism . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .... 27
2.2.2 Theoretical Analysis on Pt Activity Toward the ORR .... 28
2.3 Electrochemical Measurements . . . . . . . . . . . . . . . . . . . . . . .... 29

ix
x Contents

2.4 Further Improvement Toward Pt Activity . . . . . . . . . . . . . . . . . .. 31

2.4.1 Pt Nanoparticles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 31
2.4.2 An Organic Solvent System-Assisted Electrodeposition
of Highly Active Pt . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 33
2.5 Pt-Based Alloys . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 35
2.5.1 Theoretical Analysis on Pt-Based Alloys Activity Toward
the ORR . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 35
2.5.2 Composition Effects . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
2.5.3 The Co-Reduction of Pt with Co Ions . . . . . . . . . . . . . . . . 37
2.5.4 Crystallographic Orientation Effects . . . . . . . . . . . . . . . . . 42
2.5.5 Icosahedral Pt–Ni Nanocrystalline Electrocatalyst . . . . . . . 43
2.5.6 Stability . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48
2.5.7 Thermal Annealing Synthesis of Double-Shell Truncated
Octahedral Pt–Ni Alloys . . . . . . . . . . . . . . . . . . . . . . . .. 50
2.6 Outlook . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 57
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. 58
3 Pt-MS Electrocatalysts for ORR . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
3.2 Cu UPD Coupled with Pt2+ Galvanic Replacement . . . . . . . . . . . . 65
3.3 Pt MSs on Single-Crystal Metal Electrodes for ORR . . . . . . . . . . 66
3.3.1 Effects of Substrates on Pt MSs for ORR . . . . . . . . . . . . . 66
3.3.2 Volcano-Type Relation Between Electrocatalytic
Activity and d-Band Center for ORR on Pt MS . . . . . . . . 69
3.3.3 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
3.4 Pt MSs on Pd(111) and Carbon-Supported Pd NPs for ORR . . . . . 71
3.4.1 Pt MSs Deposited on Pd(111) and Pd NPs . . . . . . . . . . . . 72
3.4.2 Pt MS on a Pd(111) Surface for ORR . . . . . . . . . . . . . . . . 76
3.4.3 Pt MSs on Carbon-Supported Pd NPs for ORR . . . . . . . . . 79
3.4.4 Fuel Cell Tests for Pt-MS Electrocatalysts . . . . . . . . . . . . . 82
3.4.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 83
3.4.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 87
3.5 Mixed Pt–M–MS Electrocatalysts for ORR . . . . . . . . . . . . . . . . . 88
3.5.1 ORR on M0.2ML/Pt(111) . . . . . . . . . . . . . . . . . . . . . . . . . . 89
3.5.2 ORR on (M0.2Pt0.8)ML/Pd(111) . . . . . . . . . . . . . . . . . . . . . 90
3.5.3 ORR on (M0.2Pt0.8)ML/Pd NPs/C . . . . . . . . . . . . . . . . . . . 94
3.5.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 99
3.6 Pt MSs on Carbon-Supported Au–Ni NPs for ORR . . . . . . . . . . . 100
3.6.1 Electrochemical and XRD Characterization
of Au/Ni NPs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 100
3.6.2 ORR Electrocatalytic Properties . . . . . . . . . . . . . . . . . . . . 104
3.6.3 Fuel Cell Tests . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 106
3.6.4 XANES Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . 107
Contents xi

3.6.5 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108

3.6.6 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
3.7 Pt MSs on Carbon-Supported PdxNi1−x NSs for ORR . . . . . . . . . 108
3.7.1 Physicochemical Characterizations of PdxNi1−x
and PdxNi1−x@Pt NSs . . . . . . . . . . . . . . . . . . . . . . . . . . . 109
3.7.2 Greatly Enhanced Pd Utilization for the Cu UPD
on PdxNi1−x NSs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114
3.7.3 ORR Electrocatalytic Properties . . . . . . . . . . . . . . . . . . . . 115
3.7.4 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
4 Membrane Electrode Assembly (MEA) . . . . . . . . . . . . . . . . . . . . . . . 127
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 127
4.1.1 Reaction and Mass Transport Process in Fuel Cells . . . . . . 127
4.1.2 Structure in CLs and Its Characterization . . . . . . . . . . . . . 129
4.2 Proton Transport in Electrode . . . . . . . . . . . . . . . . . . . . . . . . . . . 132
4.3 Oxygen Transport in Electrode . . . . . . . . . . . . . . . . . . . . . . . . . . 137
4.3.1 Local and Bulk Oxygen Transport in CLs . . . . . . . . . . . . . 137
4.3.2 Experimental Measurement of Local and Bulk Oxygen
Transport Resistance . . . . . . . . . . . . . . . . . . . . . . . . . . . . 138
4.3.3 Influence of Ionomer Content on Local
and Bulk Oxygen Transport Resistance . . . . . . . . . . . . . . . 142
4.3.4 Mechanism of Bulk Oxygen Transport Behavior . . . . . . . . 148
4.3.5 Mechanism of Local Oxygen Transport Behavior . . . . . . . 153
4.4 Cation Contamination in Electrode . . . . . . . . . . . . . . . . . . . . . . . 156
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 163
5 Degradation of Pt-Based Catalysts in PEMFC . . . . . . . . . . . . . . . . . 167
5.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 167
5.2 Pt Degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 169
5.2.1 Pt Degradation Mechanisms . . . . . . . . . . . . . . . . . . . . . . . 169
5.2.2 Experimental Observation of Pt Degradation . . . . . . . . . . . 174
5.2.3 Mathematical Modeling of Pt Degradation . . . . . . . . . . . . 181
5.2.4 Mitigation Strategies for Pt Degradation
in the Cathode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 188
5.3 Pt-Based Alloy Degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . 199
5.3.1 Pt-Based Alloy Degradation Mechanisms . . . . . . . . . . . . . 199
5.3.2 Experimental Observation of Pt-Based
Alloy Degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 201
5.3.3 Mathematical Modeling of Pt–Co Core–Shell Catalyst
Degradation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 211
5.3.4 Mitigation Strategies for Pt-Based Alloy Degradation . . . . 217
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 218

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 223
About the Authors

Dr. Junliang Zhang is Zhiyuan Chair Professor,

Director of Institute of Fuel Cells, Executive Deputy
Dean of Zhiyuan College at Shanghai Jiao Tong
University (SJTU). He received his BS and MS from
SJTU in 1994 and 1997, respectively, and his Ph.D.
from State University of New York at Stony Brook in
2005. He then worked as Research Associate till 2007
at Brookhaven National Laboratory of US Department
of Energy in New York. From 2007 to 2011, Dr. Zhang
worked at General Motors Global Research and
Development, Electrochemical Energy Research
Laboratory in New York, as a Research Scientist and
later a Senior Scientist and Team Leader, responsible
for the development of automotive fuel cells. In 2011,
he joined the Institute of Fuel Cells at SJTU.
Dr. Junliang Zhang’s key expertise is in interfacial
electrochemistry, electrocatalysis, nanomaterials, fuel
cells, batteries as well as heat and mass transfer in
electrochemical energy system. He has published over
90 peer-reviewed journal articles on electrochemical
energy conversion and storage, which have been cited
for more than 8000 times so far. He authored one book,
one chapter, and co-invented 45 patents. His research
group includes 8 faculties as well as 50 postdocs, Ph.D.
students and Master students.

xiii
xiv About the Authors

Dr. Shuiyun Shen is Associate Professor of Institute

of Fuel Cells, School of Mechanical Engineering,
Shanghai Jiao Tong University (SJTU). She received
her Bachelor and Master degree both from Harbin
Institute of Technology and Ph.D. degree from the
Hong Kong University of Science and Technology. She
then worked as a postdoctor in the Hong Kong
University of Science and Technology. Her research
interests include electrocatalysis, high-performance
electrocatalysts for oxygen reduction reaction, high-
performance electrocatalysts for ethanol oxidation
reaction, ultralow-platinum PEMFC cathode and so
on. She has published over 40 peer-reviewed journal
articles on electrochemical energy conversion and
storage, which have been cited for more than 1000
times so far.
Chapter 1
Proton Exchange Membrane Fuel Cells
(PEMFCs)

Abstract Membrane electrode assembly (MEA) is the core component of proton

exchange membrane fuel cell, which is composed of proton exchange membrane,
cathode and anode catalytic layers and gas diffusion layers. The cost of MEA accounts
for more than 60% of that of the total system, and particularly, the cost of platinum
(Pt) catalysts accounts for nearly 70% of MEA cost. It has been well recognized
that the high cost caused by high Pt loading in MEA is one of the key issues that
hinder the commercialization of fuel cells, and thus the most direct way to reduce the
cost of fuel cells is to reduce the amount of Pt in the MEA. However, a continuous
decrease in the Pt loading in MEA will cause more serious activation overpoten-
tial and oxygen transport problems. In order to solve the above difficulties, on one
hand, low Pt electrocatalysts, such as Pt alloy catalysts, Pt core–shell catalysts and
shape-controlled Pt-based nanocrystals, have been proposed to improve the catalytic
activities in the MEA, and on the other hand, accurately measuring the mass transfer
resistance in catalytic layers is conducive to the development of low Pt membrane
electrode technology. In this section, the introduction of PEMFCs and the compo-
nents of MEA including proton exchange membrane (PEM), catalyst layers and gas
diffusion layers (GDLs) are then discussed. In addition, low Pt electrocatalysts are
our focus.

Keywords Membrane electrode assembly (MEA) · Proton exchange membrane

fuel cells (PEMFCs) · Low Pt electrocatalysts

1.1 Introduction

Similar to batteries, fuel cells convert chemical energy of fuel and oxidant into electric
energy. Yet unlike batteries, they do not need recharging as long as fuel and oxidant
are continuously supplied. When hydrogen is fed as fuel, the fuel cell only generates
electricity, water and some heat [1]. As compared to thermal engines, the advantages
of fuel cells are high efficiency, no environmental pollution and unlimited sources of
reactants. Among all of the existing fuel cells, the proton exchange membrane fuel
cells (PEMFCs) have been actively developed for use in vehicles, portable electronics

© Shanghai Jiao Tong University Press and Springer-Verlag GmbH Germany 2021 1
J. Zhang and S. Shen, Low Platinum Fuel Cell Technologies,
Energy and Environment Research in China,
https://doi.org/10.1007/978-3-662-56070-9_1
2 1 Proton Exchange Membrane Fuel Cells (PEMFCs)

and combined heat and power (CHP) systems due to their simplicity, low operating
temperature, high power density, quick start-up and quick match to the shift of power
demand. PEMFCs are especially well suited as the candidates for automobiles and
buses. Therefore, fuel cells are expected to come into widespread commercial use in
the areas of transportation stationary and portable power generation, and thus will
help solve the global problems of energy supply and clean environment [2].

1.2 PEMFC

1.2.1 PEMFC Components

A typical PEMFC unit within a cell stack showing individual components is shown
in Fig. 1.1 and contains the following components [3]:
1. The membrane electrode assembly (MEA);
2. The cathode and anode gas diffusion layers (GDLs);
3. The bipolar plates.
Each individual cell can produce a voltage of about 0.6–0.7 V; to produce a suitable
voltage, individual cells are “stacked” to form a fuel cell stack. The individual cells are
electrically connected in series by bipolar plates, and special end plates terminate
the stacks to provide the compressive forces needed for stack structural integrity.
The bipolar plates provide conducting paths for electrons between cells, distribute

Fig. 1.1 Schematic cross section of a typical PEM fuel cell [3]
1.2 PEMFC 3

the reactant gases across the entire active MEA surface area (through flow channels
integrated into the plates), remove waste heat (through cooling channels) and provide
stack structural integrity as well as barriers to anode and cathode gases.
Directly adjacent to the bipolar plates are the GDLs, which typically consist of two
layers—a macroporous substrate layer and a microporous layer (MPL). The GDLs
are gas permeable and help distribute gases to the catalyst layer, conduct electrical
current and also provide a network of paths for liquid water to move from the MEA to
the flow channel. The macroporous substrate layer consists of a carbon-fiber matrix
with a large void volume, typically 75–85%, and a primarily hydrophobic MPL
consisting of carbon black mixed with fluoropolymer. The cathode GDL normally
has an attached MPL; the anode GDL may or may not have a MPL [4].
Bipolar plates perform a number of functions within the PEMFC. They have
been used to distribute the fuel and oxidant within the cell, separate the individual
cells in the stack, carry current away from each cell, carry water away from each cell,
humidify gases and keep the cells cool. Plate topologies and materials facilitate these
functions. Topologies can include straight, serpentine or interdigitated flow fields,
internal Mani folding, internal humidification and integrated cooling. Materials have
been proposed on the basis of chemical compatibility, resistance to corrosion, cost,
density, electronic conductivity, gas diffusivity/impermeability, manufacturability,
stack volume/kW, material strength and thermal conductivity. Given the criteria
found in literature, non-porous graphite, a variety of coated metals and a number
of composite materials have been suggested for use in bipolar plates.

1.3 Membrane Electrode Assembly (MEA)

1.3.1 MEA Structure

In the early years of PEMFC developments the mid-to-late 1960s researchers defined
an MEA to be two gas diffusion electrodes (GDEs) plus a proton-conducting polymer
membrane, or ionomer [5]. This form of MEA included GDL substrates with an elec-
trocatalyst layer deposited on each GDL surface in place of what are now the MPLs.
Modern PEMFC electrocatalyst layers are usually composite structures consisting of
proton-conducting ionomer material and noble-metal (platinum) catalyst supported
on carbon. This thin-film electrode technology was pioneered initially at Los Alamos
National Laboratory in the late 1980s and early 1990s [6, 7]. Since then, the meaning
of “MEA” has changed from one based exclusively on the original PEM technology
used by NASA and General Electric Corporation (this technology contained GDLs)
to meanings which can be based on either the discrete three-layer structure (two
composite electrocatalyst layers plus electrolyte without GDLs) or the five-layer
structure including GDLs. Other developers still use CCM and MEA to distinguish
between three-layer structures and five-layer structures [3].
4 1 Proton Exchange Membrane Fuel Cells (PEMFCs)

The membrane electrode assembly (MEA) is the core component of a fuel cell that
helps produce the electrochemical reaction needed to separate electrons. On the anode
side of the MEA, a fuel (hydrogen) diffuses through the membrane and is met on the
cathode end by an oxidant (oxygen or air) which bonds with the fuel and receives the
electrons that were separated from the fuel. Catalysts on each side enable reactions
and the membrane separates the reduction and oxidation half reactions while keeping
the gases separate. The membrane allows the protons to pass through to complete the
overall reaction white forcing the electrons to pass through an external circuit. In this
way, cell potential is maintained and is drawn from the cell producing electricity.
The GDL further improves the efficiency of the fuel and oxidant to the catalyst
layer [8].
A typical MEA is composed of a polymer electrolyte membrane (PEM), two
catalyst layers and two gas diffusion layers (GDL). A MEA with this configuration
is known as a 5-Layer MEA due to its composition. An alternative version of a
membrane electrode assembly is the 3-Layer MEA which is composed of a polymer
electrolyte membrane with catalyst layers applied to both sides, anode and cathode.
An alternative name for this type of MEA is a catalyst coated membrane (CCM) [9].
A few key differences in the two types of MEAs, 5-Layer or 3-Layer, are the
presence of a pre-selected GDL, and the location of the applied catalyst layers. In the
5-Layer MEAs, the GDL is a standard carbon cloth with a microporous layer (MPL)
applied to one side of the cloth. The cloth is Teflon treated is assist in pushing the
product water away from the catalyst sites, and the MPL is made of carbon powder
that is made into a slurry with a Teflon solution and applied to the cloth to give
a smooth surface to apply the catalyst and the layer is also electrically conductive
while maintaining an open pore structure to allow the gases into the reaction sites.
This setup is easier to apply the catalyst to and allows for quicker MEA fabrication
(Fig. 1.2).
To achieve better fuel cell performance, the MEA needs to be carefully designed
and fabricated. In general, the MEA consists of five key components: the proton
exchange membrane (PEM), the anode and cathode catalyst layers, the anode and

Fig. 1.2 Schematic of MEA

1.3 Membrane Electrode Assembly (MEA) 5

cathode gas diffusion layers (GDLs) and the microporous layer (MPL), typically
contained in the GDLs. All the above-mentioned key factors play an important role
in achieving a high-performance MEA. Details of each component are as follows:

1.3.2 Proton Exchange Membrane (PEM)

Organic-based cation exchange membrane polymers were originally invented to

apply in fuel cells by William T. Grubb in 1959 [10]. The intended function of the
membrane was to provide a proton-conducting gas barrier. The standard membrane
material belongs to the fully fluorinated Teflon-based family first showed up in the
patent literature in 1966 by the DuPont De Nemours company [11]. The inventors
were apparently unaware that this polymer would have such a rich and complex
life as a membrane in electrochemical cells. But the initial effort eventually led to
the development of proton exchange membrane (PEMFC) used in today’s fuel cell
systems. The membrane functions as an ionic conductor; at the same time, it provides
a separation between the fuel and the oxidant. Electrons cannot pass through the
membrane. The ionic groups in the polymer provide active sites for ionic conduc-
tion. The conductivity of the polymer is highly dependent on the bound and free
water associated with those sites.
The PEM separates the anode and cathode in a fuel cell and acts both as a proton-
conducting medium and a barrier to avoid direct contact between the fuel and oxidant.
Many important PEMFC features depend upon the properties of the PEM. So, a PEM
needs to have specific properties [12, 13] that enable it to work well in FCs:
1. high proton conductivity and low electronic conductivity,
2. adequate mechanical strength and stability,
3. high chemical and electrochemical stability under operating conditions,
4. low fuel and oxidant permeability to maximize columbic efficiency,
5. adequate water transport (diffusion and electro-osmotic) properties because
the proton-transporting properties of most PEMs depend upon the membrane
hydration,
6. significant dimensional and morphological stability,
7. a sufficiently long lifetime and low cost.
The PEM contains many proton-conductive functional groups, which allows
protons to transfer from one group to another. Figure 1.3 shows a schematic polymer
PEM structure, which displays how the protons are transferred [14].
The degree of hydration is the main factor which governs the proton conductivity
in a polymeric membrane. The proton transport mechanism can either be described
by a hopping or by a diffusion phenomenon [15, 16]. However, proton transport
mechanism through a PEM is basically conduction through water. The dominant
intermolecular interaction in water is hydrogen bonding. The introduction of an
extra proton leads to proton defects, resulting in a contraction of hydrogen bond in
the vicinity of such defects. The binding power of a water molecule depends on the
6 1 Proton Exchange Membrane Fuel Cells (PEMFCs)

Fig. 1.3 Schematic of diagram of proton transfers in the polymer PEM [14]

number of hydrogen bonds involve in it. This also leads to relaxation effects in the
neighboring hydrogen bonds as a response to the formation and cleavage of hydrogen
bonds. When a hydrogen bond is formed, the surrounding bonds are weakened but
the cleavage of hydrogen bonds strengthens the neighboring bonds [17]. Therefore,
defects caused by the incorporation of excess protons weaken the intermolecular
interaction by means of breakage and reformation of bonds in combination with
large variations in bond length.
An understanding of the chemistry of the migration of protonic defects through
the membrane is very important. The catalytically oxidized protons, acting as defects,
migrate through the membrane from the anode to the cathode and carry water
molecules with them. The average number of water molecules carried per proton
is called the electro-osmotic drag coefficient. The carried water accumulates in
cathode/membrane interface due to this electro-osmotic drag.
Currently, PEM is typically a polymeric material with sulfonic-acid side chains
and a fluorocarbon or hydrocarbon backbone. Nafion was the first synthetic polymer
with ionic properties (ionomers) developed by modifying Teflon, a perfluorinated
sulfonic acid-based ionomer (PFSA), has been the standard membrane material as
a benchmark in the FC industry and is also the most studied fuel cell membrane
material [18–20].
The exemplified structure of Nafion is shown in Fig. 1.4, where we can see that
PFSA consists of three regions: (1) a polytetrafluoroethylene (PTFE)-like backbone,

Fig. 1.4 Structure of PFSA ionomer [13]

1.3 Membrane Electrode Assembly (MEA) 7

(2) side chains of –O–CH2 –CF–O–CF2 –CF2 – which connect the molecular backbone
and (3) ion clusters consisting of sulfonic acid ions. The sulfonic acid group is
shown in its anhydrous form, SO3 H. When exposed to water, the hydrolyzed form
(SO3 –H3 O+ ) appears, allowing for proton transport across the material [14].
Being similar to Teflon, Nafion exhibits excellent resistance to chemical attacks
and an extremely low release rate of degradation products into the surrounding
medium because the structure is based on PTFE backbone, In fact, durability of
60,000 h has been reported. It also has a relatively high operation temperature range
and may be used in many applications at temperatures up to 190 °C. Nafion has
a high proton conductivity and acts as a superacid catalyst because its sulfonic
acid groups act as an extremely strong proton donor, its protonic conductivities
can be as high as 0.2 S/cm at PEMFC operating temperatures. Commercial Nafion
membranes with thicknesses of 2, 5, 7 and 10 mil (1 mil = 25.4 µm; Nafion 112,
115, 117 and 1110, respectively) appear to be the most widely used grades of Nafion.
This material provides high proton conductivity and moderate water swelling. The
thinner membranes are generally applied to hydrogen/air applications to minimize
ohmic losses, whereas the thicker membranes are employed for DMFCs for methanol
crossover reduction [12].
Given these advantages, there are several disadvantages to the use of PFSA
membranes in PEMFCs: (a) Because the proton conductivity is determined by the
water-filled channels, Nafion is not adequate for temperatures materially lower than
0 °C or significantly over 100 °C. (b) Nafion is seemingly stable against peroxide-type
ion and radical degradation. (c) Contaminating ions may also drastically decrease
the membrane conductivity. (d) Nafion has poor mechanical and chemical stabilities
at elevated temperature. (e) The cost of Nafion is expensive because of their complex
manufacturing process.

1.3.3 Catalyst Layer

In the middle of the membrane and the GDL is the catalyst layer. Catalyst layers
are commonly formed from catalyst ink dispersions, which typically comprise a
platinum-based catalyst for both the anode and the cathode deposited onto a support
(e.g., Pt/C), ionomer and a dispersing solvent. The significance of the heterogeneous
microstructure of the dispersion on the structure and properties of the catalyst layer
formed cannot be understated. A critical component of this, though not adequately
researched to date, is the nature of the dispersion medium, which governs ink prop-
erties such as aggregation size of the catalyst/ionomer particles, viscosity, rate of
solidification and ultimately the physical and mass transport properties of the catalyst
layer. The choice of the dispersion medium depends on the method of ink deposi-
tion. For instance, screen-printing or roll-to-roll coating usually requires viscous inks
with high solids content (>5 wt%) and high boiling point additives; whereas spray
coating requires lower solids content (<2 wt%) and faster evaporating alcoholic- or
water-based solvents [21].
8 1 Proton Exchange Membrane Fuel Cells (PEMFCs)

In the catalyst layer, both the anode and the cathode electrodes consist of highly
dispersed Pt-based nanoparticles on carbon black to promote the reaction rates of
the hydrogen oxidation reaction (HOR) and the oxygen reduction reaction (ORR).
The anode reaction in a PEM fuel cell is usually referred to HOR at the surface of
the anode platinum electrocatalyst; here, hydrogen is oxidized to produce electrons
and protons that are transferred to the cathode through a PEM, respectively (H2 →
2H+ + 2e− ). The mechanism of these processes can be divided into 5 steps [22, 23],
namely (1) the transport of the H2 molecules to the Pt electrode via gas diffusion; (2)
the adsorption of hydrogen molecules from the electrolyte onto the electrode surface
through the electrolytical double layer; (3) the dissociative adsorption of hydrogen
on Pt surface via Tafel reaction (Pt–Pt + H2 ↔ Pt–H + Pt–H); (4) the ionization of H
atom giving one electron to the electrode and a hydronium ion to the solution, leaving
an empty Pt site through the Volmer reaction (or charge transfer reaction, Pt–H +
H2 O ↔ Pt + H3 O+ + e− ) and further step is the transport of the hydronium ions from
the phase boundary to the electrolyte solution phase. The cathode reaction in a PEM
fuel cell is ORR, a highly irreversible reaction even at temperatures above 100 °C,
which is typically catalyzed on nanoscale platinum electrocatalysts supported on
high surface-area carbon (Pt/C). At the cathode, oxygen is reduced by reaction with
protons and electrons to produce water (O2 + 4H+ + 4e− → 2H2 O, E = 1.23 V)
[24, 25].
As shown in Fig. 1.5, we can see that the cell polarization consists of activation
polarization, ohmic polarization and concentration polarization (caused by O2 mass
transport) [26]. The sharp decrease observed in voltage in the very low and very
high current density regimes is often due to inherent electrochemical kinetics and
restricted mass transport within the catalyst layer, respectively. Exchange current
densities for the electrochemical oxygen reduction reaction (ORR) are much smaller
(∼10−10 A cm−2 ) than for the hydrogen oxidation reaction (HOR) (10−3 A cm−2 ) and
consequently ORR is often the major cause of fuel cell power losses under load. The

Fig. 1.5 Typical fuel cell

polarization curve of the
voltage versus current
density [26]
1.3 Membrane Electrode Assembly (MEA) 9

activity of the catalyst, influenced by its particle size, surface morphology, electronic
structure and support structure, is recognized as having an overriding influence on
the electrochemical kinetics of a fuel cell [27–29]. The catalyst layer facilitates the
transport of reactant gases to catalyst sites and the egress of product water. Under
high current densities, the significant volume of water produced in the cathode must
be removed in order to prevent its flooding [30].
Thickness of the catalyst layers is typically 5–10 µm, which leads to nonuniform
reaction rate distributions caused by limitations in gas and proton transport (and
ultimately, low utilization of Pt). Significant advancements in the utilization of Pt in
the catalyst layer have been made over the past decades in the form of replacing Pt
black with Pt supported on high surface area carbon and with impregnation of the
catalyst layer with a solid proton-conducting polymer, namely PFSA ionomer. Pt
loadings have subsequently been reduced from several mg cm−2 to <0.4 mg cm−2 ,
with concomitant improvements in power density. Therefore, reducing its loading or
even completely replacing it with an abundant and cheap metal would be advanta-
geous. Low platinum loading cathode catalyst with higher activity is desirable, plat-
inum alloy with 3d transition metals were reported to have higher catalytic activity
than pure platinum particles, and carbon-supported Pt3 Co particles are commercially
available for cathode catalyst. But these alloys are usually rich in platinum. Among
those efforts, platinum monolayer catalysts developed by Adzic’s group over the
past few years showed good activity for both anode and cathode, while decreasing
the use of platinum to a very low level. Incorporating PFSA ionomer directly into
the catalyst ink prior to catalyst layer deposition extends the 3d reaction zone in the
catalyst layer and increases catalyst utilization [6, 31, 32].
The ionomer is typically combined with a supported catalyst. Contact with and
between carbon support particles provides electronic pathways for electrical current
flow to and from reaction sites. The ionomer is exposed to incoming gases, water,
catalyst and support and is in contact with the membrane. The ionomer facilitates
the transport of protons necessary for ORR and HOR reactions and plays a crucial
role in transporting protons within the catalyst layer by exchanging protons between
the catalyst layer and the membrane. It also facilitates access of reactant gases to
catalyst sites and the transport of water to and from reaction sites. Therefore, the
ionomer should be permeable to gases and water and induce pore-formation in the
catalyst layer during deposition from inks. The ionomer is designed to possess high
proton conductivity, negligible electronic conductivity, high gas/water permeability
and to serve as a physical binder for the catalyst/support particles. Moreover, it must
be physically compatible with the membrane employed and stable toward electro-
chemical redox reactions and chemical attack by free radicals. As stated above, low
voltage and power densities observed in fuel cell polarization curves in the very low
and very high current density regimes are often due to limitations in electrochemical
kinetics and poor mass transport within the catalyst layer. Likewise, ohmic losses
observed in the medium current density regime, typically attributed to protonic resis-
tance of the membrane, can be due to poor proton conductivity of the ionomer in
the catalyst layer. All these losses can be greatly exacerbated or diminished by the
10 1 Proton Exchange Membrane Fuel Cells (PEMFCs)

properties and distribution of the ionomer [33]. In addition, the preparation process
of the catalyst layer also has a great influence on the performance of the fuel cell.

1.3.4 Gas Diffusion Layer (GDL)

The gas diffusion layer (GDL), as a wet-proofed and carbon-based material, is

employed between the flow-field plates and the catalyst layers on both the anode
and cathode sides of the cell. In the MEA, the membrane and electrodes need to be
hydrated to maintain high proton conductivity and ensure adequate fuel cell perfor-
mance. However, excess water in the electrodes can result in electrode flooding,
which prevents electrochemical reactions from occurring and reduces performance;
thus, a careful balance must be maintained [34]. In this respect, the GDL plays multi-
faceted roles in cell operation by controlling mass, heat and electron transport. It also
provides robust mechanical support and protection for the delicate catalyst layer and
membrane during both assembly and operation. The GDL is specifically designed as
a porous material for the simultaneous accomplishment of these functions. The void
region mainly controls the transport of reactants and byproduct water that flow in
reverse directions, while the remaining solid structure provides pathways for electron
and heat transport.
In general, the main roles [35] of GDL in a PEMFC are as following:
1. a pathway for the reactants,
2. an effective capacity for removing the water from the CL to the channels,
3. a structural support to the MEA under the assembly pressure,
4. an electrical connection between the bipolar plates and the CL,
5. a thermal pathway to transfer the heat generated at the CL to bipolar plates.
Although there are various types of GDLs and new structural concepts are being
developed, the GDL is typically made of carbon fibers and consists of a macroporous
substrate and a microporous layer (MPL), as presented in Fig. 1.6 (Toray TGP-H-060)
[36]. The substrate, in contact with the gas flow channel, serves as a gas distributor
and a current collector. The MPL contains carbon powder and hydrophobic agent
and manages the two-phase water flow. GDL materials have a thickness in the range
of 150–350 µm and a porosity between 0.75 and 0.9 with a diameter in the order of
7 µm (wet/dry laid or woven) and have a wettability that is set by impregnation with
poly(tetrafluoroethylene) (PTFE).

1.3.4.1 Macroporous Substrate

Carbon-fiber paper or cloth has been typically employed as a substrate for the GDL
in PEM fuel cells. Conventionally, carbon fibers are graphitized at high temperature
(>2000 °C) to enhance electronic conductivity and mechanical strength and impreg-
nated with thermoset resin to manufacture carbon papers. Carbon cloths are produced
1.3 Membrane Electrode Assembly (MEA) 11

Fig. 1.6 Schematic diagram of the GDL structure [36]

by spinning and weaving of carbon yarns, followed by carbonization or graphitization

[37]. Modification of carbon-fiber cloth by phenolic resin before carbonization [38]
improved fuel cell polarization behavior without significant ohmic and mass transfer
losses. Liu et al. [39] prepared carbon-fiber paper with different yard weights (70–
320 g m2 ) for the cathode GDL. Their carbon-fiber paper with a small yard weight
which has smaller thickness and lower permeability led to a better performance,
although its electronic conductivity is relatively small. Expanded graphite manufac-
tured by perforation process from flexible graphite sheet was proposed by Yazici [40].
The performance obtained using expanded graphite at the cathode was comparable
to that of the dual-layer structured ELAT (E-TEK).

1.3.4.2 Microporous Layer

A MPL may be best described as a simple composite of a carbon powder and a

hydrophobic agent. As the name implies, the MPL is conventionally pronounced as
a stand-alone layer, but it is not a layer, which has a clear interface with the MPS.
Because the MPL structure is comprised of small-scale carbon powder/hydrophobic
agent agglomerates, some of these agglomerates can be even several orders of magni-
tude smaller than the open pores of the MPSs, and hence, they are likely to penetrate
into these comparatively larger pores, with possible modification to their architec-
tures. This penetration pattern is a strong function of several parameters, such as
specifications of the carbon powder, hydrophobic agent, MPL slurry and manufac-
turing process and can significantly affect the resulting MPL thickness. The manu-
facturing of a typical MPL involves three consecutive steps: MPL slurry prepara-
tion, deposition and sintering. In the first two steps, carbon or graphite particles are
mechanically mixed with hydrophobic agent dispersed in water and/or sometimes
organic additives, and the resulting slurry is deposited onto the MPS surface. The
12 1 Proton Exchange Membrane Fuel Cells (PEMFCs)

MPL deposition can be done through different methods, such as brushing, blading,
spraying and dipping/floating. In recent years, interestingly, the MPL has also been
separately prepared as a self-supported substrate and then assembled onto the MPS
surface [41]; however, this manufacturing technique is not a common practice in
the fuel cell field. On completion of the MPL deposition, the resulting now GDL is
subjected to heat treatment, typically around 240 °C for 30–40 min to remove any
remaining moisture, a crucial step for the accuracy of the carbon loading. The fully
dried GDL is then sintered at a higher temperature (typically 350 °C) for 30–60 min
to ensure that the hydrophobic agent distributes throughout the MPL, as this distribu-
tion is critical to imparting homogeneous surface characteristics [42]. Depending on
the specifications of the manufacturing process as well as the parameters, the MPL
thickness can show a great variation, typically between 10 and 100 µm [43]. Since
the MPL is not a stand-alone layer, its pore size is practically considered in terms of
its contribution to the known pore size distribution of the MPS.
As the key component of PEMFCs, MEA has a great influence on both perfor-
mance and cost of fuel cells. At present, the PEMFC requires a high loading of
precious metal Pt as the electrocatalysts to maintain its high energy output; however,
high loading means high cost, which will hinder the commercialization of this tech-
nology. Therefore, lowering the Pt loading without loss of performance of MEA has
become the most critical challenge restricting the commercialization of PEMFCs.

1.4 Electrocatalysts

At both the cathode and anode of a PEMFC, platinum-group metals (PGMs) are
currently required to catalyze the desired redox reactions (hydrogen oxidation at the
anode and oxygen reduction at the cathode). As these metals are commodities and
are all quite scarce, increased demand for PEMFCs will only serve to increase the
price of these catalysts if the loading is not reduced significantly from current levels.
Because of the sluggish kinetics of the oxygen reduction reaction (ORR) (∼5 orders
of magnitude slower than hydrogen oxidation kinetics) [44]. This challenge is widely
recognized in the PEMFC community and has led to a strong focus on improving the
catalysts used for the ORR at the cathode. This includes both improving the activity
and utilization as well as the durability and stability of these catalysts.
During the past decades, a wide variety of highly promising ORR catalysts
have been developed. Broadly, these catalysts can be categorized as (1) Pt/C, (2)
Pt alloy, (3) core–shell, (4) shape-controlled nanocrystals, (5) nano-frames and
(6) nonprecious metal catalysts [45]. The advantages and disadvantages from an
industrial or commercialization perspective are summarized in Fig. 1.7.
1.4 Electrocatalysts 13

Fig. 1.7 Benefits and remaining challenges for each of the primary categories of electrocatalysts
[45]

1.4.1 Pt/C Catalysts

For the past decade, the commercial PEMFC products have relied heavily on Pt/C
catalysts. When first introduced, these catalysts offered significant advantages over
unsupported Pt black because of the much smaller nanoparticles that are achiev-
able with supported catalysts. The simplicity of these catalysts is both a benefit
and a drawback. From a synthetic perspective, there is little room to tailor activity
and durability when limiting the design to a single element (Pt). In fact, further
improvements in activity and durability with conventional Pt/C now rely on advances
in catalyst supports resulting in “catalyst-support” interactions which have been
reported to enhance both activity and durability of PGM based ORR catalysts. While
promising, these approaches are unlikely to meet long-term mass activity require-
ments using conventional Pt nanoparticles [46, 47]. From a manufacturing perspec-
tive, simpler systems are advantageous. However, while having only one component
in the synthesis (Pt) may be considered an advantage, the reality is that at large scales,
catalyst manufacturing costs are minimal in comparison to PGM costs.
14 1 Proton Exchange Membrane Fuel Cells (PEMFCs)

1.4.2 Pt Alloy/C Catalysts

Since the discovery of Pt alloys as superior ORR catalysts for fuel cells at UTC in
the 1980s [48, 49], they have attracted great attention and been considered as the
second generation fuel cell catalysts after pure Pt [50, 51]. With proper posttreatment
(acid washing, for example) of Pt alloys, they are able to achieve high mass activities
while also demonstrating similar or better durability compared to Pt/C. Considerable
work has been carried out over the past decades on carbon-supported binary alloys or
ternary alloys that demonstrate 2–3 times higher mass activity versus Pt/C. Various
reasons have been proposed to explain the improved electrocatalytic activity of Pt
alloys (such as PtCo, PtNi, PtCr and PtV, etc.). They include compressive strains
due to shorter Pt–Pt bond distances, higher surface roughness caused by the transi-
tion metal dissolution, downshifting the d-band center of Pt or changing the d-band
vacancy due to strain and ligand effects, delayed formation of oxide species [52,
53], etc. It has been generally agreed that a Pt-skin like surface is formed during the
initial acid treatment and potential cycling. The structural and electronic effects of
transition metals in the core, and subsurfaces play a significant role in weakening
the adsorption of oxygen containing species. The lower coverage of these oxygen
containing species has been thought to be beneficial to enhance the ORR activity
because they poison the active sites. The membrane and ionomer contamination
caused by the transition metals dissolved during fuel cell operation delayed their
applications in PEMFCs [54].
A low dissolution potential of a transition metal resulted in a high ORR activity
but low chemical stability. During electrochemical measurements, the surfaces of Pt
alloys became pure Pt due to dissolution of non-noble metals (noted as Pt-skeleton).
After a mild thermal annealing at 1000 K, Pt atoms segregated to the surface, while
non-noble-metal atoms moved to the sublayers resulting in a Pt-skin type surface.
The Pt specific activity trends of the Pt skin and pt skeleton were slightly different
with the former following the order of Pt < Pt3 Ti < Pt3 V < Pt3 Fe < Pt3 Ni < Pt3 Co.
The specific activity enhancement factor of annealed Pt3 Co was ∼5, while it was 3
for the nonannealed one. This result emphasized the importance of the posttreatment
of Pt alloys in tuning the ORR activity by engineering the surface structures of Pt
alloys [55].
To improve the stability and durability of these catalysts, further work on
preleaching of Pt alloy catalysts has been performed to remove the base metal
(deposited on the carbon surface or poorly alloyed to the Pt). Posttreatments by
either acid leaching (starting with Co/Ni rich alloys) and/or heat treatment of Pt
alloys have also been performed, resulting in increased stability and activity, due to
the formation of a Pt-rich skin. These approaches have proven quite successful in
improving the durability of alloy catalysts, with similar or improved durability versus
Pt/C reported at the MEA/stack level. The technological maturity of these catalysts
is highlighted by Toyota’s recent announcement that a PtCo-alloy is currently used
in the Mirai.
1.4 Electrocatalysts 15

Fig. 1.8 Specific and mass

activities of Pt3 Co/C at 0.9 V
with different particle sizes
[60]

Ternary alloys Besides bimetallic alloys, ternary alloys also have been studied in
recent years. Various approaches, such as combinatorial high throughput screening,
DFT calculations and facile synthesis strategy, have been applied to optimize compo-
sitions of alloys. The combinations of Pt–MN (M, N = Fe, Co, Ni, Ti, V, Sn, Cr,
Mn, Mo, Ag, Au, Pd, Ir) have been synthesized and evaluated by different groups.
Because of the possible synergetic effects, the activity and stability of the ternary
alloys might be higher than the corresponding binary ones [56, 57].

Particle Size Effect The particle size effect of Pt alloys on ORR activity is even more
complicated than that of pure Pt. In addition to effects purely from size, other param-
eters like composition, degree of alloying, annealing temperature and shape also play
roles in determining the activity [58]. It is difficult to make a meaningful conclu-
sion if other parameters (annealing temperature, composition, etc.) are also changed
the same time besides particle size. Wang and co-workers found that the specific
activity of Pt3 Co increased with increasing particle size as smaller nanoparticles
were oxidized at a lower potential, which led to a stronger adsorption of oxygenated
species and thus a lower ORR activity. The size-dependent mass activities presented
volcano-shape behavior, and the maximum mass activity was found around 4.5 nm
due to the two opposite trends in specific surface area and specific activity with
particle size. Loukrakpam et al. also found that the mass activity of Pt3 Co/C decreased
with particle size increasing from 3 to 8 nm [59, 60] (Fig. 1.8).

1.4.3 Core–Shell Catalysts

As discussed in previous sections, the commercialization of PEM fuel cells is encoun-

tering a few problems. The slow kinetics of electrocatalytic oxygen reduction and
the high loading of Pt of the cathode material are among those major ones. Further
16 1 Proton Exchange Membrane Fuel Cells (PEMFCs)

increasing the fuel cell efficiency and reducing the cost will highly depend on
availability of better electrocatalysts with decreased Pt content and increased activity.
Driven by high activity and durability, the material structure at the nanoscale
has been explored for Pt-bimetallic catalysts. Core–shell structures especially show
superior electrocatalytic performance for ORR. Significant progress has been made
in recent years on the highly promising core–shell family of ORR catalysts. The
core–shell concept relies on having the active ORR catalyst (Pt) located only on the
surface of the nanoparticle, with another metal (most typically Pd) making up the
bulk. This unique design and concept theoretically can allow for the highest possible
Pt utilization (surface Pt to bulk) and thus from a cost perspective is highly attrac-
tive (provided less expensive cores can be developed). Additionally, the inherent
rate of the ORR can be tuned through changing the core, which has both struc-
tural and electronic impacts on the Pt shell. The core–shell is a general structure
describing the Pt-enriched surface of the catalyst particle and can be classified into
three distinct forms: Pt skin, Pt monolayer and Pt nanoporous skeleton, as shown in
Fig. 1.9 [61, 62].

Fig. 1.9 a Core–shell structure: Pt skin, Pt monolayer and Pt dealloyed Pt-bimetallic nanoparticles
[61] b Synthesis of Pt skin, Pt mono layer and Pt dealloyed Pt-bimetallic nanoparticles with a
core-shell structure
1.4 Electrocatalysts 17

Pt Skin Two different polycrystalline alloys of Pt3 Ni and Pt3 Co with 75% Pt and
100% Pt on the surfaces were prepared in an ultrahigh vacuum (UHV). The latter
is a Pt-skin structure and produced by an exchange of Pt and Co in the surface
layers. It was found that the activity for Pt skin on Pt3 Co was enhanced to 3–4
times greater than that of pure Pt. The reason behind superior performance of the
enriched Pt-skin structure is that the alloying Pt with 3d transition metals tunes the
electronic structure of catalyst surfaces. The activity correlates well with the strength
of the oxygen-metal bonded interaction, which in turn depends on the position of
the metal d states relative to the Fermi level. This is now well-known volcano-
type activity graph. The principle used for designing such catalysts is searching for
surfaces those bind oxygen a little weaker than Pt. Specifically for Pt skin, the Pt
d state must shift downward, thus giving improved ORR performance. To try and
further advance the ORR activity, Pt-ternary alloys (Pt3 (MN)1 with M, N = Fe, Co,
Mn, or Ni) were studied as electrocatalysts. These studies indicate that Pt-ternary
alloys achieve higher catalytic activities than bimetallic Pt alloys. Au@Pt, Pt@Au
and Fe3 O4 @Au@Pt nanoparticles were synthesized and are catalytically active for
ORR. Their electrocatalytic activities depended on the nanoscale spatial arrangement
of metals. An unsupported AuPt core–shell catalyst was prepared and demonstrated
an outstanding specific activity [63, 64].
Pt Monolayer An effective strategy to reduce the Pt content while retaining the
activity of a Pt-based catalyst is to deposit a few atomic layers of Pt atoms on top
of nanoscale substrates. Through his strategy, the available active surface area of Pt
catalysts is maintained but the PGM loading can decrease dramatically. Furthermore,
if the outermost top layers can be limited to only a few atomic layers, the activity
of such layers for ORR may be significantly different when compared to the bulk
materials.
Zhang and co-workers found that platinum monolayers supported on Au (111), Ir
(111), Pd (111), Rh (111) and Ru (0001) surfaces show a volcano-type dependence
on the d-band center of the platinum monolayer structures. The results elucidate
for the first time that for a good ORR electrocatalyst the kinetics of both the OO
bond breaking and the hydrogenation of reactive intermediates have to be facile at
the cathode. More importantly and from a practical perspective, we show that it is
possible to devise an ORR electrocatalyst that contains only a fractional amount
of platinum but can surpass the activity of pure platinum (e.g., Pt monolayer on
Pd(111)). Further, they investigated the complex monolayer catalysts (PtN)ML Pd (N
= Ir, Ru, Rh, Re or Os), the new electrocatalysts have: Pt mass-specific activity
up to twenty times higher than the state-of-the-art commercial catalysts, increased
stability of Pt against oxidation and very low Pt content [65]. Their superior activity
and stability reflect a low OH coverage on Pt, caused by the lateral repulsion between
the OH adsorbed on Pt and the OH or O adsorbed on neighboring, other than Pt,
late transition metal atoms. This new class of electrocatalysts promises to alleviate
some major problems of existing fuel cell technology by simultaneously decreasing
materials cost and enhancing performance provided their stability is satisfactory.
Therefore, platinum monolayers supported on appropriate metal substrates represent
18 1 Proton Exchange Membrane Fuel Cells (PEMFCs)

a viable way to reduce platinum loadings and the associated cost of fuel cell elec-
trodes. The results also point to the important possibility of fine-tuning the electrocat-
alytic activity of transition metals. The carbon-supported Pdx Ni1−x @Pt electrocata-
lysts also possess high catalytic activity and highly efficient NM utilization toward
ORR [66].

Pt Nanoporous Skeleton Under Prolonged exposure to reaction conditions, the

Pt-bimetallic catalyst with multilayered Pt-skin surface exhibited an improvement
factor of more than 1 order of magnitude in activity versus conventional Pt cata-
lysts []. Strictly, the Pt-skin surface here is actually an electrochemical dealloyed
nanoporous structure. Rough or uneven nanoporosity formation on the single cata-
lyst particle by leaching or electrochemical dealloying. It was demonstrated that the
Pt3 Ni(111) surface is 10-fold more active for ORR than the corresponding Pt(111)
surface and 90-fold more active than current state-of-the-art Pt/C catalysts [67]. The
Pt nanoporous skeleton catalyst is the most recent family of PGM ORR catalysts also
called the “nanoframe.” These catalysts consistently show significantly higher mass
activity than commercial Pt/C (up to 20-fold higher than commercial Pt/C based on
RDE studies). The largest advantage of this catalyst type is their superb stability
and durability during voltage cycling. The unique design of nanoframe catalysts
allows them to benefit from the high activity and stability typically associated with
extended platinum surfaces, while still achieving excellent ECSAs (>50 m2 /g) due
to their relatively thin frames (<2–3 nm). These catalysts are typically terminated by
highly uniform crystal surfaces, which is advantageous for durability as it is known
that defect sites in conventional nanoparticles are more susceptible to dissolution.
Additionally, having such uniform structures helps to reduce the Ostwald ripening
process typically observed for conventional nanoparticles, whereby large particles
get larger and small particles get smaller. Thus, conceptually it is very reasonable to
expect high durability from these catalysts (unlike some shape-controlled catalysts),
and there appears to be no disagreement in the literature on this topic.

While highly promising at the RDE level, little has been published on perfor-
mance at the MEA level with these catalysts. In RDE studies, the electrolyte can
fully penetrate these catalysts, and mass transport of protons to the inner surface
has not been a challenge. However, previous work has shown that it can be chal-
lenging to have the ionomer penetrate into pores <20 nm. Thus, in an MEA, it is
likely that an ionomer will have difficulty penetrating into these cages, which typi-
cally have openings <20 nm. In this case, significant proton transport limitations
would be envisioned at moderate to high current densities, leading to poor catalyst
utilization (access of reactants to the Pt surface at a given current density). Very
preliminary MEA testing data using nano-frame catalysts does appear to show these
anticipated transport problems at moderate current densities. It is possible that, for
high current density and automotive applications, these highly advanced catalysts
may require highly advanced catalyst layer design changes (e.g., incorporation of
protic ionic liquids) to help overcome this transport problem. However, as mentioned
1.4 Electrocatalysts 19

previously, this problem is far less of a concern for PEMFC products which operate
at lower current densities, and the significant activity improvement of the nano-frame
catalysts versus conventional Pt/Pt alloy would provide valuable improvements in
efficiency [68].
Before moving to the next family of catalysts, a brief comment on the ECSA
of core–shell catalysts must be made. While extraordinarily high ECSAs are often
reported for core–shell catalysts when considering Pt only, the true ECSA must
consider all PGMs in the catalyst. At the core of the most promising core–shell
catalysts is often Pd, which when included, decreases the ECSA by ∼60%. Thus, for
practical use, it is imperative that less expensive cores (PGM alloy or non-PGM) are
explored. Fortunately, this work is already underway, with highly promising results
at the RDE level.

1.4.4 Shape-Controlled Pt-Based Nanocrystals

While at an earlier stage of development than core–shell catalysts, shape-controlled

catalysts (Fig. 1.2) appear to be a highly promising class of ORR catalyst because
of their extremely high mass activities [69]. As described above, mass activity is
dependent on specific activity and Pt utilization (Pt dispersion). Core–shell catalysts
most strongly exemplify the “Pt utilization” strategy to generate high mass activities.
Conversely, shape-controlled catalysts rely primarily on achieving extraordinarily
high specific activities to generate high mass activities. In this way, these catalysts
have the closest ties to the fundamental single-crystal studies mentioned earlier in
this perspective. In principle, these catalysts attempt to recreate the ideal crystal
structure identified by single-crystal studies, but at the nanometer scale. An excellent
example of this is the 9 nm Pt2.5 Ni octahedra developed at the Georgia Institute of
Technology. Despite the relatively poor Pt utilization (Pt dispersion) afforded by the
large 9 nm particles in this study, a mass activity of 3.3 A/mg was achieved (more
than sevenfold higher than the 2020 DOE MEA target, albeit at the RDE level).
This was accomplished through maintaining the ideal Pt2.5 Ni(111) crystal structure
(which has >50× higher specific activity versus commercial Pt/C) at the nanometer
scale.
Despite the great promise shown by these catalysts, it is still too early in the
development timeline to make firm conclusions on their commercial viability (the
majority of testing on this family of catalyst has been at the RDE level only). In
particular, there are two issues which may prove highly challenging for these cata-
lysts when used in an operating fuel cell: (1) ECSA and (2) durability and stability.
ECSA: To achieve high performance at high current densities with a PGM loading of
0.1 mg/cm2 , an ECSA of ∼50 m2 /g will be required assuming no substantial advances
in ionomer (discussed later in this Perspective). Unfortunately, it has proven difficult
to maintain the desired crystal face as the particle size is decreased to 10 nm,
leaving ECSA as a key challenge for this family of catalysts for automotive applica-
tions. For nonautomotive applications (e.g., portable power, backup power, materials
20 1 Proton Exchange Membrane Fuel Cells (PEMFCs)

handling), high current densities are not as critical. For these applications, efficiency
is a key metric, and they thus operate at lower current densities. In this case, achieving
an ECSA of >50 m2 /g is less of a concern, and the much higher mass activity of these
advanced catalysts versus commercial Pt (or Pt alloy) catalysts could provide a valu-
able increase in efficiency for products used in these other important markets. It has
been reported that shape-controlled catalysts have lower stability than commercial
Pt/C when subjected to voltage cycling [70]. In fact, the above mentioned catalyst
exist only in a “metastable” state and will inevitably change to the thermodynami-
cally preferred “round” shape following voltage cycling [41]. The use of core–shell
type strategies may help to overcome this limitation, with one report showing almost
no loss in activity following an aggressive voltage cycling protocol for a PtPd-Ni
core–shell octahedral catalyst. Ultimately, MEA testing will be required to properly
understand the merits and limitations of this catalyst type. In this regard, some MEA
testing performed at GM has already indicated that some of these shape-controlled
catalysts (e.g., octahedral PtNi) are quite unstable, thus emphasizing the need for
further MEA evaluation [71].

1.4.5 Nonprecious Metal Catalysts (NPMC)

Much of the above discussion is framed around the DOE 2020 PGM loading target
of 0.125 mg/cm2 . As discussed, some challenges to achieving this goal remain.
However, as recently pointed out by GM, an even lower loading of 0.0625 mg/cm2
may be required before PEMFC vehicles can truly compete with traditional internal
combustion engine vehicles. Achieving this will clearly be a significant challenge
moving forward, highlighting the strong incentive toward developing NPMCs.
However, while these catalysts have the potential to greatly reduce cost, they are
still very far from meeting automotive requirements. Despite this, some of these
NPMCs have now demonstrated sufficient performance to be considered for use in
certain nonautomotive applications, such as backup power and/or portable power,
which have considerably lower performance, stability and durability requirements
than automotive applications. While the direct cost benefit may not be as large as for
automotive applications (due to differences in scales and sales volume), NPMCs offer
the very appealing advantage of being highly resistant to contaminants compared to
conventional PGM catalysts [72]. This has the potential to reduce overall stack and
system costs and may allow PEMFC products to operate in extreme environments that
would otherwise be very challenging. Additionally, these commercial markets could
provide an excellent “interim” target and provide useful “real-world” data for NPMC
researchers. However, prior to being truly commercialized, significant improvements
in stability are still required, as highlighted by our group in an extensive review on
this topic.
1.5 Summary 21

1.5 Summary

The high cost of PEMFCs is mainly caused by using the expensive and scarce Pt
materials, which accounts for over half of the total cost. In order to realize the DOE’s
2020 target (mass activity (0.44 A/mg) and loss in initial activity (<40% loss) targets)
and control the cost, MEA based on low Pt catalyst is necessary.

References

1. Liu CY, Sung CC (2012) A review of the performance and analysis of proton exchange
membrane fuel cell membrane electrode assemblies. J Power Sources 220:348–353
2. Yang SY, Seo DJ, Kim MJ et al (2016) Fast stack activation procedure and effective long-
term storage for high-performance polymer electrolyte membrane fuel cell. J Power Sources
328:75–80
3. Borup RL, Meyers JP, Pivovar BS et al (2007) Scientific aspects of polymer electrolyte fuel
cell durability and degradation. Chem Rev 107:3904–3951
4. Tseng CJ, Lo SK (2010) Effects of microstructure characteristics of gas diffusion layer and
microporous layer on the performance of PEMFC. Energy Convers Manage 51:677–684
5. Appleby AJ, Foulkes FR (1989) Fuel cell handbook. Van Nostrand Reinhold, New York
6. Ticianelli EA, Derouin CR, Redondo A et al (1988) Methods to advance technology of proton
exchange membrane fuel cells. J Electrochem Soc 135:2209–2214
7. Wilson MS, Gottesfeld SJ (1992) Influence of the structure in low Pt loading electrodes for
polymer electrolyte fuel cells. J Electrochem Soc 139:L28–30
8. Mehta V, Cooper J (2003) Review and analysis of PEM fuel cell design and manufacturing. J
Power Sources 114:32–53
9. Leimin X, Shijun L, Lijun Y et al (2009) Investigation of a novel catalyst coated membrane
method to prepare low platinum-loading membrane electrode assemblies for PEMFCs. Fuel
Cells 9:101–105
10. Grubb WT Jr (1959) Batteries with solid ion-exchange electrolytes. J Electrochem Soc 106:275
11. Grot W, Resnick P (1987) U.S. Patent 4,113,585
12. Kraytsberg A, Ein-El Y (2014) Review of advanced materials for proton exchange membrane
fuel cells. Energy Fuels 28:7303–7330
13. Michael AY, Matthew JL, Steven JH et al (2019) New directions in perfluoroalkyl sulfonic
acid-based proton-exchange membranes. Curr Opin Electrochem 18:90–98
14. Zhang LW, Chae SR, Hendren Z et al (2012) Recent advances in proton exchange membranes
for fuel cell applications. Chem Eng J 204:87–97
15. Kreuer KD (1996) Proton conductivity: materials and applications. Chem Mater 8:610–641
16. Haubold HG, Vad T, Jungbluth H et al (2001) Nano structure of NAFION: a SAXS study.
Electrochim Acta 46:1559–1563
17. Gebel G (2000) Structural evolution of water swollen perfluorosulfonated ionomers from dry
membrane to solution. Polymer 41:5829–5838
18. Herring AM (2006) Inorganic-polymer composite membranes for proton exchange membrane
fuel cells. J Macromol Sci Part C Polym Rev 46:245–296
19. Sun X, Simonsen SC, Norby T et al (2019) Composite membranes for high temperature PEM
fuel cells and electrolytes: a critical review. Membranes 9:1–46
20. Okazoe T, Shirakawa D, Murata K (2012) Application of liquid-phase direct fluorination: novel
synthetic methods for a polyfluorinated coating material and a monomer of a perfluorinated
polymer electrolyte membrane. Appl Sci 2:327–341
21. Holdcroft S (2014) Fuel cell catalyst layers: a polymer science perspective. Chem Mater
26:381–393
22 1 Proton Exchange Membrane Fuel Cells (PEMFCs)

22. Conway BE, Tilak BV (2002) Interfacial processes involving electrocatalytic evolution and
oxidation of H2 , and the role of chemisorbed H. Electrochim Acta 47:3571
23. Fishtik I, Callaghan CA, Fehribach RD (2005) A reaction route graph analysis of the
electrochemical hydrogen oxidation and evolution reactions. J Electroanal Chem 576:57–63
24. Debe MK (2012) Electrocatalyst approaches and challenges for automotive fuel cells. Nature
486:43–51
25. Norskov JK, Rossmeisl J, Logadottir A et al (2004) Origin of the overpotential for oxygen
reduction at a fuel-cell cathode. J Phys Chem B 108:17886–17892
26. Karen E, Swider L, Stephen A (2013) Campbell. Physical chemistry research toward proton
exchange membrane fuel cell advancement. J Phys Chem Lett 4:393–401
27. Gasteiger HA, Panels JE, Yan SG (2004) Dependence of PEM fuel cell performance on catalyst
loading. J Power Sources 127:162–171
28. Gasteiger HA, Kocha SS, Sompalli B et al (2005) Activity benchmarks and requirements for
Pt, Pt-alloy, and non-Pt oxygen reduction catalysts for PEMFCs. Appl Catal B 56:9–35
29. Alia SM, Ngo C, Shulda S et al (2016) Highly active and durable extended surface oxygen
reduction electrocatalysts. ECS Meeting Abstracts MA2016-02, 2445
30. Park J, Wang H, Vara M et al (2016) Platinum cubic nanoframes with enhanced catalytic activity
and durability toward oxygen reduction. ChemSusChem 9:2855–2861
31. Costamagna P, Srinivasan S (2001) Quantum jumps in the PEMFC science and technology
from the 1960s to the year 2000: Part I. Fundamental scientific aspects. J Power Sources
102:242–252
32. Wilson MS, Gottesfeld S (1992) High performance catalyzed membranes of ultra-low Pt
loading for polymer electrolyte fuel cells. J Electrochem Soc 139:L28–L30
33. Srinivasan S, Ticianelli EA, Derouin CR et al (1988) Advances in solid polymer electrolyte
fuel cell technology with low platinum loading electrodes. J Power Sources 22:359–375
34. Cindrella L, Kannan AM, Lin JF et al (2009) Gas diffusion layer for proton exchange membrane
fuel cells—a review. J Power Sources 194:146–160
35. Jayakumar A, Sethu S, Ramos M et al (2014) A technical review on gas diffusion, mechanism
and medium of PEM fuel cell. Ionics 21:1–18
36. Jorg R, Jens E, Federica M et al (2013) Investigation of the representative area of the water
saturation in gas diffusion layers of polymer electrolyte fuel cells. J Phys Chem C 117:25991–
25999
37. Mathias M, Roth J, Lehnert W (2003) Diffusion media materials and characterization. John
Wiley & Sons, New York
38. Liao YK, Ko TH, Liu CH (2008) Performance of a polymer electrolyte membrane fuel cell
with fabricated carbon fiber cloth electrode. Energy Fuels 22:3351–3354
39. Liu CH, Ko TH, Chang EC et al (2008) Effect of carbon fiber paper made from carbon felt with
different yard weights on the performance of low temperature proton exchange membrane fuel
cells. J Power Sources 180:276–282
40. Yazici MS (2007) Mass transfer layer for liquid fuel cells. J Power Sources 166:424–429
41. Thomas YRJ, Benayad A, Schroder M et al (2015) New method for super hydrophobic treatment
of gas diffusion layers for proton exchange membrane fuel cells using electrochemical reduction
of diazonium salts. ACS Appl Mater Interfaces 7:15068–15077
42. Laoun B, Kasat HA, Ahmad R et al (2018) Gas diffusion layer development using design
of experiments for the optimization of a proton exchange membrane fuel cell performance.
Energy 151:689–695
43. Ozden A, Shahgaldi S, Zhao J et al (2018) Assessment of graphene as an alternative microporous
layer material for proton exchange membrane fuel cells. Fuel 215:726–734
44. Barbir F (2005) PEM fuel cells: theory and practice. Elsevier Academic Press, Amsterdam
45. Banham D, Ye SY (2017) Current status and future development of catalyst materials and
catalyst layers for proton exchange membrane fuel cells: an industrial perspective. ACS Energy
Lett 2:629–638
46. Yu X, Ye S (2007) Recent advances in activity and durability enhancement of Pt/C catalytic
cathode in PEMFC: Part II: Degradation mechanism and durability enhancement of carbon
supported platinum catalyst. J Power Sources 172:145–154
References 23

47. Yu X, Ye S (2007) Recent advances in activity and durability enhancement of Pt/C catalytic
cathode in PEMFC: Part I. Physico-chemical and electronic interaction between Pt and carbon
support, and activity enhancement of Pt/C catalyst. J Power Sources 172:133–144
48. Jalan V, Taylor EJ (1983) Importance of interatomic spacing in catalytic reduction of oxygen
in phosphoric acid. J Electrochem Soc 130:2299
49. Jalan VM, Landsman DA, Lee JM (1980) U.S. Patent 4,192,907
50. Toda T, Igarashi H, Uchida H et al (1999) Enhancement of the electroreduction of oxygen on
Pt alloys with Fe, Ni, and Co. J Electrochem Soc 146:3750–3756
51. Stamenkovic V, Schmidt TJ, Ross PN et al (2002) Surface composition effects in electrocatal-
ysis: kinetics of oxygen reduction on well-defined Pt3 Ni and Pt3 Co alloy surfaces. J Phys Chem
B 106:11970–11979
52. Jia QY, Liang WT, Bates MK et al (2015) Activity descriptor identification for oxygen reduction
on platinum-based bimetallic nanoparticles: in situ observation of the linear composition-strain-
activity relationship. ACS Nano 9:387–400
53. Stephens IEL, Bondarenko AS, Gronbjerg U et al (2012) Understanding the electrocatalysis
of oxygen reduction on platinum and its alloys. Energy Environ Sci 5:6744–6762
54. Park HY, Jeon TY, Jang JH et al (2013) Enhancement of oxygen reduction reaction on PtAu
nanoparticles via CO induced surface Pt enrichment. Appl Catal B 129:375–381
55. Stamenkovic VR, Mun BS, Arenz M et al (2007) Trends in electrocatalysis on extended and
nanoscale Pt-bimetallic alloy surfaces. Nat Mater 6:241–247
56. Wang C, Li D, Chi M et al (2012) Rational development of ternary alloy electrocatalysts. J
Phys Chem Lett 3:1668–1673
57. Ozenler S, Sahin N, Akaydin B et al (2011) ECS Trans 41:1031–1042
58. Wang C, Wang G, van der Vliet D et al (2010) Monodisperse Pt3Co nanoparticles as electro-
catalyst: the effects of particle size and pretreatment on electrocatalytic reduction of oxygen.
Phys Chem Chem Phys 12:6933–6939
59. Loukrakpam R, Luo J, He T et al (2011) Nanoengineered PtCo and PtNi catalysts for oxygen
reduction reaction: an assessment of the structural and electrocatalytic properties. J Phys Chem
C 115:1682–1694
60. Shao MH, Chang QW, Dodelet JP et al (2016) Recent advances in electrocatalysts for oxygen
reduction reaction. Chem Rev 116:3594–3657
61. Hou JB, Yang M, Ke CC et al (2020) Platinum-group-metal catalysts for proton exchange
membrane fuel cells: from catalyst design to electrode structure optimization. EnergyChem
62. Wang C, Chi M, Li D et al (2011) Design and synthesis of bimetallic electrocatalyst with
multilayered Pt-Skin surfaces. J Am Chem Soc 133:14396–14403
63. Luo J, Wang L, Mott D et al (2008) Core@shell nanoparticles as electrocatalysts for fuel cell
reactions. Adv Mater 20:4342–4347
64. Hartl K, Mayrhofer KJ, Lopez M et al (2010) AuPt core-shell nanocatalysts with bulk Pt
activity. Electrochem Commun 12:1487–1489
65. Zhang JL, Vukmirovic MB, Sasaki K et al (2005) Mixed-metal Pt monolayer electrocatalysts
for enhanced oxygen reduction kinetics. J Am Chem Soc 127:12480–12481
66. Luo LX, Zhu FJ, Tian RX et al (2007) Composition-graded Pdx Ni1−x nanospheres with Pt
monolayer shells as high-performance electrocatalysts for oxygen reduction reaction. ACS
Catal 7:5420–5430
67. Stamenkovic VR, Fowler B, Mun BS (2007) Improved oxygen reduction activity on Pt3 Ni
(111) via increased surface site availability. Science 315:493–497
68. Tian X, Luo J, Nan H et al (2016) J Am Chem Soc 138:1575–1583
69. Choi SI, Shao MH, Lu N et al (2014) Synthesis and characterization of Pd@Pt-Ni core-shell
octahedra with high activity toward oxygen reduction. ACS Nano 8:10363–10371
24 1 Proton Exchange Membrane Fuel Cells (PEMFCs)

70. Li YD, Wang C, Strmcnik DS et al (2014) Functional links between Pt single crystal morphology
and nanoparticles with different size and shape: the oxygen reduction reaction case. Energy
Environ Sci 7:4061–4069
71. Kongkanand A, Mathias M (2016) The priority and challenge of high-power performance of
low-platinum proton-exchange membrane fuel cells. J Phys Chem Lett 7:1127–1137
72. Kishimoto T, Sato T, Kobayashi Y et al (2016) ECS meeting abstracts 2, 2824
Chapter 2
The Electrocatalysis of Oxygen
Reduction on Platinum and Its Alloys

Abstract The high cost and poor durability are the pressing issues that need to be
addressed before pushing the polymer electrolyte membrane fuel cells (PEMFCs)
to the market. The primary effort is made on the electrocatalysis by lowering the
loading of Pt-group which are scarce and expensive but remain as the most efficient
material until now. State-of-the-art Pt and Pt-based electrocatalysts usually offer an
exceptional electrocatalytic activity as well as durability toward the oxygen reduc-
tion reaction (ORR), but the trick lies in their well-controlled surface structure and
composition. Theoretical discussion based on the first principle about the cathodic
ORR in this chapter will provide new insights into material choice, surface element
distribution and the closed crystallographic orientation designs for electrocatalysts.
Therein, alloying Pt with 3d transition metals, Pt shells and {111}-facets are regarded
as excellent characteristics for improving the ORR performance. The experimental
strategies including chemical routes, electrodeposition and thermal annealing for
achieving those well-designed structures are then concluded. Besides, we also shed
light on the hinder for practical application of those electrocatalysts in the PEMFCs.

Keywords The oxygen reduction reaction · Low platinum-group metal

electrocatalysis · Composition effects · Shape-controlled Pt-based alloys

2.1 Introduction

In this section, we will have a brief understanding about the cathode reaction.
Compared with the anode reaction, the cathode reaction is more kinetically slug-
gish and mainly obstructs the chemical–electrical energy conversion in the PEMFCs.
High performance electrocatalysts are therefore in need to facilitate this process.

© Shanghai Jiao Tong University Press and Springer-Verlag GmbH Germany 2021 25
J. Zhang and S. Shen, Low Platinum Fuel Cell Technologies,
Energy and Environment Research in China,
https://doi.org/10.1007/978-3-662-56070-9_2
26 2 The Electrocatalysis of Oxygen Reduction on Platinum …

2.1.1 Cathode Reaction

In a typically PEMFC, the cathode reaction is referred to the oxygen reduction reac-
tion (ORR). The use of a solid organic poly-perflourosulfonic acid as the electrolyte
inside establishes a strong acidic environment (PH < 1) for the reduction reaction.
Generally, the ORR can proceed through either a direct four-electron pathway (O2
+ 4H+ + 4e− → 2H2 O) or a less efficient two-step two-electron pathway in acid
media. The later one involves the generation of H2 O2 as intermediate (O2 + 2H+ +
2e− → H2 O2 ), then the intermediate can undergo further reduction to H2 O (H2 O2 +
2H+ + 2e− → 2H2 O) or be electrochemically oxidized back to O2 (2H2 O2 → 2H2 O
+ O2 ). To achieve a high efficiency, the direct four-electron pathway is more desir-
able for PEMFCs. The ideal standard potential for this reaction is calculated to be
1.23 V. However, it is difficult to experimentally obtain the thermodynamic reversible
potential as the reaction is highly irreversible. Besides, the four-electron reduction
process is kinetically sluggish in nature, resulting in a large overpotential loss asso-
ciated with the slow reaction rate. Usually, the rest potential of the ORR under room
temperature and atmospheric pressure will not exceed 1.1 V versus the reversible
hydrogen electrode (RHE).

2.1.2 Platinum and Platinum-Based Catalysts

More importantly, among many factors influencing the chemical–electrical energy

conversion, the sluggish kinetics of the cathodic ORR is the pivotal hurdle for the
development of cost-effective PEMFCs [1, 2]. Catalysts must be applied to facilitate
the four-electron pathway and boost the efficiency of fuel cells. Traditionally, such
electrocatalyst is pure platinum, and it has been commercialized for cathodic use in
PEMFCs, but the exchange current density for the ORR at Pt surface is still limited.
In the past several years, strenuous efforts had been devoted to the exploration of
more active ORR electrocatalysts. It is convinced that shaped Pt and Pt containing
catalysts as well as Pt group metal free catalysts can offer exceptional electrocatalytic
activity toward the ORR [3–16]. However, one typical shortcoming of Pt group metal
free catalysts is their poor stability in acidic environment. Therefore, although the
high Pt loading in cathode could result in the great increase of initial catalyst cost,
which is ~41% in a PEMFC stack according to the latest analysis from Annual Merit
Review and Peer Evaluation Meeting [17], shaped Pt and Pt containing alloys remain
as the most practical and promising catalysts that are both active and stable for the
use in fuel cell technology. More detailed reason of Pt being superior for the ORR
will be given fundamentally in the next section.
As is predicted, the cathodic loading of Pt group metal should not exceed
0.05 mgPt /cm2 assuming a rated power of 1 W/cm2 to fall within the optimum cost
region for industry wide objective. Effective strategies on developing more active Pt
or Pt-based catalysts must be explored to cut down the amount of Pt for the cathodic
2.1 Introduction 27

ORR without compromising fuel cell performance. Moreover, evidence exists that Pt
could be dissolved directly or through the formed PtO intermediate under the acidic
cathodic environment and high voltage [18], which will undermine the durability and
then obstruct the marketization of PEMFCs. In this account, first, a comprehensive
discussion of oxygen reduction mechanism at Pt containing catalysts surfaces based
on computational works will be presented in this chapter. With a good understanding
of the relationship between the surface structure and the reaction performance, the
design strategies to achieve special structuralfactors for improving ORR activity and
durability are then summarized. Finally, the tricks and outlooks of using shaped Pt
and Pt-based catalysts in low Pt-loaded electrode will be given.

2.2 Reaction Mechanism

Density functional theory calculation is introduced in this section which could help to
calculate the Gibbs free energy of the adsorbed oxygen intermediates and understand
the detailed reaction mechanisms.

2.2.1 The Dissociation Mechanism and the Association

Mechanism

In general, the direct four-electron pathway predominates the oxygen reduction on

Pt-based catalysts. The whole four-electron pathway is predicted to proceed via
several detailed mechanisms, involving different adsorbed oxygen species (O*, OH*,
OOH* and HOOH*) as intermediates [19, 20]. Typically, this reaction follows either
the dissociation mechanism or the association mechanism depending on the surface
oxygen dissociation barrier. In the case of the dissociation mechanism, the oxygen
molecules dissociate on the catalyst’s surface followed by the electron transfer and
protonation to form H2 O. The reaction sequence is given by the followings, with *
denotes an active site on the surface:

O2 + 2∗ → 2O∗
2O∗ + 2H+ + 2e− → 2OH∗
2OH∗ + 2H+ + 2e− → 2H2 O + 2∗

As for the association mechanism, the first step becomes the electron and proton
transfer rather than the O–O bond cleavage. The overall reaction process is given
below:

O2 + ∗ → O2 ∗
O2 ∗ + H+ + e− → OOH∗
28 2 The Electrocatalysis of Oxygen Reduction on Platinum …

OOH∗ + H+ + e− → O∗ + H2 O
O∗ + H+ + e− → OH∗
OH∗ + H+ + e− → H2 O + ∗

Although plenty of real-time observation techniques in optical and electric filed

(in situ X-ray absorption spectroscopy, in situ X-ray scattering, in situ Raman spec-
troscopy, etc.) have shed light on the nanoscale processes nowadays, the experimental
detection of oxygen containing intermediates on catalysts’ surfaces as well as their
adsorbed configurations under the fuel cell operating conditions remain difficult.
The essential questions are the short lifetime and low coverage of those interme-
diates especially at high potential, let alone the influence from other co-adsorbed
species. Therefore, the detailed surface process of this four-electron reaction on the
atomic level is still elusive from the experimental point of view.

2.2.2 Theoretical Analysis on Pt Activity Toward the ORR

Density functional theory calculations can inform the stability of those intermediates
during the reactions and establish an thermodynamics overview of the ORR process,
so it is a suitable tool for revealing the reaction mechanism theoretically [21, 22]. The
free energies and the binding energies of each intermediates can be calculated based
on the first principle. By comparing the free energy barriers for oxygen reduction
consequence on Pt surface, the adsorbed oxygen is confirmed to be so stable that
obstructs the proton and electron transfer at high potentials. With the decrease of
potential, the stability of the adsorbed oxygen decreases, and the reaction proceeds.
This is suggested to be the origin of the overpotential for ORR on Pt surface. It can be
also concluded from density functional theory calculations that both dissociative and
associative reaction paths contribute the oxygen reduction process depending on the
type of catalysts and the potentials. It is suggested that the association mechanism is
more likely to happen on Pt and Pt-based catalysts surface.
If the theoretical catalytic activities are plotted with the obtained O binding energy
on different close-packed metals surface, the well-known “volcano” type plot with Pt
near the top is constructed, as shown in Fig. 2.1 [19]. Known as the Sabatier Principle,
an active catalyst should neither bind intermediates too strongly nor too loosely but
with the optical O binding energy at the top of the volcano plot. Figure 2.1 can be
a great explanation for Pt being superior among the given class of pure metals. For
metals that bond O strongly, such as Pt and Ni, the rate-determining step is proton
and electron transfer to O* or OH*. On the contrary, for metals that bond O too
loosely, such as Au, the proton and electron transfer to oxygen are not expected to
happen. Therefore, the rate-determining step for those metals is proton and electron
transfer to O2 * based on the association mechanism or the O–O bond cleavage based
on the dissociation mechanism.
2.2 Reaction Mechanism 29

Fig. 2.1 Trends in the oxygen reduction activity plotted as a function of the oxygen binding energy
[19]

Based on the discussions above, it may arise to us that an excellent catalyst can be
designed if the binding energies of each intermediate are ideally modified. However,
it is worth to note that even the optimal catalyst near the top of the volcano plot
has a nonzero overpotential of 0.3–0.4 V. This is because the four-electron ORR
involves various intermediates, and the binding energies of each intermediate are
strongly correlated [23–25]. A scaling relationship between OH* and OOH*, which
is GOOH = GOH + 3.2 ± 0.2 eV, is found to be applicable among almost ever
face centered cubic metals. This even leads to the observation of overpotentials on
Pt-based catalysts with higher ORR activity and is considered as the key limitation
for designing the reversible oxygen electron.
What is more, the average energy of the whole d-band, or d-band center, is found
to be a universal descriptor for the reactivity of the active sites on transition metal
surfaces [26]. Density functional theory calculations show that the binding energies
correlate well with the position of d-band center versus the Fermi level, where the
low d-band center being favorable for the desorption of adsorbed oxygenated inter-
mediates. Due to any changes in d-orbital electronic structure can result in the shift
of d-band center, modification of Pt electronic properties is believed as a feasible
way for obtaining exceptional electrocatalytic ORR activity.

2.3 Electrochemical Measurements

Electrochemical measurements of the ORR activity are often carried out by using
either rotating disk electrode (RDE) or membrane electrode assembly (MEA). The
MEA is considered as the essential component of a fuel cell stack containing a thin
film ORR catalyst layer fabrication between the cathodic electrode and membrane.
30 2 The Electrocatalysis of Oxygen Reduction on Platinum …

MEA is particularly valuable for evaluating the catalyst performance in practical

applications of PEMFCs. However, on the one hand, the mass transport limitations
will also influence the single cell performance during the measurement, on the other
hand, new Pt-based catalysts synthesized at the lab scale are quit small in quanti-
ties which are not enough for single cell tests. RDE is designed for measuring the
catalyst performance based on a half-cell configuration under ideal film eletrode
conditions to simulate the real cathodic half reaction. Benefit from its easy access
and versatility, RDE method is popular in studying the performance of new ORR
catalysts. RDE measurement under the strong acidic condition with a small amount
of Pt loading is often sensitive to measurement conditions such as oxygen and ionic
concentration in the electrolyte and preparation techniques, and thus, the precau-
tion is necessary, and clean electrode surfaces are well promised. A three-electrode
system is employed with the RDE being the working electrode to load the catalyst
thin film. Cyclic voltammetry (CV) and linear sweep voltammetry (LSV) curses are
two typical electrochemical results obtained from RDE tests.
CV experiments are conducted in the N2 -saturated 0.1 M HClO4 to evaluate the
adsorption/desorption behaviors for H+ and oxygenated products. Electrochemical
active surface area (ECSA) of Pt electrocatalyst was calculated by integrating the
hydrogen adsorption/desorption regions according to the following equation:

ECSA = Q H /q0

where QH represents the hydrogen desorption charge calculated from CV curves

between 0.07 and 0.5 V [vs. reversible hydrogen electrode (RHE)], and the desorption
pseudocapacitance of q0 is well accepted as 210 µC/cm2 for polycrystalline Pt and
Pt-based alloys. LSV for the oxygen reduction is conducted by the potential scan
from 0 to 1.1 V (vs. RHE) in O2 -saturated 0.1 M HClO4 solution at 1600 rpm
for weakening the O2 mass transport limitations. The kinetic current density (jk ) is
then calculated from the ORR polarization curve using the Koutechy–Levich (K–L)
equation.

1/j = 1/jk + 1/jd

where j (mA/cm2 ) is the measured current density at 0.9 V (vs. RHE), and jd
(mA/cm2 ) is the diffusion-limited double layer current density, which is assessed at
nearly 0.3 V (vs. RHE).
The ORR kinetics is briefly discussed based on a standard linear sweep voltam-
metry curves obtained on commercial Pt surface (Fig. 2.2). As the reduction reaction
is fast at low potentials (less than 0.65 V), the current is predominantly limited by O2
diffusion. At operating potentials (around 0.9 V), O2 chemical adsorption is found to
be the rate-determining step, and the potential dependence of the ORR rate is given
by the availability of free surface sites for O2 adsorption, which in turn depends on
the coverage of *OH (and *O).
2.4 Further Improvement Toward Pt Activity 31

Fig. 2.2 Linear sweep voltammetry curves obtained on Pt surface

2.4 Further Improvement Toward Pt Activity

To get insight into practical application of Pt in real fuel cells, the ideal facile struc-
tures employed in the ORR mechanism studying are not suitable. Instead Pt nanopar-
ticles will be discussed in this section. However, it is still important to transform those
theoretical results we obtained in Sect. 2.2 into structure designs for Pt nanoparticles.

2.4.1 Pt Nanoparticles

The theoretical analyses discussed above mainly focus on well-defined extended

surfaces with simplified facile structure. The poor dispersion of those extended
surfaces, however, will limit their use in a real fuel cell compared with nanopar-
ticles. In fact, Pt or Pt-based alloys used in real PEMFCs are dispersed on a high
surface area support in nanoparticulate forms. So, it is also important to consider
the topography factors like steps, edges, kinks and facets in real catalysts. Among
the monometallic Pt, the catalysts from E-TEK (20% Pt/C) with initial surface area
around 100 m2 /gPt and TKK (46–50% Pt/C) with the electrochemical surface area
reaching 65 m2 /gPt have often been used as a reference for comparison of activities
and in common PEMFCs in various studies. Such catalysts are typically composed
particles in the size range between about 3–5 nm.
The mass activity of catalyst is important when considering costs in PEMFCs.
There are two typical ways for improving the mass activity of Pt, one is increasing the
number of active sites, and the other is improving the intrinsic activity of those active
sites. Decreasing the mean particle size to raise the dispersion of Pt is a common way
32 2 The Electrocatalysis of Oxygen Reduction on Platinum …

for obtaining more active sites. Evidences exist that by decreasing the mean size of
Pt particles from 15 to 3 nm, 5-fold improvement in mass activity can be observed
[27]. However, with the further decrease in particle size, the mass activity cannot be
improved. This is because the intrinsic activity of single Pt atom (or active site) on
the surface of Pt nanoparticle is position dependant. As is mentioned in the previous
section, the Pt-intermediates binding energies can be changed by modifying the Pt
electronic structure and will then influence the electrochemical properties. Due to the
special electronic structures in undercoordinated sites, Pt atoms in steps, edges and
kinks are trend to bind adsorbed oxygenated intermediates strongly and consequently
showing negligible contribution to the ORR. When the particle size is smaller than
3 nm, the edge sites will make up the majority of the space in the nanoparticle, and
the greater barriers for OH* removal thus predominantly hinder the reaction [28–30].
The crystallographic orientation and the surface structure will also influence the
kinetics of the ORR. Although the {110} plants are rough with stepped atoms whereas
{111} and {100} plants are composed with flat surfaces, and there is no evidence to
suggest that under-coordinated sites on Pt surface can improve the ORR activity, a
counterintuitive phenomenon that the ORR activities on the low-index planes of Pt
give the following order: Pt{100} < Pt{111} < Pt{110} is observed experimentally
on Pt single crystal. It may indicate that the abundant {111} terraces on Pt{110}
surface can act as the active sites and boost the reaction process (Fig. 2.3).
When we further translate the knowledge of electronic and structural effects
from those extended surfaces at single crystal to real nanoparticles, the interaction
between different effects becomes more complicated. Inspired by the works on single
crystal, Narayanan and El-Sayed [32] inquired into facet-dependent shape-controlled
monometallic Pt nanocatalyst for the enhancement of the ORR activity in 2005. Until
now, it is widely accepted that nanoparticles with the polyhedron-design 3d geometric

Fig. 2.3 Illustration of various 3D polyhedral configurations shown as a function of low index
{100}, {111} and/or {110} facets and high-index {hkl} facets [31]
2.4 Further Improvement Toward Pt Activity 33

configurations usually offer an exceptional electrocatalytic activity toward the ORR.

Those configurations include the cube, the octahedron, the icosahedron, etc., which
are enclosed by different low-index facets as well as the concave structures which
can generate the high-index facets (eg. Pt{720}, Pt{730}, Pt{830}) can result in
an unexplored improvement in electrochemical performance. However, the mech-
anisms for the enhanced performance from high-index planes are still elusive as
those planes with high surface energy are hard to obtained. Porous or hollow struc-
tures can be further introduced into the polyhedron-design nanoparticles to form
nanoframe and nanocage for example. Due to the increased surface area to volume
ratio, those porous catalysts can provide more active sites without compromise of
crystallographic orientation advantages.

2.4.2 An Organic Solvent System-Assisted Electrodeposition

of Highly Active Pt

Compared with the most commonly used chemical routes to synthesize Pt electrocat-
alysts, electrodeposition is expected to be a more facile and effective method to tailor
crystallographic orientation as well as to explore related synthetic mechanism via
simply tuning applied potential, deposition current density, concentration of added
metal precursors, etc. [33]. Despite the current limitations for a large-scale applica-
tion of such electrochemical technology, these technical barriers could be overcome
in future through the fabrication of large-area electrodes and direct electrodeposition
of nanoparticles on cathode frameworks [34]. So far, most of reported electrochem-
ical syntheses of Pt electrocatalysts were performed in aqueous solutions owing to
the fact that the standard reduction potentials of noble metals (Pt, Pd, Au, etc.) are
more positive than that of hydrogen evolution reaction [35, 36]. However, since
there exist several inherent defects for aqueous solutions including narrow potential
window, low thermal stability and easy evaporation, non-aqueous systems including
organic solvents, ionic liquids and molten salts deserve to be well developed [37–
39]. Among these novel non-aqueous systems, organic solvents are more accessible
and more economical than ionic liquids, while much higher temperature is needed
for handling molten salts. Herein, our group proposed an intriguing organic solvent
system-assisted electrodeposition method to prepare Pt which was proved to be highly
active toward the ORR. The complexing agent of N,N-Dimethylformamide (DMF)
was employed as the solvent, and influences of supporting electrolytes and poten-
tiostatic deposition behavior potentials on the deposition behavior were examined.
DMF is one of the commonly used simple organic compounds that has good chem-
ical stability, high polarity, good solubility and wide electrochemical window. Aside
from being used as the solvent, DMF plays important roles as reducing, capping as
well as coordinating agents and was widely used in our studies.
Figure 2.4a presents CV curves in DMF with different supporting electrolytes.
Two reduction peaks appear between −0.5 and −1.5 V, which can be assigned to
34 2 The Electrocatalysis of Oxygen Reduction on Platinum …

Fig. 2.4 The electrodeposition process of Pt on GC electrode within different DMF solutions [40].
a CVs curves; b illustration

two continuous reduction processes from H2 PtCl6 to Pt. The small peaks at around
−1.5 V might be caused by the concomitant hydrogen evolution reaction. As shown
in Fig. 2.4b of the experimental remarks, the electrodeposition conditions on glass
carbon (GC) electrode were largely affected by supporting electrolytes. A thin, less
adhesive and fully covered membrane on the GC electrode was obtained at −1.4 V
with KClO4 while deposition occurred solely on the joint area between GC and
Teflon holder to form an unusual “Pt-ring” at any potential from −1.0 to −2.0 V with
tetrabutylammonium chloride (TBACl) as the supporting electrolyte. It is revealed
from density functional theory calculations that the surface adsorption energy of
TBA+ on graphene is −0.6 eV, far below that of K+ , which is 0.45 eV, strongly
demonstrating that TBA+ is more likely to adsorb on the electrode surface compared
with K+ , thus prohibiting the Pt electrodeposition and leading to the formation of
“Pt-ring”. The adsorption energy is expected as an important factor for choosing
suitable supporting electrolyte in future electrodeposition exploration.
The influence of electrodeposition potentials was also investigated in this research.
It is obviously confirmed in Fig. 2.5 that the morphology varies with the deposition
potential from planar to dendritic and nanorod-like structures with an average diam-
eter of 11.5 nm. The hydrogen bubbles generated at −1.8 V as shown in Fig. 2.5f

Fig. 2.5 Morphology analysis includes SEM, TEM and SEAD images [40]. a −1.4 V, b −1.5 V,
c −1.6 V, d, g, h −1.7 V, e −1.8 V, f the detachable glassy carbon electrode at −1.8 V
2.4 Further Improvement Toward Pt Activity 35

should function as the template during the formation of dendritic and nanorod-
like electrodeposits. The delicate morphology of Pt electrocatalysts characterized
by transmission electron microscopy (TEM) and selected area electron diffraction
(SAED) in Fig. 2.5g, h demonstrates the formation of polycrystalline Pt nanoparti-
cles with an average size of about 4.72 nm at the deposition potential of −1.7 V. The
Pt electrocatalyst deposited at −1.7 V exhibits the highest area-specific activity of
0.84 mA/cm2 , which is 3.23 times than commercial Pt/C (0.26 mA/cm2 ).

2.5 Pt-Based Alloys

Alloying Pt with other metals is a more effective strategy to change its electronic
structure compared with changing the exposure of Pt at different low-coordination
sites. It is proved that Pt-M (M = Fe, Co, Ni, Cu, Zr, Y, etc.) alloys are anticipated to
ideally maximize the Pt utilization. In this section, we will focus on Pt-based alloys
for the ORR.

2.5.1 Theoretical Analysis on Pt-Based Alloys Activity

Toward the ORR

Although from Fig. 2.1 Pt lies near to the top of the thermodynamic activity-binding
energy volcano plot, the proton and electron transfer to adsorbed oxygenated products
on pure Pt is still more difficult than the adsorption process. Further weakening the
binding energy of adsorbed oxygenated products (OH*, OOH* and O* which are
strongly correlated) though any effective strategy can let the catalysts get closer to
the top. It is proved that for a catalyst that follows the above scaling relationship, the
optimum GOH of this catalyst should be around 0.1 eV larger than that of pure Pt for
obtaining the best ORR activity [41–43]. Introducing heterogeneous atoms into Pt
metal to form bimetallic or multimetallic Pt-based catalysts will result in electronic as
well as the related d-band center changes in the Pt metal, thus modifying the binding
energy of intermediates. Normally, the solute metals that selected to form Pt-based
alloys for the ORR should have a positive oxide or hydroxide formation potential
than its dissolution potential, and the formed alloys should be as stable as possible.
Strenuous studies have proved experimental that the combination of Pt with other
metals, especially transition metals (Fe, Co, Ni, Cu, Y, etc.) will effectively promote
the intrinsic activity of Pt.
Among different types of Pt-based bimetallics having been considered, the Pt3 X
(with X = Co, Ni, Y, Ti, etc.) family of materials have received the most attention
because of their exceptional electrocatalytic activity. In fact, the area-specific activity
on Pt3 Ni {111} is the highest among the low-index Pt bimetallic extended surfaces.
As is shown in Fig. 2.6, the superior catalytic activities of various Pt3 X alloys are
36 2 The Electrocatalysis of Oxygen Reduction on Platinum …

Fig. 2.6 Comparison of the kinetic volcano (based on microkinetic modeling) with the limiting
potential volcano (based on thermodynamic analyses) [44]

plotted as a function of the calculated *OH binding energy (relative to that of pure
Pt), and the same volcano trends are obtained for the range of Pt-based alloys through
both microkinetic modeling and thermodynamic analyze. Similar to the dissociations
for the monometal catalysts in Fig. 2.1, the binding energies strength of adsorbed
oxygenated species acts as opposing effects on two sides of volcano plot for the Pt
alloy catalysts. Combining with d-band center theory, this effect can be interpreted
as: in the case of Pt3 V and Pt3 Ti, the alloy process leads to an excessive weakening of
the OH* binding energy, and the d-band center is far away from the Fermi level, the
proton and electron transfer to oxygen therefore limits the ORR rate, as for Pt3 Y, the
weakening of OH* binding energy is not enough, and the impediment of the reaction
rate is the same with pure Pt.

2.5.2 Composition Effects

Two critical effects namely the ligand effect and the strain effect contribute to the
modification of the electronic properties in a bimetallic surface [26, 45–47]. The
ligand effect or electronic effect is introduced by the atomic proximity of two dissim-
ilar surface metal atoms. For example, when different composition of solute metals
are used in the alloys or different arrangement modes of solute metals near the
Pt atoms. It generally involves electron transfer between the two metal atoms. The
strain effect or geometric effect is brought about by the atomic arrangement of surface
atoms. It generally includes compressed or expanded arrangements of surface atoms.
The compressed Pt–Pt bond length is suggested to bring a down shift of d-band center,
2.5 Pt-Based Alloys 37

which will bind the OH* weaker than unstrained Pt and is expected to achieve by
alloying Pt with solute metals having a smaller lattice parameter (such as Ni and Co)
than Pt according to the Vegard’s law. Both the strain and ligand effects affect the Pt
electronic structure by changing the width of the d-band which subsequently moves
up or down in energy to maintain constant band filling.
Generally, both strain and ligand effects are simultaneously observed in ordinary
Pt-based catalysts. Core–shell structures are widely used to experimentally deconvo-
lute the interplay between those two effects in Pt–Cu system. By tuning the compo-
sition of dealloyed Pt–Cux with a trained pure Pt overlayer around 1 nm thickness
over a Cu-rich core, Strasser et al. [48] successfully demonstrated the strain effect.
The ORR activity decrease was observed with the leaching of Cu, which afterward
results in the decrease of strain in Pt overlayer. It seems that the improvement in alloy
degree could enhance the influence of the strain effects which is also confirmed in
our studies in Pt–Ni system. They also produced the series of spectroscopic exper-
iments for Pt overlayer deposited on Cu{111} single crystal. It is observed that the
d-band center of Pt shifted downwards, resulting in an increased occupancy of the O
2p and Pt 5d anti-bonding states. On the contrary, to study the ligand effect, a core–
shell structure with the bulks composed of Pt but Cu in the subsurface is designed.
This kind of structure can be obtained through thermal annealing a Pt bulk with a
underpotentially deposited Cu monilayer on its surfaces [49]. Because of the lower
surface energy of Pt, Cu is expected to migrate to the subsurface. 8-fold enhancement
over pure Pt was shown after thermal annealing. Since the bulk of the crystal was
composed of Pt, the ligand effect predominant.

2.5.3 The Co-Reduction of Pt with Co Ions

The highly active Pt–Co alloy catalyst has been successfully employed in Toyota
Mirai Fuel Cell Vehicles (FCVs) which is the first kind of Pt-based catalyst achieving
commercialized [50], demonstrating an increased activity by 1.8 times compared with
that for 2008 FCV model. Indeed, before or after, tremendous attention is being paid
to design and synthesis of various highly active Pt–Co alloy catalysts for the ORR in
acidic media [51–54]. By varying the transition metal/Pt atomic ratio, a optimal Pt
compressive strain toward the ORR can be obtained. Jia et al. [55] revealed that no
matter the distribution mode of Co element, either Co content gradually decreased
from the core to near-surface region or Co uniformly distributed in the near-surface
region, the improved ORR activity of Ptx Co/C nanoparticles could be definitely
related to Pt–Pt bond length that generates a favorable Pt compressive strain. Choi
et al. [56] found that among the various Co/Pt atomic ratios, the Pt 4f7/2 binding
energy of Pt3 Co alloy nanocube could be the most desired for the ORR, which was
the descriptor of an optimal Pt d-band center. The electrochemical method is also
employed for the synthesis of Pt–Co alloy. Although very promising, the co-reduction
of Pt with Co ions remains a great challenge due to a large deposition potential gap,
i.e., the standard reduction potential for Co2+ /Co (−0.28 V vs. SHE) is much more
38 2 The Electrocatalysis of Oxygen Reduction on Platinum …

negative compared with that for (PtCl4 )2− /Pt (0.755 V vs. SHE) [57]. Our group
proposes a facile electrodeposition method that employs DMF as the selective coor-
dinating agent to assist in the co-reduction. It is very ingenious that the selective
of coordination of DMF with Pt ions rather than Co ions greatly bridges the depo-
sition potential gap between Pt and Co as high as 200 mV, thus greatly promoting
the formation of Pt–Co alloys. The series of as-electrodeposited Pt–Co alloy cata-
lysts are detailedly investigated, and the optimal Pt4 Co alloy sample demonstrates a
remarkable specific activity toward the ORR.
The deposition behavior of two series of Pt–Co alloy catalysts in the absence
or presence of DMF was carefully examined by CV measurements, illustrated in
Fig. 2.7 [(a) 1 mM K2PtCl4+ 5 mM CoCl2+ 0.1 M KCl, (b) 5 mM CoCl2+ 0.1 M
KCl, (c) 1 mM K2 PtCl4+ 0.1 M KCl, (d) 0.1 M KCl in the absence or presence
of 60 mM DMF]. Figure 2.7b shows that Co deposition remains almost unchanged
after the addition of DMF. The reduction peak potential shift of Pt caused by DMF
is demonstrated in Fig. 2.7c. A blank CV is displayed in Fig. 2.7d for comparison.
It is well known that the carbonyl ligand of DMF can be well coordinated with Pt,
which will make it harder to reduce, thus shifting the peak reduction potential of Pt to
more negative values, as shown in Fig. 2.7c. However, according to Fig. 2.7b, there
is no difference in the peak reduction potential of Co no matter the addition of DMF,

Fig. 2.7 CV curves on the GC electrode in different electrolytes [58]

2.5 Pt-Based Alloys 39

which can be attributed to the fact that the coordination of Co with Cl− is stronger
than with DMF. According to Fig. 2.7a, for the deposition solution used in this work,
when in the absence of DMF, the peak potential for the reduction of (PtCl4 )2− to Pt
is −0.34 V (vs. SCE), while −0.54 V (vs. SCE) for that in the presence of DMF,
indicating a significant coordinating effect of DMF on Pt ions. In contrast, no matter
the existence of DMF or not, the peak potential for Co deposition remains at −0.8 V
(vs. SCE), indicating a selective coordinating effect of DMF with Pt ions, which
will favorably bridge the deposition potential gap between Pt and Co, and thus being
anticipated to promote formation of PtCo alloys. It is observed that from Fig. 2.7d
that when the potential is scanned to that more negative than −0.9 V (vs. SCE),
the cathodic reduction current sharply increases due to an occurrence of hydrogen
evolution reaction (HER). For the anodic scan, the hydrogen dissolution reaction
occurs first at −0.85 V (vs. SCE), and the anodic wave between −0.7 and 0 V (vs.
SCE) corresponds to the Co oxidation. The satellite peak at −0.45 V (vs. SCE) is
attributed to the overlay of Co oxidation peak and Pt reduction peak as shown in
Fig. 2.7b, c. It is noted that the anodic peak potentials at −0.25 V (vs. SCE) for the
PtCo alloy catalysts obtained in the presence of DMF shift to more positive potentials
than those obtained in the absence of DMF, implying a stabilizing effect of DMF on
PtCo alloys.
Figure 2.8 shows the potentiostatic current–time transients during electrode-
positing PtCo alloy catalysts at different potentials in the absence or presence of
DMF. By comparison, it is identified that the corresponding nucleation and growth
mechanism of PtCo alloy electrodepositions here is in accordance with the reported
model [59–61]. Especially, similar to Simonov’s work [61], the deposition prior to the
current maximum corresponds to the primary nucleation on the GC electrode, which
is under kinetic control as shown in region I. After a current maximum, the secondary
nucleation on newly deposited crystals will appear and become dominating in region
II. As the deposition time is prolonged to region III, the nucleation and growth of
secondary crystals will continue under diffusion control. By comparison, it is found
that accompanied with the decrease in the deposition potential from −0.5 to −0.9 V

Fig. 2.8 Potentiostatic current–time transients for the electrodeposition of PtCo [58]. a In the
absence of DMF; b in the presence of DMF
40 2 The Electrocatalysis of Oxygen Reduction on Platinum …

(vs. SCE), the current maximum becomes more intense and higher, demonstrating a
faster nucleation process. It is also noted that there exists an obvious delay for the
emergence of deposition current in the presence of DMF when the deposition poten-
tials are −0.5 V, −0.6 V and −0.7 V, respectively, indicating that a slow nucleation
process under kinetic control occurs at lower overpotentials due to the coordination
between PtCl4 2− ions and DMF.
Figure 2.9 shows SEM images of the as-obtained PtCo alloy catalysts electrode-
posited on GC electrode surfaces from different potentials of −0.5 V, −0.6 V, −0.7 V,
−0.8 V, −0.9 V (vs. SCE) in the presence of DMF. As is observed, all the PtCo alloy
catalysts present aggregated granulates with coarse surfaces, and the more irregu-
larities appear gradually along with the decrease in the deposition potential. When
the deposition potential is −0.9 V (vs. SCE), special lamellar structures are formed,

Fig. 2.9 SEM images of the PtCo electrodeposited in the presence of DMF [58]. a −0.5 V; b −
0.6 V; c −0.7 V; d −0.8 V; e −0.9 V
2.5 Pt-Based Alloys 41

Fig. 2.10 High resolution SEM images of PtCo granulates deposited at −0.8 V [58]. a Before CV
activation; b after CV activation

and this might be attributed to the concomitant HER, in which large number of H2
bubbles generates, thus disturbing and blocking a uniform deposition. The atomic
compositions for the series of PtCo alloy catalysts obtained at different potentials
and in the presence of DMF before/after the CV activation are determined by EDX.
It is observed that along with the decrease of deposition potential from −0.5 to −
0.8 V (vs. SCE), the Pt/Co atomic ratio of the PtCo alloy catalyst increases, and the
irregular atomic composition (Pt43 Co57 ) for the PtCo alloy catalyst obtained at −
0.9 V (vs. SCE) can be attributed to the different diffusion rates of PtCl4 2− and Co2+
ions under limiting diffusion current and the disturbance of HER.
The corresponding high resolution SEM images of PtCo granulates before and
after the CV activation are presented in Fig. 2.10a, b, respectively. It is thus concluded
that after the CV activation, the surfaces of PtCo granulates become coarser owing
to the dissolution of large amount of Co elements. The meticulous structures of the
PtCo granulates electrodeposited at the potential of −0.6 and −0.8 V (vs. SCE) were
carefully investigated by high resolution TEM and are shown in Fig. 2.11a, b and
c, d, respectively. It is seen that the PtCo granulates are mainly composed of PtCo
nanospheres by packing crystallites, and the size of the PtCo nanospheres ranges
from 95 to 200 nm. It is also noted that compared with that with the deposition
potential of −0.6 V (vs. SCE), the PtCo nanospheres with the deposition potential of
−0.8 V (vs. SCE) have larger particle size and looser packing structures, which may
be due to the fact that a larger amount of Co is dissolved for the PtCo nanospheres
with the deposition potential at −0.8 V (vs. SCE) than that at −0.6 V (vs. SCE). It
is anticipated that the looser packing structures will lead to desired larger ECSA.
In addition, the impregnating of the series of Pt–Co alloys into HClO4 and
performing an CV activation resulted in an intensive dissolution of Co elements, thus
generating not only loose packing structures but also highly porous Pt-rich surfaces.
The synergetic effect from the above-mentioned aspects must lead to an optimal Pt–
Pt bond length for the Pt4 Co sample, thus demonstrating a remarkable ORR-specific
activity of 1.52 mA/cm2 at 0.9 V, 7 times of that for the commercial Pt/C catalyst.
It is anticipated this work indeed points out a facile strategy for efficient preparation
42 2 The Electrocatalysis of Oxygen Reduction on Platinum …

Fig. 2.11 High resolution TEM images of the PtCo nanospheres after CV activation [58]. a, b −
0.6 V, c, d −0.8 V

of highly active Pt-transition metal alloys without adding multitudinous or long-

chain organic surfactants, and besides, high-temperature annealing is also avoided.
In our future researches, the underlying mechanism on interactions between different
metal ions and coordinating agents will be systematically and detailedly explored,
thus leading to preparation of more types of highly active Pt-transition metal alloy
catalysts toward the ORR.

2.5.4 Crystallographic Orientation Effects

Crystallographic orientation advantages are also found in Pt-based catalysts and play
an extraordinary role to increasing the ORR activity. Different from Pt single crystal
surfaces, the order of ORR activity of crystal surfaces with different orientation
change after the transition metals introduced, for example, the activity order of Pt3 Ni
low-index surface is as follows: {100} < {110} < {111}. In the development of the
Pt-based catalysts, considerable efforts have been devoted to the synthesis of Pt-based
bimetallics with special facets exposed to tailor the activity. Those polyhedron-design
3d geometric configurations we discussed in Fig. 2.3 are successfully achieved in Pt-
based alloys. Among those crystal orientation, {111} facets with the lowest surface
energy are easy to obtain through various synthetic ways and play an extraordinary
role to increasing the ORR activity [3–12]. Cui et al. [5] reported the synthesis of
2.5 Pt-Based Alloys 43

octahedral PtNi nanoparticles (NPs) through a self-surfactant method, in which the

precursor ligands of acetyl acetonate determined the formation of octahedral shape
rather than the interaction of the solvent N, N-Dimethylformamide (DMF) with {111}
facets. The as-obtained PtNi octahedral with a size of 9.5 nm demonstrated a mass
activity of 1.45 A/mgPt . Wu et al. [12] generated Pt-based icosahedral nanocrystals
with surfaces enclosed by {111} facets through a gas reducing agent in liquid solution
(GRAILS) method which used both CO gas and capping agents for shape controlling.
Among the as-obtained Pt-M (M = Au, Ni, Pd) alloy, Pt3 Ni icosahedral nanocrystals
present an impressive mass-specific activity of 0.62 A/mgPt . Zhang et al. [10] in situ
produced carbon-supported octahedral Pt–Ni alloys with an improved mass activity
of 1.96 A/mgPt by reducing metal acetylacetonates impregnated on carbon support
under both CO and H2 gases.
It is worth to mention that under the strong acidic operation condition, the less
noble solute metals are unstable, and the surface layers of Pt-based catalysts are
almost composed of pure Pt. The formed Pt layers can provide kinetic stability
against the dissolution of the solute. This inspire us to synthesis Pt-based catalysts
initially coated with Pt shells, known as core–shell structure. Since the chemical–
electrical process is an interfacial reaction, the top of few catalysts surface layers
mainly contribute to the intrinsic activity of active sites, and therefore, core–shell
structure can maximize the utilization of expensive Pt. With the combination of both
advantages from component effects (the ligand effect and the strain effect brought up
by solute atoms) and utilization of Pt, core–shell structures especially show superior
electrocatalytic performance for the ORR.
Accordingly, three strategies have been proposed: (i) alloying Pt with another
transition metal (e.g., Ni, Co, Cu), (ii) synthesizing shape-controlled nanocrys-
tallines enclosed by active Pt facets, especially {111} facets and (iii) preparing core–
shell nanostructures (NCs) with few-atomic-layer Pt shells; each of which has been
demonstrated to be effective in enhancing the specific activity.

2.5.5 Icosahedral Pt–Ni Nanocrystalline Electrocatalyst

Very recently, our groups combined these three strategies to prepare the shape-
controlled, core–shell structured Pt alloy, and the as-obtained electrocatalysts, not
unexpectedly, exhibit remarkable specific activity as a result of the synergistic effects.
We firstly synthesize a sandwich-structured, icosahedral Pt2.1 Ni catalyst enclosed by
{111} facets using a hot-injection method, which shows a superior ORR activity of
0.91 mA/cm2 and 0.32 A/mgPt @ 0.9 V versus the reversible hydrogen electrode
that is 4.1 and 2.5 times higher relative to those of the commercial Pt/C, respec-
tively. Despite the successful preparation of such novel electrocatalysts, the growth
mechanism and the consequent influence of the nanostructure on the electrocatalytic
performance remain poorly understood, which makes catalyst preparation a trial-
and-error process without a general guidance. Subsequently, the growth process of
the sandwich-structured, icosahedral Pt2.1 Ni catalyst is monitored by sampling the
44 2 The Electrocatalysis of Oxygen Reduction on Platinum …

alloy nanocrystal at different time periods after hot injection and characterizing its
element-composition and structure, which clarifies the growth mechanism.
Firstly, highly uniform icosahedral Pt–Ni NCs with well-defined size and shape
were synthesized. As shown in Fig. 2.12a, the as-synthesized NCs show an obvious
narrower size distribution, and the histograms of the particle diameter distribution
indicate that the average size of the synthesized NCs is around 9.45 nm, demon-
strating that hot injection has a significant effect on particle uniformity. Scanning
transmission electron microscopy (STEM) and high resolution (HR)-TEM images

Fig. 2.12 Morphology analysis and XRD patterns of Pt–Ni/C sample [62]
2.5 Pt-Based Alloys 45

in Fig. 2.12c, d present the uniform morphology and crystal structure of the as-
synthesized NCs. The regular hexagon 2D image and the clear twin boundaries in
Fig. 2.12d prove the icosahedral morphology enclosed by {111} as confirmed by the
interplanar spacing of ≈0.227 nm. The inset in Fig. 2.12d shows the model illustra-
tion of the atomic arrangement and shape of a representative Pt–Ni icosahedron. The
selected area electron diffraction (SAED) and normalized XRD patterns in Fig. 2.12e,
f also support this argument that the as-synthesized icosahedral NCs are enclosed by
{111} facets. Compared to the commercial Pt/C catalyst, the hot-injection particles
possess more active {111} facets on surface and better crystallinity according to the
relative intensity of XRD peaks, and that was one of the main factors contributing
to its higher catalytic activity. The XRD diffraction peaks show a small right shift
relative to the standard Pt peaks, indicating a narrower lattice spacing that is caused
by the substitution of Pt atoms (r = 1.82 Å) with Ni atoms (r = 1.24 Å) as well as
the strain in this twinned crystal structure.
It can also be observed in Fig. 2.12c that there exists obvious contrast difference at
different region of particles, which is ascribed to the difference in the microstructure
and element distribution. Inductively coupled plasma-optical emission spectrometer
(ICP-OES) test was performed for the quantitative measurement of the bulk elemental
composition of the Pt–Ni NCs, while XPS was also carried out to analyze the surface
Pt–Ni atomic ratio. The Pt to Ni atom ratio determined by ICP was 2.1:1, much lower
than that from XPS (4.8:1), confirming the formation of Pt-enriched surfaces, which
is consistent with the STEM images described in the following section.
To illuminate the growth mechanism of this icosahedral Pt2.1 Ni nanocrystal, the
intermediate products at 3, 10, 15 and 30 min after injection were sampled from
the reaction system to be characterized by STEM and EDX mapping as shown in
Fig. 2.13 During the catalyst preparation process, the solvent immediately changed
from translucent yellow to black once the Pt and Ni precursors were injected, indi-
cating that a burst nucleation occurred, which is consistent with previous theoretical
simulation and experimental research [63]. Meanwhile, considering the higher redox
potential of Pt2+ /Pt than that of Ni2+ /Ni (−0.25 V vs. SHE) for Ni/Ni2+ vs. +1.18 vs.
SHE for Pt/Pt2+ ), the first step for catalyst formation should be the burst nucleation
of Pt [64]. After the burst nucleation, the existence of plenty of Pt nucleii decreased
the Ni potential barrier of nucleation and triggered the heterogeneous nucleation
of Ni along with continued Pt reduction, forming a Ni-enriched intermediate layer,
which is clearly verified by the sample at 3 and 10 min (Fig. 2.13a–h) in which the
Ni-covered Pt seeds can be observed [65]. The corresponding STEM line-scanning
profile along the centerline of a single particle at 10 min was further illustrated in
Fig. 2.14a. It is clear that the Pt intensity is several multiples of Ni in the center posi-
tion, but the Ni intensity becomes stronger than that of Pt near the edge. Moreover,
the Pt descends gradually along the growth direction, while the Ni increases, which
indicates the existence of a Ni-enriched layer.
Owing to the higher Pt content in the precursor, the remaining Pt cations are
gradually reduced to form a Pt-enriched shell. This process is slower compared to
the preceding Pt and Ni nucleation owing to the sluggish chemical reaction kinetics
of Pt reduction and deposition as well as the high cation transport resistance in an
46 2 The Electrocatalysis of Oxygen Reduction on Platinum …

Fig. 2.13 EDX element maps of representative intermediate products at different time after
injection [62]

Fig. 2.14 STEM line-scanning fitting profiles at different time after rejection along the line as
indicated in the insets [62]
2.5 Pt-Based Alloys 47

oleylamine (OAm, 70%) system, as demonstrated by the samples at 15 and 30 min. In

conclusion, the Pt2.1 Ni icosahedron formation involves three steps: (1) burst nucle-
ation of Pt atoms to form a Pt-enriched core, (2) heterogeneous nucleation of Ni
atoms onto the Pt core to form a Ni-enriched interlayer and (3) formation of a Pt-
enriched surface with {111} facets enclosed through the reduction of residual Pt
cations.
The STEM image and the corresponding EDX mappings of the sample at 15
and 30 min (Fig. 2.13i–p) depict the element distribution of a single particle, which
clearly demonstrates a sandwich structure, consisting of a Pt-enriched core, Ni-
enriched interlayer and Pt-enriched shell. As presented in Fig. 2.14b (30 min after
injection), the line-scanning cave in the center and the two symmetrical peaks of
Ni content, as well as the Pt-fitting curve, offer direct evidences of the sandwich
structure. Moreover, the nanostructure of the sample at 60 min remains the same size
and shape, which supports that the Pt cations in the reaction system were exhausted
after the reaction proceeded for 30 min.
To further verify the proposed growth mechanism, Pt1.5 Ni (molar ratio determined
by ICP-OES) was obtained by reducing the amount of Pt(acac)2 in precursor by
25%, which is the quantity required by the Pt shell as calculated by its thickness.
The STEM mapping results of the final Pt1.5 Ni and the intermediate products of
Pt2.1 Ni at 10 min after injection are given in Fig. 2.15. Clearly, both of these two
samples own a Pt@Ni structure, assuring that the Pt shell of Pt2.1 Ni is generated by
the mild reduction of excess Pt cations. Therefore, burst nucleation, reaction kinetic-
controlled growth process and CO and other mediating agents worked jointly for
the formation of {111} enclosed and composition-graded Pt2.1 Ni icosahedra. For
more intuitive understanding, a schematic illustration for the growth pathway of the
as-synthesized Pt2.1 Ni NCs is proposed and presented in Fig. 2.16.

Fig. 2.15 STEM images and the corresponding EDX element maps at 10 min after injection [62].
a–d The final products Pt1.5 Ni, e–h the intermediate products Pt2.1 Ni
48 2 The Electrocatalysis of Oxygen Reduction on Platinum …

Fig. 2.16 Illustration for the growth pathway and the final structure of the sandwich-structured
Pt2.1 Ni NCs [62]

The superior specific activity and stability highly depend on the unique sandwich
structure of Pt2.1 Ni icosahedral NCs. Firstly, the Pt-enriched shell offers more active
sites via the exposure of {111} facets, as well as separates the Ni layer from acidic
environment, and thus retards the dissolution of Ni atoms. The Ni-enriched interlayer,
located beneath the Pt-enriched shell, increases the utilization of Pt atoms, and more
importantly, induces the electronic modification of the outermost Pt shell. The Pt-
enriched core also plays a vital role on the stable sandwich structure as it can protect
the nanostructure from collapse even if the excessive dissolution of Ni atoms happens
within the interlayer after longtime cycling. Moreover, conventional Ni@Pt core–
shell nanocrystals encounter a notable internal strain accumulation caused by the
lattice mismatch between Pt and Ni (≈11% between Pt and Ni), which weakens the
structure stability. By contrast, the Pt–Ni–Pt-layered sandwich structure mitigates
the strain accumulation as the Pt core acts as a buffer layer, which enables a better
resilience and structural stability.

2.5.6 Stability

Stabilizing Pt-based nanoparticles in the cathode of PEMFCs seem to be even more

challenging than controlling their activity. Here, we only discuss the dissolution of
nanocatalysts itself, while the corrosion of the carbon support at high potentials
will also influence the durability of catalysts. The constant dissolution of solute
metals under the strong acidic condition could result in the destruction of optimal
component and therefore lead to the negative ligand effect and strain effect. The
design strategy focus on obtaining those bimetallic materials that having Pt shell as
the topmost surface layer and metal alloys that possessing low tendency to segregate.
Evidenced from our studies that the Pt2.1 Ni icosahedral NCs with sandwich structure
presents a remarkable catalytic stability after 15,000 cycles, After 3000, 10,000 and
15,000 cycles, the mass-specific activities still remain 99.7%, 88.0% and 71.1%
benchmarked with the original value for the Pt2.1 Ni/C, while only 75.9%, 54.8%
and 36.1% for that of commercial Pt/C, respectively. Some Pt-based bimetallics
are predicted to be both electrochemically active and stable for the ORR, such as Pt
2.5 Pt-Based Alloys 49

alloyed with Y and Sc. Pt-Y bimetallic catalysts prepared using a sputtering technique
experimentally achieved increasing activity and stability [66, 67]. Introducing Mo
into octahedral Pt–Ni alloys can also result in enhanced activity and stability, proved
both theoretically and experimentally. This also shed light on the multicomponent
alloys [68]. Moreover, some inertial metals such as Au are believed to stabilize Pt
particles.
Regardless the influence from element component, the stability of Pt-based alloys
also depends on their surface structures. The defective structure which is widely
accepted as one kind of high active structure in Pt-based alloys including evident
curving edges and surface depressions could lead to the exposure of many unstable
low-coordination sites, but the stable particles used in PEMFCs usually do not have
a high percentage of edge, corner and step sites. In fact, the defective particles
applied in cathode will soon loose their active features in shape and become sphere.
What is more, the heterogeneous structures usually introduced by chemical reduction
methods are also thermodynamically unstable because of the presence of defects
on grain surfaces and boundaries which will increase the internal energy. Thermal
annealing is a useful method for enhancing Pt-based alloys’ catalystic performance
as it can improve the alloy degree and accurately control the surface composition
and morphology by adjusting the degree of surface segregation of transition metals.
Beermann et al. [69] investigated the impact of annealing temperature on the
composition and shape of Pt–Ni based NPs and through annealing Pt–Ni based NPs
with Ni-rich {111} facets under 4% hydrogen at 300 °C, the NPs with Ni located more
on the inside were observed to present both enhanced reactivity (2.7 A/mgPt ) and
durability. Chen et al. [70] synthesized hollow Pt3 Ni frames consisting of 24 edges
of parent rhombic dodecahedron and the formation of Pt-skin-terminated {111}-
like surface structure led to an extraordinary high mass activity of 5.7 A/mgPt . The
Pt3 Ni nanoframes were obtained via eroding segregated NiO in PtNi3 polyhedrons in
nonpolar solvents, and the oxygen was employed in this method for the oxidation and
segregation of Ni from PtNi3 polyhedrons. Wang et al. [71] reported the achievement
of ordered Pt3 Co NPs with 2–3 atomic-layer-thick Pt shells through transferring
disordered Pt–Co alloy under H2 /N2 mixed gas atmosphere at 700 °C which presented
a 2-folds mass-specific activity compared with disordered Pt3 Co alloys as well as
commercial Pt/C. Surface-segregated Co/Co oxides were also observed during the
preparation. Moreover, literatures also pointed out that the Pt surface diffusion during
thermal annealing could tailor the surface composition, thus forming Pt-rich shells
[72, 73]. It is thus believed that thermal annealing is an effective after-treatment
strategy for achieving ideal surface elemental distribution and morphology in Pt-
M alloys. Nonetheless, surface impurities, surface segregation of the solute metal,
sintering, and degradation of the carbon support all complicate the annealing process
on nanoparticles.
50 2 The Electrocatalysis of Oxygen Reduction on Platinum …

2.5.7 Thermal Annealing Synthesis of Double-Shell

Truncated Octahedral Pt–Ni Alloys

In this regard, herein, our group first synthesized heterogeneous quasi-octahedral

Ni-rich Pt-based alloys through a one-step method. Then, the alloys were annealed
in H2 atmosphere to form a core-double-shells structure without changing their octa-
hedral morphology owing to the presence of NiO outer shell. It was proved that the
as-obtained truncated octahedral PtNi3.5 alloys were well-shaped and consisted of
homogeneous Pt–Ni alloy cores enclosed by NiO–Pt double shells. The removal of
NiO outer shell could be considered as a novel dealloying method without morpho-
logical changes in Ni-rich Pt-based alloys. It is also presumed that the formation of
NiO-Pt double shell was the result of the oxides surface segregation driven by oxides
concentration gradient. Both the formation of Pt-segregated shell exposed by acid
treatment and the high alloy degree obtained during the H2 annealing process lead to
the changes of surface composition, morphology as well as Pt electronic structures,
which strongly point toward enhancements in ORR activity and durability. After an
acid treatment, the ORR activity of the annealed truncated octahedral Pt2.1 Ni elec-
trocatalyst was 2.91 mA/cm2 and 1.24 A/mgPt in area- and mass-specific activity,
which are 4.2 and 3.4 times that of unannealed defective quasi-octahedral Pt3 Ni.
The PtNi3.5 NPs were in situ deposited on carbon powders in DMF, and the
product is labeled “as-prepared PtNi3.5 /C”. Figure 2.17a-1, a-2 shows both TEM
and HR-TEM images of as-prepared PtNi3.5 /C. It is observed that most as-prepared
PtNi3.5 NPs are composed of defective quasi-octahedrons with an average edge

Fig. 2.17 Morphology and elemental distribution analysis [74]. a As-prepared PtNi3.5 /C;
b annealed5% H2 PtNi3.5 /C
2.5 Pt-Based Alloys 51

length of 10.28 nm. The defective structure including evident curving edges and
surface depressions could result in the loss of highly active {111} facets while
lead to the exposure of many low-coordination sites. To gain an insight into deli-
cate composition profiles, HAADF-STEM was also employed to characterize the
as-prepared PtNi3.5 NPs. Figure 2.17a-5 depicts the HAADF-STEM image, and
their corresponding elemental mappings are presented in Fig. 2.17a-6–a-8. The
two-dimensional mappings of Pt and Ni elements in representative as-prepared
PtNi3.5 NPs are characterized with identical variations, demonstrating a simultaneous
distribution of Pt and Ni in the NPs.
Figure 2.17b-1 presents the morphology of the sample resulting from thermally
annealing as-prepared PtNi3.5 /C under H2 (5 vol% in Ar) at 300 °C for 90 min. A
pure Ar atmosphere was also employed to treat the as-prepared PtNi3.5 /C for compar-
ison. The above-annealed electrocatalysts are labeled as “annealedgas PtNi3.5 ” with
“gas” corresponding to the annealing atmosphere. It is observed that although the
alloy NPs both exhibit distinct structure changes after being annealed under either
5 vol% H2 in Ar or pure Ar, most annealed NPs with a size of 4 nm or larger in
edge length remain enclosed with eight {111} facets. The average edge length is
9.46 nm for annealed5% H2 PtNi3.5 whereas 8.95 nm for annealedAr PtNi3.5 . To figure
out the changes of surface structure, both the annealed5% H2 and annealedAr PtNi3.5
NPs were further analyzed under HR-TEM. The structure defects disappear after
thermal anneal, and flat {100} and {111} facets are formed instead. A thermody-
namically driven effect for the morphological restructure is confirmed as identical
structure formed after thermal anneal regardless of annealing gaseous atmosphere.
It is believed that the unsteady low-coordination edge atoms have a priority to trans-
form in order to lower the surface energy, and the edge atoms diffusing toward
{111} facets could fill the depression on surfaces, thus resulting in more focused
bulks and smooth {111} facets. The structure described as truncated octahedron is
energetically favorable with the formation of {100} facets, as predicted by Wang
et al. [75] The element distribution of Pt and Ni inside annealed5% H2 PtNi3.5 NPs
was determined by STEM-EDX elemental face-scanning. The elemental maps of
projected Pt and Ni are recorded in Fig. 2.17b-5–b-8, which implies the formation
of Pt segregation as well as a uniform distribution of Ni in annealed5% H2 PtNi3.5
NPs. On the contrary, no segregation of Pt or Ni is observed in annealedAr PtNi3.5
NPs. It is therefore concluded that the elemental distribution change is driven by a
gaseous atmosphere effect. Apart from the well-shaped truncated octahedrons, some
small particles are also visible in Fig. 2.17b-5. After being annealed under the H2
environment, they remain spheroid in morphology and monodisperse while enclosed
with Ni-rich shells.
The delicate chemical composition at the atomic scale was further observed under
STEM. Figure 2.17a-3, a-4, b-3 and b-4 show the atomic-resolution HAADF-STEM
images viewed along [110] zone axis of individual as-prepared and annealed5% H2
PtNi3.5 NPs. The structure defects can be further confirmed in as-prepared PtNi3.5
NPs with rounded vertices and loss of {111} facets (Fig. 2.17a-3, a-4). Usually, those
concave-like surface defects could act as catalytic sites with unique activities due
to their unique local structural environments [76]. However, it is also noted that the
52 2 The Electrocatalysis of Oxygen Reduction on Platinum …

as-prepared PtNi3.5 crystals can be either homogeneous or heterogeneous as the pres-

ence of anisotropic crystal orientations (region inside the yellow line). Compared
with homogeneous crystals, those heterogeneous crystals are thermodynamically
unstable because of the presence of defects on grain surfaces and boundaries which
will increase the internal energy, thus leading to the fact that the anisotropic quasi-
octahedrons are easily to be destroyed under the cathode operating condition. From
Fig. 2.17b-3, b-4, the morphological restructure under H2 (5 vol% in Ar) atmosphere
results in the formation of atomically flat {111} and {100} facets. It is noted that
the HAADF-STEM technique has a high compositional sensitivity when there are
equal thickness effects, and therefore, the segregation of Pt or Ni could be distin-
guished through Z-contrast. The uniform contrast in the center of atomic-resolution
images confirms a relative high alloy degree inside annealed5% H2 PtNi3.5 truncated
octahedrons. It is also observed that the bimetallic cores are enclosed with double
shells, evidenced by the outer two layers with weaker Z-contrast enriched by Ni and
the following four layers with stronger Z-contrast enriched by Pt. It is thus believed
that the annealed5% H2 PtNi3.5 truncated octahedron possesses a homogeneous Pt–Ni
alloy core surrounded by two shells, one is Ni-segregated outer shell, and the other
is Pt-segregated inner shell.
It is noted that the transition metal Ni-segregated outer shells of the annealed5% H2
PtNi3.5 truncated octahedrons are unstable and will dissolve during electrochemical
measurements or single cell tests. Then, the as-dissolved Ni ions can block the gas
diffusion layer pores as well as change the conductivity of membrane. Therefore,
an acid pickling was applied to the annealed5% H2 PtNi3.5 /C before electrochemical
measurements. The same acid treatment was also conducted on as-prepared PtNi3.5 /C
for comparison. After removing the surface oxide and unstable heterogeneous struc-
tures, the ICP results point out that the acid treatment for as-prepared PtNi3.5 /C
and annealed5% H2 PtNi3.5 /C lead to the formation of Pt3 Ni/C (labeled “acid-treated
Pt3 Ni/C”) and Pt2.1 Ni/C (labeled “acid-treated annealed5% H2 Pt2.1 Ni/C”). Since the
annealed5% H2 PtNi3.5 /C truncated octahedrons own Pt-segregated inner shells, the
Ni elements in the bimetallic core become more difficult to dissolve.
The elemental analyses based on XPS were performed to determine the valence
state of near-surface atoms. As shown in Fig. 2.18a–d, XPS spectra from the Ni 2p
core level regions of as-prepared PtNi3.5 , acid-treated Pt3 Ni, annealed5% H2 PtNi3.5
and acid-treated annealed5% H2 Pt2.1 Ni were fitted with four doublets corresponding
to Ni0 , Ni2+ and their satellite peaks. It is seen that the Ni element on surfaces of
as-prepared PtNi3.5 is mainly in the form of Ni2+ , which is 2 times of the amount
of Ni0 . Combining with the O 1s spectra, it can be confirmed that a large amount
of NiO was formed in as-prepared PtNi3.5 NPs during the hydrothermal process. It
is also observed that after H2 annealing, the proportion of NiO slightly decreases,
which implies that NiO has been marginally reduced during the H2 annealing process.
Undoubtedly, after the acid treatment, the majority of NiO has been removed in acid-
treated annealed5% H2 Pt2.1 Ni/C near-surface regions while there exists only a slight
decrease in NiO/Ni ratio in acid-treated Pt3 Ni/C. It can be thus convinced that a segre-
gation of NiO toward surfaces and the formed Ni-rich outer shells of annealed5% H2
Pt2.1 Ni NPs are mainly composed of NiO. It is also important to mention that the
2.5 Pt-Based Alloys 53

Fig. 2.18 Ni 2p XPS spectra (black) and the fitting results (red) [74]

migration and agglomeration of Pt–Ni alloy NPs on carbon powders is simultane-

ously accompanied with internal structural deformation, which will therefore result in
the loss of electrochemical surface area (ECSA) for ORR. It has been well acknowl-
edged that the formation of oxide shells such as iron oxides [77], MgO [78] and
NiO [79] can successfully prevent the sintering of NPs. It is thus believed that the
NiO-segregated outer shells on annealed5% H2 PtNi3.5 /C play an important role in
mitigating conglobation during internal structural deformation.
To further identify the formation and effect of formed NiO-segregated outer
shells during the H2 anneal process, the acid-treated Pt3 Ni/C was annealed under
the same conditions with that for the as-prepared PtNi3.5 /C (labeled “annealed5% H2
acid-treated Pt3 Ni/C”). The TEM images of acid-treated Pt3 Ni/C, acid-treated
annealed5% H2 Pt2.1 Ni/C and annealed5% H2 acid-treated Pt3 Ni/C are presented in
Fig. 2.19a–c, and the HR-TEM image of acid-treated annealed5% H2 Pt2.1 Ni/C is
presented in Fig. 2.19d. As shown in Fig. 2.19a, d, the morphology of the acid-
treated Pt3 Ni and acid-treated annealed5% H2 Pt2.1 Ni NPs is monodisperse and roughly
coincides with that of as-prepared PtNi3.5 and annealed5% H2 PtNi3.5 NPs, but the
average edge length of acid-treated Pt3 Ni NPs obviously declines to 5.88 nm after
the acid treatment. The TEM image of acid-treated annealed5% H2 Pt2.1 Ni/C is shown
in Fig. 2.19b and it is found that there only exists a slight decrease in the average edge
length (8.87 nm). This comparison illustrates the negative effect from Ni dissolution
without the protection of Pt-segregated shells as well as thermodynamically unstable
heterogeneous structures, thus leading to a serious destruction in morphology. The
Ni 2p XPS spectra of annealed5% H2 acid-treated Pt3 Ni confirm a promotion in the
ratio of NiO under a reducing environment, which indicates a common phenomenon
54 2 The Electrocatalysis of Oxygen Reduction on Platinum …

Fig. 2.19 TEM and HR-TEM images [74]. a Acid-treated Pt3 Ni/C, b, d acid-treated annealed5% H2
Pt2.1 Ni/C, c annealed5% H2 acid-treated Pt3 Ni/C

of NiO segregation for the thermal anneal under a H2 gaseous environment, irre-
spective to the proportion of Pt and Ni in the NPs. According to Fig. 2.19c, with
the removal of Ni element, the annealed5% H2 acid-treated Pt3 Ni/C NPs agglomerate
severely owing to an insufficient NiO protection during H2 annealing.
Figure 2.20a–e present the Pt 4f XPS spectra of as-prepared PtNi3.5 , acid-treated
Pt3 Ni, annealed5% H2 PtNi3.5 , acid-treated annealed5% H2 Pt2.1 Ni and annealed5% H2
acid-treated Pt3 Ni. The Pt 4f7/2 and Pt 4f5/2 intense doublet was then deconvoluted
to the corresponding doublets of different valence states, and the as-prepared PtNi3.5
nanoparticles with Pt partially oxidized are initially composed of Pt0 , Pt2+ and Pt4+ .
Based on XPS results, the Pt0 4f7/2 peak of as-prepared PtNi3.5 NPs appears at
71.1 eV, which is similar to that of pure Pt and may indicate their poor alloy degree.
After being annealed, the Pt0 4f7/2 peak shifts positively to around 71.72 eV, demon-
strating a change of outer electronic structures which decreases the d-band center
of Pt atoms. Accompanied with the downshift of d-band center, the adsorption of
oxygenated intermediated products on Pt decreases, resulting in an enhanced ORR
activity. A positive shift of 0.47 eV for acid-treated Pt3 Ni relative to as-prepared
PtNi3.5 mainly results from the fact that after the acid treatment, erratic grains are
2.5 Pt-Based Alloys 55

Fig. 2.20 Pt 4f XPS spectra (black) and the fitting results (red) [74]

removed and stable uniform alloys are exposed. However, the acid treatment still has
limit effect on Pt outer electronic structures compared with thermal annealing.
In contrast to bare reduction of Ni2+ , a certain amount of highly charged Pt element
has been reduced by H2 , thus resulting in the increase in the amount of Pt0 in PtNi3.5
after being annealed under H2 . This is because Ni has a much higher affinity for
oxygen as compared to Pt (NiOx formation energy of −2.54 eV per O versus PtO,
PtO2 formation energies of −0.41 and −0.63 eV per O) [80], and the reduction of
PtOx may result in both the H2 reduction effects and the replacement of Pt in PtOx with
Ni atoms [81]. Owing to the formation of Ni-segregated outer shells and the higher
oxygen affinity of Ni, the segregated Pt at inner shells mainly composed of stable Pt0 ,
evident by the same percentage of Pt0 preserved in the acid-treated annealed5% H2
Pt2.1 Ni after acid pickling. Interestingly, it is found that the proportion of Pt0 in acid-
treated Pt3 Ni/C decreases after being annealed, indicating an outward segregation
of PtOx . Combined with the Ni 2p XPS results, an outward segregation of oxides
accompanied with the reduction of highly charged Pt and Ni during H2 anneal process
is confirmed. Menning and Chen [82] predicted a thermodynamically stable 3d metal-
Pt–Pt surface structure in O2 , and it is speculated that this 3d metal-Pt–Pt surface
structure may also stable in the presence of O atoms near surface, thus forming a NiO–
Pt double shells structure. It is also noted that compared with as-prepared PtNi3.5 ,
the outward segregation of both NiO and PtOx becomes conspicuous in acid-treated
Pt3 Ni during the H2 thermal annealing. We presume that the outward segregation of
oxides in as-prepared PtNi3.5 is driven by the decrease of oxides concentration on
surface owing to the reduction by H2 . The removal of oxides on acid-treated Pt3 Ni
surface during the acid pickling results in the initial decrease of oxides concentration,
56 2 The Electrocatalysis of Oxygen Reduction on Platinum …

therefore the tendency of outward segregation of oxides increases, and the added
oxides segregation could cover up the effect from H2 reduction.
Figure 2.21 reports XRD patterns measured from as-prepared PtNi3.5 , acid-treated
Pt3 Ni, annealed5% H2 PtNi3.5 , acid-treated annealed5% H2 Pt2.1 Ni and annealedAr
PtNi3.5 as well as standard Pt, Ni and NiO crystals (reference: Pt PDF no. 04-0802, Ni
PDF no. 04-0850 and NiO PDF no. 47-1049). There exist two separate peaks for NiO
at 2θ = 37.2° and 79.4° in the as-prepared PtNi3.5 . After the subsequent thermal or
acid treatment, the patterns for NiO phase disappear. The reflections of as-prepared
PtNi3.5 are broad and asymmetric, indicating the formation of a heterogeneous struc-
ture. The removal of erratic gains and Ni-rich phase in as-prepared PtNi3.5 through the
acid pickling results in more symmetric reflections in the XRD pattern of acid-treated
Pt3 Ni. Considering the thermal annealing under different gaseous environments,
both high alloy degrees were attained, evidenced by a decrease of the full width at
half maximum (FWHM) in the {111} reflections of both annealed5% H2 PtNi3.5 and
annealedAr PtNi3.5 . Compared with annealedAr PtNi3.5 , the annealed5% H2 PtNi3.5
attains a higher alloy degree indicating an intense oxide and metal interdiffusion
under the H2 environment. It is noted that the formed double-shell structure causes
the asymmetry of XRD peaks as measured from annealed5% H2 PtNi3.5 compared
with more symmetric peaks from solid solution Pt–Ni alloy by thermal annealing
under pure Ar environment. Moreover, owing to the interdiffusion, homogeneous
Pt–Ni alloy cores are formed for the acid-treated annealed5% H2 Pt2.1 Ni samples,
which give rise to the decline of crystalline interplanar space. The lattice contraction
indicates the change in lattice strain, thus influencing the electronic structure, as
convinced by XPS results.

Fig. 2.21 XRD patterns of Pt–Ni alloys [74]

2.5 Pt-Based Alloys 57

Owing to the evaluation of Pt content on the surface of homogeneous Pt–Ni alloy

prevent Ni erosion while the thermal annealing removed the unstable structures,
the mass-specific activities of acid-treated annealed5% H2 Pt2.1 Ni/C remain 80.2%,
57.8%, 51.0%, 49.4% and 45.4% after 3000, 5000, 12,000, 20,000 and 30,000 cycles
ADT, observed a great enhancement comparing with acid-treated Pt3 Ni/C (mass-
specific activity loss by 60.9% after 5000 cycles durability test).

2.6 Outlook

The widespread use of platinum in MEAs hinders the development of PEMFCs.

Therefore, the reduction of the amount of Pt in the cathode is essential to the commer-
cialization of PEMFCs. Over the past years, Pt alloy nanocatalysts with low Pt
loadings have attracted people’s focus. Great progress has been made on dealloyed
Pt-based alloy catalysts from the aspects of shape, size and the ORR catalytic perfor-
mance by the method of a rotating disk electrode in ideal electrochemical solution.
Unfortunately, when it comes to MEAs under practice operation, the application of
Pt alloy catalysts with low Pt loadings still faces many challenges.
For practical fuel cells, however, much high local oxygen partial pressures at the
active surface sites of the catalysts result in at least three orders of magnitude limiting
currents larger than in RDE. Thus, the RDE tests are unable to entirely predict prac-
tical fuel cell catalysts. For the catalyst itself, advanced shape-controlled bimetallic
Pt nanocatalysts with carefully designed morphology, surface facet structure and
surface compositions turned out hard to be realized in electrode layers of single cell
MEAs to date. The nanoparticles shape designed for lowering Pt utilization remained
unstable, partially because of the leaching of metal atoms. What’s more, mass trans-
port including proton through the ionomer and oxygen through the micropores and
the ionomer to the surface of the active sites remain difficulties for low Pt-loaded
electrode layers.
Therefore, a combination of material design at the atomistic scale, layer compo-
sition tuning and fuel cell test engineering is necessary for the realization of high
power density in low Pt-loaded layers using advanced Pt alloys. As is metioned
before, by means of tuning the electronic structure and exposed facets, advanced Pt
alloy nanocatalysts are hoped to perform impressive catalytic Pt mass-based activi-
ties. And the alloys of Pt and the rare earth metal element catalysts engaged in MEAs
under high operation currents and temperatures combined with the larger number of
potential sweep cycles may be developed by the controlling of the intrinsic strain, the
optimization of coordination numbers of the surface active sites and the arrangements
of geometric topography.
To make sure sufficient access of protons and oxygen to the catalytic sites of the
Pt particles within the micropores in the cathode catalyst layer, a carefully designed
structure between active sites and ionomer is essential. This can be achieved by a
partial penetration of the ionomer into these pores, which, in turn, can be realized by
pore size tuning or ionomer length modifications. Besides, to achieve a more uniform
58 2 The Electrocatalysis of Oxygen Reduction on Platinum …

distribution of the ionomer, the possible chemically modified carbon supports and
the ratio of ionomer and carbon should also be optimized.
Last but not the least, the technologies which bridge the RDE and MEA tests
should be developed. To improve the mass transport and preserve the facile applica-
tion of RDE, a floating electrode technique (FET) can be put forward for ultralow
loading Pt catalyst. Using the FET, large ORR current densities are obtained, because
the ionomer or solution film is thin enough (nano scale) to enable excellent O2 mass
transfer. Then, half-cell GDE measurements under actual PEMFCs operating condi-
tions offer more accurate trends in fuel cell catalyst activity. Besides, real-time obser-
vation and visualization including in situ TEM, in situ X-ray absorption spectroscopy
and in situ X-ray scattering can provide a guide on the nanoscale processes relevant
for an improvement of Pt alloy catalysts with low Pt-loaded.

References

1. Gasteiger HA, Kocha SS, Sompalli B et al (2005) Activity benchmarks and requirements for Pt,
Pt-alloy, and non-Pt oxygen reduction catalysts for PEMFCs. Appl Catal B Environ 56:9–35
2. Peng ZM, Yang H (2009) Designer platinum nanoparticles: control of shape composition in
alloy nanostructure and electrocatalytic property. Nano Today 4:143–164
3. Choi S, Xie S, Shao M et al (2013) Synthesis and characterization of 9 nm Pt–Ni octahedra with
a record high activity of 3.3 A/mgPt for the oxygen reduction reaction. Nano Lett 13:3420–3425
4. Chou S, Lai Y, Yang YY et al (2014) Uniform size and composition tuning of PtNi octahedra
for systematic studies of oxygen reduction reactions. J Catal 309:343–350
5. Cui C, Gan L, Li H et al (2012) Octahedral PtNi nanoparticle catalysts: exceptional oxygen
reduction activity by tuning the alloy particle surface composition. Nano Lett 12:5885–5889
6. Gan L, Heggen M, Rudi S et al (2012) Core-shell compositional fine structures of dealloyed
Pt(x)Ni(1–x) nanoparticles and their impact on oxygen reduction catalysis. Nano Lett 12:5423–
5430
7. Park J, Liu J, Peng H et al (2016) Coating Pt–Ni octahedra with ultrathin Pt shells to enhance
the durability without compromising the activity toward oxygen reduction. Chemsuschem
9:2209–2215
8. Wu J, Yang H (2011) Synthesis and electrocatalytic oxygen reduction properties of truncated
octahedral Pt3Ni nanoparticles. Nano Res 4:72–82
9. Wu J, Zhang J, Peng Z et al (2010) Truncated octahedral Pt3Ni oxygen reduction reaction
electrocatalysts. J Am Chem Soc 132:4984–4985
10. Zhang C, Hwang SY, Trout A et al (2014) Solid-state chemistry-enabled scalable produc-
tion of octahedral Pt–Ni alloy electrocatalyst for oxygen reduction reaction. J Am Chem Soc
136:7805–7808
11. Zhang J, Yang H, Fang J et al (2010) Synthesis and oxygen reduction activity of shape-controlled
Pt3Ni nanopolyhedra. Nano Lett 10:638–644
12. Wu J, Qi L, You H et al (2012) Icosahedral platinum alloy nanocrystals with enhanced
electrocatalytic activities. J Am Chem Soc 134:11880–11883
13. Stamenkovic VR, Fowler B, Mun BS et al (2007) Improved oxygen reduction activity on
Pt3Ni(111) via increased surface site availability. Science 315:493–497
14. Chen M, He Y, Spendelow JS et al (2019) Atomically dispersed metal catalysts for oxygen
reduction. ACS Energy Lett 4:1619–1633
15. Zhang H, Chung HT, Cullen DA et al (2019) High-performance fuel cell cathodes exclusively
containing atomically dispersed iron active sites. Energy Environ Sci 12:2548–2558
References 59

16. Zhang H, Hwang S, Wang M et al (2018) Single atomic iron catalysts for oxygen reduction in
acidic media: particle size control and thermal activation. J Am Chem Soc 139:14143–14149
17. Papageorgopoulos D (2018) Fuel cells R&D overview. https://www.hydrogen.energy.gov/pdfs/
review18/fc01_papageorgopoulos_2018_o.pdf
18. Darling RM, Meyers JP (2003) Kinetic model of platinum dissolution in PEMFCs. J
Electrochem Soc 150:1523–1527
19. Norskov JK, Rossmeisl J, Logadottir A et al (2004) Origin of the overpotential for oxygen
reduction at a fuel-cell cathode. J Phys Chem B 108:17886–17892
20. Nilekar AU, Mavrikakis M (2008) Improved oxygen reduction reactivity of platinum
monolayers on transition metal surfaces. Surf Sci 602:89–94
21. Vang RT, Honkala K, Dahl S et al (2005) Controlling the catalytic bond-breaking selectivity
of Ni surfaces by step blocking. Nat Mater 4:160–162
22. Kandoi S, Greeley J, Sanchez-Castillo MA et al (2006) Prediction of experimental methanol
decomposition rates on platinum from first principles. Top Catal 37:17–28
23. Rossmeisl J, Logadottir A, Nørskov JK (2005) Electrolysis of water on (oxidized) metal
surfaces. Chem Phys 319:178–184
24. Greeley J, Rossmeisl J, Hellman A et al (2007) Theoretical trends in particle size effects for
the oxygen reduction reaction. Z Phys Chem 221:1209–1220
25. Hansen HA (2009) PhD thesis. Technical University of Denmark
26. Kitchin JR, Nørskov JK, Barteau MA et al (2004) Modification of the surface electronic and
chemical properties of Pt(111) by subsurface 3d transition metals. J Chem Phys 120:10240–
10246
27. Perez-Alonso FJ, McCarthy DN, Nierhoff A et al (2012) The effect of size on the oxygen
electro-reduction activity of mass-selected platinum nanoparticles. Angew Chem 51:4641–
4643
28. Nesselberger M, Ashton S, Meier JC et al (2011) The particle size effect on the oxygen reduction
reaction activity of Pt catalysts: influence of electrolyte and relation to single crystal models.
J Am Chem Soc 133:17428–17433
29. Yang Z, Ball S, Condit D et al (2011) Systematic study on the impact of Pt particle size and
operating conditions on PEMFC cathode catalyst durability. J Electrochem Soc 158:1439–1445
30. Maillard F, Pronkin S, Savinova ER (2010) Handbook of fuel cells. Wiley, Hoboken
31. Wang YJ, Long W, Wang L et al (2018) Unlocking the door to highly active ORR catalysts
for PEMFC applications: polyhedron-engineered Pt-based nanocrystals. Energy Environ Sci
11:258–275
32. Narayanan R, El-Sayed MA (2005) Catalysis with transition metal nanoparticles in colloidal
solution: nanoparticle shape dependence and stability. J Phys Chem B 109:12663–12676
33. Lai SCS, Lazenby RA, Kirkman PM et al (2015) Nucleation aggregative growth and detachment
of metal nanoparticles during electrodeposition at electrode surfaces. Chem Sci 6:1126–1138
34. Zhang J, Yan SG, Wagner FT (2009) Method of treating nanoparticles using a proton exchange
membrane and liquid electrolyte cell. US Patent 0145781
35. Lapp AS, Duan Z, Marcella N et al (2018) Experimental and theoretical structural investigation
of AuPt nanoparticles synthesized using a direct electrochemical method. J Am Chem Soc
140:6249–6259
36. Luo L, Zhu F, Tian R et al (2017) Composition-graded PdxNi1−x nanospheres with Pt mono-
layer shells as high-performance electrocatalysts for oxygen reduction reaction. ACS Catal
7:5420
37. Simka W, Puszczyk D, Nawrat G (2009) Electrodeposition of metals from non-aqueous
solutions. Electrochim Acta 54:5307–5319
38. Katayama Y, Endo T, Miura T et al (2013) Electrodeposition of gold in an amide-type ionic
liquid. J Electrochem Soc 161:87–91
39. Kumbhar PP, Lokhande CD (1995) Electrodeposition of yttrium from a nonaqueous bath. Met
Finish 93:28–31
40. Shen S, Li F, Zhao L et al (2018) Communication—an organic solvent system-assisted
electrodeposition of highly active Pt for the oxygen reduction reaction. J Electrochem Soc
165:3392–3394
60 2 The Electrocatalysis of Oxygen Reduction on Platinum …

41. Hansen HA, Viswanathan V, Nørskov JK (2014) Unifying kinetic and thermodynamic analysis
of 2 e– and 4 e– reduction of oxygen on metal surfaces. J Phys Chem C 118:6706–6718
42. Mayrhofer KJJ, Blizanac BB, Arenz M et al (2005) The impact of geometric and surface
electronic properties of Pt-catalysts on the particle size effect in electrocatalysis. J Phys Chem
B 109:14433–14440
43. Friebel D, Viswanathan V, Miller DJ et al (2012) Balance of nanostructure and bimetallic
interactions in Pt model fuel cell catalysts: in situ XAS and DFT study. J Am Chem Soc
134:9664–9671
44. Kulkarni A, Siahrostami S, Patel A et al (2018) Understanding catalytic activity trends in the
oxygen reduction reaction. Chem Rev 118:2302–2312
45. Mavrikakis M, Hammer B, Nørskov JK (1998) Effect of strain on the reactivity of metal
surfaces. Phys Rev Lett 81:2819–2822
46. Lischka M, Mosch C, Gross A (2007) Tuning catalytic properties of bimetallic surfaces: oxygen
adsorption on pseudomorphic Pt/Ru overlayers. Electrochim Acta 52:2219–2228
47. Hoster HE, Alves OB, Koper MTM (2010) Tuning adsorption via strain and vertical ligand
effects. ChemPhysChem 11:1518–1524
48. Strasser P, Koh S, Anniyev T et al (2010) Lattice-strain control of the activity in dealloyed
core-shell fuel cell catalysts. Nat Chem 2:454–460
49. Knudsen J, Nilekar AU, Vang RT et al (2007) A Cu/Pt near-surface alloy for water–gas shift
catalysis. J Am Chem Soc 129:6485–6490
50. Satyapal S (2008) DOE hydrogen and fuel cells program record. https://interface.ecsdl.org/con
tent/24/2/45.abstract
51. Choi DS, Robertson AW, Warner JH et al (2016) Low-temperature chemical vapor deposition
synthesis of Pt–Co alloyed nanoparticles with enhanced oxygen reduction reaction catalysis.
Adv Mater 28:7115–7122
52. Chen S, Ferreira PJ, Sheng W et al (2008) Enhanced activity for oxygen reduction reaction
on “Pt3 Co” nanoparticles: direct evidence of percolated and sandwich-segregation structures.
J Am Chem Soc 130:13818–13819
53. Wang C, Dennis VDV, Chang KC et al (2009) Monodisperse Pt3Co nanoparticles as a catalyst
for the oxygen reduction reaction: size-dependent activity. J Phys Chem C 113:19365–19368
54. Chen S, Sheng W, Yabuuchi N et al (2009) Origin of oxygen reduction reaction activity on
“Pt3Co” nanoparticles: atomically resolved chemical compositions and structures. J Phys Chem
C 113:1109–1125
55. Jia Q, Liang W, Bates MK et al (2015) Activity descriptor identification for oxygen reduction on
platinum-based bimetallic nanoparticles: in situ observation of the linear composition–strain–
activity relationship. ACS Nano 9:387–400
56. Choi SI, Lee SU, Kim WY et al (2012) Composition-controlled PtCo alloy nanocubes with
tuned electrocatalytic activity for oxygen reduction. ACS Appl Mater Inter 4:6228–6234
57. Sorsa O, Romar H, Lassi U et al (2017) Co-electrodeposited mesoporous PtM (M=Co Ni Cu)
as an active catalyst for oxygen reduction reaction in a polymer electrolyte membrane fuel cell.
Electrochim Acta 230:49–57
58. Shen S, Li F, Luo L et al (2018) DMF-coordination assisted electrodeposition of highly active
PtCo alloy catalysts for the oxygen reduction reaction. J Electrochem Soc 165:43–49
59. Scharifker B (1983) Theoretical and experimental studies of multiple nucleation. Electrochim
Acta 28:879–889
60. Sherstyuk OV, Pronkin SN, Chuvilin AL et al (2000) Platinum electrodeposits on glassy carbon:
the formation mechanism morphology and adsorption properties. Russ J Electrochem 36:741–
751
61. Simonov AN, Cherstiouk OV, Vassiliev SY et al (2014) Potentiostatic electrodeposition of Pt
on GC and on HOPG at low loadings: analysis of the deposition transients and the structure of
Pt deposits. Electrochim Acta 150:279–289
62. Tian T, Shen S, Zhu F et al (2018) Icosahedral Pt–Ni nanocrystalline electrocatalyst: growth
mechanism and oxygen reduction activity. Chemsuschem 11:1015–1019
References 61

63. Zhou W, Wu J, Yang H (2013) Highly uniform platinum icosahedra made by hot injection-
assisted GRAILS method. Nano Lett 13:2870–2874
64. Stamenkovic VR, Mun BS, Mayrhofer KJJ et al (2006) Effect of surface composition on
electronic structure stability and electrocatalytic properties of Pt-transition metal alloys: Pt-skin
versus Pt-skeleton surfaces. J Am Chem Soc 128:8813–8819
65. Turnbull D (1950) Kinetics of heterogeneous nucleation. J Chem Phys 18:198–203
66. Mayrhofer KJJ, Arenz M (2009) Log on for new catalysts. Nat Chem 1:518–519
67. Hwang SJ, Kim SK, Lee JG et al (2012) Role of electronic perturbation in stability and activity
of Pt-based alloy nanocatalysts for oxygen reduction. J Am Chem Soc 134:19508–19511
68. Huang XQ, Zhao ZP, Cao L (2015) High-performance transition metal-doped Pt3Ni octahedra
for oxygen reduction reaction. Science 348:1230–1234
69. Beermann V, Gocyla M, Kuhl S et al (2017) Tuning the electrocatalytic oxygen reduction
reaction activity and stability of shape-controlled Pt–Ni nanoparticles by thermal annealing-
elucidating the surface atomic structural and compositional changes. J Am Chem Soc
139:16536–16547
70. Chen C, Kang Y, Huo Z et al (2014) Highly crystalline multimetallic nanoframes with three-
dimensional electrocatalytic surfaces. Science 343:1339–1343
71. Wang D, Xin HL, Hovden R et al (2013) Structurally ordered intermetallic platinum–
cobalt core–shell nanoparticles with enhanced activity and stability as oxygen reduction
electrocatalysts. Nat Mater 12:81–87
72. Gocyla M, Kuehl S, Shviro M et al (2018) Shape stability of octahedral PtNi nanocatalysts for
electrochemical oxygen reduction reaction studied by in situ transmission electron microscopy.
ACS Nano 12:5306–5311
73. Gan L, Heggen M, Cui C et al (2016) Thermal facet healing of concave octahedral Pt–Ni
nanoparticles imaged in situ at the atomic scale: implications for the rational synthesis of
durable high-performance ORR electrocatalysts. ACS Catal 6:692–695
74. Luo X, Guo Y, Zhou H et al. Thermal annealing synthesis of double-shell truncated octahedral
Pt–Ni alloys for the oxygen reduction reaction of polymer electrolyte membrane fuel cells.
Front Energy (accepted)
75. Wang G, Van Hove MA, Ross PN et al (2005) Monte Carlo simulations of segregation in Pt–Ni
catalyst nanoparticles. J Chem Phys 122:5410
76. Callevallejo F, Pohl MD, Reinisch D et al (2017) Why conclusions from platinum model
surfaces do not necessarily lead to enhanced nanoparticle catalysts for the oxygen reduction
reaction. Chem Sci 8:2283–2289
77. Liu C, Wu X, Klemmer T et al (2005) Reduction of sintering during annealing of FePt
nanoparticles coated with iron oxide. Chem Mater 17:620–625
78. Gao M, Li A, Zhang J et al (2014) Fabrication and magnetic properties of FePt nanoparticle
assemblies embedded in MgO-matrix systems. J Sol-Gel Sci Technol 71:283–290
79. Zeynali H, Sebt SA, Arabi H et al (2012) Synthesis and characterization of FePt/NiO core–shell
nanoparticles. J Inorg Organomet P 22:1314–1319
80. Li W, Osterlund L, Vestergaard EK et al (2004) Oxidation of Pt(110). Phys Rev Lett 93:146104
81. Ahmadi M, Behafarid F, Cui C et al (2013) Long-range segregation phenomena in shape-
selected bimetallic nanoparticles: chemical state effects. ACS Nano 7:9195–9204
82. Menning CA, Chen JG (2010) Regenerating Pt–3d–Pt model electrocatalysts through oxida-
tion–reduction cycles monitored at atmospheric pressure. J Power Sources 195:3140–3144
Chapter 3
Pt-MS Electrocatalysts for ORR

Abstract Fuel cells are expected to be one of the major clean and sustainable energy
sources in the near future. However, the sluggish kinetics of ORR and the high Pt
loading in the cathode are the urgent issues to be addressed, since they determine
the efficiency and the cost of fuel cells. In this chapter, an approach is developed for
designing the electrocatalyst for ORR in fuel cells. These electrocatalysts consist of
only Pt monolayer or mixed transition-metal-Pt monolayer shells (MS) on suitable
carbon-supported metal or alloy nanoparticles. The synthesis involves depositing
a Cu shell on a suitable transition metal or metal alloy surface at underpotentials,
followed by the galvanic replacement of the Cu MS with Pt or mixed metal-Pt. It
is found that the electronic properties of Pt MS can be fine-tuned by electronic and
geometric effects introduced by the substrate metal (or alloy) and lateral effects of the
neighboring metal atoms. The role of substrates is found reflected in a “volcano” plot
of the MS activity for ORR as a function of their calculated d-band centers. The Pt
mass-specific activity of the Pt-MS electrocatalysts is up to twenty times higher than
state-of-the-art commercial Pt/C. The enhancement of the activity is caused mainly
by the decreased formation of Pt–OH (the blocking species for ORR), and to a lesser
degree by the electronic effects. Fuel cell tests show a very good long-term stability
of the electrocatalysts. Results demonstrate a viable way in designing the electrocat-
alysts which can successfully alleviate two issues regarding commercialization of
fuel cells: the costs of electrocatalysts and their efficiency.

Keywords Pt monolayer shell · Underpotential deposition · Substrate effect ·

Oxygen reduction reaction

3.1 Introduction

The burgeoning global energy demands have caused two major challenges: (1) the
rapid consumption of the traditional fossil energy leads to the rising prices and
exhausting resources of fossil fuels; (2) the emission pollutants from burning fossil

© Shanghai Jiao Tong University Press and Springer-Verlag GmbH Germany 2021 63
J. Zhang and S. Shen, Low Platinum Fuel Cell Technologies,
Energy and Environment Research in China,
https://doi.org/10.1007/978-3-662-56070-9_3
64 3 Pt-MS Electrocatalysts for ORR

fuels result in the severe environmental degradations [1–6]. Such energy and envi-
ronmental crises have greatly affected the economic, ecological and social develop-
ments of human beings, and thus, tremendous research efforts have been devoted to
the renewable energy technologies possessing high efficiencies and low emissions
[1–7]. As also discussed in Chaps. 1 and 2, among various types, proton exchange
membrane fuel cells (PEMFCs) have attracted tremendous attention as the promising
alternatives to the traditional fossil-fuel-based energy systems in the fields of portable,
transportation and residential applications, due to their reproducible fuel resources,
mild operating conditions and low emissions [1, 2, 4, 6, 8, 9]. Although greatly
developed, the commercialization of PEMFCs is still stagnated by many obstacles,
the critical one is that the oxygen reduction reaction (ORR) kinetics under natural
conditions is very sluggish, resulting in a large overpotential loss, which needs to be
significantly improved by electrocatalysts to meet the commercial requirements for
PEMFCs [4–6, 9].
In the past three decades, intensive research has been conducted on designing and
preparing highly active and durable ORR electrocatalysts [1, 5–7, 9–12]. Despite
the great potentials of non-noble metal (NM) electrocatalysts, Pt-group-metal nano-
materials are still the most practical and effective electrocatalysts in PEMFCs [6, 7,
11]. And among them, Pt is the most efficient single polycrystalline metal for ORR,
which in turn has greatly hindered the large-scale commercialization of PEMFCs,
due to its high cost (~45 $ g−1 ) and scarce reserve (37 p.p.b. in the earth crust),
which accounts for over 55% of the total cost [1, 3, 4, 6, 7, 9, 11, 13, 14]. There-
fore, plenty of strategies have been explored for improving the ORR electrocatalytic
performance of Pt-based nanomaterials and eventually reducing the NM loading in
the practical PEMFCs, for instance: ➀ optimizing the size and distribution on the
support [e.g., the optimal size for Pt nanoparticles (NPs) is ~3 nm]; ➁ alloying with
foreign elements (such as B, Co, Cu and Au); ➂ controlling their exposed facets
(i.e., morphologies and high-index facets, such as cube, octahedron, icosahedron
and dendrite); ➃ fabricating the special structures [such as aerogel, core@shell,
monolayer shell (MS) and sandwich]; ➄ utilizing the appropriate supports (such as
carbon black, graphene, metal oxide and metal nitride) [4–6, 10, 15–43]. Although
excellent progresses have been achieved, it is still challenging to design and synthe-
size high-performance Pt-based ORR electrocatalysts simultaneously possessing the
remarkable electrocatalytic activity, NM utilization and electrochemical durability
[4]. It should also be noted that, as the intrinsic ORR electrocatalytic activity [i.e.,
area-specific activity (ASA)] increases, the electrocatalyst loading in PEMFCs can
be continuously reduced. However, it will also lead to the decrease in electrochem-
ically active surface area (ECSA), and thus inevitably resulting in a new problem
of the increase in local O2 transport resistance, which can severely compromise the
performance of PEMFCs especially at high current densities [44].
Therefore, to develop the practical Pt-based ORR electrocatalysts for PEMFCs,
both the ORR electrocatalytic performance and the mass transfer problem should be
taken into consideration, and the key lies in reducing the NM loading without the
ORR electrocatalytic performance and ECSA losses. In this regard, concentrating
the active NM atoms in the electrocatalyst surface seems to be the most effective
3.1 Introduction 65

approach, which makes the well-defined core@shell structures with the non-NM or
low-cost core and the NM or high-cost shell very promising, especially the ones
possessing the MS structures [4, 17, 19, 20, 22, 24, 26, 28, 32–34, 41, 42, 45].

3.2 Cu UPD Coupled with Pt2+ Galvanic Replacement

Underpotential deposition (UPD) is an electrochemical phenomenon in which a metal

MS is deposited on a foreign metal substrate at the potential that is more positive than
its thermodynamic equilibrium one [46–50]. Such electrochemical phenomenon is
attributed to that the interaction between the deposited metal M1 and the substrate
metal M2 is greater than the one between the metal M1 and itself. UPD is not only
limited to the first MS (or sub-MS) deposition, but can also occur when the substrate
still affects the second or third MS [47–49]. Hevesy et al. firstly report the UPD
phenomenon of some radioactive elements on copper electrodes in 1912 [50, 51]. As
the research methods and theories of UPD develop, the potential applications of UPD
have also attracted enormous attention. So far, the technologies involving UPD have
already been applied in the fields of metal corrosion and protection, electrocatalysis,
semiconductor electrosynthesis, electrochemical analysis, surface modification of
electrodes and so on [4, 50, 52–68].
In order to prepare the ORR electrocatalysts with Pt-MS structures, Adzic et al.
have developed a versatile and effective electrochemical technology, which involves
the Cu UPD and Pt2+ galvanic replacement, as demonstrated in Fig. 3.1.
Taking a pure Pd electrode (i.e., substrate) as an example, the electrochem-
ical processes involving the Cu UPD and Pt2+ galvanic replacement are generally
conducted as follows: (1) The Pd electrode is first subjected to a pre-treatment, mainly
to polish and clean the electrode surface; (2) the Pd electrode is then placed into an
Ar-saturated 50 mM H2 SO4 and 50 mM CuSO4 solution, in which a steady cyclic
voltammogram (CV) of the Cu UPD is obtained at 20 mV s−1 , to determine the end
potential of the backward linear sweep voltammogram (LSV) of the Cu UPD on the
Pd substrate; (3) subsequently, a forward LSV to 0.90 V [vs. reversible hydrogen
electrode (RHE)] is performed to completely remove the UPD Cu atoms on the
electrode surface, and a backward LSV to the end potential is performed to acquire
a Cu MS on the Pd substrate; (4) immediately, the above-modified Pd electrode is
transferred into an Ar-saturated 50 mM H2 SO4 and 1.0 mM K2 PtCl4 solution and
kept for 2 min to completely replace Cu with Pt2+ [4, 36, 37, 69]. In this way, the Pd
electrode modified with a Pt MS is prepared.
66 3 Pt-MS Electrocatalysts for ORR

Fig. 3.1 Schematic diagram of the electrochemical processes involving the Cu UPD and Pt2+
galvanic replacement on a Pd substrate

3.3 Pt MSs on Single-Crystal Metal Electrodes for ORR

3.3.1 Effects of Substrates on Pt MSs for ORR

In order to confirm the idea that the Pt-MS electrocatalysts are very promising to
alleviate the problems of the high Pt usage and low electrocatalytic activity for
ORR, single-crystal metal surfaces are subjected to the Cu UPD and Pt2+ galvanic
replacement. The electrocatalytic properties of the bimetallic surfaces consisting of
metal MSs on single-crystal metal surfaces have been extensively studied in ultrahigh
vacuum systems, but to a lesser extent in electrochemical ones. In many cases, the
formation of a surface metal–metal bond significantly changes the electronic proper-
ties of the metal overlayer, and pronounced differences are observed in the reactivity
of some transition metal MSs on various substrates. The selection of the substrates
and the determination of what electronic and structural properties that the substrates
should have are the key points. The study involves the determination of the ORR
electrocatalytic activity for the Pt MSs on Ru(0001), Ir(111), Rh(111), Pd(111) and
Au(111) surfaces in 0.1 M HClO4 solutions.
The result of the Pt-MS deposition on Rh(111) surface is a good example that
illustrates the work with all types of single-crystal electrodes. Figure 3.2a shows the
typical CV with two peaks for the Cu UPD on a Rh(111) surface. The scan rate is
20 mV s−1 . The charge associated with these peaks is 540 µC cm−2 , which is close
to 524 µC cm−2 , needed for depositing a pseudomorphic Cu MS on an ideal Rh(111)
3.3 Pt MSs on Single-Crystal Metal Electrodes for ORR 67

Fig. 3.2 CVs for a Rh(111) surface and the Pt MS on a Rh(111) surface [55, 66]. a CVs for
a Rh(111) surface in N2 -saturated 50 mM H2 SO4 solutions with 50 mM Cu2+ (solid line) and
without Cu2+ (dash line). b CVs for the Pt MS on a Rh(111) surface (solid line) and for a bare
Rh(111) surface (dash line) in N2 -saturated 50 mM H2 SO4 solutions

surface. The most negative peak is very close to the bulk deposition of Cu, which
demands careful control of the Cu UPD deposition to avoid the deposition of more
than a monolayer of Cu. This is critical since the amount of deposited Pt corresponds
to the UPD Cu coverage.
The dash line shows the CV for a Rh(111) surface in the absence of Cu2+ in the
solution. Figure 3.2b depicts the CVs for the Pt MS on a Rh(111) surface (solid
line) and for a bare Rh(111) surface without the Pt MS (dash line). The scan rate
is 20 mV s−1 . As expected, the deposition of the Pt MS partially blocks the oxide
formation on Rh(111) since Pt is oxidized at more positive potentials.
Figure 3.3 is the scanning tunneling microscopy (STM) image of a Pt MS
deposited on a Rh(111) surface obtained at 0.8 V in a 0.1 M HClO4 solution. The
image size is 47 × 47 nm, and the Z range is 3 nm. The deposition consists of
two-dimensional interconnected Pt islands. With the small difference between the Pt
and Rh lattice constants, these islands are expected to be epitaxial with the Rh(111)
substrate. There are a certain number of holes between the islands and also a few
sites with Pt atoms in the second layer. Similar depositions are also obtained with
other substrates.
Figure 3.4a shows the rotating disk measurements for the ORR of the Pt MSs
on five different single-crystal surfaces, and for comparison, on a Pt(111) surface
taken from the reference [70]. The ring potential is 1.1 V, and the collection effi-
ciency is 0.20. The most active surface is PtML /Pd(111), and the least active one is
PtML /Ru(0001). Figure 3.4b shows the oxidation at the ring electrode of the H2 O2
generated by the ORR at the disk, which is measured by keeping its potential at
1.1 V where the reaction is under diffusion control. Less active electrodes generate
larger amounts of H2 O2 . However, the ORR on the five Pt surfaces all involves the
four-electron reaction pathway, including the least active Pt/Ru(0001), where 6.8%
H2 O2 is generated at the peak current. Therefore, the overall reaction can be written
as follows.
68 3 Pt-MS Electrocatalysts for ORR

Fig. 3.3 An in situ STM image of a Pt MS on a Rh(111) surface obtained at 0.8 V in a 0.1 M
HClO4 solution [55]

Fig. 3.4 LSVs for the Pt MSs on different surfaces [66, 67]. a LSVs for the Pt MSs on Ru(0001),
Ir(111), Rh(111), Au(111) and Pd(111) surfaces in N2 -saturated 0.1 M HClO4 solutions. b The
current for the H2 O2 oxidation on a ring electrode (I ring ) as a function of the disk potential

O2 + 4H+ + 4e− = 2H2 O (3.1)

The analyses involving the Koutecky–Levich (K–L) plots and Tafel slopes indicate
that the reaction is a first order in terms of the concentration of the dissolved O2 , and
that the first charge transfer step is the rate-determining step (RDS). These results are
in agreement with those observed on a pure Pt surface. Accordingly, the comparison
of the ORR rates on these surfaces is a reliable and valid procedure.
3.3 Pt MSs on Single-Crystal Metal Electrodes for ORR 69

3.3.2 Volcano-Type Relation Between Electrocatalytic

Activity and d-Band Center for ORR on Pt MS

As presented in Fig. 3.5, the ORR kinetic currents at 0.8 V for the Pt MSs on
various substrates as a function of the associated calculated d-band centers (εd )
show a volcano-type relation, in which PtML /Pd(111) possesses the maximum elec-
trocatalytic activity. The use of the kinetic current instead of the exchange current
density in the plot is preferred, because there is no reliable method for determining
the exchange current density for ORR. Other authors have used the same approach
[71]. This increase in ORR kinetics on PtML /Pd(111) compared to that on Pt(111) is
surprising, since Pt(111) and Pt(110) are the most active low-index Pt crystal planes
for ORR known in HClO4 solutions.
In addition, the finding of this volcano-type relation is novel in the fields of
both electrocatalysis and catalysis in general. For a number of catalytic reactions, it
has been shown that the activity of catalysts exhibits volcano-type dependences on
some variables, and the best of the catalyst strikes a balance between two competing
influences [72–74].
To ascertain the competing influences in the present case, it is necessary to have
an understanding of the ORR mechanism, which, however, is still be the subject of
extensive discussion [75–77]. Evidence suggests that the first charge transfer step,
or O2 adsorption with simultaneous charge and proton transfer, appears to be the
RDS on pure Pt. The exchange of a second electron with the addition of another
proton forms two OH species or hydrogen peroxide (H2 O2 ). H2 O2 can escape into the
solution phase, which terminates the reaction in a two-electron process. An additional

Fig. 3.5 ORR kinetic currents (jk : square symbols) at 0.8 V for the Pt MSs on different single-
crystal surfaces in 0.1 M HClO4 solutions as a function of calculated d-band centers (εd − εf :
relative to the Fermi level) of the respective clean Pt MSs [66, 67]
70 3 Pt-MS Electrocatalysts for ORR

transfer of two electrons and two protons in reactions with OH or H2 O2 completes

the four-electron reduction of O2 to H2 O.
Despite the lack of consensus in regard to this mechanism, the four-electron reac-
tion pathway for ORR must involve both the breaking of the O–O bond (whether in
O2 , O2 H, or H2 O2 ) and the formation of an O–H bond. A more reactive surface, such
as the one characterized by a higher-lying εd tends to bind adsorbates more strongly,
thereby enhancing the kinetics of the dissociation reactions producing these adsor-
bates. On the other hand, a surface with a lower-lying εd , tends to bind adsorbates
more weakly and facilitates the formation of the bonds among them. Thus, it may
be expected that the most active Pt MS should have a εd with an intermediate value.
Density functional theory (DFT) studies have shown that compressive strain tends
to downshift εd in energy, whereas tensile strain has the opposite effect [78]. The Pt
MSs on Ru(0001), Rh(111) and Ir(111) are compressed compared to Pt(111), whereas
PtML /Au(111) is stretched by more than 4%. Indeed, calculated εd is accordingly
either lower or higher in energy than that of Pt(111) (Fig. 3.5) [67]. The position
of the εd for the Pt MS does not correlate strictly with the amount of strain present
because the position of the εd for the Pt MS depends on both the strain (geometric
effect) and the electronic interaction between the Pt MS and its substrate (ligand
effect) [79].
PtML /Au(111) lies at the high εd end of the plot and binds oxygen much more
strongly than Pt(111), whereas PtML /Ru(0001), PtML /Rh(111) and PtML /(Ir(111) lie
at the other end and bind oxygen considerably weakly than Pt(111). If Pt(111) is taken
to be close to the optimum balance as described above, then the position to of the εd
would suggest that PtML /Ru(0001), PtML /Ir(111) and PtML /Rh(111) are less active for
ORR than Pt(111), because breaking the O–O bond is more difficult on these surfaces
than Pt(111). This situation would facilitate the H2 O2 formation at the expense of
the complete reduction of oxygen to water. On the other hand, PtML /Au(111) may
suffer from the hindered oxygen atom or OH hydrogenation kinetics, owing to the
stronger binding of oxygen atoms or oxygen-containing fragments, and is therefore
less active than Pt(111) as well. Slow O or OH hydrogenation rates may also cause
surface O or OH coverage to build up and block the O2 adsorption, dissociation or
hydrogenation sites. PtML /Pd(111) is most similar to pure Pt(111) and should possess
greater activity than the other Pt MSs. Indeed, the ORR kinetics on PtML /Pd(111)
is measured to be higher even than that on Pt(111). In addition, the production of
H2 O2 is higher on PtML /Ru(0001) and PtML /Ir(111) than on the rest of the Pt MSs
(Fig. 3.4b). Both of these experimental findings are in line with the prediction, which
are based on the premise that the optimal ORR electrocatalyst should strike a balance
between the O–O bond breaking and the O–H bond-making activity.
Experimentally, it is also found that the superior ORR electrocatalytic activity of
PtML /Pd(111) is associated with a reduced OH coverage [37]. Furthermore, CV and
X-ray absorption near edge structure (XANES) measurements have identified more
positive potentials of the OH adsorption on PtML /Pd(111) than those on Pt(111),
which can be attributed to an enhanced repulsive interaction of OH groups on
PtML /Pd(111) compared to on pure Pt(111). Pd–OH, on the other hand, forms at
3.3 Pt MSs on Single-Crystal Metal Electrodes for ORR 71

less positive potentials. More details about the ORR on PtML /Pd will be discussed in
the following sections.

3.3.3 Summary

In summary, the ORR electrocatalytic activity of the Pt MSs on Au(111), Ir(111),

Pd(111), Rh(111) and Ru(0001) surfaces show a volcano-type dependence on the d-
band center of the Pt MSs. The results elucidate for the first time that, for a good ORR
electrocatalyst, the kinetics of both the O–O bond breaking and the hydrogenation
of reactive intermediates have to be facile at the cathode. More importantly and from
a practical perspective, it is possible to devise an ORR electrocatalyst that contains
only a fractional amount of Pt but can surpass the electrocatalytic activity of pure
Pt [e.g., the Pt MS on Pd(111)]. Therefore, Pt MSs on appropriate metal substrates
represent a viable way to reduce Pt loadings and thus the associated cost of fuel
cell electrodes. The results also point to the important possibility of fine-tuning the
electrocatalytic activity of transition metals.

3.4 Pt MSs on Pd(111) and Carbon-Supported Pd NPs

for ORR

In the previous section, the existence of a volcano-type relation between the ORR
electrocatalytic activity and the d-band center (εd ) of the Pt MS is demonstrated
with PtML /Pd(111) lying on the top of the curve, indicating the best ORR electro-
catalytic activity. Further, experimental studies and DFT calculations reveal that the
PtML /Pd(111) system strikes an optimal balance between the O–O bond breaking
and the O (or OH) species hydrogenation. In this section, more detailed kinetics
of ORR is studied in acid solutions on PtML /Pd(111) and on carbon-supported Pd
nanoparticles (NPs) using the rotating ring disk electrode technique. The kinetics
of ORR shows a significant enhancement for a Pt MS on Pd(111) and on Pd NP
surfaces in comparison with the reaction on Pt(111) and on Pt NPs. Electrochemical
and XANES measurements show that these enhancements in ORR electrocatalytic
activity are partially due to the decreased formation of Pt–OH. An enhanced atomic
scale roughness and low coordination of some atoms indicated by the STM measure-
ments may also contribute to the observed improved ORR electrocatalytic activity.
The Pt mass-specific activity (MSA) of the PtML /Pd/C is 5–8 times higher than that
of the Pt/C. The noble-metal (NM) MSA is two times higher than that of Pt/C. The
long-term electrochemical durability of the PtML /Pd/C is tested in a real fuel cell.
72 3 Pt-MS Electrocatalysts for ORR

3.4.1 Pt MSs Deposited on Pd(111) and Pd NPs

Figure 3.6a shows the typical CV for the Cu UPD on a Pd(111) surface, with a single
peak at the potential of 0.50 V. The scan rate is 20 mV s−1 . The dash line shows
the CV for a Pd(111) surface in the absence of Cu2+ in the solution. Both the two
CVs agree with the data in a previous literature [80]. The electrode is held at the
potential of 0.32 V and transferred out from the solution at that potential. The charge
associated with the Cu UPD is 448 µC cm−2 , which is close to 490 µC cm−2 needed
for depositing a pseudomorphic Cu MS on a Pd(111) surface. Figure 3.6b shows the
CVs for the Pt MS on a Pd(111) surface (solid line) and for a bare Pd(111) surface

Fig. 3.6 CVs for a Pd(111) surface and the Pt MS on a Pd(111) surface [37]. a CVs for the Cu UPD
on a Pd(111) surface (solid line) in 50 mM H2 SO4 with 50 mM Cu2+ , and for a Pd(111) surface in
the absence of Cu2+ (dash line). b CVs for the Pt MS on a Pd(111) surface (solid line) obtained by
galvanic replacement of the Cu MS from (a), and for a bare Pd(111) surface (dash line)
3.4 Pt MSs on Pd(111) and Carbon-Supported Pd NPs for ORR 73

Fig. 3.7 STM images for the Pt MS on a Pd(111) surface [37]. a 250 nm × 250 nm, and b 125 nm
× 125 nm, of the Pt MS deposited on a Pd(111) surface by galvanic replacement of a Cu MS

without Pt (dash line). The electrolyte solution is 50 mM H2 SO4 , and the scan rate
is 20 mV s–1 . As expected, the deposition of the Pt MS causes a partial blocking of
the oxide formation on Pd(111), since Pt is oxidized at more positive potentials.
The STM measurement is used for the additional characterization of the Pt deposi-
tion. Figure 3.7 shows two STM images of the Pt MS deposited on a Pd(111) surface
obtained at 0.8 V in a 0.1 M HClO4 solution. The electrode potential is 0.8 V in a
0.1 M HClO4 solution, and the tunneling current is 1.24 nÅ. The deposition consists
of interconnected Pt islands, with some holes of monoatomic depth. The steps are
those of the Pd surface. A similar deposition is observed for the Pt MS on Au(111)
[81]. The Pt islands are probably epitaxial with a Pd(111) surface, given the small
mismatch between the Pt and Pd lattices. An interesting property of such a Pt MS
is an increased atomic scale roughness and the low coordination of many atoms that
are considered to have enhanced ORR electrocatalytic activity.
Figure 3.8 shows the TEM images of Pd NPs on a carbon (Vulcan XC 72) support,
while Fig. 3.9 displays the Pd particle size distribution obtained from Fig. 3.8a. The
size distribution peaks at about 9 nm. Some of the larger particles appear to be the
agglomerates of several smaller ones. The atomically resolved image in Fig. 3.8b
shows Pd atomic rows with the row spacing corresponding to the (111) facets of the
Pd NP.
Figure 3.10a shows the CV for the Cu UPD on Pd NPs (solid line) and the base
CV for Pd NPs (dash line). The Pd loading is 10 nmol Pd (60.9 nmol cm−2 ), and
the scan rate is 50 mV s−1 . Several peaks are seen positive to the onset potential of
the Cu bulk deposition at 0.21 V, and their associated charge is 2.1 mC cm−2 . This
charge is used to calculate the effective Pd surface area for the Pt deposition that will
be available for ORR (vide infra). The geometric electrode surface area is used in
this plot.
74 3 Pt-MS Electrocatalysts for ORR

Fig. 3.8 TEM images of the Pd NPs on a carbon (Vulcan XC 72) support [37]. a Low-magnification
overview showing the spread of the Pd particle size distribution on the carbon support, b high-
resolution image showing one-dimensional lattice fringes corresponding to the (111) lattice planes
in Pd

Fig. 3.9 Pd particle size distribution obtained from Fig. 3.8a [37]

Figure 3.10b shows the CV for the Pt MS on Pd NPs obtained by replacing the Cu
MS from Fig. 3.10a. The Pd loading is 10 nmol Pd (60.9 nmol cm-2 ). The electrolyte
solution is 50 mM H2 SO4 , and the scan rate is 50 mV s−1 . The insert in (b) depicts a
proposed structural model for the electrocatalyst in the form of a Pd cubo-octahedron
with a two-dimensional Pt MS on its surface. A full Pt MS is shown only on one
3.4 Pt MSs on Pd(111) and Carbon-Supported Pd NPs for ORR 75

Fig. 3.10 CVs for Pd10 /C and Pt/Pd10 /C [37]. a CVs for the Cu UPD on carbon-supported Pd NPs
(solid line) in 50 mM H2 SO4 with 50 mM Cu2+ , and for a Pd surface in the absence of Cu2+ (dash
line). b CVs for the Pt MS on Pd NPs (solid line) obtained by galvanic replacement of the Cu MS
from (a), and for the Pd NPs (dash line)

face for clarity. It depicts a characteristic hydrogen adsorption/desorption region and

the Pt–OH formation and reduction. As in the case of a Pd(111) surface (Fig. 3.6b),
the deposition of a Pt MS shifts the Pt–OH formation/reduction to more positive
potentials in comparison with the process on Pd NPs. The inset in Fig. 3.10b is a
proposed structural model for the electrocatalyst in the form of a Pd cubo-octahedron
with a two-dimensional Pt MS on its surface. A full Pt MS is shown only on one face
for the sake of clarity.
76 3 Pt-MS Electrocatalysts for ORR

3.4.2 Pt MS on a Pd(111) Surface for ORR

Figure 3.11 shows the rotating disk-ring electrode measurements for the ORR of
the Pt MS on a Pd(111) surface in a 0.1 M HClO4 solution, the rotation rates are
indicated in the graph, the scan rate is 20 mV s−1 , the ring potential is 1.27 V, the ring
and disk areas are 0.126 and 0.283 cm2 , respectively, and the collection efficiency is
24%. The ORR electrocatalytic activity of this surface is considerable, as indicated
by the very positive onset potential (0.95–1 V), a half-wave potential of 0.838 V, and
the lack of ring currents in the kinetic region. Measurable ring currents are observed
only in the potential region of diffusion control, which is unsuitable for analyzing
the disk-ring measurements. Since no peroxide is detected at the ring electrode in the
kinetic region, based on these measurements, it is impossible to conclude whether
the reaction involves a four-electron reduction in a direct or series pathway.
The K–L plots at different potentials show a linear dependence at all potentials
(Fig. 3.12). The electrode potentials are indicated in the graph, the inset shows the

Fig. 3.11 Polarization curves obtained with a rotating disk-ring electrode for the ORR of the Pt
MS on a Pd(111) surface in 0.1 M HClO4 solutions [37]
3.4 Pt MSs on Pd(111) and Carbon-Supported Pd NPs for ORR 77

Fig. 3.12 K–L plots at different potentials obtained from the data in Fig. 3.11 [37]

Tafel plot obtained from the kinetic currents jk obtained from these plots. The linearity
and the parallelism of these plots are usually taken to indicate first-order kinetics with
respect to molecular oxygen, although this criterion is not very specific [82]. From
the K–L equation as follows.

1/j = 1/n FkCO2 + 1/Bω1/2 (3.2)

where j is the current density, F is the Faraday constant, k is the reaction rate constant,
CO2 is the concentration of the dissolved O2 , B is the constant, and ω is the rotation
rate, we can calculate from the intercepts of the 1/j axis at 1/ω1/2 = 0 the ORR
kinetic currents. From the slopes of the K–L plots, i.e., the constant B, the number
of electrons exchanged in the reduction of oxygen molecule can be obtained. The
experimental value of B = 0.0451 mA rpm−1/2 , evaluated from Fig. 3.12, agrees well
with the calculated value of B = 0.044 mA rpm−1/2 . The calculation is performed
for a four-electron reduction using the published data for O2 solubility (1.26 × 10–3
mol−1 ), the solution’s viscosity (1.009 × 10–2 cm2 s−1 ), and oxygen diffusivity (1.93
× 10−5 cm2 s−1 ) [70, 83–85]. A four-electron reduction for ORR is in agreement
with the negligible currents for H2 O2 oxidation on the ring electrode.
78 3 Pt-MS Electrocatalysts for ORR

Fig. 3.13 Comparison of the polarization curves for the ORR on Pd(111), Pt(111) and Pt/Pd(111)
in 0.1 M HClO4 solutions [37]

The Tafel plot obtained from the kinetic currents jk is given as the inset in Fig. 3.12.
The plot is linear with a slope of ~−90 mV dec−1 above 0.65 V. The slope of −118 mV,
usually observed for the bulk Pt at high current densities, indicates the first electron
transfer RDS. It appears likely that the intrinsic slope for the ORR on Pt/Pd(111)
surface is −118 mV dec−1 , signifying that the first electron exchange is the RDS on
this surface, as on Pt (vide infra).
Figure 3.13 compares the ORR kinetics on Pd(111), Pt(111) and Pt/Pd(111) in
0.1 M HClO4 solutions. The LSV for Pt(111) is taken from reference, the scan rate
for Pd(111) and Pt/Pd(111) is 20 mV s−1 , while for Pt(111), it is 50 mV s−1 . The
LSV for Pt(111) is from reference [70]. The ORR electrocatalytic activity of the
Pt(111) and Pt/Pd(111) surfaces are considerably larger than that of Pd(111). The
small increase in the ORR kinetics on the Pt/Pd(111) surface compared to that on
Pt(111) is surprising, since the latter’s surface, along with Pt(110), is the most active
low-index crystal plane in HClO4 solutions. Although small, this improvement is of
considerable importance, because it shows that, with a suitable support, it is possible
to devise a very active electrocatalyst with only a monolayer amount of Pt whose
ORR electrocatalytic activity can surpass that of the bulk Pt. The studies of the ORR
on Pt3 Ni and Pt3 Co alloys may also indicate a possibility to reduce Pt to the top layers,
although the “Pt skin,” formed upon the dissolution of Ni from the surface layers in
these electrocatalysts, is usually several layers thick and supported by Pt-rich alloys
[86–89].
3.4 Pt MSs on Pd(111) and Carbon-Supported Pd NPs for ORR 79

3.4.3 Pt MSs on Carbon-Supported Pd NPs for ORR

Figure 3.14 displays the rotating disk-ring polarization curves for the Pt MSs on
carbon-supported Pd NPs. The rotation rates are indicated in the graph, the scan rate
is 10 mV s−1 , the ring potential is 1.27 V, the ring and disk areas are 0.037 and
0.164 cm2 , respectively, and the collection efficiency is 22%. The ring currents are
negligible, indicating a complete four-electron reaction pathway for ORR on the disk
electrode. The electrode consists of Pt MSs deposited on 20 nmol of Pd/C placed on
glassy carbon rotating disk electrode. Using the average size of 9 nm for the Pd NPs
determined by TEM measurements, the ratio of surface atoms to the total number
of atoms was estimated as 15% based on Benfield’s calculation using an icosahedral
particle model [90]. Therefore, for the sample with 20 nmol Pd on carbon, the amount
of Pt deposited by replacing the Cu MS on the Pd surface is approximately 3 nmol
(18 nmol cm−2 ), or 3.4 µgPt cm−2 . This estimation can be verified by calculating the
charge associated with putting Cu MSs onto Pd NPs. From this charge, the surface
area accessible for the electrochemical reaction is obtained, which is exactly what is

Fig. 3.14 Polarization curves obtained with a rotating disk-ring electrode for the ORR of the Pt
MSs on carbon-supported Pd NPs in 0.1 M HClO4 solutions [37]
80 3 Pt-MS Electrocatalysts for ORR

needed for this analysis. The charge for depositing Cu MSs, after correcting for the
double-layer charging of Pd, is 2.1 mC cm−2 for 10 nmol of Pd on the disk. Assuming
there is a one-to-one ratio between the Cu and Pd atoms in a pseudomorphic MS, this
charge indicates that Cu is deposited on 10.5 nmol cm−2 of Pd surface atoms. This
number is in good agreement with the model’s calculation (18/2 = 9 nmol cm−2 )
given above. This method of determining the real surface area of a Pt MS on Pd
surface may be more reliable than measuring H adsorption, because of the possible
interference of H absorption/desorption on Pd.
A very high ORR electrocatalytic activity of the Pt MSs on Pd NPs is indicated by
a half-wave potential of 0.853 V and by the negligible ring-electrode currents. The
K–L plots obtained from the data in Fig. 3.14 are shown in Fig. 3.15. The electrode
potentials are indicated in the graph, the inset shows the Tafel plot obtained from the
kinetic currents jk obtained from these plots. For a Pt/Pd(111) surface, a set of parallel
straight lines at different potentials is obtained. The Tafel slope of −96 mV dec−1
fits the polarization curve obtained from the plot of log(jk ) versus E, given as the
inset in Fig. 3.15. This slope is comparable to the slope of −90 mV dec−1 obtained
for a Pt/Pd(111) surface discussed above.
The size of the Pd NPs (9 nm) used to prepare the electrodes with Pt MSs is
considerably larger than the optimal size of Pt NPs for ORR, i.e., 4 nm according to
Peuckert et al.’s study [91]. A reduction in the size of the Pd NPs below 9 nm can

Fig. 3.15 K–L plots at different potentials obtained from the data in Fig. 3.14 [37]
3.4 Pt MSs on Pd(111) and Carbon-Supported Pd NPs for ORR 81

Fig. 3.16 Comparison of polarization curves for the ORR on Pd (10 nmol) and Pt (10 nmol) NPs,
and the Pt MSs on Pd NPs (10 and 20 nmol Pd) [37]

increase the ORR electrocatalytic activity of this bimetallic surface, and the amount
of 20 nmol of Pd can be decreased while increasing the number of the active Pt sites.
Figure 3.16 graphs the ORR for Pd and Pt NPs, as well as the Pt MSs on Pd NPs
(two different loadings). The electrode geometric area is 0.164 cm2 . Comparing the
ORR electrocatalytic activity of the Pt/C with an average particle size of 3.1 nm with
that of the Pt/Pd/C of 9 nm is not an adequate assessment, because of their different
surface areas, but this does not affect our main conclusion. The ORR electrocatalytic
activity of the Pt MSs on Pd NPs (10 nmol) is much higher than that of pure Pd
NPs (10 nmol) as indicated by a shift of the half-wave potential by 120 mV to
positive values. More importantly, the ORR electrocatalytic activity of this surface
is somewhat higher (25 mV in half-wave potential) than that of Pt NPs (10 nmol).
For the Pd loading of 10 nmol, or 6.4 µgPd cm−2 , the amount of Pt in the MS on this
surface is 1.5 nmol, or 1.7 µgPt cm−2 . The half-wave potential for this electrode is
0.838 V. It is important to note that the ORR electrocatalytic activity of this surface
is higher than that of 10 nmol (12 µgPt cm−2 ) of Pt NPs, despite the fact that the Pd
NPs are 9 nm and the Pt NPs are 3 nm, and the former has a smaller real surface
area. The electrode consisting of a Pt MS on 20 nmol Pd has the highest ORR
electrocatalytic activity, mainly due to the increased Pt surface area (Fig. 3.16). The
higher ORR electrocatalytic activity of the Pt-MS electrocatalysts compared to those
of Pt and Pd indicate a synergetic effect, which is particularly interesting since the
ORR electrocatalytic activity of the Pt/Pd surface surpasses that of Pt NPs with seven
times larger loading.
In addition to the polarization curves, a useful way of comparing of the activities
of various electrocatalysts is by their MSA. Figure 3.17a shows the Pt-MSA of the
three electrodes containing Pt at 0.85 and 0.80 V. The electrodes having the Pt MS
have 5–8 times higher ORR electrocatalytic activity than the electrode with Pt NPs.
82 3 Pt-MS Electrocatalysts for ORR

Fig. 3.17 ORR electrocatalytic activity. a Pt- and b NM-MSA of the Pt (10 nmol) NPs and the Pt
MSs (1.3 and 2.4 nmol) on Pd NPs (10 and 20 nmol Pd) at 0.8 and 0.85 V [37]

This finding underlines the importance of the MS electrocatalysts that can reduce
the amount of Pt in the fuel cell’s electrode down to very low levels. If the total
metal content is taken into account, i.e., the mass of Pt and Pd, the plot shows
still significantly higher ORR electrocatalytic activity of the Pt-MS electrocatalysts
(Fig. 3.17b).

3.4.4 Fuel Cell Tests for Pt-MS Electrocatalysts

The fuel cell tests for the Pt-MS electrocatalysts are conducted, and the long-term
electrochemical durability tests are carried out using the fuel cell with the electrodes
of 50 cm2 in area. The cathode is the Pt/Pd/C electrocatalyst containing 77 µg cm−2
of Pt (0.21 gPt kW−1 ) and 385 µg cm−2 of Pd.
3.4 Pt MSs on Pd(111) and Carbon-Supported Pd NPs for ORR 83

Fig. 3.18 Long-term electrochemical durability test of the Pt/Pd electrocatalyst in an operating
fuel cell at 80 °C [66]

Figure 3.18 illustrates the trace of the cell voltage at a constant current of
0.6 A cm−2 against time at 80 °C. The fuel cell voltage at constant current of
0.6 A cm−2 is plotted as a function of time with a cathode containing 385 µg cm−2 of
Pd and 77 µg cm−2 of Pt (or 0.21 g kW−1 of Pt). A commercial Pt/C electrocatalyst
(180 µg cm−2 of Pt) is used for the anode. Up to 1000 h, the cell voltage drops about
120 mV and then shows a much lower decrease rate with time. The total loss in
voltage is approximately 130 mV at 2300 h, at the end of the test. CV measurements
occasionally are made during the test to investigate the changes in active surface area
of Pt in the cathode from the charges determined by hydrogen adsorption peaks. This
indicates the losses of the active Pt surface area, by approximately 26% at 1200 h and
29% at 2300 h compared to the initial surface area before testing. These losses in Pt
can reflect the dissolution of Pt at the open circuit potential, or the embedding of Pt
atoms into Pd substrate. The latter is predicted by the anti-segregation of Pt from Pd
according to the DFT calculations [92]. It is noted that a part of the observed drop in
voltage may also be due to the losses at the anode (a commercial Pt/C electrocatalyst)
since the measurements do not separate the losses at the anode and cathode in these
fuel cells. The results are promising but further investigation is also necessary to
improve the long-term electrochemical durability of the Pt-MS electrocatalyst.

3.4.5 Discussion

The ORR electrocatalytic catalytic properties of the bimetallic surfaces consisting

of the metal MSs on metal single-crystal surfaces have been extensively studied
84 3 Pt-MS Electrocatalysts for ORR

in an ultrahigh vacuum (UHV) environment and to a lesser extent in electro-

chemical systems [93–97]. In many cases, the formation of a surface metal–metal
bond produces the large changes in the electronic properties of the metal MS, and
pronounced differences are observed in the reactivity of some transition metal MSs
on different substrates [95]. Some interpretations of these effects include the shifts
of core levels due to a charge transfer between MS and the substrate or the changes
in the density of states near the Fermi level [95, 96]. Meitzner et al.’s earlier work on
the bimetallic alloys and NP catalysts links the trends in the reactivity to the d-band
occupancy and electronegativity, which is also used to interpret the ORR electrocat-
alytic activity of several alloys [71, 98]. An increased 5d vacancy of Pt, as a result of
its interaction with a substrate metal, is believed to increase the interaction between
O2 and Pt, thereby enhancing the ORR electrocatalytic activity of a Pt “skin” (several
layers) on a PtNi, or PtFe alloy [99]. Stamenkovic et al. ascribe the enhanced ORR
electrocatalytic activity of a Pt “skin” in these alloys to a smaller formation of Pt–OH
on that surface [87, 88].
A description of the ORR electrocatalytic activity of metal MSs is proposed by
Nørskov and coworkers based on the DFT calculations [92]. According to them,
the reactivity scales well with the shifts in the center of the d-band for the strained
crystals and metal MSs. Some experimental support has been reported. For example,
the electronic factors in the bimetallic systems show a parallelism between the change
of adsorption energy and the d-band center shifts for CO and H2 [100]. The data for
a Pt sub-MS on Ru NPs and Ru(0001), where Pt is likely to be compressed (4%
difference in lattice constants), indicate a reduced adsorption of CO in comparison
with Pt [101, 102].
The above-presented results show increased ORR electrocatalytic activity of Pt
MSs on Pd substrates, surpassing that of Pt in both the single crystal and NP forms.
The possible factors that determine such activity may include a mismatch in the lattice
constants between the Pt MS and the Pd substrate, as well as the changes in the d-band
properties of Pt caused by its interaction with Pd. A mismatch between the lattice
constants is small, only 0.8%. Consequently, it may generate a very small compressive
strain in a pseudomorphic Pt MS on Pd. Given the small difference in the lattice
constants of Pt and Pd, and the fact that the Pt MS is not entirely pseudomorphic, a
decrease in activity produced by compression probably is negligible. The effect of
the d-band filling of the substrate is also expected to be slight since the fractional
filling of the d-bands of Pd and Pt is the same. However, their interaction can be
expected to result in some charge redistribution through hybridization of the states
from each atom. The DFT calculations show a small increase of ORR electrocatalytic
activity for a pseudomorphic Pt MS on a Pd(111) surface compared to Pt(111) [92].
For the inverted system, i.e., a Pd MS on a Pt(111) surface, the ORR electrocatalytic
activity decreases in acid solutions, but there is an increase in alkaline electrolytes
[97, 103].
The Tafel slope of about −90 mV dec−1 probably reflects a different state and
coverage of Pt–OH on the Pt/Pd(111) surface in comparison with bulk Pt, or Pt(111).
Slopes of −59 and −118 mV dec−1 are observed for polycrystalline Pt surfaces at
low and high current densities in non-adsorbing electrolytes, respectively [104, 105].
3.4 Pt MSs on Pd(111) and Carbon-Supported Pd NPs for ORR 85

Marković et al. report a single slope of −75 mV for Pt(111) in HClO4 solutions
[70]. Several authors have discussed the role of Pt–OH in determining the change
of the Tafel slope. Damjanovic and his colleagues attribute the “low” Tafel slope
at high potentials in HClO4 solutions to the Temkin conditions for the adsorption
of the intermediates, mainly Pt–OH [106]. Several workers propose that Pt–OH is
not derived from the reduction of O2 , but rather from the reaction of H2 O with Pt,
and that causes inhibition of O2 reduction [107, 108]. Albu and Anderson suggest
that the cathodic transfer coefficient, β, may vary with potential from about 0.5 to
near 1.0 causing changes in the Tafel slope from −118 to −59 mV with increasing
potential [76, 109]. This proposition, however, does not explain why the Tafel slope
is independent of the potential for O2 reduction on Pt(111) in H2 SO4 solutions.
Marković et al. propose a model for the kinetic current, which assumes that the
adsorbed OH can alter the adsorption energy of the ORR intermediates, thus having
an energetic effect on the ORR kinetics and on the Tafel slope in addition to blocking
the Pt sites [70].
Wang et al. demonstrate that, in addition to the site blocking, Pt–OH has a negative
electronic effect on the ORR kinetics [110]. According to these authors, a decrease
in the coverage of Pt–OH can enhance the ORR kinetics. This possibility is also
discussed for a Pt “skin” on a Pt3 Ni support [87]. Figure 3.19 shows that the formation
of Pt–OH on the Pt MS on Pd NPs is considerably smaller than on the Pt NPs’ surfaces,
which may be partly responsible for the observed enhanced ORR kinetics of the Pt-
MS electrocatalyst. The electrolyte solution is 50 mM H2 SO4 , and the scan rate is
50 mV s−1 . In addition to this effect, an enhanced atomic scale surface roughness and
the low coordination of considerable number of surface atoms may also contribute
to the observed ORR electrocatalytic activity.

Fig. 3.19 CVs for the Pt MSs on Pd NPs (solid line) obtained by galvanic replacement of the Cu
MS from Fig. 3.10a, and for Pt NPs containing 10 nmol Pt on carbon electrode (dash line) [37]
86 3 Pt-MS Electrocatalysts for ORR

A strong evidence of the delayed oxidation of a Pt MS on Pd NPs in comparison

with the oxidation of Pt NPs is obtained from the in situ XANES measurements as
a function of the potential. Figure 3.20a shows the Pt L3 edge spectra obtained with

Fig. 3.20 XANES analyses. XANES spectra obtained with the a Pt/Pd and b Pt/C electrocatalysts
at four different potentials in 1.0 M HClO4 solutions. c A comparison of the change of the absorption
peak as a function of the potential for Pt/Pd/C and Pt/C
3.4 Pt MSs on Pd(111) and Carbon-Supported Pd NPs for ORR 87

the Pt/Pd electrocatalysts at four different potentials. Only at the highest potentials,
there is an increase in the intensity of white line as a consequence of the Pt–OH
formation causing a depletion of Pt’s d-band [111]. The increase in the intensity
of white line for the Pt/C electrocatalyst commences at considerably less positive
potentials (Fig. 3.20b, c). This shows that the oxidation of the Pt MS on Pd substrate
requires higher potentials than Pt NPs on carbon support. A comparison of the spectra
at 0.47 V shows a negligible change in the electronic properties of a Pt MS on Pd in
comparison with Pt/C (Fig. 3.20c). No effect of the potential change is observed below
1 V. This corroborates the above proposition of the role of Pd–OH in suppressing a
Pt–OH formation.

3.4.6 Summary

The ORR kinetics is studied in acid solutions for the Pt MS on a Pd(111) surface
and carbon-supported Pd NPs using the rotating disk-ring electrode technique. Pt
MSs deposited on Pd surfaces can be very active ORR electrocatalysts. The ORR
kinetics shows a small but significant enhancement with the Pt MSs on Pd(111)
and Pd NP surfaces in comparison with the reaction on Pt(111) and Pt NPs. The
four-electron reaction pathway, with a first charge transfer RDS, is operative on both
Pt/Pd(111) and Pt/Pd/C surfaces, with a very small amount of H2 O2 detected on the
ring electrode in the hydrogen adsorption potential region. The observed rise in the
ORR electrocatalytic activity of the Pt MS surfaces compared with Pt bulk and NP
electrodes may be partly caused by the decreased Pt–OH adsorption.
The results presented above illustrate that placing a Pt MS on NPs of a suitable
metal substrate is an attractive way in designing the improved ORR electrocatalysts,
and in reducing the Pt loadings in fuel cell electrodes. The method for the controlled
deposition of a metal MS, which involves a galvanic replacement of an UPD metal
MS, offers a unique way of depositing a metal MS on carbon-supported metal NPs
in a surface-limited reaction. This can be achieved by using neither metal vapor
deposition in UHV, nor chemical vapor deposition. The fact that it is not producing
pseudomorphic MSs is not a drawback, but rather a useful feature of this method.
A disordered MS, having imperfections and low-coordination atoms, is likely to be
more active than a pseudomorphic or a uniform MS. An additional support for this
view comes from the surface-enhanced Raman spectroscopy (SERS) measurements
with a Pt MS on Au prepared using the UPD and galvanic replacement method. The
excellent SERS spectra reported from Pt surfaces are obtained with it, which indicates
a high reactivity of such Pt MSs. Further work utilizing this approach seems quite
promising for both reducing the loading of noble metals and increasing the ORR
electrocatalytic activity of fuel cell electrocatalysts [112].
88 3 Pt-MS Electrocatalysts for ORR

3.5 Mixed Pt–M–MS Electrocatalysts for ORR

In the previous sections, it is proved that Pt/Pd has a higher ORR electrocatalytic
activity than pure Pt due to the decreased Pt–OH formation on the Pt/Pd surface at
high potentials. Along this line, if the OH coverage is further reduced up to a clean
Pt MS surface at high potentials, the further increase in ORR electrocatalytic activity
can be expected. This objective can be realized by adding another metal (M) to the
Pt MS forming the mixed Pt–M-MS electrocatalysts, with the other metal (M) being
more oxyphilic than Pt. Since OH species or O atoms are preferentially and more
strongly adsorbed on M rather than Pt, additional OH adsorption on Pt sites becomes
more difficult due to the OH–OH or OH–O repulsion. Therefore, it is possible to
create clean Pt sites by adjusting the coverage of M (see Fig. 3.21 for the schematic
diagram of the proposed model).
Indeed, a new class of ORR electrocatalysts has been prepared, consisting of
mixed Pt–M MSs (M = Ir, Ru, Rh, Re, or Os) deposited on a Pd(111) single crystal
or on carbon-supported Pd NPs. Several of these mixed Pt–M-MS electrocatalysts
exhibit very high ORR electrocatalytic activity and increased stability of Pt against
oxidation, as well as a 20-fold increase in Pt-MSA, compared to the conventional pure
Pt electrocatalysts. Their superior ORR electrocatalytic activity and stability reflect
a low OH coverage on Pt, caused by the lateral repulsion between the OH adsorbed
on Pt and the OH or O adsorbed on neighboring, other than Pt, late transition metal
atoms. The origin of this effect is identified through a combination of experimental
and theoretical methods, employing electrochemical techniques, X-ray absorption
spectroscopy and DFT calculations. This new class of electrocatalysts promises to

Fig. 3.21 Model for the decrease of the OH coverage on Pt, caused by a high OH or O coverage
on a second metal M [65]
3.5 Mixed Pt–M-MS Electrocatalysts for ORR 89

alleviate some major problems of the existing fuel cell technology by simultaneously
decreasing the materials cost and enhancing the performance.

3.5.1 ORR on M0.2ML /Pt(111)

In order to support the hypothesis of the OH–OH repulsion, a modified Pt(111)

surface with the 0.2 MS of Ir (or Re) deposition is tested for ORR. The electrode is
prepared by depositing a submonolayer of Cu (0.2 ML) on a well-defined Pt(111)
surface in the UPD potential region, which is followed by the Ir replacement of Cu.
Figure 3.22 shows a typical CV of Pt(111) in a 50 mM H2 SO4 solution with 50 mM
Cu2+ . A couple of cathodic and anodic peaks indicate the deposition and desorption
of Cu taking place in two major steps. The dash line in Fig. 3.22 is the Pt(111) CV
in a 0.1 M HClO4 solution, which is included for comparison.
Figure 3.23 shows the anodic parts of the CVs for Pt(111) and Pt(111) with Ir,
at θ Ir = 0.2. The scan rate is 50 mV s−1 . The presence of Ir on the Pt(111) surface
has minimal effect on the hydrogen adsorption/desorption reaction, shown by the
segment of the curves between 0.0 and 0.3 V, whereas the Pt–OH formation, the
process dominating the curves at E > 0.4 V, is considerably decreased and shifted
positively when Ir is present. In the presence of Ir (θ Ir = 0.2), the OH coverage on
Pt(111) at 1.1 V, as calculated by integrating the currents between 0.4 and 1.1 V,
is θ OH = 0.45, whereas on bare Pt(111) the θ OH = 1 (see Fig. 3.24). These results
demonstrate that there is a repulsion between the adsorbed OH species, even between
the OH bound on the surface M-adatom and the OH bound to the supporting Pt(111)
surface.

Fig. 3.22 CVs for the Cu UPD on a Pt(111) surface (solid line) and for a Pt(111) surface (dash
line)
90 3 Pt-MS Electrocatalysts for ORR

Fig. 3.23 Forward scans of the CVs for a Pt(111) in a 0.1 M HClO4 solution bare and covered by
the 0.2 ML of Ir (solid line) with the subtracted double layer charging and Ir oxidation currents [65]

Figure 3.25 compares the ORR kinetics on Pt(111), Ir0.2ML /Pt(111) and
Re0.2ML /Pt(111) in 0.1 M HClO4 solutions. The scan rate is 20 mV s−1 . The ORR
electrocatalytic activity of the Ir0.2ML /Pt(111) and Re0.2ML /Pt(111) surfaces is signif-
icantly larger than that of Pt(111). The small increase in ORR kinetics on the [Ir(or
Re)]0.2ML /Pt(111) surface compared to that on Pt(111) is surprising, since the latter’s
surface has even 20% more Pt sites available than the former one. Although small,
this improvement is of considerable importance because it shows that, the repulsion
between the OH species can be used to improve the ORR electrocatalytic activity
of bulk Pt. Figure 3.26 shows a comparison of the kinetic currents taken at 0.8 and
0.85 V of the indicated two surfaces. It can be seen that at both potentials, the kinetic
currents of Ir0.2ML /Pt(111) are about twofold higher than that of Pt(111) surface.

3.5.2 ORR on (M0.2 Pt0.8 )ML /Pd(111)

Figure 3.27 shows a comparison of the polarization curves of (Ir0.2 Pt0.8 )ML /Pd(111)
and of PtML /Pd(111) at 1600 rpm in 0.1 M HClO4 solutions. The scan rate is
20 mV s−1 . A 32 mV enhancement in half-wave potential of (Ir0.2 Pt0.8 )ML /Pd(111)
is observed with respect to that of PtML /Pd(111), which is already significantly more
active than the bulk Pt(111) surface. Figure 3.28 shows the kinetic currents obtained
from the K–L plots (not shown) at 0.8 V plotted as a function of the Ir (or Ru)
coverage in the surface MS. Both metals (Ir and Ru) are known to have high OH
coverage at considerable lower potential than Pt. In both cases, the maximum ORR
electrocatalytic activity is obtained with the MSs consisting of 80% Pt and 20% of
the second metal. The kinetic current density is more than three times larger than that
3.5 Mixed Pt–M-MS Electrocatalysts for ORR 91

Fig. 3.24 OH coverage analyses. a Integrated charge of the forward scans for Ir0.2ML /Pt(111) (solid
line) and Pt(111) (dash line) indicated in Fig. 3.23. b OH coverage calculated according to (a)

for a pure Pt MS on Pd. Figure 3.29 plots the kinetic currents normalized by the Pt
MS fraction when Ru is used as the second metal. Activity rises with increasing Ru
coverage up to 0.2 and remains constant above that value, suggesting that the activity
of Pt atoms, at Ru coverages higher than 0.2, remains invariable as a consequence of
a negligible coverage of OH on the Pt.
In order to put the work in a broader perspective within the framework of the
Pt–OH repulsion by an adsorbate on the neighboring metal M, similar experiments
are performed with Au, Pd, Rh, Os or Re as the alternative M component in the
mixed Pt–M MS supported on Pd(111). Figure 3.30 shows the results of these exper-
iments for all of the mixed Pt–M MSs at the 80:20 Pt:M ratio, wherein the measured
kinetic current densities are plotted against the effective repulsion energy between
92 3 Pt-MS Electrocatalysts for ORR

Fig. 3.25 Comparison of the polarization curves for the ORR kinetics on Pt(111), Ir0.2ML /Pt(111)
and Re0.2ML /Pt(111)) in 0.1 M HClO4 solutions at 1600 rpm [69]

two OH(a)s, or an O(a) and an OH(a) for the case of M = Re or Os, as calculated from
first principles (vide infra). Positive (negative) energies indicate more (less) repulsive
interaction between adsorbed OHs, respectively, compared to PtML /Pd(111). A very
good linear correlation is observed indicating that the primary effect in enhancing
the ORR kinetics on Pt can be attributed to the reduced OH coverage on Pt, which, in
turn, reflects the repulsion experienced by Pt–OH in the presence of M–OH or M–O
formed at lower potentials. The second metal (M) used is important: Au binds OH
weakly, and, consequently, Au atoms in the mixed Pt–Au MS do not enhance the
ORR kinetics; in fact, the measured current densities for Pt–Au were slightly lower
than those of a pure Pt MS. As expected, Pt–Rh shows an intermediate behavior,
since Rh binds OH less strongly than Ru or Ir. Higher activity is observed for Pt–Ir
and Pt–Ru, in accord with the strong adsorption of OH on Ir and Ru at low potentials.
3.5 Mixed Pt–M-MS Electrocatalysts for ORR 93

Fig. 3.26 Kinetic currents at 0.8 and 0.85 V of the indicated cases calculated from the individual
K–L plots (not shown)

Fig. 3.27 Comparison of the polarization curves for the ORR kinetics on (Ir0.2 Pt0.8 )ML /Pd(111)
and PtML /Pd(111) in 0.1 M HClO4 solutions at 1600 rpm [63]

The highest activity of all of the mixed Pt–M MSs tested is recorded for M being
Os or Re, both of which, as discussed below, tend to adsorb O rather than OH. The
repulsion on OH adsorbed on Pt, caused by adsorbed O on M, is even greater than
that caused by OH adsorbed on M in the mixed surface layer, thereby enhancing
ORR the most.
94 3 Pt-MS Electrocatalysts for ORR

Fig. 3.28 ORR electrocatalytic activity of two mixed Pt–M-MS electrocatalysts expressed as the
kinetic current density at 0.80 V as a function of the M/Pt ratio in the MS [65]

Fig. 3.29 Pt coverage normalized kinetic current density at 0.80 V as a function of the M/Pt ratio
in the MS [63]

3.5.3 ORR on (M0.2 Pt0.8 )ML /Pd NPs/C

Figure 3.31 shows a comparison of the CVs of (Ir0.2 Pt0.8 )ML /Pd20 /C and of
PtML /Pd20 /C in 0.1 M HClO4 solutions. The curve shape of the H adsorption–
desorption region is slightly changed, with Hads/des on (Ir0.2 Pt0.8 )ML /Pd20 /C surface
commencing at more negative potentials. This is probably because the potential
of zero charge on this surface occurs at more negative potential. However, the
integrated H desorption charge is nearly the same for these two cases. At 0.5–
0.7 V, (Ir0.2 Pt0.8 )ML /Pd20 /C has more pronounced anodic and cathodic peaks than
PtML /Pd20 /C, indicating that Ir has more OH coverage than Pt surface at the same
3.5 Mixed Pt–M-MS Electrocatalysts for ORR 95

Fig. 3.30 Kinetic current densities at 0.80 V as a function of the calculated interaction energy
between two OHs, or OH and O [65]

Fig. 3.31 CVs for the indicated surfaces in 0.1 M HClO4 solutions, the scan rate is 20 mV s−1

potential region. While the potential is further increased up to 1.1 V, the Pt–
OH formation is considerably inhibited on (Ir0.2 Pt0.8 )ML /Pd20 /C, compared with on
PtML /Pd20 /C, as shown in Fig. 3.19. This behavior is well in the line of the prediction
about the OH–OH repulsion.
Figure 3.32 shows the polarization curves of (Ir0.2 Pt0.8 )ML /Pd20 /C and
(Re0.2 Pt0.8 )ML /Pd20 /C compared with PtML /Pd20 /C and Pt10 /C at 1600 rpm in 0.1 M
HClO4 solutions. The scan rate is 10 mV s−1 . The trend of the ORR electrocat-
alytic activity agrees well with the results measured on the corresponding MSs on
single crystals. It is (Re0.2 Pt0.8 )ML /Pd20 /C > (Ir0.2 Pt0.8 )ML /Pd20 /C > PtML /Pd20 /C >
96 3 Pt-MS Electrocatalysts for ORR

Fig. 3.32 Comparison of the polarization curves for the ORR on the indicated surfaces in
0.1 M HClO4 solutions at 1600 rpm [63, 69]. a Comparison of (Ir0.2 Pt0.8 )ML /Pd20 /C with other
electrocatalysts. b Comparison of (Re0.2 Pt0.8 )/Pd20 /C with other electrocatalysts

Pt10 /C, in the sequence of ORR electrocatalytic activity decreasing. The Tafel plots
obtained from the K–L plots (not shown) are shown in Fig. 3.33. The Tafel slope
of −118 mV dec−1 is usually observed on the clean Pt surface at the high current
density region, indicating one electron transfer as the RDS. The lower Tafel slope
on Pt surface at the region close to open circuit potential is believed to be caused
by the property of the OH adsorption coverage being a function of potential. The
−112 mV dec−1 for (Ir0.2 Pt0.8 )ML /Pd20 /C is very close to the −118 mV dec−1 for
the theoretical one electron transfer rate-determining process, implying a negligible
Pt–OH formation on this surface. The Tafel slope of PtML /Pd20 /C has a middle value
between those of Pt10 /C and (Ir0.2 Pt0.8 )ML /Pd20 /C, indicating that the OH coverage
is also between those of the corresponding two surfaces.
Figure 3.34 displays the Pt- and NM-MSA of (Ir0.2 Pt0.8 )ML /Pd20 /C and of a
Pt/C commercial electrocatalyst for the ORR at 0.8 V in 0.1 M HClO4 solutions.
3.5 Mixed Pt–M-MS Electrocatalysts for ORR 97

Fig. 3.33 Comparison of the

Tafel plots from the K–L
plots for surfaces in 0.1 M
HClO4 solutions

Fig. 3.34 Pt- and NM-MSA

of (Pt0.8 Ir0.2 )ML /Pd/C and
Pt/C commercial
electrocatalyst for the ORR
at 0.8 V in 0.1 M HClO4
solutions

This remarkably larger Pt-MSA of (Ir0.2 Pt0.8 )ML /Pd20 /C, which is about twenty
times greater than that of the state-of-the-art commercial Pt/C, points to a strong
promise inherent to this class of the mixed Pt–M-MS electrocatalysts. The NM-
MSA of (Ir0.2 Pt0.8 )ML /Pd20 /C is fourfold higher than the Pt/C commercial electro-
catalyst, which is still very satisfactory, particularly since its main constituent Pd is
considerably less expensive than Pt.
The data collected with in situ X-ray absorption near edge structure spectroscopy
(XANES) on carbon-supported (Pt0.8 Ir0.2 )/Pd20 /C NPs clearly suggest that the oxida-
tion of Ir (the formation of Ir–OH) occurs readily in the mixed Pt0.8 Ir0.2 MS, whereas
the oxidation of Pt in the same MS is highly suppressed, occurring only at signifi-
cantly higher potentials. Figure 3.35a shows the Pt L3 edge absorption spectra for the
carbon-supported (Pt0.8 Ir0.2 )/Pd20 NPs in 1.0 M HClO4 solutions at the potentials of
0.47 and 1.17 V; the spectrum for Pt/C NPs at 1.17 V is presented for comparison. The
change observed in the absorption edge peak of the (Pt0.8 Ir0.2 )/Pd20 /C (commonly
98 3 Pt-MS Electrocatalysts for ORR

Fig. 3.35 XANES analyses [65]. a XANES spectra obtained with (Pt0.8 Ir0.2 )ML /Pd/C at 0.47 and
1.17 V and with Pt/C at 1.17 V in 1.0 M HClO4 solutions. b Relative change of the peak intensity
(μE − μ0.47 )/μ0.47 for the Pt in (Pt0.8 Ir0.2 )ML /Pd/C and in Pt/C, and for the Ir in (Pt0.8 Ir0.2 )ML /Pd/C
as a function of potential

called the “white line”) between 1.17 and 0.47 V, is negligible. Given that Pt is defi-
nitely not oxidized at 0.47 V, these results suggest that there is no Pt–OH formation
on (Pt0.8 Ir0.2 )/Pd20 /C at the potential as high as 1.17 V. In contrast, the Pt/C shows
a pronounced peak at 1.17 V. The increase in the intensity of the edge peak is a
consequence of the depletion of Pt’s d-band caused by its oxidation, i.e., the Pt–
OH formation [111]. Since this reaction is not observed with (Pt0.8 Ir0.2 )/Pd20 /C, it is
concluded that in the presence of Ir, there is no Pt–OH formation up to 1.17 V. On
the other hand, the corresponding spectra for the Ir L3 edge demonstrate the facile
oxidation of Ir (Fig. 3.35b).
Furthermore, Fig. 3.35b displays the relative change of the absorption intensity
as a function of potential for Pt and Ir in a mixed Pt–Ir MS on Pd and for Pt/C,
normalized by the intensity at 0.47 V, (μE − μ0.47 )/μ0.47 . These plots clearly show
the easy oxidation of Ir in the Pt–Ir MS and of Pt in Pt/C, while the oxidation of Pt in
3.5 Mixed Pt–M-MS Electrocatalysts for ORR 99

Fig. 3.36 Pt and Ir L3 edge X-ray absorption spectra for (Pt0.8 Ir0.2 )ML /Pd/C in 1.0 M HClO4
solutions at 0.47 V [65]

the Pt–Ir MS is observed only at a much higher potential (e.g., 1.17 V), and even then
it is only minor. The Ir atoms in the Pt–Ir MS are oxidized at relatively low positive
potentials, and the resulting Ir–OH formation distinctly inhibits Pt oxidation. Hence,
these data strongly support the model sketched in Fig. 3.21 and verify the repulsion
between the OH on Ir and the OH on Pt.
The deposition of the mixed Pt–M MS on Pd by replacing a Cu MS is a uniquely
convenient method for putting a controlled number of metal atoms on a suitable
substrate [81]. However, the verification of its success is required because of the
different deposition kinetics of the two metals. The check of the composition of
(Pt0.8 Ir0.2 )/Pd20 /C, based on the differences between the absorption coefficients at
50 eV below and above the L3 edge in the spectra for Pt and Ir (Fig. 3.36), indicates
that the molar Pt:Ir ratio is 4.66:1, or 82% Pt and 18% Ir. These values are in very
good agreement with the 80:20 ion concentration ratio used in the electrocatalyst
preparation method [81].

3.5.4 Summary

A class of ORR electrocatalysts is synthesized from a mixed MS of Pt and another

late transition metal (Ir, Ru, Rh, Re or Os) deposited on Pd substrates. These elec-
trocatalysts have: (1) the Pt-MSA up to twenty times higher than the state-of-the-art
commercial electrocatalysts, (2) the increased stability of Pt against oxidation and
(3) the very low Pt content. Their superior ORR electrocatalytic activity and stability
reflect a low OH coverage on Pt, caused by the lateral repulsion between the OH
adsorbed on Pt and the OH or O adsorbed on neighboring, other than Pt, late transition
100 3 Pt-MS Electrocatalysts for ORR

metal atoms. This class of electrocatalysts promises to alleviate some major prob-
lems of the existing fuel cell technology by simultaneously decreasing the materials
cost and enhancing the performance.

3.6 Pt MSs on Carbon-Supported Au–Ni NPs for ORR

The goal is to obtain an optimized substrate for a Pt MS and to further reduce the
NM content in the ORR electrocatalysts. The method introduced in the previous
sections for depositing a Pt MS on another metal substrate involves the Cu UPD on
the substrate metal as the first step. However, in aqueous solutions, Cu UPD occurs
only on noble metals under convenient experimental conditions. There is no report
so far about the Cu UPD on non-noble metals or metal oxides in aqueous solutions,
which limits the more extensive application of this technique. For example, Pt–Ni
and Pt–Co alloys are reported more active for ORR due to the electronic interaction
between Pt and the other metal, but these alloys are still rich in Pt content, with
most of the Pt in the bulk without taking part in the electrocatalytic reaction [87,
88]. However, Cu UPD occurs neither on Ni nor on Co. Thus, a Pt MS cannot be
deposited on the two metals using the UPD method.
According to the binary alloy phase diagram of the Au–Ni system, the two compo-
nents are virtually completely immiscible in the bulk at room temperature [113].
However, it has been found by the STM that the deposition of Au on Ni single-
crystal surfaces can result in a stable surface alloy in the first atomic layer [114]. The
low energy electron diffraction of Au on Ni(110) has also shown the formation of a
surface alloy restricted to the outermost atomic layer [115]. The theoretical calcula-
tions using the DFT technique have also confirmed that the Au–Ni surface alloy is
energetically favorable [116, 117].
On the basis of these results and the knowledge of the strong segregation of the
Au–Ni system, an electrocatalyst is synthesized by co-depositing Au and Ni to form
NPs on carbon (XC-72), followed by annealing at 600 °C in H2 [118]. This results
in a surface segregation of Au. The Pt MS is deposited by the galvanic replacement
of a Cu MS deposited at underpotentials on Au/Ni surface. Figure 3.37 shows the
schematic model of this procedure. RRDE, XRD and XANES measurements are
carried out for this electrocatalyst: PtML /AuNi10 /C. A fuel cell test has also been
done on it.

3.6.1 Electrochemical and XRD Characterization of Au/Ni

NPs

Figure 3.38 shows the Cu UPD on AuNi10 NPs (solid line), and the base curve for
AuNi10 (dash line). The Au and Ni loading on the RDE are 15 and 150 nmol cm−2 ,
3.6 Pt MSs on Carbon-Supported Au–Ni NPs for ORR 101

Fig. 3.37 Schematic model for the procedures of the synthesis of PtML /AuNi/C [36]

Fig. 3.38 CVs for the Cu UPD on carbon-supported AuNi10 NPs on glassy carbon disk (solid line)
in 50 mM H2 SO4 solutions with 50 mM Cu2+ , and for a AuNi10 surface in the absence of Cu2+
(dash line) [36]

respectively. The base curve of AuNi10 does not show any Ni dissolution, indicating
a good protection by Au. Several peaks are seen positive to the onset of the Cu
bulk deposition at 0.33 V; their associated charge is 7.4 nmolCu cm−2 . This charge
was used to calculate the effective Pt surface area for the Pt deposition that will be
available for ORR. The geometric electrode surface area is used in this plot.
Figure 3.39 shows the CV for the Pt MS on AuNi10 NPs obtained by replacing
the Cu MS from Fig. 3.38. The electrolyte solution is 0.1 M HClO4 and the scan
rate is 50 mV s−1 . It depicts a characteristic hydrogen adsorption/desorption region
at E = 0–0.35 V and the Pt–OH formation and reduction at E = 0.6–1.1 V. The
Pt loading calculated from the H adsorption charge is 7 nmolPt cm−2 or 1.4 µgPt
cm−2 , which agrees well with the UPD Cu amount. It is noted that this is an ultralow
loading of Pt. Assuming there is a one-to-one ratio between the Cu and Au atoms in
a pseudomorphic MS, this charge indicates that Au is about two monolayer thick.
102 3 Pt-MS Electrocatalysts for ORR

Fig. 3.39 CVs for a Pt MS on AuNi10 NPs (solid line) obtained by the galvanic replacement of the
Cu MS from Fig. 3.38, and for AuNi10 NPs (dash line) [36]

In order to verify the surface segregation of Au, the XRD measurement is carried
out. Figure 3.40 shows the results, the XRD intensity profile obtained for the AuNi10
NPs is compared with that of pure Au NPs. The strongest peak in the AuNi10 curve
is near 26.2°, which can be assigned to the reflection of Ni(111) (26.2°) and of
Au(200) (26.1°). The symmetric peak at 30.33° is due to the Ni(200) reflection. In
contrast to the symmetric peaks of the reference curve of pure Au NPs, the Au(111)
and Au(220) peaks of AuNi10 NPs are unsymmetrical and shifted to higher angles,

Fig. 3.40 X-ray diffraction intensity profiles obtained from the AuNi10 and pure Au (reference)
NPs [36]
3.6 Pt MSs on Carbon-Supported Au–Ni NPs for ORR 103

indicating that Au atoms are subjected to compressive strain. This observation agrees
well with the fact that interatomic distance of the bulk Au is larger than that of bulk
Ni. Therefore, when Au and Ni form surface alloy, Au atoms are compressed due
to the lattice mismatch. The result verifies that Au spreads on Ni surface, and the
calculated thickness of Au is about two atomic layers.

Fig. 3.41 Polarization curves obtained with a rotating disk-ring electrode for the ORR on a Pt MS
on carbon-supported AuNi10 NPs in 0.1 M HClO4 solutions. a Ring current versus potential. b Disk
current versus potential
104 3 Pt-MS Electrocatalysts for ORR

3.6.2 ORR Electrocatalytic Properties

Figure 3.41 displays the rotating disk-ring polarization curves for a Pt MS on carbon-
supported AuNi10 NPs. The rotation rates are indicated in the graph, the scan rate
is 10 mV s−1 , the ring potential is 1.27 V, the ring and disk areas are 0.037 and
0.164 cm2 , respectively, and the collection efficiency is 22%. The ring currents are
negligible indicating a complete four-electron reduction of O2 on the disk electrode.
The electrode consists of a Pt MS deposited on 15 nmol of AuNi10 supported by
carbon nanoparticles placed on glassy carbon rotating disk electrode. A very high
activity of a Pt MS on AuNi10 NPs is indicated by a half-wave potential of 0.850 V
and by negligible ring-electrode currents. The K–L plots obtained from the data in
Fig. 3.41 are shown in Fig. 3.42. The electrode potentials are indicated in the graph,
and the insert shows the Tafel plot obtained from the kinetic currents jk obtained
from these plots. For a Pt/AuNi10 surface, a set of parallel straight lines at different
potentials is obtained. The Tafel slope of −92 mV dec−1 fits the polarization curve
obtained from the plot of log jk versus E, given as an insert to Fig. 3.42. This slope

Fig. 3.42 K–L plots at different potentials obtained from the data in Fig. 3.41
3.6 Pt MSs on Carbon-Supported Au–Ni NPs for ORR 105

Fig. 3.43 Comparison of polarization curves for the ORR on Pt (61 nmol cm−2 ) NPs, and a Pt MS
on AuNi10 NPs (15 nmolAu cm−2 ) [69]

is larger than that obtained for Pt/C (−86 mV dec−1 ), indicating less OH adsorption
on the Pt surface.
Figure 3.43 shows the polarization curves for Pt/AuNi10 and Pt10 /C. The electrode
geometric area is 0.164 cm2 . The Pt loading of Pt/AuNi10 /C is about 10 times less
than that of Pt10 /C, while the half-wave potential is still 25 mV higher than the latter
one. A more reasonable comparison is the Pt- and NM-MSA. Figure 3.44 shows this
comparison by using mass-specific kinetic currents at 0.8 and 0.85 V, as presented
in the plot. The Pt-MSA for Pt/AuNi10 is about 20 times higher than that of Pt10 /C.
Moreover, it is still four times higher than that of Pt10 /C in NM-MSA.

Fig. 3.44 Pt- and NM-MSA of Pt (61 nmolPt cm−2 ) NPs, and a Pt MS (7 nmolPt cm−2 ) on AuNi10
NPs (15 nmolAu cm−2 ) expressed as kinetic currents at 0.8 and 0.85 V [69]
106 3 Pt-MS Electrocatalysts for ORR

3.6.3 Fuel Cell Tests

The fuel cell tests are carried out using the fuel cell with the electrodes of 5 cm2 in
area. The cathode is Pt/AuNi10 /C containing 19 µg cm−2 of Pt, 37 µg cm−2 of Au
and 113 µg cm−2 of Ni. The weight percentage of the cathode electrocatalyst is: Pt
1.44%, Au 2.92%, Ni 8.69%, and the total metal content was 13% over carbon. The
anode used in the test is a commercial Pt/C containing 77 µg cm−2 of Pt (20% Pt,
Pt/C, ETEK). Figure 3.45 illustrates the voltage curve and power performance curve
against the cell current at 80 °C. The MEA area is 5 cm2 , the anode is 0.077 mgPt
cm−2 (20% Pt/C, ETEK), and the cathode is 0.169 mgNM cm−2 (19 µg Pt, 37 µg
Au, 0.113 mgNi cm−2 ). The cathode electrocatalyst is Pt: 1.44wt%; Au: 2.92wt%;
Ni: 8.69wt% (13.0% NM/C). The performance of the cathode electrocatalyst is very
promising. At the current of 0.6 A cm−2 (usually the fuel cell operating current),
the cell voltage is as high as 0.67 V, corresponding to a 55% of cell efficiency. The
calculation of this efficiency includes the potential drop coming from the anode,
since the test cannot separate the potential losses from the two electrodes, and the

Fig. 3.45 Operating fuel cell test on the Pt/AuNi10 /C cathode electrocatalyst: cell voltage and
power output as a function of cell current density [69]
3.6 Pt MSs on Carbon-Supported Au–Ni NPs for ORR 107

anode for the testing fuel cell contains significant less Pt than commercial fuel cells.
The maximum power output reaches at 0.8 A cm−2 . The noble-metal mass-specific
power output is calculated at 0.6 A cm−2 . As shown in the graph, for the cathode
alone, the results are 0.05 gPt kW−1 , 0.14 g(Pt+Au) kW−1 ; for the whole cell, they are
0.24 gPt kW−1 , 0.33 g(Pt+Au) kW−1 . The results are promising.

3.6.4 XANES Measurements

In situ XANES measurements are carried out using a specially designed electrochem-
ical cell in order to get the information about the electronic and chemical properties of
Pt in PtML /AuNi10 /C (Fig. 3.46). Figure 3.46 shows the normalized X-ray absorption
intensity near the Pt L3 edge. The inset is the relative change of the peak intensity
(μE − μ0.47 )/μ0.47 for Pt in Pt/Au/Ni/C and in Pt/C as a function of potential obtained
from the spectra. There is no obvious change of the white line up to 0.97 V, and the
peak height is only slightly increased at 1.17 V. The peak height of the white line
at L3 edge is usually taken as an indicator of the d-band orbital vacancy. Assuming
there is no oxidation taking place on Pt surface at 0.47 V double layer region, the
result shows that Pt in PtML /AuNi10 /C is very stable until 0.97 V and is slightly
oxidized at 1.17 V. Furthermore, the inset of Fig. 3.46 displays the relative change
of the absorption intensity as a function of potential for Pt in PtML /AuNi10 /C, in
PtM L/Au/C and in Pt/C, normalized by the intensity at 0.47 V, (μE − μ0.47 )/μ0.47 .
These plots clearly show the easy oxidation of Pt in Pt/C, while the oxidation of Pt
in the PtML /AuNi10 /C is observed only at a much higher potential (e.g., 1.17 V), and
even then it is only minor. Hence, these data explain the higher ORR electrocatalytic
activity of PtML /AuNi10 /C than that of Pt/C.

Fig. 3.46 XANES spectra obtained with Pt/Au/Ni/C at 0.47 and 1.17 V, and with Pt/C at 1.17 V
in 1.0 M HClO4 solutions, demonstrating a shift in Pt oxidation in Pt/Au/Ni/C [36]
108 3 Pt-MS Electrocatalysts for ORR

3.6.5 Discussion

The plot of the kinetic currents for the ORR on the Pt MSs on various substrates at
0.8 V as a function of the calculated d-band center εd shows a volcano-like curve,
with PtML /Pd(111) possessing the maximum ORR electrocatalytic activity [67]. The
Pt MS on Au(111) shows a lower ORR electrocatalytic activity, which is explained
by the strong adsorption of the intermediates in the ORR on this highly stretched
surface. The present result indicates a different interaction of Pt and Au/Ni involving
a lower tensile strain, and thus a weaker adsorption of the intermediates in ORR and
higher reaction rates.

3.6.6 Summary

The Pt MS supported on an Au–Ni alloy has an excellent ORR electrocatalytic

activity. The surface segregation of Au is verified by the asymmetry of the Au peaks
in powder X-ray diffraction data. In agreement with this result, the CV does not show
any Ni dissolution, indicating a good protection by Au. The resulting electrocatalyst
has the activity slightly higher than that of Pt/C and an exceptional Pt-MSA, which
is about 20 times that of a standard Pt/C. These results indicate that the surface-
segregated alloys may be promising supports if tailored to enhance Pt’s ORR elec-
trocatalytic activity. In addition, they indicate a different interaction of Pt and Au
involving a lower tensile strain than that for Pt on Au(111) single crystal, and thus a
weaker adsorption of the intermediates in ORR. Au and Ni alloys seem very suitable
as a support for a Pt MS. The electrocatalyst reduces the noble metal use to a very
low level.

3.7 Pt MSs on Carbon-Supported Pdx Ni1−x NSs for ORR

It is undoubtedly desirable, however, very challenging to appropriately balance the

electrocatalytic activity, electrochemical durability and NM utilization when devel-
oping the Pt-based ORR electrocatalysts. Accordingly, a versatile and effective
strategy that promises the structures of both composition-graded core and MS is
proposed to synthesize highly uniform and sub-10 nm Pdx Ni1−x @Pt nanospheres
(NSs) as high-performance ORR electrocatalysts. Highly uniform and composition-
graded Pdx Ni1−x NSs are previously obtained via a facile one-pot Ni-substitution-
based process, and then Pt MSs are coated onto them through the Cu UPD coupled
with Pt2+ galvanic replacement. Results show that carbon-supported Pdx Ni1−x @Pt
electrocatalysts possess both high electrocatalytic activity and highly efficient
NM utilization for ORR. The optimal Pd0.42 Ni0.58 @Pt/C exhibits 0.61 mA cm−2 ,
3.7 Pt MSs on Carbon-Supported Pdx Ni1−x NSs for ORR 109

0.42 A mg−1 NM and 1.45 A mg−1 Pt @ 0.9 V (vs. RHE) in ASA, NM- and Pt-MSA,
respectively, reaching 2.8, 3.3 and 11.2 times relative to those of the commercial
Pt/C. Moreover, Pd0.42 Ni0.58 @Pt/C also has a satisfactory electrochemical dura-
bility, preserving its high ORR electrocatalytic activity even after 12,000 poten-
tial cycles of the accelerated degradation test (ADT). The synthetic mechanism of
Pdx Ni1−x NS core and Pt MS, as well as their combined effects on the electrocatalytic
activity, electrochemical durability and NM utilization of Pdx Ni1−x @Pt/C for ORR
is comprehensively investigated.

3.7.1 Physicochemical Characterizations of Pdx Ni1−x

and Pdx Ni1−x @Pt NSs

A series of highly uniform and composition-graded Pdx Ni1−x NSs are synthesized
by a facile one-pot Ni-substitution-based process. In the typical synthetic protocol,
oleylamine (OAm) serves as the solvent, trioctylphosphine (TOP) works as the surfac-
tant to stabilize the NS structure [119], while the use of PdBr2 indispensably leads
to the formation of the fine spherical morphology [120]. As illustrated in Fig. 3.47,
preferred nucleation of Ni atoms results from a homemade Ni–TOP complex due to
its low decomposition temperature of ~190 °C [119, 121]. It is observed that the reac-
tant color gradually changes from light-green to dark-brown when the reactants are
heated from ~190 to 245 °C, indicating a continuous nucleation and growth of NSs.
Subsequently, the replacement of part Ni atoms with Pd2+ from PdBr2 is triggered
by the huge difference in redox potential (–0.25 V SHE for Ni/Ni2+ VS. +0.915 V SHE
for Pd/Pd2+ ) combining with a higher temperature [122]. In the meantime, Ni atoms
continue to be deposited onto the surfaces of both Ni atoms and Pd atoms via the
successive decomposition of Ni–TOP complex. These two processes will not stop
until either Ni(acac)2 or PdBr2 is depleted.
The above speculation can be further confirmed by four other similar controlled
synthetic procedures. When Ni(acac)2 is not added, the reactant color keeps hardly
changed even being heated to the temperature as high as 250 °C and maintains

Fig. 3.47 Schematic illustration for the synthetic mechanism of Pdx Ni1−x NS, the difference in
the atomic radius between Pd and Ni is neglected [4]
110 3 Pt-MS Electrocatalysts for ORR

Fig. 3.48 Structural models

for the series of
composition-graded
Pdx Ni1−x NSs [4]

for 2 h. The use of PdBr2 will not lead to the formation of the Pd–TOP complex,
which is supposed to decompose into Pd atoms at ~235 °C, as reported in a previous
work [119]. When further replacing PdBr2 with Pd(acac)2 , the nucleation occurred
at ~150 °C and led to Pd NPs. By this way, the Pd1.00 Ni0.00 NSs with the similar size
and morphology are successfully synthesized via a similar synthetic procedure as
Pdx Ni1−x NSs [4]. However, Ni NSs can still be obtained without PdBr2 or Pd(acac)2
in the precursors due to the low decomposition temperature of ~190 °C for the
homemade Ni–TOP complex. Moreover, when the holding temperature was fixed at
195 °C in the presence of both PdBr2 and Ni(acac)2 , only Ni NSs could be obtained
after a 30 min reaction, proving the preferred nucleation of Ni atoms and the higher
onset temperature of Pd2+ replacement [4]. Therefore, such Ni-substitution-based
process results in the unique composition-graded structure, in which Ni content
decreases while Pd increases along the radial direction from inner core to outer
surface. Furthermore, according to the XPS results for the series of Pdx Ni1−x /C,
the surface Ni atomic ratio increases with that in the bulk composition; thus, the
structural models for the series of Pdx Ni1−x NSs can be speculated as Fig. 3.48 [4].
The structures of the as-synthesized Pdx Ni1–x NSs are analyzed via the STEM–
EDS elemental line-scanning tests. Figure 3.49a and b, respectively, shows the Pd/Ni
elemental distribution of single Pd0.69 Ni0.31 and Pd0.42 Ni0.58 NSs, along the scanning
direction indicated by the yellow arrows in the insets. It is clearly seen that the Pd
and Ni signals arise and descend almost simultaneously, which is obviously different
from the well-defined core@shell structures which possess the clear signal-arising
and -descending position distances between the core and shell element [32, 123].
This enables Ni atoms to influence the surface crystal lattice of Pdx Ni1−x NSs in a
way that the well-defined core@shell structures cannot, and such influence will be
critical to the ORR electrocatalytic activity of Pdx Ni1−x @Pt/C. Moreover, the fitting
curve profile for Ni shows a “volcano” shape, indicating that Ni concentrates in the
core region. By contrast, the fitting curve profile for Pd presents a “valley” shape in
the center, revealing that Pd mostly distributes in the shell [38]. The overall elemental
signal intensity of Pd is stronger than that of Ni for the Pd0.69 Ni0.31 NS, while it is
opposite for the Pd0.42 Ni0.58 NS, which are consistent with the inductively coupled
plasma–optical emission spectrometer (ICP–OES) results [4]. Furthermore, in the
surface region of these two NSs, the elemental signal intensity of Pd is always stronger
than that of Ni, and the signal intensity difference between Pd and Ni decreases as Ni
in the bulk composition increases, confirming the Pd-enriched surfaces of Pdx Ni1−x
3.7 Pt MSs on Carbon-Supported Pdx Ni1−x NSs for ORR 111

Fig. 3.49 STEM–EDS elemental line-scanning results [4]. STEM–EDS elemental line-scanning
results of the as-synthesized single a Pd0.69 Ni0.31 and b Pd0.42 Ni0.58 NSs, the blue and green dash
lines are the fitting curves for the scattered signal intensity dots of Pd and Ni, respectively. The
insets are the HAADF–STEM images integrated with the corresponding fitting curves of signal
intensity dots. c Schematic illustration for the composition-graded nanostructure of Pdx Ni1−x NS

NSs and that the surface Ni atomic ratio increases with Ni in the bulk composi-
tion, which have been previously proved by the XPS results. The above elemental
distribution is consistent with the composition-graded structure of Pdx Ni1−x NS. For
better understanding, schematic illustration for the composition-graded structure of
Pdx Ni1−x NS, in which Ni content decreases while Pd increases along the radial
direction from inner core to outer surface, is presented in Fig. 3.49c.
Figure 3.50a–c shows the typical TEM images of the Pd0.69 Ni0.31 , Pd0.54 Ni0.46
and Pd0.42 Ni0.58 NSs, respectively, and it is clearly observed that the as-synthesized
Pdx Ni1−x NSs show well-defined spherical morphologies and are highly uniform.
The histograms of particle diameter distribution reveal that the series of Pdx Ni1−x
NSs have narrow particle diameter distribution. The average particle diameters
of Pd0.69 Ni0.31 , Pd0.54 Ni0.46 and Pd0.42 Ni0.58 NSs are, respectively, 8.46, 8.14 and
7.30 nm, and the corresponding relative standard deviations are only 6.5, 7.2 and
6.6%. Figure 3.50d demonstrates the uniform distribution of Pd0.69 Ni0.31 NSs on
carbon supports, which is in favor of the Cu UPD without the active surface area loss
caused by the NSs aggregation.
The structure of Pd0.69 Ni0.31 @Pt NS is also analyzed by the scanning TEM–
energy dispersive X-ray spectroscopy (STEM–EDS) elemental line-scanning test.
As shown in Fig. 3.51a, both the signal intensities and shapes for Pd and Ni show
the similar elemental distribution with the as-synthesized Pd0.69 Ni0.31 NS before the
formation of Pt shell. More importantly, the fitting curve profile for Pt shows a flat
112 3 Pt-MS Electrocatalysts for ORR

Fig. 3.50 TEM images [4]. Representative TEM images and the corresponding histograms
of particle diameter distribution (insets) for the as-synthesized a Pd0.69 Ni0.31 , b Pd0.54 Ni0.46 ,
c Pd0.42 Ni0.58 NSs and d Pd0.69 Ni0.31 /C

“plateau” shape with a much lower intensity across the central region and a “bump”
shape in the surface region, indicating that Pt is uniformly distributed in a very
thin shell, thus confirming the successful formation of Pt shell on the Pd0.69 Ni0.31
NS core. To precisely determine the thickness and uniformity of the Pt shell, an
effective analysis that combined element-sensitive electron energy loss spectroscopy
(EELS) with high-angle annular dark field (HAADF) intensity tests is employed
[124]. The background-subtracted EELS signal for the Pd peak around the L3 edge
energy of 3173 eV (red, right) and the HAADF intensity (left) were all measured
in two-dimensional mapping, to determine the sizes and shapes of the Pd0.69 Ni0.31
NS core and the entire NS, respectively. Figure 3.51b shows a nearly 7.75 nm NS
with a 7.24 nm Pd0.69 Ni0.31 NS core, which are obtained from the intensity cutoffs
in the HAADF and EELS intensities, respectively. Combined with the STEM–EDS
line-scanning result, it can be confirmed that the difference of 0.51 nm in diameter
corresponds to the existence of a monolayer thick Pt shell, since the Pt lattice spacing
is ~0.24 nm [124].
According to the XRD patterns in Fig. 3.52, the series of Pdx Ni1−x NSs all possess
the low-crystalline structure. The broaden diffraction peaks observed in the XRD
patterns may be ascribed to the bonding of TOP to Pdx Ni1−x NSs, which lowers the
crystallinity during the growth of NSs [124]. It is also noted that, with the increase
in Ni content of the Pdx Ni1−x NSs, the diffraction peak shifts slightly toward higher
2θ angle. This can be attributed to the increasing partial substitution of Pd atoms
3.7 Pt MSs on Carbon-Supported Pdx Ni1−x NSs for ORR 113

Fig. 3.51 STEM–EDS elemental line-scanning results and EELS analyses [4]. a STEM–EDS
elemental line-scanning result for the single Pd0.69 Ni0.31 @Pt NS. b Its two-dimensional mapping
results of the HAADF intensity (left) and the background-subtracted Pd EELS signal around the
L3 edge energy of 3173 eV (red, right) with 0.44 nm pixel−1 resolution, the scale bars in the insets
are 2 nm

with Ni atoms based on Vegard’s law, since Ni (1.24 Å) has a smaller atomic radius
than Pd (1.39 Å) [125, 126]. It is believed that such shifting can lead to the increased
surface lattice contraction of Pdx Ni1−x NSs [18, 126, 127].
114 3 Pt-MS Electrocatalysts for ORR

Fig. 3.52 XRD patterns for the series of Pdx Ni1−x /C, the dash lines represent the diffraction peak
positions, respectively [4]

3.7.2 Greatly Enhanced Pd Utilization for the Cu UPD

on Pdx Ni1−x NSs

The Cu MSs can be underpotentially deposited onto the surfaces of only noble metals,
including Pd, Au, Pt and so on [23, 36, 58, 128]. The use of the composition-graded
Pdx Ni1−x NSs as the substrate can greatly reduce the amount of noble metals for
Cu UPD, thus quite critical for improving the eventual NM utilization for ORR.
Figure 3.53 presents the CVs for the Cu UPD on the surfaces of carbon-supported
Pd1.00 Ni0.00 , Pd0.69 Ni0.31 , Pd0.54 Ni0.46 and Pd0.42 Ni0.58 NSs, respectively. The inset

Fig. 3.53 CVs of the Cu UPD for the series of Pdx Ni1−x /C [4]
3.7 Pt MSs on Carbon-Supported Pdx Ni1−x NSs for ORR 115

displays the comparison for their normalized amounts of underpotentially deposited

Cu atoms with respect to unit mass of Pd, which are calculated from the Cu adsorption
region. It shows that Cu MSs can be deposited and removed from the Pdx Ni1−x NSs’
surfaces during the potential cycling, which is positive to the Cu bulk deposition
potential (0.32 V) in this case. Differing from the previously reported work [14, 36,
37, 129, 130], no evident sharp peak is observed in the CV curves for the Cu UPD.
It is possibly ascribed to the low crystallinity of Pdx Ni1−x NSs, thus weakening the
electrochemical characteristic peaks. The amount of the deposited Cu atoms can
be calculated from the Cu adsorption region in the CVs as denoted in Fig. 3.53.
The inset in Fig. 3.53 shows the normalized amounts of the deposited Cu atoms for
the series of Pdx Ni1−x /C with respect to unit mass of Pd (noted as Cu/Pd, which
is normalized by that of the Pd1.00 Ni0.00 /C). It can be seen that, with the increase
in Ni content, the value of Cu/Pd for Pdx Ni1−x /C increases accordingly, indicating
an enhancement in the Pd utilization for the Cu UPD. This greatly enhanced Pd
utilization is mainly attributed to the combined effects of the composition-graded
structure (Pd-enriched surface and Ni-enriched core) and the decrease of Pd content
as the Ni content increases in the Pdx Ni1−x NSs.

3.7.3 ORR Electrocatalytic Properties

The ORR electrocatalytic activity of the series of Pdx Ni1−x @Pt/C is investigated in
O2 -saturated 0.1 M HClO4 solutions and compared with that for the commercial
Pt/C. Figure 3.54a–c shows the comparisons of the Tafel plots in terms of the ASA
(mA cm−2 ), NM-(A mg−1 NM ) and Pt-MSA (A mg−1 Pt ), respectively. All of the
ORR electrocatalytic activity are calculated without iR-correction or background
calibration, which are visually presented in Fig. 3.54d by the bar charts. The series
of Pdx Ni1−x @Pt/C exhibit significantly enhanced ORR electrocatalytic activity than
the commercial Pt/C. Moreover, there exists a distinctive composition dependence
on the Ni content in Pdx Ni1−x @Pt/C. In particular, the Pd0.42 Ni0.58 @Pt/C shows the
optimal ORR electrocatalytic activity among all of the electrocatalysts, and its ASA
reaches 0.61 mA cm−2 , while 0.42 A mg−1 NM for the NM-MSA, and 1.45 A mg−1 Pt
for the Pt-MSA, which are approximately 2.8, 3.3 and 11.2 times of those for the
commercial Pt/C, respectively.
Such distinctive composition-dependent enhancement in ASA and Pt-MSA is
mainly attributed to the increased surface lattice contraction of Pd substrate as the Ni
content increases. As illustrated in Fig. 3.55, the increased surface lattice contraction
of Pd substrate can lead to an increased substrate-induced surface strain of the Pt
MSs [23, 26, 36, 69], thus resulting in the downshift of the strain-induced d-band
center of Pt [69, 131]. This downshift of Pt d-band center will further cause a weaker
Pt–O/OH bonding, thereby decreasing the intermediate O/OH species adsorption on
the Pt MSs, thus preserving more active sites for ORR and eventually resulting in
the enhancement in ASA [23, 127, 132]. The Pt-MSA directly relates to the ASA
because of the Pt-MS structure, which ideally guarantees the involvement of every
116 3 Pt-MS Electrocatalysts for ORR

Fig. 3.54 ORR electrocatalytic activity [4]. ORR electrocatalytic activity comparisons (Tafel plots)
among the series of Pdx Ni1−x @Pt/C and the commercial Pt/C in a ASA, b NM-, and c Pt-MSA.
d Overall ORR electrocatalytic activity comparisons in the bar charts

Fig. 3.55 Schematic illustrations for the Pt MSs ideally deposited on the surfaces of the purified
Pd1.00 Ni0.00 , Pd0.69 Ni0.31 , Pd0.54 Ni0.46 and Pd0.42 Ni0.58 NSs, respectively [4]

Pt atom in ORR. Besides, there also exists increasingly strong “synergetic effects” as
the Ni content increases, which modifies the electronic structures of the Pt MSs. This
“synergetic effects” also plays an additional role in improving the ASA, because the
strong “synergetic effects” can also lower the energy of Pt d-band center [23, 34,
133, 134]. Additionally, for the enhancement in NM-MSA, it can also be ascribed
to the greatly enhanced Pd utilization for Cu UPD, which leads to an even higher
NM-MSA than that of the Pd3 Co@Pt/C in a previously reported work [14], even
3.7 Pt MSs on Carbon-Supported Pdx Ni1−x NSs for ORR 117

Fig. 3.56 Area-normalized XPS spectra of the Pt 4f7/2 for the commercial Pt/C and the series of
Pdx Ni1−x @Pt/C, respectively, which are normalized by the corresponding total Pt 4f photoelectron
intensity (area) [4]

though the particle size of the Pd3 Co@Pt/C (4.6 nm) in the previous work is much
smaller.
The XPS spectrum of Pt 4f7/2 can be utilized to probe the electronic structure of
Pt, and evaluate its d-band center shifting [14]. Figure 3.56 shows the Pt 4f7/2 XPS
spectra for the commercial Pt/C and the series of Pdx Ni1−x @Pt/C, which are obtained
from the Pt 4f XPS spectra, respectively, and further normalized by the corresponding
total Pt 4f photoelectron intensity (area) [4, 14]. A clear upshift of the binding energy
(BE) in the Pt 4f7/2 of Pdx Ni1−x @Pt/C can be observed compared with the commer-
cial Pt/C, which is also as a function of the Ni composition in the Pdx Ni1−x @Pt/C.
The Pt 4f7/2 BEs are 71.10, 71.22, 71.30, 71.45 and 71.57 eV, respectively, for
the commercial Pt/C, Pd1.00 Ni0.00 @Pt/C, Pd0.69 Ni0.31 @Pt/C, Pd0.54 Ni0.46 @Pt/C and
Pd0.42 Ni0.58 @Pt/C. The upshift of the BE in Pt 4f7/2 implies a downshift in the d-band
center of Pt [14]. This BE upshift trend in the Pt 4f7/2 for the series of Pdx Ni1−x @Pt/C
directly confirms the aforementioned downshift trend of Pt d-band center as the Ni
composition in the Pdx Ni1−x @Pt/C increases.
The electrochemical durability under highly acidic conditions of the commercial
Pt/C and Pd0.42 Ni0.58 @Pt/C are both tested utilizing the ADT protocol [4]. According
to Fig. 3.57a and c, after, respectively, 6000 and 12,000 potential cycles, the ECSAs
for the commercial Pt/C remain 91% and 83%; while still 97% and 91% for the
Pd0.42 Ni0.58 @Pt/C. The loss in ECSA for Pd0.42 Ni0.58 @Pt/C can be mainly attributed
to the preferential dissolution of Ni atoms in the Pd0.42 Ni0.58 substrate caused by
the harsh electrochemical and highly acidic environments, which can lead to the
restructuring and further dissolution of the Pt MS.
118 3 Pt-MS Electrocatalysts for ORR

Fig. 3.57 ORR electrochemical durability [4]. Evolution of CV and the corresponding ORR polar-
ization curves for the a, b Pd0.42 Ni0.58 @Pt/C and c, d commercial Pt/C after every 6000 potential
cycles of the ADT

As shown in Fig. 3.57b and d, the ORR electrocatalytic activity for the commercial
Pt/C directly decreases to 0.19 mA cm−2 and 0.11 A mg−1 NM/Pt after 6000 poten-
tial cycles, and 0.17 mA cm−2 and 0.09 A mg−1 NM/Pt after 12,000 potential cycles,
remaining only 77 and 69% for the ASA and Pt-MSA, respectively. By contrast,
after the initial 6000 potential cycles of the ADT, the ASA, NM- and Pt-MSA for
Pd0.42 Ni0.58 @Pt/C rise to 0.73 mA cm−2 , 0.48 A mg−1 NM and 1.65 A mg−1 Pt , respec-
tively, while decrease to 0.69 mA cm–2 , 0.43 A mg−1 NM and 1.46 A mg−1 Pt after
12,000 potential cycles. It is noted that, for better comparison, the calculations of
the MSA after the ADT are all based on the corresponding initial NM loading. The
enhancement in ORR electrocatalytic activity after the initial 6000 potential cycles
may be attributed to the restructuring of the Pt MS due to the slight dissolution of Ni
atoms in the substrate. However, after 12,000 potential cycles, as shown in Fig. 3.57a,
the shape of CV curve changes evidently, and the ECSA also has a relatively evident
decline compared to that after the initial 6000 potential cycles, indicating the disso-
lution of the Pt MSs triggered by the further dissolution of Ni atoms in the substrate.
The loss of the Pt MS counteracts the effect of its restructuring, thus resulting in
the decline in ORR electrocatalytic activity. By contrast, there is no such “volcano”
trend for the ORR electrocatalytic activity of the commercial Pt/C, due to its solid
and single-metal structure.
3.7 Pt MSs on Carbon-Supported Pdx Ni1−x NSs for ORR 119

Therefore, compared to the commercial Pt/C, Pd0.42 Ni0.58 @Pt/C still preserves
its high ORR electrocatalytic activity even after 12,000 potential cycles of the ADT,
confirming its satisfactory electrochemical durability in the highly acidic environ-
ments, which can be attributed to the enclosure of the Pt MSs which retard the
dissolution of the Ni atoms in the Pdx Ni1−x cores.

3.7.4 Summary

In summary, a versatile and effective strategy that promises the structures of

both composition-graded core and MS is developed to prepare the series of
Pdx Ni1−x @Pt/C. Firstly, a facile one-pot Ni-substitution-based process is employed
to synthesize the highly uniform and composition-graded Pdx Ni1−x NSs. Then, Pt
MSs are coated onto them through the Cu UPD coupled with Pt2+ galvanic replace-
ment. The composition-graded structure of Pdx Ni1−x NS results from the preferential
nucleation of Ni atoms and the following replacement of partial Ni atoms with Pd2+
from PdBr2 . Benefiting from the combined effects of composition-graded Pdx Ni1–x
NS core and Pt MS, the series of Pdx Ni1−x @Pt/C exhibits both high electrocatalytic
activity and highly efficient NM utilization for ORR. Especially, the most active
Pd0.42 Ni0.58 @Pt/C shows 0.61 mA cm−2 , 0.42 A mg−1 NM and 1.45 A mg−1 Pt in ASA,
NM- and Pt-MSA, respectively, reaching an enhancement of 2.8, 3.3 and 11.2 times
when benchmarked against the commercial Pt/C. Besides, Pd0.42 Ni0.58 @Pt/C also
maintains its high ORR electrocatalytic activity under highly acidic conditions even
after 12,000 potential cycles of the ADT, demonstrating its satisfactory electrochem-
ical durability. This strategy has been demonstrated to be an effective approach for
developing high-performance ORR electrocatalysts, which can be further expanded
to the synthesis of various kinds of nanomaterials for different electrocatalytic
applications.

References

1. Chen A, Holt-Hindle P (2010) Platinum-based nanostructured materials: synthesis, properties,

and applications. Chem Rev 110(6):3767–3804
2. Debe MK (2012) Electrocatalyst approaches and challenges for automotive fuel cells. Nature
486(7401):43–51
3. Luo L, Fu C, Yang F et al (2019) Composition-graded Cu–Pd nanospheres with Ir-doped
surfaces on N-doped porous graphene for highly efficient ethanol electro-oxidation in alkaline
media. ACS Catal 10(2):1171–1184
4. Luo L, Zhu F, Tian R et al (2017) Composition-graded Pdx Ni1−x nanospheres with Pt mono-
layer shells as high-performance electrocatalysts for oxygen reduction reaction. ACS Catal
7(8):5420–5430
5. Nie Y, Li L, Wei Z (2015) Recent advancements in Pt and Pt-free catalysts for oxygen reduction
reaction. Chem Soc Rev 44(8):2168–2201
120 3 Pt-MS Electrocatalysts for ORR

6. Wang YJ, Zhao N, Fang B et al (2015) Carbon-supported Pt-based alloy electrocatalysts for
the oxygen reduction reaction in polymer electrolyte membrane fuel cells: particle size, shape,
and composition manipulation and their impact to activity. Chem Rev 115(9):3433–3467
7. Porter NS, Wu H, Quan Z et al (2013) Shape-control and electrocatalytic activity-enhancement
of Pt-based bimetallic nanocrystals. Acc Chem Res 46(8):1867–1877
8. Gasteiger HA, Kocha SS, Sompalli B et al (2005) Activity benchmarks and requirements
for Pt, Pt-alloy, and non-Pt oxygen reduction catalysts for PEMFCs. Appl Catal B Environ
56(1–2):9–35
9. Shao M, Chang Q, Dodelet JP et al (2016) Recent advances in electrocatalysts for oxygen
reduction reaction. Chem Rev 116(6):3594–3657
10. Bing Y, Liu H, Zhang L et al (2010) Nanostructured Pt-alloy electrocatalysts for PEM fuel
cell oxygen reduction reaction. Chem Soc Rev 39(6):2184–2202
11. Wu J, Yang H (2013) Platinum-based oxygen reduction electrocatalysts. Acc Chem Res
46(8):1848–1857
12. Zhang H, Jin M, Xiong Y et al (2013) Shape-controlled synthesis of Pd nanocrystals and their
catalytic applications. Acc Chem Res 46(8):1783–1794
13. Stamenkovic VR, Fowler B, Mun BS et al (2007) Improved oxygen reduction activity on
Pt3 Ni(111) via increased surface site availability. Science 315(5811):493–497
14. Wang JX, Inada H, Wu L et al (2009) Oxygen reduction on well-defined core-shell nanocata-
lysts: particle size, facet, and Pt shell thickness effects. J Am Chem Soc 131(47):17298–17302
15. Anwar MT, Yan X, Shen S et al (2017) Enhanced durability of Pt electrocatalyst with tantalum
doped titania as catalyst support. Int J Hydrogen Energy 42(52):30750–30759
16. Cai B, Hubner R, Sasaki K et al (2018) Core-shell structuring of pure metallic aerogels towards
highly efficient platinum utilization for the oxygen reduction reaction. Angew Chem Int Ed
57(11):2963–2966
17. Chen G, Kuttiyiel KA, Su D et al (2016) Oxygen reduction kinetics on Pt monolayer shell
highly affected by the structure of bimetallic AuNi cores. Chem Mater 28(15):5274–5281
18. Choi SI, Lee SU, Kim WY et al (2012) Composition-controlled PtCo alloy nanocubes with
tuned electrocatalytic activity for oxygen reduction. ACS Appl Mater Inter 4(11):6228–6234
19. Choi SI, Shao M, Lu N et al (2014) Synthesis and characterization of Pd@Pt-Ni core-shell
octahedra with high activity toward oxygen reduction. ACS Nano 8(10):10363–10371
20. Gamler JTL, Leonardi A, Ashberry HM et al (2019) Achieving highly durable random alloy
nanocatalysts through intermetallic cores. ACS Nano 13(4):4008–4017
21. Jayasayee K, Veen JARV, Manivasagam TG et al (2012) Oxygen reduction reaction (ORR)
activity and durability of carbon supported PtM (Co, Ni, Cu) alloys: influence of particle size
and non-noble metals. Appl Catal B Environ 111–112:515–526
22. Karan HI, Sasaki K, Kuttiyiel K et al (2012) Catalytic activity of platinum monolayer
on iridium and rhenium alloy nanoparticles for the oxygen reduction reaction. ACS Catal
2(5):817–824
23. Koenigsmann C, Santulli AC, Gong K et al (2011) Enhanced electrocatalytic performance of
processed, ultrathin, supported Pd-Pt core-shell nanowire catalysts for the oxygen reduction
reaction. J Am Chem Soc 133(25):9783–9795
24. Kuttiyiel KA, Sasaki K, Choi Y et al (2012) Bimetallic IrNi core platinum monolayer shell
electrocatalysts for the oxygen reduction reaction. Energy Environ Sci 5(1):5297–5304
25. Li D, Wang C, Strmcnik DS et al (2014) Functional links between Pt single crystal morphology
and nanoparticles with different size and shape: the oxygen reduction reaction case. Energy
Environ Sci 7(12):4061–4069
26. Liu H, Koenigsmann C, Adzic RR et al (2014) Probing ultrathin one-dimensional Pd–Ni
nanostructures as oxygen reduction reaction catalysts. ACS Catal 4(8):2544–2555
27. Nan H, Tian X, Luo J et al (2016) A core–shell Pd1 Ru1 Ni2 @Pt/C catalyst with a ternary alloy
core and Pt monolayer: enhanced activity and stability towards the oxygen reduction reaction
by the addition of Ni. J Mater Chem A 4(3):847–855
28. Park J, Zhang L, Choi SI et al (2015) Atomic layer-by-layer deposition of platinum on palla-
dium octahedra for enhanced catalysts toward the oxygen reduction reaction. ACS Nano
9(3):2635–2647
References 121

29. Shen L-L, Zhang G-R, Miao S et al (2016) Core–shell nanostructured Au@Nim Pt2 elec-
trocatalysts with enhanced activity and durability for oxygen reduction reaction. ACS Catal
6(3):1680–1690
30. Tan X, Prabhudev S, Kohandehghan A et al (2015) Pt–Au–Co alloy electrocatalysts demon-
strating enhanced activity and durability toward the oxygen reduction reaction. ACS Catal
5(3):1513–1524
31. Tian R, Shen S, Zhu F et al (2018) Icosahedral Pt-Ni nanocrystalline electrocatalyst: growth
mechanism and oxygen reduction activity. ChemSusChem 11(6):1015–1019
32. Tian X, Luo J, Nan H et al (2016) Transition metal nitride coated with atomic layers of Pt as
a low-cost, highly stable electrocatalyst for the oxygen reduction reaction. J Am Chem Soc
138(5):1575–1583
33. Tian X, Tang H, Luo J et al (2017) High-performance core–shell catalyst with nitride nanopar-
ticles as a core: well-defined titanium copper nitride coated with an atomic Pt layer for the
oxygen reduction reaction. ACS Catal 7(6):3810–3817
34. Wang X, Vara M, Luo M et al (2015) Pd@Pt core-shell concave decahedra: a class of catalysts
for the oxygen reduction reaction with enhanced activity and durability. J Am Chem Soc
137(47):15036–15042
35. Zhang H, Jin M, Wang J et al (2011) Synthesis of Pd-Pt bimetallic nanocrystals with a
concave structure through a bromide-induced galvanic replacement reaction. J Am Chem Soc
133(15):6078–6089
36. Zhang J, Lima FH, Shao MH et al (2005) Platinum monolayer on nonnoble metal-noble metal
core-shell nanoparticle electrocatalysts for O2 reduction. J Phys Chem B 109(48):22701–
22704
37. Zhang J, Mo Y, Vukmirovic MB et al (2004) Platinum monolayer electrocatalysts for O2
reduction: Pt monolayer on Pd(111) and on carbon-supported Pd nanoparticles. J Phys Chem
B 108(30):10955–10964
38. Zhang J, Sasaki K, Sutter E et al (2007) Stabilization of platinum oxygen-reduction
electrocatalysts using gold clusters. Science 315(5809):220–222
39. Zhang N, Feng Y, Zhu X et al (2017) Superior bifunctional liquid fuel oxidation and oxygen
reduction electrocatalysis enabled by PtNiPd core-shell nanowires. Adv Mater 29(7):1603774
40. Zhang S, Hao Y, Su D et al (2014) Monodisperse core/shell Ni/FePt nanoparticles and their
conversion to Ni/Pt to catalyze oxygen reduction. J Am Chem Soc 136(45):15921–15924
41. Zhao X, Chen S, Fang Z et al (2015) Octahedral [email protected] Ni core-shell nanocrystals with
ultrathin PtNi alloy shells as active catalysts for oxygen reduction reaction. J Am Chem Soc
137(8):2804–2807
42. Zhou M, Wang H, Elnabawy AO et al (2019) Facile one-pot synthesis of Pd@Pt1L octahedra
with enhanced activity and durability toward oxygen reduction. Chem Mater 31(4):1370–1380
43. Zhou M, Wang H, Vara M et al (2016) Quantitative analysis of the reduction kinetics respon-
sible for the one-pot synthesis of Pd-Pt bimetallic nanocrystals with different structures. J Am
Chem Soc 138(37):12263–12270
44. Shen S, Cheng X, Wang C et al (2017) Exploration of significant influences of the operating
conditions on the local O2 transport in proton exchange membrane fuel cells (PEMFCs). Phys
Chem Chem Phys 19(38):26221–26229
45. Zhang L, Zhu S, Chang Q et al (2016) Palladium–platinum core–shell electrocatalysts for
oxygen reduction reaction prepared with the assistance of citric acid. ACS Catal 6(6):3428–
3432
46. Herrero E, Buller LJ, Abruna HD (2001) Underpotential deposition at single crystal surfaces
of Au, Pt Ag and other materials. Chem Rev 101(7):1897–1930
47. Kokkinidis G (1986) Underpotential deposition and electrocatalysis. J Electroanal Chem Inter
Electrochem 201(2):217–236
48. Kolb DM, Przasnyski M, Gerischer H (1974) Underpotential deposition of metals and work
function differences. J Electroanal Chem Inter Electrochem 54(1):25–38
49. Pangarov N (1983) Thermodynamics of electrochemical phase formation and underpotential
metal deposition. Electrochim Acta 28(6):763–775
122 3 Pt-MS Electrocatalysts for ORR

50. Xu Z, Chen Y, Zhang Z et al (2015) Progress of research on underpotential deposition-I:

theory of underpotential deposition. Acta Phys Chim Sin 31(7):1219–1230
51. Von Hevesy G (1912) Radioaktive methoden in der elektrochemie. Zeitschrift für Elektro-
chemie und angewandte physikalische Chemie 18(13):546–549
52. Alanyalıoğlu M, Bayrakçeken F, Demir Ü (2009) Preparation of PbS thin films: a new
electrochemical route for underpotential deposition. Electrochim Acta 54(26):6554–6559
53. Herzog G, Arrigan DW (2003) Comparison of 2-mercaptoethane sulfonate and mercaptoacetic
acid disorganized monolayer-coated electrodes for the detection of copper via underpotential
deposition-stripping voltammetry. Electroanalysis 15(15–16):1302–1306
54. Herzog G, Arrigan DW (2005) Determination of trace metals by underpotential deposition–
stripping voltammetry at solid electrodes. TrAC Trend Anal Chem 24(3):208–217
55. Lima FHB, Zhang J, Shao MH et al (2007) Catalytic activity−d-band center correlation for
the O2 reduction reaction on platinum in alkaline solutions. J Phys Chem C 111(1):404–410
56. Lima FHB, Zhang J, Shao MH et al (2007) Pt monolayer electrocatalysts for O2 reduction:
PdCo/C substrate-induced activity in alkaline media. J Solid State Electrochem 12(4):399–407
57. Lin S-Y, Tsai T-K, Lin C-M et al (2002) Structures of self-assembled monolayers of n-alkanoic
acids on gold surfaces modified by underpotential deposition of silver and copper: odd−even
effect. Langmuir 18(14):5473–5478
58. Luo LX, Shen SY, Zhu FJ et al (2016) Formic acid oxidation by Pd monolayers on Pt3 Ni
nanocubes. Acta Phys Chim Sin 32(1):337–342
59. Nişancı FB, Öznülüer T, Demir Ü (2013) Photoelectrochemical properties of nanostructured
ZnO prepared by controlled electrochemical underpotential deposition. Electrochim Acta
108:281–287
60. Oyamatsu D, Kanemoto H, Kuwabata S et al (2001) Nanopore preparation in self-assembled
monolayers of alkanethiols with use of the selective desorption technique assisted by
underpotential deposition of silver and copper. J Electroanal Chem 497(1–2):97–105
61. Popov B, Zheng G, White RE (1994) The underpotential deposition of zinc for mitigation of
hydrogen absorption and penetration into HY-130 steel. Corros Sci 36(12):2139–2153
62. Şişman İ, Demir Ü (2011) Electrochemical growth and characterization of size-quantized
CdTe thin films grown by underpotential deposition. J Electroanal Chem 651(2):222–227
63. Vukmirovic MB, Zhang J, Sasaki K et al (2007) Platinum monolayer electrocatalysts for
oxygen reduction. Electrochim Acta 52(6):2257–2263
64. Xu Z, Qi DM, Jiang L et al (2015) Progress of research on underpotential deposition-
II: research techniques and application of underpotential deposition. Acta Phys Chim Sin
31(7):1231–1250
65. Zhang J, Vukmirovic MB, Sasaki K et al (2005) Mixed-metal Pt monolayer electrocatalysts
for enhanced oxygen reduction kinetics. J Am Chem Soc 127(36):12480–12481
66. Zhang J, Vukmirovic MB, Sasaki K et al (2005) Platinum monolayer electrocatalysts for
oxygen reduction: effect of substrates, and long-term stability. J Serb Chem Soc 70(3):513–525
67. Zhang J, Vukmirovic MB, Xu Y et al (2005) Controlling the catalytic activity of platinum-
monolayer electrocatalysts for oxygen reduction with different substrates. Angew Chem Int
Ed 44(14):2132–2135
68. Zheng G, Popov BN, White RE (1994) Use of underpotential deposition of zinc to mitigate
hydrogen absorption into Monel K500. J Electrochem Soc 141(5):1220
69. Adzic RR, Zhang J, Sasaki K et al (2007) Platinum monolayer fuel cell electrocatalysts. Top
Catal 46(3–4):249–262
70. Marković N, Gasteiger H, Grgur B et al (1999) Oxygen reduction reaction on Pt (111): effects
of bromide. J Electroanal Chem 467(1–2):157–163
71. Mukerjee S, Srinivasan S, Soriaga MP et al (1995) Effect of preparation conditions of Pt alloys
on their electronic, structural, and electrocatalytic activities for oxygen reduction-XRD, XAS,
and electrochemical studies. J Phys Chem 99(13):4577–4589
72. Macquarrie D, Chorkendorff I, Niemantsverdriet JW (2005) Concepts of modern catalysis
and kinetics. Appl Organomet Chem 19(5):696–696
References 123

73. Nørskov JK, Bligaard T, Logadottir A et al (2002) Universality in heterogeneous catalysis. J

Catal 209(2):275–278
74. Somorjai GA, Li Y (2010) Introduction to surface chemistry and catalysis. Wiley, Hoboken
75. Clouser S, Huang J, Yeager E (1993) Temperature dependence of the Tafel slope for oxygen
reduction on platinum in concentrated phosphoric acid. J Appl Electrochem 23(6):597–605
76. Sidik RA, Anderson AB (2002) Density functional theory study of O2 electroreduction when
bonded to a Pt dual site. J Electroanal Chem 528(1–2):69–76
77. Yeager E, Razaq M, Gervasio D et al (1993) Dioxygen reduction in various acid electrolytes.
Department of Chemistry, Case Western Reserve University, Cleveland, OH
78. Mavrikakis M, Hammer B, Nørskov JK (1998) Effect of strain on the reactivity of metal
surfaces. Phys Rev Lett 81(13):2819
79. Kitchin J, Nørskov JK, Barteau M et al (2004) Modification of the surface electronic
and chemical properties of Pt(111) by subsurface 3d transition metals. J Chem Phys
120(21):10240–10246
80. Chierchie T, Mayer C (1988) Voltammetric study of the underpotential deposition of copper
on polycrystalline and single crystal palladium surfaces. Electrochim Acta 33(3):341–345
81. Brankovic S, Wang J, Adzic R (2001) Metal monolayer deposition by replacement of metal
adlayers on electrode surfaces. Surf Sci 474(1–3):L173–L179
82. Anastasijević N, Vesović V, Adžić R (1987) Determination of the kinetic parameters of the
oxygen reduction reaction using the rotating ring-disk electrode: part I: theory. J Electroanal
Chem Inter Electrochem 229(1–2):305–316
83. Haynes WM (2014) CRC handbook of chemistry and physics. CRC Press, Boca Raton
84. Markovic NM, Gasteiger HA, Ross PN Jr (1995) Oxygen reduction on platinum low-index
single-crystal surfaces in sulfuric acid solution: rotating ring-Pt(hkl) disk studies. J Phys Chem
99(11):3411–3415
85. Wang C, Daimon H, Onodera T et al (2008) A general approach to the size- and shape-
controlled synthesis of platinum nanoparticles and their catalytic reduction of oxygen. Angew
Chem Int Ed 47(19):3588–3591
86. Paulus U, Wokaun A, Scherer G et al (2002) Oxygen reduction on high surface area Pt-based
alloy catalysts in comparison to well defined smooth bulk alloy electrodes. Electrochim Acta
47(22–23):3787–3798
87. Stamenković V, Schmidt T, Ross P et al (2002) Surface composition effects in electrocatalysis:
kinetics of oxygen reduction on well-defined Pt3 Ni and Pt3 Co alloy surfaces. J Phys Chem B
106(46):11970–11979
88. Stamenković V, Schmidt T, Ross P et al (2003) Surface segregation effects in electrocatalysis:
kinetics of oxygen reduction reaction on polycrystalline Pt3 Ni alloy surfaces. J Electroanal
Chem 554:191–199
89. Toda T, Igarashi H, Watanabe M (1999) Enhancement of the electrocatalytic O2 reduction on
Pt–Fe alloys. J Electroanal Chem 460(1–2):258–262
90. Benfield RE (1992) Mean coordination numbers and the non-metal–metal transition in
clusters. J Chem Soc Faraday Trans 88(8):1107–1110
91. Peuckert M, Yoneda T, Dalla Betta R et al (1986) Oxygen reduction on small supported
platinum particles. J Electrochem Soc 133(5):944–947
92. Nørskov J (2001) Catalysis calculations and concepts. Adv Catal 45:71
93. Baldauf M, Kolb D (1996) Formic acid oxidation on ultrathin Pd films on Au(hkl) and Pt(hkl)
electrodes. J Phys Chem 100(27):11375–11381
94. Naohara H, Ye S, Uosaki K (2000) Electrocatalytic reactivity for oxygen reduction at epitax-
ially grown Pd thin layers of various thickness on Au(111) and Au(100). Electrochim Acta
45(20):3305–3309
95. Rodriguez J (1996) Physical and chemical properties of bimetallic surfaces. Surf Sci Rep
24(7–8):223–287
96. Ruckman MW, Strongin M (1994) Monolayer metal films on metallic surfaces: correlation
between electronic structure and molecular chemisorption. Acc Chem Res 27(8):250–256
124 3 Pt-MS Electrocatalysts for ORR

97. Schmidt T, Stamenkovic V, Arenz M et al (2002) Oxygen electrocatalysis in alkaline elec-

trolyte: Pt(hkl), Au(hkl) and the effect of Pd-modification. Electrochim Acta 47(22–23):3765–
3776
98. Meitzner G, Via G, Lytle F et al (1992) Analysis of X-ray absorption edge data on metal
catalysts. J Phys Chem 96(12):4960–4964
99. Toda T, Igarashi H, Watanabe M (1998) Role of electronic property of Pt and Pt alloys on
electrocatalytic reduction of oxygen. J Electrochem Soc 145(12):4185
100. Christoffersen E, Liu P, Ruban A et al (2001) Anode materials for low-temperature fuel cells:
a density functional theory study. J Catal 199(1):123–131
101. de Mongeot FB, Scherer M, Gleich B et al (1998) CO adsorption and oxidation on bimetallic
Pt/Ru(0001) surfaces—a combined STM and TPD/TPR study. Surf Sci 411(3):249–262
102. Sasaki K, Mo Y, Wang J et al (2003) Pt submonolayers on metal nanoparticles—novel
electrocatalysts for H2 oxidation and O2 reduction. Electrochim Acta 48(25–26):3841–3849
103. Climent V, Marković N, Ross P (2000) Kinetics of oxygen reduction on an epitaxial film of
palladium on Pt(111). J Phys Chem B 104(14):3116–3120
104. Adzic R (1998) Recent advances in the kinetics of oxygen reduction. Electrocatalysis 197
105. Tarasevich M, Sadkowski A, Yeager E (1983) Oxygen electrochemistry. In: Comprehensive
treatise of electrochemistry. Springer, Berlin, pp 301–398
106. Damjanovic A, Genshaw M, Bockris JOM (1966) Distinction between intermediates produced
in main and side electrodic reactions. J Chem Phys 45(11):4057–4059
107. Scherson D (1992) Structural effects in electrocatalysis and oxygen electrochemistry:
workshop. Electrochem Soc
108. Uribe F, Wilson M, Springer T et al (1992) Structural effects in electrocatalysis and oxygen
electrochemistry. In: Scherson DD, Tryk D, Daroux M, Xing X (eds), p. 494
109. Albu TV, Anderson AB (2001) Studies of model dependence in an ab initio approach to
uncatalyzed oxygen reduction and the calculation of transfer coefficients. Electrochim Acta
46(19):3001–3013
110. Wang J, Markovic N, Adzic R (2004) Kinetic analysis of oxygen reduction on Pt(111) in
acid solutions: intrinsic kinetic parameters and anion adsorption effects. J Phys Chem B
108(13):4127–4133
111. Mukerjee S, Srinivasan S, Soriaga MP et al (1995) Role of structural and electronic properties
of Pt and Pt alloys on electrocatalysis of oxygen reduction: an in situ XANES and EXAFS
investigation. J Electrochem Soc 142(5):1409–1422
112. Mrozek MF, Xie Y, Weaver MJ (2001) Surface-enhanced Raman scattering on uniform
platinum-group overlayers: preparation by redox replacement of underpotential-deposited
metals on gold. Anal Chem 73(24):5953–5960
113. Massalski TB, Murray J, Bennett L et al (1986) Binary alloy phase diagrams. American
Society of Metals, Metals Park, OH
114. Nielsen LP, Besenbacher F, Stensgaard I et al (1993) Initial growth of Au on Ni(110): surface
alloying of immiscible metals. Phys Rev Lett 71(5):754
115. Boerma D, Dorenbos G, Wheatley G et al (1994) Atomic positions of Au atoms on a Ni(110)
surface. Surf Sci 307:674–679
116. Molenbroek AM, Nørskov JK, Clausen BS (2001) Structure and reactivity of Ni−Au
nanoparticle catalysts. J Phys Chem B 105(23):5450–5458
117. Ruban A, Skriver HL, Nørskov JK (1999) Surface segregation energies in transition-metal
alloys. Phys Rev B 59(24):15990
118. Hammer B, Hansen LB, Nørskov JK (1999) Improved adsorption energetics within density-
functional theory using revised Perdew-Burke-Ernzerhof functionals. Phys Rev B 59(11):7413
119. Son SU, Jang Y, Park J et al (2004) Designed synthesis of atom-economical Pd/Ni
bimetallic nanoparticle-based catalysts for Sonogashira coupling reactions. J Am Chem Soc
126(16):5026–5027
120. Mazumder V, Chi M, Mankin MN et al (2012) A facile synthesis of MPd (M = Co, Cu)
nanoparticles and their catalysis for formic acid oxidation. Nano Lett 12(2):1102–1106
References 125

121. Metin Ö, Ho SF, Alp C et al (2012) Ni/Pd core/shell nanoparticles supported on graphene as
a highly active and reusable catalyst for Suzuki-Miyaura cross-coupling reaction. Nano Res
6(1):10–18
122. Zoski CG (2006) Handbook of electrochemistry. Elsevier, Amsterdam
123. Wang C, van der Vliet D, More KL et al (2011) Multimetallic Au/FePt3 nanoparticles as
highly durable electrocatalyst. Nano Lett 11(3):919–926
124. Liu Y, Wang C, Wei Y et al (2011) Surfactant-induced postsynthetic modulation of Pd
nanoparticle crystallinity. Nano Lett 11(4):1614–1617
125. Cordero B, Gomez V, Platero-Prats AE et al (2008) Covalent radii revisited. Dalton Trans
21:2832–2838
126. Wu J, Gross A, Yang H (2011) Shape and composition-controlled platinum alloy nanocrystals
using carbon monoxide as reducing agent. Nano Lett 11(2):798–802
127. Gan L, Heggen M, Rudi S et al (2012) Core-shell compositional fine structures of deal-
loyed Pt(x) Ni(1−x) nanoparticles and their impact on oxygen reduction catalysis. Nano Lett
12(10):5423–5430
128. Shao M, Peles A, Shoemaker K et al (2011) Enhanced oxygen reduction activity of platinum
monolayer on gold nanoparticles. J Phys Chem Lett 2(2):67–72
129. Shao M, Shoemaker K, Peles A et al (2010) Pt monolayer on porous Pd−Cu alloys as oxygen
reduction electrocatalysts†. J Am Chem Soc 132(27):9253–9255
130. Shao M, Smith BH, Guerrero S et al (2013) Core-shell catalysts consisting of nanoporous
cores for oxygen reduction reaction. Phys Chem Chem Phys 15(36):15078–15090
131. Chen Y, Liang Z, Yang F et al (2011) Ni–Pt core–shell nanoparticles as oxygen reduction
electrocatalysts: effect of Pt shell coverage. J Phys Chem C 115(49):24073–24079
132. Guo S, Zhang S, Su D et al (2013) Seed-mediated synthesis of core/shell FePtM/FePt (M =
Pd, Au) nanowires and their electrocatalysis for oxygen reduction reaction. J Am Chem Soc
135(37):13879–13884
133. Choi S-I, Shao M, Lu N et al (2014) Synthesis and characterization of Pd@Pt–Ni core–shell
octahedra with high activity toward oxygen reduction. ACS Nano 8(10):10363–10371
134. Xie S, Choi SI, Lu N et al (2014) Atomic layer-by-layer deposition of Pt on Pd nanocubes
for catalysts with enhanced activity and durability toward oxygen reduction. Nano Lett
14(6):3570–3576
Chapter 4
Membrane Electrode Assembly (MEA)

Abstract With a continuous decrease in Pt loading, mass transport resistances

within MEA are severely aggravated, especially that the oxygen transport resis-
tance is becoming an obstacle for commercialization of fuel cells. In this chapter,
the proton transport in the ultrathin perfluorosulfonic acid ionomer film covered on
catalyst surface is investigated first. Next, the oxygen transport behavior including
gas diffusion in porous electrode and gas permeation in ultrathin perfluorosulfonic
acid ionomer film is probed; thus, the underlying transport mechanism is discussed
detailedly. In addition, the application of numerous novel electrocatalysts leads to
high ORR performances but also cation contamination, which is also assessed in
the last part of this section. Herein, we shed light on transport and reaction mech-
anism inside the MEA, as well as provide key insights for optimization design for
appropriate electrode structures.

Keywords Membrane electrode assembly · Proton transport · Oxygen transport ·

Cation contamination

4.1 Introduction

High-performance H2 fueled proton exchange membrane fuel cells (PEMFCs) for

automotive propulsions have been regarded as an ideal and the most promising alter-
native to replace fossil fuel-based internal combustion engines. Membrane electrode
assemblies (MEAs) function as critical components in PEMFCs, whose cost accounts
for over 41% of that of fuel cell stacks [1].

4.1.1 Reaction and Mass Transport Process in Fuel Cells

In a typically MEA structure, a proton exchange membrane (PEM) is sandwiched

by two electrodes, i.e., anode catalyst layer (ACL) and cathode catalyst layer (CCL),
which consist of proton exchange ionomer and electrocatalyst. Further, gas diffusion

© Shanghai Jiao Tong University Press and Springer-Verlag GmbH Germany 2021 127
J. Zhang and S. Shen, Low Platinum Fuel Cell Technologies,
Energy and Environment Research in China,
https://doi.org/10.1007/978-3-662-56070-9_4
128 4 Membrane Electrode Assembly (MEA)

layers (GDLs) are fabricated onto both anode and cathode sides of MEA, and bipolar
plates (BPs) are fixed on the outside of GDLs. A schematic of MEA and the mass
transport process occurred within a single PEM fuel cell is shown in Fig. 4.1.
As previously mentioned, the reaction mechanism is not complex, and the
mass transport behavior within the single cell is not mysterious. As demonstrated
in Fig. 4.1, the electrochemical reaction and mass transport process occurred in
PEMFCs can be divided to five categories:
1. Reactant supplies: In order to ensure the continuous current, it is necessary to
continuously provide fuel (hydrogen) and oxides (oxygen) for the fuel cell, which
passes through the flow channel and GDL; then, enter CL for reaction. A large
amount of reaction gas is required for the hydrogen oxidation reaction (HOR) and
oxygen reduction reaction (ORR), especially at large current; thus, it is necessary
to design efficient flow fields in BP and construct suitable microstructure in GDL
and CL. It is worth to note that the reaction gas relies on convection transport in
the flow channel and diffusion transport in GDL and CL. Therefore, the different
mass transfer characteristics should be taken into consideration when developing
optimization strategies.
2. Oxidation–reduction reactions: Once the reaction gas comes to the catalyst layer,
the HOR and ORR begin immediately. The current of PEMFC depends on the
reaction rate of the oxidation–reduction reactions, i.e., the larger the reaction
rate the higher the current. To obtain higher current and better performance, it is
crucial to develop efficient electrocatalyst as well as increase the utilization of
the electrocatalyst in the electrode.
3. Proton transport: The protons produced by HOR would be transport through the
PEM to CCL; then, they would be transport through the ultrathin ionomer film to
access the electrocatalysts. It is generally accepted that the vehicular mechanism
dominants at low hydration, while the hopping mechanism dominants at high
hydration. Transport of proton in ionomer is not only dependent on water content,

Fig. 4.1 Schematic of MEA

and the mass transport
processes
4.1 Introduction 129

but also governed by the its interaction with the SO3− sites, by the length and
hydrophilicity of side chain, as well as the segmental motions of the polymer
chains [2].
4. Electron transport: Same to the protons, the electrons also need bringing from
anode to cathode to take part in the reaction. Since the PEM is non-conductive,
the electrons are forced round the circuit in the opposite direction, which ensure
the continuous output of current.
5. Product discharging: As a representative of clean energy, the reaction products of
PEMFCs are only water that does not pollute the environment. However, if these
product water cannot be discharged from the cell in time, it will seriously affect
the performance of fuel cell. If the product accumulates in the flow channel or
electrode, the phenomenon of “water flooding” will occur, blocking the transport
path of the reaction gas, then preventing the reaction gas from being transported
into the catalytic layer, thus causing insufficient reactants, and eventually causing
“performance decrease” in MEAs.

4.1.2 Structure in CLs and Its Characterization

A typical CL consists of a carbon-supported Pt electrocatalyst (Pt/C), perfluorosul-

fonic acid ionomer (PFSI) binder and pore space, which are illustrated in Fig. 4.2 [3].
It is noted that under certain conditions (e.g., high relative humidity or high current
density), there appears liquid water within the pore or between electrocatalyst and
PFSI.

Fig. 4.2 Representation of

the principal components of
CLs and two types of mass
transport behavior
130 4 Membrane Electrode Assembly (MEA)

Further, electrocatalysts and ionomer form agglomerates; thus, there exit two types
of pores inside the CLs: small-scale pores (primary pores) between the particles in
the agglomerates and large-scale pores (secondary pores) between the agglomerates.
On the other hand, electrocatalysts are mostly covered by ultrathin ionomer film
(5–10 nm), which functions as a proton conductor but oxygen transport barrier.
The electrochemical reactions occur at the triple-phase (C, Pt, ionomer) bound-
aries where electrons, protons and reactants meet at the Pt catalyst surface. Within
this complex geometrical and chemical environment, C support and ionomer binder
create networks for conduction electrons and protons, respectively, and the pore space
generates pathways for reactant transport and product removal.
There are a large number of intricate and entangled phenomena in CLs, including
electron and ion conduction, electrical contact resistance, gas diffusion, heat trans-
port, ohmic heating, oxidation–reduction reaction, gas sorption and desorption, water
condensation and evaporation, as well as ionomer swelling. To clarify these electro-
chemical reaction and mass transport phenomena, it is critical to characterize the
nanostructure inside CLs.
One effective method to investigate the morphology of CLs is scanning elec-
tron microscopy (SEM) which has been widely used to visualize the cross-section
of MEAs in past studies. A typical SEM image of cross-sectional MEA is shown
in Fig. 4.3. Accordingly, various techniques to obtain cross-section are developed,
such as freeze-fracture method [4], microtomy [5], focusing ion beam (FIB) [6]
and cross-sectional polishing (CP) method [7]. For further insights, tomographic
investigation by the FIB/SEM technique has been widely used to reconstruct the 3D
microstructures in CLs [8, 9].
However, the SEM images can only reflect the nanomorphology on the catalyst
surface (as shown in the left images of Fig. 4.4), which is insufficient to characterize
the high surface carbon (HSC)-supported Pt electrocatalysts, for that Pt particles

Fig. 4.3 Typical SEM image

of cross-sectional MEA
4.1 Introduction 131

Fig. 4.4 Typical SEM and TSEM images of catalysts [10]. a Pt/N-modified Ketjenblack;
b Pt/Ketjenblack

distribute not only on the carbon surface but also inside the primary pore of carbon
materials. To observe the Pt location on these electrocatalysts, SEM and TSEM
(transmission SEM) images are taken simultaneously from the same area. The SEM
signal gives information about the Pt particles deposited on the exterior surface of
the carbon support only. On the other hand, as shown in the right images of Fig. 4.4,
the TSEM signal of the bright field transmission electron detector shows both the Pt
particles on the exterior and interior surface.
As mentioned above, gas diffusion electrodes (GDEs) are porous structure, and
the pores inside CLs range from nanometers to submicrometers. To characterize
the porosity and pore size distribution over such a wide range, mercury intrusion
porosimetry (MIP) and gas adsorption (Brunauer–Emmett–Teller, BET) studies are
employed, which measures pore size from 3 to 1000 nm and 2 to 100 nm, respectively.
In MIP, for example, mercury is intruded in the pores with the help of external
pressure, and the bottleneck size of pores is calculated by using Washburn’s equation.
Figure 4.5 demonstrates a typical MIP result of CLs.
132 4 Membrane Electrode Assembly (MEA)

Fig. 4.5 A typical MIP result of CLs. a Pore size distribution; b cumulative intrusion versus pressure

4.2 Proton Transport in Electrode

The PFSA ionomer films that curled around catalysts particles within CLs have an
ultrathin thickness, which has different proton conduction mechanisms with proton
exchange membranes (>10 µm) due to the confinement effect.
Herein, in order to further investigate effect of ionomer equivalent weight (EW)
on proton conductivity in CCLs, self-assembly method is employed to obtain a good
representation of ionomer structure and electrolyte resistances are measured using
electrochemical impedance spectroscopy (EIS). An interdigital array (IDA) micro-
electrode was fabricated onto Si/SiO2 substrates, and Fig. 4.6a and b show the 3D
model and scanning electron microscope (SEM) image of IDA substrate.
Specifically, the substrate is immersed in Nafion solution with certain concentra-
tion for a well-controlled time to obtain the required thickness. Figure 4.6c displays a
schematic depiction of the adsorbed Nafion on Si/SiO2 substrate, and the insert image
of Fig. 4.7a shows the typical 3D view of the ultrathin Nafion film characterized by
atomic force microscopy (AFM). The Nafion film thickness is measured based on
step height of AFM probe during the scanning from film surface to substrate surface.
For the proton conductivity measurement, samples are placed in an environmental
chamber with relative humidity (RH) of 98% to assure the fully hydration of Nafion

Fig. 4.6 IDA microelectrode and adsorbed Nafion [11]. a 3D model of IDA microelectrode, b SEM
image of IDA substrate, c adsorbed nafion on Si/SiO2 substrate
4.2 Proton Transport in Electrode 133

Fig. 4.7 Typical impedance results and proton conductivities [11]. a An example of impedance
result belonging to the EW-1100 ionomer with a film thickness of 40 nm (the insert image is the
3D view of AFM-tested film); b tested proton conductivities of EW-1100 ionomer; c tested proton
conductivities of EW-1000 ionomer; d tested proton conductivities of EW-750 ionomer with various
thicknesses

ionomer. The film proton conductivity can be calculated as:

1 d
σ = · (4.1)
R f l(N − 1)δ

where σ is the proton conductivity, Rf is the resistance got from the impedance
measurement, d represents the distance between the microelectrode array, l is the
length of single microelectrode, N is the number of microelectrodes, and δ is the
thickness of Nafion film.
Figure 4.7a shows an example of an impedance result that belongs to an EW-1100
Nafion film with a thickness of 40 nm. The impedance diagram consists of an oblique
line at low-frequency region and a semicircle at high-frequency region whose radius
the film resistance (Rf ) can be fitted from. It is clearly observed that Rf decreases
as environmental temperatures increases since the ion mobility is larger at higher
temperature.
Figure 4.7b–d shows the proton conductivity of Nafion films with different thick-
nesses, and it is clear that the proton conductivity of ultrathin Nafion film is much
134 4 Membrane Electrode Assembly (MEA)

smaller than that of thick membrane. As a result, the proton conduction within the
PEMFC catalyst layer is more sluggish than that in the thick membrane. In addition,
for EW-1100 ionomer at 20 °C, as the film thickness increases from 40 to 120 nm,
the corresponding proton conductivity increases significantly, suggesting a potential
thickness-dependent microstructure of ultrathin Nafion films.
In terms of bulk membranes, lower EW has a slight increase in proton conduc-
tivity, and the reason behind it is that the lower EW (i.e., higher ionic exchange
concentration) would induce larger charge carrier concentration which influences
the conductivity as shown in Eq. (4.2):

σ (T ) = c(T ) · Z · μ(T ) (4.2)

where σ is the conductivity, c is the charge carrier concentration, Z represents

the valence of the carrier and fundamental charge unit respectively, μ denotes the
mobility of charge carriers.
In order to clarify the EW effect of ultrathin Nafion film, proton conductivity of
Nafion films with same thickness but various EWs is shown in Fig. 4.8a–c. It is clear
that lower EW ionomer leads to higher conductivity and the deviation enlarges as
temperature increases. The observed EW effect in ultrathin film is consistent with

Fig. 4.8 Proton conductivity as a function of EW [11]. a Film thickness of 40 nm; b film thickness
of 80 nm; c film thickness of 120 nm
4.2 Proton Transport in Electrode 135

Fig. 4.9 2D GISAXS patterns [11]. a EW-750-45 nm ionomer; b EW-1000-45 nm ionomer; c EW-
1100-45 nm ionomer

that in bulk membrane; however, the reason behind might be different since there
is plausible difference in proton conduction mechanism between nanosized film
and bulk membrane. The conductivity of EW-750-40 nm film is approximately two
times higher than EW-1100-40 nm film at each temperature level. However, such
wide disparity cannot be a result of the relatively small difference in charge carrier
concentration which can be calculated using following equation:

0.01 × IEC · ρ
c= (4.3)
1 + 0.01x

where ρ is the polymer density, x represents volume-based water uptake. Therefore,

it is reasonable to deduce that ion mobility plays a more significant role on such EW
effect than ionic exchange concentration does.
Grazing incidence small-angle X-ray scattering (GISAXS) measurements were
employed to deepen the understanding the microstructure of ultrathin Nafion film. As
shown in Fig. 4.9, the obtained GISAXS patterns of EW-750 ionomer display a typical
scattering ring which is originated from correlations among ionic domains [12], and
the scattering ring becomes weaker for ionomers with increased EW especially the
one of EW-1100 that the scattering ring almost disappears.
Figure 4.10a shows the line profiles from GISAXS patterns wherein the ionic peak
at q = 1.5 to 2 nm−1 stands for the phase separation. Nafion ionomer film with lower
EW presents stronger peak intensity, suggesting higher degree of phase separation.
Contrast to bulk membrane, the ultrathin film has limited phase separation due to the
confinement effect, which could explain the huge gap on proton conductivity between
bulk membrane and ultrathin film. As shown in Fig. 4.10b, the confinement effect is
alleviated as the film thickness increases, demonstrating that the higher fraction of
ionic groups in the lower EW ionomer is able to induce better phase separation. Thus,
proton-conducting pathway is improved and in turn facilitates the proton mobility.
Polarization curves of PEMFCs with various ionomer equivalent weights in
cathode catalyst layer are shown in Fig. 4.11a. While the cell temperature, pres-
sure, RH, ionomer/carbon ratio, reactant gas, cathode/anode Pt loading are 80 °C,
136 4 Membrane Electrode Assembly (MEA)

Fig. 4.10 Line cuts from the GISAXS patterns for qp [11]. a 40 nm film with various EWs;
b EW-1000 ionomer with different film thicknesses

Fig. 4.11 Fuel cell performance of MEAs with different ionomer EW value [11]. a Polarization
curves, b complex-plane impedance

ambient pressure, 100%, 0.75, H2 /O2 , and 0.4/0.1 mgpt cm−2 , respectively. MEA
with EW-750 ionomer shows higher performance. Complex-plane impedance plots
for H2 /N2 operation on single cells are shown in Fig. 4.11b. While reactant gas,
ionomer/carbon ratio, cathode/anode Pt loading, testing potential, frequency range,
peak-to-peak perturbation is H2 /N2 , 0.8, 0.4/0.1 mgpt cm−2 , 0.2 V, 1 MHz–10 Hz, and
±0.01 V. respectively. The membrane resistance (Rmembrane ) could be omitted; then,
the resistance of the electronic the plot would be simply shifted along the real axis.
A Warburg-like response (45° slope) is observed at high frequencies, corresponding
to ion migration through the catalyst layer, while the impedance plot curves up to a
limiting capacitance response (vertical) at low frequencies, corresponding to the total
capacitance and resistance of the CCLs. The ionic resistance (Rionic ) can be obtained
from the length of the Warburg-like region projected onto the real impedance axis
(Rionic /3), and it is obvious that the electrolyte resistance of MEA with EW-1000 is
35% higher than that with EW-750. These fuel cell performance and Rionic both show
4.2 Proton Transport in Electrode 137

that lower ionomer EW not only drives higher conductivity in ultrathin Nafion films
but is benefit to promote the performance of single cell and low down the electrolyte
resistance in cathode catalyst layer.

4.3 Oxygen Transport in Electrode

4.3.1 Local and Bulk Oxygen Transport in CLs

Although the reduction of Pt loading lowers the cost of PEMFCs, it increases the gas
transport resistance within CLs and in turn greatly sacrifices the cell performance.
It is noted that gas transport resistance is considered as the stumbling block of high
performance of PEMFCs at high current densities, especially for the oxygen transport
at cathode, which is much slower than the hydrogen transport at anode.
The oxygen transport resistance in CL includes two parts, i.e., local resistance
and bulk resistance, both of which are directly influenced by ionomer [13, 14]. The
local resistance (RLocal ), derived from the oxygen transport into/through the ionomer
film covered on Pt surface, is dominated by the coverage, thickness and interfacial
properties of ionomer [2, 15–17]. Suzuki et al. [18] demonstrated that the oxygen
dissolution resistance at the gas/ionomer interface contributes to a significant part
of the transport resistance when the ionomer film is thinner than 100 nm. Greszler
[19] also proposed the hypothesis that the sluggish oxygen adsorption at ionomer
surface is the limiting step of local transport process. Kudo et al. [20] measured the
local resistance in a two chamber cell, and they found that the ionomer/Pt interfacial
delivers a transport resistance similar to that of 30–70 nm ionomer film, while the
actual thickness of ionomer film in CLs is ca. 10 nm only.
The bulk resistance (RBulk ), derived from the Knudsen diffusion and molecule
diffusion in the porous structure of CL, is also influenced by the ionomer that occu-
pies the pore space formed by carbon agglomerates [21, 22]. To investigate the
ionomer effect on porous structure, Suzuki et al. [22] developed a three-dimensional,
mesoscale model to visualize the percolating paths of CL, which showed a ca. 25–
30% decrease in oxygen diffusion coefficient when the ionomer content increased
from 14 to 50%. Okumura et al. [23] used the FIB-SEM technique to obtain the
3D visualization of CL structure, which showed that gas pathway was significantly
occupied by ionomer and the corresponding porosity decreased from ca. 0.7 to ca.
0.4 as the ionomer content increased from 8 to 38%. Obviously, the ionomer plays
an essential role in the oxygen transport process within the CL, and thus illuminating
the effect of ionomer is the basis for CL improvement to further reduce Pt loading
without sacrificing performance.
138 4 Membrane Electrode Assembly (MEA)

4.3.2 Experimental Measurement of Local and Bulk Oxygen

Transport Resistance

As shown in Fig. 4.12, a dual-layer cathode catalyst layer is designed to experimen-

tally measure the oxygen transport resistance in CLs. The cathode consists of several
components; thus, the total oxygen transport resistance (RTotal ) in cathode can be
divided into different parts [24]:

RTotal = RDCL + RCCL + RGDL + RChannel (4.4)

where RCCL , RDCL , RGDL and RChannel denote the oxygen transport resistances in real
catalyst layers (CCLs), dummy catalyst layers (DCLs), gas diffusion layers (GDLs)
and the flow channels.
RTotal is measured by the limiting current density method based on the combination
of Fick’s law and Faraday’s law:

dcO2
NO2 = −DOeff2 (4.5)
dδ
i
NO2 = (4.6)
4F

Fig. 4.12 Structure and

resistance distribution of the
cell with dual-layer CL [25]
4.3 Oxygen Transport in Electrode 139

where NO2 is the molar flux of oxygen and DOeff2 is the effective diffusion coefficient
of oxygen. cO2 is the molar concentration of oxygen. δ and F denote the transport
length and Faraday constant, respectively.
Combining Eqs. (4.5) and (4.6), we can get:

i cO − cPt,surf
= DOeff2 2 (4.7)
4F δ
where cO2 is the oxygen concentration in the channels, cPt,surf is the oxygen
concentration on the Pt surface.
At the limiting current density, Eq. (4.7) turns to:

cO2 − cPt,surf 1 4F xO2 · pCH 1

RTotal = 4F = 4FcO2 · · (4.8)
i i lim RT i lim

where pCH is the partial pressure of oxygen in channels, R is the universal gas
constant, and T is the absolute temperature of cells. Therefore, RTotal can be calculated
by Eq. (4.8) with certain xO2 and pCH at limiting current density.
RChannel can be estimated based on empirical equation [24] and calculated to be
0.258 s cm−1 .
RGDL can be directly measured by changing the GDL thickness (i.e., different
numbers of carbon papers are used), as shown in Eq. (4.9)
With several layers of GDL used, the total resistance can be expressed as

RTotal = RCCL + rGDL · n + RChannel (4.9)

where rGDL is the resistance of a single GDL and n is the number of GDLs. As shown
1
in Fig. 4.13a, ilim increases proportionally as n increases; thus, rGDL can be calculated
(Fig. 4.13b).
Further, since DCL has the same structure as CCL but has no Pt particles, there is
not electrochemical reaction happening in DCL, RDCL includes bulk resistance only.
RTotal can be further expressed as following equation

RTotal = rBulk · h + RCCL + RGDL + RChannel (4.10)

where rBulk is defined as the bulk resistance per unit thickness, h is the thickness
of DCL. Meanwhile, the DCL thickness is much higher than that of CCL; thus, the
bulk resistance in CCL is negligible. Therefore, Eq. (4.10) is further expressed as a
function of the DCL thickness:
 
RTotal = rBulk · h + RLocal + RGDL + RChannel δDCL  δCL
eff
(4.11)

1 RT RT (RLocal + RGDL + RChannel )

= · rBulk · h + (4.12)
i lim 4F xO2 · pCH 4F xO2 · pCH
140 4 Membrane Electrode Assembly (MEA)

Fig. 4.13 GDL resistance

[25]. a Linear regression
method, b test results

According to Eq. (4.12), rBulk and RLocal can be obtained by the slope and the
intercept, respectively, as shown in Fig. 4.14. As expected, 1/i lim increases with the
increase in DCL thickness, for that thicker DCL prolongs the transport length of
oxygen diffusion. Table 4.1 summarizes the calculation results in which RLocal is
found to be 0.99 s cm−1 . rBulk is calculated by the slope of the plots in Fig. 4.14
and is found to be 1800 s cm−2 . As rBulk is the reciprocal of the effective diffusion
coefficient in CL (DOeff2 ), DOeff2 could be obtained as 5.55 × 10–4 cm2 s−1 . RBulk in
CCL is estimated as 0.071 s cm−1 , almost 15 times smaller than RLocal .

Fig. 4.14 Plot of Eq. (4.16)

which illustrates the
relationship between 1/i lim
and DCL thickness [25]
 4.3 Oxygen Transport in Electrode

Table 4.1 Calculated values of bulk and local resistances

x0 Intercept (cm2 A−1 ) RLocal (s cm−1 ) Slope (µm A−1 ) rBulk (s cm−2 ) RBulk (s cm−1 ) eff
DO2CL (cm2 s−1 )
0.01 10.5884 1.001 1.12847 1776.141 0.0700 5.63 × 10–4
0.02 5.01915 0.914 0.59728 1880.163 0.0740 5.32 × 10–4
0.04 2.74307 1.061 0.27791 1749.652 0.0691 5.72 × 10–4
Average – 0.992 – 1801.985 0.0713 5.55 × 10–4
141
142 4 Membrane Electrode Assembly (MEA)

It is clearly observed that the local oxygen resistance is much larger than the
bulk oxygen resistance in CCLs, suggesting that the local oxygen transport domains,
especially in low Pt loading MEAs. Hence, it is essential to reduce RLocal for achieving
higher fuel cell performance.

4.3.3 Influence of Ionomer Content on Local and Bulk

Oxygen Transport Resistance

The dual-layer CL design could be employed to investigate the effect of different

factors, such as ionomer content, which is a key material in CLs of PEMFCs. Herein,
we measure the local and bulk resistances of MEAs with various ionomer contents,
which provide quantitative information on the respective effect of ionomer on the
local and bulk transport. In detail, the ionomer/carbon ratio (I/C) in CCL is set as
0.65, 0.8 and 0.95, while I/C in DCL remains 0.8, which are named 6580, 8080 and
9580, respectively. Moreover, I/C in DCLs are set as 0.65, 0.8 and 0.95, while I/C in
CCL remains 0.8, which are named 8065, 8080 and 8095, respectively. The I/C in
CCL and DCL, as well as the sample name is summarized in Table 4.2.
Figure 4.15a–c show the reciprocal of limiting current density (1/i lim ) versus
DCL thickness (h), measured with 1%, 2% and 4% O2 /N2 –O2 mixed gas. RLocal and
rBulk is calculated via the intercept and the slop, respectively.
In fact, local transport resistance per electrode area (RLocal ) could be further
normalized as local transport resistance per unit Pt surface area (rLocal ):
rLocal
RLocal = (4.13)
f Pt

where f Pt is the roughness factor of the electrode, defined as the effective Pt

surface area per unit electrode area.
In order to deepen the understanding of local transport behavior, all the local
transport resistances are expressed as rLocal , which are summarized in Table 4.3 and
Fig. 4.15d. rLocal increases from 6.15 to 58.88 s cm−1 as I/C increases from 0.65 to
0.9, corresponding to a ten times increment. The unexpected augment implies that
the thickened ionomer film on Pt surface is not the only reason, and it is suspected
that such a high sensitivity derives from the change of ionomer morphology with
increased ionomer fraction.

Table 4.2 I/C in CL and DCL of the tested MEAs

6580 8080 9580 8065 8095
I/C of CL 0.65 0.8 0.95 0.8 0.8
I/C of DCL 0.8 0.8 0.8 0.65 0.95
Table 4.3 Resistance properties of 6580, 8080 and 9580 with different O2 concentration
6580 8080 9580
4.3 Oxygen Transport in Electrode

Intercept xO2 (%) 1/i (cm2 A−1 ) rLocal (s cm−1 ) 1/i (cm2 A−1 ) rLocal (s cm−1 ) 1/i (cm2 A−1 ) rLocal (s cm−1 )
1.00 6.028 7.185 10.59 25.41 19.32 60.31
2.00 2.769 5.222 5.019 23.21 9.181 56.48
4.00 1.435 6.040 2.743 26.95 4.801 59.85
Average – – 6.149 – 25.19 – 58.88
Slope xO2 (%) 1/(i*δ) (cm2 A−1 µm−1 ) rBulk (s cm−2 ) 1/(i*δ) (cm2 A−1 µm−1 ) rBulk (s cm−2 ) 1/(i*δ) (cm2 A−1 µm−1 ) rBulk (s cm−2 )
1.00 1.135 1786 1.128 1776 1.173 1846
2.00 0.5809 1829 0.597 1880 0.585 1841
4.00 0.2912 1833 0.278 1750 0.275 1730
Average – – 1816 – 1802 – 1805
143
144 4 Membrane Electrode Assembly (MEA)

Fig. 4.15 Influence of I/C on local transport resistance [26]. a–c Limiting current density of MEA
samples; d local resistance of specific Pt surface area with I/C of 0.65, 0.8 and 0.95

It is well acknowledged that large ionomer fraction could result in non-

homogeneous distribution of ionomer and even ionomer aggregation on catalyst
surface which may block some secondary pores and imped gas diffusion [27]. Hence,
on the one hand, extra oxygen transport barriers are caused by the ionomer incrassa-
tion and aggregations on Pt surface, e.g., “ionomer pocket” [27]. On the other hand,
as shown in Fig. 4.16, the diffusion path is blocked by ionomer; thus, oxygen has to
diffuse through the ionomer phase rather than the gas phase. These complex states
of ionomer morphology on Pt surface, resulted from high ionomer fraction, should
be responsible for the high sensitivity of local resistance to ionomer content.
As for bulk resistances, MEAs with different I/C in DCL and constant I/C in CCL
are measured by limiting current methods, while 1/i lim versus h measured in 1%,
2% and 4% O2 /N2 –O2 is shown in Fig. 4.17a–c, and detailed values are summarized
in Table 4.4.
As shown in Fig. 4.17d rBulk increases significantly as I/C increases. rBulk is also
significantly influenced by I/C, for that rBulk increases from 360 to 13,000 s cm−2
when I/C varies from 0.65 to 0.95, corresponding to a 36 times increment.
Such high sensitivity is mainly caused by an extra increment of bulk transport
resistance due to the structural variations of pores. The cross-sectional SEM images
of DCL (Fig. 4.18) show that the macropores are gradually filled by the ionomer
as I/C increases, while mercury porosimetry measurement (Fig. 4.19) shows that
the pore diameter and porosity decrease as I/C increases. The mean pore diameter
4.3 Oxygen Transport in Electrode 145

Fig. 4.16 Possible changes of local transport process with increased ionomer content [26]

Fig. 4.17 Influence of I/C on bulk transport resistance [26]. a–c Limiting current density of MEA
samples; d bulk resistance per unit thickness with I/C of 0.65, 0.8 and 0.95

is measured to be 50.3 nm, 40.3 nm and 29.8 nm for I/C-0.65, 0.8 and 0.95, and
the corresponding porosity is 64.0%, 53.3% and 51.4%, respectively, confirming the
occupation of macropores by ionomer phase, which is responsible for the enlarged
bulk resistance.
 146

Table 4.4 Resistance properties of 8065, 8080 and 8095 with different O2 concentration
8065 8080 8095
Intercept xO2 (%) 1/i (cm2 A−1 ) rLocal (s cm−1 ) 1/i (cm2 A−1 ) rLocal (s cm−1 ) 1/i (cm2 A−1 ) rLocal (s cm−1 )
1.00 9.911 22.70 10.59 25.41 – –
2.00 4.601 19.87 5.019 23.21 0.3207 −0.565
4.00 2.436 22.04 2.743 26.95 −0.6605 −1.082
Average – – 21.54 – 25.19 – –
Slope xO2 (%) 1/(i*δ) (cm2 A−1 µm−1 ) rBulk (s cm−2 ) 1/(i*δ) (cm2 A−1 µm−1 ) rBulk (s cm−2 ) 1/(i*δ) (cm2 A−1 µm−1 ) rBulk (s cm−2 )
1.00 0.2443 335 1.128 1776 – –
2.00 0.1149 362 0.5972 1880 3.916 12,326
4.00 0.0614 386 0.2783 1750 2.166 13,636
Average – – 360 – 1802 – 12,981
4 Membrane Electrode Assembly (MEA)
4.3 Oxygen Transport in Electrode 147

Fig. 4.18 Cross-sectional SEM images of DCL with I/C of 0.65, 0.8 and 0.95 [26]

Fig. 4.19 Pore size distribution of DCL with I/C of 0.65, 0.8 and 0.95 [26]

The influence of I/C on transport resistance is verified by 5 cm × 5 cm MEAs

(0.1 mgPt cm−2 at cathode without DCL), and the cell temperature, back pressure,
RH and stoichiometric ratio of H2 /Air are 80 °C, 150 kPaabs , 67% RH and1.5:2,
respectively. As shown in Fig. 4.20, MEA with lower I/C has significantly better
performance relative to the other two, due to the low transport resistance, especially
the low local resistance as indicated in Fig. 4.15.
148 4 Membrane Electrode Assembly (MEA)

Fig. 4.20 Fuel cell

performances with different
I/C [26]

4.3.4 Mechanism of Bulk Oxygen Transport Behavior

As mentioned above, it is very critical to understand the transport mechanism in

CCL and in turn to explore strategies to greatly mitigate the transport resistance for
developing high-performance MEAs with ultralow Pt loading.
One way to minimize the O2 transport resistance is to tailor the structure of
CCLs through changing the composition, amount and preparation technologies. A
number of different pore-forming additives have been added into the catalyst ink
to increase the porosity of CCLs, which bring better performances of single cells
[28, 29]. Nevertheless, so far, the detailed mechanism of the pore-forming step has
been rarely investigated, due to the complex structure and indistinguishable transport
resistances in CCLs.
In CCLs, whose typical pore size is usually less than 100 nm, the transport resis-
tance comes from the diffusion of O2 in these primary and secondary pores (the
bulk resistance, RBulk ) and permeation through the ultrathin ionomer film covered
on catalyst (the local resistance, RLocal ) [15, 30]. Commonly, the gas diffusivity in
CCLs is described by the Bruggeman relation, which can be given as [31]:

DOeff2 = Do2 · ε1.5 (4.14)

where ε is the porosity, DOeff2 is the O2 effective diffusivity in porous media, and Do2 is
the O2 free space diffusivity, which can be estimated by the Bosanquet approximation
as [32]:

1 1 1
= + (4.15)
Do2 DM DK
4.3 Oxygen Transport in Electrode 149

where DM and DK are correspondingly molecular and Knudsen diffusivities, respec-

tively. Based on these two basic equations, various studies of the gas diffusion have
been conducted, including computational simulations [22, 33, 34] and experimental
measurement [35–37].
However, to deepen the understanding of bulk transport resistance, we use MgO
as a pore-forming agent to build ordered microstructures and explore the detailed
transport mechanism.
The SEM images are shown in Fig. 4.21, and we can see there are more and
larger pores in the pore-formed electrodes than the original ones. The MgO added
into the ink stays in DCL and occupies a certain volume after the spraying. After the
MgO dissolving in acid, numbers of mesoscale cavities forms in DCL, which can be
observed in SEM images (Fig. 4.21a, c).
The mercury intrusion porosimetry (MIP) results are depicted as plots of pore
diameter versus differential volume to elucidate the pore size distributions in
Fig. 4.22. It is clear that the most probable pore diameter enlarges from 40.3 to
85.3 nm after pore forming, while the porosity increases from 53.33 to 65.35%.
ICP elemental analysis was performed to determine the amount of MgO residual in
pickled electrode, and the corresponding MgO content is 0.18 wt% only. Considering
the initial mass ratio of MgO in electrode (13 wt%), it can be concluded that over
98.6% pore-forming agent has been removed that would have negligible impact on
catalysts activity.
Figure 4.23 shows the limiting current densities of electrodes containing DCLs
with various thicknesses were measured by LSV. The reaction gas, RH, operating
pressure and temperature are H2 /8 mol% O2 –N2 gas mixture, 67%, 150 kPaabs and
80◦ C, respectively. Standard deviations are indicated by the error bar. Not surpris-
ingly, RTotal increases as the increase in h, which is in approximately linear relation.
According to the intercepts in Fig. 4.23, RLocal of pore-formed and non-pore-formed
electrodes can be calculated to be 0.32 and 0.37 s cm−1 for the two electrode remains
same Pt loading and electrode structure in CLs.

Fig. 4.21 SEM images [38]. a Pore-formed CLs; b original CLs

150 4 Membrane Electrode Assembly (MEA)

Fig. 4.22 MIP results for

pore-formed and original
DCLs [38]

Fig. 4.23 Plot RTotal versus

h for both pore-formed and
original electrodes [38]

In terms of the bulk resistance, it is clearly that the change of RTotal versus h is
gentler in pore-formed electrodes than in original electrodes. rBulk can be calculated
according to the slopes in Fig. 4.23, and the values are summarized in Table 4.5.
It is easy to understand that the pore former leaves a number of cavities in DCLs
after its removal, which provide more and wider pathways for O2 diffusion. The
pore-forming increases not only the porosity, but also the mean pore size, and the

eff , calculated ε eff and calculated κ

Table 4.5 Summary of ε from MIP, measured DO 2

ε (%) eff (cm2 s−1 )

DO εeff (%) κ
2

Pore-formed electrodes 65 11.6 × 10−4 3.22 0.05

Original electrodes 53 6.3 × 10−4 2.14 0.04
4.3 Oxygen Transport in Electrode 151

larger the pore size and porosity are, the smaller the overpotential due to the mass
transport limitation is.
Particularly, as mentioned above, DOeff2 can be figured out easily and the value
is also summarized in Table 4.5. The great increment disaccords the quantitative
relations in Eq. (4.14). Combining Eq. (4.14) and the MIP results, when the porosity
increases from 53.33 to 65.35%, there is only a 35.6% increase in DOeff2 . The difference
in the calculated and experimental increment suggests that the porosity is not the only
factor influencing DOeff2 , even not the major one. It is believed that the pore distribution
also plays a pivotal role in O2 diffusion, which need further investigation.
In fact, the experimental results [35, 39, 40] are always lower than the calculated
results [34, 41, 42]. This is because that the widely used Bruggeman approximation
assumes that all the pores are well connected and the non-continuum events would
not occur [43]. However, the complex microstructure in CCL does not satisfy such an
ideal situation, as there are numbers of “dead pores” and “narrow corridors” in CCL.
As schematic described in Fig. 4.24, the non-effective pores and poorly effective
pores lead to huge diffusion barriers and cause such unacceptable errors [44, 45].
Gasteiger et al. attributed the discrepancy between the experimental and calculated
results to the difference of particle morphology and modified the Bruggeman relation.
Two additional proportionality factors, f and α, were introduced, and the revised
equation can be expressed as Eq. (4.16). However, this modification attributes to all
the complex structure to the effective tortuosity τ eff , which don’t work well in the
gas diffusion in CLs of PEMFCs, the one with numerous non-effective pores and
low-effective pores.

ε 1 1+α
DOeff2 = Do2 · = Do2 · ·ε (4.16)
τ eff f

Herein, we would like to propose another modification via introducing an effective

porosity (εeff ) which could be expressed as the product of porosity ε and an effective
coefficient (κ):

εeff = ε · κ (4.17)

For that the Bruggeman approximation is commonly used for porous electrodes
and implemented in commercial software packages (e.g., COMSOL Multiphysics
and ANSYS Fluent), the approximation is still employed into the modificatory
expression:
1.5
DOeff2 = Do2 · (ε eff ) (4.18)

Table 4.5 summaries the measured DOeff2 , ε from MIP, calculated εeff using
Eq. (4.18), and calculated κ using Eq. (4.17). The calculated εeff is far less than the
ε from MIP, which verifies the presence of the dead-end pores. And although ε only
152 4 Membrane Electrode Assembly (MEA)

Fig. 4.24 Schematic illustration on four kinds of pores. a Occluded pores; b non-effective pores;
c poorly effective pores; d highly effective pores

increases by 22.5%, while εeff increases by 50.5%. The disproportionate increase

implies that the porosity is not the only factor influencing the oxygen diffusion, even
not the major one. κ can be effected by various parameters, and it is believed that
the pore distribution also plays a pivotal role. But the change rule of κ needs further
investigation.
Furthermore, to verify the effect of pore-forming process on fuel cell performance,
ultralow Pt MEAs are measured. The cathode Pt loading of the two MEAs is 0.08 mgPt
cm−2 and 0.1 mgPt cm−2 , respectively. The reaction gas, stoichiometric ratio, RH,
operating pressure and temperature are H2 /air, 2/2, 100%, 150 kPaabs and 80 °C,
respectively.
As shown in Fig. 4.25, pore-forming process improves the performance signifi-
cantly, especially at high current densities for that the peak power density increases
from 436 to 592 mW cm−2 . It is worth to notice that at 1200 mA cm−2 , the oxygen
concentration polarization degrades the performance dramatically in the original
4.3 Oxygen Transport in Electrode 153

Fig. 4.25 Performance

comparison between the
pore-formed and original
electrodes [38]

electrodes, while the performance of the pore-formed electrodes is still in the ohmic
polarization region.

4.3.5 Mechanism of Local Oxygen Transport Behavior

Indeed, the dual-layer CCL design is an effective method to measure RLocal and RBulk ;
however, it needs heavy workload which could be avoid in certain situation. Here,
we would like to introduce a new method to obtain RLocal combining the limiting
current measurements with mathematical calculations. Based on the new method,
we systematically investigate on the influences of operating conditions on RLocal in
CCLs, i.e., dry mole fraction O2 (xO2 ).
Relationship between RCCL and RLocal can be expressed as Eqs. (4.19) and (4.20),
which is deduced by Greszler et al. [19]
 
4FcO2 h h
RCCL = = RLocal · · coth (4.19)
i lim

ψ = DOeff2 · h · RLocal (4.20)

Here, since there is no DCL at cathode, Eq. (4.4) turns to:

RTotal = RCCL + RGDL + RChannel (4.21)

where RGDL and RChannel can be measured [39] and calculated [24] as mentioned
above.
Combining Eqs. (4.19), (4.20) and (4.21) with the limiting current measurement,
we can obtain a more accurate RLocal simply. It is noted that, to figure out RLocal ,
154 4 Membrane Electrode Assembly (MEA)

DOeff2 and h should be provided. Herein, an appropriate value of 0.0006 cm2 s−1 was
employed for further calculation. Then, to deepen the understanding of influences of
operating conditions on local oxygen transport behavior, limiting current is measured
at various conditions, and the corresponding RLocal are calculated subsequently.
RLocal under different xO2 is shown in Fig. 4.26, and the RH, operating pressure and
temperature, are, respectively, 67% RH, 150 kPaabs and 80 °C. Standard deviations
are indicated by the gray error bar. It is clearly observed that RLocal increases with as
xO2 increases.
In fact, it could be analyzed in terms of the solution–diffusion model which
has been well acknowledged to describe the oxygen permeation through the PFSA
ionomer film covered on catalyst surface.
As shown in Fig. 4.27a, b, the local oxygen transport can be divided into three
processes, (1) oxygen adsorption at gas/ionomer surface, (2) oxygen diffusion inside
the ionomer film and (3) oxygen adsorption from the ionomer to Pt surfaces. The
adsorption at gas/ionomer surface can be described as Eqs. (4.22) and (4.23), while
processes (2) and (3) can be expressed as Eqs. (4.24) and (4.25), respectively.
gas/i
ceq = SO2 ·P O2 (4.22)

JO2 = −kgas/i · (ceq − cionomer ) (4.23)

cionomer − cPt
JO2 = −DOionomer · (4.24)
2
δi

JO2 = −ki/Pt · cPt (4.25)

where ceq , cionomer , cPt denote the oxygen equilibrium concentration, oxygen
gas/i
concentration at ionomer surfaces and oxygen concentration at Pt surfaces, SO2

Fig. 4.26 RLocal under

different xO2 [46]
4.3 Oxygen Transport in Electrode 155

Fig. 4.27 Schematic illustration on local transport [46]. a, b Process of oxygen permeation through
ionomer film; c typical oxygen concentration profile in ionomer film

denotes the oxygen solubility of oxygen, JO2 denotes the oxygen flux, kgas/i and
ki/Pt denote the apparent rate constant for the oxygen adsorption at the gas/ionomer
interface and ionomer/Pt interface, DOionomer
2
denotes the effective oxygen diffusion
coefficients inside ionomer, δi denotes the average thickness of ionomer film.
Combined Eqs. (4.22)–(4.25), rLocal can be expressed as:
  
δi 1 1 1 1
rLocal = + + · · (4.26)
DOionomer
2
kgas/i ki/Pt RT SDgas/i

Since the Nafion film covered on catalyst surface is ultrathin, it is believed that
oxygen adsorption at gas/ionomer surface dominates the local transport processes. In
low Pt loading MEAs, the effective Pt surface area decreases as Pt loading decreases;
thus, the oxygen adsorption on the ionomer is easier to become saturated, which
leads to a limit for the increase in the oxygen flux. According to the Faraday’s law
(Eq. 4.27), the limiting current increases proportionally as JO2 , but not xO2 .

i = 4F JO2 (4.27)
156 4 Membrane Electrode Assembly (MEA)

Table 4.6 Different kinds of

Isotherms Equations
isotherms summarized from gas
literatures Langmuir qe = qm K L C e
gas
1+K L Ce
gas 1/n
Freundlich qe = K F (Ce )
gas
Temkin qe = qm ln(K T Ce )
Dubinin–Radushkevich qe = qm exp(−Dε2 ),
gas −1
ε = RT ln(1 + Ce )
gas
Redlich–Peterson qe = ARP Ce
gas
1+BRP Ce
gas p
Koble–Corrigan qe = AKC Ce
gas p
1+BKC Ce
θ
Flory–Huggins c0 = K FH (1 − θ)n FH
gas n
Hill q SH C e H
qe = gas n
K D +Ce H

Sips gas β
K S Ce S
qe = gas β
1+K L Ce S
gas
Toth qe = K T Ce
gas 1/t
(aT +Ce )
gas
Khan qe =
qm b K C e
gas a
(1+bK C e ) K

Radke–Prausnitz gas β
aRP r R Ce R
qe = gas β
aRP +r R Ce R −1
gas
BET qe = qm CBET Ce
gas
(Cs −Ce )[1+((CBET −1)(Ce /C s )]
gas
FHH Ce α qm r
ln Cs = − RT ( qe d )

MET 1/3
qe = k
gas
ln(Cs /Ce )

gas/i
The saturation of oxygen on ionomer surface results in a decrease in SO2 , which
should be a constant according to Henry’s law. Actually, the experimental results
precisely suggest that Henry’s law is not as suitable in CCLs as researchers used
to regard. There exist a multiple of adsorption equilibrium isotherms which fit to
the oxygen adsorption at gas/ionomer surface as listed in Table 4.6. Several typical
adsorption equilibrium isotherms are also shown in Fig. 4.28.

4.4 Cation Contamination in Electrode

Dealloyed Pt-M nanoparticles have been paid more and more attention due to
their enhanced ORR activity resulting from the unique surface features such as Pt
sandwich-segregation, core–shell or spongy structure. However, the non-noble-metal
4.4 Cation Contamination in Electrode 157

Fig. 4.28 Different kinds of adsorption isotherms [46]. a Langmuir; b Freundlich; c Temkin;
d Dubinin–Radushkevich

cation tends to dissolve in to ionomer in CLs and further deteriorate the performance
and durability of MEAs.
On the one hand, the dissolution of non-noble-metal (M) deteriorates the charge
transfer kinetics of ORR and gas transport so that the fuel cell performance decreases
monotonically with increased contamination [47–52]. On the other hand, the disso-
lution of M cations suppresses anodic hydrogen oxidation reaction (HOR) kinetics
for that the dissolved M cations may deposit on the anodic electrocatalysts, then
cover the active reaction sites and further decrease the HOR kinetics [53]. Yu et al.
revealed that Cu2+ from cathode PtCu3 electrocatalysts could migrate through the
proton exchange membrane and plate as Cu metal at anode [54]. In addition, the disso-
lution of M cations exacerbates the ionic conductivity in both the proton exchange
membrane and PFSA ionomer film in CLs. The underlying reason is that proton trans-
port depends on the interaction between H+ and –SO3 − group in ionomer; however,
the dissolved M cations such as Fe3+ , Ni2+ and Cu2+ prone to replace the H+ and
occupy the –SO3 − group sites resulting, to a rapidly decreased proton transport
[55, 56].
Thus, it is imperative to deepen the understanding of cation contamination mecha-
nism in electrodes. Herein, dealloyed PtCu3 catalysts with various levels of Cu residue
158 4 Membrane Electrode Assembly (MEA)

Fig. 4.29 XRD profiles of

D-25-PtCu3 /HSC and
D-80-PtCu3 /HSC [57]

are employed to systematically investigate the effect of Cu contamination. The deal-

loyed PtCu3 catalysts supported on high surface area carbon (HSC) are synthesized
using impregnation–reduction method followed by acid treatment. And the catalysts
are named by as D-25-PtCu3 /HSC and D-80-PtCu3 /HSC based on different acid
treatment temperature, which leads to different Cu residue.
XRD profiles of two types of catalysts are shown in Fig. 4.29, while Pt and Co
are also shown in XRD profiles. It is clear that all reflections show positive shifts
with respect to Pt, but the shift of D-25-PtCu3 /HSC is higher relative to that of D-80-
PtCu3 /HSC, suggesting a larger lattice strain in D-25-PtCu3 /HSC. The XRD results
manifest that D-25-PtCu3 /HSC contains more Cu residue.
The ICP results further confirm the abovementioned conclusion. As shown in
Table 4.7, D-25-PtCu3 /HSC catalyst contains 22.55 wt% Pt and 3.89 wt% Cu, while
D-80-PtCu3 /HSC contains 22.40 wt% Pt and 1.72 wt% Cu. The atomic ratios of Pt
to Cu in D-25-PtCu3 /HSC and D-80-PtCu3 /HSC are calculated to be 1.89 and 4.24,
respectively.
Fuel cell performance of MEAs made with D-25-PtCu3 /HSC and D-80-
PtCu3 /HSC are shown in Fig. 4.30, and the Pt loading, reactant gas, operation temper-
ature, pressure and relative humidity are 0.07 ± 0.01 mgPt cm−2 at the anode and
0.18 ± 0.02 mgPt cm−2 at the cathode, H2 /air, 80 °C, 150 kPaabs , and 35% RH,
respectively. At low current densities (<200 mA cm−2 ), the MEA made with D-25-
PtCu3 /HSC catalyst present a higher performance, which is caused by the better ORR
mass activity. However, MEA made with D-80-PtCu3 /HSC gains better performance
at medium-to-high current densities (>400 mA cm−2 ), due to the different Cu residue

Table 4.7 Atomic

Catalysts Pt (wt %) Cu (wt %) Pt/Cu atomic ratio
compositions by ICP for
D-25-PtCu3 /HSC and D-25-PtCu3 /HSC 22.55 3.89 1.89
D-80-PtCu3 /HSC D-80-PtCu3 /HSC 22.40 1.72 4.24
4.4 Cation Contamination in Electrode 159

Fig. 4.30 Fuel cell performance for D-25-PtCu3 /HSC and D-80-PtCu3 /HSC [57]

in the used D-PtCu 3 /HSC catalysts which bring different level of Cu contamination
in the operating cell.
Figure 4.31 shows the SEM images of MEA cross sections and the corresponding
energy dispersive spectrometer (EDS). The thicknesses of anode and cathode are
measured to be 5.3 ± 0.7 µm and 13.5 ± 1.5 µm, respectively, and the signal of copper
element retained in the cathode. However, it is observed that the signal of copper
element in MEA with D-25-PtCu3 /HSC is stronger than that with D-80-PtCu3 /HSC,
due to the comparative Pt loading at the cathode of different MEAs and different
Pt/Cu atomic ratio between D-80-PtCu3 /HSC catalyst and D-25-PtCu3 /HSC catalyst.

Fig. 4.31 Cross-sectional SEM and EDS analysis of MEAs after operation [57]. a, b D-25-
PtCu3 /HSC catalysts; c, d D-80-PtCu3 /HSC catalyst
160 4 Membrane Electrode Assembly (MEA)

In addition, the copper element is also detected in both the membrane and anode,
where the distinction between D-25-PtCu3 /HSC and D-80-PtCu3 /HSC become more
evident.
It is noted that there is more Cu element in anode relative to membrane, which
is ascribe to the Cu/Cu2+ standard oxidation/reduction potential (0.34 V vs. normal
hydrogen electrode). Since the cathode electrocatalysts are usually exposed to air
atmosphere, Cu would dissolve as Cu2+ with different rate depending on the stability
of catalysts, then migrate through the membrane to the anode. After a long time of
operation, an increasing amount of Cu2+ dissolved out from cathode and went into
membrane and into anode then deposit as Cu in H2 atmosphere [54].
The current density versus applied voltage plot for hydrogen pump performed on
MEAs made with D-25-PtCu3 /HSC and D-80-PtCu3 /HSC are shown in Fig. 4.32,
and operation temperature and relative humidity 80 °C and 35% RH, respectively.
The applied voltage versus current density up to 1500 mA cm−2 presents linear
relationship, and the slope is about 67 and 49 m cm2 for D-25-PtCu3 /HSC and
D-80-PtCu3 /HSC, respectively.
Figure 4.33a, b show the electrochemical impedance spectroscopy (EIS) results,
and the operation temperature, reactant gas and relative humidity are 80 °C, H2 /N2
and 120% RH and 35% RH, respectively. Distinct differences are observed at both
RHs in the whole frequency region, while the high frequency resistance (HFR) can
be obtained from the intercept by extrapolating the high frequency end of EIS curve
to the real axis. The HFR includes both the ionic resistance of the membrane and
the electronic resistances [58] and is dependent on RH significantly, but independent
on catalyst types, indicating that the copper contamination in the membrane is not
significant due to the electric field effect between the anode and the cathode.
The value of HFR is about 46 m cm2 and 215 m cm2 at 120% RH and 35% RH,
respectively. In terms of MEA made with D-80-PtCu3 /HSC, the 46 m cm2 HRF at

Fig. 4.32 Applied voltage versus current density for a hydrogen pump [57]
4.4 Cation Contamination in Electrode 161

Fig. 4.33 EIS results of D-PtCu3 /HSC [57]. a Tested at 120% RH; b tested at 35% RH

120% RH (Fig. 4.33a) is close to the 49 m cm2 hydrogen-pump-measured resistance

(Fig. 4.32), suggesting that the MEA made with D-80-PtCu3 /HSC has a low level of
copper contamination at both membrane and electrode, and the HOR/HER overpo-
tential and the proton resistance in the electrode are negligible in the whole MEA at
the high RH. As for MEA made with D-25-PtCu3 /HSC, the higher hydrogen-pump-
measured resistance of about 67 m cm2 indicates that the HOR/HER overpotential
are non-negligible. In hydrogen pump experiments, Cu2+ would deposit to be Cu
since hydrogen gas flowing at the both electrodes. When the voltage is not applied,
the Cu2+ ions in the membrane will diffuse to the electrodes and be continuously
depleted. Once voltage being applied during the data collection, the residual Cu2+ in
the membrane will migrate from anode to cathode, and then deposit at the cathode.
Hence, Cu2+ content is kept at a very low level in the membrane and in the elec-
trodes, thus leads to unconspicuous influence on proton transport. But in the case of
D-25-PtCu3 /HSC, more Cu2+ dissolved out relative to that of D-80-PtCu3 /HSC, thus
more amount of Cu would deposit on HER active sites accordingly, which affect the
HER kinetics apparently.
The sheet resistance (Rsheet ) is calculated by fitting the AC impedance to a
transmission-line model [59], which is listed in Table 4.8. It can be obtained that
Rsheet in MEA made of D-25-PtCu3 /HSC is 3–5 times higher relative to that of D-
80-PtCu3 /HSC. At 120% RH, Rsheet is comparable to HFR in D-25-PtCu3 /HSC, but
at 35% RH, the proton resistance in the electrode is ~8 times higher than HFR in
D-25-PtCu3 /HSC.
Based on the above results, potential in the ionomer phase is depicted in
Fig. 4.34 to evaluate the influence of Cu2+ on the performance of cathode, anode

Table 4.8 Rsheet at different

Catalysts 120% RH (m cm2 ) 35% RH (m cm2 )
relative humidity condition
D-25-PtCu3 /HSC 54 1750
D-80-PtCu3 /HSC 11 507
162 4 Membrane Electrode Assembly (MEA)

Fig. 4.34 Schematic of MEA with Cu2+ contamination operating in H2 /air condition [57]

and membrane in operating PEMFCs. As mentioned previously, Cu2+ dissolves out

from D-PtCu3 /HSC at cathode, then migrate through the membrane to anodic cata-
lyst layer, concomitantly catalytically plate onto the anode Pt surface with the aid of
H2 reductant.
As shown in Fig. 4.35a, when the fuel cell is not in operation, the potential in the
ionomer phase keeps in equilibrium throughout the MEA. But once the cell begins
to work, an electric current runs through the MEA, as shown in Fig. 4.35b and c.
Since there is a proton transfer resistance in both electrodes and the membrane, a
potential drop along the direction from the anode through the cathode occurs.
In addition, the proton transport in the ionomer phase establishes an electric field
which leads to a Cu2+ accumulation at the lower potential region toward the cathode.

Fig. 4.35 Potential in the ionomer phase [57]. a No operating; b high humidity condition; c low
humidity condition
4.4 Cation Contamination in Electrode 163

Since the mobility of much less than proton, the potential decreasing amplitude in the
ionomer phase of the cathode is higher than that in membrane. As shown in Fig. 4.35c,
when the relative humidity keeps in a low level, the proton transfer resistance becomes
larger, causing a quicker potential drop along the current flow direction. Thus, there
are more serious Cu2+ accumulating toward the cathode, leading to an even steeper
potential drop in the cathode ionomer phase. As a result, Rsheet rises sharply and the
effective kinetic current density of ORR decreases accordingly [51], especially at
high current density.
Herein, it is essential to lower the negative effects of cation contamination in
PEMFCs, and several mitigation strategies are proposed as following:
(1) Increasing the proton content per unit volume of the cathode, i.e., optimizing
the mass ratio of ionomer to carbon in the electrode, or adopting an ionomer
with lower EW.
(2) Improving the Pt load on the anode to achieve an optimal balance between the
total Pt load and performance of the PEMFCs.
(3) Ameliorating the catalyst synthesis to reduce cation residual.

References

1. USDOE DOE hydrogen and fuel cells program record. https://www.hydrogen.energy.gov/pdfs/

15006_separation_distance_reduction.pdf.
2. Kusoglu A, Weber AZ (2017) New insights into perfluorinated sulfonic-acid ionomers. Chem
Rev 117(3):987–1104
3. Cetinbas FC, Ahluwalia RK et al (2017) Hybrid approach combining multiple characterization
techniques and simulations for microstructural analysis of proton exchange membrane fuel cell
electrodes. 344:62–73
4. Helmly S, Hiesgen R et al (2013) Microscopic investigation of platinum deposition in PEMFC
cross-sections using AFM and SEM. J Electrochem Soc 160(6):F687–F697
5. Klingele M, Breitwieser M et al (2015) Direct deposition of proton exchange membranes
enabling high performance hydrogen fuel cells. J Mater Chem A 3(21):11239–11245
6. Katayanagi Y, Shimizu T et al (2015) Cross-sectional observation of nanostructured catalyst
layer of polymer electrolyte fuel cell using FIB/SEM. J Power Sources 280:210–216
7. Suzuki T, Tsushima S et al (2010) Characterization of the PEMFC catalyst layer by cross-
sectional visualization and performance evaluation. ECS Trans 33(1):1465
8. Zils S, Timpel M et al (2010) 3D visualisation of PEMFC electrode structures using FIB
nanotomography. Fuel Cells 10(6):966–972
9. Singh R, Akhgar AR et al (2014) Dual-beam FIB/SEM characterization, statistical reconstruc-
tion, and pore scale modeling of a PEMFC catalyst layer. J Electrochem Soc 161(4):F415–F424
10. Ott S, Orfanidi A et al (2020) Ionomer distribution control in porous carbon-supported catalyst
layers for high-power and low Pt-loaded proton exchange membrane fuel cells. Nat Mater
19(1):77–85
11. Shen S, Han A et al (2019) Influence of equivalent weight of ionomer on proton conduction
behavior in fuel cell catalyst layers. J Electrochem Soc 166(12):F724–F728
12. Bass M, Berman A et al (2010) Surface structure of nafion in vapor and liquid. J Phys Chem
B 114(11):3784–3790
13. Iden H, Takaichi S et al (2013) Relationship between gas transport resistance in the catalyst
layer and effective surface area of the catalyst. J Electroanal Chem 694:37–44
164 4 Membrane Electrode Assembly (MEA)

14. Liu H, Epting WK et al (2015) Gas transport resistance in polymer electrolyte thin films on
oxygen reduction reaction catalysts. Langmuir 31(36):9853–9858
15. Nonoyama N, Okazaki S et al (2011) Analysis of oxygen-transport diffusion resistance in
proton-exchange-membrane fuel cells. J Electrochem Soc 158(4):B416–B423
16. Novitski D, Holdcroft S (2015) Determination of O2 mass transport at the Pt | PFSA Ionomer
Interface under reduced relative humidity. ACS Appl Mater Interfaces 7(49):27314–27323
17. Weber AZ, Kusoglu A (2014) Unexplained transport resistances for low-loaded fuel-cell
catalyst layers. J Mater Chem A 2(41):17207–17211
18. Suzuki T, Kudo K et al (2013) Model for investigation of oxygen transport limitation in a
polymer electrolyte fuel cell. J Power Sources 222:379–389
19. Greszler TA, Caulk D et al (2012) The impact of platinum loading on oxygen transport
resistance. J Electrochem Soc 159(12):F831–F840
20. Kudo K, Jinnouchi R et al (2016) Humidity and temperature dependences of oxygen transport
resistance of nafion thin film on platinum electrode. Electrochim Acta 209:682–690
21. Karan K (2017) PEFC catalyst layer: recent advances in materials, microstructural characteri-
zation, and modeling. Curr Opin Electrochem 5(1):27–35
22. Suzuki A, Sen U et al (2011) Ionomer content in the catalyst layer of polymer electrolyte
membrane fuel cell (PEMFC): effects on diffusion and performance. Int J Hydrogen Energy
36(3):2221–2229
23. Okumura M, Noda Z et al (2016) An FIB-SEM study on correlations between PEFC
electrocatalyst microstructure and cell performance. ECS Trans 75(14):347–354
24. Baker DR, Caulk DA et al (2009) Measurement of oxygen transport resistance in PEM fuel
cells by limiting current methods. J Electrochem Soc 156(9):B991–B1003
25. Wang C, Cheng X et al (2017) The experimental measurement of local and bulk oxygen
transport resistances in the catalyst layer of proton exchange membrane fuel cells. J Phys
Chem Lett 8(23):5848–5852
26. Wang C, Cheng X et al (2019) Respective influence of ionomer content on local and bulk oxygen
transport resistance in the catalyst layer of PEMFCs with low Pt loading. J Electrochem Soc
166(4):F239–F245
27. Lopez-Haro M, Guétaz L et al (2014) Three-dimensional analysis of Nafion layers in fuel cell
electrodes. Nat Commun 5(1):5229
28. Zhao J, He X et al (2007) Addition of NH4HCO3 as pore-former in membrane electrode
assembly for PEMFC. Int J Hydrogen Energy 32(3):380–384
29. Mu S, Xu C et al (2010) Accelerated durability tests of catalyst layers with various pore
volume for catalyst coated membranes applied in PEM fuel cells. Int J Hydrogen Energy
35(7):2872–2876
30. Uchida M, Fukuoka Y et al (1996) Effects of microstructure of carbon support in the catalyst
layer on the performance of polymer-electrolyte fuel cells. J Electrochem Soc 143(7):2245–
2251
31. Perry RH, Green DW et al (1997) Perry’s chemical engineers’ handbook, 7th edn. McGraw-Hill,
New York, pp 323R–325R
32. Pollard WG, Present RD (1948) On gaseous self-diffusion in long capillary tubes. Phys Rev
73(7):762–774
33. Lange KJ, Sui P-C et al (2011) Pore scale modeling of a proton exchange membrane fuel cell
catalyst layer: effects of water vapor and temperature. J Power Sources 196(6):3195–3203
34. Nam JH, Lee K-J et al (2009) Microporous layer for water morphology control in PEMFC. Int
J Heat Mass Transf 52(11):2779–2791
35. Yu Z, Carter RN et al (2012) Measurements of pore size distribution, porosity, effective oxygen
diffusivity, and tortuosity of PEM fuel cell electrodes. Fuel Cells 12(4):557–565
36. Beuscher U (2006) Experimental method to determine the mass transport resistance of a
polymer electrolyte fuel cell. J Electrochem Soc 153(9):A1788–A1793
37. Stumper J, Haas H et al (2005) In situ determination of MEA resistance and electrode diffusivity
of a fuel cell. J Electrochem Soc 152(4):A837–A844
References 165

38. Cheng X, Wang C et al (2019) Insight into the effect of pore-forming on oxygen transport
behavior in ultra-low Pt PEMFCs. J Electrochem Soc 166(14):F1055–F1061
39. Wang C, Cheng X et al (2017) The experimental measurement of local and bulk oxygen
transport resistances in the catalyst layer of proton exchange membrane fuel cells. J Phys
Chem Lett 8(23)
40. Shen J, Zhou J et al (2011) Measurement of effective gas diffusion coefficients of catalyst
layers of PEM fuel cells with a Loschmidt diffusion cell. J Power Sources 196(2):674–678
41. Berning T, Lu D et al (2002) Three-dimensional computational analysis of transport phenomena
in a PEM fuel cell. J Power Sources 106(1–2):284–294
42. Wang C-Y (2004) Fundamental models for fuel cell engineering. Chem Rev 104(10):4727–
4766
43. Pisani L (2011) Simple expression for the tortuosity of porous media. Transp Porous Media
88(2):193–203
44. Yu C (2011) Numerical analysis of heat and mass transfer for porous materials a theory of
drying. Tsinghua University Press, Beijing
45. Brownell L, Gami D et al (1956) Pressure drop through porous media. AIChE J 2(1):79–81
46. Shen S, Cheng X et al (2017) Exploration of significant influences of the operating conditions
on the local O2 transport in proton exchange membrane fuel cells (PEMFCs). Phys Chem
Chem Phys 19(38):26221–26229
47. Okada T, Dale J et al (1999) Unprecedented effect of impurity cations on the oxygen reduction
kinetics at platinum electrodes covered with perfluorinated ionomer. Langmuir 15(24):8490–
8496
48. Okada T, Ayato Y et al (2000) Oxygen reduction kinetics at platinum electrodes covered with
perfluorinated ionomer in the presence of impurity cations Fe3+, Ni2+ and Cu2+. Phys Chem
Chem Phys 2(14):3255–3261
49. Okada T, Ayato Y et al (2001) The effect of impurity cations on the oxygen reduction kinetics at
platinum electrodes covered with perfluorinated ionomer. J Phys Chem B 105(29):6980–6986
50. Li H, Gazzarri J et al (2010) PEM fuel cell cathode contamination in the presence of cobalt ion
(Co2+). Electrochim Acta 55(20):5823–5830
51. Kienitz B, Baskaran H et al (2007) A half cell model to study performance degradation of a
PEMFC due to cationic contamination. ECS Trans 11(1):777–788
52. Kienitz B, Pivovar B et al (2011) Cationic contamination effects on polymer electrolyte
membrane fuel cell performance. J Electrochem Soc 158(9):B1175–B1183
53. Pourbaix M (1966) Atlas of electrochemical equilibria in aqueous solutions. Pergamon press,
Oxford, 644 p
54. Yu Z, Zhang J et al (2012) Comparison between dealloyed PtCo3 and PtCu3 cathode catalysts
for proton exchange membrane fuel cells. J Phys Chem C 116(37):19877–19885
55. Kelly MJ, Fafilek G et al (2005) Contaminant absorption and conductivity in polymer electrolyte
membranes. J Power Sources 145(2):249–252
56. Okada T, Ayato Y et al (1999) The effect of impurity cations on the transport characteristics
of perfluorosulfonated ionomer membranes. J Phys Chem B 103(17):3315–3322
57. Zhu F, Wu A et al (2020) The asymmetric effects of Cu2+ contamination in a proton exchange
membrane fuel cell (PEMFC). Fuel Cells 20(2):196–202
58. Neyerlin KC, Gu W et al (2007) Cathode catalyst utilization for the ORR in a PEMFC: analytical
model and experimental validation. J Electrochem Soc 154(2):B279–B287
59. Makharia R, Mathias MF et al (2005) Measurement of catalyst layer electrolyte resistance in
PEFCs using electrochemical impedance spectroscopy. J Electrochem Soc 152(5):A970–A977
Chapter 5
Degradation of Pt-Based Catalysts
in PEMFC

Abstract Degradation of Pt-based catalysts is a major cause for the performance

loss of PEMFCs during normal operational conditions. To improve the durability of
PEMFCs, it is critical to fully understand the degradation process and mechanisms of
Pt-based catalysts. In this chapter, several degradation mechanisms for Pt catalysts are
investigated in detail, including particle growth due to Ostwald ripening and particle
agglomeration, Pt mass loss due to migration and particle detachment, and impurity
contamination, while for Pt-based alloy catalysts, the leaching of transitional metal
out of catalysts is discussed emphatically, which is another vital degradation mecha-
nism resulting in the loss of catalytic activity here. Furthermore, in-situ experimental
methods for observation of these processes and the recent mathematical methods for
modeling the degradation behaviors of Pt-based catalysts are concluded for further
understanding the degradation processes. In addition, several significant strategies
are proposed for mitigating the above mentioned degradation processes of Pt-based
catalysts, such as Pt catalyst alloyed or modified by proper metals, gradient design of
cathode catalyst layer, and optimizing the morphology, size and structure of Pt-based
alloy catalysts.

Keywords Degradation · Pt-based catalysts · XAFS analysis · Mathematical

modeling

5.1 Introduction

Durability is one of the biggest barriers to the commercialization of proton exchange

membrane fuel cells (PEMFCs) in vehicles. The degradation of PEMFC performance
has been widely proved to exist during the operation, especially for cathode catalyst
layer (CCL) with low or ultralow Pt loading. The CCL with low or ultralow Pt loading
would experience larger performance loss with the decay of electrochemical surface
area (ECSA) due to the sharply increased O2 transport resistance [1, 2]. However,
decreasing Pt loading is inevitable for reducing the cost, because as the catalyst cost
accounts for the largest proportion of the total cost of PEMFC stack. Thus, it brings

© Shanghai Jiao Tong University Press and Springer-Verlag GmbH Germany 2021 167
J. Zhang and S. Shen, Low Platinum Fuel Cell Technologies,
Energy and Environment Research in China,
https://doi.org/10.1007/978-3-662-56070-9_5
168 5 Degradation of Pt-Based Catalysts in PEMFC

Fig. 5.1 Schematic

representing the structure of
a single fuel cell

a bigger challenge for the durability of PEMFCs with comprehensive consideration

of the cost.
The performance loss during a long-term operation can be resulted from the
degradation of MEA, GDL, bipolar plate, etc. The schematic structure of a single
fuel cell is shown in Fig. 5.1. While MEA degradation is the main contributor for
fuel cell voltage loss during normal operation, here we mainly discuss some funda-
mental degradation processes in MEA structures, and conclude the in situ observation
researches and mathematical simulations for predicting and mitigating the catalyst
degradation in this chapter.
The degradation of MEA can be further divided into three parts, catalyst degrada-
tion, carbon corrosion, and ionomer and/or membrane degradation. Catalyst degra-
dation is the major cause during normal load cycling, as the accompanied voltage
cycling will lead to Pt dissolution and re-deposition, and subsequently Pt mass loss
by Pt ion diffusion into membrane, finally resulting in the ECSA loss. While during
startup/shutdown, the upper potential can reach as high as 1.5 V, which would lead to
pronounced carbon corrosion and further detachment of catalyst nanoparticles into
the ionomer. Without electronic contact with conductors, these Pt-based nanopar-
ticles are useless for catalytic reaction, so that causes a loss of effective catalytic
active sites. On the other hand, oxidation of carbon support can also lead to changes
in surface hydrophobicity, which would cause gas transport difficulties. In addition,
peroxide (HO·) and hydroperoxide (HOO·) radicals will be produced during the
chemical reaction on the surface of catalysts, which are widely believed to be the
critical issues for durability and reliability of the membrane, as these radicals can
attack the membrane and lead to membrane thinning or the formation of pinholes.
Otherwise, the presence of foreign cationic ions from the corrosion of metal bipolar
plates or leaching out of Pt-based catalysts during the operation can contaminate the
membrane, which results in attenuated water flux and proton conductivity.
5.2 Pt Degradation 169

5.2 Pt Degradation

While the surface of catalysts is the major electrochemical reaction site, a lot of
researches have been carried out to understand the fundamental mechanisms of
cathode catalyst degradation. Pt has excellent activity for the oxygen reduction reac-
tion (ORR) and therefore has been widely used as the cathodic catalyst in PEMFCs.
However, these catalysts tend to degrade under the conditions of high humidity, low
pH values and dynamic loads, which will then lead to the decay of PEMFC perfor-
mance. More seriously, the catalyst degradation will be further exacerbated when the
fuel cell vehicles are operating under more aggressive working conditions, such as
start-up/shutdown and freeze/thaw. The commercialization of PEMFC has brought a
big challenge for the durability of catalysts, which is required to withstand more than
5000 h for cars, 20,000 h for buses and 40,000 h for stationary applications. Unfortu-
nately, the lifetime of most MEA provided by manufacturers and research institutes
cannot reach these values at present. Therefore, it is critical to fully understand the
degradation mechanisms for improving catalyst durability.

5.2.1 Pt Degradation Mechanisms

Several degradation mechanisms have been proposed for pure Pt catalysts: (1)
particle growth due to Ostwald ripening and/or particle migration and coales-
cence/agglomeration on carbon support; (2) Pt mass loss due to Pt ion migration and
re-precipitation in the membrane by crossover hydrogen and/or detachment of cata-
lysts off carbon support induced by carbon corrosion; and (3) electrochemical active
sites poisoned by impurities from the fuel and air, and/or system-derived contami-
nants. These processes will lead to a loss of effective catalytic active sites (similar
to the meaning of ECSA) and a loss of effective catalyst mass, finally resulting in
apparent ORR activity loss in the CCL and voltage loss of PEMFC stacks during
long-term operation.

5.2.1.1 Ostwald Ripening and Particle Agglomeration

Nanoparticle coalescence/agglomeration and Ostwald ripening of Pt on the surface

of particles have been demonstrated to be one of the most dominant mechanisms for
catalyst degradation in PEMFCs, as shown in Fig. 5.2. Since lowering the Pt loading
is the major method for reducing the cost of fuel cells, the size of Pt catalyst used
in PEMFCs is usually required to be in nanometer scale with a diameter of 2~5 nm,
as the effective active surface area could be high enough for ORR in this range.
However, this brings another problem for the stability of nanoparticles. Nano-sized
particles are widely believed to have a strong tendency to agglomerate into bigger
particles due to their high surface energy [3]. And the smaller the nanoparticles are,
170 5 Degradation of Pt-Based Catalysts in PEMFC

Fig. 5.2 Schematic representing Ostwald ripening and Pt agglomeration on the surface of carbon
support

the higher the surface energy will be, thus the easier to agglomerate. It is worth to
mention that particle agglomeration happens not only in the preparation of MEA,
but also on the carbon surface during long-term or accelerated stress tests. While
particles agglomerate, the Pt effective active surface area decreases.
On the other side, Ostwald ripening of Pt on particle surface can also cause
the particle growth, which is resulted from Pt dissolution and re-precipitation. As
suggested by Darling and Meyers [4], Pt atoms can dissolve through two pathways,
directly electrochemical dissolution and/or chemical dissolution of Pt oxide. The
(electro)chemical reactions are shown as the following:
Pt electrochemical dissolution

Pt  Pt2+ + 2e− (5.1)

Pt oxide formation

Pt + H2 O  PtO + 2H+ + 2e− (5.2)

PtO + H2 O  PtO2 + 2H+ + 2e− (5.3)

Chemical dissolution of Pt oxide

PtO + 2H+  Pt2+ + H2 O (5.4)

PtO2 + 4H+ + 2e−  Pt2+ + 2H2 O (5.5)

5.2 Pt Degradation 171

So, which one plays the major role of Pt dissolution? It is mainly determined
by the cathode potential in PEMFC. Below 1.0 V, direct electrochemical dissolu-
tion contributes more, while Pt dissolves mainly through chemical dissolution of Pt
oxide when the potential is above 1.1 V, which can be concluded from the previous
researches [4–6]. The chemical dissolution rate is found to be much slower than elec-
trochemical dissolution rate with low Pt oxide coverage under 1.0 V [7, 8]. However,
with the increase of environment potential, the Pt oxide coverage will increase to
be even larger than 1.0 at 1.1 V [9, 10], which means that there is no pure Pt atoms
exposed to be directly dissolved and chemical dissolution becomes the only pathway.
Therefore, due to the voltage cycling accompanied by load cycle during normal
operation and the variance of particle size among catalysts, small particles will tend
to dissolve in the ionomer phase and then the dissolved Pt2+ will re-deposit on the
larger particles nearby, which then makes the average diameter of particles grow and
that is the so-called Ostwald ripening.

5.2.1.2 Platinum Loss and Migration

Pt loss during operation is another major source of catalyst degradation. Basically,

this is caused by Pt dissolution and deposition into the membrane by crossover
hydrogen (as shown in Fig. 5.3) or Pt particle detachment of carbon support into the
ionomer nearby. The process of Pt dissolution has been discussed above, which is
unavoidable in the cathode under the drive of high potential. While Pt atoms have
been dissolved in the hydrated ionomer, Pt ion transport would occur subsequently

Fig. 5.3 Schematic drawing of platinum deposition in the membrane [11]

172 5 Degradation of Pt-Based Catalysts in PEMFC

Fig. 5.4 TEM images of Pt band formation. a At the CCL/membrane interface [11] and b in the
membrane [16]

due to the gradient of Pt ion concentration between CCL and membrane, electrical
potential across the thickness of cathode and/or water flux-induced convection. Loss
of platinum across the CCL/GDL interface was usually ignored [12–15], as no Pt
can be detected in the effluent product water significantly. But in the membrane, Pt2+
could react with crossover hydrogen from the anode. The reaction is expressed as
the following:

Pt2+ + H2  Pt + 2H+ (5.6)

These processes would then result in the formation of a Pt band at the

CCL/membrane interface or inside of membrane after long-term operation. The TEM
images of Pt band are shown in Fig. 5.4. It is worthwhile to mention that the posi-
tion of Pt band is mainly determined by hydrogen and oxygen partial pressure, as
suggested by Zhang et al. [16].
On the other hand, the detachment of catalysts off carbon support can be attributed
to two reasons: (1) carbon corrosion at the site supporting Pt particles, where the
reaction enthalpy can be lowered in the presence of catalysts, and (2) a partial oxide
layer gradually forms between the carbon support and Pt nanoparticles during voltage
cycling, which results in the reduction of adhesion [17]. The IL-TEM micrographs
of the detachment of Pt particles from carbon support can be seen in Fig. 5.5.

5.2.1.3 Active Site Contamination

Contamination absorbing onto the electrocatalyst surface is another possible cause

of severe catalyst degradation by reducing the catalytic activity of surface Pt sites
or making some of them inactivated. Based on the sources, contamination can be
classified into two categories: The first source is the impurities presented in hydrogen
fuel and intake air; the second source is those derived from the fuel cell system, such as
metal ions or silicon from the corrosion of bipolar metal plates, catalysts and sealing
5.2 Pt Degradation 173

Fig. 5.5 IL-TEM micrographs of a catalyst region during voltage cycling between 0.4 and 1.4
V RHE [18]. a Before cycling; b after 3600 cycles; and c after 7200 cycles

gaskets. Actually, cations including metal ions mainly contaminate the ionomer in
CCL or membrane by ion exchange with protons combined with sulfonic acid group,
which increases the gas and proton transport resistance. For contaminants from the
fuel and air, the most common impurities and their sources are listed in Table 5.1.
The degradation mechanisms due to these impurities depend a lot on their type
and chemical characteristics, and correspondingly, the conclusion of whether the
performance loss due to these impurities is irreversible or reversible varies a lot, too

Table 5.1 Origin of common

Fuel for hydrogen Potential impurities
fuel and air impurities [19]
Hydrogen fuel impurities
Crude oil CO, NH3 , H2 S, HCN
Gasolines Hydrocarbons and aldehydes
Natural gas CO, NH3 , H2 S, HCN,
hydrocarbons, mercaptans
Methanol/DME CO, odorants, alcohols
Biomass Cations, aldehydes, alcohols,
formic acid, NH3 , H2 S, HCN
Water electrolysis Anions and cations
Air impurities
Fuel combustion pollution SOx , NOx , hydrocarbons,
soots and particulates
Ambient air, farming NH3
Natural sources Ocean salts and dust
Others
Deicers NaCl and CaCl2
FC system corrosion products Cations and anions
174 5 Degradation of Pt-Based Catalysts in PEMFC

[20]. For example, CO can form a strong bond with Pt atom and then the chemisorbed
CO will inhibit the hydrogen adsorption onto Pt sites for hydrogen electro-oxidation
(HOR), but it is reversible as this effect can be eliminated by enhancing the potential
to an acceptable value. However, for H2 S, even though the degradation is also due to
a strong chemisorption with Pt catalysts like CO, this poisoning process is relatively
irreversible, as it is hard to be electrochemically desorbed in a continuously operating
H2 /air fuel cell.

5.2.1.4 Effects of Working Conditions

Operational conditions can also have significant impact on the PEMFC durability,
such as the working temperature, humidity and cell voltage (the upper potential
limit and the potential cycling mode applied on the electrode). First of all, when the
UPL is lower than 1.1 V, evaluated temperature and/or upper potential limit (UPL)
will greatly promote the Pt dissolution, which then leads to an increase of catalyst
coarsening and further loss of Pt active surface. The corrosion of carbon support can
also be exacerbated with increasing temperature. Above 1.1 V, the Pt dissolving rate
decreases due to the formation of a more protective Pt oxide film. But meanwhile,
carbon corrosion is further deteriorated at higher temperature or UPL, which would
cause a severe detachment and agglomeration of catalyst nanoparticles, and plays a
major role in the performance loss. On the other hand, higher humidity can increase
Pt ion transport in the water channel networks and then the faster diffusivity of Pt
ions would result in a larger Pt mass loss, but carbon corrosion will increase with
decreasing humidity, and lower humidity will also induce larger voltage loss due to
the increasing ion transfer resistance.

5.2.2 Experimental Observation of Pt Degradation

Various experimental techniques have been applied to investigate the degradation

mechanisms of Pt catalysts. Even though we have concluded several possible degra-
dation mechanisms for catalysts in PEMFC, we still need more researches to reveal
which one plays a major role in the degradation of catalytic activity and the another
potential degradation processes which we have not found or understood before yet.
Ex situ techniques have been widely applied in PEMFC for over three decades,
which includes transmission electron microscopy (TEM), X-ray diffraction (XRD),
X-ray photoelectron spectrometer (XPS), etc. For instance, Meyer and co-workers
employed potential cycling experiments coupled with small-angle X-ray scattering to
study the ORR activity degradation for Pt nanoparticles and suggested that the atomic
dissolution of Pt is the dominant factor for the activity degradation [21]. Similar
results were also reported by Markovic et al. by using inductively coupled plasma
(ICP)-mass spectroscopy (MS) measurements [22]. Agglomeration and migration
of nanoparticle were also considered as an important degradation mechanism as
5.2 Pt Degradation 175

Fig. 5.6 XRD patterns for Pt/C electrocatalysts in MEAs at the 0 and 30,000 ADT cycles [25]

revealed by transmission electron microscopy (TEM) [23]. Further studies by X-ray

photoelectron spectrometer (XPS) in conjunction with electrochemical impedance
spectroscopy (EIS) concluded that the activity loss consists of two parts [24]. The
recoverable loss is attributed to the reduction of Pt oxide, while the unrecoverable
loss is related to the Pt dissolution/re-deposition, agglomeration and detachment.

5.2.2.1 Structure Evolution of the Pt/C Catalysts by XRD and TEM

The Pt/C electrocatalyst powders were assembled to MEAs and then subjected to an
activation process before accelerated degradation test (ADT) cycling tests. The ADT
was based on a steplike potential cycling (0.6–1.0 V vs. RHE at a triangle profile
with a scan rate of 50 mV s−1 ) using H2 /N2 (anode/cathode) with 100% RH at 80 °C.
Figure 5.6 presents the XRD patterns for Pt/C electrocatalysts. It is shown that the
Pt/C remains single-phase fcc structure at the 0 and 30,000 ADT cycles. Compared
with 0 cycles, the feature peaks shift to lower 2θ after 30,000 cycles, indicating an
increased lattice parameter and a possible compositional change during cycling.
The particle morphologies at the 0 and 30,000 ADT cycles are shown in Fig. 5.7.
The particle size increases after 30,000 ADT cycles, which indicates the occurrence
of an Ostwald ripening process during cycling.

5.2.2.2 In Situ Pt L3 -Edge XAFS Analysis of Pt/C Electrocatalysts

Although various mechanisms have been proposed previously, less in situ structural
observation, especially including the electronic and geometric structure information,
was performed to directly evident the ORR activity degradation mechanisms of Pt-
based electrocatalysts. X-ray absorption spectroscopy (XAFS) is a powerful tool to
probe the oxidation states and local chemical environments around the absorbing
atoms. Here, we present an investigation of the degradation of Pt/C electrocatalysts’
176 5 Degradation of Pt-Based Catalysts in PEMFC

Fig. 5.7 Representative TEM images and the corresponding histograms of particle diameter distri-
bution (insets) for Pt/C electrocatalyst in MEAs [25]. a At the 0 cycles and b 30,000 ADT
cycles

ORR activity under operating conditions of PEMFCs by in situ XAFS combined with
other ex situ techniques. The setup of in situ XAS measurement fuel cell is shown in
Fig. 5.8. The incident X-ray beam hits the MEA sample, which is composed of counter

Fig. 5.8 Schematic representing the setup of in situ XAS measurement fuel cell [25]
5.2 Pt Degradation 177

Fig. 5.9 XANES spectra at the Pt L3 -edge for Pt/C electrocatalysts at the 0 cycles and 30,000
cycles with different potentials [25]

electrode (CE), nafion membrane and working electrode (WE), and creates an absorp-
tion signal at the detector. WE, CE and reference electrode (RE) were connected
with the extra electrochemical workstation through Pt film due to investigation of
the potential dependence structure of the sample.

1. XANES spectra at the Pt L3 -edge.

The XANES spectra at the Pt L3 -edge for Pt/C electrocatalysts were recorded in situ at
selected voltage potentials (0.65, 0.9 and 1.2 V). Three potentials (0.65, 0.9 and 1.2 V)
were, respectively, selected according to the initial oxidation of catalyst surface, a
working state of fuel cell with a small current and low working load, and a full
oxidation of catalyst surface.

At 0 Cycles The spectra of Pt/C at various potentials are all similar to that of the
Pt foil, as demonstrated in Fig. 5.9a, and this indicates that Pt/C after cycling still
remains the fcc structure, which agrees with the XRD results (Fig. 5.6). The intensity
of the white line peak at 11,564 eV, correlating to the level of vacancy in the Pt 5d
orbitals near the Fermi level, increases slightly with increasing potential (inset in
Fig. 5.9a). The increasing vacancies in the 5d orbital may originate from the adsorp-
tion of oxygen species from the electrolyte following a Temkin isotherm behavior.
Additionally, an isosbestic point for Pt L3 -edge XAFS spectra illustrated the exis-
tence of only two phases for this system. The appearance of the isosbestic point
at 11,572 eV with the different voltage operations directly verified transformation
between an initial electronic state of Pt atoms at 1.2 V and a final electronic state at
0.65 V without any stable intermediate states.

At 30,000 Cycles The Pt L3 -edge XANES spectra at 30,000 ADT cycles for Pt/C
electrocatalysts are shown in Fig. 5.9b. Similar features as the 0 cycles can be
observed: (1) The white line intensity increases with increasing potential, and 2)
178 5 Degradation of Pt-Based Catalysts in PEMFC

there exists an isosbestic point at 11,572 eV. The intensity difference between 1.2
and 0.65 V reduces from 0.04 at the 0 cycles to 0.03 at 30,000 ADT cycles. This
may be related to the growth of nanoparticles, which decreases the specific surface
area and makes the structural changes and the chemisorption of oxygen species less
sensitive. This may be the reason for the slightly increased SA in Pt/C at the 30,000
ADT cycles.

2. Data fitting, coordination number and bond length.

In order to gain further structural information of the potential-dependent effects,
Pt L3 -edge XAFS data fitting for Pt/C electrocatalysts at various ADT cycles was
performed to extract the structural parameters of the metal clusters and the average
coordination number (CN) around Pt. Pt–Pt and Pt–O scatter paths were used to fit
the first shell of Pt absorber for Pt/C XAFS data. The fitting results are shown in
Table 5.2.

At 0 Cycles The Pt–Pt, Pt–O CN and R as a function of potential for Pt/C electro-
catalyst at the 0 cycles are shown in Fig. 5.10a. The CN for Pt–Pt drops from 10.5 at
0.65 V to 9.7 at 0.9 and 1.2 V, while that for Pt–O increases from 0.9 (0.65 and 0.9 V)
to 1.1 (1.2 V). The R for Pt–Pt and Pt–O has no significant variation with potential.

At 30,000 cycles Variations of CN and R of Pt–Pt and Pt–O at 30,000 ADT cycles
as a function of potential are shown in Fig. 5.10b. Similar features as the 0 cycles
can be observed: ➀ The total metal CN around Pt shows a decrease with increasing
potential, and ➁ the bond lengths for Pt–M and Pt–O have no significant variation
with potential. In contrast to the 0 cycles, the variations were moderate for Pt/C due
to the particle growth, which is in consistence with the Pt L3 -edge XANES results.
CNs are strongly dependent on particle size and therefore larger particles exhibiting
smaller changes in coordination number.

5.2.2.3 Degradation Mechanism of the Pt/C Electrocatalysts

As evidenced in Table 5.2, in contrast to the 0 cycles at 0.9 V, the significant decrease
in CN for Pt–Pt from 9.7 ± 0.4 to 8.8 ± 0.4 and the obvious increase in coordination
number of Pt–O from 0.9 ± 0.4 to 1.6 ± 0.3 can be investigated in the 30,000 ADT
cycles. There is no obvious variation in the distance of Pt–Pt (2.77 Å), whereas the
distance of Pt–O bonds significantly decreased from 1.98 to 1.92 Å after the 30,000
ADT cycles. These behaviors possibly indicate the oxidation of Pt during voltage
cycle.
From the surface composition point of view, for Pt/C electrocatalyst at the 0
cycles, the size of Pt nanoparticles with 4.37 nm (Fig. 5.7a) is regarded to constitute
approximately nine Pt layers, assuming sphere structures with the fcc arrangement,
and hence CN(Pt–Pt) is estimated to be 11.1 by the equation [12 * (12 + 22 + 32 +
42 + 52 + 62 + 72 + 82 ) + 9 * 92 ]/[12 + 22 + 32 + 42 + 52 + 62 + 72 + 82 +
 5.2 Pt Degradation

Table 5.2 Structure parameters for Pt/C electrocatalyst as obtained from in situ Pt L3 -edge XAFS experiments with different potentials at the 0 and 30,000
ADT cycles [25]
Pt/C Pt–Pt Pt–O R-factor
ADT cycles Potentials (V) CN R (Å) σ 2 (10−3 Å−1 ) CN R (Å) σ 2 (10−3 Å−1 )
0 cycles 1.2 9.7 ± 0.4 2.765 ± 0.002 5.13 1.1 ± 0.3 1.973 ± 0.023 6.30 0.008
0.9 9.7 ± 0.4 2.765 ± 0.002 5.12 0.9 ± 0.4 1.977 ± 0.032 6.30 0.010
0.65 10.5 ± 0.4 2.764 ± 0.002 5.08 0.9 ± 0.4 1.975 ± 0.036 6.30 0.009
30,000 cycles 1.2 8.4 ± 0.4 2.764 ± 0.002 5.25 1.7 ± 0.3 1.925 ± 0.016 10.77 0.004
0.9 8.8 ± 0.4 2.766 ± 0.001 5.20 1.6 ± 0.3 1.917 ± 0.015 10.60 0.002
0.65 8.9 ± 0.4 2.766 ± 0.005 5.30 1.4 ± 0.8 1.919 ± 0.039 10.10 0.004
Amp fixed at 0.84 for Pt, as obtained by fitting the Pt foil. Fits were done in R-space, K 1,2,3 weighting. For Pt, 1.5 < R < 3.2 Å and K = 3.6–13.5 Å−1 was used
179
180 5 Degradation of Pt-Based Catalysts in PEMFC

Fig. 5.10 Coordination number and bond length of the first shell of Pt absorber at the Pt L3 -edge
EXAFS spectra as a function of potential for Pt/C electrocatalysts at the 0 and 30,000 cycles [25]

92 ], where 12 is the CN of Pt atoms in fcc Pt bulk and 9 is the CN of Pt atoms at the

outermost surface. However, in the 0 cycles for Pt/C at 0.9 V, the CN of Pt–Pt and
Pt–O is, respectively, 9.7 and 0.9. Considering the formation of PtO phase (assuming
the percentage is X) in the surface layer of Pt nanoparticles, the CN (Pt–Pt) with 9.7
is estimated by the equation 11.1 * (1 − X) + [12 * (12 + 22 + 32 + 42 + 52 +
62 + 72 ) * X + 9 * 82 * X]/[12 + 22 + 32 + 42 + 52 + 62 + 72 + 82 + 92 ], and
the CN (Pt–O) with 0.9 is estimated by the equation [4 * 92 * X + 2 * 82 * X +
2 * 92 * (1 − X)]/[12 + 22 + 32 + 42 + 52 + 62 + 72 + 82 + 92 ]. The obtained
X was about 0.44, implied the formation of partial-surface PtO layer. After 30,000
ADT cycles, the size of Pt nanoparticles with 4.96 nm (Fig. 5.7b) is considered
approximately ten Pt layers. When adsorbed oxygen atoms are located only at the Pt
surface, almost no CN(Pt–Pt) should change, but actually the CN(Pt–Pt) decreased
drastically from 9.7 to 8.8 for Pt/C electrocatalyst at the 30,000 ADT cycles, which
indicates the presence of both surface oxygen and subsurface oxygen, forming a
full-surface PtO layer. By assuming a full coverage of tetragonal PtO phase at the
Pt nanoparticle surface, the CN(Pt–O) = [4 * 102 * X + 2 * 92 * X + 2 * 102
* (1 − X)]/[12 + 22 + 32 + 42 + 52 + 62 + 72 + 82 + 92 + 102 ] = 1.6, where the
obtained X is about 1.1 (>1.0). This result verified the formation of structure with
the full-coverage surface and partial-coverage subsurface PtO layer after the 30,000
ADT cycles. These results are consistent with XPS results for Pt/C electrocatalyst,
which indicated the shifted to higher energy at the 30,000 ADT cycles relative to
the 0 cycles (Fig. 5.11). The formed PtO at high potential (1.2 V) cannot reduce the
initial state at low potential (0.65 V) and therefore can be reduced at a lower potential
(<0.65 V). The similar hysteresis behavior in voltage up and down process also was
reported in the previous literature [26]. Therefore, the oxidation of Pt during the ADT
process induced the structure transformation from the partial-coverage surface PtO
layer to the full-coverage surface and partial-coverage subsurface PtO layer, resulting
in the drastic reduction of Pt active sites on Pt/C NP surface and consequently in the
decreased ORR activity.
5.2 Pt Degradation 181

Pt 4f 7/2
Pt 4f 5/2

30,000 cycles

Intensity/a.u 0 cycles

80 78 76 74 72 70 68
Binding Energy/ eV

Fig. 5.11 XPS spectra in the Pt 4f regions for Pt/C electrocatalysts at the 0 (black line) and 30,000
(red line) ADT cycles [25]

Fig. 5.12 Schematic showing the oxidation of Pt during the ADT process [25]

In conclusion, For Pt/C, XAFS results revealed that the oxidation of Pt during the
ADT process induced the structure transformation from the partial-coverage surface
PtO layer to the full-coverage surface and partial-coverage subsurface PtO layer,
resulting in the loss of Pt active site and consequently the decreased ORR activity,
as shown in Fig. 5.12.

5.2.3 Mathematical Modeling of Pt Degradation

Degradation of the Pt catalyst can be a complicated process, which is related to

various variables, such as humidity [27], temperature [28], upper potential limit
[28, 29] and voltage profiles [30–32]. However, many experiments are based on
postmortem observations under microscope; therefore, it is difficult to identify the
significance of each factor. Mathematical modeling is an alternative to study the
catalyst degradation and can provide many details that are difficult to be obtained
from experiments. It is effective, fast and efficient, and can provide many details
that are difficult to be obtained from experiments. It has been successfully applied
to investigate the loss of electrochemical surface area (ECA) of pure Pt catalyst
[4, 12–15, 33].
182 5 Degradation of Pt-Based Catalysts in PEMFC

However, the detailed processes are still unclear when it comes to the impact of
cycling profile on catalyst degradation, as it is related to two competitive processes:
Pt dissolution and Pt oxide formation/reduction. Pt oxide coverage could protect
the surface Pt atoms from dissolving, especially at high potentials [33, 34]. But Pt
dissolution can still happen when the oxide coverage fraction is lower than 100%
under 1.0 V. Therefore, different cycle profiles may accelerate/suppress the Pt disso-
lution by decreasing/increasing oxide coverage during voltage cycling in the range
of 0.6–1.0 V and then influence the Pt degradation.
Here, mathematical method is introduced to explore the fundamental processes
among these phenomena, and the evolutions of Pt dissolution and Pt oxide formation,
and ECSA loss due to Pt mass loss and Ostwald ripening are evaluated in detail,
respectively. The simulation results are validated by comparing with the previous
experimental results under a consistent condition.

5.2.3.1 Model Description

A one-dimensional Pt degradation model is employed for the simulation. The

following fundamental processes are mainly considered: Pt dissolution and re-
precipitation; Pt oxide formation and removal; diffusion of Pt2+ in the CCL; and
deposition of diffused Pt2+ by crossover hydrogen. In this model, the whole CCL
region is divided into N control volumes along the thickness. And in each control
volume i, particles are divided into M discrete particle size groups. For each particle
group, the diameter is assumed to be the same, in d i,j , and the corresponding particle
number is Numi,j , which can be calculated according to the experimental ECSA and
Pt loading before degradation. The oxide coverage of each particle group is defined
as θ i,j . The net Pt dissolution rate and the net oxidation rate of particle group d i,j can
be expressed as:
 
−H 1  
vPt (z, d, t) = v1  exp 1 − min(1, θi, j )
RT
⎛ ⎞
n 1 F(1 − β1 ) 4Pt γPt
⎜ exp − RT
Ueq −
di, j n 1 F
−V ⎟
⎜ ⎟
×⎜ ⎟ (5.7)
⎝ v2 cPt2+ n 1 Fβ1 4Pt γPt ⎠
− ref
exp Ueq − −V
v1 cPt2+ RT di, j n 1 F
−1
voxide (z, d, t) = v1∗  exp (H 2 + λθi, j )
RT
⎛ θi, j n 2 F(1 − β2 ) ωθi, j ⎞
1− exp − U2 + −V
⎜ 2 RT n2 F ⎟
⎜
×⎝ ⎟
v∗ θi, j n 2 Fβ2 ωθi, j ⎠
− 2∗ (10−2pH ) exp U2 + −V
v1 2 RT n2 F
(5.8)
5.2 Pt Degradation 183

where the equilibrium oxide potential (U 2 ) for nanoparticles is dependent on the

particle size d i,j , given by:

1 γPtO PtO γPt Pt

U2 = Ufit + − (5.9)
F di, j di, j

Thus, according to Eqs. (5.7–5.9), the evolutions of particle diameter d i,j and
oxide coverage θ i,j of each particle group in every control volume can be tracked by:

d(di, j )
= −vPt Pt (5.10)
dt
d(θi, j ) voxide 2θi, j d(di, j )
= − (5.11)
dt  di, j dt

In Eq. (5.7), the Pt2+ concentration in each control volume cPt2+ is solved from a
material balance equation:

∂cPt2+  
ε = ∇ · ε1.5 DPt2+ ∇cPt2+ + SPt2+ (5.12)
∂t

where the left term presents the evolution of Pt2+ concentration and the second term
shows the Pt ion diffusion loss, while SPt2+ is the source term produced by Pt net
dissolution, which can be written as:


M π
(di, j )2 Numi, j vPt
Si = 2
(5.13)
j=1
L A/N

where L and A represent the thickness and the area of CCL, respectively. To solve the
above differential equations, initial and boundary conditions should be given. The
initial oxide coverage θi, j for each particle group and the initial Pt2+ concentration
cPt2+ across the CCL are both assumed to be zero. At the CCL/GDL interface, we
assumed that there is no flux of Pt ions into the GDL:

∂cPt2+ 
=0 (5.14)
∂ x x=0

While at the CCL/membrane interface, a Pt band could be found under the H2 |N2
(anode/cathode) condition [11, 35]. Thus, Pt2+ concentration at this interface is
assumed to remain zero. But when under the H2 |air (anode/cathode) condition, a
Pt band was usually found in the membrane after long-term cycling and its distance
from the CCL/GDL interface can be calculated by [16]:

2K O2 pO2 δM
δPt = (5.15)
K H2 pH2 + 2K O2 p O2
184 5 Degradation of Pt-Based Catalysts in PEMFC

where pO2 and pH2 are the partial pressure of O2 and H2 , K O2 and K H2 are the
permeability of O2 and H2 in the membrane and δM is the thickness of the membrane.
As there is no Pt2+ production in the membrane, the Pt2+ concentration varies linearly
from the CCL/membrane interface to the Pt band; therefore, the Pt2+ concentration
at the CCL/membrane interface can be calculated by:


1.5 ∂cPt2+ c 2+
ε  = − Pt (5.16)
∂ x x=L δPt

Finally, the GSA and ECSA during the degradation can be calculated with the
derived particle diameter and number by:

N ,M
 2
di, j
GSA(t) = 2π Numi, j (5.17)
i=1, j=1
2

ECSA(t) ECSA(0)
= (5.18)
GSA(t) GSA(0)

It should be mentioned that in the actual simulation process, the corresponding

particle number Numi,j is assumed to be zero when its diameter is smaller than 0.5 nm
because the particles are too small to stay stable, which can also be speculated from
Eq. (5.7). Accordingly, the total Pt mass remaining can be calculated by:

N ,M

1
m Pt (t) = πρPt Numi, j di,3 j (5.19)
12 i=1, j=1

The ECSA losses due to Pt mass loss and Ostwald ripening are evaluated by:

ECSAloss,Pt = ECSA(0) × (m Pt (0) − m Pt (t)) (5.20)

ECSAloss,O.R. = (ECSA(0) − ECSA(t)) − ECSAloss,Pt (5.21)

5.2.3.2 Model Validation

The model validation is performed by comparing our simulation results with the
experimental data from Ferreira et al. [36] under 0.6–1.0 V versus RHE, 20 mV/s,
80 °C, fully humidified H2 |N2 (anode/cathode) condition, while the Pt loading also
keeps consistent at 0.4 mgPt /cm2 . The comparison of simulated and experimental
ECSA evolutions and Pt mass remaining on carbon support is shown in Fig. 5.13a
and b, respectively. It is found that the predicted ECSA evolution fits the experi-
mental one reasonably well, while a better agreement between the experimental and
5.2 Pt Degradation 185

Fig. 5.13 Comparison of the simulation results with the experimental data [39]. a ECSA evolutions;
b remaining Pt mass distribution; and c remaining particle number distribution

predicted Pt mass distribution after cycling is reached in the present work than Li
et al. [37]. The results show that Pt particles suffer more serious corrosion when
closer to the CCL/membrane interface, which then leads to a severe particle number
loss in the region near the CCL/membrane interface, as shown in Fig. 5.13c. While
few Pt particles can be found in this region, a Pt depletion region forms and then
causes a further voltage loss by increasing ion and mass transfer resistance during
the degradation [38].

5.2.3.3 Models for Pt Particle Detachment and Agglomeration

As discussed above, under the normal operational conditions (80 °C, fully humidified
with potential <1.0 V), the effects of nanoparticle detachment and agglomeration
on Pt degradation can be negligible. However, these processes should be taken into
consideration in mathematical modeling for some specific situations, such as for high-
temperature PEMFC, and/or with higher UPL during start-up/shutdown. Figure 5.14
shows an interaction of individual degradation mechanisms proposed by Kregar
et al. [40]. In their models, the number of detached particles was assumed to increase

Fig. 5.14 Schematics of Pt degradation mechanisms in the MEA [40]. a Interacting and
b intertwined
186 5 Degradation of Pt-Based Catalysts in PEMFC

linearly with the amount of corroded carbon, which is written as:

dNdet,i ṁ C,corr,i
= −kdet Ni (5.22)
dt m C,max,i

where k det is the physically motivated calibration parameter, m C,max,i is the mass of
corroded carbon required to detach particles, ṁ C,corr,i is the corrosion rate of carbon
mass and N i is particle number.
The detached particles merged with the attached particles were assumed to be
independent of their size, and the rate of attachment to particles with a size of Ri is
expressed by:

dNatt,i Ṅdet,tot
= M Ni (5.23)
dt i=1 Ni

And the merged particles appear in size class Ri obtained by:

dNmer,i Ri X s  Ṅdet,i Ṅatt,k

= (5.24)
dt n i R √ Ṅdet,tot
3 3 3
R j +Rk ≈Ri

On the other hand, on the basis of Smoluchowski’s coagulation equation [41],

random collisions are assumed and the coagulation rate kernel k cgl is constant. The
creation and extinction terms can be written by:

r
+ 1 r2  
J = kcgl f N (r 3 − r̃ 3 )1/3 , t f N (r̃ , t)d r̃ (5.25)
2 (r 3 − r̃ )
3 2/3
0+

r
−
J = kcgl f N (r, t) f N (r̃ , t)d r̃ (5.26)
0+

5.2.3.4 Effects of Pt Degradation Processes on PEMFC Performance

1. Effects of CO and CO2 contaminations on performance degradation.

During actual fuel cell operation, the presence of CO and CO2 in hydrogen fuel is
a normal phenomenon. A multiphase agglomerate catalyst model was proposed by
Abdollahzadeh et al. [42] for considering the effects of CO and CO2 presence in
the anode. Then, the kinetic mechanisms on the surface of anode catalysts in the
presence of CO and CO2 should take the processes of CO adsorption (Eq. 5.27) and
CO electro-oxidation (Eq. 5.28) into consideration, which are given by:
5.2 Pt Degradation 187

Pt + CO  Pt--CO (5.27)

Pt--CO + H2 O  Pt + CO2 + 2H+ + 2e− (5.28)

Therefore, the exchange current density in the anode should sum the currents
of electro-oxidation reaction for hydrogen, carbon monoxide and water, which are
given by:

ηFαa ηFαa
JH2 = 2keh θH aeff,Pt exp − exp − (5.29)
RT RT
ηFαCO ηFαCO
JCO = 2kec θCO aeff,Pt exp − exp − (5.30)
RT RT
ηF
JOH = aeff,Pt kl θHPt2 O θCO exp
2RT
  ηF
+k−l 1 − θHPt2 O (1 − θCO − θH )PCO2 exp − (5.31)
2RT

The impact of anode feeding different CO and CO2 concentrations on PEMFC

performance and durability is shown in Fig. 5.15. Small concentration of carbon
monoxide present is demonstrated to act as a poison of the Pt electrocatalyst in the
anode. Adsorbed CO can affect the activity of the electrode surface by preventing the
hydrogen adsorption. In addition, it is able to find that higher CO concentration will
not only enhance the performance loss but also decrease the durability of PEMFC.

2. Effects of ECSA loss on performance degradation.

By coupling the Pt degradation models with a transport model for fuel cell perfor-
mance, the effects of Pt degradation processes on PEMFC performance loss can be

Fig. 5.15 Impact of anode feeding a different CO on PEMFC performance and durability [42]
188 5 Degradation of Pt-Based Catalysts in PEMFC

Fig. 5.16 Polarization curves through current cycling test of 5000 cycles between 0.05 and
2.0 A/cm2 [43]

evaluated. The detailed model description can refer to the works by Li and Wang
[43] and Zhang et al. [44]. Accompanied with the ECSA loss in CCL (in Fig. 5.13),
the ORR kinetic loss and the microscale transport loss occur and both contribute to
the total voltage loss under moderate current condition, as demonstrated in Fig. 5.16.
While under high current density, the microscale oxygen transport resistance was
found to be enhanced and induces an additional interfacial concentration overpoten-
tial, which then plays the major role on performance loss. Meanwhile, for lower Pt
loading, particle degradation will become more serious and result in a higher increase
ratio of gas transport resistance [44].

5.2.4 Mitigation Strategies for Pt Degradation in the Cathode

Based on the Pt degradation mechanisms demonstrated above, numerous efforts

have been devoted to mitigating Pt degradation and maintaining the catalytic activity
during the operation.

5.2.4.1 Pt Catalyst Alloyed or Modified by Proper Metals

To mitigate the losses of catalytic activity due to Pt particle growth and migration, Pt
catalyst alloyed or modified by proper metals has attracted the interest of enormous
researchers. Especially, this strategy is also helpful for lowing the Pt loading and
the cost of fuel cell. However, both increased and decreased stabilities of Pt-based
electrocatalysts have been reported in the literatures for PEMFC application. After
preliminary screening, the proper elements to alloy with Pt are limited to a variety
5.2 Pt Degradation 189

Fig. 5.17 Comparison of ADF mapping and mass activity of dealloyed PtCo3 and PtCu3 catalysts
after cycling [45]

of transition metals, such as Co, Cu, Cr, Fe, Ni and Pd. Similar to Pt-based alloy
catalysts, modifying the surface of Pt catalysts can also enhance the durability of Pt/C
catalyst. Here, we would like to introduce these two strategies by sharing results from
our research group.

1. Comparison of the stability of alloying Pt with Co and Cu.

Alloying Pt with transition metals to fabricate Pt–M precursor and then dealloying
them by acid leaching are widely used processes to receive high and stable Pt–M
catalysts. In this part, the durabilities of dealloyed PtCo3 and PtCu3 catalysts were
compared by Yu et al. [45]. Figure 5.17 showed the comparison of annular dark field
(ADF) mapping and mass activity of the two catalysts after cycling. Dealloyed PtCo3
is shown to have lower pristine mass activity but higher durability.

2. Pt electrocatalysts modified using gold clusters.

Pt dissolution is a fundamental process which plays an important role in several

degradation mechanisms. Here, we demonstrate that platinum (Pt) oxygen reduction
fuel cell electrocatalysts can be stabilized against dissolution under potential cycling
regimes (a continuing problem in vehicle applications) by modifying Pt nanoparticles
with gold (Au) clusters. This behavior was observed under the oxidizing conditions
of the O2 reduction reaction and potential cycling between 0.6 and 1.1 V in over
30,000 cycles. The Au clusters on Pt nanoparticles were generated by underpotential
depositing on carbon-supported Pt nanoparticles through galvanic displacement by
Au of a Cu monolayer on Pt. The Pt surface area blocked by Au was determined by
the adsorption and oxidative desorption of CO. The CO stripping measurements on
Pt/C and Au/Pt/C were performed in N2 -saturated 0.1 M HClO4 , after the electrodes
190 5 Degradation of Pt-Based Catalysts in PEMFC

Fig. 5.18 CO stripping from the Au/Pt/C and Pt/C electrode surfaces [46]

were immersed in CO-saturated 0.1 M HClO4 and held at 0.25 V for 5 min, as shown
in Fig. 5.18, which revealed that the Au clusters in Au/Pt/C covered about 30 to 40%
of the Pt surface.
The stabilizing effect of Au clusters on Pt was determined in an accelerated
stability test by continuously applying linear potential sweeps from 0.6 to 1.1 V,
which caused surface oxidation–reduction cycles of Pt. The surface reaction involves
the formation of PtOH and PtO derived from the oxidation of water that causes the
dissolution of Pt via the Pt2+ oxidation state. We conducted the test by applying
potential sweeps at the rate of 50 mV/s to a thin-layer rotating disk electrode in an O2 -
saturated 0.1 M HClO4 solution at room temperature. For comparison, a Pt/C catalyst
with the same Pt loading as that in Au/Pt/C was subjected to the same potential cycling
conditions. After 30,000 cycles, changes in the Pt surface area and electrocatalytic
activity of the ORR were determined. The catalytic activity of Au/Pt/C, measured as
the currents of O2 reduction obtained before and after potential cycling, showed only
a 5-mV degradation in half-wave potential over the cycling period (Fig. 5.19a); in
contrast, the corresponding change for Pt/C amounts to a loss of 39 mV (Fig. 5.19c).
Voltammetry was used to determine the Pt surface area of the Au/Pt/C and Pt/C
electrodes by measuring H adsorption before and after potential cycling. Integrating
the charge between 0 and 0.36 V associated with H adsorption for Au/Pt/C shows
no change, indicating no recordable loss of Pt surface area (Fig. 5.19b). However,
for Pt/C, only ~55% of the original Pt surface area remained after potential cycling
(Fig. 5.19d).

In Conclusion Despite blockage of approximately one-third of the Pt sites on

Au/Pt/C by Au, the Au clusters still show huge stabilizing effect and lack of ORR inhi-
bition. Insignificant changes happen in the activity and surface area of Au-modified
5.2 Pt Degradation 191

Fig. 5.19 ORR electrochemical durability [46]. Polarization curves for the O2 reduction reaction
on a Au/Pt/C and c Pt/C catalysts on a rotating disk electrode, before and after 30,000 potential
cycles; sweep rate, 10 mV/s; and rotation rate, 1600 rpm. Voltammetry curves for Au/Pt/C (b) and
Pt/C (d) catalysts before and after 30,000 cycles, and sweep rate, 50 and 20 mV/s, respectively

Pt over the course of cycling, in contrast to sizable losses observed with the pure Pt
catalyst under the same conditions. In situ X-ray absorption near-edge spectroscopy
(Fig. 5.20) and voltammetry data suggested that the Au clusters confer stability by
raising the Pt oxidation potential.

5.2.4.2 Gradient Design of Particle Size and Pt Loading in CCL

Gradient design is one of the effective strategies to mitigate the Pt dissolution and
improve the performance and durability, too. In the previous researches, graded
catalyst loading in the CCL with higher loading at the inlet was demonstrated to
improve the performance and catalyst utilization [47–49]. Along the thickness direc-
tion, gradient CCLs with varying catalyst loading, particle size and ionomer content
have also been proved to be effective in improving the performance and/or durability
of fuel cells [35, 49–53]. And the effects of gradient design of particle size and Pt
loading on the ECSA loss and distributions of Pt active surface area and Pt mass
192 5 Degradation of Pt-Based Catalysts in PEMFC

Fig. 5.20 XANES analysis [46]. a XANES spectra obtained with the Au/Pt/C catalyst at the Pt L3
edge at four different potentials. b A comparison of the change of the absorption edge peaks of the
XANES spectra for Au/Pt/C and Pt/C as a function of potential, obtained with the electrocatalysts
at four different potentials in 1 M HClO4

during voltage cycling were explored in our recent researches [39]. The two-layer
CCL and three-layer CCL with gradient particle size, with larger Pt nanoparticles
closer to the membrane (MEM) but uniform Pt loading along the thickness, are shown
in Fig. 5.21a and b, respectively. Figure 5.21c presents the three-layer CCL with a
combination of gradient particle size and gradient Pt loading, while Pt loading in the
medium-particle region is the average value and in the smaller-particle region next
to GDL is larger. The properties of all the CCL structures studied in the following
discussion are summarized in Table 5.3.

1. Effect of increasing particle size near the CCL/membrane interface.

Fig. 5.21 Schematic views of the gradient CCL structures investigated in this part [39]. a Two-layer
CCL with gradient particle size; b three-layer CCL with gradient particle size; and c three-layer
CCL with a combination of gradient particle size and gradient Pt loading
5.2 Pt Degradation 193

Table 5.3 Description of the CCL structures in this work [39]

Model Characteristic
Layer number Region Pt loading Average particle
size (nm)
Uniform—3.0 nm 1 Total (12 µm) Uniform Pt 3.0
(12 µm) loading
3 nm (8 µm) + 2 GDL side (8 µm) Uniform Pt 3.0
4.0 nm (4 µm) MEM side (4 µm) loading 4.0
3 nm (8 µm) + 2 GDL side (8 µm) Uniform Pt 3.0
5.0 nm (4 µm) MEM side (4 µm) loading 5.0
3 nm (6 µm) + 2 GDL side (6 µm) Uniform Pt 3.0
4.0 nm (6 µm) MEM side (6 µm) loading 4.0
3 nm (6 µm) + 2 GDL side (6 µm) Uniform Pt 3.0
5.0 nm (6 µm) MEM side (6 µm) loading 5.0
3 nm (4 µm) + 3 GDL side (4 µm) Uniform Pt 3.0
4.0 nm (4 µm) + Middle (4 µm) loading 4.0
5.0 nm
(4 µm)/uniform Pt MEM side (4 µm) 5.0
loading
Pt loading—1.2 + 3 GDL side (4 µm) 1.2 times the 3.0
1.0 + 0.8 average
Middle (4 µm) 1.0 times the 4.0
average
MEM side (4 µm) 0.8 times the 5.0
average
Pt loading—1.4 + 3 GDL side (4 µm) 1.4 times the 3.0
1.0 + 0.6 average
Middle (4 µm) 1.0 times the 4.0
average
MEM side (4 µm) 0.6 times the 5.0
average

As larger catalyst nanoparticles have better stability and are less prone to the corrosion
during voltage cycling [15, 54], particles with average size of 4 and 5 nm are modeled
to be deposited near the CCL/membrane interface to mitigate the Pt degradation in
this region, respectively, while the average particle size in other regions is fixed at
3 nm. In order not to sacrifice the initial ECSA and therefore keep higher performance
at the beginning, particles with bigger size than 5 nm are not considered here and
these larger Pt catalysts are only applied in the Pt depletion region near membrane.
Figure 5.22 shows the effect of particle size in the region next to the membrane
on ECSA and normalized ECSA evolutions. As expected, using larger nanoparti-
cles decreases the initial ECSA but results in higher ECSA retention after cycling.
Compared with uniform cathode, applying 5-nm size catalyst nanoparticles near the
194 5 Degradation of Pt-Based Catalysts in PEMFC

Fig. 5.22 Effect of particle size in the region (4 µm) near the CCL/membrane interface [39].
a ECSA evolutions and b normalized ECSA evolutions

CCL/membrane interface shows the maximum ECSA of 30.4 m2 /gPt after cycling
(Fig. 5.30a), and the ECSA retention increases from 33 to 55% (Fig. 5.30b). However,
it is worthwhile to note that the final ECSA value for 3 nm (8 µm) + 5.0 nm (4 µm)
CCL is similar to that for 3 nm (8 µm) + 4.0 nm (4 µm) CCL while smaller at most of
the time during the degradation, which suggests that not bigger catalyst nanoparticles
near the membrane are better for optimal performance and durability.
The distributions of ECSA, remaining Pt active surface and Pt mass after cycling
along the cathode thickness for these CCLs are evaluated in Fig. 5.23a–c, respectively.
It is found that depositing bigger size Pt nanoparticles near the membrane can greatly
mitigate the ECSA loss and Pt mass loss in the region near CCL/membrane interface
and improve the total Pt mass retention in the meanwhile, as presented in Fig. 5.23d.
For catalyst layer with uniform-sized Pt particles, only 6.90% Pt active surface and
6.92% Pt mass remain in the nearest layer (1 µm) next to the membrane after 10k
cycles, whereas the remaining Pt active surface and Pt mass are increased to 90.07%
and 86.21%, respectively, by employing the gradient layer with 5-nm size Pt particles
next to the membrane. However, albeit not catastrophic, the loss of ECSA and Pt
mass is exacerbated in the smaller size region, which would be unfavorable to the
oxygen mass transport and then cause another performance loss. This may be the
main limitation for the cathode with a gradient particle size design. Consequently, to
improve the durability and performance of the catalyst layer, Pt nanoparticles with
suitable particle size should be used near the membrane.

2. Effect of layer number of gradient CCL.

Higher local Pt active surface near the CCL/membrane interface is helpful for
reducing ohmic loss due to proton transport, while improving total ECSA retention
is beneficial to decrease ORR activation overpotential [50]. As discussed in Sect. 3.2,
employing larger size Pt nanoparticles near the membrane can enhance the local and
overall ECSA retention. However, decreasing Pt active surface and Pt mass in the
GDL side and medium regions would increase the oxygen mass transport loss, which
plays a very important role in performance loss at high current density (>1.4 A cm−2 )
5.2 Pt Degradation 195

Fig. 5.23 Effects of enhancing the particle size in the region (4 µm) near CCL/membrane interface
on Pt degradation in the MEA after 10k voltage cycles [39]

[50, 55]. Consequently, the ideal Pt active surface distribution for gradient CCLs is
expected to keep more uniform during the degradation. We explored the effect of
layer number on ECSA and Pt mass loss during the degradation by dividing the CCL
into two or three parts, and each part deposited different size Pt catalysts.
Figure 5.24a shows the effect of layer number of gradient CCL on ECSA evolution.

Fig. 5.24 Effects of gradient layer number on Pt degradation during cycling [39]. a ECSA
evolutions and b normalized ECSA evolutions
196 5 Degradation of Pt-Based Catalysts in PEMFC

Increasing layer number of gradient CCL by using medium size catalysts in medium
region (4 µm) is found to decrease the initial ECSA slightly from 56 m2 /gPt for
3 nm (6 µm) + 4.0 nm (6 µm) CCL to 50 m2 /gPt , but lowers the rate of ECSA
decay and then enhances the final ECSA from 32.5 to 35.4 m2 /gPt . In Fig. 5.24b,
the normalized ECSA evolutions of the gradient CCLs with different layer numbers
during 10k cycles are presented, where the ECSA retention is further improved from
58.50 to 70.71% obviously. In this regard, employing three-layer gradient CCL can
improve the durability and performance by decreasing ORR activation overpotential.
Figure 5.25 compares the distributions of ECSA, percentage of remaining Pt
active surface and Pt mass across the CCL for gradient CCLs with different layer
numbers after 10k voltage cycles. Inserting a new layer with medium size catalyst
nanoparticles in the medium region of 3 nm (6 µm) + 5.0 nm (6 µm) CCL is
found to have little impact on the degradation of particles in the membrane side
region, but mitigates the loss of ECSA and Pt mass in the medium region, which is
supposed to be the main reason for the enhancement of total ECSA in Fig. 5.25a. The
minimum ECSA can be found to increase from 13.76 to 15.22 m2 /gPt , which then is
supposed to improve the fuel cell durability. While the remaining Pt active surface
and Pt mass distributions are more uniform with higher Pt active surface and Pt
mass concentrating in the medium region for the three-layer CCL after cycling from

Fig. 5.25 Effects of gradient layer number on Pt degradation in the MEA after 10k voltage cycles
[39]
5.2 Pt Degradation 197

Fig. 5.26 Effects of gradient Pt loading on Pt degradation in the MEA during cycling for the 3 nm
(4 µm) + 4 nm (4 µm) + 5.0 nm (4 µm) CCL [39]

Fig. 5.25b, c. Thus, the performance loss will be further reduced by minishing the
ohmic loss and oxygen mass transport loss meanwhile, even at high current density.
However, the Pt degradation in the GDL side region is accelerated and the EOT Pt
active surface here is very low comparing with other position. Therefore, the lifetime
of this three-layer gradient CCL is limited, once Pt mass near the GDL depletes or
reduces to a certain degree, and the performance of this CCL after cycling is still not
the optimum.

3. Effect of gradient Pt loading across the CCL.

To get a better performance during the degradation, gradient Pt loading is applied to

combine with three-layer gradient particle size in this section. Here, we further put
forward that higher Pt loading closer to the GDL should be applied for the gradient
CCL with larger particle size near the CCL/membrane interface.
Effects of gradient Pt loading on ECSA and normalized ECSA evolutions for the
3 nm (4 µm) + 4 nm (4 µm) + 5.0 nm (4 µm) CCL are illustrated in Fig. 5.26. With
higher Pt loading in the smaller-particle size region near the GDL, the initial ECSA
is expected to be increased. But the rate of ECSA degradation can be speculated to
be faster, as smaller particles are easier to dissolve. Increasing the Pt loading closer
to the GDL results in higher ECSA loss, but the ECSA after 10k cycles is found to be
similar in Fig. 5.26a. While according to Fig. 5.26b, the retention of ECSA decreases
from 70.71 to 65.59% for the Pt loading—1.4 + 1.0 + 0.6 CCL at the EOT. In fact,
the initial ECSA of fabricated gradient-Pt-loading CCL would be bigger, as the local
roughness factor (Pt active surface/GSA, m2 Pt /m2 ) is bigger in the region of high
density of nanoparticles, so that the final ECSA may be improved, too.
As discussed above, the EOT distributions of remaining Pt active surface, Pt mass
and particle number across the CCL for the cathode with different kinds of gradient
Pt loading are also critical to evaluate the cathode performance reasonably, which are
presented in Fig. 5.27. The Pt active surface and mass distributions after 10k cycles
are observed to be more uniform by increasing the Pt loading in the smaller size
region, while the final Pt active surface near the CCL/membrane interface seems to be
198 5 Degradation of Pt-Based Catalysts in PEMFC

Fig. 5.27 Effects of gradient Pt loading on Pt degradation for the 3 nm (4 µm) + 4 nm (4 µm) +
5.0 nm (4 µm) CCL after 10k voltage cycles [39]

relatively low for the cathode with only 0.6 times the average Pt loading in this region
in Fig. 5.27a. It may be more obvious that the 3 nm (4 µm) + 4 nm (4 µm) + 5.0 nm
(4 µm) CCL with 1.4 times the average Pt loading near the GDL performs worse
for the distribution of remaining particle number in Fig. 5.27c, as the initial particle
number decreases too much when the Pt loading is too low, especially in the larger
size region near the membrane. Consequently, in comprehensive consideration of the
distribution of remaining Pt active surface, Pt mass and particle number, employing
1.2 times the average Pt loading close to the GDL for the 3 nm (4 µm) + 4 nm (4 µm)
+ 5.0 nm (4 µm) CCL would be better for the performance after cycling.

In Conclusion The optimum gradient design is achieved by manipulating the

particle size gradient and Pt loading gradient simultaneously, which can gain
more uniform distribution of Pt active surface for improving the end-of-test (EOT)
performance and durability.
5.3 Pt-Based Alloy Degradation 199

5.3 Pt-Based Alloy Degradation

To lower the Pt loading and further enhance the ORR activity, Pt–M alloys (M is
the base metal, i.e., Co, Ni, Pd, etc.) are usually considered as the most promising
catalysts in the future. However, the Pt–M/C electrocatalysts usually show worse
durability than pure Pt/C catalysts, as the leaching of M metals for Pt–M alloys during
operation will induce the evolution of nanoparticle structure/composition resulting in
decreasing catalytic activity, and the M ions dissolved in ionomer would accentuate
the proton transport resistance meanwhile. Additionally, local gas transport resistance
can also be deteriorated for the CCL with these low or ultralow Pt loading Pt-based
electrocatalysts [56–58]. In the past decades, tremendous efforts have been made
to understanding the degradation of Pt–M alloy cathode catalyst. The formation or
growth of Pt shell is suggested to be responsible for the specific activity loss for
Pt–M catalysts in previous researches [59–61]. The variation of composition of the
Pt–M catalysts across the CCL after cycling is also considered important for the
performance degradation, which draws a lot of attentions [62–64].

5.3.1 Pt-Based Alloy Degradation Mechanisms

In addition to the degradation mechanisms discussed in Sect. 5.2.1, the dissolution

process of M metal out of Pt–M alloys is another vital degradation mechanism for
Pt-based electrocatalysts, which is then taken into consideration emphatically here.
However, it is far more complicated than Pt dissolution, which is still not fully
understood, as the leaching process of M atoms occurs not only on the surface but
also inside the nanoparticles during long-term operation [25]. To sum up, there are
three possible pathways for the dissolution of transition metal M. Firstly, for M atoms
exposed on the surface of nanoparticles, directly electrochemical dissolution would
happen in a much faster rate than Pt atoms:

Co  Co2+ + 2e− (5.32)

While for M atoms inside nanoparticles, they need to segregate from the bulk core
to the surface firstly, and the rate would be very slow because it requires substantial
activation energy [65], but this process can still play an important role in the Pt-
based alloy degradation during long-term voltage cycling. The second and third
dissolution pathways of M atoms then come from the reasonable speculations of
different segregation processes:
1. The second dissolution pathway is expressed by:

➀ Firstly, oxide species [O] absorbs on the particle surface and forms Pt–[O] (as
shown in Eq. 5.2), and then the oxidized species will gradually migrate into the
Pt particles by exchanging positions with the Pt atoms:
200 5 Degradation of Pt-Based Catalysts in PEMFC

Fig. 5.28 Model for transition development of oxide film in monolayer and subsequently
place-exchanged domains, initially prior to complete occupancy of the free metal surface by
electrodeposited OH and later O species [66]

[O]outside --Pt  Pt--[O]inside (5.33)

The place exchanging between Pt atoms and [O] species has been proved by
Conway et al. [66], as shown in Fig. 5.28.
➁ Inside the particles, [O] species would tend to combine with M atoms at higher
potential [67]:

Pt--[O]inside + M  Pt + M--[O]inside (5.34)

➂ Under lower potential, M–[O] then may be slowly dragged outside particles.
➃ Chemical dissolution of M–[O] on the particle surface:

[O]outside --M + 2H+  M2+ + H2 O (5.35)

2. The third dissolution pathway is expressed as:

➀ With [O] coverage on the particle surface, Co could migrate out of the core under
the dragging force of [O] and electric potential, due to the strong binding energy
between [O] species and Co atoms.
➁ While Co atom reaches the particle surface, it will then dissolve quickly
electrochemically (Eq. 5.32).
5.3 Pt-Based Alloy Degradation 201

Fig. 5.29 Segregation model and energy of a Pt nanoparticle in oxygen atoms [70]

There are also some researches supporting the third dissolution pathway. For
instance, Menning [68, 69], Noh [70] and Han [71] et al. put forward that the transition
metal M inside the particles will tend to displace the Pt on the surface and migrate to
the particle surface when the surface of alloy particles is covered by oxidizing species
in the oxidation environment, like in Fig. 5.29. Meanwhile, Dai et al. [65] reported
the surface composition and the dynamics involved in facet-dependent oxidation of
equilibrium-shaped Pt3 Co nanoparticles in oxygen atmosphere at 350 °C by in situ
TEM, which also indicates this Co segregation process. This was further verified by
Mayrhofer et al. [72] under normal PEMFC conditions.

5.3.2 Experimental Observation of Pt-Based Alloy

Degradation

Different from Pt/C electrocatalysts, the degradation processes of Pt-based alloys are
much more complicated, and the major degradation mechanism will also vary with the
structure, composition and morphology of these Pt-based electrocatalysts. Therefore,
to explore the detailed degradation processes of different Pt-based electrocatalysts,
it is optimal to combine in situ XAFS analysis with other ex situ techniques.
202 5 Degradation of Pt-Based Catalysts in PEMFC

In situ XAFS has been applied to investigating the effects of alloying of Pt with
3d transition metals (i.e., Co, Ni, Rh, etc.) on the oxidation states and local chem-
ical environments, which then would influence the ORR activity of catalysts. For
example, in situ XAFS analysis for Pt-based electrocatalyst indicated that the surface
Pt of catalysts was metallic at 0.4 V (vs. RHE) and oxidized to form disordered PtO
phase at 1.4 V (vs. RHE) [73]. Ishiguro et al. performed in situ time-resolved XAFS
to investigate the kinetics of Pt3 Ni/C and Pt3 Co/C compared to Pt/C catalysts at
molecular level. They revealed that the addition of the 3d transition metal hindered
the oxidation of Pt and significantly enhanced the reaction rates of Pt discharging,
Pt–O bond breaking and Pt–Pt bond reforming in the reductive process from 1.0 V
(vs. RHE) to 0.4 V (vs. RHE) [74]. Combined XAFS with DFT calculations, Huang
and co-workers revealed that the compressive strain and ligand effect in Rh-doped
Pt nanowires optimize the adsorption energy of hydroxyl and consequently enhance
the specific activity [75]. However, the evolution of Pt-based alloys during normal
voltage cycling is hardly probed by in situ XAFS. Here, we present a systemic inves-
tigation of the degradation of PtCo/C electrocatalysts under operating conditions of
PEMFCs.

5.3.2.1 Structure Evolution of the PtCo/C Catalysts by XRD and TEM

The PtCo/C electrocatalyst powders were degraded during potential cycling between
0.6 and 1.0 V at a triangle profile with scan rate of 50 mV s−1 under the conditions
of H2 /N2 (anode/cathode) with 100% RH at 80 °C. All the measurements were
performed at room temperature and under atmospheric pressure. The XRD patterns
for PtCo/C electrocatalysts at the 0 and 30,000 ADT cycles are shown in Fig. 5.30.
First of all, PtCo/C also remains single-phase fcc structure after 30,000 ADT cycling,
the same as Pt/C electrocatalysts. Compared with 0 cycles, the feature peaks also

Fig. 5.30 XRD patterns for PtCo/C electrocatalysts in MEAs at the 0 and 30,000 ADT cycles [25]
5.3 Pt-Based Alloy Degradation 203

Fig. 5.31 Representative TEM images and the corresponding histograms of particle diameter
distribution (insets) for PtCo/C electrocatalyst in MEAs before and after 30,000 ADT cycles [25]

shift to lower 2θ after 30,000 cycles. According to the particle morphologies at the
0 and 30,000 ADT cycles (in Fig. 5.31), the particle size increases as expected.

5.3.2.2 In Situ Pt L3 -edge XAFS Analysis of Pt/C Electrocatalysts

1. XANES spectra at the Pt L3 -edge.

Three potentials (0.65, 0.9 and 1.2 V) were, respectively, selected according to the
initial oxidation of catalyst surface, a working state of fuel cell with a small current
and low working load, and a full oxidation of catalyst surface.

At 0 Cycles The spectra for PtCo/C electrocatalyst are presented in Fig. 5.32a, which
show similar behavior to Pt/C electrocatalyst. The Pt L3 -edge white line peak for
PtCo/C at each potential is much stronger than that of Pt/C (in Fig. 5.9a), suggesting
a higher level of vacancy in the Pt 5d-band orbitals in PtCo/C. An increased 5d-orbital
vacancy may depend on the electronic configuration of the alloying element, and the
Co atom with more d vacancies than Pt. Furthermore, when the potential decreases
from 1.2 to 0.65 V, there is a change of 0.04 in the white line peak intensity for
Pt/C and 0.03 for PtCo/C. The peak intensity difference at these two potentials has
been reported as a surface-sensitive method to identify surface/adsorbate interactions
[76, 77]. A smaller value for PtCo/C indicates a lower affinity of oxygen species
chemisorption on the Pt in PtCo/C, in contrast to that in Pt/C. Therefore, higher Pt
5d-orbital vacancies and lower oxygen species chemisorption in Pt atom contributed
to the significant enhancement of the ORR activity for PtCo/C relative to Pt/C.

At 30,000 Cycles Figure 5.32b presents the Pt L3 -edge XANES spectra at 30,000
ADT cycles for PtCo/C electrocatalysts. The white line intensity increases with
204 5 Degradation of Pt-Based Catalysts in PEMFC

Fig. 5.32 XANES spectra at the Pt L3 -edge for Pt/C electrocatalysts before and after the 30,000
cycles with different potentials [25]

increasing potential, and there exists an isosbestic point at 11,572 eV, which are
similar to the features at 0 cycles. The intensity difference does not change with
cycling number (inset figures), confirming the lower ability to absorb oxygen species.

2. Data fitting, coordination number and bond length.

Pt–Pt, Pt–Co and Pt–O scatter paths were used to fit the first shell of Pt absorber for
PtCo/C XAFS data. The fitting result is shown in Table 5.4. The R-space experiment
and fitting data for PtCo/C electrocatalyst are shown in Fig. 5.33, respectively. The
corresponding CN and bond length (R) around Pt for Pt/C and PtCo/C electrocatalysts
as a function of potential at different ADT cycles were obtained and discussed below.

At 0 Cycles For PtCo/C at the 0 cycles (Fig. 5.34a), the total metal (Pt and Co)
CN of the first shell of Pt absorber (denoted as Pt–M) shows a slight decrease with
increasing potential, varying from 8.3 (0.65 V), 8.2 (0.9 V) to 7.9 (1.2 V), while that
for Pt–O increases from 0.8 (0.65 V), 0.8 (0.9 V) to 1.0 (1.2 V). The R for Pt–M and
Pt–O has no significant variation with potential from 0.65 to 1.2 V.

At 30,000 Cycles Variations of CN and R of Pt–Pt and Pt–O at 30,000 ADT cycles as
a function of potential are shown in Fig. 5.34b. The CN change of Pt–M was negligible
as a result of this potential change at the 30,000 ADT cycles, which indicated the
higher stability of PtCo/C electrocatalyst in contrast to Pt/C.

Notably, the Pt–Pt bond length in Pt/C and PtCo/C electrocatalysts at all ADT
cycles both are shorter than that in bulk platinum (2.78 Å). This behavior ascribed to
the compressive stress between Pt atoms which acts to modify its electronic properties
[78, 79]. However, the Pt–Pt bond distance in PtCo/C markedly reduced that of
Pt/C (Tables 5.2 and 5.4), which is consistence with XRD result. Besides, the Pt–O
distance is lengthened in PtCo/C compared to Pt/C. These behaviors in the PtCo/C
catalyst are indicative of a Co-induced change in the Pt electronic structure, which
Table 5.4 Structure parameters for PtCo/C electrocatalyst as obtained from in situ Pt L3 -edge XAFS experiments with different potentials at the 0 and 30,000
ADT cycles [25]
5.3 Pt-Based Alloy Degradation

PtCo/C Pt–Pt Pt–Co Pt–O R-factor

ADT Potentials CN R (Å) σ2 (10−3 Å−1 ) CN R (Å) σ2 (10−3 Å−1 ) CN R (Å) σ2 (10−3 Å−1 )
cycles (V)
0 1.2 6.3 ± 0.2 2.712 ± 0.002 6.39 1.6 ± 0.1 2.683 ± 0.004 6.88 1.0 ± 0.1 2.001 ± 0.004 3.47 0.001
cycles 0.9 6.5 ± 0.2 2.709 ± 0.002 6.07 1.7 ± 0.1 2.688 ± 0.003 5.97 0.8 ± 0.1 2.001 ± 0.006 3.47 0.002
0.65 6.8 ± 0.6 2.716 ± 0.006 6.39 1.5 ± 0.3 2.687 ± 0.016 6.88 0.8 ± 0.2 2.007 ± 0.017 3.47 0.009
30,000 1.2 6.5 ± 0.2 2.724 ± 0.002 5.88 1.3 ± 0.1 2.681 ± 0.006 6.00 0.9 ± 0.1 2.007 ± 0.006 3.92 0.001
cycles 0.9 6.5 ± 0.2 2.722 ± 0.002 6.20 1.3 ± 0.1 2.687 ± 0.004 6.00 0.8 ± 0.1 2.006 ± 0.005 3.67 0.003
0.65 6.8 ± 0.4 2.725 ± 0.002 5.75 1.3 ± 0.1 2.681 ± 0.005 6.00 0.7 ± 0.1 2.010 ± 0.007 3.92 0.001
Amp fixed at 0.84 for Pt, as obtained by fitting the Pt foil. Fits were done in R-space, K 1,2,3 weighting. For Pt, 1.4 < R < 3.1 Å and K = 2.8–2.5 Å−1 was used
205
206 5 Degradation of Pt-Based Catalysts in PEMFC

PtCo/C 0 cycles Exp (solid line) PtCo/C 30,000 cycles Exp (solid line)
Fit (dash line) Fit (dash line)
Amplitude (a.u.)

Amplitude (a.u.)
1.2V 1.2V

0.9V 0.9V

0.65V 0.65V

1 2 3 4 1 2 3 4
R(Angstrom) R(Angstrom)

Fig. 5.33 R-space experiment (solid line) and fitting data (dash line) of Pt L3 -edge XAFS for
PtCo/C electrocatalyst with the different potentials at the 0 and 30,000 ADT cycles [25]

12 12
PtCo/C 0 cycles (a) (b)
2.8 2.8
Coordination number

Coordination number

10 R(Pt-M) 10
2.7 2.7
Bond length (Å)

Bond length (Å)

8 8
CN(Pt-M)
2 R(Pt-O) 2
2.0 2.0

CN(Pt-O)
0 1.9 0 1.9
0.6 0.9 1.2 0.6 0.9 1.2
Potential (V vs RHE) Potential (V vs RHE)

Fig. 5.34 Coordination number and bond length of the first shell of Pt absorber at the Pt L3 -edge
EXAFS spectra as a function of potential for PtCo/C electrocatalysts at the 0 cycles and the 30,000
cycles [25]

results in a weakened Pt–O bond [80]. Additionally, the Pt–O coordination numbers
in PtCo/C electrocatalyst lower than those of Pt/C electrocatalyst at all ADT cycles,
which also indicated that the existence of Co inhibited the oxidation of Pt atom, and
then the decreased coverage of oxygen species. Therefore, the shorter Pt–Pt bond
distance induced by the addition of Co also enhanced the ORR activity of PtCo/C
electrocatalyst relative to Pt/C electrocatalyst.

5.3.2.3 Degradation Mechanism of the PtCo/C Electrocatalysts

Direct structure parameter information related to the CN and R around Pt atom for
two electrocatalysts at 0.9 V function as the ADT cycles is necessary to investigate the
correlation between the structure and ORR activity of electrocatalyst as evidenced in
5.3 Pt-Based Alloy Degradation 207

Table 5.4. With the increase in number of ADT cycles from 0 to 30,000, the reduction
in the CN of Pt–Co bond from 1.7 ± 0.1 to 1.3 ± 0.1 and the relaxation in the distance
of Pt–Pt from 2.71 to 2.72 Å were found, which possibly illustrated the oxidation
and dissolution of Co atoms.
To further indicate the oxidation and dissolution of Co atoms, XANES spectra
at the Co K-edge for PtCo/C electrocatalyst at various ADT cycles as function of
different potentials (1.2 V and 0.9 V, whereas 0.65 V not collected) were performed.
Co–Co, Co–Pt and Co–O scatter paths were used to fit the first shell of Co absorber
for Co K-edge XAFS data in the PtCo/C electrocatalyst at all ADT cycles at 0.9 V,
and results are shown in Table 5.5. The R-space experiment and fitting data of Co K-
edge XAFS for PtCo/C electrocatalyst are shown in Fig. 5.35. Co K-edge XAFS data
for PtCo/C electrocatalyst at 1.2 V cannot be successfully fit due to the poor-quality
spectra.
Figure 5.36 shows XANES spectra at the Co K-edge for PtCo/C electrocatalyst
at 0.9 V as function of ADT cycles. As shown in Fig. 5.36, there are four feature
peaks in Co K-edge XAFS for PtCo/C electrocatalyst, respectively, labeled as A,
B, C and D, which identified the formation of Pt–Co alloy. Compared with the Co
foil, features C and D are shifted to lower energies in the PtCo/C electrocatalyst
due to the expansion of the Co–M (M: Co and Pt) bond distances, indicating that
Co involved in a bimetallic phase and alloying with Pt possessing a larger atomic
radius. These feature peaks at 7713 eV (A) and 7725 eV (B) for Co K-edge XANES
mainly describe to 3d and hybridized 4s, and 4p band occupancy, respectively. In
contrast to the Co K-edge XAFS of Co foil, the shape of these feature peaks (B and
C) for PtCo/C electrocatalyst with all ADT cycles is similar to that of CoO reference,
which indicated the significant oxidation of Co and the formation of CoO. A gradual
decrease in the intensity of A peak and a pronounced increase in the tensity of B
and C peaks for Co K-edge XANES with the increase in number of ADT cycles
verified a progressive Co oxidation. To obtain detailed Co local environment, CN
and R around Co for PtCo/C electrocatalyst at 0.9 V with different ADT cycles are
shown in Table 5.5. The simultaneously decreased Co–Co and Co–Pt CN and the
increased Co–O CN indicated the oxidation of Co and the formation of CoO during
the ADT cycles going from 0 to 30,000 ADT cycles. This behavior also reflected
by the decreased distance of Co–O bonds from 2.14 to 2.13 Å, which indicated the
enhanced interaction between Co and O. Therefore, the attenuation of strain and
ligand effects derived from the formation of CoO resulted in the decreased ORR
activity for PtCo/C.
The Co K-edge XAFS at various ADT cycles function as the potentials were
performed and results shown in Fig. 5.37. As can be seen in Fig. 5.37a, the amplitude
of Co K-edge white line peak significantly increased with the increased potential,
which induces a progressive Co oxidation at the higher potential. Besides, it should
be also noted that, in the Co K-edge XANES range, potential dependence is so
pronounced as in the case of Pt L3 -edge, which verified that Pt local environment
was stable, whereas the local geometric and electronic structures of Co showed a
remarkable change. These results also demonstrated that cobalt atoms provide a
better center for oxygen species (some oxygen-containing functional groups, such
 208

Table 5.5 Structure parameters for PtCo/C electrocatalyst as obtained from in situ Co K-edge XAFS experiments with different potentials at the 0 and 30,000
ADT cycles [25]
PtCo/C Co–Co Co–Pt Co–O R-factor
ADT Potentials CN R (Å) σ 2 (10−3 Å−1 ) CN R (Å) σ 2 (10−3 Å−1 ) CN R (Å) σ 2 (10−3 Å−1 )
cycles (V)
0 0.9 1.0 ± 0.1 2.666 ± 0.006 1.90 6.1 ± 0.4 2.688 ± 0.003 5.97 0.9 ± 0.1 2.141 ± 0.014 6.22 0.002
cycles
30,000 0.9 0.8 ± 0.1 2.664 ± 0.008 0.98 5.6 ± 0.5 2.687 ± 0.004 6.00 1.5 ± 0.1 2.134 ± 0.010 3.94 0.003
cycles
5 Degradation of Pt-Based Catalysts in PEMFC
5.3 Pt-Based Alloy Degradation 209

PtCo/[email protected] Exp (solid line)

Fit (dash line)

Amplitude (a.u.)
0 cycles

30,000 cycles

1 2 3 4
R(Angstrom)

Fig. 5.35 R-space experiment (solid line) and fitting data (dash line) of Co K-edge XAFS for
PtCo/C electrocatalyst at 0.9 V at the 0 and 30,000 ADT cycles [25]

1.8
PtCo/[email protected]
1.5
Normalized (E)

1.2 D
BC
0.9

0.6 Co foil
0 cycles
A
0.3 30,000 cycles
CoO
0.0
7700 7710 7720 7730 7740 7750 7760
Energy (eV)

Fig. 5.36 XANES spectra at the Co K-edge for PtCo/C electrocatalyst at 0.9 V with the different
ADT cycles [25]

as O, OH or OOH) adsorption than Pt, which leads to more Pt active sites for the
adsorption of molecular oxygen (O2 ) and then enhances their ORR activity. Further,
the same tendency related to the potential dependence on the amplitude of Co K-edge
XANES was observed on the 30,000 ADT cycles and results are shown in Fig. 5.37b.
However, in contrast to the 0 cycles, the amplitude difference at the two potential of
0.9 and 1.2 V significantly decreased after 30,000 ADT cycles, which indicated a
slight reduction of cobalt oxide with the potential going from 1.2 to 0.9 V for PtCo/C.
The formed CoO at high potential would be dragged to the particle surface and then
was dissolved in acid environment resulting in CoO segregation during ADT process.
These results are fair agreement with ICP-AES results for PtCo/C electrocatalyst,
which exhibited that the molar Co: Pt ratio decreased after 30,000 ADT cycles.
Therefore, after 30,000 ADT cycles, the dissolution and segregation of CoO caused
the increased Pt outer layer thickness (> a few Pt atomic layers) also determined
210 5 Degradation of Pt-Based Catalysts in PEMFC

(a) (b)
1.8 1.8
PtCo/C 0 cycles PtCo/C 30,000 cycles
1.6 1.6
1.4 1.4

Normalized (E)
Normalized (E)

1.2 D 1.2 D
1.0 B C 1.0 B C
0.8 0.8
0.6 Co foil 0.6 Co foil
A CoO A CoO
0.4 0.4
1.2 V 1.2 V
0.2 0.9 V 0.2 0.9 V
0.0 0.0
7700 7710 7720 7730 7740 7750 7760 7700 7710 7720 7730 7740 7750 7760
Energy (eV) Energy (eV)

Fig. 5.37 XANES spectra at the Co K-edge for PtCo/C electrocatalyst at the 0 and 30,000 ADT
cycles with different potentials [25]

the drastic drop of ORR activity for PtCo/C electrocatalyst. Previous literatures also
reported that the ORR activities of core–shell nanoparticles gradually decreased with
the increasing Pt shell thickness resulting from the leaching of the non-precious metal
core [81, 82].

In Summary For Pt–Co/C electrocatalyst, XAFS results indicated the reduced Pt–
Co CN and the relaxation Pt–Pt bond after 30,000 ADT cycles, which illustrated the
oxidation and dissolution of Co atoms, as shown in Fig. 5.38. The oxidation of Co
atom induced the attenuation of strain and legend effects, and the dissolution and
segregation of CoO caused the increased Pt outer layer thickness during the ADT
process mainly determined the ORR activity degradation for PtCo/C.

Fig. 5.38 Schematic showing the degradation processes of PtCo/C electrocatalysts during the ADT
process [25]
5.3 Pt-Based Alloy Degradation 211

5.3.3 Mathematical Modeling of Pt–Co Core–Shell Catalyst

Degradation

5.3.3.1 Model Description

A schematic diagram of the Pt3 Co degradation model and the detailed corrosion
mechanism for Pt3 Co particles is shown in Fig. 5.39. A two-phase core–shell struc-
ture, first proposed by Strasser et al. [83], is applied for Pt3 Co particles. According
to previous reports, each Pt–M catalyst particle has one single core [85] and could
remain non-porous after potential cycling if the particles are smaller than 10 nm [86].
Therefore, in the present model, all the particle sizes are below 10 nm. Additionally,
all the Pt3 Co particles are assumed to be semispherical and uniform across the plane.
The initial thickness of the Pt shell in each particle is assumed to be independent of
the particle size [61].
Three parts are included in this model: Pt shell oxidation model, Pt and Co dissolu-
tion model and complete degradation model. Pt shell oxidation model and Pt and Co
dissolution model are used to predict the oxide fractional coverage, particle size and
Pt shell thickness for each particle diameter group. A complete degradation model is
implemented to explain how these two parts collaboratively influence the evolution
of catalyst structure and how the model is initialized, evolved and analyzed.

1. Pt shell oxidation and dissolution models.

For Pt on the shell, the Pt oxidation and dissolution processes will be similar to those
for pure Pt catalysts; therefore, the oxide formation and reduction rate can also be
written as Eq. (5.8), and the Pt dissolution rate is expressed by Eq. (5.7), too.
2. Co dissolution model.

Fig. 5.39 Schematic of a 1D Pt3 Co degradation model and corrosion mechanism for Pt3 Co particles
in a PEMFC cathode [86]
212 5 Degradation of Pt-Based Catalysts in PEMFC

For the bimetallic core–shell catalyst, particle sizes and structures will evolve during
electrochemical aging. The change of Pt shell thickness will influence the surface
electronic structure and the lattice strain on the shell to affect the catalytic activity
and the equilibrium potential for Pt dissolution on the Pt shell. Consequently, the
equilibrium dissolution potential for a Pt–Co particle with particle size D and Pt
shell thickness L Pt is modified to be [86]:
 4Pt γPt
Ueq − ξ L Pt − , L Pt ≤ 6 ML
U1 = Di, j n 1 F
4Pt γPt (5.36)
Ueq − ξ L Pt (= 6 ML) − Di, j n 1 F
, L Pt ≥ 6 ML

And for simplification, the Co dissolution process is assumed to mainly proceed

through the first pathway. When Co atom is exposed to the surface, it will dissolve
quickly as it is thermodynamically unstable (dissolution equilibrium voltage U 0 =
0.28 V). Therefore, when L Pt ≤ 1 ML, Co dissolution rate vCo can be expressed as:

d(di, j )
= −(vCo Co + nvCo Pt ) = −vPt Pt (5.37)
dt
where d is the diameter of the core and n is the atomic ratio of Pt/Co in the core.
When L Pt ≥ 1 ML, vCo is assumed to be zero. Then, vCo is:

vPt · (Co+n
Pt
Pt )
, L Pt ≤ 1 ML
vCo = (5.38)
0, L Pt ≥ 1 ML

3. Models for the evolution of particle structure.

The Pt shell for the starting PSD is 2.6 ML for every particle diameter group L i, j
(1 ML is about 2.25 Å). Then, the evolution of particle structure can be expressed as
[86]:
For L Pt ≥ 1 ML,

di, j (t) = di, j (t − 1) (5.39)

d(Di, j (t))
= −vPt Pt (5.40)
dt
and

L i, j (t) = 1/2 · (Di, j (t) − di, j (t)) (5.41)

For L Pt ≤ 1 ML,

L i, j (t) = 1 ML (5.42)
5.3 Pt-Based Alloy Degradation 213

d(Di, j (t))
= −(vPt Pt + vCo Co ) (5.43)
dt
and

di, j (t) = Di, j (t) − 2 × L i, j (t) (5.44)

5.3.3.2 Model Validation

The model validation is conducted by comparing the model predictions with exper-
imental results of Xin et al. [61]. The unknown parameters in the Pt shell oxidation
model are firstly fitted by comprehensive consideration of the oxide onset voltage,
leveling anodic current and cathodic peak voltage. A comparison between the exper-
imental [60] and the simulated cyclic voltammetries (CVs) before and after electro-
chemical aging is shown in Fig. 5.40. All the experimental CV data are postprocessed
by removing a constant double-layer current, and simulations are operated above
0.6 V, at which platinum is oxidized. It is found that the simulated CV curves in
this model fit the experimental ones reasonably well during the anodic sweeps, not
only before but also after degradation. While significant errors in the oxide reduction
peak may come from the complexity of the Pt–O reduction behaviors, the removal
of the oxide from particle surface has both thermodynamic and kinetic properties
[15]. Additionally, higher-order oxide (PtOx , where x ≥ 1) could form at potentials
higher than 1.0 V [87, 88]. The PtOx species needs more time to be removed by the
place-exchange reaction during the cathodic sweep, which then leads to a slower
oxide reduction kinetics and broader oxide reduction peak than the PtO species in

Fig. 5.40 Experimental [60] (solid lines) and simulated (dashed lines) CV data with a scan rate of
50 mV s−1 at room temperature before and after electrochemical aging [86]
214 5 Degradation of Pt-Based Catalysts in PEMFC

the present Pt oxide model. However, we believe that this model works quite well for
the present purposes of predicting the PtO coverage, especially at potentials below
1.0 V.
For understanding the Pt dissolution behavior and fitting the parameters in
Eq. (5.7), equilibrium Pt2+ concentration is used. The data of Pt dissolution for
commercial TKK Pt3 Co catalyst are extracted from a DOE report [19], where the
equilibrium concentration is about 0.09 µM at room temperature and 1.0 V when
the mean diameter is 5.7 nm. According to the report, Pt dissolution bulk equilib-
rium voltage Ueq was measured to be 1.217 V, and the surface tension for Pt shell
is almost 1.1 times the value of pure Pt catalysts [19]. ξ is fitted according to the
assumption that Ueq should be equal to the dissolution bulk equilibrium voltage of
pure Pt catalyst, when the Pt shell thickness ≥6 ML. As the net Pt dissolution rate
ref
is ought to be zero at equilibrium, the value of v2 can be fitted when v1 and cPt 2+ are

assumed firstly. Eventually, the Pt dissolution activation enthalpy H 1 is obtained by

fitting the experimental ECA, PSD and Pt shell thickness under the same conditions
with sweeping rate of 50 mV s−1 from 0.6 to 1.0 V versus RHE [60, 61].
Figure 5.41a shows the evolutions of Pt2+ concentration and Pt oxide coverage in
one voltage cycle between 0.6 and 1.0 V. Upon positive sweep, the Pt2+ concentration
below 0.85 V is negligible as it is far smaller than the maximum one. But the Pt2+
concentration increases rapidly with increasing voltage between 0.85 and 1.0 V and

Fig. 5.41 Model validation by comparison of the simulated and experimental Pt2+ concentration,
ECA, PSD and Pt shell thickness [86]
5.3 Pt-Based Alloy Degradation 215

reaches its maximum (~0.054 µM) at 0.4 s after the sweep is reversed because the
reaction rate is not fast enough to reach equilibrium at each temporal potential [19].
Thus, the dynamic maximum Pt2+ concentration is still lower than the thermodynamic
equilibrium concentration at 1.0 V (~0.09 µM [43]). PtO coverage, using a small
particle (D = 4.08 nm) as an example, is negligible below 0.7 V. It then increases
with increasing voltage from 0.7 to 1.0 V. When the voltage is reversed, PtO coverage
remains stable at its maximum value (~25%) until the voltage is below 0.8 V. PtO
coverage on a large particle (D = 10.08 nm) shows similar behavior during cycling,
but the magnitude is lower than that on a small particle. A comparison of the simulated
ECA and the experimental results during the degradation is shown in Fig. 5.41b. By
adjusting the Pt dissolution activation enthalpy H 1 , a good agreement can be reached.
Figure 5.41c shows a comparison between the simulated and experimental PSDs
near the CCL/GDL interface. The simulated PSD shows a larger mean diameter
than the experimental result. There are two possible reasons for the discrepancy: (1)
the sampling bias for the initial and final measured PSD, and (2) not counting the
particles of carbon support in the simulation. While Fig. 5.41d compares the predicted
and experimental Pt shell thickness at different particle size near the CCL/GDL
interface after 30k (1k = 1000) voltage cycles. For large particles, i.e., D > 7 nm, the
predicted Pt shell thickness agrees with the experimental data reasonably well. Small
difference between the simulated and the experimental data, if any, may originate
from the coalescence between the nanoparticles, which results in particles that contain
multiple Pt–Co cores and thicker Pt shells but this contribution to the ECA is small
[61]. For small particles, i.e., D ≤ 7 nm, the simulated Pt shell thickness shows
large deviations from the experimental data. This is attributed to the oversimplified
assumption that Co dissolution only occurs when the Pt shell is less than 1 ML (from
particle surface). In fact, it is possible that Co leaching could happen when the Pt
shell thickness is 1.0–4.5 ML, as there could be sufficient time for the bulk Co to
segregate to the surface during long-term voltage cycles. Two mechanisms may exist
in Co segregation: ➀ There are three steps for the first mechanism. Firstly, surface
oxide species (OH or O) transfer into subsurface by place exchange of “OPt” and
“PtO” dipoles under high potential, and then the oxide species would tend to bond
with Co atoms by replacement reaction in the core, while under low potential, the
“CoO” species would be dragged to the particle surface with decreasing surface oxide
coverage; ➁ for the second mechanism, it was believed that the binding energy of
adsorbed oxygen onto Co atoms underneath the Pt shell would provide sufficient
driving force for Co segregation. However, the kinetics of Co diffusion from the bulk
to the surface is still unclear. Moreover, it is far slower than Co dissolution on the
surface when Pt shell is thinner than 1ML. Therefore, it is a reasonable simplification
above in the present model.
216 5 Degradation of Pt-Based Catalysts in PEMFC

Fig. 5.42 Evolution of particle structure during the cycling [86]

5.3.3.3 Evolutions of Pt3 Co Particle Structure and Composition

The evolution of particle structure, composed of the Pt shell thickness and particle
diameter, is a distinctly important information for further understanding of the perfor-
mance degradation of the core–shell catalyst. Specifically, three particles near the
CCL/GDL interface are tracked during the voltage cycles, as shown in Fig. 5.42a.
The particle of medium size with initial diameter of 7.6 nm is found to grow first,
then dissolve and finally return to the initial state by mainly changing the Pt shell
thickness. The biggest particle tends to grow with thicker Pt shell during the whole
cycles, while the smallest one dissolves totally in the end. Since larger particles are
shown to have thicker Pt shell in one control volume, the Pt on the shell of smaller
particles is supposed to dissolve easily during the anodic sweep, while Pt ion re-
deposits slowly during the cathodic sweep. Consequently, it is concluded that, in the
control volume near the CCL/GDL interface, particles (>7.6 nm) will finally grow to
be bigger while the small ones (<7.6 nm) will dissolve to be smaller or even disappear
after 30k cycles.
Evolution of diameter of the particles with the initial Pt shell thickness (2.6 ML
= 0.585 nm) in different control volumes is illustrated in Fig. 5.42b. First of all, the
diameter of particles with 2.6 ML Pt shell is found to be smaller near the GDL/CCL
interface during the cycling, because more small particles dissolve closer to the
CCL/membrane interface, as demonstrated in Fig. 5.42b. According to Eq. (5.7) and
Eq. (5.36), lower Pt2+ concentration will give rise to more serious dissolution of
Pt shell for small particles; therefore, the 2.6 ML Pt shell during the cycling will
correspond to a larger particle near the CCL/membrane interface. Moreover, it can
be speculated that if the initial particle diameter is in the region above the line of one
control volume in Fig. 5.42b (e.g., particle A in region 1 for the control volume of x/L
= 0.65), it will grow to be bigger after the corresponding cycling numbers. While
the initial particle diameter lies in the region below the line of one control volume
(e.g., particle B in region 2 for the control volume of x/L = 0.65), this particle will
dissolve to be smaller or even disappear.
5.3 Pt-Based Alloy Degradation 217

Fig. 5.43 Evolutions of the Pt/Co atomic ratio, Co mass remaining and total Co mass loss during
the cycling [86]

Once the structure of each particle is determined during cycling, the composi-
tion of particles in different control volumes can be calculated. Although the Pt/Co
atomic ratio at each position is identical before cycling, the Pt/Co atomic ratio along
the cathode thickness differs very much during the potential cycling, as shown in
Fig. 5.43a. The Pt/Co atomic ratio after certain potential cycles is found to decrease
from the GDL/CCL interface to the CCL/membrane interface, as the Pt/Co atomic
ratio grows faster when closer to the GDL/CCL interface. This trend is in agreement
with the finding of Chen et al. [63] and further confirms the validity of the model.
It is worthwhile to note that the Pt/Co atomic ratio near the membrane decreases
with increasing cycle number. This can be artificial because Co dissolution from
the subsurface layers is neglected in the present model. In Fig. 5.43b, we plot the
percentage of Co mass remaining in different positions and the normalized total
Co mass loss after certain cycles. The Co mass loss is significantly faster near the
CCL/membrane interface, which is due to the thinner Pt shell for protecting the inner
Co atom. So during the cycling, the remaining Co mass near the GDL/CCL interface
is higher than that near the CCL/membrane interface. The model calculates that 72%
Co mass remains near the GDL/CCL interface after 30k cycles, while only 44.5%
near the CCL/membrane interface. Additionally, 40% Co mass is lost in the whole
CCL after cycling. To sum up, the Pt loss and Co loss both increase from GDL/CCL
interface to the CCL/membrane interface, but as the Pt/Co atomic ratio is larger near
the GDL/CCL interface, the loss rate of Co is higher than Pt.

5.3.4 Mitigation Strategies for Pt-Based Alloy Degradation

Before summing up the strategies for mitigating Pt-based alloy degradation, it is

worthwhile to mention that the mitigation strategies proposed for pure Pt catalysts
are also suitable for the Pt-based alloys. But there are some other special mitigation
strategies for Pt-based alloys only, which then would be concluded here: ➀ optimizing
218 5 Degradation of Pt-Based Catalysts in PEMFC

the morphology and structure of the Pt-based catalysts. There are mainly three struc-
tures synthesized for Pt-based alloy application in PEMFC: core–shell, spongy and
skeleton. Comparing with spongy and skeleton structure, core–shell catalysts usually
show a higher stability, as it is able to control the thickness of Pt shell for suppressing
the dissolution of transitional metals; ➁ establishing control over the particle size.
For many M-enriched Pt bimetallic catalysts, nanoporosity usually forms during the
preparation or voltage cycling, which can significantly influence the catalyst stability
by deteriorating the leaching of based metals. Gan et al. [85] and Han et al. [89] have
suggested that the formation process of nanoporosity can be avoided for particle
size smaller than 10 nm. However, the other degradation processes, agglomeration
and Ostwald ripening, will be enhanced for smaller particles. Therefore, to receive
a better durability, we should control the particle size of Pt bimetallic catalysts to a
suitable distribution range.

References

1. Wang C, Cheng X, Lu J, Shen S et al (2017) The experimental measurement of local and bulk
oxygen transport resistances in the catalyst layer of proton exchange membrane fuel cells. J
Phys Chem Lett 8(23):5848–5852
2. Shen S, Cheng X, Wang C, Yan X et al (2017) Exploration of significant influences of the oper-
ating conditions on the local O2 transport in proton exchange membrane fuel cells (PEMFCs).
Phys Chem Chem Phys 19(38):26221–26229
3. Wang GX, Yang L, Wang JZ, Liu HK et al (2005) Enhancement of ionic conductivity of PEO
based polymer electrolyte by the addition of nanosize ceramic powders. J Nanosci Nanotechnol
5(7):1135–1140
4. Darling RM, Meyers JP (2003) Kinetic model of platinum dissolution in PEMFCs. J
Electrochem Soc 150(11):A1523–A1527
5. Zana A, Speder J, Roefzaad M, Altmann L et al (2013) Probing degradation by IL-TEM:
the influence of stress test conditions on the degradation mechanism. J Electrochem Soc
160(6):F608–F615
6. Schneider P, Sadeler C, Scherzer AC et al (2019) Fast and reliable state-of-health model of a
PEM cathode catalyst layer. J Electrochem Soc 166(4):F322–F333
7. Tang L, Han B, Persson K, Friesen C et al (2010) Electrochemical stability of nanometer-scale
Pt particles in acidic environments. J Am Chem Soc 132(2):596–600
8. Matsumoto M, Miyazaki T, Imai H (2011) Oxygen-enhanced dissolution of platinum in acidic
electrochemical environments. J Phys Chem C 115(22):11163–11169
9. Ahluwalia RK, Arisetty S, Wang X, Wang X et al (2013) Thermodynamics and kinetics of plat-
inum dissolution from carbon-supported electrocatalysts in aqueous media under potentiostatic
and potentiodynamic conditions. J Electrochem Soc 160(4):F447–F455
10. Ahluwalia RK, Arisetty S, Peng JK, Subbaraman R et al (2014) Dynamics of particle growth and
electrochemical surface area loss due to platinum dissolution. J Electrochem Soc 161(3):F291–
F304
11. Yasuda K, Taniguchi A, Akita T, Ioroi T, Siroma Z (2006) Platinum dissolution and deposition
in the polymer electrolyte membrane of a PEM fuel cell as studied by potential cycling. Phys
Chem Chem Phys 8(6):746–752
12. Rinaldo SG, Stumper J, Eikerling M (2010) Physical theory of platinum nanoparticle
dissolution in polymer electrolyte fuel cells. J Phys Chem C 114(13):5773–5785
13. Holby EF, Sheng W, Shao-Horn Y, Morgan D (2009) Pt nanoparticle stability in PEM fuel cells:
influence of particle size distribution and crossover hydrogen. Energy Environ Sci 2(8):865–871
References 219

14. Bi W, Fuller TF (2008) Modeling of PEM fuel cell Pt/C catalyst degradation. J Power Sources
178(1):188–196
15. Holby EF, Morgan D (2012) Application of Pt nanoparticle dissolution and oxidation modeling
to understanding degradation in PEM fuel cells. J Electrochem Soc 159(5):B578–B591
16. Zhang J, Litteer BA, Gu W, Liu H, Gasteiger HA (2007) Effect of hydrogen and oxygen
partial pressure on Pt precipitation within the membrane of PEMFCs. J Electrochem Soc
154(10):B1006–B1011
17. Mayrhofer KJ, Meier JC, Ashton SJ, Wiberg GK et al (2008) Fuel cell catalyst degradation on
the nanoscale. Electrochem Commun 10(8):1144–1147
18. Mayrhofer KJ, Ashton SJ, Meier JC, Wiberg GK et al (2008) Non-destructive transmission
electron microscopy study of catalyst degradation under electrochemical treatment. J Power
Sources 185(2):734–739
19. Garzon F, Brosha E, Pivovar B, Rockward T et al (2006) 2006 Annual DOE fuel cell program
review
20. Borup R, Meyers J, Pivovar B, Kim YS et al (2007) Scientific aspects of polymer electrolyte
fuel cell durability and degradation. Chem Rev 107(10):3904–3951
21. Gilbert JA, Kariuki NN, Subbaraman R, Kropf AJ et al (2012) In situ anomalous small-angle
X-ray scattering studies of platinum nanoparticle fuel cell electrocatalyst degradation. J Am
Chem Soc 134(36):14823–14833
22. Lopes PP, Strmcnik D, Tripkovic D, Connell JG et al (2016) Relationships between atomic
level surface structure and stability/activity of platinum surface atoms in aqueous environments.
ACS Catal 6(4):2536–2544
23. Sui S, Wang X, Zhou X, Su Y et al (2017) A comprehensive review of Pt electrocatalysts for
the oxygen reduction reaction: nanostructure, activity, mechanism and carbon support in PEM
fuel cells. J Mater Chem A 5(5):1808–1825
24. Zhang Y, Chen S, Wang Y, Ding W et al (2015) Study of the degradation mechanisms of
carbon-supported platinum fuel cells catalyst via different accelerated stress test. J Power
Sources 273:62–69
25. Jiang F, Zhu F, Yang F, Yan X et al (2020) Comparative investigation on the activity degradation
mechanism of Pt/C and PtCo/C electrocatalysts in PEMFCs during accelerate degradation
process characterized by an in-situ X-ray absorption fine structure. ACS Catal 10(1):604–612
26. Nagamatsu SI, Arai T, Yamamoto M et al (2013) Potential-dependent restructuring and
hysteresis in the structural and electronic transformations of Pt/C, Au (core)-Pt (shell)/C, and
Pd (core)-Pt (shell)/C cathode catalysts in polymer electrolyte fuel cells characterized by in situ
X-ray absorption fine structure. J Phys Chem C 117(25):13094–13107
27. Bi W, Sun Q, Deng Y, Fuller TF (2009) The effect of humidity and oxygen partial pressure on
degradation of Pt/C catalyst in PEM fuel cell. Electrochim Acta 54(6):1826–1833
28. Alsabet M, Grden M, Jerkiewicz G (2006) Comprehensive study of the growth of thin oxide
layers on Pt electrodes under well-defined temperature, potential, and time conditions. J
Electroanal Chem 589(1):120–127
29. Zihrul P, Hartung I, Kirsch S, Huebner G et al (2016) Voltage cycling induced losses in elec-
trochemically active surface area and in H2 /air-performance of PEM fuel cells. J Electrochem
Soc 163(6):F492–F498
30. Harzer GS, Schwämmlein JN, Damjanović AM et al (2018) Cathode loading impact on voltage
cycling induced PEMFC degradation: a voltage loss analysis. J Electrochem Soc 165(6):F3118–
F3131
31. Kneer A, Wagner N, Sadeler C, Scherzer AC et al (2018) Effect of dwell time and scan
rate during voltage cycling on catalyst degradation in PEM fuel cells. J Electrochem Soc
165(10):F805–F812
32. Uchimura M, Kocha SS (2007) The impact of cycle profile on PEMFC durability. ECS Trans
11(1):1215–1226
33. Darling RM, Meyers JP (2005) Mathematical model of platinum movement in PEM fuel cells.
J Electrochem Soc 152(1):A242–A247
220 5 Degradation of Pt-Based Catalysts in PEMFC

34. Redmond EL, Trogadas P, Alamgir FM, Fuller TF (2013) The effect of platinum oxide growth
on platinum stability in PEMFCs. ECS Trans 50(2):1369–1376
35. Yu H, Baricci A, Casalegno A, Guetaz L et al (2017) Strategies to mitigate Pt dissolution in low
Pt loading proton exchange membrane fuel cell: II. A gradient Pt loading design. Electrochim
Acta 247:1169–1179
36. Ferreira PJ, la O’ GJ, Shao-Horn Y et al (2005) Instability of Pt/C electrocatalysts in
proton exchange membrane fuel cells—a mechanistic investigation. J Electrochem Soc
152(11):A2256–A2271
37. Li Y, Moriyama K, Gu W, Arisetty S, Wang CY (2015) A one-dimensional Pt degradation
model for polymer electrolyte fuel cells. J Electrochem Soc 162(8):F834–F842
38. Baricci A, Zago M, Casalegno A (2016) Modelling analysis of heterogeneity of ageing in
high temperature polymer electrolyte fuel cells: insight into the evolution of electrochemical
impedance spectra. Electrochim Acta 222:596–607
39. Zheng Z, Yang F, Lin C et al (2020) Design of gradient cathode catalyst layer (CCL) struc-
ture for mitigating Pt degradation in proton exchange membrane fuel cells (PEMFCs) using
mathematical method. J Power Sources 451:227729
40. Kregar A, Tavčar G, Kravos A, Katrašnik T (2020) Predictive system-level modeling framework
for transient operation and cathode platinum degradation of high temperature proton exchange
membrane fuel cells. Appl Energy 263:114547
41. Von Smoluchowski M (1917) Investigation of a mathematical theory on the coagulation of
colloidal suspensions. Z Physik Chem (Ger) 92:155
42. Abdollahzadeh M, Ribeirinha P, Boaventura M, Mendes A (2018) Three-dimensional modeling
of PEMFC with contaminated anode fuel. Energy 152:939–959
43. Li Y, Wang CY (2017) Modeling of transient platinum degradation in a low Pt-loading PEFC
under current cycling. J Electrochem Soc 164(2):F171–F179
44. Zhang R, Min T, Chen L, Kang Q et al (2019) Pore-scale and multiscale study of effects of
Pt degradation on reactive transport processes in proton exchange membrane fuel cells. Appl
Energy 253:113590
45. Yu Z, Zhang J, Liu Z et al (2012) Comparison between dealloyed PtCo3 and PtCu3 cathode
catalysts for proton exchange membrane fuel cells. J Phys Chem C 116(37):19877–19885
46. Zhang J, Sasaki K, Sutter E, Adzic RR (2007) Stabilization of platinum oxygen-reduction
electrocatalysts using gold clusters. Science 315(5809):220–222
47. Wilkinson DP, St-Pierre J (2003) In-plane gradients in fuel cell structure and conditions for
higher performance. J Power Sources 113(1):101–108
48. Zhang Y, Smirnova A, Verma A, Pitchumani R (2015) Design of a proton exchange membrane
(PEM) fuel cell with variable catalyst loading. J Power Sources 291:46–57
49. Cetinbas FC, Advani SG, Prasad AK (2014) Investigation of a polymer electrolyte membrane
fuel cell catalyst layer with bidirectionally-graded composition. J Power Sources 270:594–602
50. Baricci A, Bonanomi M, Yu H, Guetaz L et al (2018) Modelling analysis of low platinum
polymer fuel cell degradation under voltage cycling: gradient catalyst layers with improved
durability. J Power Sources 405:89–100
51. Yu H, Baricci A, Bisello A, Casalegno A et al (2017) Strategies to mitigate Pt dissolution
in low Pt loading proton exchange membrane fuel cell: I. A gradient Pt particle size design.
Electrochim Acta 247:1155–1168
52. Wang Q, Eikerling M, Song D et al (2004) Functionally graded cathode catalyst layers for
polymer electrolyte fuel cells I. Theoretical modeling. J Electrochem Soc 151(7):A950–A957
53. Xie Z, Navessin T, Shi K, Chow R et al (2005) Functionally graded cathode catalyst layers
for polymer electrolyte fuel cells II. Experimental study of the effect of nafion distribution. J
Electrochem Soc 152(6):A1171–A1179
54. Gummalla M, Ball SC, Condit DA et al (2015) Effect of particle size and operating conditions
on Pt3 Co PEMFC cathode catalyst durability. Catalysts 5(2):926–948
55. Jomori S, Nonoyama N, Yoshida T (2012) Analysis and modeling of PEMFC degradation:
effect on oxygen transport. J Power Sources 215:18–27
References 221

56. Kongkanand A, Mathias MF (2016) The priority and challenge of high-power performance of
low-platinum proton-exchange membrane fuel cells. J Phys Chem Lett 7(7):1127–1137
57. Shen S, Cheng X, Wang C et al (2017) Exploration of significant influences of the operating
conditions on the local O2 transport in proton exchange membrane fuel cells (PEMFCs). Phys
Chem Chem Phys 19(38):26221–26229
58. Weber AZ, Kusoglu A (2014) Unexplained transport resistances for low-loaded fuel-cell
catalyst layers. J Mater Chem A 2:17207–17211
59. Carlton CE, Chen S, Ferreira PJ et al (2012) Sub-nanometer-resolution elemental mapping
of “Pt3 Co” nanoparticle catalyst degradation in proton-exchange membrane fuel cells. J Phys
Chem Lett 3(2):161–166
60. Yu Y, Xin HL, Hovden R, Wang D et al (2012) Three-dimensional tracking and visualiza-
tion of hundreds of Pt−Co fuel cell nanocatalysts during electrochemical aging. Nano Lett
12(9):4417–4423
61. Xin HL, Mundy JA, Liu Z et al (2012) Atomic-resolution spectroscopic imaging of ensembles
of nanocatalyst particles across the life of a fuel cell. Nano Lett 12(1):490–497
62. Rasouli S, Godoy RO, Yang Z et al (2017) Surface area loss mechanisms of Pt3 Co nanocatalysts
in proton exchange membrane fuel cells. J Power Sources 343:571–579
63. Chen S, Gasteiger HA, Hayakawa K et al (2010) Platinum-alloy cathode catalyst degradation
in proton exchange membrane fuel cells: nanometer-scale compositional and morphological
changes. J Electrochem Soc 157(1):A82–A97
64. Dubau L, Maillard F, Chatenet M, Guetaz L et al (2010) Durability of Pt3 Co/C cathodes in a 16
cell PEMFC stack: macro/microstructural changes and degradation mechanisms. J Electrochem
Soc 157(12):B1887–B1895
65. Dai S, Hou Y, Onoue M, Zhang S et al (2017) Revealing surface elemental composition and
dynamic processes involved in facet-dependent oxidation of Pt3 Co nanoparticles via in situ
transmission electron microscopy. Nano Lett 17(8):4683–4688
66. Conway BE, Barnett B, Angerstein-Kozlowska H, Tilak BV (1990) A surface-electrochemical
basis for the direct logarithmic growth law for initial stages of extension of anodic oxide films
formed at noble metals. J Chem Phys 93(11):8361–8373
67. Myers D (2016) FC-PAD—Electrodecatalysts and supports. US Department of Energy
Hydrogen and Fuel Cells Program 2016 Peer Evaluation Report. https://www.hydrogen.ene
rgy.gov/pdfs/review16/fc136_borup_2016_o.pdf
68. Menning CA, Hwu HH, Chen JG (2006) Experimental and theoretical investigation of the
stability of Pt-3d-Pt (111) bimetallic surfaces under oxygen environment. J Phys Chem B
110(31):15471–15477
69. Menning CA, Chen JG (2008) Thermodynamics and kinetics of oxygen-induced segrega-
tion of 3d metals in Pt–3d–Pt (111) and Pt–3d–Pt (100) bimetallic structures. J Chem Phys
128(16):164703
70. Noh SH, Seo MH, Seo JK et al (2013) First principles computational study on the
electrochemical stability of Pt–Co nanocatalysts. Nanoscale 5(18):8625–8633
71. Han BC, Van der Ven A, Ceder G, Hwang BJ (2005) Surface segregation and ordering of alloy
surfaces in the presence of adsorbates. Phys Rev B 72(20):205409
72. Stamenkovic VR, Fowler B, Mun BS, Wang G et al (2007) Improved oxygen reduction activity
on Pt3Ni (111) via increased surface site availability. Science 315(5811):493–497
73. Chen S, Sheng W, Yabuuchi N, Ferreira PJ et al (2009) Origin of oxygen reduction reaction
activity on “Pt3 Co” nanoparticles: atomically resolved chemical compositions and structures.
J Phys Chem C 113(3):1109–1125
74. Ishiguro N, Kityakarn S, Sekizawa O, Uruga T et al (2014) Rate enhancements in struc-
tural transformations of Pt–Co and Pt–Ni bimetallic cathode catalysts in polymer electrolyte
fuel cells studied by in situ time-resolved X-ray absorption fine structure. J Phys Chem C
118(29):15874–15883
75. Huang H, Li K, Chen Z, Luo L et al (2017) Achieving remarkable activity and durability
toward oxygen reduction reaction based on ultrathin Rh-doped Pt nanowires. J Am Chem Soc
139(24):8152–8159
222 5 Degradation of Pt-Based Catalysts in PEMFC

76. Arruda TM, Shyam B, Ziegelbauer JM et al (2008) Investigation into the competitive and site-
specific nature of anion adsorption on Pt using in situ X-ray absorption spectroscopy. J Phys
Chem C 112(46):18087–18097
77. Teliska M, O’Grady WE, Ramaker DE (2005) Determination of O and OH adsorption sites
and coverage in situ on Pt electrodes from Pt L23 X-ray absorption spectroscopy. J Phys Chem
B 109(16):8076–8084
78. Mukerjee S, Srinivasan S, Soriago MP, McBreen J (1995) Role of structural and electronic
properties of Pt and Pt alloys on electrocatalysis of oxygen reduction. J Electrochem Soc
142:1409–1422
79. Kitchin JR, Nørskov JK, Barteau MA, Chen JG (2004) Role of strain and ligand effects in the
modification of the electronic and chemical properties of bimetallic surfaces. Phys Rev Lett
93(15):156801
80. Kim HJ, Choi SM, Nam SH et al (2009) Carbon-supported PtNi catalysts for electrooxidation
of cyclohexane to benzene over polymer electrolyte fuel cells. Catal Today 146(1–2):9–14
81. Xie S, Choi SI, Lu N, Roling LT et al (2014) Atomic layer-by-layer deposition of Pt on Pd
nanocubes for catalysts with enhanced activity and durability toward oxygen reduction. Nano
Lett 14(6):3570–3576
82. Oezaslan M, Hasche F, Strasser P (2013) Pt-based core–shell catalyst architectures for oxygen
fuel cell electrodes. J Phys Chem Lett 4(19):3273–3291
83. Strasser P, Koh S, Anniyev T et al (2010) Lattice-strain control of the activity in dealloyed
core–shell fuel cell catalysts. Nat Chem 2(6):454
84. Oezaslan M, Heggen M, Strasser P (2012) Size-dependent morphology of dealloyed bimetallic
catalysts: linking the nano to the macro scale. J Am Chem Soc 134(1):514–524
85. Gan L, Heggen M, O’Malley R, Theobald B, Strasser P (2013) Understanding and controlling
nanoporosity formation for improving the stability of bimetallic fuel cell catalysts. Nano Lett
13(3):1131–1138
86. Zheng Z, Luo L, Zhu F, Cheng X et al (2019) Degradation of core-shell Pt3 Co catalysts in proton
exchange membrane fuel cells (PEMFCs) studied by mathematical modeling. Electrochim Acta
323:134751
87. Ahluwalia RK, Papadias DD, Kariuki NN, Peng JK et al (2018) Potential dependence of Pt and
Co dissolution from platinum-cobalt alloy PEFC catalysts using time-resolved measurements.
J Electrochem Soc 165(6):F3024–F3035
88. Imai H, Izumi K, Matsumoto M et al (2009) In situ and real-time monitoring of oxide growth
in a few monolayers at surfaces of platinum nanoparticles in aqueous media. J Am Chem Soc
131(17):6293–6300
89. Han B, Carlton CE, Kongkanand A et al (2015) Record activity and stability of dealloyed
bimetallic catalysts for proton exchange membrane fuel cells. Energy Environ Sci 8(1):258–266
Index

C 188, 190, 191, 194, 195, 198, 201,

Cation contamination, 127, 156, 157, 163 203, 204, 207, 209, 210
Composition effects, 36 Oxygen transport, 1, 127, 137, 138, 142, 144,
148, 153, 154, 188
D
Degradation, 7, 49, 64, 109, 167–169, 171–
175, 178, 181, 182, 184–190, 192, P
194–199, 201, 202, 206, 210, 211, Proton Exchange Membrane Fuel Cells
213, 215, 217 (PEMFCs), 1, 57, 64, 127, 167
Proton transport, 5, 7, 9, 18, 127, 128, 157,
L 161, 162, 173, 194, 199
Low platinum-group metal electrocatalysis, Pt-based catalysts, 1, 17, 26–30, 35, 37, 42,
v 43, 167, 168, 217
Low Pt electrocatalysts, 1 Pt monolayer shell, 63

M
Mathematical modeling, 168, 181, 185, 210 S
Membrane Electrode Assembly (MEA), 2– Shape-controlled Pt-based alloys, 19
5, 10, 12, 18–21, 29, 58, 106, 127, Substrate effect, 63
128, 130, 136, 144, 145, 147, 158–
162, 168–170, 185, 195–197

U
O Underpotential Deposition (UPD), 65, 87,
Oxygen Reduction Reaction (ORR), 8, 9, 89, 100, 108
12–19, 25–33, 35–38, 41–43, 48, 50,
53, 54, 57, 58, 63–71, 73, 76–85, 87–
90, 92–97, 99–101, 103–105, 107–
110, 114–116, 118, 119, 128, 156– X
158, 163, 169, 174, 175, 180, 181, XAFS analysis, 175, 201–203

© Shanghai Jiao Tong University Press and Springer-Verlag GmbH Germany 2021 223
J. Zhang and S. Shen, Low Platinum Fuel Cell Technologies,
Energy and Environment Research in China,
https://doi.org/10.1007/978-3-662-56070-9