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


3.3.1 Physicochemical Characterizations 77


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


3.6 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


4.2.3 X-ray Absorption Spectroscopic Measurements 111


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


4.7 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


5.3.1 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


5.6 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).


3.3.1 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


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


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


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


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.


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


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


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


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


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


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


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


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


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.


4.3.1. 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


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


I R [A

]

0

1e-6

2e-6

3e-6

4e-6

5e-6

0.1M NaOH0.1M HClO4

(a) ORR Polarization Curves


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


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


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


0 10 20 30 40

EC

ell [

V]

0.0

0.2

0.4

0.6

0.8

1.0


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


0 20 40 60 80 100

Pow

er D

ensi

ty [m

W/c

m2 ge

o]

0

5

10

15

20

25


Polarization Curves50oC - 100% RH


0 20 40 60 80

EC

ell [

V]

0.0

0.2

0.4

0.6

0.8

1.0


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


0 10 20 30 40 50

EC

ell [

V]

0.0

0.2

0.4

0.6

0.8

1.0


Power Density Curves


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


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


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


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


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


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


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

1 STUDY OF PERFLUORINATED IONOMER MEMBRANE DEGRADATION A ...703/fulltext.pdf · electrostatic...

Documents

Transcript of 1 STUDY OF PERFLUORINATED IONOMER MEMBRANE DEGRADATION A ...703/fulltext.pdf · electrostatic...