1 STUDY OF PERFLUORINATED IONOMER MEMBRANE DEGRADATION A ...703/fulltext.pdf · electrostatic...
Transcript of 1 STUDY OF PERFLUORINATED IONOMER MEMBRANE DEGRADATION A ...703/fulltext.pdf · electrostatic...
1
ELECTROCATALYSIS OF OXYGEN REDUCTION IN ALKALINE MEDIA AND A
STUDY OF PERFLUORINATED IONOMER MEMBRANE DEGRADATION
A dissertation presented
by
Nagappan Ramaswamy
to
The Department of Chemistry and Chemical Biology
In partial fulfillment of the requirements for the degree of
Doctor of Philosophy
in the field of
Chemistry
Northeastern University
Boston, Massachusetts
April 2011
2
©Nagappan Ramaswamy 2011
All Rights Reserved
3
ELECTROCATALYSIS OF OXYGEN REDUCTION IN ALKALINE MEDIA AND A
STUDY OF PERFLUORINATED IONOMER MEMBRANE DEGRADATION
by
Nagappan Ramaswamy
ABSTRACT OF DISSERTATION
Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Chemistry
in the Graduate School of Arts and Sciences of Northeastern University, April, 2011
4
Abstract
Oxygen Reduction Reaction (ORR) is an overarching technological and fundamental
challenge in the context of electrochemical energy conversion. Recent developments in alkaline
ionomer membranes that transport hydroxide anions have spurred research interests in Alkaline
Membrane Fuel Cell (AMFC) technology. Among several challenges such as improving OH
conductivity and anodic alcohol oxidation, a key source of overpotential loss is the cathodic
ORR process. In this scenario, the fundamental understanding and design of novel
electrocatalysts for ORR assume a crucial role. While Pt-based materials have been the mainstay
in acidic media, the dispensability of the so called “acid-stability” criterion permits the use of
several non-noble metal-based electrocatalysts for ORR in alkaline media. In this dissertation, a
combination of catalyst synthesis, electrochemical studies and X-ray absorption spectroscopy
(XAS) investigations have been carried out to understand ORR in alkaline media. Such a study is
expected to provide a detailed understanding of oxygen reduction mechanisms, pathways and
ultimately methods to design novel electrocatalysts.
Chapter 1 begins with a broad introduction to the niche position that electrochemical
energy conversion enjoys and the fundamentals of electrocatalysis and XAS techniques. Pt-based
materials, the subject of Chapter 2, act as model systems to understand the
mechanisms/pathways under more or less ideal conditions. Typically, electrocatalytic reactions
are treated as inner-sphere processes whereas the possibilities of a surface-independent outer-
sphere electron transfer component in the overall inner-sphere electrocatalytic process have not
come to the fore of the discussion. Such a scenario is observed during ORR in alkaline media,
where the specifically adsorbed hydroxide anions are found to mediate/promote outer-sphere
electron transfer to solvated molecular oxygen.
5
Chapter 3 discusses ORR on chalcogen-modified ruthenium nanoparticles (Ru/C,
Se/Ru/C, Se/RuMo/C, S/Ru/C, S/RuMo/C) synthesized via aqueous routes. This class of material
was chosen since the ability of chalcogens to suppress oxide formation on transition metals is
likely to promote direct molecular adsorption of O2. Chapter 4 delves into biomimetic materials
such as iron porphyrin. Here, the origin of electrocatalytic activity of heat-treated metal
macrocycles has been studied using square wave voltammetry, X-ray absorption near edge
spectroscopy (XANES) and Delta-Mu (∆µ) techniques. I report 1) an anodic shift in the metal
center redox potential, 2) stabilization of a ferrous-hydroperoxyl adduct due to double-layer
electrostatic interactions, and 3) that atomic vacancy defects on graphitic carbon surfaces play a
key role in improving ORR activity upon heat treatment.
In Chapter 5, the durability of perfluorinated sulfonic acid proton exchange ionomer
membranes is investigated under fuel cell operating conditions using a novel array-electrode
assembly setup. Correlation of membrane degradation with the peroxide yield is obtained. A
Fenton-type mechanism of peroxide radical generation from H2O2 formed due to a two-electron
pathway of ORR is found to be the dominant membrane degrading factor.
Final Chapter 6 presents an evaluation of electrocatalysts in Alkaline Membrane Fuel
Cells (AMFC). The initial results are very promising and warrant further intensive research. The
importance of certain challenges such as electrocatalyst design, specific adsorption of quaternary
ammonium cations, and study of alkaline anode-membrane double layer strucure are pointed out
for future directions.
6
Acknowledgements
Firstly, I would like to share my deep eternal gratitude to my wonderful parents Mrs.
Chittu Ramaswamy and Mr. N. Ramaswamy. You have given me an extremely memorable
childhood and moral education of the highest standard. You have led me by example, put me into
the best schools, given me the best environment to grow up and always made sure I progress in
my life.
I express my deep gratitude to my advisor Prof. Sanjeev Mukerjee for taking me under
his wings as a graduate student in his vibrant research group at Northeastern University and for
having molded a naïve and shy student into a confident and unreserved individual. He made it a
practice to constantly push the limits and intellectually kindle me to think at different
fundamental levels. For having received direct lessons from him on writing proposals, it
wouldn’t be wrong to say that his innate knack of putting together excellent proposals has rubbed
onto me too. His vision, nonchalance, and determination in conducting his research group have
sown and will continue to sow several seeds for the future and I am immensely glad to be one of
those.
I thank my dissertation committee members, Professors Eugene S. Smotkin, David E.
Budil, Geoffrey Davies and Dr. K.M Abraham for their invaluable time and advice. My sincere
thanks to the ever-growing Northeastern University, in particular to the Department of Chemistry
for having given an opportunity to pursue my higher studies. All members of the Chemistry
department, in particular Jean Harris, Richard Pumphrey, Nancy Weston, Dr. Paul Dimilla, Dr.
Kevin Millea and Sheila Magee Beare at the Graduate School have been of valuable help. I
would like to acknowledge the help offered by the International Student and Scholar Institute at
Northeastern University for having made the lives of foreign students less troublesome.
7
Former members of the Mukerjee group have all played notable roles by setting the stage
for future students. Drs. Craig Urian and Madhu Saha for imparting their fuel cell testing skills,
Drs. Vivek S. Murthi and Lajos Gancs for training me on electroanalytical techniques, Drs.
Joseph Ziegelbauer and Thomas Arruda for the training on X-ray Absorption Spectroscopy
during those sleepless nights at Brookhaven National Laboratory. Dr. Lei Zhang, whose
segmented cell assembly I inherited, Drs. Wen Wen and Cormac O’Laoire are gratefully
acknowledged. Dr. Qinggang He with whom I worked closely on several projects, Dr. Badri
Shyam (George Washington University), Brian Hult, and Chris Boggs have made my graduate
study very memorable. Current members of the Mukerjee group are acknowledged: Christopher
J. Allen - for brewing me coffee every afternoon and bringing home-made cookies, Mathew
Trahan, Daniel Abbott, Mike Bates, Kara Strickland, Urszula Latosiewicz, Mehmet Nurullah
Ates, Keeve Gurkin, Jaehee Hwang, Drs. Aditi Halder, Qingying Jia, Iromie Gunasekara and
Braja Ghosh have made the working atmosphere very vibrant and memorable.
Special thanks to Dr. Nazih Hakim without whose help my first two years in this group
would have been more difficult. The subject of Chapter 5 in this dissertation on Nafion®
membrane degradation was conducted in collaboration with Dr. Nazih Hakim. Looking back, I
also had a rare experience of working with Robert J. Allen who has long history starting from the
1960’s in the synthesis of platinum based fuel cell electrocatalysts. Working with Robert was
made possible by funding from an Israeli start-up venture named EnStorage Incorporation during
which time period I also had the experience of interacting with collaborators Prof. Emanuel
Peled, and Dr. Nina Travitsky from Tel-Aviv University, Israel. Interactions with Dr.
Mohammad Enayetullah, who brought the interest in me to design electrochemical cells, were
very informative and scintillating. Discussions with collaborators Prof. Paul Kohl and Dr. Murat
8 Unlu from Georgia Tech University have been very fruitful. Scientists at the National
Synchrotron Light Sources (NSLS), in Brookhaven National Laboratory (BNL), namely Drs.
Kaumudi Pandya, Syed Khalid, Nebojsa Marinkovic, have been very selflessly helpful during
those demanding moments at X11A and X19A beamlines.
Funding during the academic years 2005-2006 and 2006-2007 were provided via
Teaching Assistantship from the Department of Chemistry at Northeastern University. Funding
via Research Assistantship during the academic years 2007 to 2011 were provided by the US
Army Research Office. I would also like to bring to memory my undergraduate institution,
Central Electrochemical Research Institute (CECRI) at Karaikudi in India, where I received the
initial exposure to the science and engineering of electrochemistry. Dr. R. Pattabiraman at
CECRI was particularly helpful in motivating me to pursue higher studies in electrochemistry.
My wonderful younger sister Dr. Annapoorani who raced past me in getting her doctoral
degree is deeply acknowledged for her support. Sincere thanks to my paternal grandfather Mr. C.
RM. Nagappan, and maternal grandfather Mr. VR. Thiayagarajan, both of whom have played
very influential roles in my life and to whom I looked up to as role models. Finally to my
wonderful wife Mathangi for her extreme patience, love and support all of which she would need
more in the years to come.
9 TABLE OF CONTENTS
Abstract 3
Acknowledgements 6
Table of Contents 9
List of Abbreviations and Symbols 14
Chapter 1 Introduction 19
1.1 Global Energy Scenario and the Need for Renewable Energy 19
1.2 Electrochemical Energy Conversion 20
1.2.1 Proton Exchange Membrane (PEM) Fuel Cells 20
1.2.2 Alkaline Anion Exchange Membrane (AAEM) Fuel Cells 22
1.3 Electrocatalysis of Oxygen Reduction Reaction 24
1.3.1 Electrochemistry Fundamentals 26
1.3.2 Electrochemical Double Layer 29
1.3.3 Rotating Ring Disk Electrode Technique 30
1.3.4 Noble-Metal Electrocatalysts 32
1.3.5 Non-Noble Metal Electrocatalysts 34
1.4 X-ray Absorption Spectroscopy 36
1.5 Scope of Dissertation 41
1.6 References 41
Chapter 2 Impact of Double-Layer Structure and Mechanistic Changes
During Electrocatalysis of Oxygen Reduction Reaction in
Alkaline Medium 43
10 2.1 Introduction 43
2.2 Experimental 50
2.2.1 Electrochemical Characterization 50
2.3 Results and Discussions 50
2.3.1 Structural Aspects of the Double-layer 50
2.3.2 ORR on Pt/C Nanoparticles: Acid vs. Alkaline Medium 54
2.4 Conclusions 66
2.5 Acknowledgements 67
2.6 References 68
Chapter 3 Chalcogen (S/Se) Modified Ruthenium Catalysts for ORR in
Alkaline Medium: Electrochemical Kinetics and X-ray
Absorption Spectroscopy Investigations 71
3.1 Introduction 71
3.2 Experimental 75
3.2.1 Catalyst Synthesis 75
3.2.2 Physicochemical Characterizations 76
3.2.3 Electrochemical Characterizations 76
3.2.4 X-ray Absorption Spectroscopic Measurements 77
3.3 Results and Discussions 77
3.3.1 Physicochemical Characterizations 77
3.3.2 Electrochemical Characterizations 80
3.3.2.1 Cyclic Voltammetry 80
3.3.2.2 ORR Measurements 82
11 3.3.3 In situ X-ray Absorption Spectroscopic Measurements 94
3.3.3.1 XANES and EXAFS 94
3.4 Conclusions 103
3.5 Acknowledgements 104
3.6 References 104
Chapter 4 Redox Potential Tuning and Influence of Graphitic Defects
on the ORR Activity of Pyrolyzed Iron Porphyrin
Electrocatalysts 107
4.1 Introduction 107
4.2 Experimental 110
4.2.1 Catalyst Preparation 110
4.2.2 Electrochemical Characterizations 110
4.2.3 X-ray Absorption Spectroscopic Measurements 111
4.3 Results and Discussions 112
4.3.1 Electrochemical Characterization and ORR 112
4.3.2 X-ray Absorption Spectroscopy 120
4.3.2.1 EXAFS 120
4.3.2.2 XANES 123
4.3.2.3 Delta-Mu Studies 126
4.3.4 ORR Reaction Mechanism 131
4.5 Conclusions 133
4.6 Acknowledgements 134
4.7 References 135
12 Chapter 5 Degradation Mechanism Study of Perfluorinated Proton
Exchange Membrane under Fuel Cell Operating Conditions 138
5.1 Introduction 138
5.2 Experimental 142
5.2.1 Physicochemical Characterization 142
5.2.2 Electrochemical Characterization 143
5.2.3 Segmented Cell Design 144
5.2.4 Durability Test Design Criteria 145
5.2.5 Anode Side Durability Test 146
5.2.6 Cathode Side Durability Test 148
5.2.7 MEA Fabrication 149
5.2.8 Post-Mortem Characterization Techniques 150
5.2.9 Fourier Transform Infrared Spectroscopy 150
5.2.10 Conductivity Measurements 151
5.2.11 Ion Exchange Capacity 151
5.3 Results and Discussions 151
5.3.1 Physicochemical Characterization 151
5.3.2 Electrochemical Measurements: Cyclic Voltammetry 154
5.3.3 Oxygen Reduction Reaction and Peroxide Yield 157
5.3.4 Assignment of Nafion® Absorption Bands 163
5.3.5 Effect of Radical Initiated Membrane Degradation 164
5.3.6 Anode-side Durability Test Results 173
5.3.7 Interpretation Based on Mechanically Coupled
13
Vibrational Modes 175
5.4 Conclusions 176
5.5 Acknowledgements 177
5.6 References 178
Chapter 6 Alkaline Anion Exchange Membrane Fuel Cell Studies,
Thesis Summary and Future Directions 181
6.1 Alkaline Membrane Fuel Cell Studies 181
6.1.1 H2/O2 Alkaline Membrane Fuel Cell 182
6.1.2 Direct Ethanol Alkaline Membrane Fuel Cell 184
6.2 Preliminary Investigations of the Alkaline Anode-Membrane
Interface 187
6.2.1 Effect of Specific Adsorption of Quaternary Ammonium
Cations 191
6.3 Concluding Remarks 193
6.4 References 194
Curriculum Vita 195
14
List of Abbreviations and Symbols
α Charge Transfer Coefficient
η Overpotential/Ion Exchange Capacity
χ Peroxide Yield
ω Rotation Rate
υ Kinematic Viscosity of the Electrolyte
φM Inner-potential of the Metal Electrode
φ1 Potential at the Inner-Helmholtz Plane
φ2 Potential at the Outer-Helmholtz Plane
σi Excess Charge Density in Inner-Helmholtz Plane
σd Excess Charge Density in the Diffuse Layer
θ Angle of Incidence/Coverage
λ Wavelength of X-ray beam
µ Absorption Coefficient
µ(E) Absorption Coefficient at energy E
µo(E) Absorption Coefficient at energy Eo
µm Micrometer (10-6 meters)
σ2 Debye-Waller Factor
σ Ionic Conductivity
Ω Ohms
∆µ Delta-Mu
γ Electrosorption Valency
∆µt Theoretical Delta-Mu
∆H° Standard Enthalpy Change
∆G° Standard Gibbs Free Energy
∆G Gibbs Free Energy
∆S° Standard Entropy Change
Å Angstroms (10-10 meters)
atm Atmospheric Pressure
15
A Atomic Mass
B Levich Constant
C Coulombs
CO2 Concentration of Molecular Oxygen
d Distance between Diffraction Planes
D Diffusion Coefficient
e Electron
eV Electron Volts
Eo Binding Energy
E° Standard Reduction Potential
E°’ Formal Potential
E Electrode Potential
ECell Cell Voltage/Potential
ECathode Cathode Potential
EAnode Anode Potential
EP Peak Potential
Epa Anodic Peak Potential
Epc Cathodic Peak Potential
ERing Ring Electrode Potential
EDisk Disk Electrode Potential
F Faraday’s Constant
ħ Planck’s Constant
i Current Density
ik Kinetic Current Density
ilim Limiting Current Density
io Exchange Current Density
I Current
IR Ring Current
ID Disk Current
Io Intensity of Incident X-ray Beam
16
It Intensity of Transmitted X-ray Beam
Ir Intensity of Reference X-ray Beam
me Mass of an Electron
n Number of Electrons Transferred
nm Nanometer (10-9 meters)
N Collection Efficiency/Coordination Number
Ρ Element Density
qM Charge on the Metal Electrode
R Bond Length/Universal Gas Constant
t Sample Thickness
T Temperature
Tg Glass Transition Temperature
V Volts
Z Atomic Number
ads adsorbate
AAEM Alkaline Anion Exchange Membrane
AFC Alkaline Fuel Cell
AMFC Alkaline Membrane Fuel Cell
ATR Attenuated Total Reflectance
BNL Brookhaven National Laboratory
BPC Black Pearl Carbon
CE Counter Electrode
CV Cyclic Voltammetry
DFT Density Functional Theory
DHE Dynamic Hydrogen Electrode
EDAX/EDS Energy Dispersive Analysis of X-rays
EtOH Ethanol
EXAFS Extended X-ray Absorption Fine Structure
FCC Face Centered Cubic
17
FeTPPCl Iron (III) meso-tetraphenylporphyrin chloride
FeTPP Iron tetraphenylporphyrin
FePc Iron Phthalocyanine
FER Fluoride Emission Rate
FTIR Fourier-Transform Infrared Spectroscopy
FRA Frequency Response Analyzer
GC Glassy Carbon
GDE Gas Diffusion Electrode
GDL Gas Diffusion Layer
GOS Guoy-Chapman-Stern
HCP Hexagonal Close Packed
Hg/HgO Mercury/Mercury(II)Oxide Reference Electrode
HT Heat Treated
HRR Hydrogen Peroxide Reduction Reaction
HRSEM High Resolution Scanning Electron Microscope
IEC Ion Exchange Capacity
IHP Inner-Helmholtz Plane
IR Infrared Spectroscopy
LCA Linear Combination Analysis
LEED Low Energy Electron Diffraction
MEA Membrane Electrode Assembly
NSLS National Synchrotron Light Source
OCP Open Circuit Potential
OCV Open Circuit Voltage
OHP Outer-Helmholtz Plane
ORR Oxygen Reduction Reaction
PAFC Phosphoric Acid Fuel Cells
PGM Platinum Group Metal
PZC Potential of Zero Charge
PZTC Potential of Zero Total Charge
18
PEM Proton Exchange Membrane
PEMFC Proton Exchange Membrane Fuel Cell
PIPS Passivated Implanted Planar Silicon
RH Relative Humidity
RHE Reversible Hydrogen Electrode
RPM Revolutions Per Minute
RDE Rotating Disk Electrode
RDS Rate Determining Step
RE Reference Electrode
RRDE Rotating Ring-Disk Electrode
SEM Scanning Electron Microscope
SHE Standard Hydrogen Electrode
SWV Square Wave Voltammetry
UPD Underpotential Deposition
WE Working Electrode
XANES X-ray Absorption Near Edge Spectroscopy
XAS X-ray Absorption Spectroscopy
XRD X-ray Diffraction
19
Chapter 1
Introduction
1. Introduction
1.1. Global Energy Scenario and the Need for Renewable Energy
One of the grand challenges facing the scientific community at present is to ensure energy
security for the exponentially growing human population. The supply and demand of fossil-fuel
based energy resources are so lop-sided that several natural resources are predicted to be
exhausted within a single human lifespan.1 Since the beginning of the industrial revolution, the
use of fossil fuels has exacerbated this scenario by increasing the atmospheric carbon dioxide
content, which is considered to be a root cause of global warming.2 Taking a historical
perspective, it has been recently shown that atmospheric CO2 emissions have been less than 300
ppm for the past four hundred thousand years or so.3-4 Human activity, primarily due to burning
of fossil fuels in the last hundred years, has pushed atmospheric CO2 levels to about 380 ppm.3-4
While the global warming theory has its opponents, the exhaustible nature of fossil fuels such as
petroleum products, natural gas, coal etc., is irrefutable.1,5 This state of affairs is aggravated by
the unhindered growth of human population all over the world, and the rapidly growing Asian
economies. Fortunately, there is a growing awareness regarding the necessity of clean, safe, and
secure energy sources as recently pointed out by the U.S. Department of Energy (DOE) report.6
Given this circumstance, there has been increasing research and development activity in various
renewable energy sources, in particular electrochemical energy storage and conversion devices
such as fuel cells and batteries.7 A brief introduction of their fundamental principles and their
modus operandi are given in this chapter.
20 1.2 Electrochemical Energy Conversion
Electrochemistry is a special discipline that deals with the interplay of electrical energy and
chemical energy. Fuel cells are electrochemical energy conversion devices that convert chemical
energy stored in fuels directly into electrical energy. Fuel cells can be classified on various bases,
although the temperature of operation is most often used. Proton Exchange Membrane (PEM)
and Alkaline Anion Exchange Membrane (AAEM) fuel cells typically operate below 100°C,
whereas Phosphoric Acid Fuel Cells (PAFC) operate at medium temperatures of 150-220°C.
High temperature operation is typically carried out in Molten Carbonate and Solid Oxide fuel
cells, where temperatures of about 600-1000°C are usually encountered. The subject of
discussion here is limited to low temperature devices such as PEM and AAEM fuel cells.
1.2.1 Proton Exchange Membrane (PEM) Fuel Cells
A schematic design of a PEM fuel cell is shown in Figure 1.1(a). Two electrodes, an
anode and a cathode, are separated by a solid ionomer membrane that transports protons. The
electrodes are electronically conducting whereas the membrane is an ionic conductor that
physically separates the two electrodes. Typical fuels such as compressed gaseous hydrogen are
fed to the anode and an oxidant such as atmospheric oxygen is fed to the cathode. At the anode,
hydrogen undergoes oxidation to yield protons and electrons. Protons are then transported across
the membrane to the cathode under the influence of the so-called electrochemical potential
gradient. Electrons are transported across the external circuit to the cathode. Protons and
electrons recombine at the cathode and involve in oxygen reduction reaction to water. The
anodic oxidation and the cathodic reduction reactions are as shown below:
Anode: H2 → 2H+ + 2e E° = 0.00 V vs. SHE (1.1)
Cathode: 1/2O2 + 2H+ + 2e → H2O E° = 1.23 V vs. SHE (1.2)
21
SHE represents the Standard Hydrogen Scale of electrode potentials. The involvement of protons
in the above reactions causes PEM fuel cells to operate under acidic conditions. While the above
reactions exemplify the half-cell reactions, the overall cell reaction (ECell = ECathode – EAnode) is
the spontaneous conversion of hydrogen and oxygen to water, heat and electricity.
Overall: H2 + 1/2O2 → H2O ECell = 1.23 V (1.3)
Having written these reactions, it is interesting to note the origin of electrical energy
which is thermodynamically considered as net work done (work other than PV work). The
enthalpy associated with reaction (1.3) is ∆H° = -286 kJ mol-1 at 25°C and atmospheric pressure.
Consider a chemical reaction wherein H2 and O2 are directly mixed in a calorimeter at 25°C and
1 atm. The amount of heat liberated will be equal to 286 kJ mol-1. This basically implies that all
energy is dissipated as heat in a spontaneous chemical combustion
Figure 1.1: Schematic illustrations of (a) PEM and (b) AAEM fuel cells
22
reaction. The standard Gibbs free energy change is ∆G° = -237 kJ mol-1. Given ∆G° = ∆H° -
T∆S°, it can be calculated that T∆S° = -49 kJ mol-1. Now, consider a case where H2 and O2 are
physically separated and subjected to discharge in a fuel cell configuration as shown in Figure
1.1(a) where the external circuit constitutes a load resistance. Here the energy corresponding to
the term T∆S° = -49 kJ mol-1 is inevitably liberated as thermal energy, whereas ∆G° = -237 kJ
mol-1 is theoretically converted into electrical energy. So this ∆G term is identified as the
maximum useful work done.8
The combination of anode, cathode and the ionomer membrane is called the Membrane
Electrode Assembly (MEA). State-of-the-art anode and cathode catalysts for PEM fuel cell
consist of platinum or platinum alloy nanoparticles supported on high surface area carbon
supports. Ionomer membranes are solid polymers that are permeable to either anions or cations.
In PEM fuel cells, sulfonated tetrafluroethylene based ionomer membranes (Nafion®) are used
as a proton conductor. The structure of Nafion® is shown in Figure 1.2(a).
1.2.2 Alkaline Anion Exchange Membrane (AAEM) Fuel Cells
AAEM fuel cells, schematically shown in Figure 1.1(b) are alkaline counterparts of the
PEM fuel cell systems. The primary difference is the chemical composition and function of the
Figure 1.2: Chemical structures of (a) perfluorinated sulfonic acid proton exchange membrane (Nafion®) and (b) polysulfone based anion exchange membranes with quaternary ammonium anion exchange groups.
23 alkaline ionomer membrane used between the electrodes. In AAEM fuel cells, ionomer
membranes that transport hydroxide anions from cathode to anode are used. Figure 1.2(b) shows
the chemical structure of a polysulfone based anion exchange membrane with quaternary
ammonium anion exchange groups. Involvement of hydroxide anions causes the cell to operate
under alkaline conditions. The anodic oxidation of hydrogen and cathodic reduction of oxygen
take place under alkaline conditions as shown below:
Anode: H2 + 2OH → 2H2O + 2e E° = -0.828 V vs. SHE (1.4)
Cathode: 1/2O2 + H2O + 2e → 2OH E° = 0.401 V vs. SHE (1.5)
The overall cell reaction in alkaline fuel cell is the conversion of hydrogen and oxygen to heat,
water and electricity.
H2 + 1/2O2 → H2O ECell = 1.23 V (1.6)
AAEM fuel cells have the following intrinsic potential advantages as compared to their acidic
counterparts.9-10
1) Electrocatalysis of anodic methanol oxidation and cathodic oxygen reduction in alkaline
media is facile as compared to acidic media.
2) Alkaline media open the possibility of using inexpensive non-noble electrocatalysts based
on (i) supported noble/non-noble metal clusters, and (ii) metal-organic and inorganic
complexes.
3) Transport of OH ions and subsequent electro-osmotic drag of H2O opposes the crossover
of liquid fuel such as methanol leading to intrinsically lower crossover problems (lower
cathodic overpotential loss).
4) Better water management, since it is consumed at the cathode as a stoichiometric
component and is produced at the anode, leading to decrease in cathode flooding.
24
5) At high pH, attack of the ionomer membrane by peroxide radicals is suppressed, allowing
the use of hydrocarbon backbone membranes and avoiding expensive fluorinated
polymers.
6) Use of metal-cation free hydroxyl ion exchange membranes implies that the carbonate
precipitation found in liquid electrolyte AFC is overcome.
7) AAEM fuel cells allow wider choices of fuel feed such as ethanol, xylitol, oxalic acid,
formic acid, gasoline, ammonia etc., apart from hydrogen and methanol used in acidic
environments due to enhanced C-C bond cleavage possibilities in alkaline media.
1.3 Electrocatalysis of Oxygen Reduction Reaction:
Hydrogen oxidation reaction under acidic and alkaline conditions is thermodynamically
reversible and kinetically less demanding, implying minimal overpotential losses at the anode
during fuel cell operation. On the other hand, Oxygen Reduction Reaction (ORR) is irreversible
and exhibits significant kinetic barriers. An illustrative overall scheme for ORR is given in
Figure 1.3 in acidic and alkaline media. ORR, a multi electron transfer process, involves various
reaction intermediates. For example, in alkaline media, ORR could proceed to the 4e¯ product
either via the direct (k1) or series (k2+k3) path. These two pathways represent efficient 4e¯
reduction of O2 to OH . However, an inefficient 2e¯ pathway (k5) leads to the intermediate
peroxide anion as the stable product. The analogous case of ORR in acidic media is also shown
in Figure 1.3. O2/H2O represents the redox couple in acidic media whose standard reduction
potential is 1.23 V vs. SHE. Due to the involvement of 4H+ and 4e in the ORR process, there is
a 59 mV per pH unit decrease in potential upon changing the pH from zero to fourteen. This
causes the standard reduction potential of the O2/OH¯ redox couple in an alkaline electrolyte to
25
be 0.401 V vs. SHE. Similar changes are observed in the H+/H2, H2O/H2 redox couples. The
reason to point this out is that the potentials of redox couple such as O2/O2¯ that do not involve
protons does not show any change in potentials. The bulk of this dissertation is concerned with
electrochemical ORR reaction mechanisms/pathways on electrocatalyst materials. A very brief
introduction to the fundamentals of electrochemistry is given here, followed by a literature
survey of ORR on noble and non-noble metal electrocatalyst materials. A more comprehensive
treatises could be found in standard text books.8,11-12
Figure 1.3: (Left) Potential scale showing the H+/H2, H2O/H2, O2/H2O, O2/OH¯ redox couples on a Standard Hydrogen Electrode scale. (Right) Summary of Oxygen Reduction Reaction in acidic and alkaline medium.
26 1.3.1 Electrochemistry Fundamentals
Electrochemical processes are heterogeneous electron transfer reactions that occur
between an electronically conducting electrode and an electroactive species dissolved in an
ionically conducting electrolyte. The primary thermodynamic equation that relates the standard
free energy (∆G°) to the standard reduction potential (E°) is given by
∆G° = -nFE° (1.7)
Consider a redox couple (O/R) characterized by the following electrochemical reaction
O + ne → R (1.8)
where n is the number of electrons transferred and, O and R are the oxidized and reduced
species, respectively. The equilibrium condition is given by the Nernst equation that relates the
electrode potential (E) to the bulk concentration of the redox couple (O/R) as shown below,
E = E°’ + [(RT/nF) ln(CO*/CR
*)] (1.9)
where R is the universal gas constant, T is the temperature, F is Faraday’s constant (96485 C
mol-1), and (CO*/CR
*) represents the ratio of the bulk concentrations of O and R. The parameter
E°’ is called the formal potential which is an expedient definition that overcomes the use of
inconvenient parameters such as activity coefficients. E°’ is related to the standard reduction
potential as follows:
E°’ = E° + [(RT/nF) ln([O]/[R])] (1.10)
where [O]/[R] represents the ratio of activities of O and R.
For the electrochemical reaction shown in equation (1.8), a fundamental current-overpotential
relation known as the Butler-Volmer equation under absence of mass-transport limitations is:
i = i 0 [exp(-αfη) – exp((1-α)fη)] (1.11)
27 where ‘i’ is the current density, ‘i0’ is the exchange current density, ‘α’ is the charge transfer
coefficient, ‘η’ is the overpotential, and the constant f = F/RT.
For a one-electron transfer reaction, at high overpotential regions (i.e. η >> ±50 mV),
equation (1.11) breaks down to the following equation known as the Tafel equation:
η = [(RT/αF) ln(i0) – (RT/nF) ln(i)] (1.12)
A plot of η vs. ln(i) yields information on the electron transfer kinetics, such as the charge
transfer coefficient, equilibrium exchange current density, and the Tafel slope (2.3RT/αF).
Cyclic Voltammetry (CV) is a technique that is routinely used in characterizing and
activating electrocatalysts materials in electrolytes. CV offers qualitative study of the cleanliness
of the electrode surface, and quantitative study of the charge transfer involved in electrochemical
reactions. Figure 1.4 shows an illustrative example using a platinum catalyst supported on high
surface area carbon black in 0.1M HClO4 electrolyte. Saturating the electrolyte with an inert gas
like argon permits the study of platinum surface characteristics. At potentials below 0.45 V,
hydrogen adsorption/desorption takes place wherein the protons from the electrolyte deposit on
the Pt surface up to a monolayer level. In the potential region between 0.45 V and 0.75 V, no
charge transfer across the electrode/electrolyte interface takes place. In this region, the current is
capacitive in nature and is characterized by double layer charging. At potentials above 0.75 V,
water in the electrolyte undergoes oxidation on Pt sites and forms Pt-OH species. Beginning with
the pioneering work by Conway13 and others, it is now well established that oxide formation on
Pt in aqueous electrolytes progress according to following steps (a) through (f) with increasing
anodic potentials; (a) initial adsorption of OH¯(Pt-OH(γ-1) with γ=0), (b) charge transfer to form
electro-neutral OHads (Pt-OH(γ-1) with γ=1), (c) deprotonation to form Pt-Oads, (d) place exchange
to O-Pt to form a quasi-2D lattice, and (f) finally the progressive formation of 3D-bulk oxide
28
phase.13-15 Interesting changes take place upon purging the electrolyte with O2 or H2 gases.
Saturating the electrolyte with O2 leads to the oxygen reduction reaction taking place on the Pt
catalyst. The onset potential for oxygen reduction is typically ~ 1 V on a Pt catalyst. At
potentials above 1 V due to Pt-OH formation, oxygen reduction reaction is hindered on Pt
catalysts. This is due to the non-availability of Pt sites required to reduce O2. At potentials above
1 V, CV of Pt under argon and oxygen saturated conditions merged together giving prima facie
evidence that oxide formation and oxygen reduction on Pt share similar crystallographic sites.
Upon purging the electrolyte with H2, it is observed that the CV is shifted positively over the
Figure 1.4: Cyclic voltammetry of Pt/C catalyst in 0.1M HClO4 purged with different gases. Scan rate: 20 mV/s.
29 entire potential region by a certain constant which is characterized by the limiting current for H2
oxidation. This indicates that H2 oxidation is independent of under-potentially deposited (UPD)
hydrogen and oxide formation on Pt.
1.3.2 Electrochemical Double Layer
Understanding the structure of the double-layer was an aspect of extensive research in the
electrochemistry society in the first half of the twentieth century.8,11 Immersion of an electrode
into an electrolyte gives rise to the formation of the so-called electrochemical double-layer. This
interfacial property is due to the creation of excess charges or alternatively due to the difference
in electrochemical potentials of the two phases. Metallic electrodes being excellent conductors
with high concentrations of mobile electrons restrict the excess charge formation to the surface
of thickness about 1Å. On the electrolyte side, a more extensive response is observed due to the
low conductivity and concentration of ionic charge carriers. Ionic species in the electrolyte
respond both structurally and electrically. In an aqueous electrolyte, water molecules orient
themselves on the surface based on their dipole interaction with the excess charge on the metal
surface. Anionic species such as halides, chlorides etc tend to directly chemisorb on metal
surfaces such as Pt. These chemisorbed species are said to be specifically adsorbed due to an
interplay between the free energy of adsorption and free energy of solvation of the anions. On
the other hand, cations such as sodium and potassium tend to be typically well solvated and do
not adsorb on the electrode surface. Exceptions do exist. On Pt surfaces, anions such as
perchlorates do not adsorb, and cations such as quaternary ammonium ions adsorb strongly.
Figure 1.5(a) shows a schematic of the electrochemical double layer. The loci of species that
chemisorb on the electrode surface constitute the Inner-Helmholtz Plane (IHP). These species
populating the IHP covalently interact with the metal surface. On the other hand, the Outer-
30
Helmholtz Plane (OHP) consists of solvated species interacting with the electrode via long range
electrostatic forces. Figure 1.5(b) shows the potential drop across the electrode-electrolyte
interface with and without specific adsorption of anions. The Potential of zero charge (PZC) is an
electrode and electrolyte specific parameter that characterizes a region where there are no excess
charges present on the electrode surface.
1.3.3: Rotating Ring-Disk Electrode Technique:
The Rotating Ring Disk Electrode (RRDE) technique has been extensively used to
understand the kinetics of ORR.8 A schematic of the electrode is shown in Figure 1.6. This
technique involves the convective transport of dissolved molecular oxygen from the bulk to the
electrode surface prior to diffusive transport within the diffusion layer to the catalyst site. This
Figure 1.5: Schematic of (a) electrochemical double layer and (b) potential drop across the electrode/electrolyte interface. Redrawn with permission from reference 8.
31
technique allows the detection of stable reaction intermediates generated at the disk during ORR
by potential control of the ring electrode. RRDE typically consists of a glassy carbon disk
electrode around which a gold ring electrode is concentrically placed. Oxygen reduction reaction
(ORR) is carried out at the catalyst deposited on the disk. The electrode is immersed into the O2
saturated electrolyte and physically rotated to draw molecular O2 to the electrode surface.
Oxygen reduction takes place at the disk and any stable intermediate formed during ORR is
detected at the ring electrode which is potentiostatically controlled at a suitable potential. Current
regions obtained from RRDE experiments can be divided into two regions, namely the kinetic
region and diffusion limited region. In the kinetically controlled region, the ORR process is
limited by the reaction kinetics such as the activation energy. In the diffusion limited region, the
ORR process is limited by the mass transport of dissolved O2 to the electrode surface. The total
Figure 1.6: Schematic of Rotating Ring-Disk Electrode
32 current density (i) can then be written as the sum of the reciprocals of the kinetic current (ik) and
diffusion limited current (ilim) density values shown below:
(1/i) = (1/ik) + (1/ilim) (1.13) Kinetic current density can be extracted from the total current density using the following
equation:
ik = (ilim * i)/(i lim-i) (1.14) Limiting current is characterized by the Levich equation as a function of rotation rate as shown
below:
i lim = Bω1/2 (1.15)
The Levich constant ‘B’ is typically used to extract the number of electrons (n) transferred
during ORR based on the equation below:
B = 0.62nFD2/3υ
-1/6CO2 (1.16)
where ‘υ’ is the kinematic viscosity of the electrolyte, D and CO2 are the diffusion coefficient and
solubility of O2. The ring current measured is due to the stable reaction intermediates that are
generated during ORR at the disk electrode. Only a certain fraction of the intermediate generated
at the disk is detected at the ring due to geometric limitations such as the disk-ring design and the
spacing between them. This necessitates the use of a parameter called the Collection Efficiency
(N), which is typically less than 40%. Using the ring current and the collection efficiency values,
the mole fraction (χ) of peroxide intermediate generated can be quantified as shown below:
χ = [(2*IR/N)/(ID+(IR/N))] (1.17)
1.3.4 Noble-Metal Electrocatalysts
Under electrochemical conditions, most non-noble metal electrodes undergo passivation,
implying that the surface is covered with a thin layer of oxide film. This oxide film prevents the
strong chemisorption of molecular O2, which is a prerequisite for efficient ORR process. This
33
has led to the use of Pt-based electrocatalysts materials for fundamental studies as well as fuel
cell applications. The dissociation energy of O2 double bond is 495 kJ mol-1 and the valence
electronic configuration of molecular O2 in its ground state is
(1σg)2(1σu)
2(2σg)2(2σu)
2(3σg)2(1πu)
4(1πg)2. The O2 reduction reaction causes electron transfer to
Pt + O2 + e-
+ H2O + OH-
+ e-
H2O
+ 2OH-
+ e- Pt + OH-
+ e-
Pt
OO
Pt
OO
Pt
O
O
H
Pt
O
O
H
Pt
O
O
H
Pt
O
O
H
Pt
O
H
Pt
O
H
Figure 1.7: (Top) Proposed adsorption modes of O2 on metal sites, and (bottom) Plausible ORR mechanism assuming end-on adsorption of O2 on a single Pt site.
34 the antibonding orbitals of O2 and thus a progressive weakening of the O=O bond. Figure 1.7
shows the various proposed adsorption modes of O2 on metal sites. These interactions primarily
involve the σ or π orbitals of molecular oxygen with the d-orbitals of the metal. ORR on Pt
surfaces is typically considered to be an efficient 4e¯ process with minimal quantities of peroxide
intermediate being formed. This peroxide intermediate is primarily due to electrolyte and/or
surface impurities present during ORR. Figure 1.7 shows a plausible mechanism involved in
ORR on Pt sites assuming end-on adsorption of O2 (Pauling model) on Pt. This 4e¯ reduction
process shows that ORR proceeds via the formation of superoxo, hydroperoxo, peroxide and
hydroxyl adsorbed intermediates finally leading to the hydroxide anion as the product.
Fundamental to the electrocatalysts theory is the Sabatier principle that dictates that the adsorbed
intermediates should neither adsorb too strongly nor too weakly. To be an efficient
electrocatalyst, all the reaction intermediates should remain adsorbed with an optimum binding
energy until the final stable product is formed that desorbs into the electrolyte. More details on
the electrochemical kinetics of ORR on Pt are given in chapter 2.
1.3.5 Non-noble Metal Electrocatalysts
Cost, scarcity and scientific curiosity are factors that push researchers to search for non-
precious group metal (non-PGM) based electrocatalysts for ORR. Currently available non-PGM
electrocatalysts can be broadly classified into pyrolyzed metal macrocycles with metal-nitrogen
reaction centers, first row transition metal based chalcogenides, and electron conducting polymer
based structures. Figure 1.8 provides structures of the broadly investigates non-PGM fuel cell
electrocatalysts. The so-called “acid stability” criterion has thus far necessitated the transition
metal reactive sites to be safely ensconced within a protective shell provided by ligands,
chalcogens etc., to prevent them from oxidative dissolution. Non-PGM transition metal based
35
monomeric phthalocyanine, and porphyrin systems are widely investigated, inexpensive
electrocatalysts for ORR. The metal atom rather than any other part of the macrocycle is the
active site whereas chelation predominantly serves to preserve the metal atom in a stable form
for ORR. Prior to heat treatment, these materials conduct ORR via a redox electrocatalytic
mechanism which requires the active metal sites to be previously reduced in order to later reduce
the adsorbed O2 molecule. While most studies in the literature involve in understanding the
active site structure obtained after heat treatment of metal macrocycles, no clear experimental
proof exists that elucidates the reaction mechanism/pathway of ORR on heat-treated metal
macrocycles. Although very attractive, these materials conduct ORR with significant two
electron transfer, leading to undesired peroxide generation. Facile kinetics on monomeric
macrocycles is highly improbable given the demands of spin and solvent shell reorganization
along with 4e transfer. This leads to the use of binucleating ligands capable of holding two or
more metal center within a certain distance (~4Å) and geometry, such that each metal center
shares the onus of O2 binding and electron transfer.
Figure 1.8: Structures of various non-PGM catalysts used for ORR.
36
Recent quantum mechanical computations have shown that Co-based chalcogenide
systems with pentlandite (Co9S8 and Co9Se8) have activity for ORR in acidic media and more
interestingly that the sulfide has a lower overpotential of about ~250 mV compared to its Se
brethren. A thin film method was established to investigate such chalcogen materials based on
Co, Cr, and Fe sulfides and selenides. Although the inability to determine the electrochemically
active surface area of these catalysts obscures its direct comparison to Pt, they have significantly
low open circuit potential and severe kinetic and stability limitations.
Recently, heteroatomic conducting polymers such as polyaniline, polypyrrole etc., have
been used as templates for supporting transition metals and also provide metal-nitrogen reaction
sites for adsorption and reduction of O2. These classes of materials provide the advantage of
synthesizing these materials without the pyrolysis step and yet prepare ORR active non-PGM
catalysts. Electronic conductivity of these materials arises from the delocalization of the pi-
electrons along the chain. These materials again provide the base metal with a viable
environment in the high potential regime. However, it is too early to review their long term
stability, peroxide yield and substantiate them as alternatives for Pt group materials.
1.4 X-ray Absorption Spectroscopy:
X-ray absorption spectroscopy (XAS) is an element specific, atomic level probe
involving the excitation of tightly bound core level electrons, near and above the binding energy,
by incident x-ray photons from a high intensity, energy tunable x-ray source such as the
synchrotron.16-17 This technique has been applied widely to study heterogeneous catalyst
materials under ex situ conditions, and it is now well known that the true active state of the fuel
cell catalyst exist only during potential control under electrochemical conditions. In recent years,
37
XAS has attracted great attention due to the ability to study fuel cell electrocatalysts and battery
electrode materials under in situ electrochemical conditions.18 Development of modern
synchrotron light sources has made high intensity of x-ray photon fluxes possible. For instance,
X19A beamline (NSLS) supports a flux of 1011 photons second-1 at 10 keV incident energy on a
sample spot size of 2 mm2. Figure 1.9 shows a sketch of experimental setup for XAS
measurements. Light source from a synchrotron consists of light with various wavelengths across
the spectrum and for this reason it is typically called the ‘white-light’. This light source passes
through a monochromator (Si(111) single crystal) which operates based on Bragg’s law (2dsinθ
= nλ) and permits the passage of only a narrow band of wavelength at X-ray energies. By
continuously changing the angle of incidence (θ) of the white-light on the monochromator, an
energy tunable source of high intensity X-rays is thus achieved. Three gas ionization detectors
for the incident (Io), transmission (It), and reference (Ir) chambers are used. Electrode material is
placed between Io and It, whereas a reference foil made of the same element under study is
placed between It and Ir. Figure 1.10 shows a typical XAS spectrum taken at Pt L3 edge whose
Synchrotron
Monochromator
Io
Spectro-electrochemicalCell
Fluorescence detector
ItIr
Referencefoil
Figure 1.9: Sketch of synchrotron based XAS experiments.
38
binding energy is 11564 eV. Absorption event is governed by the Beer-Lambert’s law as given
below:
It = Io exp(-µt)
where ‘Io’ and ‘It’ are the intensities of incident and transmitted x-ray beams, ‘µ’ is the
absorption coefficient, and ‘t’ is the sample thickness. The element specificity of XAS technique
arises due to the fact that the absorption coefficient exhibits strong dependence on the atomic
number (Z) and the x-ray energy (E) as follows:
µ ≈ ρZ4/AE3
where ‘ρ’ is the element density, ‘A’ is its atomic mass. Dependence of adsorption coefficient on
the fourth power of atomic number Z gives rise to the element specificity of XAS technique. As
shown in Figure 1.10, XAS spectra consist of two parts, (i) the X-ray Absorption Near Edge
Spectra (XANES) and (ii) the Extended X-ray Absorption Fine Structure (EXAFS). XAS
Figure 1.10: Typical normalized XAS spectra obtained at Pt L3 edge (11564 eV)
39 measurement is primarily a study of the energy dependence of µ at and above the binding energy
of core electrons in an element. As long as the energy of the incident x-ray beam is lower than
the binding energy (E0) of the core electrons in the element, no absorption event takes place.
Once the incident x-ray energy matches the binding energy of a certain core atomic level, a sharp
increase in absorption coefficient is observed due to electronic excitation of the core electrons to
empty states near the Fermi level. This emitted electron is now called a ‘photoelectron’ since it
carries the energy of the incident photon and this photoelectron is considered to leave an empty
state in the core level, now called as a ‘hole’. Increasing the energy of incident beam above the
binding energy of the core level electrons leads to the transfer of the excess energy to the emitted
photoelectrons as the kinetic energy. This kinetic energy propels the photoelectron to travel in
the continuum and probe the nearby atomic environment. XANES region consists of localized
transitions caused by the excitation of core level electrons to the low lying empty states near the
Fermi level where as EXAFS region is a photoelectron interference phenomenon caused by the
interaction of outgoing photoelectron with the small fraction of the backscattered photoelectrons
from the nearest atomic neighbors. XANES region yields information regarding the electronic
properties of the absorber atom and surface adsorbates where as the EXAFS region can yield
information regarding the structural and geometric properties (bond lengths and coordination
numbers) of the system under investigation. The oscillations in the EXAFS region are due to the
constructive and destructive interference due to the x-ray absorption event of the incident photon
modulated by the back-scattered photoelectrons. This EXAFS fine structure is defined by the
following equation:
χ(E) = [µ(E)-µo(E)]/[∆µo(E)]
40 where µ(E) is the absorption coefficient at energy E, µo(E) is the absorption of an isolated atom,
and ∆µo(E) is the absorption at the binding energy Eo. The mathematical treatment of EXAFS
and the data analysis becomes convenient while treating the problem in terms of the emitted
photoelectron and not the energy of incident x-ray. This is performed by converting the x-ray
energy E into the so-called k-space, where ‘k’ represents the wave number of the photoelectrons
as given below:
k = [2me(E-Eo)/ħ2]1/2
where ‘me’ is the mass of an electron, ‘ħ’ is the Planck’s constant. The k-space exhibits
oscillations of different frequencies characteristic of different near-neighbor coordination shells.
Based on a single backscatter phenomenon, the EXAFS equation is derived as follows16:
where Nj is the coordination number, Bj(k) is the scattering property of the atoms neighboring
the excited atom, rj is the interatomic distance, λ(k) is the mean-free path, σ2 is the disorder in the
near-neighbor distance, and δj(k) is the central atom phase shift. EXAFS is also sensitive to the
identity of the neighboring atom since the scattering parameters in the EXAFS equation are
dependent on the atomic number of the neighboring atom. XANES region yields information
regarding the electronic properties of the absorber atom such as its site symmetry, and valence
state. On the other hand, EXAFS yields information regarding the short-range atomic structural
properties (bond lengths and coordination numbers) of the absorber atom.
41 1.5 Scope of Dissertation
This dissertation deals with three broad areas from an electrochemists’ perspective: 1)
Understand electron transfer mechanisms and reaction pathways during electrocatalytic oxygen
reduction in alkaline media as distinct from acidic conditions, (2) Develop non-noble metal
based catalysts for oxygen reduction in alkaline media, and (3) Understand the degradation
pathways of perfluorinated sulfonic acid based proton exchange membranes. A combination of
catalyst synthesis, electrochemical and in situ X-ray Absorption Spectroscopy studies have been
carried out. Pt-based catalysts have been used as model systems to understand reaction
mechanisms/pathways. This information was then used in the design of non-PGM based
catalysts such as chalcogen (S,Se) modified ruthenium nanoparticles and heat-treated metal
macrocycles. While ruthenium is a precious metal, it represents a reasonable shift away from the
expensive Pt based catalysts. The bulk of the dissertation is concerned with understanding the
initial activity of these catalyst materials under more or less ideal RRDE conditions. Preliminary
fuel cell studies under real-life conditions are treated in Chapter 6. The stability of these catalysts
is certainly important in translating them into a fuel cell electrode; however, this is beyond the
scope of the dissertation.
1.6 References
(1) Weisz, P. B. Physics Today 2004, 57, 47. (2) Lewis, N. S.; Nocera, D. G. Proceedings of the National Academy of Sciences of the United States of America 2006, 103, 15729. (3) Petit, J. R.; Jouzel, J.; Raynaud, D.; Barkov, N. I.; Barnola, J. M.; Basile, I.; Bender, M.; Chappellaz, J.; Davis, M.; Delaygue, G.; Delmotte, M.; Kotlyakov, V. M.; Legrand, M.; Lipenkov, V. Y.; Lorius, C.; Pepin, L.; Ritz, C.; Saltzman, E.; Stievenard, M. Nature (London) 1999, 399, 429. (4) Siegenthaler, U.; Stocker, T. F.; Monnin, E.; Luethi, D.; Schwander, J.; Stauffer, B.; Raynaud, D.; Barnola, J.-M.; Fischer, H.; Masson-Delmotte, V.; Jouzel, J. Science (Washington, DC, U. S.) 2005, 310, 1313. (5) Bartlett, A. A. Physics Today 2004, 57, 53.
42 (6) Trends in Renewable Energy Consumption and Electricity 2009, U.S. Department of Energy, 2011. (7) Srinivasan, S.; Mosdale, R.; Stevens, P.; Yang, C. Annual Review of Energy and the Environment 1999, 24, 281. (8) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications; Second ed.; John Wiley & Sons, Inc., 2000. (9) Slade, R. C. T.; Varcoe, J. R. Solid State Ionics 2005, 176, 585. (10) Varcoe, J. R.; Slade, R. C. T. Fuel Cells 2005, 5, 187. (11) Bockris, J. O. M.; Reddy, A. K. N. Modern Electrochemistry; Second ed.; Springer, 1998; Vol. 1 & 2. (12) Hamann, C. H.; Hamnett, A.; Vielstich, W. Electrochemistry; Second ed.; John Wiley & Sons, Inc., 2007. (13) Conway, B. E. Progress in Surface Science 1995, 49, 331. (14) Angerstein-Kozlowska, H.; Conway, B. E.; Sharp, W. B. A. J. Electroanal. Chem. Interfacial Electrochem. 1973, 43, 9. (15) Jerkiewicz, G.; Vatankhah, G.; Lessard, J.; Soriaga, M. P.; Park, Y.-S. Electrochimica Acta 2004, 49, 1451. (16) Teo, B. K. Exafs: Basic Principles and Data Analysis; Springer: New York, 1986. (17) X-ray absorption : principles, applications, techniques of EXAFS, SEXAFS, and XANES; Koningsberger, D. C.; Prins, R., Eds.; Wiley: New York, 1988; Vol. 92. (18) Mukerjee, S.; Arruda, T. Modern Aspects of Electrochemistry 2010, 50, 503.
43
Chapter 2
Impact of Double-Layer Structure and Mechanistic Changes during Electrocatalysis of
Oxygen Reduction in Alkaline Medium
2.1 Introduction
Oxygen Reduction Reaction (ORR) on noble and non-noble metal surfaces remains to be
one of the well investigated electrochemical processes. This interest stems from both
technological and fundamental standpoints. While under acidic conditions Pt and Pt alloys
remain the mainstay as catalyst materials for ORR due to the acid stability criterion, in alkaline
electrolytes a wide range of non-noble metals and their oxides are understood to be stable
enough for practical applications. Alkaline Fuel Cells (AFC), once a very promising renewable
energy source, failed to attract continued research interest primarily due to issues such as
carbonate precipitation and electrolyte leakage.1 Further, the rapid growth of Proton Exchange
Membrane (PEM) fuel cells shifted research interest into the acidic counterpart. Despite a
worldwide research effort in PEMFC, widespread commercialization is strongly predicated on
component costs and striking an optimum balance between performance and durability.
Appleby2 in 1970 envisaged the improbability of increasing the rate constants for Oxygen
Reduction Reaction (ORR) on noble metals due to the compensating changes between the pre-
exponential factors and heat of activation, and apparently this scenario has not changed
significantly since then. Further, the poor selectivity of Pt materials in the presence of impurities
and fuel crossover aggravates electrocatalysis in PEMFC. However, realization of the fact that a
wider range of non-Pt based catalyst materials can be employed in high pH environments and
recent research efforts in developing metal cation-free alkaline anion exchange membranes for
44 hydroxide anion transport that does not suffer from carbonate precipitation have spurred research
activities in AFCs.3-6 This is despite i) the two to three orders of magnitude lower hydroxide
anion conductivity in state-of-the-art alkaline membranes compared to proton transport in acidic
membranes, 7 and ii) manifestation of poisoning due to carbonate anion exchange in alkaline
membranes.8 The claim that alkaline fuel cell performs better than PEMFC remains quite
unjustified since such claims have usually been based on comparisons between phosphoric acid
fuel cells and AFCs.9 On the contrary, comparison of performance between PEMFC and liquid
electrolyte based AFC technologies roughly exhibit equivalent performances.9-10 However, the
fact that wider range of non-Pt catalyst materials can be employed in alkaline media is
completely justified given ample evidence in the literature employing several non-noble
materials as electrocatalysts and the dispensability of the so called acid-stability criterion at high
pH environments.1,11-12 Although several ambitious engineering designs were tested to prevent
leakage and carbonate precipitation issues while improving AFC performances, this did not
revive the prospects of AFC technology.1,13-14 However, replacement of liquid electrolytes in
conventional AFC with metal cation free Alkaline Anion Exchange Membrane (AAEM) that can
transfer hydroxide ions (OH−) can revitalize the AFC technology and impart a new momentum to
it.3-4
ORR pathway rather than ORR mechanism has typically been addressed in the literature
due to the easy accessibility of the former from rotating ring-disk electrode (RRDE) based
studies and the complexity in understanding the latter based on electrochemical and
spectroscopic results.15 Pt and Pt alloy based catalysts remain well investigated primarily from
electrochemical kinetic studies on single crystals and in situ X-ray absorption spectroscopic
studies.16-17 ORR pathways is found to be similar in both acid and alkaline media on Pt based
45 materials.15,18 Based on the initial propositions by Damjanovic et al,18-20 the rate determining step
on Pt electrodes is widely agreed to be the first electron transfer step to the adsorbed molecular
O2 with or without rapid proton transfer. These kinetic studies involve the elucidation of
relationships of dE/d(log i), dE/d(log pO2), and dE/d(pH).18-19 It is also understood that the ORR
kinetics are qualitatively different on pre-reduced, oxide free electrodes and electrodes that
contain a thin film of oxide coverage on them.21-23 Two Tafel slope regions in both acid and
alkaline medium are typically observed. Tafel slopes of -2.3RT/F (-60 mV/dec) obtained in the
low current density region follows the Temkin adsorption isotherm due to intermediate oxide
coverage arising from ORR reaction intermediates. In the high current density region the Tafel
slope of -2*2.3RT/F (-120mV/dec) is governed by the Langmuir isotherm since significant oxide
coverage ceases to exist at these potentials.19 However Tarasevich et al24 later pointed out that
the adsorbed OH species on Pt surfaces that inhibit O2 adsorption arise primarily from water
activation, whereas the ORR reaction intermediates typically exhibit lower coverage values
compared to water activation products. Over all pH ranges and entire current density regions it is
known that the reaction order with respect to molecular O2 is unity.19 Also, from the kinetic
studies it is now known that the reaction order of the rate determining step with respect to H3O+,
and OH¯ (i.e. dE/d(pH)), is respectively 3/2 and -1/2 in the low current density region.19 In the
high current density region, it is observed that the reaction order with respect to H3O+, and OH¯
(dE/d(pH)), is respectively 1 and 0. This unusual “fractional” reaction order with respect protons
or hydroxyl ions is the manifestation of the dependence of free energy of activation (∆G*) on the
intermediate oxide coverage (as governed by Temkin adsorption isotherms) and hence on the pH
of the electrolyte. For a recent review of the kinetics, see references Spendelow et al6,
Gottesfeld25, and Adzic et al26. Further, a combination of electrochemical and spectroscopy
46 studies has shown that Pt-M alloys (M = Co, Ni, Cr, etc.) show improved kinetics due to a
combination of structural and electronic factors that serve to inhibit OH poisoning of the Pt
sites.17,27-31
A major alternative viewpoint to the rate determining step in ORR was proposed by
Yeager et al. 32 Accordingly, it was proposed that ORR on Pt surfaces is likely to involve
dissociative chemisorption of molecular O2 with the initial adsorption of O2 (with or without an
electron transfer) as the rate determining step. Also, based on hydrogen-deuterium kinetic
isotope studies the rate determining step was not likely to involve proton transfer. In gas phase
studies there are several evidences for dissociative chemisorption of molecular O2.15 However,
under electrochemical conditions where solvation effects and other surface adsorbed species
(OHads, electrolyte anions) are present, adsorption of molecular O2 is likely to be weakened.32
ORR on Pt based catalysts is understood to proceed via “parallel” routes with the 4e
“direct” or “series” pathway as the predominant route and a minor route involving a 2e pathway
to peroxide. This parallel generation of peroxide is mainly related to the oxides, anions and
impurities on the surface that weaken the adsorption of the peroxide intermediate.15,31 The
Rotating Ring Disk Electrode (RRDE) technique has been intensively used to understand the
kinetics of ORR.33 This technique involves the convective transport of dissolved molecular
oxygen from the bulk to the electrode surface prior to diffusive transport within the diffusion
layer to the catalyst site. This technique allows the detection of stable reaction intermediates
generated at the disk during ORR by potential control of the ring electrode. Various kinetic
models have been developed to understand the reaction pathways involved in ORR. The first
model was developed by Damjanovic et al34 following which Wroblowa et al,35 and
Anastasijevic et al,36-37 proposed extensive models. Briefly there are two 4e¯ pathways and one
47 2e pathway. 4e transfer could either be a direct 4e or a series 4e¯ pathway. The “direct”
pathway involves the concerted transfer of 4e to adsorbed molecular oxygen to form H2O
without the formation of peroxide intermediates. The “series” 4e pathway involves sequential
transfer of electrons to adsorbed molecular oxygen to form adsorbed peroxide species which
without desorbing from the surface involves in another 2e transfer to form water. While these
two pathways are rather ideal definitions of the two possible 4e pathways, it is important to
understand the following aspects of this distinction as “direct” and “series”. The direct 4e
transfer requires the concerted transfer of 4e and 4H+ to the dissociatively chemisorbed
molecular oxygen. However evidence for dissociative adsorption of O2 under electrochemical
conditions is not available in the literature, making this pathway very less likely. So it is quite
possible that the so called “direct” 4e pathway proceeds via the peroxide pathway such that the
peroxide intermediate does not desorb from the catalyst surface to any appreciable extent. In this
particular case, the ring-disk kinetic studies are incapable of making the distinction between a
direct 4e pathway and a series pathway involving 4e transfer.24 The 2e pathway involves the
transfer of two electrons to the adsorbed molecular oxygen forming the peroxide species and the
peroxide intermediate diffuses to the electrolyte bulk without any further reduction. This is the
case where peroxide is the final product and the catalyst is incapable of reducing the peroxide
intermediate any further. An “interactive” pathway was also defined in order to treat catalyst
materials exhibiting heterogeneity in active sites where intermediate species undergo surface
diffusion and further reduction at more active sites.36-37 Given the pKa values for the first and
second ionization of H2O2 at 25˚C (pK1 = 11.69 and pK2 = ~20) the predominant peroxide
species for pH>12 is HO2. 24 Further, another important distinction in alkaline media is that in a
4e “series” pathway the lower working electrode potential range on an absolute scale causes the
48 HO2 intermediate to desorb from the catalyst surface, however, in the analogous case of acidic
media the higher working electrode potential range decreases the possibility of H2O2 desorption
from the electrode surface.6 As mentioned by Appleby,38 an alkaline electrolyte essentially acts
as a homogeneous catalyst due to its stabilization effect on the intermediate product HO2.
For electrocatalytic reactions proceeding via inner-sphere electron transfer mechanisms,
it is typically assumed that either molecular adsorption of reactant species (dissociatively or non-
dissociatively) or an electron transfer is the first step.39 However, for neutral, non-polar species
like molecular O2, direct molecular O2 adsorption is likely to be inhibited relative to say, for
example adsorption of superoxide radical anion (O2•¯) unless the free energy of adsorption of the
O2 molecule is very exothermic on a specific catalytic surface. This is especially true under fuel
cell conditions where the cathodic reaction typically occurs at potentials well positive of the
potential of zero charge (pzc). Multi-step, multi-electron transfer processes like ORR that
involve many adsorbed intermediates undoubtedly classifies as an inner-sphere electron transfer
reaction. However, among the many elementary reaction steps involved in ORR there could be a
surface-independent outer-sphere electron transfer component in the overall electrocatalytic 4e¯
inner-sphere electron transfer reaction. In that perspective O2 reduction by one-electron transfer
to superoxide (O2•¯) is observed at E° = -0.3±0.03 V vs. SHE corresponding to ∆G°=30±2 kJ
mol-1 with both O2 and O2•¯ remaining in the aqueous phase.40-41 Given the pH independence of
this redox couple (O2/ O2•¯), the potential of this reaction does not change as the pH is varied
from zero to fourteen.42 However, due to the occurrence of four proton transfer steps in oxygen
reduction to H2O/OH , its standard reduction potential changes by 0.828 V from 1.229 V to
0.401 V vs. SHE as the pH value changes from zero to fourteen. This causes the overpotential for
the first electron transfer step (O2/ O2•¯) to decrease from 1.53 V at pH=0 to 0.7 V at pH=14,
49 indicating a sharp decrease in overpotential at alkaline pH conditions. Markovic et al42 argued
based on a modified Pourbaix diagram approach that the above mentioned decrease in
overpotential is the primary thermodynamic reason for the applicability of a wide range of non-
noble materials in alkaline media. Due to the high overpotential required for O2/O2•¯ redox
couple in acidic media, only certain specific catalyst surfaces such as platinum that offer high
free energy of adsorption for O2 can catalyze ORR in acidic media. However, in alkaline media,
decrease in the overpotential for O2/O2•¯ causes almost all electronically conducting electrode
materials to be ORR active at alkaline pH.42 While the decrease in overpotential for the first
electron transfer is certainly significant, this argument is primarily of thermodynamic origin. The
concept of involving the possibilities of outer-sphere electron transfer during ORR in alkaline
media bears importance and it was pointed out earlier by Bockris1, and Appleby 38 that the
exchange current density values in alkaline media exhibit near independence on a large number
of electrode materials including silver, gold, manganese oxides, perovskites and various carbon
surfaces. So certain steps in the overall ORR process in alkaline media could proceed via a non-
electrocatalytic pathway.38
In this report we investigate the kinetics of ORR in alkaline media from the perspective
of the reaction mechanisms and the double layer structure. While the fundamental
electrochemical aspects of ORR on Pt in acidic media have been thoroughly investigated by
various research groups and continue to be a subject of intense study, oxygen reduction on Pt in
alkaline medium exhibit certain interesting behavior that has not been discussed previously in
detail in the literature. A combination of pertinent review of the literature along with
experimental results shown here are used to unravel the various possible ORR reaction
mechanisms in alkaline media.
50
2.2 Experimental
2.2.1 Electrochemical characterizations:
30% Pt/C catalyst from BASF-ETEK (Somerset, NJ) was used as received. All
electrochemical measurements were made at room temperature using a rotating ring-disk
electrode (RRDE) setup from Pine Instruments connected to an Autolab (Ecochemie Inc., model-
PGSTAT 30) bi-potentiostat. Alkaline (0.1 M NaOH) and acidic (0.1 M HClO4) electrolytes
were prepared using sodium hydroxide pellets (semiconductor grade, 99.99%, Sigma-Aldrich)
and double-distilled 70% perchloric acid (GFS Chemicals) respectively. Catalyst inks were
typically prepared by dispersing 25mg of the catalyst in 10 mL of a 1:1 Millipore H2O:isopropyl
alcohol mixture along with 100µL of 5wt% Nafion(R) solution as a binder. 10µL aliquot of the
catalyst ink was dispensed on Glassy Carbon (GC) disk of 5.61mm dia. Gold ring electrode was
held at 1.1 V vs. RHE in alkaline electrolyte and at 1.3 V vs. RHE in acidic electrolyte to detect
stable peroxide intermediates. The collection efficiency of the disk-ring electrode was 37.5%. All
potentials are refered to the Reversible Hydrogen Electrode (RHE) scale prepared from the same
solution as the bulk electrolyte unless otherwise stated. More details on the RRDE methodology
can be found in an earlier publication.31
2.3 Results and Discussions
2.3.1 Structural Aspects of the Double-layer:
As mentioned above in traversing a pH range of 14 from acidic to alkaline media the
working electrode potential decreases by about ~0.83 V based on the Nernst equation.
Spendelow et al6 pointed out that the change in electrode potential is likely to have significant
51 consequences on the energetics of adsorption of reactants, intermediates, and products. Further
this decrease in potential also affects the strength of adsorption of spectator species,
contaminants, and anions from the supporting electrolytes.6 Strong poisons in acidic media such
as halide anions do not poison the platinum active sites in alkaline media. This is primarily due
to the fact that the lower working electrode potential in alkaline media causes the excess surface
charge on the electrode to be relatively more negative than that in acidic media. This excess
negative charge repels the chloride anions away from the inner-Helmholtz plane (IHP). This is
also one reason why typically higher peroxide anion (HO2¯) intermediate is detected at the ring
in alkaline media compared to the neutral H2O2 intermediate in acidic media.6
Besides the effect of working electrode potential on adsorption strength, it is likely that
there is a significant change in the double-layer structure at the electrode/electrolyte interface as
the pH is changed from zero to fourteen. As shown in Figure 2.1, in a typical aqueous acidic
electrolyte of oxygen saturated 0.1 M HClO4 the primary constituents are hydronium ions
(H3O+), perchlorate anions (ClO4¯), solvated molecular oxygen, and the solvent water molecules.
A brief description of the various constituents of the compact part of the electrochemical double
layer and their influence on ORR is given below. Given that the concentration of the
acidic/alkaline supporting electrolyte is typically ≥ 0.1 Molar, the diffuse part of the double layer
is not considered. Cathodic potentials of oxygen reduction in an operating fuel cell typically
occur at potentials well positive of the potential of zero charge (pzc). At these conditions in
acidic media, chemisorbed molecular O2 (either dissociatively or non-dissociatively adsorbed),
specifically adsorbed hydroxyl species (OHads arising from water activation), and solvent water
dipoles constitute the IHP. Solvated molecular O2, and ClO4¯ anions populate the outer-
Helmholtz plane (OHP). The distance of closest approach of H3O+ ions to the electrode surface is
52
Figure 2.1: Schematic illustration of the double-layer structure during ORR in acidic (left) and alkaline (right) media. Insets (a) and (b) illustrate the inner and outer-sphere electron transfer processes.
limited to the outer-Helmholtz plane (OHP) owing to its net positive charge. Consider the first
electron transfer step to the adsorbed O2,ads species to form the superoxide intermediate (O2•¯)ads.
Only after this first electron transfer step has taken place, protonation of the (O2•¯)ads intermediate
by the transfer of proton from OHP to the IHP could take place. Since protons in acidic media
have very high mobility this step is not rate limiting. To be an efficient electrocatalyst, the ORR
reaction intermediates should remain adsorbed on the catalyst site until 4e and 4H+ are
transferred followed by desorption of stable H2O molecule as the final product.
53
In the case of alkaline media (0.1 M NaOH), although the double layer structure is not
dramatically different there are some important aspects that need to be taken into account as
shown in Figure 2.1. At high pH water molecules act not only as solvent but also serve as the
source of protons required in ORR. IHP is now populated by specifically adsorbed hydroxyl
species (OHads arising from OH anion adsorption), solvent water dipoles, and chemisorbed O2.
Alkali metal ions are typically well solvated and are classically expected to populate the OHP. In
the case of a typical electrocatalytic inner-sphere mechanism, electron transfer to O2,ads to form
(O2•¯)ads followed by proton transfer from water molecules take place. Proton transfer could
occur from water molecules co-adsorbed on the electrode surface or could also be from the
partial solvation shell of adsorbed O2/(O2•¯)ads. Increasing alkalinity of the supporting electrolyte
(pH>12) causes the rate of proton transfer from water to decrease concomitant to the decrease in
water activity. This is primarily the reason for increased stability of the superoxide radical anion
O2•¯ in strongly alkaline electrolytes.43 The above ideas primarily reflect the electrocatalytic
inner-sphere electron transfer mechanism. The outer-sphere electron transfer to form superoxide
O2•¯ is characterized by the following equation: 40
O2,aq + e → (O2•¯)aq E0 = -0.33 V vs. SHE (2.1)
As mentioned above, the overpotential for this reaction in acidic media (pH = 0) is ~1.53 V
which decreases significantly to ~0.7 V in alkaline media (pH = 14). This decrease in
overpotential implies that strong chemisorption of O2 to the electrode surface is not a
prerequisite. Other non-covalent forces such as long-range dipole-dipole interactions or the free
energy associated with hydrogen bonding (typically < 35 kJ mol-1) could be sufficient enough to
overcome the overpotential required for this reaction (vide infra).
54 2.3.2 ORR on Pt/C Nanoparticles: Acid vs. Alkaline media:
ORR on Pt based catalysts in aqueous electrolytes have been debated intensively in the
literature 16-17,44-48 and only some new information relevant to the context of this thesis will be
presented here. Figure 2.2 provides a snapshot of the cyclic voltammetry (CV) and ORR of 30%
Pt/C in 0.1 M NaOH and 0.1 M HClO4 electrolytes. Table 2.1 shows the electrochemical kinetic
parameters extracted from this plot. The electrochemical surface area due to H
adsorption/desorption was measured to be 70±2 m2/gPt and 43±1 m2/gPt in 0.1 M HClO4 and 0.1
M NaOH electrolytes, respectively, which is in good agreement with the literature.47,49 The CV
of 30% Pt/C catalyst features the typical hydrogen underpotential deposition/stripping region
below 0.5 V vs. RHE in both electrolytes followed by oxide formation on Pt at potentials above
0.7 V vs. RHE. Oxide formation on Pt in acidic media is due to oxidation of the solvent water
molecules (water activation)44 and in alkaline media is due to specific adsorption of hydroxide
anions from the supporting electrolyte.24,50 The onset potential of Pt-OH formation is similar in
both electrolytes although in alkaline media oxide formation current exhibits a characteristic
peak shape where as in acidic media the oxide formation current is relatively more flat. The half-
wave potential (E1/2) of Pt-OH in 0.1 M NaOH is 0.775 V whereas in 0.1 M HClO4 E1/2 of Pt-OH
formation is shifted slightly more positive to 0.810 V. Figure 2.2(b) shows the charge density
due to Pt-OH formation in the two electrolytes. Below 1 V vs. RHE, which is the potential region
of interest to ORR, charge density on Pt/C was found to be marginally higher in 0.1 M NaOH
than in 0.1 M HClO4. Figure 2.2(c&d) shows the ORR polarization curve at 900 rpm and mass-
transport corrected Tafel plots respectively taken at 20 mV/s. The onset potential of ORR in both
the electrolytes is ~1 V vs. RHE, which is followed by a mixed kinetic-diffusion region between
55 the potentials 0.7 V and 1 V. A well defined diffusion limited current density region is observed
below 0.6 V vs. RHE.
(a) Cyclic Voltammetry
E Vs RHE0.0 0.2 0.4 0.6 0.8 1.0 1.2
Cur
rent
Den
sity
[A/c
m2 ge
o]
-4e-4
-2e-4
0
2e-4
4e-4
0.1M HClO40.1M NaOH
(c) Polarization Curves
Potential [V Vs RHE]0.0 0.2 0.4 0.6 0.8 1.0
Cur
rent
Den
sity
[A/c
m2 ge
o]-5e-3
-4e-3
-3e-3
-2e-3
-1e-3
0
0.1M HClO40.1M NaOH
(e) Ring Current
Potential [V Vs RHE]
0.0 0.2 0.4 0.6 0.8 1.0
Rin
g C
urre
nt [A
]
0
1e-5
2e-5
3e-5
4e-5
0.1M HClO4
0.1M NaOH
(b) Oxide Charge Density
0.2 0.4 0.6 0.8 1.0 1.2 1.4
Cha
rge
Den
sity
[ µµ µµC
/cm
2 ]
0
1000
2000
3000
0.1M HClO4
0.1M NaOH
E Vs RHE
(d) Tafel Plots
log ik [A/cm2geo]
1e-4 1e-3 1e-2
E V
s R
HE
0.80
0.85
0.90
0.95
0.1M HClO4
0.1M NaOH
25mV
Figure 2.2: Comparison of electrochemical characteristics of BASF-ETEK 30% Pt/C in 0.1 M HClO4 and 0.1 M NaOH electrolytes at a loading of 15 µgPt/cm2 on 5.61 mm Glassy Carbon disk electrode. (a) Cyclic voltammetry in de-oxygenated electrolytes, (b) charge density due to oxide formation on Pt, (c) ORR polarization curves at 900 rpm in O2 saturated electrolytes, (d) mass transport corrected Tafel plots, and (e) ring current measured during ORR at 900 rpm. Scan rate: 20 mV/s. ERing = 1.1 V vs. RHE in 0.1 M NaOH and 1.3 V vs. RHE in 0.1 M HClO4.
56 Table 2.1: Electrochemical kinetic parameters
a Catalyst Loadings: Pt - (15 µgPt/cm2geo)
b io – Exchange current obtained by extrapolating the Tafel plot in the region 0.95 V to 0.90 V b Tafel slopes were measured in the potential range (0.95 V-0.90 V/0.90 V-0.80 V) b From the slope of Koutecky-Levich plot (ilim vs. ω0.5)
Two Tafel slope regions could be discerned as shown in Table 2.1. A Tafel slope of
approximately -60±3 mV/dec was observed in the potential range of 0.95 V to 0.90 V vs. RHE
and another slope of approximately -100 to -120 mV/dec were observed in the potential range of
0.90 V to 0.80 V in both 0.1 M NaOH and 0.1 M HClO4 electrolytes. This is in good agreement
to the literature data on supported polycrystalline Pt nanoparticles.46 This indicates similar ORR
pathways on Pt/C at the extreme pH conditions measured here. However, the marginally higher
Pt-OH coverage in alkaline medium causes a penalty of ~25 mV higher overpotential at a kinetic
current density of ik = 1 mAcm-2geo for ORR in 0.1 M NaOH relative to 0.1 M HClO4. As shown
in Table 2.1, lower kinetic current density values and a one order of magnitude lower exchange
current density is obtained on Pt in 0.1 M NaOH compared to 0.1 M HClO4 electrolyte. Similar
observations were made on Pt low index single crystal surfaces where the higher OHad coverage
on Pt in KOH electrolyte relative to HClO4 caused the ORR activity to be higher in acidic media
than in alkaline media.50-51
The most interesting observation in this experiment using Pt/C is in the ring currents
measured during ORR at 900 rpm rotation rate. Figure 2.2(e) shows the ring current due to the
oxidation of peroxide intermediates generated during ORR at the Pt/C catalyst deposited on the
Catalysta Electrolyte ik x103 [A/cm2
geo] @ 0.9 V/0.8 V
iox109 [A/cm2
geo]b
Tafel Slopes
[mV/dec]c
Number of Electrons
Transferredd
Peroxide Yield (%)
0.7 V 0.8 V
30% Pt/C 0.1 M NaOH 1.15/11.6 7.00 61/105 3.7 0.35 0.73 30% Pt/C 1.0 M NaOH 0.72/10.5 0.06 44/80 3.6 1.45 1.73 30% Pt/C 0.1 M HClO4 2.32/23.0 47 70/101 4.0 0.24 0.27
57 GC disk. Firstly in acidic electrolyte the ring current due to peroxide oxidation is lower in the
entire potential region and does not show any significant increase until the disk potential enters
the hydrogen under potentially deposited (UPD) region. On the contrary, the ring current due to
peroxide oxidation in 0.1 M NaOH electrolyte shows a sharp increase at 0.9 V in the cathodic
scan. The sharp increase in ring current at 0.9 V during ORR in alkaline media is closely related
to Pt-OH formation from specific adsorption of hydroxide anions as seen from the cyclic
voltammetry in 0.1 M NaOH electrolyte. As shown in Figure 2.2(a), in 0.1 M NaOH electrolyte,
Pt-OH formation commences at ~0.7 V vs. RHE, reaches a peak current at 0.81 V and a plateau
at ~0.9 V. On the cathodic direction the peak potential for Pt-OH reduction is ~0.75 V vs. RHE.
As seen in Figures 2.2 (c&e), on the cathodic scan beginning ORR on the disk at 1.00 V, the
increase in ring current commences at 0.9 V and reaches a peak potential of 0.75 V vs. RHE.
This clearly indicates that there is an interaction between molecular oxygen and the hydroxyl
species on the surface.
Looking back at the double layer structure in alkaline medium depicted in Figure 2.1, the
solvated molecular O2, represented here in the text as a cluster O2.(H2O)n, can interact with the
surface hydroxyl species (OHads) via a hydrogen bond between the H atom in OHads and the O
atom in the solvent water molecule. Such hydrogen bond energies (<35 kJ mol-1) are typically
much lower than the energy associated with covalent bond strength such as in the case of direct
chemisorption of O2 on Pt (>300kJ mol-1).52 Such low interaction energies due to hydrogen bond
formation are sufficient enough to overcome the overpotential for the first electron transfer
reaction in alkaline media according to equation (2.1) shown above. This hydrogen bond
formation stabilizes the solvated molecular oxygen O2.(H2O) cluster in the OHP and promotes an
outer-sphere electron transfer to form the superoxide species in alkaline media. Contrarily in
58 acidic media, although hydrogen bond formation could still take place, the overpotential for the
first electron transfer via an outer-sphere process is too large to be overcome by this hydrogen
bond formation.
So, in the case of ORR on Pt in alkaline media, two mechanisms are proposed here. First
is the well-known electrocatalytic inner-sphere electron transfer mechanism where molecular O2
undergoes direct chemisorption on an oxide-free Pt site leading to a direct/series 4e¯ pathway
without the desorption of reaction intermediates such as peroxide from the surface according to
the following well known reaction scheme.15
O2 → O2,ads (2.2a)
O2,ads + H2O + 2e → (HO2¯)ads + OH (2.2b)
(HO2¯)ads + H2O + 2e → 3OH (2.3)
Second is the outer-sphere electron transfer mechanism where solvated molecular O2 cluster
O2.(H2O)n weakly interacts with adsorbed hydroxyl species to promote a 2e¯ reaction pathway to
HO2¯ anion as a reaction product which desorbs from the surface and is eventually detected at the
ring electrode. This reaction is formulated as shown here:
Pt-OH + [O2.(H2O)n]aq + e → Pt-OH + (HO2•)ads + OH + (H2O)n-1 (2.4)
(HO2•)ads + e → (HO2¯)ads (2.5)
(HO2¯)ads → (HO2¯)aq (2.6)
The first step in the above reaction shown in equation (2.4) involves electron transfer (or
tunneling) from the electrode surface across a thin oxide film and at least one layer of solvation
shell to solvated O2. Equation (2.4) above involves several elementary steps as written below:
Pt-OH + [O2.(H2O)n]aq + e → Pt-OH + [O2•¯.(H2O)n]aq (2.4a)
[O2•¯.(H2O)n]aq → (O2
•¯)ads + nH2O (2.4b)
59
(O2•¯)ads + H2O → (HO2
•)ads + OH (2.4c)
First electron transfer to O2,aq forms (O2•¯)aq which then undergoes desolvation and subsequent
adsorption on the oxide sub-structure of the Pt surface to form (O2•¯)ads, followed by proton
transfer to form adsorbed hydroperoxyl radical, (HO2•)ads. Second electron transfer to (HO2
•)ads
yields (HO2•¯)ads. The binding energy of (HO2
•¯)ads on the oxide sub-structure of Pt is likely to be
lower than that on an oxide free Pt site. This leads to the facile desorption of HO2¯ anion into the
electrolyte which is eventually detected at the ring electrode. The interaction between the
O2.(H2O)n cluster and the surface hydroxyl species causes certain non-specificity to the identity
of the underlying electrode metal. This non-specificity opens the gate to use a wide-range of non-
noble metals and their oxides as electrode materials for ORR in alkaline media. Further, as
identified in Figure 2.2(e), we consider that this peak-shaped ring current in the potential range
of 0.6 V to 0.9 V in alkaline media to be a characteristic signature of the outer-sphere electron
transfer reaction mechanism. In acidic media the adsorbed OHads species from water activation
primarily serve only to block/inhibit the adsorption of molecular O2 and other reaction
intermediates via the well known site-blocking effect.16,53 However, as shown here in alkaline
media the OHads species not only block the direct adsorption of O2 but also serves to promote the
2e outer-sphere electron transfer reaction to peroxide.
There are several precedents for electrochemical reactions which are mediated or
promoted by specifically adsorbed anions and surface groups.54 Catalysis of solvated metal
cations Maq2+/3+ (M = Fe, Eu, V) by surface specific carbonyl groups was proposed earlier by
McCreery et al,54-55 wherein formation of a hydrogen bond between the carbonyl groups and a
complexed water molecule was speculated. Similar phenomena are also observed with
specifically adsorbed halide anions promoting the outer-sphere redox reactions of cobalt
60 complexes.56-57 The anions or the surface groups act as an outer-sphere bridge between the
reactant and the electrode surface. Here, in the case of ORR in alkaline media the surface
hydroxyl species catalyzes or promotes the outer-sphere reaction and this mechanism is likely to
be faster in rate than the corresponding parallel inner-sphere reaction. So, Tafel slopes of ORR
on Pt in alkaline media are similar to those in acidic media because the rate determining step is
still the first electron transfer to adsorbed molecular O2 via the inner-sphere mechanism at an
oxide free Pt site.
It is important to understand the nature of the hydroxyl species that promotes/mediates
the outer-sphere electron transfer reaction. Beginning with the pioneering work by Conway58 and
others, it is now well established that oxide formation on Pt in aqueous electrolytes progresses
according to the following steps (a) through (f) with increasing anodic potentials; (a) initial
adsorption of OH(Pt-OH(γ-1) with γ=0), (b) charge transfer to form electro-neutral OHads (Pt-
OH(γ-1) with γ=1), (c) deprotonation to form Pt-Oads, (d) place exchange to O-Pt to form a quasi-
2D lattice, and (f) finally the progressive formation of 3D-bulk oxide phase.53,58-59 While place
exchange and later steps are understood to occur at potentials >1.15 V, the primary oxide species
of interest to ORR conditions below 1.0 V is typically (OHads)γ-1 (0<γ<1) and Oads. Ideally,
deprotonation (step c) is expected to commence only after complete charge transfer (step b) of all
adsorbed OH species indicating discrete potential windows for each step. However in reality
due to surface heterogeneity there is sufficient overlap of potential regions for steps (b) and (c).
In this context, the recent study by Jerkiewicz et al53 shows that Pt-Oads formation takes place at
potentials <1.15 V. In a study of ORR on Pt low index single crystals in alkaline media
Markovic et al 50-51 showed that this peroxide detected on the ring in the mixed kinetic-diffusion
region is observed only on Pt(100) and Pt(110) surfaces. No such peroxide intermediate was
61 detected on the Pt(111) surface. This difference was attributed to the state of reversibility of the
OHads species based on the anodic and cathodic charges obtained due to oxide formation.
Pt(111)-OHads was found to be reversible whereas on Pt(100) and Pt(110), the OHads species was
found to be irreversible. While this could be an important criterion, besides this factor the state of
discharge of the adsorbed hydroxyl species, i.e. the electrosorption valency (γ) need to be taken
into account. The interaction of O2.(H2O) cluster via a hydrogen bond between the H atom in
(OHads)γ-1 species and O atom in the water molecule would be inhibited unless the
electrosorption valency (γ) is close to unity, meaning almost complete charge transfer from the
OHads species. This inhibition would likely arise due to electrostatic repulsion between the net
negative charge on (OHads)γ-1 and the negative oxygen atom dipole in water thereby increasing
the distance between OHads and O2.(H2O)n cluster leading to a decrease in the electron
transmission coefficient of the outer-sphere process. So, a γ value of close to unity would
facilitate the hydrogen bond formation required for outer-sphere electron transfer. In this context,
during the anodic scan the potential at which the reaction step OH¯ → OHads + e is complete
could trigger the outer-sphere electron transfer. Similarly on the cathodic scan the potential at
which the reaction Oads + H+ + e → OHads begins could trigger the outer-sphere electron
transfer. Recently it was pointed out by Markovic et al60 that in alkaline media certain non-
covalent forces of interaction (such as H-bond formation) between the solvated alkali metal
cation in OHP and the adsorbed hydroxyl species in IHP are observed leading to quasi-specific
adsorption of clusters such as OHad-M+(H2O)x. Such quasi-specific adsorption was found to
stabilize the OHads species on the Pt surface and influence the kinetics of electrocatalytic
reactions. However, such an effect was primarily due to the site-blocking nature of the OHad-
62 M+(H2O)x species and primarily deals with solvated alkali metal cations that are not redox active
in the potential window of aqueous electrolytes.60
In Figure 2.2(e), peroxide intermediate detected at the ring in 0.1 M NaOH within the
potential window 0.6 V to 0.35 V primarily arises from the 2e¯ reduction of O2 to HO2¯ from the
carbon support.11 The carbon support promotes outer-sphere electron transfer via mediation by
the quinone/hydroquinone surface functional groups.61-63 A mechanism similar to the one
proposed above for the mediation of outer-sphere electron transfer by Pt-OHads species can be
extended to the mediation reaction by the quinone/hydroquinone surface functional groups on
carbon. Hydrogen bond formation between the hydroquinone species and the oxygen in one of
the solvent water molecules in the O2.(H2O)n cluster causes an electron transfer via an outer-
sphere mechanism to form (O2¯)aq. This superoxide species could either undergo
disproportionation reaction to form O2 and HO2¯ or undergo adsorption on the carbon surface
followed by proton and electron transfer to form HO2¯ anion. The relative rate of these two
routes depends on the electrolyte pH and surface area of the carbon support as discussed in detail
by McCreery et al.41 Finally, there is a sharp increase in ring current as the disk potential enters
the H adsorption/desorption region. There are several reasons for this behavior. Peroxide
detected in this high overpotential region could be of relevance in an operating fuel cell where
hydrogen permeates from the anode across the membrane to the cathode (also the alternate
situation of oxygen diffusion to the anode) and skews the ORR route to a 2e pathway generating
peroxide intermediates. In this high overpotential region there are two primary reasons for
increased peroxide yield. Firstly, the presence of adsorbed hydrogen species deprives molecular
O2 of the ensemble of Pt sites required for chemisorption.50 The second reason is closely related
to the potential of zero total charge (pztc) of Pt in aqueous electrolytes. Pt pztc is typically found
63
to be 0.3±0.05 V depending on the methodology used to measure it, impurity effect and specific
adsorption.53,64 As the disk potential is swept across ~ 0.3 V, the Pt catalyst traverses through the
pztc region. Under this condition, the solvent water dipoles localized on the electrode surface
undergo molecular rotation by changing the orientation from oxygen end (flip-up state of water)
facing the electrode to an orientation with the hydrogen atoms facing (flip-down state of water)
(c) Ring Current
Disk Potential [V Vs RHE]0.0 0.2 0.4 0.6 0.8 1.0
Rin
g C
urre
nt [A
]
0
5e-6
1e-5
2e-5
2e-5
3e-5
3e-5
4e-5
(a) Polariztion Curves
Disk Potential [V Vs RHE]0.0 0.2 0.4 0.6 0.8 1.0
Cur
rent
Den
sity
[A/c
m2 ge
o]
-5e-3
-4e-3
-3e-3
-2e-3
-1e-3
0
0.1 M NaOH1.0 M NaOH
(b) Tafel Plots
log ik [A/cm2geo]
1e-5 1e-4 1e-3 1e-2
E V
s R
HE
0.85
0.90
0.95
Figure 2.3: Comparison of electrochemical characteristics of BASF-ETEK 30% Pt/C in 0.1 M NaOH and 1.0 M NaOH electrolytes at a loading of 15 µgPt/cm2 on 5.61 mm Glassy Carbon disk electrode. (a) ORR polarization curves at 900 rpm in O2 saturated electrolytes, (b) mass transport corrected Tafel plots, and (c) ring current measured during ORR at 900 rpm. Scan rate: 20 mV/s. ERing = 1.1 V vs. RHE.
64 the electrode surface.64-65 As a result the hydrogen atoms in the water molecules are not freely
available for ORR process. This causes a mechanistic disadvantage to ORR.
Figure 2.3 compares the results of ORR experiments performed on 30% Pt/C catalyst in
0.1 M and 1.0 M NaOH electrolytes. Figure 2.3(a) shows that the ORR polarization curves in
both the electrolytes are qualitatively similar but a lower diffusion limited current density in 1.0
M NaOH compared to 0.1 M NaOH electrolyte is observed. This is due to the lower O2 solubility
(CO2), diffusion coefficient (DO2) and higher kinematic viscosity (υ) in going from 0.1 M to 1.0
M concentration of NaOH. In 0.1 M NaOH CO2 = 1.22x10-6 mol cm-3, DO2 = 1.90x10-5 cm2s-1,
and υ = 8.70x10-3 cm2s-1.50 In 1.0 M NaOH CO2 = 8.40x10-7 mol cm-3, DO2 = 1.65x10-5 cm2s-1,
and υ = 1.06x10-2 cm2s-1.66 Finally as seen in Figure 2.3(c), the peroxide detected on the ring in
0.1 M NaOH between the potential window 0.6 V to 0.9 V is observed only in the cathodic scan
but not in the anodic scan. This is primarily associated with the state of the adsorbed oxide
species in terms of its chemical nature (either OHads, or Oads), coverage, and electrosorption
valency. However at a higher concentration of alkaline electrolyte, i.e. at 1 M NaOH, peroxide is
detected at the ring electrode in both the anodic and the cathodic scan. This is observed because
at any given potential the fraction of electroneutral OHads coverage on Pt is higher in 1.0 M
NaOH relative to a lower concentration of 0.1 M NaOH.
Finally an effort is made here to review and reconcile some difference regarding the ORR
characteristics of low-index gold single crystals in aqueous electrolytes. Low index single
crystals of gold exhibits remarkable structure sensitivity towards ORR in alkaline media.67-69
ORR on Au(hkl) low-index single crystals in acidic media proceeds predominantly to a two
electron H2O2 product.70-71 In alkaline media Au(111) and Au(110) crystal faces perform only
2e reduction whereas Au(100) crystal face exhibits 4e¯ reduction of O2 to OH . This 4e
65 reduction on Au(100) in alkaline media is observed only within a narrow potential window of
mixed kinetic-diffusion region between 0.05 V to -0.35 V vs. Hg/HgO, OH¯. In the diffusion
limited current region on Au(100) in alkaline media ORR reverts back to a 2e peroxide anion
pathway. The change in pathway from 4e¯ to 2e process was initially attributed by McIntyre et
al 72 to the potential induced Au(100)-(1x1) to “hex” surface reconstruction. The Au(100)-(1x1)
surface exhibits unique four-fold adsorption sites where stronger binding of molecular O2 and the
reaction intermediates could lead to 4e¯ reduction, whereas the (1x1)-“hex” reconstruction
removes the four-fold sites. However it was later understood that the (1x1)-“hex” transition
could not be the predominant mechanism because the change in the reaction pathway from 4e to
2e is complete within a narrow potential range of ~300mV whereas the potential-induced
structural transition is not complete even after the potential was extended into the negative limit
of -1.1 V. 73 It was also observed that the presence of specifically adsorbed hydroxyl species on
the gold surface was critical to the 4e¯ pathway. On the Au(111) surface, coverage due to OH¯
was less and the ORR proceeded only to peroxide anion. However on the Au(100) crystal face
higher OHads coverage along with significant charge transfer from OH¯ads species to the
electrode was observed. It was also found that the adsorbed hydroxyl species was more
discharged on the Au(100)-(1x1) face with four-fold symmetry compared to the Au(110) and
Au(111). The 4e pathway on Au(100) was observed only within the potential range where
adsorbed hydroxyl species were observed. The ORR pathway reverts back to 2e route in the
diffusion controlled region on Au(100) where adsorbed hydroxyl species ceases to exist. In order
to reconcile these differences it was proposed that further reduction of HO2¯ was catalyzed by a
state of adsorbed hydroxyl species (AuOH) unique to this Au(100) surface. 73-75 Specific
66 adsorption of OH anions in alkaline media on the Au(100)-(1x1) surface was observed to
stabilize this crystal face from potential induced “hex” reconstruction.76
One limitation in the above attempts in the literature to understand the higher activity of
the Au(100)-(1x1) surface within the potential region where AuOH species exists is the
assumption that molecular O2 adsorbs prior to reduction via an inner-sphere mechanism.
However, the possibility of promotion of an outer-sphere bridge formed by adsorbed hydroxyl
species has not been considered. Using this modified approach, solvated molecular oxygen
[O2.(H2O)n] species can undergo outer sphere reduction to form O2•¯ or (HO2
•)ads which then
adsorbs on the unique four-fold symmetry sites of the Au(100)-(1x1) face not available with the
other crystal faces. Higher discharge of adsorbed hydroxyl species on Au(100) compared to the
other crystal faces would actually promote the outer sphere reaction because of less repulsion
between the adsorbed hydroxyl species and the negative dipole (oxygen end) of the water dipole
that is solvating molecular O2. The lower hydroxyl species coverage and higher negative
electrosorption valency on Au(111) and Au(110) in alkaline media inhibits the promotion of an
outer sphere mechanism and hence only the 2e¯ reduction pathway is observed. Finally, this
outer sphere mechanism being inhibited in acidic media gives rise primarily to 2e product on all
Au(hkl) surfaces due to weak adsorption of molecular O2 directly on various gold facets.
2.4 Conclusions
Performing ORR on Pt in alkaline media is disadvantageous not only from the
perspective of cost but also from a kinetics point of view due to significant peroxide generation
at typical operating potential of fuel cells. The following reasons are understood to play an
important role during ORR in alkaline media. The presence of adsorbed hydroxyl species on Pt
67 catalyst sites during ORR not only inhibits direct molecular adsorption of O2 but also promotes a
2e reduction of O2 to HO2¯ by an outer-sphere electron transfer mechanism. This outer-sphere
process is made possible by the low but finite free energy associated with the non-covalent
interaction between the solvated molecular O2 cluster and the specifically adsorbed surface
hydroxyl groups possibly via a hydrogen bond formation. The formation of oxide species on the
metal surface and its interaction with the solvated molecular O2 causes a certain non-specificity
to the underlying metal surface. This non-specific outer-sphere mechanism is understood to be
the reason for the possibility of using a wide range of non-noble metals and their oxides as
electrode materials for oxygen reduction in alkaline media. More quantitative details such as the
accurate picture of this non-covalent interaction and the adsorption energy associated with it
needs to investigated in the future. However this outer-sphere mechanism in most cases, with the
possible exception of Au(100) crystal face, is observed to promote only the two electron
reduction process. In order to increase the faradaic efficiency of the oxygen reduction reaction it
is important to promote the inner-sphere electrocatalytic mechanism by adsorbing the reaction
intermediates on specific catalyst surfaces.
2.5 Acknowledgements:
The authors deeply appreciate financial assistance from the Army Research Office under the
Single Investigator grant. The authors also gratefully acknowledge the supply of electrocatalysts
from BASF fuel cells (Somerset, NJ, USA)
68 2.6 References:
(1) Bockris, J. O.; Appleby, J. In Assessment of Research Needs for Advanced Fuel Cells.; Penner, S. S., Ed. 1986; Vol. 11, p 95. (2) Appleby, A. J. Catal. Rev. 1970, 4, 221. (3) Varcoe, J. R.; Slade, R. C. T. Fuel Cells 2005, 5, 187. (4) Varcoe, J. R.; Slade, R. C. T.; Yee, E. L. H. Chemical Communications 2006, 1428. (5) Slade, R. C. T.; Varcoe, J. R. Solid State Ionics 2005, 176, 585. (6) Spendelow, J. S.; Wieckowski, A. Phys. Chem. Chem. Phys. 2007, 9, 2654. (7) Varcoe, J. R. Physical Chemistry Chemical Physics 2007, 9, 1479. (8) Piana, M.; Boccia, M.; Filpi, A.; Flammia, E.; Miller, H. A.; Orsini, M.; Salusti, F.; Santiccioli, S.; Ciardelli, F.; Pucci, A. J. Power Sources 2010, 195, 5875. (9) Penner, S. S. Assessment of Research Needs for Advanced Fuel Cells. [In: Energy (Oxford), 1986; 11(1-2)], 1986. (10) McLean, G. F.; Niet, T.; Prince-Richard, S.; Djilali, N. Int. J. Hydrogen Energy 2002, 27, 507. (11) Yeager, E. Electrochim. Acta 1984, 29, 1527. (12) Wang, B. J. Power Sources 2005, 152, 1. (13) Guelzow, E. J. Power Sources 1996, 61, 99. (14) Kordesch, K.; Hacker, V.; Gsellmann, J.; Cifrain, M.; Faleschini, G.; Enzinger, P.; Fankhauser, R.; Ortner, M.; Muhr, M.; Aronson, R. R. J. Power Sources 2000, 86, 162. (15) Adzic, R. In Electrocatalysis; Lipkowski, J., Ross, P. N., Eds.; Wiley-VCH: 1998, p 197. (16) Markovic, N. M.; Ross, P. N. Surface Science Reports 2002, 45, 117. (17) Mukerjee, S. Catalysis and Electrocatalysis at Nanoparticle Surfaces 2003, 501. (18) Sepa, D. B.; Vojnovic, M. V.; Damjanovic, A. Electrochim. Acta 1980, 25, 1491. (19) Sepa, D. B.; Vojnovic, M. V.; Vracar, L. M.; Damjanovic, A. Electrochimica Acta 1987, 32, 129. (20) Damjanovic, A.; Dey, A.; Bockris, J. O. M. Electrochim. Acta 1966, 11, 791. (21) Damjanovic, A.; Bockris, J. O. M. Electrochimica Acta 1966, 11, 376. (22) Bockris, J. O. M.; Huq, A. K. M. S. Proc. Roy. Soc. (London) 1956, A237, 277. (23) Reddy, A. K. N.; Genshaw, M. A.; Bockris, J. O. M. Journal of Chemical Physics 1968, 48, 671. (24) Tarasevich, M. R.; Sadkowski, A.; Yeager, E. In Comprehensive Treatise of Electrochemistry; Conway, B. E., Bockris, J. O. M., Yeager, E., Eds.; Plenum Press: New York, 1983; Vol. 7, p 301. (25) Gottesfeld, S. In Fuel Cell Catalysis A Surface Science Approach; Koper, M. T. M., Ed.; John Wiley & Sons, Inc.: 2009. (26) Xu, Y.; Shao, M.; Mavrikakis, M.; Adzic, R. R. In Fuel Cell Catalysis A Surface Science Approach; Koper, M. T. M., Ed.; John Wiley & Sons, Inc.: 2009. (27) Lima, F. H. B.; Zhang, J.; Shao, M. H.; Sasaki, K.; Vukmirovic, M. B.; Ticianelli, E. A.; Adzic, R. R. J. Phys. Chem. C 2007, 111, 404. (28) Mukerjee, S.; Srinivasan, S.; Soriaga, M. P. J. Electrochem. Soc. 1995, 142, 1409.
69 (29) Lima, F. H. B.; Salgado, J. R. C.; Gonzalez, E. R.; Ticianelli, E. A. J. Electrochem. Soc. 2007, 154, A369. (30) Lima, F. H. B.; Sanches, C. D.; Ticianelli, E. A. J. Electrochem. Soc. 2005, 152, A1466. (31) Murthi, V. S.; Urian, R. C.; Mukerjee, S. Journal of Physical Chemistry B 2004, 108, 11011. (32) Yeager, E.; Razaq, M.; Gervasio, D.; Razaq, A.; Tryk, D. Proc. - Electrochem. Soc. 1992, 92-11, 440. (33) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications; Second ed.; John Wiley & Sons, Inc. (34) Damjanovic, A.; Genshaw, M. A.; Bockris, J. O. M. Journal of Chemical Physics 1966, 45, 4057. (35) Wroblowa, H. S.; Pan, Y.-C.; Razumney, G. J. Electroanal. Chem. Interfacial Electrochem. 1976, 69, 195. (36) Anastasijevic, N. A.; Vesovic, V.; Adzic, R. R. J. Electroanal. Chem. Interfacial Electrochem. 1987, 229, 305. (37) Anastasijevic, N. A.; Vesovic, V.; Adzic, R. R. J. Electroanal. Chem. Interfacial Electrochem. 1987, 229, 317. (38) Appleby, A. J. Compr. Treatise Electrochem. 1983, 7, 173. (39) Bard, A. J. J. Am. Chem. Soc. 2010, 132, 7559. (40) Bard, A. J.; Parsons, R.; Jordan, J.; Editors Standard Potentials in Aqueous Solution, 1985. (41) Yang, H.-H.; McCreery, R. L. J. Electrochem. Soc. 2000, 147, 3420. (42) Blizanac, B. B.; Ross, P. N.; Markovic, N. M. Electrochimica Acta 2007, 52, 2264. (43) Zhang, C.; Fan, F.-R. F.; Bard, A. J. J. Am. Chem. Soc. 2009, 131, 177. (44) Murthi, V. S.; Urian, R. C.; Mukerjee, S. J. Phys. Chem. B 2004, 108, 11011. (45) Paulus, U. A.; Wokaun, A.; Scherer, G. G.; Schmidt, T. J.; Stamenkovic, V.; Markovic, N. M.; Ross, P. N. Electrochim. Acta 2002, 47, 3787. (46) Paulus, U. A.; Wokaun, A.; Scherer, G. G.; Schmidt, T. J.; Stamenkovic, V.; Radmilovic, V.; Markovic, N. M.; Ross, P. N. Journal of Physical Chemistry B 2002, 106, 4181. (47) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Applied Catalysis, B: Environmental 2005, 56, 9. (48) Mukerjee, S.; Srinivasan, S.; Soriaga, M. P. Journal of the Electrochemical Society 1995, 142, 1409. (49) Sheng, W.; Gasteiger, H. A.; Shao-Horn, Y. J. Electrochem. Soc. 2010, 157, B1529. (50) Markovic, N. M.; Gasteiger, H. A.; Ross, P. N., Jr. J. Phys. Chem. 1996, 100, 6715. (51) Markovic, N.; Gasteiger, H.; Ross, P. N. J. Electrochem. Soc. 1997, 144, 1591. (52) Janin, E.; von Schenck, H.; Gothelid, M.; Karlsson, U. O.; Svensson, M. Phys. Rev. B Condens. Matter Mater. Phys. 2000, 61, 13144. (53) Jerkiewicz, G.; Vatankhah, G.; Lessard, J.; Soriaga, M. P.; Park, Y.-S. Electrochimica Acta 2004, 49, 1451. (54) Chen, P.; McCreery, R. L. Analytical Chemistry 1996, 68, 3958.
70 (55) Chen, P.; Fryling, M. A.; McCreery, R. L. Analytical Chemistry 1995, 67, 3115. (56) Wakabayashi, N.; Kitamura, F.; Ohsaka, T.; Tokuda, K. Journal of Electroanalytical Chemistry 2001, 499, 161. (57) Wandlowski, T.; de Levie, R. Journal of Electroanalytical Chemistry 1995, 380, 201. (58) Conway, B. E. Progress in Surface Science 1995, 49, 331. (59) Angerstein-Kozlowska, H.; Conway, B. E.; Sharp, W. B. A. J. Electroanal. Chem. Interfacial Electrochem. 1973, 43, 9. (60) Strmcnik, D.; Kodama, K.; van der Vliet, D.; Greeley, J.; Stamenkovic, V. R.; Markovic, N. M. Nature Chemistry 2009, 1, 466. (61) Sarapuu, A.; Vaik, K.; Schiffrin, D. J.; Tammeveski, K. J. Electroanal. Chem. 2003, 541, 23. (62) Tammeveski, K.; Kontturi, K.; Nichols, R. J.; Potter, R. J.; Schiffrin, D. J. J. Electroanal. Chem. 2001, 515, 101. (63) Vaik, K.; Sarapuu, A.; Tammeveski, K.; Mirkhalaf, F.; Schiffrin, D. J. J. Electroanal. Chem. 2004, 564, 159. (64) Iwasita, T.; Xia, X. Journal of Electroanalytical Chemistry 1996, 411, 95. (65) Bockris, J. O. M.; Habib, M. A. Electrochimica Acta 1977, 22, 41. (66) Taylor, E. J.; Vilambi, N. R. K.; Gelb, A. Journal of the Electrochemical Society 1989, 136, 1939. (67) Adzic, R. R.; Markovic, N. M.; Vesovic, V. B. J. Electroanal. Chem. Interfacial Electrochem. 1984, 165, 105. (68) Markovic, N. M.; Adzic, R. R.; Vesovic, V. B. J. Electroanal. Chem. Interfacial Electrochem. 1984, 165, 121. (69) Adzic, R. R.; Markovic, N. M. J. Electroanal. Chem. Interfacial Electrochem. 1982, 138, 443. (70) Adzic, R. R.; Strbac, S.; Anastasijevic, N. Mater. Chem. Phys. 1989, 22, 349. (71) Strbac, S.; Adzic, R. Journal of the Serbian Chemical Society 1992, 57, 835. (72) McIntyre, J. D. E.; Peck, W. F., Jr. Proc. - Electrochem. Soc. 1984, 84-12, 102. (73) Markovic, N. M.; Tidswell, I. M.; Ross, P. N. Langmuir 1994, 10, 1. (74) Strbac, S.; Adzic, R. R. Electrochim. Acta 1996, 41, 2903. (75) Blizanac, B. B.; Lucas, C. A.; Gallagher, M. E.; Arenz, M.; Ross, P. N.; Markovic, N. M. J. Phys. Chem. B 2004, 108, 625. (76) Tidswell, I. M.; Markovic, N. M.; Lucas, C. A.; Ross, P. N. Physical Review B: Condensed Matter and Materials Physics 1993, 47, 16542.
71
Chapter 3
Electrochemical Kinetics and X-ray Absorption Spectroscopic Investigations of Oxygen
Reduction on Chalcogen Modified Ruthenium Catalysts in Alkaline Media
3.1 Introduction
Development of non-noble metal electrocatalysts, reduction in overpotential and loading
of the Pt based catalysts for Oxygen Reduction Reaction (ORR) has significant fundamental and
technological implications. Acidic electrolytes narrow down the choice of catalyst materials due
to stringent requirements based on the ‘acid-stability’ criterion. Alkaline electrolyte based fuel
cells provide the much required window of opportunity to develop non-noble metal based
catalysts. Alkaline Fuel Cells (AFC), once a very promising renewable energy source failed to
attract continued research interest primarily due to issues such as effect of carbonate
precipitation, and electrolyte leakage.1 Recent efforts in developing alkaline anion exchange
membranes (AAEM) for OH anion conductivity brings promise to this technology of AFC.2-3
ORR pathway on Pt based catalysts in both acid and alkaline media is understood to be
similar based on electrochemical kinetic studies on polycrystalline and single crystal electrodes.3-
4 The First e transfer step to the adsorbed molecular O2 is widely agreed to be the rate
determining step in the overall 4e¯ multi-step process. Parallel routes consisting of
predominantly the 4e¯ transfer either via a “direct” or “series” pathway and a minor 2e pathway
leading to peroxide is typically observed based on Rotating Ring-Disk Electrode (RRDE)
studies. In Chapter 2 it was shown that on metal surfaces such as Pt, certain important changes in
ORR mechanism and double-layer structure take place in alkaline media. These mechanistic
changes involve the possibility of outer-sphere e¯ transfer at oxide covered metal sites at the
72 higher pH. O2 reduction by one-electron transfer to form superoxide (O2
•¯) is observed at E° = -
0.3±0.03 V vs. SHE corresponding to ∆G°=30±2kJ mol-1 with both O2 and O2•¯ remaining in the
aqueous phase.5-6 Given the pH independence of this redox couple (O2/O2•¯), the potential of this
reaction does not change as the pH is varied from zero to fourteen.7 However, due to the
occurrence of 4H+ transfer steps in ORR, its standard reduction potential changes by 0.828 V
from 1.229 V to 0.401 V vs. SHE as the pH value changes from zero to fourteen. This causes the
overpotential for the first electron transfer step (O2/O2•¯) to decrease from 1.53 V at pH=0 to 0.7
V at pH=14 indicating a sharp decrease at alkaline pH conditions.7 It was also shown in Chapter
2 that in alkaline media the specifically adsorbed hydroxyl species (OHads) on the catalyst surface
mediates an outer-sphere electron transfer to solvated molecular O2 species to form superoxide
radical anion. This superoxide undergoes adsorption on the oxide sub-structure of the catalyst
surface, followed by proton transfer to form adsorbed hydrperoxyl species (HO2•)ads. This is
followed by electron transfer to form HO2¯ anion. This mechanistic change implies that the
mediation by the surface adsorbed hydroxyl species causes certain non-specificity to the identity
of the underlying electrode metal surface. Hence this non-specificity characteristic of all outer-
sphere electron transfer reactions was found to be the rationale behind the possibility of using
any electronically conducting electrode material for ORR in alkaline media.
In alkaline media, apart from Pt based materials a gamut of ORR electrocatalyst materials
have been tested ranging from various carbons, 8-9 first row transition metals, 10-11 metal oxides,
12-14 metal macrocycles, 9,15-17 perovskites, 18-19 pyrochlores, 20 and spinels.21 Among this wide
variety of catalysts materials in alkaline media, only the transition metal macrocycles and
transition metal oxides seem to be thoroughly investigated for ORR. Although these metal
macrocycles and metal oxides, in particular MnOx are inexpensive alternatives, the amount of
73 peroxide generated leading to loss in efficiency of the 4e- process and persistent stability issues
hinder the translation of these materials into practical electrodes. For ORR in acid media, a class
of alternative catalyst materials studied over the last two decades with promising ORR activity
and fuel crossover tolerance is the transition metal based chalcogenides.22
Initially reported for ORR in their semi-conducting crystalline Chevrel phases (RuxMo6-
xSe8) synthesized via high temperature and high pressure routes, these materials exhibited ORR
activity due to the transition metal clusters acting as reservoirs of electronic charge although the
electrochemical functionality of Se was relatively unknown.23-26 Further developments led to the
synthesis of amorphous Ru-based transition metal chalcogenides supported on high surface area
carbon synthesized via low temperature non-aqueous routes.27 It is now well known that these
Ru based chalcogenides Ru-X (X = S, Se, Te) (i) consist of a metallic core of Ru cluster with the
chalcogen coordinated to the periphery of the cluster,28 (ii) such a Ru-X coordination provides
sufficient “coexistence” of Ru and Se on the surface exposing molecular O2 to the Ru sites, since
Ru is the active site for ORR and the chalcogen primarily functions to preserve this Ru core in a
metallic state by preventing oxide formation in aqueous electrolytes,29 (iii) the Se derivatives
exhibit higher ORR activity compared to their S brethren, (iv) selenium exists in a metallic state
with a certain charge transfer from Ru to Se that also renders Ru less susceptible to oxide
formation 30 (v) irrespective of the chalcogen, all these RuxXy derivatives are inherently tolerant
to crossover methanol,31-32 (vi) also, these chalcogenide nanoparticles have been reported to be
Se-decorated Ru catalysts as opposed to stoichiometric RuxSey binary catalysts,33 (vii) although
less addressed, the stability of these catalysts due to anodic dissolution of its individual
components is a major limiting factor 26 and (viii) transition metal additives such as Mo improve
the ORR activity of these chalcogenides. Although this enhancing effect of transition metal
74 additive has been ascribed to the oxygen binding properties of Mo, there is no fundamental
insight into this effect in the literature. 32,34-35 Pt and Pt-based bimetallic systems have been well
understood using single crystal studies for elucidating the ORR process,36 whereas RuxSy and
RuxSey do not lend themselves to single crystal studies. Technological advancements have led to
the use of various spectroscopic methods to study the nature and chemistry of adsorption as it
pertains to ORR under in situ electrochemical conditions.37-38
In this chapter, we have systematically investigated carbon supported Ru based
chalcogenides (Ru, S/Ru, S/RuMo, Se/Ru, and Se/RuMo) synthesized via aqueous routes for
ORR in alkaline media. A two-pronged strategy of using electrochemical characterizations and
in situ X-ray Absorption Spectroscopy (XAS) has been employed to elucidate the ORR
pathway/mechanism and its dependence on structural/electronic aspects of these catalysts. An
extensive in situ (XAS) investigation was conducted at Ru, Se, and Mo K-edges in order to
understand the short-range atomic order of this class of material. The element specificity of XAS
studies imparts the ability to investigate the individual elemental components of these binary and
ternary chalcogenide systems and their behavior under electrochemical conditions. These
structural characteristics are interpreted in conjunction with standard electrochemical kinetic
studies. The objective of this research effort is to understand i) the reasons for kinetic facility
during ORR in alkaline media, ii) elucidate the ORR pathway/mechanism of chalcogen modified
transition metal catalysts in alkaline media as distinct from acidic conditions, iii) the influence of
chalcogen (S, Se) in modifying the transition metal surface, and the molybdenum additive in
improving the ORR activity and stability. To the best of the authors’ knowledge this is the first
detailed report of this class of materials in alkaline media as it pertains to ORR.
75 3.2 Experimental:
3.2.1 Catalyst Synthesis
All ruthenium based chalcogen modified catalysts discussed in this thesis were
synthesized under aqueous conditions according to the references.39-40 Vulcan XC72R carbon
was used as the catalyst support material. All ruthenium based catalysts had a nominal metal
loading of 20% by weight of ruthenium on the carbon support. Selenium modified ruthenium
catalysts were synthesized according to procedure reported by Campbell et al.40 Details of the
synthesis procedure are available in the original patent literature. Briefly Se/Ru/C was
synthesized by dissolving 1.12 g (4.45 mmol Ru) of RuCl3, 0.247 g (2.225 mmol) of SeO2 in 100
ml of Millipore water along with 1.05 g of Vulcan Carbon (Cabot Corporation). The reaction
mixture was dispersed by sonication followed by heating to 80°C for 30 minutes under stirring.
After cooling the reaction solution to room temperature, 0.673 g of NaBH4 (corresponding to 4
times the moles of Ru) in 0.1 M NaOH was added to reduce the metal ions. The reaction mixture
was re-heated at 80°C under stirring for an hour to complete the reduction process followed
sequentially by cooling to room temperature, vacuum filtration, washing with copious amounts
of de-ionized (Millipore) water and drying at 60°C overnight under vacuum. Heat treatment of
the catalyst was typically performed at 500°C for 2 hours under an argon atmosphere. To
synthesize Ru/C no SeO2 precursor was utilized. To synthesize Se/Ru-Mo/C 169 mg (0.89 mmol
Mo) of phosphomolybdic acid was added along with the precursors for ruthenium and selenium.
Synthesis of sulfur modified ruthenium based catalysts involved the procedure of Allen et
al.39 S/Ru/C was synthesized using ammonium tetrathiosulfate as the precursor for sulfur which
involves in a direct metathesis reaction with the transition metal and the reaction proceeds
without the presence of free sulfide ion in the aqueous solution. S/RuMo/C catalyst was
76 synthesized using ammonium tetrathiomolybdate as the precursor for both sulfur and
molybdenum. Heat treatment of the sulfur modified catalysts were also performed at 500°C for 2
hours under an argon atmosphere. 30% Pt/C catalyst procured from BASF fuel cells (Somerset
NJ, USA) was used for comparative purposes.
3.2.2 Physicochemical characterization:
X-ray diffraction (XRD, Rigaku-model D/MAX-2200T) was used to characterize the
crystal structure and particle size of the catalysts. The measurements were made at 46 kV and 40
mA fitted with Cu Kα = 1.5406 Å. The diffraction patterns were recorded with a scan rate of
0.5°/min between 10° and 100°. The analysis of the XRD data was carried out using the MDI
Jade 5.0 package. Energy dispersive analysis of X-rays (EDAX) was carried out using an EDS-
GENESIS HITACHI S-4800 equipped with a Cu filter and a liquid nitrogen cooled Si(Li)
detector and it was used for element speciation of the catalysts at an acceleration voltage of 25
kV.
3.2.3 Electrochemical characterization
All electrochemical measurements were made at room temperature using a rotating ring-
disk electrode (RRDE) setup from Pine Instruments connected to an Autolab (Ecochemie Inc.,
model-PGSTAT 30) bi-potentiostat. Alkaline (0.1 M NaOH) and acidic (0.1 M HClO4)
electrolytes were prepared using sodium hydroxide pellets (semiconductor grade, 99.99%,
Sigma-Aldrich) and double-distilled 70% perchloric acid (GFS Chemicals), respectively.
Catalyst inks were typically prepared by dispersing 25mg of the catalyst in 10 ml of 1:1
Millipore H2O:isopropyl alcohol mixture along with 100 µL of 5 wt% Nafion(R) solution as a
binder. 5 µL aliquot of the catalyst ink was dispensed on a Glassy Carbon (GC) disk of 5.61mm
dia. Gold ring electrode was held at 1.1 V vs. RHE in alkaline electrolyte and at 1.3 V vs. RHE
77 in acidic electrolyte to detect stable peroxide intermediate. Collection efficiency of the disk-ring
electrode setup was 37.5%. All potentials are refered to a reversible hydrogen electrode (RHE)
scale made out of the same solution as the bulk electrolyte unless otherwise stated.
3.2.4 X-ray Absorption Spectroscopic (XAS) Measurements
The in situ XAS studies at Ru (K edge – 22117 eV) , Se (K edge – 12658 eV), Mo (K
edge – 20000 eV) binding energies were performed at the National Synchrotron Light Source
(NSLS, Brookhaven National Laboratories, NY) beamline X-11A. Detailed information on the
spectro-electrochemical cell design are given elsewhere.41 The total geometric loading of Ru
metal on the electrode was chosen to give a transmission absorption height of unity. All data at
Ru K-edge was collected in transmission mode and the Se and Mo K-edge data were collected in
fluorescence mode using a PIPS detector. Data were collected using the typical three gas
ionization detector (I0, It and Iref) setup for 10% photon absorption in I0 and 50-70% in It and Iref.
Argon or oxygen saturated 0.1 M NaOH was used as the electrolyte along with a RHE made out
of the same electrolyte as the reference electrode. Details of EXAFS analysis are available in an
earlier publication.41 Briefly, the IFEFFIT suite Version 1.2.9 42 was used for background
subtraction using AUTOBK algorithm and normalization. The typical k-range window during
EXAFS fit was 2.500-13.500 Å-1 (Kaiser-Bessel).
3.3 Results and Discussions
3.3.1 Physicochemical Characterizations (XRD/EDS)
The elemental speciation of the in-house synthesized catalysts were carried out using
EDS and the results are shown in Table 3.1. All chalcogen modified catalysts had a nominal
ruthenium metal loading of 20% by weight on carbon support. Ruthenium to chalcogen (S/Se)
78
Table 3.1: Physicochemical characterization of binary and ternary heat treated chalcogenide catalysts using EDS and XRD
Catalysta Stoichiometry (Atomic ratio from
EDS) (Ru:S/Se:Mo)
XRD Crystallite Sizeb (nm)
Ru/C -- 13
Se/Ru/C 1:0.6:0 7
Se/RuMo/C 1:0.6:0.08 9
S/Ru/C 1:0.48:0 7
S/RuMo/C 1:0.6:0.36 7 a All catalysts were synthesized with a nominal Ru metal loading of 20% by weight on carbon b Crystallite size from Debye-Scherrer of Ru(101) diffraction peak
atomic ratio was typically observed to be 1:0.5±0.15. Atomic content of molybdenum in the
Se/RuMo/C catalyst, where phosphomolybdic acid was used as the Mo precursor, was found to
be always low with a ratio of typically Ru:Mo of 1:0.1±0.02. It is also noted that similarly low
Mo content in the catalyst was also reported earlier using the same precursor.43 Higher Mo
content (Ru:Mo = 1:0.36) was observed in the case of S/RuMo/C catalyst where
tetrathioammonium molybdate was used as the Mo precursor. Representative XRD profiles of
carbon supported Ru/C and Se/Ru/C recorded from 2θ values of 10° to 90° as shown in Figure
3.1 and the Debye-Scherrer crystallite size from the main Ru(101) diffration peak is shown in
Table 3.1. Ruthenium in Ru/C-Heat Treated (HT) catalyst exists in its most commonly found
hexagonal close packed (hcp) structure (Space group: P63/mmc) with a lattice constant of
a=b=2.7106, c=4.2911 in good agreement with JCPDS powder diffraction patterns. All crystal
planes contributing to the hcp structure expected within the measured 2θ range are indexed in
Figure 3.1(a) for the Ru/C-HT catalyst. No diffraction peaks due to oxides of ruthenium were
79 observed since the catalyst was always stored under inert ambient conditions and the XRD
patterns were measured immediately after synthesis. Ruthenium particle size was calculated to
be 13 nm based on Ru(101) crystallite line broadening. As shown in Figure 3.1(a), prior to heat
treatment the crystal planes of Ru/C-Non-Heat Treated (NHT) catalyst were poorly developed
and not observed in its XRD pattern attesting to its amorphous nature.
2θ [Degrees]20 40 60 80
Inte
ntis
y [a
.u]
(a) Ru/C - P63/mmc
Ru(100)
Ru(002)
Ru(101)
Ru(102)Ru(110)Ru(103)
Ru(112)
Ru(201)
2θ [Degrees]20 40 60 80
Inte
nsity
[a.u
]
Before Heat Treatment
After Heat Treatment
(b) Se/Ru/C - After Heat Treatment
Figure 3.1: XRD patterns of (a) Ru/C before and after heat treatment, (b) Se/Ru/C. XRD patterns of Se/RuMo/C, S/Ru/C, S/RuMo/C were similar to Se/Ru/C.
80
Figure 3.1(b) shows the XRD pattern for Se/Ru/C-HT catalyst. All primary diffraction
peaks corresponding to hcp Ru structure (P63/mmc space group) are also observed in Se/Ru/C
catalyst. A minor diffraction peak due to RuO2 at 2θ value of ~54° was observed.43 Lattice
constants of Se/Ru/C-HT catalyst were found to be a=b=2.7080, c=4.2808, indicating minor
lattice contractions relative to Ru/C-HT. Particle sizes based on line broadening of Ru(101)
crystallite size was found to 7 nm to 9 nm for the chalcogen modified catalysts, indicating that
the chalcogen prevents sintering of the underlying transition metal nanoparticle during the heat
treatment step. A stable compound of ruthenium and selenium is ruthenium diselenide (RuSe2)
which exhibits a pyrite structure with a cubic lattice (space group: Pa-3). However, no signs of
RuSe2 compound were observed in XRD. The XRD pattern of Se/Ru/C-HT catalyst clearly
shows that the underlying Ru lattice has undergone little change even after modifying with ~50%
atomic content of selenium. Similar XRD patterns were obtained for the other chalcogen
modified catalysts in this study and hence are not discussed any further to avoid redundancy.
3.3.2 Electrochemical Characterizations
3.3.2.1 Cyclic Voltammetry - Electrochemical surface characterestics
Figure 3.2(a) shows the cyclic voltammetry (CV) of Ru/C catalyst in de-oxygenated 0.1
M NaOH electrolyte before and after heat treatment (HT). Before HT, Ru/C catalyst exhibits an
anodic hydrogen desorption peak centered at 0.1 V vs. RHE. After HT, in concordance with the
evolution of Ru crystal planes as observed by XRD, a more well resolved hydrogen desorption
peak with a shoulder at ~0.18 V is observed. The double layer capacitance decreases after HT,
possibly indicating the removal of surface oxides during heat treatment. Closer inspection
revealed that Ru-OH formation commences on the Ru surface even at potentials as low as 0.3-
0.4 V (vs. RHE) immediately after desorption of hydrogen from the surface, in agreement with
81
the literature. 44-45 Reduction of ruthenium oxides in the cathodic scan shows a broad wave up to
0.4 V. In Figure 3.2(b), CV for Se/C shows a redox couple with anodic and cathodic peak
potentials of Epa = 0.448 V and Epc = 0.301 V vs. RHE, respectively. Based on the
thermodynamic data this redox couple is assigned to the oxidation of Se0 to Se4+ written below as
a standard reduction reaction:5
SeO32¯ + 4e + 3H2O → Se + 6OH E0 = -0.357 V vs. SHE (0.41 V vs. RHE) (3.1)
(a) Ru/C
Potential [V Vs RHE]
0.0 0.2 0.4 0.6 0.8 1.0
Cur
rent
Dens
ity [A
/cm
2 geo]
-6e-4
-4e-4
-2e-4
0
2e-4
4e-4
6e-4
Before HTAfter HT
(b) Se/Ru/C
Potential [V Vs RHE]
0.0 0.2 0.4 0.6 0.8 1.0
Cur
rent
Dens
ity [A
/cm
2 geo]
-4e-4
-2e-4
0
2e-4
4e-4
Before HTAfter HTSe/C
0.1M NaOH20 mV/s
Figure 3.2: Cyclic voltammetry in de-oxygenated 0.1 M NaOH electrolyte at 20 mV/s of (a) Ru/C and (b) Se/Ru/C catalysts before (solid line) and after (dashed) heat treatment (HT) at 500°C in inert atmosphere. Also shown in (b) is the cyclic voltammetry of 15% Se/C under similar conditions. Cyclic voltammetry of Se/RuMo/C, S/Ru/C, and S/RuMo/C were similar to that of Se/Ru/C both before and after heat treatments.
82
Modifying ruthenium surface with selenium shows interesting behavior as seen in Figure
3.2(b). No hydrogen adsorption/desorption current is observed on the Se/Ru/C catalyst similar to
the observations made in acidic electrolytes.33 Prior to HT of Se/Ru/C, a broad oxidation peak
between 0.3 V and 0.7 V in the double-layer region followed by a sharp increase in current at 0.8
V vs. RHE is observed. The broad oxidation peak between 0.3 V and 0.7 V is due to combination
of both Ru-OH formation and the oxidation of elemental selenium that is present in Se/Ru/C
catalyst. Interaction of selenium with ruthenium stabilizes selenium from oxidation and shifts the
Se0/Se4+ oxidation potential to more positive potential which is observed as a sharp rise in
current above ~0.8 V. However, this oxidation current of Se/Se4+ above 0.8 V is not
compensated by a concomitant cathodic reduction current, clearly indicating either a dissolution
of selenium oxide or a passivation effect due to oxide formation. However, HT of Se/Ru/C
catalyst removes any elemental selenium and improves the alloying between selenium and
ruthenium, thereby stabilizing both Se and Ru from oxidation. Oxidation currents due to Ru-OH
formation and Se/Se4+ reaction are significantly muted in Se/Ru/C catalyst after HT.
3.3.2.2 ORR measurements on Ruthenium based Catalysts:
Figure 3.3(a&b) shows the ORR on Ru/C-HT catalyst in O2 saturated 0.1 M NaOH and
0.1 M HClO4 electrolytes at 900 rpm. The onset potential for Ru/C-HT in 0.1 M HClO4 is 0.78 V
vs. RHE whereas in 0.1 M NaOH is 0.9 V vs. RHE. In 0.1 M NaOH electrolyte the mixed
kinetic-diffusion region between 0.9 V and 0.5 V is ensued by a well defined limiting current
region below 0.5 V Vs. RHE. By contrast, in 0.1 M HClO4 electrolyte ORR on Ru/C-HT is
kinetically controlled even at very high overpotentials and no clear diffusion limited current
region could be identified. At a kinetic current density of 0.1 mAcm-2geo Ru/C-HT exhibits ~125
83
mV lower overpotential in 0.1 M NaOH than in 0.1 M HClO4. Primary reasons for this lower
overpotential in alkaline medium are delineated here.
In Chapter 2, it was shown that the specifically adsorbed hydroxyl species (OHads) on the
electrode surface mediate an outer-sphere electron transfer reaction. This mediation process
(a) PolarizationCurves
Potential [V Vs RHE]
0.0 0.2 0.4 0.6 0.8 1.0
i D [A
/cm
2 geo]
-5e-3
-4e-3
-3e-3
-2e-3
-1e-3
0
0.1M NaOH0.1M HClO4
(c) Ring Currents
Potential [V Vs RHE]
0.0 0.2 0.4 0.6 0.8 1.0
I R [A
]
0
2e-6
4e-6
6e-6
8e-6
1e-5
1e-5
1e-5
2e-50.1M NaOH0.1M HClO4
(b) Tafel Plots
log ik [A/cm2
geo]
1e-6 1e-5 1e-4 1e-3
E V
s R
HE
0.7
0.8
0.9
125 mV
E [V Vs RHE]
0.2 0.4 0.6 0.8 1.0
i D [A
/cm
2 geo]
-5e-3
-4e-3
-3e-3
-2e-3
-1e-3
0
1e-3ORRHRR
900 rpm20 mV/s
0.1M HClO4
0.1M NaOH
(d) H2O2 Reduction
Figure 3.3: ORR on Ru/C Heat Treated (HT) catalyst in 0.1 M NaOH and 0.1 M HClO4 electrolytes. (a) ORR Polarization Curves at 900 rpm and 20 mV/s, (b) Tafel plots, (c) Ring current measured during ORR at 900 rpm. Ering = 1.1 V vs. RHE in 0.1 M NaOH and Ering = 1.3 V vs. RHE in 0.1 M HClO4 and (d) H2O2 reduction reaction in comparison to ORR in 0.1 M HClO4 and 0.1 M NaOH. H2O2 reduction is shown in oxygen-free electrolytes containing externally added HO at a concentration of 3.5 mM.
84 involves electron transfer from the electrode to a solvated molecular O2 species across an outer-
sphere bridge constituted by the surface adsorbed hydroxyl species. Hydrogen bond formation
between the OHads species and the solvated molecular O2 cluster (i.e. O2.(H2O)n) localizes the
solvated molecular O2 cluster in the outer-Helmhlotz plane (OHP). This outer-sphere reaction is
primarily a 2e reduction of O2 to HO2¯ as the product. This outer-sphere reaction is formulated
as shown below where M corresponds to the underlying metal site which in this case is
ruthenium:
M-OH + [O2.(H2O)n]aq + e → M-OH + (HO2•)ads + OH + (H2O)n-1 (3.2)
(HO2•)ads + e → (HO2¯)ads (3.3)
(HO2¯)ads → (HO2¯)aq (3.4)
Equation (3.2) written above involves electron transfer (or tunneling) to solvated O2 from the
electrode surface across a thin oxide film and at least one layer of solvation shell. Equation (3.2)
above involves several elementary steps as written below:
M-OH + [O2.(H2O)n]aq + e → M-OH + [O2•¯.(H2O)n]aq (3.2a)
[O2•¯.(H2O)n]aq → (O2
•¯)ads + nH2O (3.2b)
(O2•¯)ads + H2O → (HO2
•)ads + OH (3.2c)
This (HO2¯)aq formed in step (3.4) above is detected at the ring electrode and appears as a
shoulder in ring current between 0.7 V and 0.8 V as shown in Figure 3.3(c). In fact, such a ring
current profile was used in our earlier report as a characteristic signature of the outer-sphere
electron transfer reaction mechanism on Pt surface. In acidic media peroxide detected is very
minimal and about an order of magnitude lower than that in 0.1 M NaOH. It should be noted that
in 0.1 M NaOH this outer-sphere electron transfer occurs only at oxide covered Ru sites, whereas
at oxide-free Ru sites direct molecular O2 adsorption should take place leading to efficient 4e¯
85 reduction of O2 to OH via an inner-sphere electrocatalytic pathway. So, in alkaline media a
combination of both inner-sphere and outer-sphere electron transfer mechanism is operative. The
consequence of this outer-sphere electron transfer in alkaline media is that this mechanism leads
higher concentration of HO2¯ to be generated near the electrode surface, i.e. the double layer;
however no evidence for such an outer-sphere reaction is observed in acidic media. Higher
activity of HO2¯ effectively shifts the potential of the electrode from that of the O2/HO2¯ couple
to that of the HO2¯/OH¯ redox couple by carrying out HO2¯ reduction to OH on oxide free Ru
sites. This is shown in Figure 3.3(d) where ORR and hydrogen peroxide reduction reaction
(HRR) is shown on Ru/C-HT catalyst in both 0.1 M NaOH and 0.1 M HClO4 at 900 rpm. 2e
reduction reaction of hydrogen peroxide in acidic (H2O2/H2O) and alkaline media (HO2¯/OH¯) is
written below:5
Acidic Medium: H2O2 + 2H+ + 2e → 2H2O E0 = 1.763 V vs. SHE (3.5)
Alkaline Medium: HO2¯ + H2O + 2e → 3OH E0 = 0.867 V vs. SHE (3.6)
As is well known, the standard reduction potentials of the above reactions are well positive of the
4e reduction of molecular O2 in both acidic and alkaline electrolytes. So from a thermodynamic
perspective any peroxide intermediate formed should be immediately reduced further. This is
also kinetically true on ruthenium as shown in Figure 3.3(d). Half-wave potential (E1/2) of HRR
on Ru/C-HT in 0.1 M NaOH is 40 mV positive compared to that of the E1/2 of ORR.44 So the
kinetics of the reaction in alkaline medium favors the immediate reduction of any peroxide
intermediate generated during ORR. This is also true in acidic media although the shape of the
HRR profile in acidic medium on Ru/C-HT requires more explanation. The onset potential of
HRR on Ru/C-HT in 0.1 M HClO4 is as high as 0.88 V which is only 20 mV lower than ORR
and HRR onset potentials in 0.1 M NaOH electrolyte. However, as is known previously46 H2O2
86 undergoes decomposition to O2 and H2O in acidic electrolyte at Ru/C surface. This
decomposition reaction that generates O2 near the electrode surface skews the HRR profile in
acidic medium to higher overpotentials characteristic of ORR in acidic medium. However the
aspect of relevance to the discussion here is that the kinetics of the system favors further
reduction of hydrogen peroxide intermediate in both acidic and alkaline medium. Once the
H2O2/HO2¯ stable intermediate is generated, this species undergoes adsorption at oxide free Ru
sites and further reduces to H2O/OH according to equation (3.5)/(3.6). So, any in situ generation
of hydrogen peroxide intermediate should shift the potential to more positive values and indeed
this is what is observed in alkaline medium. The in situ parallel generation of HO2¯ anion
intermediate via the outer-sphere electron transfer reaction scheme shown above in equations
(3.2)-(3.4) serves to shift the ORR potential to more positive values in alkaline medium Figure
3.3(a). On the contrary, this excess parallel generation of H2O2 via the outer-sphere reaction
mechanism is not observed in acidic medium and hence the higher overpotential for ORR in
acidic medium. As mentioned above, this outer-sphere reduction at oxide covered metal site is
totally independent of the direct molecular O2 reduction on oxide-free metal site. In acidic
medium the surface oxides primarily serve to inhibit the direct adsorption of molecular O2 but in
alkaline medium the surface oxides not only block O2 adsorption but also promote outer-sphere
electron transfer. This is the rationale behind the so-called kinetic facility on Ru/C catalyst in
alkaline medium compared to acidic conditions. In general, this can also be further extended to
other catalyst systems that exhibit lower ORR overpotential in alkaline medium compared to that
in acidic electrolytes.
On Ru/C-HT, as seen in Figure 3.3(c), the ring current in the potential region 0.65 V to
0.3 V is due to 2e reduction of O2 to HO2¯ on the carbon support mediated by the
87 quinone/hydroquinone surface functional groups. Finally a sharp increase in ring current in the
hydrogen adsorption/desorption region is also observed due to change in orientation of water
molecules near the electrode surface as the potential traverses into the H adsorption region.
These two aspects have already been discussed in Chapter 2 and not explained further here.
Potential [V Vs RHE]
0.0 0.2 0.4 0.6 0.8 1.0
Cur
rent
Den
sity
[A/c
m2 ge
o]
-5e-3
-4e-3
-3e-3
-2e-3
-1e-3
0
Ru/C Non-HTRu/C HT500C
(c) Ring Current
Potential [V Vs RHE]
0.0 0.2 0.4 0.6 0.8 1.0
Rin
g C
urre
nt [A
]
2e-6
4e-6
6e-6
8e-6
1e-5
1e-5
1e-5
Ru/C Non-HTRu/C HT500C
O2 Satd.0.1M NaOH
900 rpm20 mV/s
ERing = 1.1V Vs RHE
(b) Tafel Plots
log ik [A/cm2geo]1e-5 1e-4 1e-3
E V
s R
HE
0.8
0.9
(a) Polarization Curves
Figure 3.4: ORR on Ru/C catalyst before and after heat treatment. (a) Anodic ORR Polarization Curves in O2 saturated 0.1 M NaOH electrolyte at 900 rpm and 20 mV/s, (b) Mass transport corrected Tafel plots, and (c) Ring current measured during ORR at 900 rpm. Ering = 1.1 V vs. RHE.
88
Figure 3.4 compares the ORR process on Ru/C before and after the heat treatment
process measured in 0.1 M NaOH electrolyte at 900 rpm. In brief, as seen in Figure 3.4 (a&b)
both the ORR polarization curves and the Tafel plots show that the kinetics of ORR is not
affected by the heat treatment process. However as seen in Figure 3.4 (c) the shoulder in the ring
current between the potential ranges of 0.65 V to 0.8 V appears as a well resolved peak after heat
treatment of the Ru/C catalyst. This is in excellent agreement with the XRD results shown above
wherein after heat treatment process the crystal planes are found to be very well resolved.
Evolution of the crystal facets after heat treatment enhances the specific adsorption of hydroxide
anions. Subsequently it is this specifically adsorbed hydroxyl (OHads) species that
promote/mediate the 2e¯ outer-sphere electron transfer reaction mechanism of O2 to HO2¯
according to the reaction mechanism shown in equations (3.2)-(3.4). This result is another
testimony to the outer-sphere electron transfer hypothesis in alkaline medium. Another
interesting aspect from Figure 3.4(a&b) is that the kinetics of the 4e inner-sphere
electrocatalytic reduction of O2 at oxide-free Ru site is totally unaffected by the parallel 2e
outer-sphere electron transfer at the oxide-covered Ru site. So, the Tafel slope that signifies the
slow rate determining step is not affected by the parallel outer-sphere step.
Based on the above interesting evidences found on Ru/C catalyst in alkaline medium the
effect of modification of Ru/C by chalcogen (S/Se) is discussed. Figure 3.5 shows ORR
characteristics of heat-treated Se modified Ru/C based binary and ternary catalysts measured at
900 rpm in 0.1 M NaOH electrolyte. As can be seen in Figure 3.5(a), on all the three catalysts i.e.
Ru/C, Se/Ru/C, and Se/RuMo/C a mixed kinetic-diffusion region is observed between 0.9 V to
0.6 V followed by a well-defined diffusion limited current density region below 0.6 V vs. RHE.
The ORR polarization curves of the selenium modified catalysts are shifted to more positive
89
potentials compared to that of the Ru/C catalyst. As quantified in Figure 3.5(b), Se/Ru/C catalyst
is shifted positively by 40 mV relative to Ru/C. From the literature data, in acidic medium
modification of Ru/C by selenium is typically observed to yield an anodic potential shift of ~100
to 150 mV.33,47 The anodic shift upon selenium modification of Ru/C in alkaline medium is
lower than in acidic medium because the activity of Ru/C is higher in alkaline medium than in
acidic medium even prior to selenium modification. However the Tafel slopes of Ru/C, Se/Ru/C,
and Se/RuMo/C, as shown in Table 3.2, typically exhibit a similar two slope region indicating
(a) PolarizationCurves
Potential [V Vs RHE]0.0 0.2 0.4 0.6 0.8 1.0
Cur
rent
Den
sity
[A/c
m2 ge
o]
-5e-3
-4e-3
-3e-3
-2e-3
-1e-3
0
Ru/CSe/Ru/CSe/RuMo/C
(c) Ring Current
Potential [V Vs RHE]0.0 0.2 0.4 0.6 0.8 1.0
Rin
g C
urre
nt [A
]
0
2e-6
4e-6
6e-6
8e-6
1e-5
1e-5
1e-5
Ru/CSe/Ru/CSe/RuMo/C
(b) Tafel Plots
log ik [A/cm2geo]1e-4 1e-3 1e-2
E V
s R
HE
0.75
0.80
0.85
0.90
ERing = 1.1V Vs RHE
40 mV
Figure 3.5: Comparison of ORR activity on heat treated Ru/C, Se/Ru/C, Se/RuMo/C catalysts in O2 saturated 0.1 M NaOH electrolyte at 900 rpm and 20 mV/s. (a) ORR Polarization curves, (b) Tafel Plots, and (c) Ring Current
90 that Se modification and ternary Mo additive has not affected the ORR reaction pathway
significantly. Tafel slopes shown in Table 3.2 are lower than the typical 60/120 mV/dec slopes
obtained on Pt primarily because of higher influence of the surface oxides on Ru based catalysts
that is known to be more oxophilic than the Pt surface. However, as is known from experiments
in acidic medium, selenium modification prevents the oxide formation on the underlying
transition metal active site. So, it is expected that selenium will suppress the outer-sphere
reaction and promote the direct molecular O2 adsorption on oxide-free Ru metal sites. In
agreement with this, the ring current peak observed between 0.6 V and 0.8 V which is
characteristic of the outer-sphere electron transfer process is slightly lower after selenium
modification. However the non-specificity characteristic of all outer-sphere electron transfer
reactions does not disallow the possibility of any peroxide formation on the oxides of selenium
in alkaline media.
Table 3.2: Electrochemical kinetic parameters of the various heat treated Ru based catalyst studied in comparison to commercial ETEK-BASF 30% Pt/C. Data obtained from RRDE experiments in O2 saturated 0.1 M NaOH electrolyte at 900 rpm and 20 mV/s.
a Catalyst Loadings: Pt - (15 µgPt/cm2geo); Ru based Chalcogenides – (12 µgRu/cm2
geo) b ik – Kinetic current density based on geometric area c io – Exchange current density based on geometric area d Tafel slopes were measured in the potential range (0.95 V-0.90 V/0.90 V-0.80 V) for platinum and (0.89 V-0.86 V/0.86 V-0.81 V) for Ru based catalysts e From the slope of Koutecky-Levich plot (ilim vs. ω0.5)
Catalysta ik x103 [A/cm2
geo]b @
0.9 V/0.8 V
iox109 [A/cm2
geo]c
Tafel Slopes [mV/dec]d
Number of Electrons
Transferrede
Peroxide Yield (%)
0.7 V 0.6 V
30% Pt/C 1.15/11.6 7.00/702 61/105 3.7 0.35 0.67 Ru/C --/1.6 --/4.00 39/76 3.5 5.51 5.01
Se/Ru/C --/4.4 0.3/77 59/91 3.7 2.55 2.29 Se/RuMo/C --/5.4 2/144 64/95 3.6 2.34 2.13
S/Ru/C --/2.6 0.02/52 52/93 3.6 2.27 1.72 S/RuMo/C --/5.0 0.18/22 48/79 3.8 1.23 9.61
91
Figure 3.6 shows the effect of modifying Ru/C catalyst with sulfur in 0.1 M NaOH
electrolyte at 900 rpm. As seen in Table 3.2, the kinetic current density of Se/Ru/C and S/Ru/C
are of the same order whereas the Mo additive increases the ORR activity marginally. Figure
3.6(c) shows representative Levich plots (Ilim vs. ω1/2), the slope of which is used to determine
the number of electrons transferred as shown in Table 3.2. Theoretical Levich slope (B) was
calculated using the parameters for diffusivity DO2 = 1.90x10-5 cm2s-1, solubility CO2 = 1.22x10-6
mol cm-3, and kinematic viscosity ν = 8.70x10-3 cm2s-1.48 As seen from both Table 3.2 and Figure
3.6 (c), the number of electrons transferred was typically found to be ≥3.6 on all the chalcogen
modified Ru/C catalysts. Figure 3.7 compares ORR and HRR on S/Ru/C-HT catalyst in 0.1 M
NaOH electrolyte. As expected HRR on S/Ru/C-HT occurs at more positive potential compared
to ORR. As shown in Figure 3.6(c) MoS/C catalyst was found to be predominantly a 2e catalyst.
As will be seen below based on EXAFS measurements, the structure of S/RuMo/C catalyst does
(c) Levich Plots
ωωωω1/2 [rpm 1/2]
10 20 30 40 50 60
I lim [m
A]
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
Ru/CSe/RuMo/CPt/CMoS/C
(a) ORR Polarization Curves
Potential [V Vs RHE]
0.0 0.2 0.4 0.6 0.8 1.0
Cur
rent
Den
sity
[A/c
m2 ge
o]
-5e-3
-4e-3
-3e-3
-2e-3
-1e-3
0
S/Ru/CS/RuMo/C
(b) Tafel Plots
log ik [A/cm2geo]1e-4 1e-3 1e-2
E V
s R
HE
0.8
0.9
n = 2
n = 4
Figure 3.6: (a) ORR Polarization curves, (b) Tafel Plots of S/Ru/C and S/RuMo/C in O2 saturated 0.1 M NaOH electrolyte at 900 rpm and 20 mV/s and (c) Levich plots.
92
not involve any alloy formation between Ru and Mo. Further it exists as distinct clusters of
sulfur decorated ruthenium and sulfur decorated molybdenum. The presence of sulfur/selenium
decorated molybdenum is to primarily generate HO2 intermediate which then diffuses on the
surface of the catalyst or spills over to ruthenium sites to further reduce to OH. This seems to be
the primary rationale behind the effect of the ternary Mo additive in improving the ORR activity
of binary X/Ru/C (X=S, Se) catalysts since no direct electronic or structural modification of the
binary catalysts are observed upon addition of Mo. (vide infra)
Figure 3.8 shows the effect of extending the potential cycling region of Se/Ru/C catalyst
to higher potentials of about 1.2 V vs. RHE. This set of experiments was performed to
understand chalcogen coordination in the catalysts and such unrealistic potentials of 1.2 V does
not arise under fuel cell operating conditions. Upon potentiostatically cycling the electrode to 1.2
V, a sharp increase in irreversible Se oxidation currents is observed above 0.85 V according to
Potential [V Vs RHE]
0.0 0.2 0.4 0.6 0.8 1.0
Cur
rent
Den
sity
[A/c
m2 ge
o]
-5e-3
-4e-3
-3e-3
-2e-3
-1e-3
0
1e-3O2 Reduction
HO2- Reduction
0.1M NaOH900 rpm20 mV/s
O2 Satd. 0.1M NaOH
3.5mM H2O2 in 0.1M NaOH
S/Ru/C
Figure 3.7: O2 and HO2¯ reduction on S/Ru/C-HT catalyst in 0.1 M NaOH. ORR performed in O2 saturated 0.1 M NaOH. HO2¯ reduction performed in 3.5 mM H2O2 comtaining O2-free 0.1 M NaOH. Conditions: 900 rpm and 20 mV/s
Potential [V Vs RHE]
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Cur
rent
Den
sity
[A/c
m2 ge
o]
-2e-3
-1e-3
0
1e-3
2e-3
3e-3
4e-3
E Vs RHE
0.05 0.10 0.15 0.20 0.25
i [A
/cm
2 geo]
-2e-4
0
2e-4
4e-4
6e-4
Scan 2
Scan 25
Scan 80
Figure 3.8: CV of Se/Ru/C in argon saturated 0.1 M NaOH at 50 mV/s up to an extended positive potential of 1.2 V. Inset is a magnification of the hydrogen desorption region.
93
reaction shown above in equation (3.1) followed by reaction in equation (3.7) shown below
leading to the formation of soluble selenate (SeO42-) ions.
SeO32¯ + 2OH → SeO4
2¯ + H2O + 2e (3.7)
The removal of selenium coordinated to the periphery of the Ru core subsequently causes the H
UPD peak on Ru to reappear slowly indicating that as Se gets removed from the surface, the Ru
surface sites exhibit its individual characteristic electrochemical properties. Also, after every ten
scans up to 1.2 V, the ORR activity of the electrode was measured and it is observed that once
the Se on the surface is continuously removed the activity decreases approaching that of the
Ru/C catalyst. (Data not shown)
In RRDE studies of ORR activity of catalysts under flooded electrolyte conditions the
potential of the electrode is typically scanned from ~1 V vs. RHE to potentials into the hydrogen
adsorption/desorption region. The activity of these catalysts is typically measured under such
Potential [V Vs RHE]0.0 0.2 0.4 0.6 0.8
Cur
rent
Den
sity
[A/c
m2 ge
o]
-5e-3
-4e-3
-3e-3
-2e-3
-1e-3
0Long RegionShort Region
Potential [V Vs RHE]0.2 0.4 0.6 0.8 1.0
Cur
rent
Den
sity
[A/c
m2 ge
o]
-5e-3
-4e-3
-3e-3
-2e-3
-1e-3
0
Initial ORR ScanAfter 1000 Cycles
Durability Test1000 Cycles were made in O2 Satd.
0.1M NaOH at 900 rpm and 20 mV/sbetween 0.9V & 0.6V
(a) (b)
Figure 3.9: (a) ORR on Se/RuMo/C catalyst showing the effect of negative potential limits on ORR activity. (b) Durability test of Se/RuMo/C catalyst. Electrolyte: O2 saturated 0.1 M NaOH, 20 mV/s scan rate, and 900 rpm rotation rate.
94 ideal conditions where scanning the potential into the H UPD region reduces all surface oxides
and upon the subsequent anodic scan a relatively high number of oxide free sites are present for
molecular O2 reduction. However as shown in Figure 3.9(a) restricting the negative potential
limit to about 0.6 V vs. RHE, the activity of Se/RuMo/C is found to be lower compared to a scan
where the negative potential limit was swept up to 0.05 V. This experiment bears relevance to an
operating fuel cell where galvanostatic control of the MEA typically leads to a cathode potential
of 0.7 V to 0.8 V and no opportunity is given to reduce the surface oxides. So care must be taken
while translating ORR activity from RRDE experiments to fuel cell conditions. Presence of Mo
in the Se/RuMo/C catalyst was found to significantly enhance the stability of the catalyst as
shown in Figure 3.9(b). Long term ORR on Se/RuMo/C catalyst was studied by cycling the
potential between 0.9 V and 0.6 V for 1000 scans. ORR activity shows only marginal decrease in
activity. This decrease in activity is primarily attributed to the oxidation of selenium at potentials
close to 0.9 V during potential cycling experiments.
3.3.3 In situ X-ray Absorption Spectroscopic Measurements
3.3.3.1. XANES and EXAFS:
X-ray Absorption Spectroscopic (XAS) measurements were performed under in situ
electrochemical conditions in oxygen free and oxygen saturated 0.1 M NaOH electrolyte at Ru
K-edge (22117 eV), Se K-edge (12658 eV) and Mo K-edge (20000 eV). Figure 3.10(a&b) shows
the X-ray Absorption Near Edge Spectra (XANES) and the non-phase corrected Fourier
transformed Extended X-ray Absorption Fine Structure (EXAFS) obtained at Ru K-edge on
Ru/C-HT catalyst in O2 saturated 0.1 M NaOH electrolyte at various potentials. The
corresponding EXAFS fit results are shown in Table 3.3. Ru K-edge spectra corresponds to the
electronic excitation of core level electrons from 1s levels to empty states with p-orbital
95
Table 3.3: Insitu EXAFS fit results for Ru/C-HT catalyst obtained from experiments performed at Ru K-edge (22117 eV) as a function of potential in O2 saturated 0.1 M NaOH electrolyte. Phase-corrected bond lengths are shown.
E [V vs. RHE]
Ru-Ru Ru-O Eo [eV] ∆RRu-Ru [Å]
N (R[Å]) N (R[Å]) N (R[Å])
0.05 3.0 (2.575) 4.3 (2.693) 0.10 (1.965) 5.866
-0.0531
0.3 2.9 (2.596) 4.1 (2.675) 0.30 (1.998) 6.099
-0.0478
0.7 2.8 (2.607) 4.0 (2.680) 0.45 (1.969) 6.348 -0.0422
0.8 2.6 (2.607) 3.7 (2.680) 0.80 (1.999) 6.348 -0.0432
1.0 2.1 (2.608) 3.0 (2.692) 1.11 (1.999) 7.063
-0.0412
character near the Fermi level. XANES fingerprint and oscillations in the extended region of the
Ru/C catalyst concur well with that of Ru metal. Arrows in Figure 3.10(a) indicate the change in
Ru K-edge position and intensity of white line and in Figure 3.10(b) the arrows indicate the
behavior of Ru-O and Ru-Ru interactions with increasing potentials. It is observed that there is
an increase in the white line intensity as the potential is increased from 0.05 V to 1.0 V
indicating that the Ru/C-HT catalyst is continuously oxidized under electrochemical conditions
as the potential increases. With increasing potential, the formation of electronegative oxide
species on Ru from specific adsorption of hydroxide anions causes a shift in the electron density
near the Fermi level of Ru atoms towards the oxide species, thus increasing the electron
vacancies. This leads to a higher probability of X-ray absorption and subsequent increase in the
white line intensity. This is also reflected in the Fourier transformed EXAFS spectrum shown in
Figure 3.10(b) where the Ru-Ru interaction centered at 2.5Å is seen to be decreasing and the Ru-
O peak centered on ~1.6 Å is found to be increasing continuously with increasing potentials.
Except for the spectrum at 0.05 V, a Ru-O peak centered on ~1.6 Å is clearly discernable. These
Ru-Ru and Ru-O interactions are quantified in terms of their bond lengths and coordination
96
number as shown in Table 3.3. Although Ru-O at ~1.6 Å is observed, peak corresponding to Ru-
Ru bond at 3.110 Å characteristic of Ru-O-Ru bridge bonds in RuO2 is absent indicating no
RuO2 formation but only formation of oxides/hydroxides on the surface of Ru. The coordination
number of Ru-Ru interaction decreases and Ru-O interaction increases systematically with
increasing potential. This behavior of Ru/C with increasing potential is expected since the
oxophilic character of Ru causes surface oxide films to form at potentials in, or close to the H
desorption region and have been observed before on bulk ruthenium electrodes in aqueous
electrolytes.44,49-50 In particular Ertl et al45 using LEED experiments showed that on a single
crystal Ru(0001) complex surface oxide phases of (2x2) and (3x1) are formed at potentials as
low as 0.2 V, 0.3 V vs. Ag/AgCl. At potentials less than 0.3 V vs. Ag/AgCl, O atoms occupy 3-
fold coordinated sites with θ<0.3 and at higher potential (1.1 V vs. Ag/AgCl), O atoms reside in
threefold hcp hollow sites with θ=1.
R[Å]
1.0 1.5 2.0 2.5 3.0 3.5 4.0
| χχ χχ (
R)|
[Å-2
]
0.0
0.1
0.2
0.3
0.4
0.05V0.3V0.6V0.8V1.0V
Ru-O
Ru-Ru
E (eV)
22100 22150 22200 22250 22300
Nor
mal
ized
µµ µµ(E
)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
0.05 V0.6 V0.8 V 1.0 V
Ru/C Ru/C(a)(b)
Figure 3.10: (a) X-ray absorption near edge region (XANES) and (b) non-phase corrected Fourier transformed EXAFS for Ru/C catalyst obtained at Ru K-edge insitu in O2 saturated 0.1 M NaOH electrolyte at various potentials as indicated. Corresponding EXAFS fit results are shown in Table 3.3.
97
Figure 3.11 shows a similar set of experiments performed at Ru K-edge on Se/Ru/C-HT
catalyst which clearly exhibits the electrochemical functionality of selenium on Ru. XANES and
Fourier transformed EXAFS results in Figure 3.11 (a&b) were obtained at Ru K-edge of
Se/Ru/C under in situ conditions at potentials from 0.05 V to 1.0 V vs. RHE in O2 saturated 0.1
M NaOH electrolyte and the corresponding EXAFS fit results are shown in Table 3.4. Similar
spectra were obtained at Ru K-edge on S/Ru/C catalyst also and hence not shown here. Two
important observations are made. Firstly, the underlying Ru metal nanoparticle is predominantly
intact with only minor changes in its bond length indicated by the Ru-Ru interaction centered at
2.5 Å. Ru-Ru bond length prior to chalcogen modification is observed at 2.596 Å at 0.3 V
whereas after modification by selenium an expansion of the Ru-Ru bond length to 2.643 Å at 0.3
V is observed. This is in contrast to the ruthenium diselenide RuSe2 compound where the
E (eV)
22080 22120 22160 22200 22240 22280
Nor
mal
ized
µµ µµ(E
)
0.0
0.2
0.4
0.6
0.8
1.0
0.05 V0.6 V0.8 V1.0 V
(a) Se/Ru/C
R [Å]
1.0 1.5 2.0 2.5 3.0 3.5 4.0
| χχ χχ (
R)|
[Å-3
]
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
0.05 V 0.6 V0.8 V 1.0 V
Ru-Ru
Ru-SeRu-O
(b) Se/Ru/C
Figure 3.11: (a) X-ray absorption near edge region and (b) non-phase corrected Fourier transformed EXAFS for Se/Ru/C catalyst obtained at Ru K-edge insitu in O2 saturated 0.1 M NaOH electrolyte at various potentials as indicated in the figure. Corresponding EXAFS fit results are shown in Table 3.4. Qualitatively similar results were obtained with S/Ru/C catalyst at the Ru K-edge.
98
Table 3.4: Insitu EXAFS fit results for Se/Ru/C-HT catalyst obtained from experiments performed at Ru K-edge (22117 eV) as a function of potential in O2 saturated 0.1 M NaOH electrolyte. Phase-corrected bond lengths are shown.
E [V vs. RHE]
Ru-Ru Ru-Se N (R[Å])
Ru-O N (R[Å])
Eo [eV] ∆RRu-Ru [Å] N (R[Å]) N (R[Å])
0.05 1.9 (2.639) 3.0 (2.694) 0.8 (2.415) -- 1.2367 -0.0101
0.3 2.4 (2.643) 2.6 (2.697) 0.9 (2.415) -- 1.2482 -0.0066
0.6 2.0 (2.641) 2.9 (2.695) 0.9 (2.400) -- 1.2811 -0.0083
0.8 1.8 (2.634) 3.1 (2.693) 0.9 (2.417) -- 1.5252 -0.0119
1.0 0.9 (2.710) 0.6 (2.765) -- 2.3 (2.033) 3.7020 -0.0014
shortest Ru-Ru bond distance is found at 4.197 Å. This clearly attests that in Se/Ru/C catalyst,
the Ru particle is clearly in its metallic state and the selenium atoms have not significantly
permeated the Ru hcp metal lattice. Secondly, the modification by chalcogen shifts the oxide
formation on Ru to potentials above 0.8 V. It is observed that the chalcogen (S and Se) modified
Ru catalysts does not exhibit any increase in white line intensity up to 0.8 V vs. RHE and also,
the Ru-Ru coordination number relatively remains constant up to 0.8 V in contrast to unmodified
Ru/C electrodes. At 1.0 V, there is a sudden rise in Ru-O interaction and a concomitant decrease
in Ru-Ru coordination numbers. The corresponding variations of Ru-Ru and Ru-O coordination
numbers and bond lengths are shown in Table 3.4. The Ru-Se bond length in Se/Ru/C catalyst is
observed at 2.415 Å at 0.3 V which is marginally lower than that in RuSe2 compound at 2.472 Å.
At 1.0 V, the selenium clusters coordinated to the periphery of the Ru cluster is subjected to
irreversible oxidation in alkaline medium according to equations (3.1) and (3.7) shown above.51
Consequently, the Ru sites deprived of protective Se are also subject to oxidation. Figure 3.12(a)
shows the Se K-edge XANES region at various potentials in deoxygenated 0.1 M NaOH
electrolyte. The increase in white line magnitude in the near edge region with increasing
99
Table 3.5: Representative insitu EXAFS fit results of Se/Ru/C-HT catalyst obtained from experiments performed at Se K-edge (12658 eV) at 0.5 V and 1.1 V vs. RHE in O2 saturated 0.1 M NaOH. Phase-corrected bond lengths are shown.
Potential [V vs. RHE]
Se-Ru Interaction
Se-Se Interaction
Se-O Interaction
N (R[Å]) N (R[Å]) N (R[Å]) 0.5 V 2.04 (2.384) 0.40 (2.437) 0.1 (1.749) 1.1 V 1.79 (2.384) 0.31 (2.437) 1.1 (1.745)
potential indicate that Se overlayers undergo minor oxidation when the potential is increased
from 0.05 V to 0.8 V above which the Se structure significantly breaks down forming selenium
oxides according to reactions in equation (3.1) and (3.7) shown above. This is in contrast to the
behavior in acidic medium wherein it was earlier reported that no change in XANES region was
observed with increasing potentials up to 800 mV in 0.1 M H2SO4.52 Figure 3.12(b) shows the
Fourier transformed EXAFS spectra of Se/Ru/C at Se K-edge (12658 eV) under in situ
conditions at 0.5 V in O2 saturated 0.1 M NaOH and the corresponding k2-weighted first shell fit.
The corresponding fit results are shown in Table 3.5. This fit was obtained by using the
ruthenium diselenide RuSe2 (space group: p a-3) lattice parameters for the Se-Ru and Se-Se bond
interactions. Elemental trigonal selenium (space group: P 32 2 1) content was observed prior to
heat treatment but not in the heat treated catalysts.53-54 Presence of elemental selenium was
observed using XPS analysis in the study by Savinova et al.47 The main peak due to Se-Ru and
Se-Se interactions are observed in the R-space centered on 2.38 Å. A resonant peak
corresponding to the same Se-Ru interactions is also observed at around 1.8 Å.52 As shown in
Table 3.5, Se-Ru bond length in the Se/Ru/C catalyst sample is 2.384 Å which is marginally
lower than that of 2.471 Å in RuSe2 standard compound. Se-Se bond length at 2.437 Å with a
low coordination number of NSe-Se = 0.4 as shown in Table 3.5. Also, as shown in Figure 3.12(c)
100 the Se-O peak significantly grows at 1.1 V vs. RHE indicating the breakdown of Se structure on
the Ru surface at higher potentials.
Se K-edgeSe/Ru/C
Energy (eV)
12655 12660 12665 12670 12675
Norm
alized
µµ µµ(E
)
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
0.05 V0.50 V0.80 V 1.00 V
Se K-edgeSe/Ru/C
|Chi
(R)| [Å
]
0.2
0.4
0.6
0.8
1.0
1.2
Experimental Theoretical Fit
Se - RuSe - Se
Se - RuSe - Se
R [Å]0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
|Chi
(R)| [Å
-3]
0.0
0.2
0.4
0.6
0.8
1.0
0.5 V1.1 V
Se K-edgeSe/Ru/C
Se-RuSe-Se
Se - RuSe - O
(a)
(b)
(c)
Figure 3.12: (a) Se K-edge XANES spectra at various potentials, (b) Fourier transformed Se K-edge EXAFS spectra taken at 0.5 V, and (c) Comparison of Se K-edge EXAFS at 0.5 V and 1.1 V. Experiments performed in O2 saturated 0.1 M NaOH.
101
Figure 3.13 shows the Ru K-edge Fourier transformed EXAFS spectra of S/RuMo/C
catalysts in O2 saturated 0.1 M NaOH at 0.6 V vs. RHE and the corresponding coordination
numbers and bond lengths are shown in Table 3.6. It is observed that Ru-Mo interaction is absent
in S/RuMo/C synthesized via the aqueous route which is in agreement with the literature for
RuMoX (X=S, Se) catalysts synthesized via non-aqueous routes also.37-38,55 Comparing Tables
3.4 & 3.6, it is found that the bond lengths of Ru-Ru and Ru-X (X=S, Se) in both S/Ru/C and
X/RuMo/C catalysts are similar and do not show any systematic difference although the Ru-Ru
coordination numbers are significantly larger for the X/RuMo/C catalyst. Interestingly, for the
S/RuMo/C catalyst the Ru-Ru and Ru-S coordination numbers are very stable and do not show
any signs of corrosion even up to 1.0 V vs. RHE in alkaline medium. This is shown in Table 3.6
R [Å]
1 2 3 4 5
| χ
χ
χ
χ (
R)|
[Å-3
]
0.0
0.5
1.0
1.5
2.0
2.5
Exptl.Fit
Ru-Ru
Ru-Se
Ru-O
R [Å]1.0 1.5 2.0 2.5 3.0 3.5
| χχ χχ (
R)|
[Å-3
]
S/RuMo/CMo Foil
(b) Mo K-Edge(a) Ru K-EdgeS/RuMo/C
Mo-SeMo-O
Mo-Mo
Mo-Mo
Figure 3.13: Fourier transformed EXAFS spectra of S/RuMo/C catalyst under insitu conditions at 0.6 V vs. RHE in O2 saturated 0.1 M NaOH electrolyte at (a) Ru K-edge and (b) Mo K-edge. Also shown in inset (b) is the reference Mo foil at Mo K-edge.
102 where the Ru-Ru and Ru-S interactions are found to be very stable even up to 1.0 V vs. RHE
attesting to the stability of the ternary S/RuMo/C catalyst. This is in good agreement with the
improved stability of the ternary catalyst observed in acidic medium.43 Figure 3.13(b) and Table
3.7 shows the Mo K-edge spectra obtained using a Mo foil (ex situ - 6 micron thick) and
S/RuMo/C catalyst in situ at 0.6 V in deoxygenated 0.1 M NaOH electrolyte. Reference Mo foil
was fit using Mo IM3M (body centered cubic) crystal structure data. Two Mo-Mo interactions at
2.741 Å and 3.163 Å are observed. In the case of S/RuMo/C, the Mo-Mo interaction at 2.741 Å
is absent whereas the second Mo-Mo interaction at higher bond length is observed at 3.154 Å.
Mo K-edge data of S/RuMo/C was fit using the MoS2 hexagonal crystal structure (space group
P63/mmc). Mo-Ru interaction is not observed. Further, Mo is observed to exist as clusters of
MoxXy and MoOz. This is in accordance with similar materials synthesized via non-aqueous
routes.55
Table 3.6: Insitu EXAFS fit results for S/RuMo/C-HT catalyst obtained from experiments performed at Ru K-edge (22117 eV) as a function of potential in O2 saturated 0.1 M NaOH electrolyte. Phase-corrected bond lengths are shown.
E [V vs. RHE]
Ru-Ru Ru-S N (R[Å])
Ru-O N (R[Å])
Eo [eV] ∆RRu-Ru [Å] N (R[Å]) N (R[Å])
0.05 4.9 (2.589) 5.6 (2.677) 2.0 (2.324) 0.1 (1.965) 5.190 -0.0607
0.3 4.2 (2.635) 3.9 (2.679) 1.9 (2.335) 0.1 (1.965) 6.147 -0.0143
0.6 5.7 (2.630) 3.9 (2.681) 2.0 (2.345) 0.1 (1.968) 4.888 -0.0190
0.8 5.8 (2.621) 4.3 (2.680) 2.0 (2.343) 0.1 (1.966) 5.497 -0.0280
1.0 5.7 (2.631) 3.8 (2.679) 2.0 (2.349) 0.1 (1.965) 5.499 -0.0178
Table 3.7: Representative insitu EXAFS fit results of S/RuMo/C-HT catalyst obtained from experiments performed at Mo K-edge (20000 eV) at 0.6 V Vs. RHE in O2 saturated 0.1 M NaOH in comparison to exsitu Mo foil. Phase-corrected bond lengths are shown.
Mo-Mo Mo-S Mo-O N (R[Å]) N (R[Å]) N (R[Å]) N (R[Å])
Mo Foil 6.7 (2.741) 5.0 (3.163) -- -- RuMoSe @ 0.6 V -- 3.5 (3.154) 5.3 (2.411) 1.1 (2.254)
103 3.4 Conclusions
It is understood that on unmodified Ru/C catalyst ORR activity in alkaline medium is
more facile compared to acidic medium. This kinetic facility arises due to parallel in situ
generation of peroxide anion intermediate in alkaline medium via an outer-sphere electron
transfer mechanism mediated by the specifically adsorbed hydroxyl species. This shifts the
potential of the electrode to more positive values by effectively carrying out peroxide reduction
at oxide-free Ru sites. However no evidence for outer-sphere reaction is observed in acidic
medium thereby leading to higher overpotential for direct molecular O2 reduction. Upon
selenium or sulfur modification of Ru/C catalyst a decrease in the overpotential for ORR is
observed, although this positive shift in alkaline medium is lower than what is typically observed
in acidic medium. Structural characterizations indicate that the chalcogen primarily remains
coordinated to the surface of the ruthenium nanoparticle and preserves the underlying transition
metal in its metallic state. Presence of the chalcogen prevents oxidation of the transition metal up
to potentials of about 0.8 to 0.9 V vs. RHE in alkaline medium. Inclusion of a ternary element
such as Mo improves the stability of the X/Ru/C (X=S, Se) catalyst. S/RuMo/C catalyst
primarily exists as a composite of clusters of sulfur modified ruthenium and molybdenum
nanoparticles. Since no direct alloy formation or interaction between Ru and Mo is observed, it is
proposed that any increase in activity of the ternary catalyst is primarily based on a spillover
effect of peroxide intermediate from molybdenum sulfide clusters to the ruthenium sites.
104 3.5 Acknowledgements:
The authors deeply appreciate the financial assistance of the Army Research Office under the
Single Investigator grant. Use of the National Synchrotron Light Source (NSLS), Brookhaven
National Laboratory (BNL), was supported by the U.S. Department of Energy, Office of Basic
Energy Sciences. Support from beamline personnel Dr. Kaumudi Pandya (X11A) is gratefully
acknowledged. Assistance from Robert J. Allen on catalyst synthesis is deeply appreciated.
3.6 References:
(1) Bockris, J. O.; Appleby, J. In Assessment of Research Needs for Advanced Fuel Cells.; Penner, S. S., Ed. 1986; Vol. 11, p 95. (2) Varcoe, J. R.; Slade, R. C. T. Fuel Cells 2005, 5, 187. (3) Spendelow, J. S.; Wieckowski, A. Phys. Chem. Chem. Phys. 2007, 9, 2654. (4) Adzic, R. In Electrocatalysis; Lipkowski, J., Ross, P. N., Eds.; Wiley-VCH: 1998, p 197. (5) Bard, A. J.; Parsons, R.; Jordan, J.; Editors Standard Potentials in Aqueous Solution, 1985. (6) Yang, H.-H.; McCreery, R. L. J. Electrochem. Soc. 2000, 147, 3420. (7) Blizanac, B. B.; Ross, P. N.; Markovic, N. M. Electrochimica Acta 2007, 52, 2264. (8) Morcos, I.; Yeager, E. Electrochim. Acta 1970, 15, 953. (9) Yeager, E. Electrochim. Acta 1984, 29, 1527. (10) Lima, F. H. B.; de Castro, J. F. R.; Ticianelli, E. A. J. Power Sources 2006, 161, 806. (11) Ryabukhin, A. G.; Ershov, A. I. 1971, No. 17, 60. (12) Lima, F. H. B.; Calegaro, M. L.; Ticianelli, E. A. J. Electroanal. Chem. 2006, 590, 152. (13) Lima, F. H. B.; Calegaro, M. L.; Ticianelli, E. A. Electrochim. Acta 2007, 52, 3732. (14) Ohsaka, T.; Mao, L.; Arihara, K.; Sotomura, T. Electrochem. Commun. 2004, 6, 273. (15) Wang, B. J. Power Sources 2005, 152, 1. (16) Zagal, J. H. Coordination Chemistry Reviews 1992, 119, 89. (17) Chang, C. J.; Deng, Y.; Nocera, D. G.; Shi, C.; Anson, F. C.; Chang, C. K. Chemical Communications (Cambridge) 2000, 1355. (18) Santos, D. M. F.; Sequeira, C. A. C. Diffusion and Defect Data--Solid State Data, Pt. A: Defect and Diffusion Forum 2006, 258-260, 327.
105 (19) Sequeira, C. A. C.; Santos, D. M. F.; Baptista, W. Journal of the Brazilian Chemical Society 2006, 17, 910. (20) Horowitz, H. S.; Longo, J. M.; Horowitz, H. H. J. Electrochem. Soc. 1983, 130, 1851. (21) Singh, R. N.; Tiwari, S. K.; Chartier, P. Indian J. Chem., Sect. A 1990, 29A, 837. (22) Alonso-Vante, N. In Catalysis and Electrocatalysis at Nanoparticle Surfaces; Wieckowski, A., Savinova, E. R., Vayenas, C. G., Eds. 2003, p 931. (23) Vante, N. A.; Tributsch, H. Nature (London) 1986, 323, 431. (24) Vante, N. A.; Schubert, B.; Tributsch, H.; Perrin, A. J. Catal. 1988, 112, 384. (25) Vante, N. A.; Jaegermann, W.; Tributsch, H.; Hoenle, W.; Yvon, K. J. Am. Chem. Soc. 1987, 109, 3251. (26) Alonso-Vante, N.; Schubert, B.; Tributsch, H. Mater. Chem. Phys. 1989, 22, 281. (27) Solorza-Feria, O.; Ellmer, K.; Giersig, M.; Alonso-Vante, N. Electrochim. Acta 1994, 39, 1647. (28) Alonso-Vante, N.; Malakhov, I. V.; Nikitenko, S. G.; Savinova, E. R.; Kochubey, D. I. Electrochim. Acta 2002, 47, 3807. (29) Dassenoy, F.; Vogel, W.; Alonso-Vante, N. The Journal of Physical Chemistry B 2002, 106, 12152. (30) Babu, P. K.; Lewera, A.; Chung, J. H.; Hunger, R.; Jaegermann, W.; Alonso-Vante, N.; Wieckowski, A.; Oldfield, E. J. Am. Chem. Soc. 2007, 129, 15140. (31) Alonso-Vante, N.; Bogdanoff, P.; Tributsch, H. J. Catal. 2000, 190, 240. (32) Reeve, R. W.; Christensen, P. A.; Hamnett, A.; Haydock, S. A.; Roy, S. C. J. Electrochem. Soc. 1998, 145, 3463. (33) Cao, D.; Wieckowski, A.; Inukai, J.; Alonso-Vante, N. Journal of the Electrochemical Society 2006, 153, A869. (34) Trapp, V.; Christensen, P.; Hamnett, A. J. Chem. Soc., Faraday Trans. 1996, 92, 4311. (35) Alonso-Vante, N.; Tributsch, H.; Solorza-Feria, O. Electrochim. Acta 1995, 40, 567. (36) Markovic, N. M.; Ross, P. N. Surface Science Reports 2002, 45, 117. (37) Ziegelbauer, J. M.; Murthi, V. S.; O'Laoire, C.; Gulla, A. F.; Mukerjee, S. Electrochim. Acta 2008, 53, 5587. (38) Malakhov, I. V.; Nikitenko, S. G.; Savinova, E. R.; Kochubey, D. I.; Alonso-Vante, N. J. Phys. Chem. B 2002, 106, 1670. (39) Allen, R. J.; Gulla, A. F.; (De Nora Elettrodi S.p.A., Italy). US 20050164877, 2005. (40) Campbell, S. A.; (Ballard Power Systems Inc., Can.). US 2004096728, 2004. (41) Arruda, T. M.; Shyam, B.; Lawton, J. S.; Ramaswamy, N.; Budil, D. E.; Ramaker, D. E.; Mukerjee, S. J. Phys. Chem. C 2010, 114, 1028. (42) Newville, M. Journal of Synchrotron Radiation 2001, 8, 322. (43) Guinel, M. J. F.; Bonakdarpour, A.; Wang, B.; Babu, P. K.; Ernst, F.; Ramaswamy, N.; Mukerjee, S.; Wieckowski, A. ChemSusChem 2009, 2, 658. (44) Anastasijevic, N. A.; Dimitrijevic, Z. M.; Adzic, R. R. J. Electroanal. Chem. Interfacial Electrochem. 1986, 199, 351. (45) Zei, M. S.; Ertl, G. Phys. Chem. Chem. Phys. 2000, 2, 3855.
106 (46) Anastasijevic, N. A.; Dimitrijevic, Z. M.; Adzic, R. R. Electrochimica Acta 1986, 31, 1125. (47) Zaikovskii, V. I.; Nagabhushana, K. S.; Kriventsov, V. V.; Loponov, K. N.; Cherepanova, S. V.; Kvon, R. I.; Boennemann, H.; Kochubey, D. I.; Savinova, E. R. J. Phys. Chem. B 2006, 110, 6881. (48) Markovic, N. M.; Gasteiger, H. A.; Ross, P. N., Jr. Journal of Physical Chemistry 1996, 100, 6715. (49) Prakash, J.; Joachin, H. Electrochim. Acta FIELD Full Journal Title:Electrochimica Acta 2000, 45, 2289. (50) Hadzi-Jordanov, S.; Angerstein-Kozlowska, H.; Vukovic, M.; Conway, B. E. J. Electrochem. Soc. 1978, 125, 1471. (51) Pourbaix, M. Atlas of Electrochemical Equilibria in Aqueous Solutions, 1966. (52) Inukai, J.; Cao, D.; Wieckowski, A.; Chang, K.-C.; Menzel, A.; Komanicky, V.; You, H. J. Phys. Chem. C 2007, 111, 16889. (53) Cherin, P.; Unger, P. Inorganic Chemistry 1967, 6, 1589. (54) Cherin, P.; Unger, P. Acta Crystallographica Section B 1972, 28, 313. (55) Malakhov, I. V.; Nikitenko, S. G.; Savinova, E. R.; Kochubey, D. I.; Alonso-Vante, N. Nucl. Instrum. Methods Phys. Res., Sect. A 2000, 448, 323.
107
Chapter 4
Redox Potential Tuning and Influence of Graphitic Defects on the Origin of ORR Activity
of Pyrolyzed Iron Porphyrin Electrocatalysts
4.1 Introduction:
The drive to replace expensive and scarce Pt based catalysts for cathodic Oxygen
Reduction Reaction (ORR) has led to a class of electrocatalysts composed of first row transition
metal ions stabilized by surface nitrogen functionalities on graphitic surfaces.1-5 While the non-
pyrolyzed versions of these catalysts primarily yield 2e¯ reduction products, heat treatment of
metallomacrocycles has always been resorted to in order to increase the stability, activity and
selectivity of the 4e- reduction route.6-8 Although some authors observed that ORR is conducted
by sites comprised of surface nitrogen groups devoid of any metal ion centers,9-10 it is now
widely accepted that the transition metal ion centers coordinated to the surface nitrogen groups
(Me-Nx) constitute the active site,7-8,11-12 whereas the ligand primarily serves to prevent the
metal center from passivation/corrosion under electrochemical conditions.13 The nature of the
active site in terms of its location on the carbon support (edge vs. basal plane),7 coordination
number (Fe-N4 vs. non-Fe-N4 environment),8 chemical identity of the nitrogen functional groups
(pyridinic, pyrrolic, quaternary)14 have remained a key aspect of intense discussion. Several
theories exist to explain the nature of the active site. van Veen15-17, McBreen18 and Schulenburg19
et al hypothesize that the partial destruction of the metal-macrocycle complexes during pyrolysis
and the formation of the secondary structures containing M-N4/C are responsible for the catalytic
activity. Yeager 2,11 and Scherson20-22 et al suggested that metallic iron or oxide is the primary
heat treatment product which dissolves when brought into contact with acidic electrolyte and re-
108 coordinates to surface nitrogen groups to provide the active sites of the form C-Nx-M. Following
a series of studies using separate precursors for the metal and nitrogen, Dodelet et al 7,23-34
proposed the active site to consist of iron metal cation coordinated by four pyridinic nitrogens
attached to the edges of two graphitic sheets in the micropores (width ≤ 20 Å) of the carbon
support. Dodelet et al7 and Dahn et al35 reported a direct correlation between catalytic activity
and surface nitrogen content. Besides the exact structure of the active site, the low active site
density or the metal loading that are obtained in these catalysts eludes clear understanding. A
maximum in catalytic activity is obtained at a very low metal loading (~5000 ppm by weight for
inorganic precursors and ~2 wt% for macrocycle precursors).24 Any higher loading does not lead
to increase in activity but rather to formation of inactive metallic or metal carbide clusters.24
Dodelet et al36 suggested that only the microporosity generated during pyrolysis upon
gasification of disordered carbon content act as host for active sites and the various grades of
carbon support with varying microporous surface areas are immaterial. As pointed out by
Gasteiger et al, 37 such low cost electrocatalysts when utilized in higher loadings (~not more than
10 times higher than that of the present Pt based cathodes i.e., not more than 100µm thick
electrodes) may be acceptable in a fuel cell cathode if a nominal but continuous loss in activity
of these materials leads to minimal impact on fuel cell performance. Although higher activities
have been recently demonstrated, the stability of these Fe-based catalysts is unacceptably inferior
to translate them into an operating fuel cell cathode.26
One drawback on most studies in this class of materials is that, irrespective of whatever
the exact nature of the active site is, this does not necessarily explain the fundamental origin and
the causes for limitations in the ORR activity of these catalysts. It was recently pointed out in
reviews by Bezerra38 and Dahn35 et al that most studies on this class of materials have focused on
109 the optimal synthesis conditions and structure necessary for maximum activity, whereas a more
fundamental understanding will be of great help in designing alternative and innovative routes
for a new catalyst synthesis. Based on a cross laboratory study, the authors proposed that the
nature of the catalytic site from various laboratories is of fundamentally the same nature.35 For
non-heat treated metallomacrocycles, Zagal et al6 showed that the d-orbital character of the metal
center dictates the course of the reaction. Several relationships between the ORR activity, redox
potential of the metal center and the intermolecular hardness of the macrocycle were established
for the non-heat treated macrocycles.3-4,39-41 No such definite correlations could be obtained for
the heat treated catalysts given the obscurity surrounding the nature of the active site. van Veen
et al16 recently showed that upon heat treatment of iron porphyrin, the Fe2+/3+ redox reaction
shifted from 0.2 V to 0.4 V vs. RHE. However, a redox potential of 0.4 V after heat treatment
does not explain an ORR onset potential of 0.8 V in acidic medium.
It was shown in Chapter 2 that electrocatalysis of O2 reduction in alkaline medium
involves both inner- and outer-sphere electron transfer mechanisms. While the inner-sphere
mechanism involves the well known direct molecular O2 adsorption on the active site, outer-
sphere electron transfer involves the reduction of O2 to peroxide intermediate promoted/mediated
by the specifically adsorbed hydroxide anions on the active site. The involvement of surface
adsorbed hydroxide species causes a certain non-specificity to the identity of the underlying
electrode material and opens the gateway to use wide range of non-noble metal electrodes in
alkaline medium. One of the objectives of this research effort is to develop non-noble
electrocatalysts for oxygen reduction that suppress the outer-sphere electron transfer and
promote direct molecular adsorption of O2 so that efficient 4e reduction can be achieved. A
combination of electrochemical and X-ray absorption spectroscopic measurements have been
110 used in order to understand the structure/property relationships of the pyrolyzed class of
catalysts. While understanding the nature of the active site is important, the primary objectives of
this chapter are to understand the fundamental reasons for the origin of ORR activity and the
causes for the limitation in the active site density in this class of heat treated catalysts.
4.2 Experimental
4.2.1 Catalyst Preparation
Iron(III) meso-tetraphenylporphyrin chloride (FeTPPCl) was procured from Alfa Aesar
and used as received. The molecular weight of FeTPPCl is 704 g/mol, implying 7.95% by weight
of iron content in the original macrocycle. FeTPPCl was mixed with Black Pearl carbon (BPC)
in the mass ratio 1:4 and ball milled for 2 hours at 400rpm followed by pyrolysis at various
temperatures ranging from 300°C to 1100°C for 2 hours under argon atmosphere. The loading of
iron on the carbon support was typically 3±0.15% by weight as determined by Energy Dispersive
Analysis of X-rays using a EDS-GENESIS HITACHI S-4800 instrument.
4.2.2. Electrochemical characterization
All electrochemical measurements were made at room temperature using a rotating ring-
disk electrode (RRDE) setup from Pine Instruments connected to an Autolab (Ecochemie Inc.,
model-PGSTAT 30) bi-potentiostat. Alkaline (0.1 M NaOH) and acidic (0.1 M HClO4)
electrolytes were prepared using sodium hydroxide pellets (semiconductor grade, 99.99%,
Sigma-Aldrich) and double-distilled 70% perchloric acid (GFS Chemicals) respectively. Catalyst
inks were typically prepared by dispersing 25 mg of the catalyst in 10 ml of 1:1 Millipore
H2O:Isopropyl alcohol mixture along with 100 µL of 5 wt% Nafion(R) solution as a binder. 5 µL
aliquot of the catalyst ink was dispensed on Glassy Carbon (GC) disk of 5.61mm dia. Gold ring
111 electrode was held at 1.1 V vs. RHE in alkaline electrolyte and at 1.3 V vs. RHE in acidic
electrolyte to detect stable peroxide intermediate. Collection efficiency of the disk-ring electrode
was 37.5%. All potentials are refered to a reversible hydrogen electrode (RHE) scale made out of
the same solution as the bulk electrolyte unless otherwise stated. Square Wave voltammetry
(SWV) was performed at frequency of 10 Hz using 5 mV step potential and 20 mV amplitude.
4.2.3 X-ray Absorption Spectroscopic (XAS) Measurements
The in situ XAS studies at Fe K-edge (7112 eV) was performed at the X19A beamline of
the National Synchrotron Light Source (NSLS, Brookhaven National Laboratory, NY). Detailed
information on the spectro-electrochemical cell design are given elsewhere.42 Spectra at Fe K-
edge were collected in fluorescence mode using a PIPS detector. Argon or oxygen saturated 0.1
M NaOH was used as the electrolyte. Details on data analysis of X-ray Absorption Near Edge
Spectrum (XANES) and Extended X-ray Absorption Fine Structure (EXAFS) are also available
in an earlier publication.42 Briefly, the IFEFFIT suite Version 1.2.9 43 was used for background
subtraction using AUTOBK algorithm and normalization. Typical k-range window during
EXAFS fit was 2.500-12.500 Å-1 (Kaiser-Bessel). Data analysis for Delta-Mu (∆µ) studies at Fe
K-edge involved specific normalization procedures detailed elsewhere.42,44 This involves careful
calibration of edge energy (Fe K-edge 7112 eV), alignment to standard reference scan to account
for any drift in the beam energy. A postedge normalization procedure was then applied to the
aligned scans via a cubic spline function which normalizes the oscillations over a specific energy
range (typically 25 to 200 eV with respect to E0) on a per-atom basis. Difference spectra were
obtained using the equation ∆µ = µ(V) - µ(0.1 V), where µ(V) is the XANES spectra of the
catalyst at various potentials and µ(0.1 V) is the reference XANES signal at 0.1 V at which
potential no evidence for electrochemical adsorbates (Hupd, Oads, OHads) were found on these iron
112 based catalysts. Theoretical delta mu curves (∆µt) were constructed using the FEFF 8.0 code.45
This was accomplished using the relationship ∆µt = µ(Oads-Fe-Nx-C) - µ(Fe-Nx-C), where the
oxide species (Oads or OHads) is in a specific binding site on Fe. It should be noted that theoretical
∆µ spectra are generally shifted by upto 10-15 eV and scaled by a multiplication factor, if
neccessary, for optimal comparison with experimental data.
4.3 Results and Discussions
4.3.1. Electrochemical Characterization and Oxygen Reduction Reaction:
Figure 4.1 shows a comparison in dilute acidic and alkaline electrolytes the ORR activity
of FeTPP/C catalyst pyrolyzed at 800°C. The maximum activity was observed at heat treatment
temperature of 800°C as shown in Figure 4.2. As observed in Figure 4.1(a), the onset potential
for ORR in 0.1 M NaOH is 0.95 V vs. RHE whereas in 0.1 M HClO4 electrolyte it is 0.80 V vs.
RHE. This 150 mV lower overpotential in alkaline medium is clearly reflected over the entire
mixed kinetic-diffusion region. In 0.1 M NaOH electrolyte, the mixed kinetic-diffusion region is
ensued by a well-defined diffusion limited region. In 0.1 M HClO4, no clear diffusion limited
region could be discerned, which is indicative of kinetic control in acidic medium even at high
overpotentials. Electrochemical kinetic parameters are summarized in Table 4.1. At a potential of
0.80 V vs. RHE the ORR kinetic current density of FeTPP/C (pyrolyzed at 800°C) is clearly four
orders of magnitude higher in 0.1 M NaOH electrolyte than that in 0.1 M HClO4. Similar Tafel
slopes for FeTPP/C catalyst pyrolyzed at 800°C is indicative of the same rate determining step in
both acidic and alkaline electrolytes. However, for a given catalyst the four-orders of magnitude
difference in kinetic activity between acidic and alkaline medium is intriguing and requires
further detailed investigations as discussed below. Figure 4.1(c) shows the hydrogen peroxide
113 reduction activity of FeTPP/C catalyst (pyrolyzed at 800°C) in both acidic and alkaline medium
in comparison to the corresponding ORR polarization curves.
The H2O2 reduction study was carried out in oxygen-free electrolytes containing 3.5 mM
H2O2. The onset potential for peroxide reduction in 0.1 M HClO4 is 0.84 V vs. RHE whereas in
(c) Peroxide Reduction
Potential [V Vs RHE]
0.2 0.4 0.6 0.8 1.0
i [A
/cm
2 geo]
-2e-3
-1e-3
0
1e-3
2e-3
ORRHRR
(b) Ring Current
Potential [V Vs RHE]0.0 0.2 0.4 0.6 0.8 1.0
I R [A
]
0
1e-6
2e-6
3e-6
4e-6
5e-6
0.1M NaOH0.1M HClO4
(a) ORR Polarization Curves
Potential [V Vs RHE]0.0 0.2 0.4 0.6 0.8 1.0
i D [A
/cm
2 geo]
-5e-3
-4e-3
-3e-3
-2e-3
-1e-3
0
0.1M NaOH0.1M HClO4
0.1M HClO4
0.1M NaOH
Figure 4.1: ORR activity of FeTPP/C catalyst (pyrolyzed at 800°C) in O2 saturated acidic and alkaline electrolytes. (a) ORR polarization curves, (b) ring current, and (c) H2O2 Reduction Reaction (HRR) in comparison to ORR. All measurements were performed at 900 rpm rotation rate and 20 mV/s scan rate. Ering = 1.1 V vs. RHE in 0.1 M NaOH and Ering = 1.3 V vs. RHE in 0.1 M HClO4. HRR is shown in oxygen-free electrolytes containing externally added H2O2 at a concentration of 3.5mM.
114 0.1 M NaOH it is 1.01 V vs. RHE. Besides this onset potential difference, in 0.1 M NaOH
electrolyte the mixed kinetic-diffusion region for peroxide reduction is more anodic compared to
that of ORR in the same electrolyte which is then followed by a reasonably discernable diffusion
limited current density region. This clearly indicates that peroxide reduction in alkaline medium
is kinetically favored such that any peroxide intermediate formed during ORR in 0.1 M NaOH
will be immediately reduced to the 4e¯ product. On the contrary, the reduction of hydrogen
peroxide in acidic medium is kinetically unfavorable due to the following reasons: weak binding
of peroxide intermediate on the active site leads to desorption into the bulk electrolyte or
catalytic decomposition to molecular O2. This clearly indicates that stabilizing the peroxide
intermediate on the active site is important in effectively carrying out ORR. However, the
reasons for the stability of peroxide intermediate on the active site in alkaline medium but not in
acidic medium is intriguing. This difference is clearly electrochemical in origin and could be
attributed to a double-layer Frumkin effect.46-47 Given the pKa values for the first and second
ionization of H2O2 at 25°C (pK1 = 11.69 and pK2 = ~20) the predominant peroxide species for
pH<12 is H2O2 whereas at pH>12 it is HO2¯.48 Accordingly, the cationic nature of the Fe2+ active
site (vide infra) electrostatically stabilizes the anionic HO2¯ species in alkaline medium. Such an
electrostatic double-layer Frumkin effect is absent in acidic medium given the neutral charge on
H2O2. Figure 4.1(b) shows the ring current due to peroxide oxidation measured during ORR at
900 rpm in both acidic and alkaline electrolytes. Clearly the peroxide generated during ORR on
FeTPP/C (pyrolyzed at 800°C) is higher in acidic medium compared to that in alkaline medium
(Table 4.1).
115
Table 4.1: Electrochemical kinetic parameters of the FeTPP/C catalyst (pyrolyzed at 800°C) in comparison to ETEK-BASF 30% Pt/C. Data obtained from RRDE experiments in O2 saturated 0.1 M HClO4 and 0.1 M NaOH electrolytes at 900 rpm and 20 mV/s.
a Catalyst Loadings: Pt - (15 µgPt/cm2geo); FeTPP – (0.75 µgFe/cm2
geo) b From the slope of Levich plot (ilim vs. ω0.5);
Catalysta Electrolyte ik x103 [A/cm2
geo] @
0.9 V/0.8 V
iox109 [A/cm2
geo] Tafel
Slopes [mV/dec]
Number of Electrons
Transferredb
Peroxide Yield (%)
0.7 V 0.6 V
30% Pt/C 0.1 M NaOH 1.15/11.6 7.00 61/105 3.7 0.35 0.67 30% Pt/C 0.1 M HClO4 2.32/23.0 47 70/101 4.0 0.25 0.23 FeTPP/C 0.1 M NaOH 0.52/10.3 0.30 55/90 4.0 0.40 0.67 FeTPP/C 0.1 M HClO4 --/0.008 0.007 60/99 3.7 6.54 6.75
(a) ik and % HO2-
Pyrolysis Temperature [oC]300 400 500 600 700 800 900 1000 1100
i k @
0.9
V [A
/cm
2 geo]
-1e-4
0
1e-4
2e-4
3e-4
4e-4
5e-4
6e-4
HO
2- yie
ld @
0.6
V [%
mol
e fr
actio
n]
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Figure 4.2: (a) Kinetic current (ik) for FeTPP/C and percent mole fraction of peroxide detected at the ring electrode as a function of pyrolysis temperature of the catalyst. Experiments performed in O2 saturated 0.1 M NaOH at 900 rpm and 20 mV/s. ERing = 1.3 V vs. RHE.
116
The onset potential for ORR in 0.1 M NaOH is 0.95 V vs. RHE whereas the
corresponding peroxide oxidation current does not begin until 0.8 V. In 0.1 M HClO4, the onset
potential for both ORR and peroxide oxidation is 0.8 V. This is further proof for the instability of
peroxide intermediate on the active site in acidic medium because the weak binding of the H2O2
intermediate on the Fe2+ active site in acidic medium facilitates its desorption into the bulk
electrolyte and subsequently to be detected at the ring electrode (vide infra). This desorption of
H2O2 into the bulk electrolyte due its weak binding on the active site is the primary source of
peroxide detected at the ring in acidic medium. At very high overpotentials (< 0.3 V) the ORR
process in acidic electrolyte becomes very efficient and consequently the peroxide detected
decreases.49-50 The ring current profile in 0.1 M NaOH is more complex and requires careful
considerations as explained in detail in chapter 2. Briefly, in alkaline electrolyte, the peak in ring
current at 0.50 V is due to the 2e¯ O2 reduction process proceeding via an outer-sphere electron
transfer mechanism promoted by the quinone/hydroquinone surface functional groups on the
carbon support. The increase in ring current at potentials below 0.3 V is due to reorientation of
the water molecules on the surface as the electrode traverses into the hydrogen under-potentially
deposited (UPD) region.51-52 In acidic medium neither water is involved in the ORR process nor
is the proton discharge on the Fe2+ site evidenced. Therefore no such increase in ring current is
observed at potentials below 0.3 V vs. RHE. Finally a shoulder in the ring current in 0.1 M
NaOH from 0.6 V to 0.7 V is observed. This shoulder was shown to be a characteristic signature
for the 2e outer-sphere electron transfer ORR mechanism (vide infra) promoted by the
specifically adsorbed hydroxide species on the Fe2+ active site as detailed in chapter 2.
117
The nature of the active site obtained upon pyrolysis and the fundamental origin of the
activity is investigated here in detail. Figure 4.3(a) shows the square wave voltammetry (SWV)
of non-heat treated Fe(III)TPPCl/C in argon saturated 0.1 M NaOH and 0.1 M HClO4
electrolytes and Figure 4.3(b) shows the corresponding cyclic voltammetry (CV) profile. As seen
in both the SWV and the CV profiles the redox transition involving the metal center Fe2+/3+ is
observed at 0.314 V vs. RHE in 0.1 M NaOH and at 0.155 V vs. RHE in 0.1 M HClO4. 11,20,53 As
seen in the SWV profile, the peaks at 1.260 V in 0.1 M HClO4 and 1.508 V in 0.1 M NaOH
corresponds to the one electron redox transition involving the delocalized π-electron system in
the macrocycle ligand.53 Both the anodic and cathodic Fe2+/3+ redox peaks in the CV do not
E [V vs. RHE]0.0 0.3 0.6 0.9 1.2 1.5 1.8
i [A
/cm
2 geo]
0
1e-3
2e-3
3e-3
4e-3
0.1 M NaOH0.1 M HClO4
(b) Cyclic Voltammetry
E Vs RHE
-0.2 0.0 0.2 0.4 0.6 0.8 1.0
0.1
mA
/cm
2 geo
FeII /FeIII
LigandOxidation
(a) Square Wave Voltammetry
FeTPPCl/C Non-pyrolyzed
Figure 4.3: (a) SWV and (b) CV of as-received Fe(III)TPPCl/C. Experiments were performed in argon saturated 0.1 M NaOH and 0.1 M HClO4. SWV parameters: 5 mV step potential, 20 mV amplitude, and 10 Hz frequency. CV: 20 mV/s.
118 exhibit any peak separation indicating that the redox couple is confined to the electrode surface.
The Fe2+/3+ redox peak in the CV profile is superimposed on a large capacitive current due to the
high carbon support content in the catalyst whereas the SWV profile exhibits only the Faradaic
charge transfer processes.
As observed in the literature,11 CV of the heat-treated catalysts do not yield useful
information since in most cases after heat-treatment the double-layer charging current
overwhelms the Faradaic currents. In order to overcome this limitation, SWV was performed as
shown in Figure 4.4 in order to understand the evolution of the active site in these catalysts with
increasing heat treatment temperatures. Figure 4.4 is divided into two panels due to different
potential ranges that are required for baseline correction of SWV profiles. Panel 1 shows the
evolution of redox couples below 1.0 V whereas the panel 2 shows above 1.0 V vs. RHE in
argon saturated 0.1 M NaOH electrolyte. After 300°C pyrolysis the Fe2+/3+ redox couple at 0.31
E [V vs. RHE]1.0 1.2 1.4 1.6 1.8
Cur
rent
[A
]
-1e-4
0
1e-4
2e-4
3e-4
4e-4
5e-4
E [V vs. RHE]0.0 0.2 0.4 0.6 0.8 1.0
Cur
rent
[A
]
-1e-4
0
1e-4
2e-4
3e-4
4e-4
5e-4
300oC 300oC
500oC500oC600oC
800oC
900oC
600oC800oC
900oC
(a) Square Wave Voltammetry in 0.1 M NaOH at 10 Hz
Fe2+/Fe3+ PeakPotential
Pyrolysis Temperature [oC]0 200 400 600 800 1000
EP
eak [
V]
0.3
0.6
0.9
1.2
Panel 1 Panel 2
Figure 4.4: SWV profiles of FeTPP/C catalyst as a function of heat treatment temperature. Inset shows the peak potential of the Fe2+/3+ redox couple as a function of the heat treatment temperature. All SWV experiments were performed in argon saturated 0.1 M NaOH electrolyte with a step potential of 5 mV, amplitude of 20 mV, and scan frequency of 10 Hz.
119 V and the ligand oxidation peak at 1.5 V are similar to the case of non-heat treated catalyst
shown in Figure 4.3. After heat treatment at 500°C and 600°C the magnitude of the peak currents
decrease significantly due to the sublimation of some fraction of the metal macrocycle complex.
More importantly, the Fe2+/3+ peak potentials have shifted to more anodic potentials. As seen in
Panel 1 of Figure 4.4, after pyrolysis at 500°C and 600°C the peak potential (Ep) of the Fe2+/3+
couple has shifted to 0.405 V and 0.427 V respectively. Correspondingly, the ligand oxidation
peak at ~1.5 V, shown in Panel 2 of Figure 4.4, decreases in magnitude significantly due to the
sublimation and/or destruction of the macrocycle. The ligand oxidation peak is not observed after
heat treatment at temperatures greater than 600°C indicating that the delocalized π-electron
system of the macrocycle does not survive such high temperatures on the carbon support. After
800°C pyrolysis, Fe2+/3+ couple has shifted more anodically to Ep = 0.48 V. As shown in Panel 2
the most interesting observation is that after pyrolysis at 600°C a new shoulder begins to emerge
at ~1.2 V. This shoulder resolves into a clear peak after 800°C pyrolysis with a peak potential of
1.25 V vs. RHE. This new peak is attributed to an anodic shift in the Fe2+/3+ redox couple based
on the in situ XANES experimental results shown later (vide infra). This indicates the presence
of two Fe2+/3+ redox couples, one at a low potential (~0.48 V) and another at a high potential
(1.25 V). The anodic shift in Fe2+/3+ redox peak potentials shown in Figure 4.4 (inset) clearly
indicates that after 600°C pyrolysis the metal center exists in two redox environments.
Qualitatively similar behavior was observed in acidic medium as shown in Figure 4.5. It is
interesting to note that after heat treatment at 600°C the Fe2+/3+ redox peak potential in 0.1 M
HClO4 was shifted to an anodic potential of only 0.80 V vs. RHE compared to the 1.25 V in
alkaline medium. Thus there are two reasons for the lower ORR overpotential for FeTPP/C
catalyst in alkaline medium: 1) The higher redox potential of the Fe2+/3+ metal center in alkaline
120
medium and 2) the improved stability of the peroxide intermediate on the active site. This
translates into efficient 4e¯ ORR reaction with lower overpotential in alkaline medium compared
to acidic medium.
4.3.2. X-ray Absorption Spectroscopy:
4.3.2.1. EXAFS: Figure 4.6 shows the representative in situ EXAFS spectra of FeTPP/C catalyst
pyrolyzed at 300°C and 800°C. Spectra were taken at a potential of 0.1 V vs. RHE in O2
saturated 0.1 M NaOH electrolyte. Table 4.2 shows the corresponding EXAFS fit results for the
as-received macrocycle and the catalysts pyrolyzed at various temperatures. The as-received
FeTPPCl compound exhibits Fe-N4 coordination environment at a Fe-N bond length of 2.054 Å
and Fe-Cl bond at 2.228 Å in good agreement with the literature.54-56 Upon adding the carbon
E [V vs. RHE]
-0.10 0.00 0.10 0.20 0.30 0.40 0.50
Cur
rent
[A]
-1e-4
0
1e-4
2e-4
3e-4
4e-4
5e-4
6e-4
E [V vs. RHE]
0.50 0.60 0.70 0.80 0.90 1.00
Cur
rent
[A]
-1e-5
0
1e-5
2e-5
3e-5
4e-5
5e-5
300oC
600oC
500oC
800oC
900oC
600oC
500oC
800oC
300oC
Fe2+/Fe3+ PeakPotential
Pyrolysis Temperature [oC]
0 200 400 600 800 1000
EP
eak
[V v
s. R
HE
]
0.0
0.2
0.4
0.6
0.8
1.0
Panel 1 Panel 2
Figure 4.5: SWV profiles of FeTPP/C catalyst as a function of pyrolysis temperature. Inset shows the peak potential of the Fe2+/3+ redox couple as a function of the pyrolysis temperature. Experiments performed in argon saturated 0.1 M HClO4 with 5 mV step potential, 20 mV amplitude, and 10 Hz scan frequency. For the as-received FeTPPCl in 0.1 M HClO4 electrolyte, ligand oxidation peak was observed at potentials greater than 1.2 V vs. RHE and is not shown in the plot below. Note that the scaling for the left and right current axes is different.
121
support and pyrolyzing to 300°C temperature, the Cl atom is replaced with an O atom at a bond
length of 2.055 Å (ex situ) whereas the Fe-N4 environment is preserved. In situ experiments in
0.1 M NaOH show that (Table 4.2) the Fe-N bond length changes from 1.996 Å at 0.10 V to
2.059 Å at 0.90 V. At 0.10 V the iron metal center is in the reduced valence state (Fe2+)
surrounded by four nitrogen atoms and no oxygen ligand at the axial position. At 0.90 V the iron
metal center is oxidized (Fe3+) and has an O atom in the axial position.21 This indicates that in
order to accommodate the O atom in the axial position, the Fe-N4 bonds in the square-planar
300oC
R [Å]0 1 2 3 4 5 6
| χχ χχ(R
)| [Å
-3]
0.1VFit
800oC
R [Å]0 1 2 3 4 5
| χχ χχ(R
)| [Å
-3]
0.1VFit
Fe-N4
Fe-O
Fe-N4
Fe-OMetallic Fe
Figure 4.6: Fe K-edge non-phase corrected Fourier transformed EXAFS spectra of FeTPP/C catalysts pyrolyzed at 300°C (Top) and 800°C (Bottom) temperatures. EXAFS experiment performed under in situ conditions of O2 saturated 0.1 M NaOH electrolyte at 0.1 V vs. RHE. See Table 4.2 for the corresponding EXAFS fit results.
122 environment show a minor expansion or this could also imply a potential dependent out of plane
movement of the iron atom in the porphyrin cavity. Increasing the pyrolysis temperature to
800°C, the Fe-N4 environment is still preserved. The change in Fe-N bond length in varying the
potential from 0.10 V to 0.90 V is 1.976 Å to 1.990 Å. As seen in Figure 4.6, metallic iron
particles begin to form due to the decomposition of some fraction of Fe-N bonds. So after 800°C
pyrolysis Fe-N4 and metallic Fe begin to coexist. However, it is known that these metallic iron
particles are inactive as they are encapsulated by a layer of graphite.7,57-58 This excludes the
direct participation of the metallic iron content in any electrochemical reactions. As seen in
Table 4.2, further increase in the pyrolysis temperature leads only to the growth of metallic iron
clusters at the cost of the Fe-N bonds.
Table 4.2: In situ Fe K-edge EXAFS fit results for FeTPP/C catalysts heat treated (HT) at 300°C and 800°C. Data collected as a function of potential in O2 saturated 0.1 M NaOH. Coordination number (N) and phase-corrected bond length (R) in angstrom are shown for each interaction. Also shown are the Debye-Waller factor (σ2) and edge shifts (E0).
E [V] Fe-N N (R[Å])
Fe-Cl N (R[Å])
Fe-Fe N (R[Å])
σ2 (Fe-N) Eo [eV]
FeTPPCl – As Received
Ex situ 4.01 (2.054) 0.99 (2.228) -- 0.005 1.38
E [V] Fe-N N (R[Å])
Fe-O N (R[Å])
Fe-Fe N (R[Å])
σ2 (Fe-N) Eo [eV]
FeTPP/C HT300°C
0.10 V 4.03 (1.996) 0.10 (2.055) -- 0.005 -4.34 0.90 V 4.00 (2.059) 1.00 (1.873) -- 0.0009 -5.23
FeTPP/C HT800°C
0.10 V 4.00 (1.976) 0.61 (1.797) 0.79 (2.569) 0.0023 -6.97 0.90 V 4.00 (1.990) 2.20 (1.848) 1.23 (2.576) 0.0035 -9.42
FeTPP/C HT900°C
0.10 V 1.00 (1.990) 0.61 (1.797) 4.20 (2.568) 0.008 -4.72 0.90 V 0.95 (2.008) 1.20 (1.845) 4.15 (2.572) 0.008 8.00
123
4.3.2.2. XANES: X-ray absorption near edge spectra (XANES) region is a local atomic probe that
is sensitive to the formal oxidation state and coordination geometry of the metal center. Figure
4.7(a&c) show the normalized Fe K-edge XANES region and the corresponding first derivative,
respectively, for the FeTPP/C catalyst heat treated at 300°C. Ex situ XANES spectra of standard
macrocyclic compounds iron (II) phthalocyanine and the as-received iron (III)
tetraphenylporphyrin chloride are also shown in Figure 4.7(a&c) for comparison. In Fe(II)Pc the
Fe2+ center is in a square planar environment coordinated to four nitrogen atoms and with no
E-Eo [eV]
-20 0 20 40 60 80
Inte
nsity
[a.u
]
Fe(II)Pc
0.10 V
0.90 V
Fe(III)TPPCl
E-Eo [eV]
0 10 20 30 40 50
Inte
nsity
[a.u
]
E-Eo [eV]-20 0 20 40 60 80
Inte
nsity
[a.u
]
E-Eo [eV]0 10 20 30 40 50
Inte
nsity
[a.u
]
Fe(III)TPPCl
0.10 V
0.90 V
Fe Metal
Fe(III)TPPCl
0.10 V
0.90 V
Fe Metal
1s to 3dtransition
at 7112.5 eV
1s to 4pztransition
at 7117 eV
(a)
(d)
Fe(II)Pc
0.10 V
0.90 V
Fe(III)TPPCl
7117 eV7112.5 eV
7112.5 eV7112 eV(b)
(c)
Figure 4.7: Fe K-edge in situ normalized XANES region of FeTPP/C catalyst heat treated at (a) 300°C and (b) 800°C. Corresponding first derivative XANES are shown in (c) and (d) respectively. Shown also for comparison are the ex situ spectra of iron (II) phthalocyanine - Fe(II)Pc, iron (III) tetraphenylporphyrin chloride - Fe(III)TPPCl and Fe Metal (4µm thick iron metal foil). In situ experiments were performed in argon saturated 0.1 M NaOH at 0.1 V and 0.9 V.
124
oxygen atom at the axial position. Due to the center-of-symmetry in this complex only the dipole
allowed transition at 7117 eV (Fe 1s to 4pz) is observed.21 In Fe(III)TPPCl besides the four
nitrogen atoms in the square planar environment the axial oxygen ligand coordinated to the metal
center disrupts the center-of-symmetry slightly and gives rise to the forbidden pre-edge
electronic transition at 7112.5 eV (Fe 1s to 3d).21 Comparing the XANES region of these two
standard compounds to the FeTPP/C catalyst (pyrolyzed at 300°C) leads to the following
conclusion. The metal center in the catalyst, in terms of its oxidation state and coordination
geometry, is reminiscent of Fe(II)Pc at 0.10 V and Fe(III)TPPCl at 0.90 V vs. RHE. As shown in
Figure 4.7(b&d), after pyrolyzing the FeTPP/C catalyst to 800°C the XANES region at both 0.10
V and 0.90 V is overwhelmed by the characteristics of metallic iron content. Figure 4.8 further
corroborates the growth of metallic iron content at heat treatment temperatures ≥700°C. In order
to avoid the influence of metallic iron content and understand the underlying redox processes,
E-Eo [eV]-20 0 20 40 60 80
Inte
nsity
[a.u
]
Fe(III)TPPCl
300oC
500oC
700oC
800oC
1000oC
Fe Metal
E [V vs. RHE]
0 200 400 600 800 1000 1200F
ract
ion
of F
e V
alen
ce S
tate
s
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Fe0
Fe3+
(a) Ex situ XANES
(b) Linear Combination Fitting
Figure 4.8: (a) Ex situ Fe K-edge XANES region in FeTPP/C catalyst pyrolyzed at various temperatures, and (b) corresponding LCA analysis.
125
Linear Combination Analysis (LCA) fitting was performed. LCA fitting determines the fraction
of various valence states of the metal center present in the catalyst. For the 300°C heat treated
catalyst shown in Figure 4.9(a) only Fe2+ and Fe3+ components were used to fit the data. As
predicted from SWV experiments shown above, a redox transition between Fe2+ and Fe3+ is
clearly observed in Figure 4.9(a) indicating that after 300°C heat treatment the two valence states
are mutually exclusive of each other. Interestingly, as shown in Figure 4.9(a) the transition point
of 0.7 V between the two redox states marks the ORR onset potential for the 300°C heat treated
catalyst. For the 800°C pyrolyzed sample, zero valent iron component was also included in the
fit as shown in Figure 4.9(b). Besides this minor metallic iron content, clearly the Fe2+ and Fe3+
fractions co-exist with each other up to a potential of 1.0 V vs. RHE. The Fe2+ fraction shown in
Figure 4.9(b) corresponds to the reduced component of the high potential Fe2+/3+ redox couple
observed at 1.25 V vs. RHE in Figure 4.4(b). This clearly shows that after 800°C heat treatment
(b) Pyrolysis at 800oC
E [V vs. RHE]
0.2 0.4 0.6 0.8 1.0 1.2 1.4
Fra
ctio
n of
Fe
Val
ence
Sta
tes
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
Fe0
Fe2+
Fe3+
(a) Pyrolysis at 300oC
E [V vs. RHE]
0.0 0.2 0.4 0.6 0.8
Fra
ctio
n of
Fe
Val
ence
Sta
tes
0.0
0.2
0.4
0.6
0.8
1.0
Fe2+
Fe3+
Figure 4.9: LCA fitting of FeTPP/C showing the fraction of various iron oxidation states of iron present in the sample after (a) 300°C and (b) 800°C heat treatment temperatures. Data obtained in argon saturated 0.1 M NaOH as a function of potential at Fe K-edge.
126 the anodic shift of Fe2+/3+ redox couple ensures the availability of reduced Fe2+ component at a
higher potential to bind molecular O2 and initiate the reduction process.
4.3.2.3 Delta-Mu (∆µ) studies: Both EXAFS and XANES being bulk averaged techniques
overlook the critical electrochemical reactions occurring on the catalyst surface. Delta Mu (∆µ)
is a surface sensitive, spectral subtraction technique where the bulk structure of the catalyst is
effectively removed leading to information on the nature and site-specificity of the surface
adsorbates.42,44,59-60 Figure 4.10(a) shows the XANES region of FeTPP/C catalyst pyrolyzed at
300°C taken at two different potentials of 0.1 V and 0.9 V vs. RHE. As discussed above, at 0.1 V
the metal center exists in the reduced Fe2+ state with no adsorbates (neither hydrides nor oxides)
at the axial position and immediate coordination environment is reminiscent of the iron (II)
phthalocyanine complex where the pre-edge peak is muted. At 0.9 V the metal center is oxidized
to the Fe3+ state with an oxygen atom at the axial position and the metal coordination
environment is similar to that of the original porphyrin complex where the pre-edge Fe(1s→3d)
forbidden transition at 7112.5 eV is observed. The delta mu spectra is obtained by subtracting the
XANES regions according to the equation ∆µ = µ(0.90 V) - µ(0.10 V). In the delta-mu spectra of
Figure 4.10(a), the positive peak feature (boxed portion) indicates the difference in absorption
probability at the pre-edge energy (7112.5 eV). With the information known above, this positive
peak feature could be safely assigned as a signature for the existence of the metal center in a
centro-symmetric environment undergoing a transition from Fe2+-N4 coordination geometry at
0.10 V to O-Fe3+-N4 coordination geometry at 0.90 V. Ensuing this positive peak is a steep
negative dip featuring a split peak below 20 eV. This negative dip characterizes charge transfer
from the metal center to the adsorbed oxygen species.61 As shown in Figure 4.10(b), the XANES
127
spectra of the 800°C pyrolyzed catalyst is predominantly characteristic of metallic iron at both
0.10 V and 1.10 V that precludes proper analysis of the active site. However, careful analysis of
the corresponding delta mu spectra clearly indicates the positive peak feature at the pre-edge
energy indicative of the fact that the active site is Fe2+-N4 where the metal center is in a centro-
symmetric environment which is mildly disrupted by the presence of an axial oxygen atom. This
clearly indicates that the Fe2+ metal center surrounded by four nitrogen atoms is the active site
that binds oxygen in the axial position and the redox transition from Fe3+ to Fe2+ triggers oxygen
adsorption according to the redox mechanism.62 Figure 4.11 shows the theoretical delta mu (∆µt)
spectra obtained using FEFF8.0 code.45 The structural models used are shown in the insets of
Energy [eV]
-40 -20 0 20 40 60 80
Inte
nsity
[a.u
]
XANES @ 0.10 VXANES @ 0.90 VDelta Mu
Energy [eV]
-40 -20 0 20 40 60 80 100
Inte
nsity
[a.u
]
XANES @ 0.10 VXANES @ 1.10 VDelta Mu
(a) 300oC Pyrolysis (b) 800oC Pyrolysis
7112.5 eV 7112.5 eV
Figure 4.10: Experimental XANES (µ) and Delta-Mu (∆µ) signatures of FeTPP/C catalyst heat treated at (a) 300°C and (b) 800°C. ∆µ signatures were obtained by subtracting the XANES regions according to ∆µ = µ(0.90 (or) 1.10 V) - µ(0.10 V). Experiments were conducted at Fe K-edge under in situ conditions in argon saturated 0.1 M NaOH electrolyte. Vertical dotted line indicates the pre-edge position at 7112.5 eV and the boxed region focusses the pre-edge region. Delta-Mu spectra have been multiplied by a factor of 5 for visual comparison of the line shapes.
128
Figure 4.11. Only the atoms encircled were used in the theoretical FEFF8.0 modeling. As
outlined in the experimental section, these spectra were calculated from Fe-N4-Cx models
derived from prior crystallographic data adjusted to the EXAFS fitting results according to the
relation: ∆µt = µ(Oads-Fe-N4-C) - µ(Fe-N4-C) where the oxide species (Oads or OHads) is in a
specific binding site. In all cases the positive peak feature at the pre-edge energy was observed
only when the adsorbed oxygen atom was placed in the axial position of the metal center. No
E-Eo [eV]
-20 0 20 40 60
E-Eo [eV]
-20 0 20 40 60
(a) 300oC Pyrolysis (b) 800oC Pyrolysis
Graphitic Cluster Monovacancy Divacancy N-doped Active site
(c)
Figure 4.11: Theoretical FEFF8 ∆µ = µ(Fe-N4-Cx-Oads) - µ(Fe-N4-Cx) signatures obtained for (a) 300°C and (b) 800°C pyrolysis conditions. The insets in (a&b) show the corresponding structural models utilized. Only the atoms encircled in these structural models were used for FEFF8 simulation. (c) Schematic illustration of the mono- and di-vacanct defective pockets in amorphous carbon acting as anchors for active site formation during pyrolysis. Color Codes: Pink – Fe, Blue – N, and Black – C.
129 successful theoretical delta mu fits could be obtained for Fe-N coordination numbers less than
four or for oxygen adsorption modes other than at the axial position. (See Figure 4.12 – for a
compilation of some of the unsuccesful delta mu fits).
As shown in the inset of Figure 4.11(a), after 300°C pyrolysis the immediate coordination
environment of the original precursor porphyrin macrocycle is clearly retained. This corresponds
to a FeN4C12 cluster where the metal center is coordinated to four nitrogen atoms and each
nitrogen atom in turn bonded to two carbon atoms. Finally inclusion of the four methine carbon
bridges gives a tally of C12. Figure 4.11(b) shows the theoretical delta mu spectrum that likely
mimics the line shape of the experimental delta mu shown in Figure 4.10(b) for the 800°C heat
treated catalyst. As shown in the inset of Figure 4.11(b) the molecular cluster used to simulate
the theoretical delta mu spectrum consisted of FeN4C10. While compositionally this cluster is not
very different from the 300°C pyrolyzed sample, the immediate coordination environment of the
metal active site after 800°C heat treatment is found to be reminiscent of the crystallographic
atomic defects such as the divacancy on the graphitic surfaces. This is schematically depicted in
Figure 4.11(c).63-64 Atomic defects such as monovacancy and divacancy on microporous carbon
and carbon nanotubes are known to thermodynamically exist or can be induced via various
chemical or physical processes.65-69 The presence of a monovacancy in carbon atom creates three
dangling bonds whereas divacancies create four dangling bonds. These dangling bonds gives rise
to unsaturated valences that then become favorable for nitrogen doping.64 Consequently these
nitrogen doped sites constitute defective pockets for metal coordination. Such atomic vacancies
are either already present on the graphite surfaces or can be created during the heat treatment
step.65,67 During heat treatment under inert atmosphere, carbothermic reaction causes desorption
of oxygen functional groups along with creation of vacancy defects.67 It has also been observed
130 earlier that heat treated Fe-Nx catalysts showed higher activity when supported on carbon that
was previously treated in concentrated inorganic acids.35 Since such acid treatment steps lead to
oxygen functional groups on the carbon support, it is likely that this yields higher number of
defective sites during subsequent heat treatment. These defective pockets are likely the favorable
zones for anchoring FeN4 active sites. Although the Fe-N4 site in the divacant defective pocket is
similar to the Fe-N4 site in the precursor porphyrin cavity, it should the noted that the original
Fe-N4 moiety is not preserved. Presumably, the low concentration of the defective sites limits the
active site density. Therefore increasing the defect density likely holds the key to increasing the
metal loading in this class of catalysts. It is noted that the defect sites in surface chemistry
generally exhibit higher chemical potential and this increase in chemical potential translates into
an anodic shift in the redox potential of the metal center.65,70
E-Eo [eV]
0 10 20 30 40 50
Inte
nsity
[a.u
]
Equatorial-O FeN3C6
E-Eo [eV]
0 10 20 30 40 50
Equatorial-O FeN2C4
E-Eo [eV]
0 10 20 30 40 50
Monovacant DefectAxial O-FeN3C9
Figure 4.12: Some of the unsuccessful delta mu fits
131 4.3.4 ORR Reaction Mechanism
The following observations are made in an attempt to correlate the SWV, XANES and
the delta mu results. The low potential Fe2+/3+ redox couple is characteristic of the original FeN4
porphyrin moiety whereas the high potential Fe2+/3+ redox couple is characteristic of the FeN4
cluster found in the microporous carbon defect sites. At a pyrolysis temperature of 600°C both
these sites seem to coexist and the higher ORR activity obtained for heat treatment temperatures
≥600°C is due to the high potential Fe2+/3+ redox couple seated in the defective pockets. Based on
the above experimental results the following reaction scheme is proposed for ORR in dilute
alkaline medium on heat treated FeTPP/C catalyst. A similar set of reactions can be developed,
mutatis mutandis, for dilute acidic medium.7,49-50,71 Equation (4.1) below shows the redox
reaction involving the metal center that is a prerequisite for adsorption of molecular oxygen on
the active site [N4-FeII-OH¯].
[N4-FeIII-OH] + e → [N4-FeII-OH¯] (4.1)
While the FeII valence state favors a square-planar tetracoordinate environment, the high
potential of this redox reaction causes the OH¯ species to poison the active site at the axial
position. This poisonous OH¯ species prevents direct molecular adsorption of O2 on the active
site. Further, the adsorbed OH¯ species mediates the 2e¯ outer-sphere electron transfer reduction
of solvated O2 molecule as shown below in the reaction schemes (4.2a)-(4.2c).
[N4-Fe(II)-OH ] + [O2.(H2O)n]aq + e → [N4-Fe(II)-OH ] + (HO2•)ads + OH + (H2O)n-1 (4.2a)
(HO2•)ads + e → (HO2¯)ads (4.2b)
(HO2¯)ads → (HO2¯)aq (4.2c)
132 In equation (4.2a), [O2.(H2O)n]aq cluster represents the solvated molecular oxygen. The adsorbed
OH¯ species acts as an outer-sphere bridge between the FeII active site and the solvated
molecular oxygen. The (HO2¯)aq species formed in equation (4.2c) via this outer-sphere
mechanism gives rise to the shoulder in ring current (Figure 4.1(b)) between the potential regions
0.6 V and 0.7 V vs. RHE in 0.1 M NaOH. It should be noted that the case of outer-sphere
mechanism does not arise in acidic medium for reasons explained in chapter 2.
Finally the 4e electrocatalytic inner-sphere electron transfer mechanism is shown in
Figure 4.13 where the molecular O2 displaces the OH species and chemisorbs directly on the
FeII active site.71 Based on our experimental results the electrocatalytic process in Figure 4.13 is
shown to take place via a redox mechanism involving the FeII/III couple.62 Once molecular O2
adsorbs on the FeII active site, the reaction proceeds to the ferrous-hydroperoxyl adduct via the
O2H2O
e-
+ OH -
H2O
e-
e-e-
-OH -
FeII
Active Site
O
H
FeIII
Ferric-Hydroperoxyl
O
OH
FeIII
FerricHydroxyl
O
H
+ 2OH -
FeIII
FerricSuperoxo
O
O
FeII
Adsorbed O2
OO
FeII
Ferrous-Hydroperoxyl
O
OH
Figure 4.13: Catalytic cycle showing the redox mechanism involved in ORR on heat treated iron porphyrin macrocycles in dilute alkaline medium. Nitrogen atoms in the square planar positions have been omitted for clarity.
133 superoxo and the ferric hydroperoxyl states. The ferrous-hydroperoxyl adduct is very critical
since its stability determines the product distribution. Peroxide anion (HO2¯) is a stable
intermediate. So, weak binding of HO2¯ on the active site will lead to its desorption into the bulk
electrolyte. As mentioned above, the positive charge on the FeII active site and the anionic nature
of the peroxide intermediate (HO2¯) gives rise to a Frumkim-type double-layer effect where the
electrostatic attractive interaction stabilizes the ferrous-hydroperoxyl adduct. This ensures that
the catalytic cycle shown in Figure 4.13 regenerates the active site via the formation of ferric-
hydroxyl species. However, in acidic medium the analogous ferrous-hydrogen peroxide adduct is
FeII-(OHOH).71 Clearly the absence of any Frumkin-type electrostatic effect causes desorption of
the stable peroxide intermediate (H2O2) into the bulk electrolyte. This leads to higher peroxide
yield in acidic medium as shown above in Figure 4.1.
4.5 Conclusions:
A combination of square wave voltammetry, XANES and delta mu studies has been used
to unravel the nature of the active site, fundamental origin of ORR activity, and plausible reasons
for the low density of active sites in heat treated iron porphyrin catalysts for oxygen reduction in
aqueous electrolytes. In alkaline medium, Fe-N4 sites promote direct molecular O2 adsorption
(inner-sphere process) and suppress the outer-sphere electron transfer (although not completely
eliminated). The shift in Fe2+/3+ redox transition to higher potential of 1.25 V vs. RHE in 0.1 M
NaOH electrolyte and 0.8 V vs. RHE in 0.1 M HClO4 is found to be the fundamental reason for
the increased ORR activity upon heat treatment. No spectroscopic proof for the involvement of
FeIV valence state of the metal center was observed which is in good agreement with the
computational study of Anderson et al71. While the Fe-N4 coordination is preserved upon heat
134 treatment, it is observed that the divacant defective centers on the amorphous carbon support act
as anchors for the active sites. Such divacant atomic defects could already be present on the
carbon surface or can be generated during heat treatment step due to carbothermic reactions. The
low density of these defect sites is likely the reason for the limitation in the active site density.
Increasing the defect site density without penalizing the stability of the carbon support is likely
to increase the active site density and thus decrease the thickness of the electrode made out of
these catalysts for fuel cell applications. Studies correlating the redox potential of the metal
center to more fundamental physical parameters such as the Lewis basicity/acidity of the carbon
basal/edge planes, chemical potential of defective centers need to be developed and will be of
interest for future studies. Future studies will involve understanding the durability issues related
to this class of catalysts along with efforts to increase the active site density and fuel cell cathode
performance.
4.6 Acknowledgements:
The authors deeply appreciate financial assistance from the Army Research Office under the
Single Investigator grant. Use of the National Synchrotron Light Source (NSLS), Brookhaven
National Laboratory (BNL), was supported by the U.S. Department of Energy, Office of Basic
Energy Sciences. Support from beamline personnel Drs. Syed Khalid and Nebojsa Marinkovic
(X19A) are gratefully acknowledged.
135 4.7 References: (1) Jasinski, R. J. Nature (London, U. K.) 1964, 201, 1212. (2) Yeager, E. Electrochim. Acta 1984, 29, 1527. (3) Zagal, J. H. Coord. Chem. Rev. 1992, 119, 89. (4) Macrocycles; Zagal, J. H., Ed., 2003; Vol. 2. (5) Wang, B. J. Power Sources 2005, 152, 1. (6) Zagal, J. H. In N4-Macrocyclic Metal Complexes; Zagal, J. H., Bedioui, F., Dodelet, J.-P., Eds.; Springer: 2006. (7) Dodelet, J.-P. In N4-Macrocyclic Metal Complexes; Zagal, J. H., Bedioui, F., Dodelet, J.-P., Eds.; Springer: 2006, p 83. (8) Ziegelbauer, J. M.; Olson, T. S.; Pylypenko, S.; Alamgir, F.; Jaye, C.; Atanassov, P.; Mukerjee, S. J. Phys. Chem. C 2008, 112, 8839. (9) Franke, R.; Ohms, D.; Wiesener, K. J. Electroanal. Chem. Interfacial Electrochem. 1989, 260, 63. (10) Gruenig, G.; Wiesener, K.; Gamburzev, S.; Iliev, I.; Kaisheva, A. J. Electroanal. Chem. Interfacial Electrochem. 1983, 159, 155. (11) Tanaka, A.; Gupta, S. L.; Tryk, D.; Fierro, C.; Yeager, E. B.; Scherson, D. A. In Proc. - Electrochem. Soc. 1992; Vol. 92-11, p 555. (12) Van Veen, J. A. R.; Van Baar, J. F.; Kroese, K. J. Journal of the Chemical Society, Faraday Transactions 1: Physical Chemistry in Condensed Phases 1981, 77, 2827. (13) Vasudevan, P.; Santosh; Mann, N.; Tyagi, S. Transition Met. Chem. (London) 1990, 15, 81. (14) Artyushkova, K.; Levendosky, S.; Atanassov, P.; Fulghum, J. Topics in Catalysis 2007, 46, 263. (15) Bouwkamp-Wijnoltz, A. L.; Visscher, W.; Van Veen, J. A. R. Electrochim. Acta 1998, 43, 3141. (16) Bouwkamp-Wijnoltz, A. L.; Visscher, W.; van Veen, J. A. R.; Boellaard, E.; van der Kraan, A. M.; Tang, S. C. J. Phys. Chem. B 2002, 106, 12993. (17) Bouwkamp-Wijnoltz, A. L.; Visscher, W.; Van Veen, J. A. R.; Tang, S. C. Electrochim. Acta 1999, 45, 379. (18) McBreen, J.; O'Grady, W. E.; Sayers, D. E.; Yang, C. Y.; Pandya, K. I. Proc. - Electrochem. Soc. 1987, 87-12, 182. (19) Schulenburg, H.; Stankov, S.; Schuenemann, V.; Radnik, J.; Dorbandt, I.; Fiechter, S.; Bogdanoff, P.; Tributsch, H. J. Phys. Chem. B 2003, 107, 9034. (20) Bae, I. T.; Tryk, D. A.; Scherson, D. A. J. Phys. Chem. B 1998, 102, 4114. (21) Kim, S.; Bae, I. T.; Sandifer, M.; Ross, P. N.; Carr, R.; Woicik, J.; Antonio, M. R.; Scherson, D. A. J. Am. Chem. Soc. 1991, 113, 9063. (22) Kim, S.; Tryk, D.; Bae, I. T.; Sandifer, M.; Carr, R.; Antonio, M. R.; Scherson, D. A. Journal of Physical Chemistry 1995, 99, 10359. (23) Lefevre, M.; Dodelet, J. P.; Bertrand, P. J. Phys. Chem. B 2002, 106, 8705. (24) Lefevre, M.; Dodelet, J. P.; Bertrand, P. J. Phys. Chem. B 2000, 104, 11238. (25) Lefevre, M.; Dodelet, J.-P. Electrochim. Acta 2003, 48, 2749. (26) Lefevre, M.; Proietti, E.; Jaouen, F.; Dodelet, J.-P. Science (Washington, DC, U. S.) 2009, 324, 71.
136 (27) Lalande, G.; Cote, R.; Guay, D.; Dodelet, J. P.; Weng, L. T. Electrochim. Acta 1997, 42, 1379. (28) Lalande, G.; Faubert, G.; Cote, R.; Guay, D.; Dodelet, J. P.; Weng, L. T.; Bertrand, P. J. Power Sources 1996, 61, 227. (29) Alves, M. C. M.; Dodelet, J. P.; Guay, D.; Ladouceur, M.; Tourillon, G. Journal of Physical Chemistry 1992, 96, 10898. (30) Faubert, G.; Cote, R.; Dodelet, J. P.; Lefevre, M.; Bertrand, P. Electrochim. Acta 1999, 44, 2589. (31) Faubert, G.; Lalande, G.; Cote, R.; Guay, D.; Dodelet, J. P.; Weng, L. T.; Bertrand, P.; Denes, G. Electrochim. Acta 1996, 41, 1689. (32) He, P.; Lefevre, M.; Faubert, G.; Dodelet, J. P. Journal of New Materials for Electrochemical Systems 1999, 2, 243. (33) Medard, C.; Lefevre, M.; Dodelet, J. P.; Jaouen, F.; Lindbergh, G. Electrochim. Acta 2006, 51, 3202. (34) Jaouen, F.; Proietti, E.; Lefevre, M.; Chenitz, R.; Dodelet, J.-P.; Wu, G.; Chung, H. T.; Johnston, C. M.; Zelenay, P. Energy & Environmental Science 2011, 4, 114. (35) Time to move beyond transition metal-N-C catalysts for oxygen reduction; Garsuch, A.; Bonakdarpour, A.; Liu, G.; Yang, R.; Dahn, J. R., Eds., 2009; Vol. 5. (36) Jaouen, F.; Lefevre, M.; Dodelet, J.-P.; Cai, M. J. Phys. Chem. B 2006, 110, 5553. (37) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Appl. Catal., B 2005, 56, 9. (38) Bezerra, C. W. B.; Zhang, L.; Lee, K.; Liu, H.; Marques, A. L. B.; Marques, E. P.; Wang, H.; Zhang, J. Electrochim. Acta 2008, 53, 4937. (39) Zagal, J.; Paez, M.; Fierro, C. Proc. - Electrochem. Soc. 1987, 87-12, 198. (40) Zagal, J. H.; Cardenas-Jiron, G. I. Journal of Electroanalytical Chemistry 2000, 489, 96. (41) Zagal, J. H.; Gulppi, M.; Isaacs, M.; Cardenas-Jiron, G.; Aguirre, M. J. Electrochim. Acta 1998, 44, 1349. (42) Arruda, T. M.; Shyam, B.; Lawton, J. S.; Ramaswamy, N.; Budil, D. E.; Ramaker, D. E.; Mukerjee, S. J. Phys. Chem. C 2010, 114, 1028. (43) Newville, M. Journal of Synchrotron Radiation 2001, 8, 322. (44) Arruda, T. M.; Shyam, B.; Ziegelbauer, J. M.; Mukerjee, S.; Ramaker, D. E. Journal of Physical Chemistry C 2008, 112, 18087. (45) Ankudinov, A. L.; Ravel, B.; Rehr, J. J.; Conradson, S. D. Physical Review B: Condensed Matter and Materials Physics 1998, 58, 7565. (46) Bard, A. J.; Faulkner, L. R. In Electrochemical Methods: Fundamentals and Applications; Second ed.; John Wiley & Sons, Inc.: 2000, p 571. (47) Krischer, K.; Savinova, E. R. Handbook of Heterogeneous Catalysis (2nd Edition) 2008, 4, 1873. (48) Tarasevich, M. R.; Sadkowski, A.; Yeager, E. In Comprehensive Treatise of Electrochemistry; Conway, B. E., Bockris, J. O. M., Yeager, E., Eds.; Plenum Press: New York, 1983; Vol. 7, p 301. (49) Boulatov, R. In N4-Macrocyclic Metal Complexes; Zagal, J. H., Bedioui, F., Dodelet, J.-P., Eds.; Springer: 2006, p 1.
137 (50) Boulatov, R. In Fuel Cell Catalysis; Koper, M. T. M., Ed.; John Wiley & Sons, Inc.: 2008, p 637. (51) Iwasita, T.; Xia, X. Journal of Electroanalytical Chemistry 1996, 411, 95. (52) Jerkiewicz, G.; Vatankhah, G.; Lessard, J.; Soriaga, M. P.; Park, Y.-S. Electrochim. Acta 2004, 49, 1451. (53) Kadish, K. M.; Van Caemelbecke, E.; Royal, G. In Porphyrin Handbook; Kadish, K. M., Smith, K. M., Guilard, R., Eds.; Academic Press: 2000; Vol. 8, p 1. (54) Hoard, J. L.; Cohen, G. H.; Glick, M. D. J. Am. Chem. Soc. 1967, 89, 1992. (55) Scheidt, W. R.; Finnegan, M. G. Acta Crystallographica, Section C: Crystal Structure Communications 1989, C45, 1214. (56) Paulat, F.; Praneeth, V. K. K.; Naether, C.; Lehnert, N. Inorganic Chemistry 2006, 45, 2835. (57) Gojkovic, S. L.; Gupta, S.; Savinell, R. F. J. Electrochem. Soc. 1998, 145, 3493. (58) Gojkovic, S. L.; Gupta, S.; Savinell, R. F. Journal of Electroanalytical Chemistry 1999, 462, 63. (59) Teliska, M.; Murthi, V. S.; Mukerjee, S.; Ramaker, D. E. Journal of the Electrochemical Society 2005, 152, A2159. (60) Teliska, M.; O'Grady, W. E.; Ramaker, D. E. Journal of Physical Chemistry B 2005, 109, 8076. (61) Lewis, E. A.; Segre, C. U.; Smotkin, E. S. Electrochimica Acta 2009, 54, 7181. (62) Beck, F. Journal of Applied Electrochemistry 1977, 7, 239. (63) Silva, S. R. P. Properties of Amorphous Carbon. [In: EMIS Datarev. Ser., 2003; 29], 2003. (64) Cho, Y. J.; Kim, H. S.; Baik, S. Y.; Myung, Y.; Jung, C. S.; Kim, C. H.; Park, J.; Kang, H. S. J. Phys. Chem. C 2011, 115, 3737. (65) Hahn, J. R.; Kang, H. Physical Review B: Condensed Matter and Materials Physics 1999, 60, 6007. (66) Mielke, S. L.; Troya, D.; Zhang, S.; Li, J.-L.; Xiao, S.; Car, R.; Ruoff, R. S.; Schatz, G. C.; Belytschko, T. Chemical Physics Letters 2004, 390, 413. (67) Luo, Y.; Heng, Y.; Dai, X.; Chen, W.; Li, J. J. Solid State Chem. 2009, 182, 2521. (68) Kondo, T.; Suzuki, T.; Nakamura, J. Journal of Physical Chemistry Letters 2011, 2, 577. (69) Nemes-Incze, P.; Konya, Z.; Kiricsi, I.; Pekker, A.; Horvath, Z. E.; Kamaras, K.; Biro, L. P. J. Phys. Chem. C 2011, 115, 3229. (70) Igami, M.; Nakada, K.; Fujita, M.; Kusakabe, K. Czechoslovak Journal of Physics 1996, 46, 2715. (71) Anderson, A. B.; Sidik, R. A. J. Phys. Chem. B 2004, 108, 5031.
138
Chapter 5
Degradation Mechanism Study of Perfluorinated Proton Exchange Membrane under Fuel
Cell Operating Conditions
5.1 Introduction:
Widespread commercialization of PEMFC is strongly predicated on component costs and
striking an optimum balance between performance and durability. Although extensive research
seeking enhancement in performance of PEMFC’s is available (decreasing noble metal catalyst
loading, improving reactant and catalyst utilization, developing new polymer electrolytes, non-
noble metal catalysts, improving stack and flow field designs) extensive investigations of
durability issues has been relatively recent mainly because of the test duration requirements and
complexity of analysis brought on by existence of parallel processes and inability to perform in-
situ, nondestructive analysis of the key components.1 Since the MEA is the heart of an operating
PEMFC where electrochemical energy conversion takes place, it is more prone to chemical and
electrochemical degradation and the biggest determinant in the extent of losses in fuel cell
performance. Degradation of MEA components are broadly understood to be due to (i)
electrocatalyst sintering via (a) thermal induced coalescence and growth following surface
migration over the carbon support material, and (b) ‘Ostwald Ripening’ which follows a
dissolution-redeposition mechanism, (ii) platinum particle agglomeration triggered by corrosion
of carbon support, (iii) electrocatalyst poisoning, surface segregation and morphology changes
due to presence of strong surface chemisorption by species such as CO, sulfur compounds,
products of methanol oxidation etc.,, (iv) self segregation of elements in a mixed metal oxides or
alloys brought on by potential excursions etc., and (v) degradation of ion conducting compoment
139 including membrane and smaller aggregates present alongside the electrocatalyst in the reaction
layer due to free radical species generated at the interface. All these degradation processes are a
strong function of operating conditions such as temperature, partial pressures, relative humidity,
overpotentials etc.
Perfluorosulphonated Nafion membranes shown in Scheme 5.1 have received priority in
durability studies2-3 although sulfonated nonfluorinated aromatic membranes4 and composite
membranes5 represent one large group of promising candidates.
Apart from physical, and mechanical degradations such as membrane thinning and pinhole
formations at elevated temperature and induced stresses, peroxide led free radical attack and
subsequent degradations are of more immediate importance due to potential of rapid irreversible
damage. Oxygen Reduction Reaction (ORR) is a multi electron transfer process which involves
several elementary steps generating of intermediate species. A more typical scheme representing
the overall oxygen reduction reaction in acid medium is shown below6
O2 O2, ads H2O2, ads H2O H2O2
k1
k2 k3
k4 k5
Scheme 5.1: Chemical Structure of Nafion® 112.
140 A noteworthy feature in the above scheme is that the overall mechanism of direct
electrochemical reduction of O2 to water (‘direct’ 4e- pathway) is a parallel process part of which
proceeds via the formation of H2O2 intermediate (2e- reduction pathway). Formation of H2O2 has
been confirmed using a microelectrode in an operating fuel cell7 and Scherer8 detected the
presence of H2O2 in the outlet stream of operating PEM fuel cells with Nafion membranes.
Besides the formation of peroxide intermediate due to the incomplete reduction of oxygen9-11
(i.e., two electron reduction occurring in a parallel pathway to the four electron reduction) at a
fuel cell cathode, it can also be generated at open circuit conditions with interaction of hydrogen
radicals and crossover oxygen at the anode catalyst-membrane interface8,12-14. It is noted that,
membrane degradation is initiated by the free radicals generated by Fenton type metal cation
catalyzed decomposition of hydrogen peroxide15-18 as shown below
M2+ + H2O2 M3+ + •OH + OH−
M3+ + H2O2 M2+ + •OOH + H+
PEM easily absorbs ionic contaminants due to the stronger affinity of foreign cations with the
sulfonic acid group than that of H+.18 Possible sources of ionic contaminants are the carbon
support, gas diffusion electrodes, humidifying bottles, corrosion of tubing or stack materials, and
other fuel cell hardware. Especially, the iron contamination from the fuel cell end plate has been
found to be the key supplier of foreign ion contamination. 15
From the perspective of peroxide radical attack of the membrane, hydroxyl (OH.) and
hydroperoxyl (OOH.) radicals are the most likely initiators of membrane chemical
decomposition12 as they are some of the most reactive chemical species known.19-21 Radical-
initiated attack leads to the breakage of perfluorocarbon backbone in Nafion membranes and
141 sulphonate groups which directly affect mechanical strength and proton conductivity of the
membranes,22 leading to increase in total cell resistance and a loss in net power output. Also, the
degradation of a polymer membrane is very much dependent on the operating conditions such as
temperature,23 and humidity,1,24 freeze-thaw cycling, transient operation, fuel or oxidant
starvation, start-up and shutdown. Earlier Fenton’s tests or similar other tests have been used to
study membrane degradation, in which the membrane is directly exposed to hydrogen peroxide
and ppm quantities of cationic contaminants. Although the Fenton’s test is straightforward and
has been considered as a benchmark for PEM durability evaluation, it has inherent limitations.
Deterioration of membrane in such a test involves no electrode process and has nothing to do
with variations in fuel cell operating conditions such as operating potential, relative humidity,
fuel and oxidant starvation etc. Such tests are controversial as they do not simulate accurate fuel
cell operating environments. An alternative approach that has more practical relevance is to run
a long-term fuel cell test and conduct post-mortem analysis to study the changes in membrane
properties. However, in a conventional sense, this method requires a minimum of hundreds of
hours in order to obtain detectable degradation. Testing fuel cells for such lengthy periods of
times is expensive and generally impractical; further, the stability of other fuel cell components
could become the dominating source of performance degradation during such tests.
A systematic investigation of degradation of polymer membrane at fuel cell operating
conditions is highly warranted in order to further the fundamental understanding of the
technology and substantiate PEMFC technology as an alternative renewable energy source. So,
in this work we attempt to analyze the durability of MEA from the perspective of radical initiated
chemical attack of the membrane. From an overall perspective the objectives of this investigation
were to understand the following: (i) Extent of membrane degradation as a function of cathode
142 and anode electrode polarization conditions, effect of (ii) temperature, (iii) catalysts loading, and
(iv) comparison of Pt vs. Pt alloys, in the case of the latter distinction between catalysts
containing alloying element on the surface vs. alloys possessing a Pt rich outer layer have been
investigated. These were studied using our novel segmented cell design4 for durability
characterization. The unique segmented fuel cell design and experimental protocol as described
in detail earlier4 enables multiple working electrodes to be analyzed on the same membrane, such
that specific half-cell (anode and cathode) conditions and choice of electrocatalysts as well as
overpotentials can be invoked under actual fuel cell operating environments. This segmented cell
experimental protocol therefore allows for measurement of membrane degradation in terms of
losses in its ionic conduction as well as within the bulk of its structure from identification of
point of chain scission. These are achieved by comparison of pre and post mortem data from
four point proton conductivity measurements (in-plane and through plane), ion exchange
capacity and infrared data, respectively. Membrane degradation is then correlated with directly
measured peroxide yield values for various electrocatalysts during oxygen reduction using the
rotating ring disk electrode technique. The objective is to therefore correlate peroxide yields with
membrane degradation and understand membrane durability as a function of temperature,
catalyst loading, electrode overpotentials, pure Pt vs. Pt-alloy electrocatalyst, nature of polymer
chain scission (point of radical initiated attack) and overall polymer breakdown.
5.2 Experimental
5.2.1 Physicochemical characterization: X-ray diffraction (XRD, model D/MAX-2200T) was
used to characterize the crystal structure, phase purity, and particle size of the catalysts. The
measurements were made with a Rigaku diffractometer, at 46 kV and 40 mA, fitted with Cu Kα
radiation source, λCu Kα =1.5406 Å. The diffraction patterns were recorded with a scan rate of
143 0.400o/min between 10 to 100o. The analysis of the XRD data was carried out using the “Cell
Refinement” package. The average crystal size of the catalyst was determined using a Scherrer
crystallite size broadening model. Micro-structural characterizations were carried out with a cold
field emission high resolution scanning electron microscope (HRSEM), Hitachi (model S-4800),
to take high resolution topographical micrographs of the catalyst samples. HRSEM micrographs
are very useful in analyzing the morphology of the carbon support, the crystallite size
distribution of the catalysts, and the coverage of the catalyst nano-particles over the support.
Attached to the SEM unit is an electron dispersion x-ray spectrometer (EDS, EDS-GENESIS
HITACHI S-4800), equipped with a Cu filter and a liquid nitrogen-cooled Si(Li) detector, and it
was used to measure the composition of the alloyed catalysts at an acceleration voltage of 25
kV.
5.2.2 Electrochemical Characterization: All electrochemical measurements were made at room
temperature using a rotating ring-disk electrode setup from Pine Instruments connected to an
Autolab (Ecochemie Inc., model – PGSTAT 30) potentiostat equipped with a bi-potentiostat
interface. All potentials in acidic and alkaline solutions were measured with respect to reversible
hydrogen electrode (RHE) and Hg/HgO reference electrode, respectively. Detailed
methodologies are given elsewhere.25 Briefly, ink formulation consisted of sonicating the
electrocatalyst powder with an appropriate quantity of water, isopropanol and small quantity of 5
wt% Nafion® solution as a binder. 4 µL aliquot of catalysts ink with a target Pt metal loading of
14 µg/cm2 was dropped on glassy carbon (GC) disk (0.196 cm2) substrate. 1 M HClO4 and 1 M
KOH were used as the electrolytes. In a separate experiment, a study for determining optimal
electrocatalyst loading on GC disk was performed. Ideal loading of 14 µg/cm2 (Pt) was
144 determined from the inflection of a plot of mass transport normalized current (at 900 rpm) for
ORR vs. loading (Pt metal).
5.2.3 Segmented Cell Design: Durability experiments were performed using a modified fuel cell
hardware based on a “high throughput screening fuel cell assembly” (NuVant System, Inc., IL,
USA).26 Figure 5.1 shows the original look of the fuel cell assembly. Its key components include
an electronically conducting flow field block and an electronically insulating array block on the
opposite side of the MEA. The array block has 25 sensors glued into the block on the opposite
side facing one of the testing spots on the array MEA and a pin jack on the other side used for
electrical connection. The heating control and gas supplies to this fuel cell were built in-house to
enable the cell to run at ambient pressure and constant temperature up to 80oC. 27 Gases were
passed through humidification bottles, which were kept at a temperature 10oC higher than that of
the cell in order to ensure the 100% humidification of the MEA.
The MEAs in this work were customized for the purpose of the durability tests. As shown
in Figure 5.2, the MEA consists of the membrane of interest with a size of 11cm×11cm.
Figure 5.1: Segmented array fuel cell assembly designed for testing multiple working electrodes, each with individual reference electrodes and a common counter.
145
Attached to one side of the membrane is a titanium mesh common counter electrode (CE). On
the other side, the testing area of the MEA is divided into five testing units. Each unit has a strip
of electrode as the working electrode (WE), and two disk electrodes for building the reference
electrode (RE) of this unit. This design enables a simultaneous evaluation of five catalyst
samples of geometric area 5 cm2 (5 cm x 1 cm) at each run under same operating conditions. A
multi-channel Arbin (BT2000) Testing System (Arbin Instruments, TX) was used for
polarization of the individual working electrodes.
5.2.4 Durability test design criteria: In prior publications concerning the susceptibility of PEM’s
to radical-initiated chemical attack, fuel cell experiments performed with either single cell or
multi-cell stack played an important role. These extended life testing reflected the combined
impact from various sources ( fuel cell component configuration, MEA fabrication, operating
conditions, thermal and load cycles, impurities, and uniformity etc.) on the lifetime of the
membrane. However the interplay of these factors leads to inevitable difficulties in interpreting
and reproducing the data and inability to assign the observed membrane failure to one particular
factor without taking other possible triggers and/or enhancers into account. From this point of
Figure 5.2: Design of MEA for durability test showing the five individual working electrodes (WEs) each with their reference electrode (RE) arrangement and counter electrode (CE)
146 view, membrane durability test was designed so as to enable data interpretation. To understand
the radical-induced membrane degradation in fuel cell operation, two types of durability test
were tailored for examining the proposed mechanisms: (a) Cathode side durability test, (b)
Anode side durability test.
5.2.5 Anode side durability test: The ‘anode (hydrogen) side degradation mechanism’ as
proposed earlier 12 is based on the premise that during fuel cell operation, molecular O2
permeates through the membrane and reacts with atomic hydrogen chemisorbed on the surface of
the anode platinum catalyst, thus producing hydrogen peroxide or free radicals. Any peroxides
formed at this interface in conjunction with traces of transition metal ions [Fe2+, Cu2+, …found in
MEA and/or catalyst support (carbon black)] results in formation of free radicals. The possible
reactions involved in this mechanism are12
H2 2Hads (on Pt or Pt/M catalyst)
Hads + O2 (diffused from cathode side) HOO.
HOO. + H+ H2O2
M2+ + H2O2 M3+ + .OH + OH¯
M3+ + H2O2 M2+ + .OOH + OH¯
This proposed mechanism has been suggested on the basis of tests in regular fuel cell setups
under open-circuit potential (OCP) conditions in prior publications.13,28 However, the parallel
process involving interaction of adsorbed oxygen at the cathode (at or near OCP conditions) and
cross over hydrogen resulting in free radical formation cannot be ruled out. This has been
pointed out earlier28 therefore, it is imperative for appropriate experimental design to enable
proper data interpretation. An earlier attempt to understand the extent of this mechanism
involved providing the electrolyte/catalyst interface a predominantly H2 environment consisting
147 of a small amount of O2. For example, to induce degradation of water soluble polystyrene
sulfonic acid [used as a model compound for hydrated polystyrenesulfonic acid (PSSA)
membrane], Hodgen et al.29 purged hydrogen gas containing 5% oxygen through a PSSA
polymer solution at the rate of 1.5 cm3 s-1 in the presence of platinized platinum. Clearly this
method deviated from fuel cell configuration; also it failed to mirror the O2 crossover behavior,
because O2 permeability through the membrane changes dramatically with the chemistry of the
polymer as well as temperature and hydration.30
In order to investigate this mechanism under fuel cell-like conditions in the absence of
interference from reactions of O2 with crossover hydrogen, our approach involved conducting
tests in the aforementioned fuel cell device running with pure hydrogen and pure oxygen (in a
normal fuel cell mode) at ambient pressure. As shown in Figure 5.3(a), humidified hydrogen was
passed through the catalyzed working electrode which also provided reference electrode for the
MEA, and humidified oxygen was passed through the noncatalyzed counter electrode side in
order to enable oxygen diffusion through the membrane to the working electrode side depending
on its permeability at the operating condition of the fuel cell. After full humidification of the
Figure 5.3: Half cell configuration of anode and cathode side degradation tests.
148 MEA (on reaching equilibrium conditions), the working electrodes were either held at the usual
anode potential of the PEMFC [0.1 – 0.2 V, vs. RHE] or left in OCP condition for a fixed period
of time. After the test, membranes in contact with the working electrodes were detached for
postmortem analysis. The results were then compared with the corresponding properties of the
non-degraded membrane before the test (Table 5.1). Further it is also noted that the possibility of
H2 cross over to the cathode side to cause analogous degradation did not find favor in the
literature because of the fact that H2 utilization efficiency on the anode side is sufficiently high
enough as evidenced by the low magnitude of H2 crossover current density (~ 1 to 3 mA/cm2)
reported widely in the literature31-34 and measured in our lab as well. Furthermore the counter
electrode side with O2 flow was a non-catalyzed carbon electrode and hence lacked the reaction
centers afforded conventionally by Pt.
Table 5.1: Basic membrane properties of Nafion®112 showing ion exchange capacity (η), membrane thickness, glass transition temperature (Tg), proton conductivity (σ). Membrane Ion Exchange Proton Thickness Tg Capacity(η) Conductivity(σ) (mm) (C)
(mEq g-1) (S cm-1 at 22C, 100%RH)
Nafion®112 0.91 0.097 0.0508 140
5.2.6 Cathode side durability test: Recent publications 11-12,15,22,35-36 suggest that the vulnerable
location for radical attack in a MEA is at the cathode (oxygen) side. This mechanism is based on
the proposition of oxygen reduction reaction (ORR) at the cathode of PEMFC proceeding via a
parallel pathway where a two-electron reduction of oxygen occurs simultaneously with the
formation of H2O2 intermediates37 along with the predominant four electron reduction to H2O;
149 the peroxides then react with trace transition metals ions (Fe2+, Cu2+… found in membrane
and/or carbon black catalysts support) to form radicals:
O2+2H+ + 2e- H2O2 [5.1]
M2+ + H2O2 M3+ + ·OH + OH- [5.2]
M3+ + H2O2 M2+ + ·OOH + H+ [5.3]
It has been pointed out that the metal ion and H2O2 concentrations necessary for the occurrence
of hydroxyl radical can be very low ( <10-25 mg L-1 H2O2 and 1 part Fe per 5-25 parts of H2O2
(wt/wt)38). Figure 5.3(b) shows our cathode durability test arrangement. The cell is operated on
pure oxygen and pure nitrogen at ambient pressure. Humidified oxygen is passed through the
working electrode side of the cell; humidified nitrogen, which is inert therefore only functions to
hydrate the MEA, is passed through the counter electrode side so that water molecules carried by
nitrogen undergo oxidation at the counter electrode and provide protons that pass through the
membrane to the working electrode side purged with humidified oxygen. This design emulates
an operating fuel cell except for suppressing the passage of hydrogen on the anode side. The
reference electrode is a solid-state dynamic hydrogen electrode (DHE)39, which is constructed by
connecting two disk electrodes (E-TEK 30% Pt/C electrode) to a power supply of 1.7 V. After
full hydration of the MEA, the potential of the working electrode could be set (vs. DHE) at
different potentials, for 24-48hr. Pre- and post test analysis of the membrane in contact with the
working electrode is conducted. The results were then compared with the corresponding
properties of the membrane before the test (Table 5.1).
5.2.7 MEA Fabrication: Working electrodes were selected from commercial ETEK-BASF 30%
Pt/C, 60% Pt/C, 30% Pt2Co/C and 30% Pt3Co/C. Material for the reference electrode was 30%
Pt/C electrode (E-TEK-BASF). The counter electrode for the anode side durability test was
150 noncatalyzed carbon gas diffusion electrode (E-TEK-BASF). For the cathode degradation test
0.6 mgPt/cm2 of 80% Pt/C (PEMEAS) was used as the catalyst and titanium mesh for current
collection. Working electrodes were prepared by sonicating for 20 minutes an appropriate
amount of water-wetted catalyst with 5 wt% Nafion® 1100EW (Ion-Power, Inc. - New Castle,
DE) and isopropanol. All working electrode catalyst loading was 0.4 mgPt/cm2. The resulting
catalyst ink was then sprayed on the commercial carbon gas diffusion electrode (single sided
ELAT, E-TEK) and then dried in an oven at 60°C for 60 minutes. For electrodes obtained from
commercial vendors, a thin layer of Nafion® ionomer solution was brushed on the electrode
surface and then dried in an oven. Typical loading of ionomer layer was in the range of 0.8 – 1.0
mg/cm2. MEAs were prepared by hot-pressing the electrodes to the polymer membrane
according to procedures described in detail earlier.27
5.2.8 Post-Mortem Characterization Techniques: After the durability experiment, the MEA was
uninstalled from the cell, and the working electrode portions were carefully cut off from the
MEA with the appropriate working electrode side carefully labeled. The samples were then
dipped in anhydrous methanol for a short fraction of a second to enable peeling of the electrodes
from the membrane. The membrane samples so obtained, typically 1 x 5 cm, were then washed
thoroughly with deionized water before performing the following analysis.
5.2.9 Fourier-transform infrared spectroscopy: Fourier transform infrared spectroscopy (FTIR)
is a handy, nondestructive technique to probe changes in membrane chemistry due to
degradation, used in numerous studies40-41 and to determine the microstructure of Nafion® in
prior PEM stability studies.13-14 Attenuated total reflectance (ATR) mode was used in an attempt
to study the interface characterestics. IR spectra were recorded with Bio-Rad FTS6000 FTIR
instrument with 45° Ge ATR crystal. For measurement, the dried sample (24hrs in vacuum at
151 60oC) was pressed against the ATR crystal with the help of a force-sensing pressure applicator.
All spectra were collected from 400 scans at 4-cm-1 resolution. Dry nitrogen gas was purged
around the sample during the measurement to eliminate moisture in the air. Linear background
correction in the spectra was attained manually.
5.2.10 Conductivity measurement: Proton conductivity was determined from a fully humidified
membrane at room temperature using four-probe conductivity cell setup described in our prior
publication.42 Measurements were carried out with a digital potentiostat/galvanostat (AUTOLAB
model PGSTAT30 equipped with FRA model, Ecochemie B. V.)
5.2.11 Ion exchange capacity (IEC): Ion Exchange Capacity (IEC) defined as the ratio of moles
of sulfonate ion exchange sites to the dry weight of Nafion® is expressed in mEq/g and were
measured using standard methods, which involved equilibrating known amount of H+ form of the
membrane in a measured volume of a standard solution of 3 M NaCl at 100°C for 10 hrs to allow
for the exchange with H+ ions. This solution was then titrated to a phenolphthalein end-point
with a standard NaOH solution.
5.3 Results and Discussion
5.3.1 Physicochemical Characterization (SEM & XRD): HRSEM micrographs for Pt/C, and Pt-
Co/C are shown in Figure 5.4. These images show that a statistically overwhelming number of
observable particles have sizes in the range <3 nm for 30%Pt/C and in the range of 2 to 5 nm for
60%Pt/C. Particle size distribution for Pt-Co alloys exhibited a wider range, particle size
variation from 2 to 10 nm was observed as compared to Pt. Except for the 60%Pt/C, where the
nano-particles of this catalyst aggregate, the catalyst particles for the rest of the samples are well
dispersed on the carbon support. Along with the SEM patterns the fluorescence signal was
152
analyzed using an EDAX analyzer (EDS-GENESIS HITACHI S-4800). Table 5.2 shows the
atomic composition of the two supported Pt alloys in comparison to Pt3Co whose bulk
composition is 75:25, the Pt2Co sample exhibits ∼70:30 ratio, thereby showing very good
correspondence with the nominal composition. X-ray diffraction patterns of the supported
electrocatalysts, Pt, Pt –Co mixtures scanned at 2θ angles over a range of 10o to 100o are shown
in Figure 5.5. The broad-based diffraction peak at ~ 2θ = 24.6o for all types of catalysts arises
from that of the carbon support. In Figure 5.5, the two Pt catalysts show the typical platinum
peaks at 2θ position of <111>, <200>, <220>, <311>, and <222>; whereas the Pt-Co peaks
position are shifted slightly higher in 2 θ values showing an appropriate reduction in lattice
parameters as a result of addition of Co in the unit cell structure showing a clear indication of Pt-
Co alloy formation.
Table 5.2: Physicochemical characterization of Pt and Pt alloy catalysts using SEM/EDS Catalyst* EDS Elemental
Composition Average Size (Å)
(XRD) Lattice Parameter (Å)
30% Pt/C -- 27 3.934 (Fm3m) 60% PtC -- 36 3.934 (Fm3m)
30% Pt3Co/C
Co = %24±1 Pt = %76 ±1
55 3.855 (Fm3m)
30% Pt2Co/C
Co= %32 ±1 Pt= %68 ±1
70 3.810 (Fm3m)
*All are commercial PEMEAS catalysts supported on Vulcan XC-72 Carbon
PEMEAS 30% Pt/C PEMEAS 60% Pt/C ETEK 30% Pt2Co/C PEMEAS 30% Pt3Co Figure 5.4: SEM images of Pt and Pt-alloy catalysts employed in this study.
153
The approximate average particle sizes of the carbon-supported catalysts were determined by
using the Scherrer equation43, which relates crystallite size to the line broadening of the peaks.
Although not accurate for particles either (<5nm) or above (>50nm), XRD still constitutes one
effective tool for estimating catalyst particle sizes usually in the range of 5 to 50 nm. However
these particle sizes reflect exclusively the diffracting domains and so all amorphous components
are excluded. The average particle sizes, based on the peak width of the <111>, <200>, and
<220 > diffraction lines are presented in Table 5.2. There is a broad agreement between the
particle sizes obtained from SEM and XRD analysis, thereby indicating a high degree of
crystalline character to the supported Pt and Pt alloy naonoparticles. The lattice parameter
obtained from the XRD patterns indexed to a face-center cubic (fcc) structure was 3.9238Å
which is in good agreement with the literature value15 of 3.9239 Å. The Pt-Co binary mixtures
have lower lattice parameter values than the corresponding pure platinum, and increasing the
2ΘΘΘΘ, degrees
20 30 40 50 60 70 80 90
Inte
nsity
, a.u
.
30wt% Pt/C
60wt% Pt/C
32wt% Pt3Co/C
30wt% Pt2Co/C
Pt<111>
Pt<200> Pt<220> Pt<311>
Figure 5.5: X-ray diffractograms of the Pt and Pt-alloy catalysts used in this study.
154 atomic content of cobalt in the mixture decreases the lattice parameter values, since the atomic
radius of cobalt is smaller than platinum and the decrease in lattice parameter indicates alloying
of the metals where cobalt enters the platinum lattice by substitution in to octahedral sites 44.
5.3.2 Electrochemical Measurements - Cyclic Voltammetry: Cyclic voltammetry (CV) was used
to characterize the catalysts in argon purged 1 M HClO4 at room temperature by cycling between
0.05 V and 1.2 V vs. RHE. Also CVs were recorded in 1 M KOH electrolyte to investigate the
electrochemical behavior of cobalt in PtCo alloys used in this study. This was especially useful
in determining presence of surface Co as characterized by typical redox peaks in alkaline
electrolytes. The resulting voltammograms in oxygen-free acidic and alkaline electrolytes taken
at 20 mV/s with a loading of 14 µgPt/cm2 are shown in Figure 5.6 and 5.7. The electrochemically
active surface area of the catalysts was also estimated from the integrated charge in the H
adsorption/desorption region of the CVs and are shown in Table 5.3. Cyclic voltammograms
show that the carbon supported Pt particles possess some degree of low coordinated crystal
Potential [V Vs RHE]
0.0 0.2 0.4 0.6 0.8 1.0 1.2
Cur
rent
[mA
/cm
2 ]
-0.4
-0.3
-0.2
-0.1
0.0
0.1
0.2
30% Pt/C60% Pt/C
(a) 1M HClO4 @ 20 mV/s
Potential [V Vs RHE]
0.0 0.2 0.4 0.6 0.8 1.0 1.2-0.3
-0.2
-0.1
0.0
0.1
0.2
30% Pt2Co/C
30% Pt3Co/C
(b) 1M HClO4 @ 20 mV/s
Figure 5.6: CV in argon saturated 1 M HClO4. (a) Pt/C and (b) Pt-alloy/C catalysts on glassy carbon disk at 20 mV/s. Current densities based on geometric electrode area.
155 planes, and hence the hydrogen adsorption/desorption features between 0.4 and 0.0 V vs. RHE
are different from the CV expected of a bulk pc-Pt electrode. The area in the hydrogen
underpotential deposition (HUPD) region decreases with decreasing Pt surface sites available.
Also an anodic shift in the reduction peak is observed for the two PtCo alloys relative Pt/C which
can be attributed to a decrease in desorption free energy of Pt-OH, Pt-O or Pt-O2 due to the
presence of the alloying element, implying that the reduction of oxygen containing intermediate
species is more facile. It is also interesting to compare the HUPD region for the two supported
PtCo catalysts. The shape of the HUPD region for Pt3Co (76% Pt) is very similar to that of Pt
catalyst and shows typical hydrogen adsorption/desorption features in the potential range of 50 to
400 mV, whereas Pt2Co (68% Pt) exists in alloy form with a poorly resolved HUPD region.
Figure 5.7 shows typical cyclic voltammograms of PtCo alloys in alkaline media using 1
M KOH at 10 mV/s. The redox couple observed between 0 V and -0.2 V [vs. Hg/HgO] is due to
the redox processes involving metallic Co in alkaline medium leading to the formation of Co3O4
and/or CoOOH as indicated by the Pourbaix diagram of Co and detailed electrochemical
investigations.45 Co oxidation peaks are not discernable due to large double layer current but the
corresponding reduction peaks are evident. It is observed from these cyclic voltammograms that
Pt3Co shows higher Co redox peaks compared to that of Pt2Co for the same loading of 14
µgPt/cm2 on the glassy carbon disk. On the contrary, EDS measurement, which is a bulk
averaged technique, shows a cobalt composition of only 25% for Pt3Co compared to that of 30%
for Pt2Co. This implies that in Pt3Co more Co is present on the surface compared to Pt2Co, in
turn observed as higher metallic Co redox currents in Pt3Co than Pt2Co.
156
Table 5.3: Electrokinetic parameters for the different electrocatalysts used in this study in 1 M HClO 4 at room temperature from RDE measurements at 900 rpm. Catalyst ik @ 0.9 Vc ik @ 0.8 V diox109 Tafel Slopea ECA (m2/gPt)
b (mA/cm2
Pt) (mA/cm2Pt) (A/cm2
Pt) (mV/dec) [HUPD region] 30% Pt/C 0.078 0.918 0.230 60/122 66 60% Pt/C 0.141 0.960 0.852 63/125.5 38 30% Pt2Co/C 0.408 2.325 1.400 60/139 35 30% Pt3Co/C 0.314 3.065 0.853 59.5/129 37 a Extracted from the anodic sweep of ORR from 0.35 V to 1.2 V vs. RHE b Based on 210 µC/cm2 for atomic hydrogen oxidation on a smooth Pt surface. c ik – kinetic current density; dio – equilibrium exchange current density;
Potential [V Vs Hg/HgO]
-1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 0.4
Cur
rent
Den
sity
[mA
/cm
2 ]
-2
-1
0
1
2
30% Pt2Co/C
30% Pt3Co/C
1M KOH @ 20 mV/s
Figure 5.7: Cyclic voltammograms in oxygen free 1 M KOH at room temperature for 30% Pt2Co/C and 30% Pt3Co/C catalysts at 20 mV/s; Current densities are based on geometric electrode surface area of the glassy carbon disk.
157
5.3.3 Oxygen Reduction Reaction (ORR) Kinetics and Peroxide Yield Measurements: Figure 5.8
shows a representative set of rotating ring-disk experiments performed with Pt and Pt alloy
catalysts in O2 saturated 1 M HClO4 at room temperature using a constant Pt metal loading of 14
µgPt/cm2 on the glassy carbon (GC) disk. The voltammograms measured at a scan rate of 20
mV/s, are shown for a rotation rate of 900 rpm. Rerpesentative scans shown in the bottom left
hand side of Figure 5.8(a), represents the anodic sweep. The anodic sweep represents a true
(a)
Rin
g C
urre
nt, I
R [m
A]
0
1
2
3
4
5
Disk Potential ED [V] Vs RHE0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Dis
k C
urre
nt, I
D [m
A/c
m2 ]
-4
-3
-2
-1
0
30% Pt/C60% Pt/C30% Pt2Co/C
30% Pt3Co/C
(B) Ring Current ER = 1.3 V
(A) Disk Current
ik [mA/cm2Pt]
0.01 0.1 1
E [V
] Vs
RH
E
0.70
0.75
0.80
0.85
0.90
0.95
1.00
30% Pt/C60% Pt/C30% Pt2Co/C30% Pt3Co/C
(C) Tafel Plots
60 mV/dec
120 mV/dec
O2 Satd. 1 M HClO4
(D) Levich Plot
ωωωω0.5 [s-0.5]
4 6 8 10 12 14 16
i lim
[m
A/c
m2 ]
0
2
4
6
8
Theo. n = 4Theo. n = 230% Pt3Co/C n = 3.72
Figure 5.8: Disk (A) and ring (B) currents on 30% Pt/C, 60% Pt/C, 30% Pt2Co/C, and 30% Pt3Co/C during ORR in the anodic sweep in 1 M HClO4 electrolyte at a disk rotation rate of 900 rpm and scan rate of 20 mV/s using a glassy carbon disk of 5mm diameter and ring collection efficiency of 39%. (C) Tafel plots for the ORR at room temperature extracted from anodic sweep at 10 mV/s, 900 rpm. (D) Representative Levich Plot for the ORR on 30% Pt/C at various potentials; current densities normalized to the electrochemical surface area of platinum on 5mm glassy carbon disk unless otherwise indicated in the plot.
158 comparison of ORR activity as it is taken immediately after the corresponding cathodic sweep
and hence represent ORR activity after removal of the oxide layer formed at or near open circuit
conditions. While it may be argued that this may not represent true fuel cell operating
environment it does represent a more accurate picture of the reaction center (Pt site) activity for
ORR. In the potential region of 0.9 V to 0.7 V vs. RHE, mixed kinetic-diffusion controlled
currents are observed ensued by a well-defined diffusion limiting current beyond 0.6 V. For the
same Pt loading of 14 µg/cm2 on the GC disk, 30% Pt/C and 60% Pt/C exhibit essentially similar
ORR activity in the mixed kinetic-diffusion controlled region followed by a significantly lower
diffusion controlled current for 30% Pt/C as compared to 60% Pt/C which can be attributed to
the increased thickness of the catalyst layer on the GC disk for 30% Pt/C relative to 60% Pt/C,
since the supporting carbon determines oxygen diffusion through the porous catalyst loaded on
the GC disk. An exactly similar argument can also be extended to the two supported PtCo alloys
with different Pt compositions, such that 30% Pt3Co/C has a smaller catalyst layer thickness and
hence higher diffusion limited current relative to 30% Pt2Co/C. On comparing the Pt/C catalysts
with PtCo/C alloy catalysts, it is seen that the two PtCo/C catalysts exhibit an anodic shift of
about 30 mV in the mixed kinetic-diffusion potential region and hence 30 mV lower
overpotential for ORR. This lower overpotential of the alloy catalysts towards ORR is due to the
fact that the presence of alloying element decreases the desorption free energy (∆Gdes) of Pt-OH,
Pt-O or Pt-O2 such that the adsorption of oxygen containing intermediate species on Pt surface
sites is inhibited in the supported alloy catalysts compared to Pt/C.46 This discussion based on the
ORR profile for each catalyst involves interference from diffusion limited current densities ilim,
and hence Tafel plots are extracted using the following equation (5.4) by eliminating ilim and
obtaining a clearer picture based on kinetic currents.
159
ik = (ilim * i)/(i lim – i) (5.4)
where ik is the kinetic current density, i is the measured current density during oxygen reduction
polarization, and i lim is the diffusion limited current density. Figure 5.8(c) shows the
corresponding Tafel plots of the Pt and Pt alloy catalysts where the kinetic current densities are
normalized on the basis of the electrochemical surface area of Pt. Firstly, it is noted that the
activity of Pt alloy catalysts are better than the Pt/C catalysts due to the unique effect of the
surface Co species enabling lower oxide formation on Pt47-48. Taking into consideration that the
intial adsorption of molecular oxygen on Pt is part of the rate determining step (rds)25, the
coverage of oxides at or near the open circuit potentials represents a surface poison. Hence
preferential oxide formation on surface Co as shown earlier is attributed to freeing Pt sites for
intiating ORR49. However the comparison of the relatively oxide free anodic scans exhibiting
enhanced ORR is significant. This shows the concomitant oxide formation on Pt sites in the
supported PtCo electrocatalysts is significantly lower than the corresponding Pt/C catalysts (both
30 and 60% on carbon loading). This in the background of previous observations of
experimentally derived activation energies on these class of supported catalysts and agreement
with theoretical calculations on transition states50 indicates that the observed enhancement is
direct effect of surface oxide coverage. Also, it is seen from the Tafel plot that the Tafel slope is
constantly changing, which is due to the continuously varying charge transfer coefficient (α)
value from 0 to 1 with overpotential; however, it is possible as a gross approximation to obtain
Tafel slope [Table 5.3] representing two distinct regions, in good comparison with previously
reported literature values 51, yielding two different Tafel slopes of -2.3RT/F, i.e., 60 mV/decade
at low overpotentials (E > 0.85 V) and -2.2.3RT/F, i.e., 120 mV/decade at high overpotentials (E
< 0.85 V) which agree very well with prior reports on single-crystal Pt electrodes52, carbon
160 supported Pt53 and pc-Pt plug25. This change from 60 mV/decade to 120 mV/decade is closely
related to earlier contention of the adsorbed OH species at potentials beyond 0.85 V vs. RHE
corresponding to 60 mV/decade of Tafel slope, whereas at higher overpotentials, 120 mV/decade
indicates a clean catalytic surface devoid of any oxygen containing adsorbed intermediate
species that can affect the adsorption of molecular O2 from the solution to the active surface site
for subsequent reduction25. Cyclic voltammogram measured in alkaline electrolyte shown in
Figure 5.7 as explained earlier indicate that the Pt3Co/C surface is rich with cobalt which can
inhibit the adsorption of oxygenated intermediate species on the Pt surface sites thereby avoiding
H2O2 generation pathways and providing a direct route for H2O formation. Subsequently, ring
current (also peroxide yield values shown in Table 5.3) of Pt3Co catalyst imply a very low
amount of peroxide generation presumably attributed to the rich cobalt on the surface.
Figure 5.8(d) shows the representative Levich plot for 30% Pt3Co/C used in this study at
various rotation rates from 100 rpm to 2500 rpm. Similar plots were obtained with the other three
catalysts and hence not shown here. Levich plot yields the so called Levich constant B,
according to relation54 (5.5) given below for ORR limiting current ilim, from which the number of
electrons transferred was calculated to be 3.72 for Pt3Co indicating a predominant 4 electron
transfer.
i lim = Bω1/2 (5.5)
where B = 0.62nFD2/3ν
-1/6C, where n is the number of electrons transferred, F is the Faradays
Constant, D is the diffusion coefficient of O2 in the electrolyte, Co is the oxygen concentration in
the electrolyte, ν is the kinematic viscosity and ω is the rotation rate in rpm.
Hydrogen peroxide yield due to the parallel pathway for ORR was also analyzed by
classical rotating ring-disk electrode (RRDE) technique. The formation of relative amount of
161 H2O2 and H2O can be determined quantitatively with the RRDE experiment by holding the
potential of the ring at 1.3 V vs. RHE, where H2O2 formed at the disk during oxygen reduction is
readily oxidized at the ring. Figure 5.8(b) shows the currents measured at the ring during the
cathodic sweep of the disk potential shown in figure at room temperature in 1 M HClO4 at 900
rpm. Table 5.4 shows the comparison of peroxide yields measured at the gold ring when the
corresponding disk electrode potentials were 0.4 V, 0.6 V and 0.7 V vs. RHE and calculated
using the following relation,
χ(H2O2) = 2(IR/N)/(ID + (IR/N) (5.6)
where N is the collection efficiency of the ring, χ is the mole fraction of peroxide formed, ID and
IR are the disk and ring currents, taking into account the total disk currents for the oxygen
reduction as the sum of reduction currents of O2 to H2O and H2O2 and the collection efficiency N
for the ring electrode25.
These data in Table 5.4 are an indication of potential dependence of peroxide yield at
disk electrode for the various catalysts used in these experiments. Peroxide currents are
negligible for disk potentials above 0.65 V indicating that ORR predominantly proceeds via four
electron transfer process without significant peroxide generation, which is relevant for the
operating potential of fuel cell cathodes. This is the region where the cathode potential of a
normal operating fuel cell falls, and alludes to the importance of maintaining a stable cell
potential with regard to the interfacial stability of the MEA, especially in the case of
discontinuous fuel cell operation in which considerable voltage fluctuations take place
frequently. Below the normal fuel cell operating potential of 0.6 V to 0.7 V vs. RHE, peroxide
generation begins to increase significantly followed by much higher ring currents at a diffusion
162
Table 5.4: Peroxide yields (mole fraction) measured using a RRDE technique with a rotation rate of 900 rpm and various disk potentials of 0.7 V, 0.6 V, 0.4 V vs. RHE in conjunction with a gold ring electrode polarized at 1.3 V in 1 M HClO 4 at room temperature. Catalyst %H2O2 @0.7 V %H2O2 @0.6 V %H2O2 @0.4 V 30% Pt/C 0.111 0.257 2.330 60% Pt/C 0.394 0.753 2.408 Pt2Co/C 0.342 0.652 3.523 Pt3Co/C 0.160 0.260 0.510
controlled disk potential of 0.4 V. This potential dependence of peroxide yield is correlated to
fuel cell membrane degradation in the next section. It is also observed that as the disk potential
further extends into the HUPD region, the ring currents of the two Pt/C samples keeps increasing.
In this case of the Pt/C catalysts, as the disk potential enters HUPD region enhanced peroxide
generation occurs as HUPD blocks the Pt surface sites necessary to split the O2 molecule leading
to increased peroxide generation. This is important in cases where H2 permeation from the anode
feed to the cathode catalyst through the membrane is significant as this is also a possible pathway
of peroxide generation and subsequent membrane degradation.
Table 5.5: Selective list of IR absorption peak assignments of H-Nafion® 112 based on pure vibrational modes. Check Section 5.3.7 for an interpretation based on mechanically coupled vibrational modes. Index Wavenumbera (cm-1) Peak assignments (A) 969 m υs(C-O-C), Ether band ‘A’, symmetric
(B) 982 m υs(C-O-C), Ether band ‘B’, symmetric (C) 1059 m υs(SO3
-), sulfonate group, symmetric
(D) 1142 vs υs(CF2), CF2 Backbone stretch a Relative Intensity: m-medium; vs-very strong; vb-very broad; sh-shoulder;
163 5.3.4 Assignment of the Absorption bands of ATR-IR Spectra of Nafion®112 (H-form): Infrared
absorption studies, along with small angle X-ray, neutron scattering investigation and scanning
probe microscopy imaging, have been used widely to elucidate the nanostructure of Nafion®
membrane. Figure 5.9 & Table 5.5 indicates the assignment of various vibrational absorption
bands,40-41,55-58 in the region between 900 cm-1 to 1350 cm-1 wavenumbers as relevant to this
study, associated with the chemical structure of Nafion®112 membranes (DuPont Corp.) in H-
form shown in Scheme 5.1. Based on pure vibration mode assignments, symmetric stretching of
the sulfonate group is observed at 1059 cm-1. The twin peak at 969 cm-1 and 982 cm-1 is due to
the presence of two ether linkages (-C-O-C-) in the Nafion® side chain. Of the two ether
vibrational absorption bands in Nafion® membrane, the higher frequency band (i.e., the one at
982 cm-1) is attributed to the ether linkage directly attached to the fluorocarbon backbone and is
labeled as ether band ‘B’ the corresponding lower frequency component at 969 cm-1 is labeled as
ether band ‘A’, due to its proximity to the sulfontae group and it electron withdrawing character.
This assignment of vibrational absorption bands of ether linkages is due to the work done by
Moore et al55, who contrasted the two ether absorption peaks in Nafion® against a single ether
absorption band present in Dow perflourosulfonate ionomers (PFSI). Dow PFSI has only one
ether linkage in its side chain and exhibits a single absorption peak centered at around ~969 cm-1.
Gierke et al59, proposed the ion-cluster network theory for the morphology of Nafion®
membranes according to which sulfonate groups with terminating the pendant side chains stretch
out into approximately spherical clusters also consisting of water and hydrated cations,
interconnected to each other by channels for ionic transport, and supported by hydrophobic
fluorocarbon backbone material. Meanwhile, Yeager and Steck60 corroborated the conclusions of
Falk et al.61 (that the ionic clusters are non-spherical in shape and have intrusions of side chain
164 ether linkages), by proposing three phase morphology for Nafion® consisting of the hydrophobic
fluorocarbon phase, hydrophilic ionic clusters and an interfacial region between these two. This
interfacial region is largely a void volume containing pendant side chain materials, a small
amount of sorbed water and trace level of sulfonate exchange sites and counter-ions. From the
results of this present study, this interfacial region is of importance since they turn out to be the
vulnerable site for radical species attack during fuel cell operation as discussed in the following
sections.
5.3.5 Effect of radical-initiated degradation on membrane properties: Segmented cell durability
tests were conducted to investigate the mechanism of degradation as a function of potential,
temperature, choice of electrocatalyst, catalyst loading, and test duration. Cathode operating
potentials of 0.4 V, 0.6 V and 0.7 V were employed to study the influence of peroxide yield,
Wavenumber (cm-1)
900950100010501100115012001250130013501400
Abs
orba
nce
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
Nafion 112
Symmetric SulfonateStretching group (C) O=S(O)=O
1059cm-1
(C)
Ether Linkage (A) -C-O-C-
969 cm-1
Ether Linkage (B) -C-O-C-
982cm-1
(B)(A)
Backbone CF2
Stretching (D)
1142 cm-1
(D)
Figure 5.9: ATR Spectrum of Nafion®112 (H-form) based on pure vibrational modes. Check Section 5.3.7 for an interpretation based on mechanically coupled vibrations.
165 subsequent membrane deterioration and also to correlate segmented cell durability test results to
RRDE data discussed above. Tests were conducted at room temperature, 40°C, 60°C, and 80°C
to study the influence of temperature for various time scales of 24 or 48 hours. At the end of each
durability test, the membrane was carefully separated from the electrode; the changes in its
proton conductivity (σ) and ion exchange capacity (η) were determined and compared with that
of the non-degraded pure Nafion® membrane properties shown in Table 5.1, to estimate the
extent of degradation quantitatively; ATR-IR absorption spectra of degraded membranes were
compared to spectra of pristine membrane (Figure 5.9) to determine the site of membrane
degradation.
Table 5.6 shows the results of cathode side degradation tests conducted at 0.4 V and 0.6
V, 40°C, 1 atm pressure conditions for duration of 24 hours with four parallel electrocatalysts
30%Pt/C, 60%Pt/C, 30%Pt2Co and 30%Pt3Co.
Table 5.6: Effect of cathode side durability tests with O2/N2, 40°°°°C, 1 atm, for a duration of 24 hours using 30% Pt/C, 60% Pt/C, 30% Pt2Co, 30% Pt3Co at two different working electrode (WE) polarization potentials of 0.4 V and 0.6 V Vs. DHE. Listed are the decreases in proton conductivity σ (σ (σ (σ (S cm-1) and ion exchange capacity η (η (η (η (mEq g-1).
Catalysts
WE @ 0.4 V WE @ 0.6 V
(%)
after before
before
η - η
η
(%)
(%)
after before
before
η - η
η
(%)
30% Pt/C 34 26 27 19 60% Pt/C 57 48 49 43
30% Pt2Co/C 50 40 44 34 30% Pt3Co/C 27 22 16 9
after before
before
σ -σ
σ
after before
before
σ -σ
σ
166
Figure 5.10 compares the FTIR spectrums of the Nafion® membranes before and after
degradation test at 0.4 V, 40°C, 1 atm pressure conditions for duration of 24 hours using Ge
crystal with IR penetration depth of 0.65 µm to 0.7 µm. The most significant difference after this
24 hour test is observed at the ether band ‘A’ peak at 969cm-1 as a decrease in its intensity of
absorption. There is a concomitant effect of cleavage of ether band ‘A’ and the relatively small
Wavenumber (cm-1)
90010001100120013001400
Abs
orba
nce
ν(SO3)
Ether
Band IEther
Band II
(b) Nafion After Degradation
(a) Dry Nafion before degradation
30% Pt/C
60% Pt/C
30% Pt2Co/C
30% Pt3Co/C
Figure 5.10: ATR-FTIR Spectrums of Nafion® obtained using Ge ATR crystal before and after cathode side degradation tests with four parallel samples as indicated in the plot, operated for 24 hours at 0.4 V, 40°C cell temperature and 1 atm pressure conditions. (a) IR Spectrum (900 cm-1 to 1400 cm-1) before degradation. (b) IR Spectrum after degradation. Check Section 5.3.7 for an interpretation based on mechanically coupled vibrations.
167 decrease in intensity of sulfonate vibrational band at 1059 cm-1, and the reason for this is
discussed below. Vibrational absorption peak at 969 cm-1, as discussed earlier in this chapter, is
due to the ether bond present closer to the sulfonate ion exchange site in the pendant side chain.
The vibrational ether band ‘B’ at 982 cm-1, due to the ether linkage directly attached to the
fluorocarbon backbone is relatively unaffected compared to the band at 969 cm-1, indicating a
preferential cleavage of the ether linkage directly attached to the sulfonate exchange groups.
Quantitative results from IR absorbance plot of the membrane can be obtained only if an internal
reference is incorporated in the membrane that does not degrade during the durability test.62
Since such an internal reference might interfere with durability testing, it wasn’t preferred in this
study. For the purpose of semi-quantitative comparison the concept of relative absorbance has
been used here. For example with the 30% Pt/C catalyst sample, ratio of absorbance intensity of
ether band ‘A’ to the symmetric -CF2 stretching peak (Hυ(-C-O-C-)I/Hυ( 2CF− )) decreased to 0.256
(after test) as compared to the initial ratio of (Hυ(-C-O-C-)I/Hυ( 2CF− )) = 0.312. –CF2 symmetric
stretching peak is used as the reference peak for normalization since it did not decrease in
intensity during the course of the experiment. It is noted here that the spectra in Figure 5.10 is
obtained using Ge ATR crystal that provides a penetration depth of only 0.65 to 0.7 µm and same
test performed with ZnSe ATR crystal (penetration depth 1.7 to 2 µm) did not exhibit any
significant decrease in any of the IR absorption bands indicating that this cleavage is highly
localized on the surface and has not probably affected the membrane beyond a distance of ~1 µm
from the surface. Table 5.7 shows the decrease in relative intensity of ether band A in
Nafion®112 after cathode side durability test at the indicated conditions with the four Pt and Pt-
alloy catalysts.
168 Table 5.7: Percentage decrease in intensity of IR absorption of Ether Band 1 at 969 cm-1 and sulfonate symmetric stretching band at 1059 cm-1 after cathode side degradation test performed at 40°C, ambient pressure, for a duration of 24 hours at a WE potential of 0.4 V. Aratio[ν(C-O-C)I] = (H υυυυ(-C-O-C-)I/Hυυυυ( 2CF− )) and Aratio[ν(SO3)] = (Hυυυυ( SO3)/Hυυυυ( 2CF− )). % A ratio = (Aratio(before) – Aratio(after) ) / Aratio(before).
Catalyst
% Aratio ν(C-O-C) I
% Aratio ν(SO3)
H2O2 Yield @
0.4 V (%)a
30% Pt/C 18 % 22% 2.330
60% Pt/C 39 % 37% 2.408
30%Pt2Co/C 21 % 23% 3.523
30%Pt3Co/C 12 % 13% 0.510
a H2O2 obtained from ring current using eqn. 5.6.
Further, the decrease in intensity of ether band ‘B’ (Hυ(-C-O-C-)II/Hυ( 2CF− )) was negligible
and not shown here. Subsequently, loss in conductivity and IEC, shown earlier in Table 5.6 is
also due to the preferential cleavage of ether band ‘A’ in the side chain, because this cleavage
scissions off the sulfonate ion exchange sites present at the terminal end of the pendant side
chain and directly attached to ether band ‘A’. Also a very minor shift in the vibrational frequency
of the sulfonate exchange sites to higher values is observed. Vibrational stretching frequency of
sulfonate group υs(SO3) in pure Nafion®-H form is localized in the spectral region around 1059
cm-1. υs(SO3) of the degraded membranes is observed around 1062 cm-1 to 1064 cm-1 and this is
due to the modest contamination of membrane by counter ions (such as Na+, Rb+, Li+, Cs+, Ca+),
because minor cationic impurities from the carbon support, gas diffusion electrodes,
humidification bottles, other fuel cell hardware are inevitable in the fuel cell operation. These
foreign cations, usually have stronger affinity with the sulfonic acid group compared to H+, and
thereby replace the protons (H+) attached to the sulfonate ion exchange sites; this replacement of
protons by metal impurities causes a polarization of S-O dipoles and subsequently shifts υs(SO3)
169
to higher frequencies40,55,63. Also IEC indicated in Table 5.6 suffers a loss similar to the decrease
in conductivity due to the loss of ion exchange site. A comparison between peroxide yield in
Table 5.4 for the various catalysts, loss in conductivity and IEC in Table 5.6 and decrease in ratio
of absorbance in Table 5.7 indicates a direct one-on-one relationship between peroxide yield
obtained on the ring electrode and the level of membrane degradation. 60% Pt/C and 30%
Pt2Co/C catalysts exhibited significantly higher peroxide yield on the ring at all potentials above
0.7 V vs. RHE and is reflected in the high levels of degradation observed in membrane
characteristics in the durability experiment. Further 30% Pt3Co/C that consistently yielded low
peroxide on the ring at all potentials gives rise to a low level of loss in membrane properties after
cathode side durability testing. This clearly indicates that the degradation on the cathode side is
Wavenumber (cm-1)96098010001020104010601080
Abs
orpt
ion
(a.u
.)
0.4 V0.6 V0.7 V
EtherBand I
EtherBand II
30% Pt/C
Figure 5.11: Potential dependent cleavage of Ether band ‘A’. Shown are the IR Spectra of Nafion® after cathode side degradation test at 40°C, 1 atm pressure conditions for duration of 24 hours using 30% Pt/C catalyst.
170 most likely due to peroxide radicals generated from local interfacial Fenton type catalysis of
H2O2 in turn generated from the 2e- pathway of ORR.
Figure 5.11 shows the IR plot of degraded membranes in the spectral region 1100cm-1 to
940cm-1 after cathode side degradation test performed at 40°C, 1 atm pressure conditions for
duration of 24 hours using three parallel samples of 30% Pt/C as the cathode catalyst at various
cathode operating potentials of 0.4 V, 0.6 V, 0.7 V vs. DHE. At normal cathode operating
potentials of 0.6 V and 0.7 V the cleavage of ether band ‘A’ is relatively less intense than at an
accelerated cathode operating potential of 0.4 V, due to significantly higher peroxide formation
at 0.4 V. Potential dependent cleavage of ether band ‘A’ shows that the peroxide radical
generation at various operating fuel cell potential is directly correlated to the subsequent polymer
membrane.. It is also seen that the intensity of ether band ‘B’ is relatively unaffected during the
durability test. This result from durability experiments in segmented cell is compared and
correlated to peroxide yield experiments performed using RRDE technique with 30%Pt/C
catalyst at room temperature in 1 M HClO4 and shown in Figure 5.8. At a diffusion controlled
fuel cell operating potential of 0.4 V, peroxide yield at the ring electrode is higher than at 0.6 V
and 0.7 V and is reflected in the higher intensity of cleavage of ether band ‘A’ at 0.4 V vs. DHE.
Although a quantitative relation cannot be obtained, there is a one-on-one trend between the
peroxide yield obtained on the ring in RRDE and Fenton type degradation on the cathode side of
an operating fuel cell.
As discussed above, the small decrease in intensity of υ(SO3), shown in Table 5.8, is
attributed to cleavage of ether band ‘A’. This results in the scission of the sulfonate group
present at the terminal end of the pendant chain. Scheme 5.2 shows the vulnerable region for
radical attack based on the result of this study. It is also seen that 60% Pt/C which has higher
171
Table 5.8: Loss in conductivity and Peroxide yield from 30% Pt/C after cathode side degradation test at 40°C, ambient pressure, for 24 hours at various WE potentials of 0.4 V, 0.6 V, 0.7 V vs. RHE.
WE@ 0.7 V
WE @ 0.6 V
WE @ 0.4 V
Peroxide Yield (%)
0.111 0.257 2.330
Loss in Conductivity
(%)
16% 27% 34%
catalyst loading and higher particle size (Table 5.2) as compared to 30% Pt/C gives higher
peroxide yield observed from RRDE results (Figure 5.8) and correspondingly higher loss in
membrane conductivity (Table 5.6) from durability testing. As shown in Scheme 5.1, based on
the chemical structure of Nafion®, the repeating unit of the fluorocarbon backbone, characterized
by the value ‘m’, determines the dry weight of the membrane and also its ion exchange
capacity41,58(number of moles of sulfonic acid ion exchange sites per gram of dry polymer
membrane) which varies from 0.55 to 1.05 mEq/g. Ion exchange capacity (IEC) of the degraded
membranes, as shown in Table 5.6, shows a loss similar to that of the proton conductivity. Since
the vibrational absorption bands at 1142 cm-1 and 1208 cm-1 represent the fluorocarbon backbone
(-CF2 groups), the relative intensity of absorption is determined by the factor ‘m’. This remains
unaffected after 24 hours of cathode side degradation test performed at 0.4 V vs. DHE for the
four cathode catalysts used in these experiments, thus the decrease in IEC can be attributed
directly to the loss of sulfonate exchange sites. Decrease in proton conductivity and IEC values
followed by relative decrease in IR intensity of one of the ether bands and no decrease in
172 fluorocarbon backbone IR absorption signature implies that the radical species did not attack the
fluorine groups of the hydrophobic backbone.
Following the discussions of Falk et al61 and Moore et al55 regarding the chemical
nanostructure of Nafion® membrane briefly summarized above, it is likely that part of the side
chain ether linkage intrudes into the hydrophilic ionic clusters. Consequently, during this period
of 24 hours of cathode side degradation tests, it is observed that the radical species generated
during the course of fuel cell operation initiates the polymer chain breakage by attacking the
hydrophilic ionic cluster region (specifically the ether linkage intruding into the hydrophilic ionic
cluster region) and virtually does not degrade the hydrophobic backbone of the membrane.
Previous results using data of membrane degradation after exposure to Fenton’s solutions such as
those reported by Inaba et al64 concluded that both main chain and side chains are decomposed at
similar rates by radical attack. The experimental results however have no relation to
electrochemical environment of an operating fuel cell. In addition there is no direct correlation
with cathode or anode interface. By contrast our fuel cell setup shows that the hydrophilic
regions within the membrane structure are more prone to radical species attack in the initial stage
Scheme 5.2: Vulnerable sites for radical attack in Nafion® 112 [Circle indicates the vulnerable site]
173 of fuel cell operation followed by decomposition of the hydrophobic main chains during
prolonged exposure to these accelerated degradation conditions. Previous study12,17 has shown
that SO2 and CO2 were observed in the outlet stream of an operating fuel cell which is
presumably due to the loss of ether linkage and sulfonate groups from the membrane side chain
and terminal groups respectively. In the literature Fluoride Emission Rate (FER) has been
frequently used as a measure of Nafion® membrane degradation, but such studies were not done
here because in our segmented cell experimental setup four parallel samples run simultaneously
with a common outlet for the gas stream and fluoride ions detected at the outlet stream represents
a total loss from the membrane due the combined effect of the four samples. Further, the
fluoride release represents the extent of the polymer degradation without any detailed
information on the nature of attack and the regions of the membrane undergoing degradation at
any given operating condition.
5.3.6 Anode-side durability tests: As explained in the experimental section, anode side durability
test involved the passage of H2 over catalyzed anode and O2 over non-catalyzed gas diffusion
layer (GDL) in order to investigate membrane degradation on the anode side due to O2
permeation through the membrane. In this test only 30% Pt/C and 60% Pt/C catalysts were
chosen for anode side durability tests since the PtCo/C alloys are cathode relevant catalysts. The
membrane samples were obtained after subjecting them to two types of conditions: (i) holding
the potential of the working electrode in the MEA at possible anode electrode overpotential of
0.1 V for 48 hours and (ii) in open circuit voltage (OCV) condition for 48 hours.
174 Table 5.10: Effect of anode side durability with O2/H2 for 48 hours at 1 atmosphere pressure as a function of temperature (60°°°°C and 80°°°°C), polarization potential (OCV, 0.1 V), and choice of anode electrocatalysts (30% Pt/C, 60% Pt/C). Listed are the decreases in proton conductivity σσσσ (S/cm-1) and ion exchange capacity ηηηη (mEq/g) of Nafion® 112 membrane.
Table 5.10 shows the decrease in proton conductivity and ion exchange capacity after
anode side durability tests performed at 60°C and 80°C at OCV and 0.1 V vs. RHE. Decreases in
conductivity and ion exchange capacity are significantly lower than those observed in the
cathode side tests as shown in Table 5.10 after this 48 hours test indicating that for the duration
of the experiment performed there is no significant degradation on the anode side. This result is
not surprising since the O2 permeation rate through Nafion® 112 membrane is relatively low and
our previous study on hydrocarbon membranes such as sulfonated poly ether sulfone (SPES) and
Nafion 1135 indicate similar low level of degradation on the anode side4. Also, considering the
fact in this experiment the non-catalyzed GDL on the cathode side does not consume O2, as a
result the test is specific to probing the interaction of adsorbed hydrogen on a working catalyzed
GDL and the effect of crossover oxygen. A separate test wherein hydrogen oxidation occurs in
significant rate (higher current density) in the same O2 crossover environment is a case for future
Catalyst T (°C) WE Potential
after before
before
σ σ
σ
−
(%)
after before
before
η η
η
−
(%)
30% Pt/C
60 OCV --- ---
0.1 V 2 4
80 OCV 2 3
0.1 V 3 3
60% Pt/C
60 OCV 4 4 0.1 V 5 6
80 OCV 3 7 0.1 V 4 10
175 work. In a prior long-term fuel cell performance study on radiation-grafted- FEP-g-polystyrene-
type membranes, Buchi et al.14 reported that the rate of radical initiated degradation increases
with increasing gas crossover. They also claimed that gas crossover is a prominent factor for the
degradation process, especially under OCV conditions; however, whether oxygen or hydrogen
crossover is the predominant contributor could not be ascertained in their regular fuel cell setup.
In our case, the suspected formation of radicals due to hydrogen crossover to the cathode side
may be eliminated as a source of peroxides at the cathode, because no Pt reaction sites were
present in the O2 side of our experimental setup (non catalyzed GDL). It has also been shown
earlier that degradation of membrane due to either H2 crossover to the cathode and/or O2
crossover to the anode side result in only less than 3% loss in efficiency due to slower diffusion
rates of these gases through the membrane12. Also Liu et al.31 showed that for short periods of
durability tests, H2 crossover to the cathode side is very minimal and thereby ruling out the
possibility of significant degradation of the membrane at the anode side. Finally, regarding the
anode side durability tests the present results cannot rule out the possible occurrence of radical
catalyzed membrane degradation at the anode side when anode side is subject to higher current
density (hydrogen oxidation conditions) in longer testing periods and it is here found that the rate
of degradation on the cathode side is much higher than that on the anode side under the
accelerated conditions used in this study.
5.3.7: Interpretation Based on Mechanically Coupled Vibrational Modes:
Recently Smotkin et al 65-66 showed based on a combination of IR spectroscopy and DFT
calculations that the sulfonate functional groups and the ether side chains cannot be considered
as pure vibrational modes. Rather, the internal vibrational coordinates of the sulfonate group and
176 the ether band ‘A’ are mechanically coupled to each other. In the light of new evidence from
Smotkin et al65-66, the vibrational band at 969 cm-1 which was classically assigned, based on pure
vibration mode, to the ether band ‘A’ represents the sulfonate symmetric stretching as the
dominant mode and the ether band ‘A’ is only a minor contributor.66 Further, the 1059 cm-1 band
is assigned to the asymmetric stretching of ether band ‘A’ as the dominant mode and the
sulfonate symmetric stretch is only a minor contributor.66 Due to their mechanical coupling, the
ether band ‘A’ and the sulfonate symmetric bands always shift and diminish together. This is in
fact what is observed in the degradation study shown above. Given the mechanical coupling
between the ether band ‘A’ and the sulfonate functional groups, durability experiments clearly
show that both these are affected proportionately. Ether band ‘A’ and the sulfonate groups are
observed to be the most vulnerable sites for peroxide radical initiated degradation.
5.4 Conclusions:
A Novel accelerated technique was used to investigate and correlate the peroxide
generation at an electrode/electrolyte interface from the perspective of radical initiated
perfluorinated membrane degradation as a function of choice of electrocatalyst, catalyst loading
on carbon support, operating overpotential, temperature, and presence of alloying element on the
surface against Pt rich outer layer. Membrane degradation process was also separately studied in
half cell configurations so that the two formerly proposed PEM degradation mechanisms could
be evaluated individually without interference. Peroxide generation observed on the ring
electrode of a RRDE for various electrocatalysts used in this study showed a one-on-one relation
with the level of degradation of perfluorinated membrane via local Fenton type reactions at the
cathode-membrane interface of an operating fuel cell due to the simultaneous 2e- pathway of
177 ORR along with the more predominant 4 e- reduction to water. Cleavage of the side chain ether
linkage, which intrudes into the hydrophilic ionic cluster, is found to be the key initiator of
conductivity and ion exchange capacity loss. Prolonged durability testing leads to breakage of
certain sections of fluorocarbon backbone as observed in the stabilization of ion exchange
capacity vs. a more linear decline in proton conductivity. Normal operating cathode
overpotentials of 0.6 and 0.7 V vs. RHE lead to lower membrane degradation relative to higher
overpotentials of 0.4 V. this is directly correlated to peroxide yields measured independently in
RRDE. Higher the loading of catalyst on carbon support and corresponding larger the particle
size results in higher peroxide yield and consequent higher membrane degradation as shown by
the comparison between 60% Pt/C and 30% Pt/C used in this study. PtCo/C alloy catalyst with
enrichment of surface Co gives lower peroxide current and maintains lower level of membrane
degradation as shown by the comparison between Pt2Co/C and Pt3Co/C. Temperature effects on
membrane degradation was found to be linear with higher ORR activity and consequent higher
peroxide generation at the interface. Degradation at the anode side due to oxygen crossover
through the membrane was found to be insignificant relative to cathode side degradation within
the duration of the experiments performed here. However these tests represent the narrow
confines of interaction of absorbed hydrogen on Pt and its interaction with crossover oxygen.
5.5 Acknowledgements
The authors deeply appreciate the financial assistance of the Army Research Office under
a single investigator grant. The authors also gratefully acknowledge the supply of
electrocatalysts from BASF-fuel cells (Somerset, NJ, USA). Assistance from Dr. Freeman Chen
is acknowledged for assistance provided during IR measurements for membrane samples.
178 5.6 References
(1) Xie, J.; Wood III, D. L.; Wayne, D. M.; Zawodzinski, T. A.; Atanassov, P.; Borup, R. L. Journal of The Electrochemical Society 2005, 152, A104. (2) Steck, A. E. Montreal, 1995, p 74. (3) Baldwin.R; Pham.M; Leonida.A; McElroy.J; Nalette.T Journal of Power Sources 1990, 29, 399. (4) Zhang.L; Mukerjee.S Journal of The Electrochemical Society 2006, 153, A1062. (5) Ramani, V.; Kunz H, R.; Fenton, J. M. Journal of Power Sources 2005, 152, 182. (6) Yeager, E. Journal of Molecular Catalysis 1986, 38, 5. (7) Liu, W.; Zuckerboard, D. Journal of The Electrochemical Society 2005, 152, A1165. (8) Scherer, G. G. Berichte der Bunsen-Gesellschaft 1990, 94, 1008. (9) Watanabe, M.; Uchida, H.; Emori, M. Journal of The Electrochemical Society 1998, 145, 1137. (10) Tarasevich, M. R.; Sadkowski, A.; Yeager, E. Oxygen Electrochemistry; Plenum Press: New York, 1983; Vol. 7. (11) Inaba, M.; Yamada, H.; Tokunaga, J.; Tasaka, A. Electrochemical and Solid State Letters 2004, 7, A474. (12) Laconti, A. B.; Hamdan, M.; McDonald, R. C. Mechanisms of membrane degradation; John Wiley & Sons, Ltd: New York, 2003; Vol. 3. (13) Wang, H.; Capuano, G. A. Journal of The Electrochemical Society 1998, 145, 780. (14) Buchi, F. N.; Gupta, B.; Haas, O.; Scherer, G. G. Electrochimica Acta 1995, 40, 345. (15) Pozio, A.; Silva, R. F.; De Francesso, M.; Giorgi, l. Electrochimica Acta 2003, 48, 1543. (16) Qiao, J.; Saito, M.; Hayamizu, K.; Okada, T. Journal of The Electrochemical Society 2006, 153, A967. (17) Kinumoto, T.; Inaba, M.; Nakayama, Y.; Ogata, K.; Umebayashi, R.; Tasaka, A.; Iriyama, Y.; Abe, T.; Ogumi, Z. Journal of Power Sources 2006, 158, 1222. (18) Okada.T In Handbook of Fuel Cells - Fundamentals, Technology and Applications; Vielstich, W., Gasteiger, H. A., Lamn, A., Eds.; John Wiley & Sons: New York, 2003; Vol. 3, p 627. (19) Maletzky, P.; Bauer, R.; Lahnsteiner, J.; Pouresmael, B. Chemosphere 1999, 38, 2315. (20) Niki, E. Chemistry of Active Oxygen Species; Center of Academic Publications of Japan: Tokyo, 1990. (21) Schumb, W. C.; Satterfield, C. N.; Wentworth, R. L. Hydrogen Peroxide; Reinhold Pub. Co.: New York, 1955. (22) Wilkinson, D. P.; St-Pierre, J. In Handbook of Fuel Cells - Fundamentals, Technology and Applications; Vielstich, W., Gasteiger, H. A., Lamn, A., Eds.; John Wiley & Sons, Ltd: New York, 2003; Vol. 3, p 611. (23) Knights, S. D.; Colbow, K. M.; St-Pierre, J.; Wilkinson, D. P. Journal of Power Sources 2004, 127, 127.
179 (24) Endoh, E.; Terazono, S.; Widjaja, H.; Takimoto, Y. Electrochemical and Solid State Letters 2004, 7, A209. (25) Murthi, V. S.; Urian, C. R.; Mukerjee, S. Journal of Physical Chemistry B 2004, 108, 11011. (26) Liu, R.; Smotkin, E. S. Journal of Electroanalytical Chemistry 2002, 535, 49. (27) Urian, R. C.; Gulla, A. F.; Mukerjee, S. Journal of Electroanalytical Chemistry 2003, 554-555, 307. (28) Endoh, E.; Terazono, S.; Widjaja, H.; Takimoto, Y. Electrochemical and Solid State Letters 2004, 7, A209. (29) Hodgdon, R. B.; Boyack, J. R.; LaConti, A. B. Report TIS 65DE 5, General Electric Company, 1966. (30) Zhang, L.; Ma, C.; Mukerjee, S. Journal of Electroanalytical Chemistry 2004, 568, 273. (31) Liu, W.; Ruth, K.; Rusch, G. Journal of Materials for Materials for Electrochemical Systems 2001, 4, 227. (32) Gasteiger, H. A.; Kocha, S. S.; Sompalli, B.; Wagner, F. T. Applied Catalysis B: Environmental 2005, 56, 9. (33) Gasteiger, H. A.; Panels, J. E.; Yan, S. G. Journal of Power Sources 2004, 127, 162. (34) Williams, M. V.; Kunz, H. R.; Fenton, J. M. Journal of The Electrochemical Society 2005, 152, A635. (35) Guo, Q.; Pintauro, P. N.; Tang, H.; A.O'Connor Journal of Membrane Science 1999, 154, 175. (36) Yu, J.; Yi, B.; Xing, D.; Liu, F.; Shao, Z.; Fu, Y.; Zhang, H. Physical Chemistry Chemical Physics 2003, 3, 611. (37) Yeager, E. Electrochimica Acta 1984, Vol. 29, 1527. (38) Maletzky, P.; Bauer, R.; Lahnsteiner, J.; Pouresmael, B. Chemosphere 1999, 10, 2315. (39) Parthasarathy, A.; Martin, C. R.; Srinivasan, S. J. Electrochem. Soc. 1991, 138, 916. (40) Mauritz, K. A.; Moore, R. B. Chemical Review 2004, 104, 4535. (41) Heitner-Wirguin, C. Journal of Membrane Science 1996, 120, 1. (42) Ma, C.; Zhang, L.; Mukerjee, S.; Ofer, D.; Nair, B. Journal of Membrane Science 2003, 219, 123. (43) Suryanarayana, C.; Norton, M. G. X-Ray Diffraction A Practical Approach; Plenum Press: New York, 1998. (44) Fuller, T. F.; Luczak, F. J.; Wheeler, D. J. Journal of the Electrochemical Society 1995, 142, 1752. (45) Abd El Rehim, S. S.; El Basosi, A. A.; El Zein, S. M.; Osman, M. M. Collect. Czech. Chem. Commun. 1994, 59, 2383. (46) Lima, F. H. B.; Janete Giz, M.; Ticianelli, E. A. Journal of Brazilian Chemical Society 2005, 16, 328. (47) Mukerjee, S.; Srinivasan, S. J. Electroanal. Chem. 1993, 357, 201. (48) Mukerjee, S.; McBreen, J.; Srinivasan, S. Proc. - Electrochem. Soc. 1996, 95-26, 38.
180 (49) Roques, J.; Anderson, A. B.; Murthi, V. S.; Mukerjee, S. Journal of the Electrochemical Society 2005, 152, E193. (50) Anderson, A. B.; Roques, J.; Mukerjee, S.; Murthi, V. S.; Markovic, N. M.; Stamenkovic, V. Journal of Physical Chemistry B 2005, 109, 1198. (51) Paulus, U. A.; Wokaun, A.; Scherer, G. G.; Schmidt, T. J.; Stamenkovic, V.; Radmilovic, V.; Markovic, N. M.; Ross, P. N. Journal of Physical Chemistry B 2002, 106, 4181. (52) Grgur, B. N.; Markovic, N. M.; Ross, P. N. Canadian Journal of Chemistry 1997, 75, 1465. (53) Paulus, U. A.; Schmidt, T. J.; Gasteiger, H. A.; Behm, R. J. Journal of Electroanalytical Chemistry 2001, 495, 134. (54) Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications; Second ed.; John Wiley & Sons, Inc., 2001. (55) Cable, K. M.; Mauritz, K. A.; Moore, R. B. Journal of Polymer Science: PArt B: Polymer Physics 1995, 33, 1065. (56) Liang, Z.; Chen, W.; Liu, J.; Wang, S.; Zhou, Z.; Li, W.; Sun, G.; Xin, Q. Journal of Membrane Science 2004, 233, 39. (57) Gruger, A.; Regis, A.; Schmatko, T.; Colomban, P. Vibrational Spectroscopy 2001, 26, 215. (58) Heitner-Wirguin, C. Polymer 1979, 20, 371. (59) Hsu, W. Y.; Gierke, T. D. Journal of Membrane Science 1983, 13, 307. (60) Yeager, H. L.; Steck, A. Journal of The Electrochemical Society 1981, 128, 1880. (61) Falk, M. Canadian Journal of Chemistry 1980, 58, 1495. (62) Conley, R. T. Infrared Spectroscopy; Allyn and Bacon, Inc.: Boston, MA, 1972. (63) Perfluorinated ionomer membranes; Eisenberg, A.; Yeager, H. L., Eds.; American Chemical Society: Washington. DC., 1982. (64) Inaba, M. In 14th International Conference on the Properties of Water and Steam in Kyoto Kyoto, Japan. (65) Kendrick, I.; Kumari, D.; Yakaboski, A.; Dimakis, N.; Smotkin, E. S. J. Am. Chem. Soc. 2010, 132, 17611. (66) Webber, M.; Dimakis, N.; Kumari, D.; Fuccillo, M.; Smotkin, E. S. Macromolecules (Washington, DC, United States) 2010, 43, 5500.
181
Chapter 6
Alkaline Anion Exchange Membrane Fuel Cell Studies, Thesis Summary and Future
Directions
6.1. Alkaline Membrane Fuel Cell (AMFC) Studies:
In alkaline medium, Oxygen Reduction Reaction (ORR) has several mechanistic
advantages; given the high pH condition several non-noble metals are also stable. Although the
development of electrocatalysts for ORR has its intrinsic challenges, from a materials
perspective a wider range of catalysts are available in alkaline medium. On the contrary, in
acidic medium acid-stability criterion limits the catalyst choice to expensive and noble platinum
based materials. Most electrocatalysts are typically studied using rotating ring-disk electrode
techniques (RRDE), where the so called ‘flooded-electrolyte’ condition is established. In this
case the liquid electrolyte wets the catalyst layer and promotes almost complete utilization of the
active sites. On the contrary, under fuel cell operating conditions solid ionomer membranes that
transport either protons or hydroxide anions are used as electrolytes. In this case, the absence of
liquid electrolyte gives rise to severe limitations in ionic transport within the catalyst layer. In
order to extend the reaction zone deep into the catalyst layer the design of appropriate electrode
architectures plays a crucial role. This limitation has typically been overcome by using
solubilized form of ionomers as a binder in the catalyst layer. This ionomer layer establishes the
so called three-phase boundary in the electrode structure by promoting electronic, ionic and
reactant transport to the active site. While these ionomer solutions have been very successfully
employed in conventional electrodes composed of 20-60% loaded Pt/C catalysts, there are some
severe drawbacks when they are used in the present generation of non-platinum group metal
(non-PGM) catalysts such as pyrolyzed macrocycles. For example, 30% Pt/C catalyst at a typical
182 cathode loading of 0.5 mgPt/cm2 yields a catalyst layer thickness of ~10 µm. The present
generation of pyrolyzed macrocycle based catalyst has a metal loading of <1% by weight on
carbon support. To obtain sufficient metal content while fabricating the electrode, the thickness
of the catalyst layer tends to reach ~50-100 µm. Such high thicknesses cause significant mass
transport issues within the catalyst layer. This requires ionomer solutions with high conductivity
and/or development of radically different electrode architectures that promote mass transport. An
alternative route to overcome this drawback would be to increase the active site density in the
present generation of non-PGM catalysts. Therefore a catalyst, irrespective of its performance
under RRDE conditions, does not necessarily qualify for fuel cell applications unless its
performance can be translated into a successful electrode material.
Recent successes in the development of novel alkaline anion exchange membranes
(AAEM) that transport hydroxide anions from the cathode to anode have reinvigorated the
alkaline fuel cell (AFC) technology. The catalysts developed for oxygen reduction in alkaline
medium have been tested using Tokuyama anion exchange membrane in alkaline fuel cells.
While the preliminary results shown here are very promising, there are several significant
challenges that need to be overcome. A brief summary of the alkaline membrane fuel cell
(AMFC) results is given here along with the possible future directions.
6.1.1. H2/O2 Alkaline Membrane Fuel Cell:
Figure 6.1 shows fuel cell performance of Pt/C catalysts used in both anode and cathode
at a loading of 0.5 mgPt/cm2 using 100% humidified H2 and O2 gas feeds. Tokuyama A201
alkaline membrane and Tokuyama AS4 ionomer solutions were used. The hydroxide anion
conductivity of Tokuyama A201 membrane is 38 mS cm-1 at 90% relative humidity and 23°C
183
temperature.1 At an operating cell voltage of 0.7 V, power density values obtained correspond to
~0.15 W/cm2geo. A peak power density of ~0.24 W/cm2
geo was obtained. While this performance
is promising, it is lower than an analogous proton exchange membrane (PEM) fuel cell, where
power densities of ~0.5 to 0.7 W/cm2geo is typically obtained. Further, even at high current
densities of 1 A/cm2geo ohmic resistance is dominating and no evident mass transport region
could be discerned. This indicates that the membrane resistance or the low OH conductivity is
one of the major limiting factors. Figure 6.2 shows the AMFC performance of non-PGM
catalysts such as FeTPP/C and CuFe/C discussed in detail in Chapter 4. At 50°C, power density
of 0.06 W/cm2geo at 0.6 V is obtained with FeTPP/C cathode. This is a promising result with non-
PGM cathodes in AMFC, considering the fact that alkaline membrane and ionomer solutions are
only in their nascent stage of development. Further, this performance of FeTPP/C cathode in
i [A/cm2geo]
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
EC
ell [
V]
0.0
0.2
0.4
0.6
0.8
1.0
1.2
PD
[mW
/cm
2 geo
]
0
50
100
150
200
250
Polarization CurvePower Density Curve
Figure 6.1: AMFC polarization and power density curves taken at 50°C cell temperature, H2/O2 gas feeds at 28/28 psig, and 100% relative humidity. Anode and Cathode: 0.5 mgPt/cm2, Membrane: Tokuyama A-201 and Ionomer: Tokuyama AS4.
184
AMFC equals the activity of similar class of cathode materials in PEMFC from the literature.2
From an electrocatalyst standpoint, there are two aspects that need future investigations:
i) Active site density of FeTPP/C catalyst needs to be increased at least by three- to four-
fold.
ii) Long term stability of FeTPP/C catalyst and reasons for loss in activity over time.
6.1.2: Direct Ethanol Alkaline Membrane Fuel Cell:
Compressed gaseous hydrogen is a very attractive fuel proposed to be used for fuel cells.
From an electrochemistry standpoint, hydrogen oxidation/reduction reactions are kinetically very
facile. However, there are long-standing safety considerations related to storage and
transportation of compressed hydrogen cylinders. High energy density liquid fuels such as
methanol and ethanol are alternative fuels that are very safe to handle. Further, the existing
i [A/cm2geo]
0.0 0.1 0.2 0.3 0.4 0.5
EC
ell [
V]
0.0
0.2
0.4
0.6
0.8
1.0
1.2
FeTPP/CCuFe/C
i [A/cm2geo]
0.0 0.1 0.2 0.3 0.4 0.5 0.6
PD
[mW
/cm
2 geo]
0
20
40
60
80
100
120
FeTPP/CCuFe/C
Polarization Curves Power Density Curves
2.9mgFeTPP/cm2
3.1mgCuFe/cm2
Figure 6.2: AMFC polarization (Left) and power density (Right) curves taken at 50°C cell temperature, H2/O2 at 28/28 psig, and 100% relative humidity. Anode: 0.5 mgPt/cm2, Cathode: FeTPP/C or CuFe/C. Membrane: Tokuyama A-201 and Ionomer: Tokuyama AS4.
185 infrastructure for the storage and distribution of petroleum products can be easily modified to
suit alternative fuels such as methanol and ethanol. Use of such liquid fuels is highly sought
despite the sluggishness in electrochemical oxidation of methanol/ethanol.
Current Density [mA/cm2geo]
10 20 30 40
PD
[mW
/cm
2 geo]
0
1
2
3
4
5
6
0.25M KOH + 1M EtOH1M EtOH
Polarization Curves
Current Density [mA/cm2geo]
0 10 20 30 40
EC
ell [
V]
0.0
0.2
0.4
0.6
0.8
1.0
0.25M KOH + 1M EtOH1M EtOH
Anode Fuel: 8ml/minCathode: O2
50oC, 100% RH
Pt Anode/A-201/Pt Cathode
Figure 6.3: AMFC polarization (Left) and power density (Right) curves at 50°C temperature, and 100% RH. Anode: 2 mgPt/cm2 and Cathode: 1.0 mgPt/cm2, Membrane: Tokuyama A-201, Ionomer: Tokuyama AS4. Anode Feed: Ethanol with & without KOH.
Power Density Curves
Current Density [mA/cm2geo]
0 20 40 60 80 100
Pow
er D
ensi
ty [m
W/c
m2 ge
o]
0
5
10
15
20
25
0.25M KOH + 1M EtOH1M EtOH
Polarization Curves50oC - 100% RH
Current Density [mA/cm2geo]
0 20 40 60 80
EC
ell [
V]
0.0
0.2
0.4
0.6
0.8
1.0
0.25M KOH + 1M EtOH1M EtOH
Anode Fuel: 8ml/minCathode: O2
PtRu Anode/A201/Pt Cathode
Figure 6.4: AMFC polarization (Left) and power density (Right) curves taken at 50°C cell temperature, and 100% relative humidity conditions. Anode: 4 mgPtRu/cm2 and Cathode: 1.0 mgPt/cm2, Membrane: Tokuyama A-201 and Ionomer: Tokuyama AS4. Anode Feed: Ethanol with & without KOH.
186
Figure 6.3 shows the AMFC performance using ethanol as the anode feed. A high loading of 2
mgPt/cm2 on the anode was used to overcome the electrochemical sluggishness in ethanol
oxidation. The cell was operated at 50°C, 100% relative humidity conditions with a cathode
loading of 1 mgPt/cm2. Anode feed was 1 M ethanol with or without 0.25 M KOH as a
supporting electrolyte. Several interesting observations are made. The open circuit potential
(OCP) in the presence of KOH is ~0.93 V whereas in the absence of KOH only 0.6 V is
obtained. There is a steep drop in OCP value in the absence of the supporting alkaline electrolyte
in the anode feed. Further, the performance of the cell is very poor and a peak power density of
only ~5mW/cm2 is achieved even with the presence of KOH in the anode feed. Figure 6.4 shows
a direct ethanol fuel cell performance curve using an anode consisting of 4 mgPtRu/cm2 catalyst
with all other parameters similar to Figure 6.3. While the cell performance has increased to a
PtRu Anode/A201/FeTPP Cathode
Current Density [mA/cm2geo]
0 10 20 30 40 50
EC
ell [
V]
0.0
0.2
0.4
0.6
0.8
1.0
0.25M KOH + 1M EtOH1M EtOH
Power Density Curves
Current Density [mA/cm2geo]
0 10 20 30 40 50
Pow
er D
ensi
ty [m
W/c
m2 ge
o]
0
2
4
6
8
10
12
14
0.25M KOH + 1M EtOH1M EtOH
Anode Fuel: 8ml/minCathode: O2Polarization Curves
50oC - 100% RH
Figure 6.5: AMFC (Left) and power density (Right) curves taken with different anode feeds at 50°C cell temperature, and 100% relative humidity conditions. Anode: 4 mgPtRu/cm2 and Cathode: 3.0 mgFeTPP/cm2, Membrane: Tokuyama A-201 and Ionomer: Tokuyama AS4.
187 peak power density of ~20 mW/cm2 in the presence of KOH, the most intriguing factor is the
low OCP value and the poor performance in the absence of KOH. These are some serious issues
that need future studies. Figure 6.5 shows a similar fuel cell performance with FeTPP/C catalyst
in the cathode. This experiment was performed in order to ensure that the poor performance does
not arise from platinum based cathode catalysts, because ethanol crossover from anode to
cathode could poison the cathode catalyst. Non-PGM cathode catalysts such as FeTPP/C are
known to be tolerant to poisoning by methanol/ethanol.3 Unfortunately, the use of non-PGM
cathode did not change the prospects of fuel cell performance, clearly indicating that the limiting
factor is anodic oxidation of ethanol.
6.2. Preliminary Investigations of the Alkaline Anode-Membrane Interface:
Our preliminary studies indicate that understanding the electrical and structural aspects of
the electrochemical double-layer at an alkaline anode-membrane interface is extremely crucial.
Figure 6.6 shows the cyclic voltammetry (CV) profile of Pt and PtRu based anodes in an AMFC.
Experiments were performed in argon saturated 0.25 M KOH on the anode. H2 feed on the Pt
cathode was used as reference electrode. As seen in Figure 6.6, in the presence of liquid KOH
electrolyte electrochemical features of Pt and PtRu catalysts such as hydrogen
adsorption/desorption and oxide formation are clearly evidenced. Figure 6.7 shows the ethanol
oxidation profiles on Pt and PtRu anodes in AAEM fuel cell using H2 on Pt cathode as the
reference electrode. In the presence of KOH electrolyte, high current densities due to ethanol
oxidation on both Pt and PtRu are obtained. However, in the absence of KOH electrolyte drastic
decrease in the activity is observed. For instance in the presence of KOH, on the Pt anode the
onset potential of ethanol oxidation is 0.4 V vs. RHE and a well defined limiting current region
188
is obtained at 1.0 V vs. RHE. On the other hand in the absence of KOH, onset potential is shifted
to > 0.8 V and no diffusion limited current region is observed even at very high potentials.
Understanding the source of such a high onset potential in the absence of excess KOH is very
important.
Ethanol crossover from anode to cathode is typically considered to be a significant
problem in PEM fuel cells. Experiments were performed in order to understand the effect of
ethanol crossover in an operating AMFC. Figure 6.8 summarizes these results. Interestingly,
ethanol crossover was found to be significant only in the presence of KOH but not in the absence
of KOH. While higher crossover currents are observed in the presence of KOH, this could be due
to either i) increased ethanol crossover to the cathode per se or 2) due to improved activity of the
cathode catalyst in oxidizing ethanol. While these two points are debatable, the important aspect
Cyclic Voltammetry of Anode at 5mV/sAnode Feed: 0.25M KOH
Cathode: H2 Ref.
E Vs RHE0.0 0.2 0.4 0.6 0.8 1.0 1.2
i [m
A/c
m2 ge
o]
-30
-20
-10
0
10
20
30
Pt AnodePtRu Anode
Cell Temp: 50oC
27 m2/gPt
17 m2/gPtRu
Figure 6.6: Cyclic voltammetry (CV) of Pt and PtRu anodes in AAEM fuel cell taken at 50°C cell temperature using H2 feed on the Pt cathode as a reference electrode. Anode feed was 0.25 M KOH electrolyte at 5 ml/min flow rate. Scan rate: 5 mV/s
189
to understand here is that the poor AMFC cell performance (evidenced in Figures 6.3 to 6.5) in
the absence of KOH in the anode feed is not due to ethanol crossover. This further corroborates
the necessity to understand the alkaline anode-membrane interface.
In order to understand the anode-membrane interface further, AC impedance study of an
operating AMFC was performed. These results are summarized in Figure 6.9. AMFC was
operated with 1 M ethanol on the anode with and without 0.25 M KOH as a supporting
electrolyte. Humidified O2 was passed on the cathode side. Impedance data was measured at a
cell potential of 0.4 V and at 50 °C temperature. For a 5 cm2 electrode area, the membrane
resistance from the high frequency intercept decreased from 0.591 Ω (absence of KOH) to 0.109
Ω (presence of KOH). This is as expected since the presence of KOH increases the conductivity
of the membrane. More importantly, the charge transfer resistance increased significantly from
1.31 Ω (presence of KOH) to 16.85 Ω (absence of KOH). This clearly indicates that the sluggish
charge transfer kinetics of ethanol oxidation is the root cause for poor performance of AMFC
Ethanol Oxidation on Pt Anode
E Vs RHE
0.0 0.2 0.4 0.6 0.8 1.0 1.2
i [m
A/c
m2 g
eo]
-20
0
20
40
60
80
100
0.25M KOH + 1M EtOH1M EtOH
Anode: 2mgPt/cm2
Ethanol Oxidation on PtRu Anode
E Vs RHE
0.2 0.4 0.6 0.8 1.0
i [m
A/c
m2 g
eo]
-40
-20
0
20
40
60
80
100
120
140
160
0.25M KOH + 1M EtOH1M EtOH
Anode: 4mgPtRu/cm2
Figure 6.7: Ethanol oxidation CV on Pt and PtRu anodes in AMFC. Experiments performed at 50°C temperature using H2 feed on the Pt cathode as a reference electrode. Anode feed: 1 M ethanol (EtOH) with and without 0.25 M KOH electrolyte. Scan rate – 5 mV/s.
190 cells in the absence of KOH. The reason for this poor performance in the absence of KOH is due
to the specific adsorption of quaternary ammonium cations on the active site.
Ethanol CrossoverCurrent
at the Cathode
Potential [V Vs RHE]
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Cur
rent
Den
sity
[m
A/c
m2 geo
]
-10
0
10
20
30
40
0.25M KOH + 1M EtOH1M EtOH
50oC20 mV/s
Cathode: N2
Figure 6.8: Ethanol crossover current at the cathode in an AMFC. Experiments performed at 50° C temperature, 20 mV/s scan rate. Anode feed: 1 M ethanol (EtOH) with and without 0.25 M KOH electrolyte. Cathode: Humidified N2.
Z' [Ohm]
0 2 4 6 8 10 12 14 16 18 20
Z''
[Ohm
]
0
1
2
3
4
5
6
7
0.25M KOH + 1M EtOH1M EtOH
Z' [Ohm]
0.0 0.3 0.6 0.9 1.2 1.5
Z''
[Oh
m]
0.0
0.2
0.4
0.6
0.8
1.0
1.2
Impedance Data Measured at 0.4VPtRu Anode Feed: 8 ml/min
Cathode: O2 - Cell Temp: 50oCImpedance Frequency - 20kHz to 0.1mHz
Figure 6.9: AC impedance profile measured in an AMFC. Experiments performed at 50° C temperature, 20 mV/s scan rate. Anode feed: 1 M ethanol (EtOH) with and without 0.25 M KOH electrolyte. Cathode: Humidified O2.
191 6.2.1. Effect of Specific Adsorption of Quaternary Ammonium Cations:
The sulfonate anions in Nafion® were shown in the literature to specifically adsorb on
the active site in PEM fuel cells.4-5 Based on our on-going studies it is clearly understood that the
quaternary ammonium cations adsorb very strongly on Pt sites in alkaline medium, than do
sulfonate anions in acidic medium. These interpretations are based on the loss in
electrochemically active surface area and the effect on ethanol oxidation shown above. This is
not surprising considering the fact that in alkaline medium the working electrode potential range
on an SHE scale decreases by 59 mV per pH unit. This causes the excess charge on the electrode
surface to be more negative in alkaline medium compared to that in acidic medium. Therefore,
the positively charged quaternary ammonium cations are expected to adsorb strongly on Pt in
alkaline medium compared to the negatively charged sulfonate anions in acidic medium.
The effect of specific adsorption of quaternary ammonium ions is two-fold:
i) Loss in active electrochemical surface area
ii) Electrostatic double-layer effect
Figure 1.10 shows an illustration of the potential drop in the double layer in the absence and
presence of specific adsorption of quaternary ammonium cations. In the absence of specific
adsorption, double-layer structure and the potential drop across it is given by the Guoy-
Chapmen-Stern (GOS) description. At any potential, say E1, above the potential of zero charge
(EPZC) the plane of closest approach for the hydroxide anions is the electrode surface. According
to GOS treatment, the compact part of the double-layer is characterized by a linear drop in
potential from the electrode surface to the OHP. In this case, let the potential at the OHP be equal
192
to φ2(E1). However, in the presence of specific adsorption of quaternary ammonium cations a
steep increase in potential is likely to be present between the electrode surface and the IHP. In
this case, let the potential at the OHP be equal to φ2(E2), such that φ2(E2) > φ2(E1). This higher φ2
potential at the OHP in the presence of specific adsorption would cause a higher population of
OH¯ anions at the OHP compared to that in the absence of specific adsorption. However, the
potential drop between the electrode surface and IHP in the presence of specific adsorption is not
favorable for the migration of OH anions to the electrode surface. This prevents the supply of
Figure 6.10: Schematic illustration of the electrochemical double-layer potential drop in the presence and absence of quaternary ammonium cations in alkaline medium.
193 hydroxide anions required for ethanol oxidation at the electrode surface. In order for the ethanol
oxidation reaction to take place, the electrode potential has to be increased to values higher than
that of E2. This difference E2-E1 is so high that very high anodic potentials are required to
oxidize ethanol in an operating AMFC in the absence of excess KOH as shown in Figure 6.7.
Further, specific adsorption of cations is likely to shift the EPZC of the catalyst further positive
which will exacerbate the situation. However, when excess KOH is present in the anode feed, the
potential drop between the electrode surface and the IHP is overcome or equalized by the higher
concentration of OH anions. This attenuates the effect of specific adsorption quaternary
ammonium cations on alcohol oxidation.
However the effect on Φ2 is likely to be significant only at low coverage of quaternary
ammonium groups. At higher coverage of quaternary ammonium cations and also in the presence
of KOH electrolyte, the effect of specific adsorption on diminishing the electroactive surface
area is likely to be more dominant.
In the analogous situation of PEMFC, sulfonate anions are known to specifically adsorb
on Pt. However, this adsorption is very weak compared to quaternary ammonium cations in
alkaline medium. Moreover the higher mobility of protons in acidic medium equalizes
neutralizes the potential drop between the electrode surface and the IHP.
6.3. Concluding Remarks
Developing alkaline membranes with high hydroxide anion conductivity and durability is
of utmost importance to substantiate the viability of AMFC technology. However, given the
mobility 6-7 of protons (36.25x10-4 cm2 V-1 s-1) and hydroxide anions (20.50 cm2 V-1 s-1) in water
at 25 °C, a pragmatic view would be that the conductivity values of the alkaline membranes are
unlikely to increase significantly over the current state-of-the-art. So, the task of understanding
194 and solving the following challenges from electrocatalyst standpoint assume tremendous
potential for fundamental research. A few of them are pointed out below:
1) Specific adsorption of quaternary ammonium cations
a. Effect on the electrochemically active surface area
b. Losses due to electrostatic double layer effect in Inner-Helmholtz plane
2) Suppressing outer-sphere electron transfer during ORR and promoting direct molecular
O2 adsorption on the active site
3) Increasing active site density in the present class of heat treated Fe-N4 catalyst systems
4) Effect of carbonate formation on membrane conductivity and electrocatalysis.
6.4 References
(1) Yanagi, H.; Fukuta, K. In Book of Abstracts, The Electrochemical Society 214th Meeting Honolulu, Hawaii, 2008. (2) Bashyam, R.; Zelenay, P. Nature (London, U. K.) 2006, 443, 63. (3) Sun, G. Q.; Wang, J. T.; Savinell, R. F. J. Appl. Electrochem. 1998, 28, 1087. (4) Kendrick, I.; Kumari, D.; Yakaboski, A.; Dimakis, N.; Smotkin, E. S. J. Am. Chem. Soc. 2010, 132, 17611. (5) Subbaraman, R.; Strmcnik, D.; Stamenkovic, V.; Markovic, N. M. J. Phys. Chem. C 2010, 114, 8414. (6) Bockris, J. O. M.; Reddy, A. K. N. Modern Electrochemistry; Second ed.; Springer, 1998; Vol. 1 & 2. (7) Duso, A. B.; Chen, D. D. Y. Analytical Chemistry 2002, 74, 2938.
195
Curriculum Vita Name: Nagappan Ramaswamy Birthplace: Vilupuram, Tamil Nadu, India Birth Date: September 3rd, 1984 Education: Northeastern University, Boston, MA Ph.D. in Physical Chemistry, April 2011 Central Electrochemical Research Institute, Anna University, India B.Tech in Chemical and Electrochemical Engineering, May 2005 Professional Experience
• Research Assistant, 2007-2011, Northeastern University • Teaching Assistant, 2005-2007, Northeastern University
Awards
• Outstanding Teaching Assistant Award - Academic year 2006-2007 - Department of Chemistry and Chemical Biology at Northeastern University
• Travel Grant Award - Polymer Electrolyte Fuel Cell (PEFC) 2010 Symposium - 218th The Electrochemical Society (ECS) Meeting – Las Vegas, NV.
Professional Memberships
• The Electrochemical Society Publications
1) 'Non-Noble CuFe Bimetallic Catalyst for Oxygen Reduction in Alkaline Medium: Electrochemical and X-ray Absorption Spectroscopic Investigations', N. Ramaswamy, H. Miller, X. Ren, S. Mukerjee, Submitted (2011)
2) 'A Novel CuFe/C Catalyst for the Oxygen Reduction in Alkaline Media', Q. He, X. Yang, X. Ren, B. E. Koel, N. Ramaswamy, S. Mukerjee, R.M. Kostecki, Accepted for publication in Journal of Power Sources (2011)
3) 'Understanding the Origin of Electrochemical Oxygen Reduction Activity of Pyrolyzed Iron Porphyrin Catalysts: Redox Potential Tuning of the Active Metal Center', N. Ramaswamy, and S. Mukerjee, In Preparation (2011)
4) 'Electrochemical Kinetics and X-ray Absorption Spectroscopic Investigations of Oxygen Reduction on Chalcogen Modified Ruthenium Catalysts in Alkaline Medium', N. Ramaswamy, R. Allen, and S. Mukerjee, Submitted to The Journal of Physical Chemistry C, (2011)
196
5) 'Impact of Double-Layer Structure and Mechanistic Changes during Electrocatalysis of Oxygen Reduction in Alkaline Medium', Nagappan Ramaswamy, and Sanjeev Mukerjee, Submitted to The Journal of Physical Chemistry C, (2011)
6) 'Electrocatalysis of Oxygen Reduction on Non-Precious Metallic Centers at High pH Environments', N. Ramaswamy, and S. Mukerjee, ECS Transactions, 33 (1, Polymer Electrolyte Fuel Cells 10), 1777-1785 (2010)
7) 'Fundamental Aspects of Spontaneous Cathodic Deposition of Ru onto Pt/C Electrocatalysts and Membranes under Direct Methanol Fuel Cell Operating Conditions. An In Situ X-ray Absorption Spectroscopy and Electron Spin Resonance Study', Arruda, Thomas; Shyam, Badri; Lawton, Jamie; Ramaswamy, Nagappan; Budil, David; Ramaker, David; Mukerjee, Sanjeev, The Journal of Physical Chemistry C, 114 (2), 1028, (2009)
8) 'Local Structure of Nanocrystalline Ru1-xNixO2-δ Oxide and its Implications to Electrocatalytic Behavior – an XPS and XAS Study', V. Petrykin, Z. Bastl, K. Macounova, J. Franc, M. Makarova, S. Mukerjee, N. Ramaswamy, I. Spirovova, P. Krtil, The Journal of Physical Chemistry C, 113 (52) 21657, (2009)
9) 'Enhanced Activity and Interfacial Durability Study of Ultra Low Pt Based Electrocatalysts Prepared by Ion Beam Assisted Deposition (IBAD) Method', N. Ramaswamy, T.M. Arruda, W. Wen, N. Hakim, M. Saha, A.Gulla, and S. Mukerjee, Electrochimica Acta, 54 (26), 6756, (2009)
10) 'Carbon-Supported, Selenium-Modified Ruthenium-Molybdenum Catalysts for Oxygen Reduction in Acidic Media', M.J-F. Guinel, A. Bonakdarpour, B. Wang, P. Babu, F. Ernst, N. Ramaswamy, S. Mukerjee and A. Wieckowski, ChemSusChem, 2, 1-8 (2009)
11) 'Degradation Mechanism Study of Perfluorinated Proton Exchange Membrane Under Fuel Cell Operating Conditions', N. Ramaswamy, N. Hakim and S. Mukerjee, Electrochimica Acta, 53, 3279 (2008)
Invention Disclosures 1) 'Catalyst for Electrochemical Applications', E. Schwab, S. Brauninger, A. Panchenko, C.
Querner, O. Unsal, M. Vogt, Q. He, N. Ramaswamy, and S. Mukerjee, Filed in USA on June 02, 2009; Application Number 61/18 3251 (BASF, Ludwigshafen, Germany – Ref# PF 0000062182/RS)
Oral Presentations at Conferences 1) 'Electrocatalysis of Oxygen Reduction on Nonprecious Metallic Centers at High pH
Environments', N. Ramaswamy, and S. Mukerjee, 218th Electrochemical Society Meeting, October 10-15, 2010, Las Vegas, Nevada, United States
2) Novel Electrocatalysts for Direct Oxidation of Alcohols and Oxygen Reduction Reaction in High pH Environments', Q. He, N. Ramaswamy, and S. Mukerjee, 214th Electrochemical Society Meeting, October 12-17, 2008, Honolulu, Hawaii, United States
3) 'Novel Electrocatalysts for Oxygen Reduction in High pH Environments', N. Ramaswamy, Q. He, J. Ziegelbauer, A. Gulla, and S. Mukerjee, 212th Electrochemical Society Meeting, October 7-12, 2007, Washington, DC, United States
4) 'Degradation Mechanism Study of Perfluorinated Polymer Electrolyte Membrane under Fuel Cell Operating Conditions', N. Ramaswamy, N. Hakim and S. Mukerjee, 211th Electrochemical Society meeting, May 6-10, 2007, Chicago, Illinois, United States