W

,.

 

“H
m.

Airman?»

.5: .f w

 

FEM”...

.2 a!
“av-{5...}? Lu“
t xt‘lJvtvi-E.
‘ havéfiuu

. , a L...»

.. a a 4
3:42;... . ..;..“vu -
. ‘1... . #1325}. ii-

. A 1.2.ur’xfijhvh
1!...“
1:;- 3“ . .1...
1 K.Y‘!I
.- .‘ 3|

z
ӣ51."?

#1.,

c
8:: I":
5....

.3
‘
3...).
5‘ 3“

3.5.5... .
l}. 12:).- 1‘
x. Pin 54..» .1.
3a.”... 3 3.:
i=3?! ‘9.
5.11.1 a... I.:.:
3 .- 731:6!!!
hm 533.52
5.3.1.! it
E. .93.)...

.51.»!
.5}?

II :1 .: 31:3.

2: -11) «.I«x..(nn.

; n L’.313v11~.v
u);

.33.... 3:13;...
3: 3:42.21
38:11:... 7....

7455‘ .I.

z... ‘
.1113}.
1...?-
n: 1: a
In}: 3.11.01

5: ‘

I III. 1.5!;

.rlnn.,s.. ‘ .....!.:.v-
I; I5.

 

a

1 Wears

46-51

This is to certify that the
dissertation entitled

THE FORMATION AND ACTIVITY OF PLATINUM
ADLAYERS ON DIAMOND ELECTRODES FOR
ELECTROCATALYSIS

presented by
JASON ALAN BENNETT

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Chemistry

4M W449“;

- flMajor Pfofessor’s'Signature
fix W } A3! 9006

Date

 

MSU is an Affirmative Action/Equal Opportunity Institution

 

LIBRARY
Michigan State
University

 

 

 

-.-.-,-.-.-.-.-.-.-.-.-,-.- ...

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 p:lCIRC/DateDue.indd-p.1

THE FORMATION AND ACTIVITY OF PLATINUM ADLAYERS ON DIAMOND
ELECTRODES FOR ELECTROCATALYSIS

By

Jason Alan Bennett

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Department of Chemistry

2006

ABSTRACT

THE FORMATION AND ACTIVITY OF PLATINUM ADLAYERS ON DIAMOND
ELECTRODES FOR ELECTROCATALYSIS

By

Jason Alan Bennett

The research described in this dissertation evaluates the potential of diamond as
an advanced carbon electrocatalyst support material. This includes both assessing the
physical and electrochemical properties of the material as well as a comprehensive

investigation into the nature of metal adlayer formation on the material.

The physical and electrochemical properties of boron-doped polycrystalline
diamond thin films, prepared with varying levels of spz-bonded nondiamond carbon
impurity, were systematically investigated. This impurity was introduced through
adjustment of the methane-to-hydrogen source gas ratio used for the deposition.
Increasing the methane-to-hydrogen ratio resulted in an increase in the fraction of grain
boundary, the extent of secondary nucleation, and the amount of spz-bonded nondiamond
carbon impurity. The effect of the source gas ratio on the electrochemical response
towards several well known redox analytes and the oxygen reduction reaction in both
acidic and alkaline media was also investigated. The results demonstrate that the grain
boundaries, and the spz-bonded nondiamond carbon impurity presumably residing there,
can have a significant impact on the electrode reaction kinetics for certain redox systems

and little influence for others.

The morphological and microstructural stability of microcrystalline and
nanocrystalline boron-doped diamond thin film electrodes under conditions observed in
phosphoric acid fuel cells was investigated. The electrodes were exposed to a 2 h period
of anodic polarization in 85% H3PO4 at ~180 °C and 0.1 A/cmz. Catastrophic
degradation was not observed for either type of diamond. The oxidation of the
microcrystalline diamond was limited to the surface, and the effects could be reversed
upon exposure to a hydrogen plasma. The nanocrystalline diamond exhibited similar
responses to well known redox analytes after anodic polarization, however an irreversible
increase in the film capacitance was also observed. This is possibly due to surface pitting

or heterogeneous conductivity in the film resulting in "hot spots."

Electrodeposition of Pt adlayers as a function of both diamond morphology and
diamond surface chemistry was also investigated using a pulsed galvanostatic deposition
approach. Our goal was to minimize the particle size and size dispersion (ca. 5-15 nm)
while maximizing the particle density (ca. 1010 cm’z) and catalytic activity towards the
oxygen reduction reaction. Depositing Pt on clean H-terminated diamond surfaces using
10 pulses of l-s (50% duty cycle) at a pulse current of 1.25 mA/cm2 produced the most
ideal particles in terms of size (ca. 30 nm), density (ca. 2 x 1010 cm'z), and catalytic
activity. Additionally the nucleation and growth mechanism of Pt on microcrystalline
and nanocrystalline diamond thin film studied using chronoamperometry indicated that
the nucleation moves from instantaneous to progressive with increasing deposition
overpotential. Finally, an introductory study of catalyst loss from the diamond surface

indicated that the loss of Pt surface area is likely due to Ostwald ripening.

Copyright by
JASON ALAN BENNETT
2006

 

A CKNO WLEDGMEN T S

There have been so many pe0ple who have influenced my life here at Michigan
State University, both professionally and personally. First and foremost I would like to
thank my principal advisor, Dr. Greg Swain. I am thankful for the guidance you have
given me over the last 5 years. I am additionally grateful that you allowed me to grow as
a scientist while still maintaining my livelihood. I will always remember "the proof is in

the pudding."

I would also like to thank my second reader, Dr. Merlin Bruening for his
additional guidance and many memorable interactions. Additionally, I thank my
remaining guidance committee members, Dr. Jim McCusker and Dr. Donnie Reinhard for
their roles in this degree. Thank you as well to the MSU chemistry department staff for
all their efforts as well as Ewa Danielewicz who trained me on and often provided

assistance with the scanning electron microscope.

Of course graduate school would have been much more difficult if it weren't for
the love and support of coworkers, friends, and family. I extend a special thanks to past
and present Swain group members. Working in the Swain lab has been an enjoyable and
rewarding experience. I am especially grateful to those of you who have directly helped
me in this pursuit: Jian, Matt, Maggie, and Shannon for their guidance and leadership
during my initial years in the group as well as Stotter, Fish, Elizabeth, and Audrey for

fruitful discussions and help during these last remaining years.

The friendships I have made here have been the most rewarding part of this
journey. To Stotter, Carl, Fish, Mel, Elizabeth, Audrey, Jesse, Bahvik, and Little J:
Thank you all for the memories and remember that you are all "cool fish in the nerd
pond." Additionally I owe a special thank you to Sarah for her continuous friendship and

enduring love. You have added an array of vibrant colors to every aspect of my life.

Finally, I would never have even made it to graduate school without the
continuing support and encouragement from my loving family. You guys have always
been there in good times and bad. I have cherished every moment and look forward to
what the future has in store for us all. Thank you again to everyone who helped me

complete this journey.

—- jason —

vi

 

TABLE OF CONTENTS

LIST OF TABLES .............................................................................................................. x

LIST OF FIGURES ........................................................................................................ xiv
CHAPTER 1

INTRODUCTION ............................................................................................................. 1

1.1 Background ..................................................................................................... l

1.2 Carbon Material Properties ............................................................................. 2

1.3 Electrocatalyst Support Materials .................................................................. 10

1.4 Introduction to Metal Phase Formation ......................................................... 15

1.5 The Oxygen Reduction Reaction ................................................................... 22

1.6 Dissertation Overview ................................................................................... 27

REFERENCES ..................................................................................................... 29
CHAPTER 2

EXPERIMENTAL METHODS ........................................................................................ 38

2.1 Diamond Thin-Film Deposition ..................................................................... 38

2.2 Electrochemical Methods ............................................................................... 43

2.2.1 Electrochemical Measurements ...................................................... 43

2.2.2 Pulsed Galvanostatic Deposition of Pt ............................................ 44

2.2.3 Reagents .......................................................................................... 47

2.3 Physical Characterization Methods ................................................................ 48

2.3.1 Atomic Force Microscopy .............................................................. 48

2.3.2 Scanning Electron Microscopy ....................................................... 49

2.3.3 Raman Spectroscopy ....................................................................... 49

2.3.4 X-ray Diffraction Spectroscopy ...................................................... 50

2.4 Chapter-Specific Experimental Methods ....................................................... 50

2.4.1 Chapter 3 ......................................................................................... 50

2.4.2 Chapter4 ......................................................................................... 51

2.4.3 Chapter 5 ......................................................................................... 52

2.4.4 Chapter 6 ......................................................................................... 54

REFERENCES ..................................................................................................... 57
CHAPTER 3

EFFECT OF spz-BONDED NONDIAMOND CARBON IMPURITY ON
BORON-DOPED POLYCRYSTALLINE DIAMOND THIN-FILM ELECTRODES... 59

3.1 Introduction .................................................................................................... 59
3.2 Results & Discussion ..................................................................................... 62
3.2.1 Effect of CH4/H2 Ratio on the Physical Properties ......................... 62
3.2.2 Effect of CH4/H2 Ratio on the Electrochemical Properties ............ 73
3.2.3 Effect of CH4/H2 Ratio on the Oxygen Reduction
Reaction (ORR) .............................................................................. 84
3.2.4 Probing the Oxygen Reduction Reaction Mechanism
on Carbon ....................................................................................... 93

vii

3.3 Conclusions .................................................................................................... 99
REFERENCES ................................................................................................... l 02

CHAPTER 4
INVESTIGATING MICROCRYSTALLINE AND NANOCRYSTALLINE
BORON-DOPED DIAMOND AS A CORROSION RESISTANT ADVANCED

ELECTROCATALYST SUPPORT MATERIAL ......................................................... 106

4.1 Introduction ................................................................................................... 106

4.2 Microcrystalline Diamond Results & Discussion ......................................... 110

4.2.1 Physical Properties ......................................................................... 110

4.2.2 Surface Chemistry .......................................................................... 117

4.2.3 Electrochemical Response ............................................................ 118

4.3 Nanocrystalline Diamond Results & Discussion ......................................... 124

4.3.1 Physical Properties ........................................................................ 124

4.3.2 Surface Chemistry ......................................................................... 130

4.3.3 Electrochemical Response ............................................................ 130

4.3.4 Electrical Conductivity Maps ....................................................... 136

4.4 Conclusions .................................................................................................. 138

REFERENCES ................................................................................................... 140
CHAPTER 5

PULSED GALVANOSTATIC DEPOSITION OF Pt PARTICLES ON
MICROCRYSTALLINE AND NANOCRYSTALLIN E DIAMOND THIN-FILM

ELECTRODES ............................................................................................................... 144
5.1 Introduction .................................................................................................. 144
5.2 Results .......................................................................................................... 148
5.2.1 Pt Deposition on Microcrystalline Diamond ................................ 148
5.2.2 Pt Deposition on Nanocrystalline Diamond ................................. 157
5.2.3 Comparison Studies with Glassy Carbon ..................................... 161
5.2.4 Investigating Pt Nucleation and Growth Mechanism ................... 163
5.2.5 Secondary Diamond Growth ......................................................... 169
5.3 Discussion .................................................................................................... 173
5.4 Conclusions .................................................................................................. 178
REFERENCES ................................................................................................... 180
CHAPTER 6
THE EFFECT OF SURFACE CHEMISTRY ON METAL PHASE FORMATION
AND PARTICLE STABILITY ...................................................................................... 186
6.1 Introduction .................................................................................................. 186
6.2 Results & Discussion ................................................................................... 189
6.2.1 Physical Characterization .............................................................. 189
6.2.2 Electrochemical Response ............................................................ 194
6.2.3 Probing the Oxygen Reduction Reaction ...................................... 198
6.2.4 Investigating the Oxygen Reduction Mechanism ......................... 205
6.2.5 Investigating Particle Stability ...................................................... 211

6.3 Conclusions .................................................................................................. 213

viii

REFERENCES ................................................................................................... 218

CHAPTER 7

KEY FINDINGS AND FUTURE RESEARCH DIRECTIONS .................................... 222
7.1 Key Findings ................................................................................................ 222
7.2 Future Research Needs ................................................................................. 224
REFERENCES .................................................................................................... 228

APPENDIX A ................................................................................................................. 229

ix

Table 1.1

Table 1.2

Table 2.1

Table 2.2

Table 3.1

Table 3.2

Table 3.3

Table 3.4

Table 3.5

Table 3.6

Table 3.7

Table 3.8

Table 3.9

Table 3.10

LIST OF TABLES

A Physical and Electrochemical Property Comparison of Glassy
Carbon and Boron-Doped Diamond ........................................................... 4

Summary of Potential Pt-Based and Non-Pt-Based PEMF C
Electrocatalysts Currently Being Investigated .......................................... 26

Conditions Used for Microcrystalline and Nanocrystalline
Boron-Doped Diamond Thin Film Deposition on Si ................................ 40

Summary of Physical Properties of Microcrystalline and
Nanocrystalline Boron-Doped Diamond Thin Films Deposited on Si ..... 40

Nominal Diamond Crystallite Diameter and Height for Films
Deposited From Different Methane-to-Hydrogen Source Gas Ratios ...... 62

Raman Spectroscopic Data for Films Deposited from Different
Methane-to-Hydrogen Source Gas Ratios ................................................ 66

Comparison Between Measured Powder X-ray Diffraction Data
and Standard Values from ASTM 6—0675, 23-0064, 46-0943 ................. 71

Full Width at Half Maximum (FWHM) and Peak Intensity Ratio
for the Diamond (111) and (220) Peaks .................................................... 72

CV Working Potential Window (1' = 21:50 uA) for Films Deposited
from Different Methane-to—Hydrogen Source Gas Ratios ........................ 74

Summary of the Cyclic Voltammetric Peak Potential Separations
(AEp) for Different Redox Analytes at Films Deposited from
Different Methane-to-Hydrogen Source Gas Ratios ................................. 80

Data for ORR in 0.1 M HClO4 for Films Deposited from Different
Methane-to-Hydrogen Source Gas Mixtures ............................................ 88

Data for OR in 0.1 M NaOH for Films Deposited from Different
Methane-to-Hydrogen Source Gas Mixtures ............................................ 92

OR Parameters Obtained from Linear Sweep Voltammetric i-E
curves in Acidic and Alkaline Media at 5% CH4/H2 Film ........................ 96

OR Parameters Obtained from Linear Sweep Voltammetric i-E
curves in Acidic and Alkaline Media at a 5% CH4/H2 Film After
Acid-Washing and Rehydrogenating ........................................................ 99

Table 4.1

Table 4.2

Table 4.3

Table 4.4

Table 4.5

Table 4.6

Table 4.7

Table 4.8

Table 4.9

Table 4.10

Table 4.11

Comparison of the Real and Geometric Areas of Boron-Doped
Microcrystalline Diamond Films Before and Afier Anodic
Polarization ............................................................................................. 114

Summary of Contact Angle Measurements for Ultrapure Water on
Boron-Doped Microcrystalline Diamond Thin Film as a Function
of the Surface Chemistry ........................................................................ 118

Potential Window and Background Current Measured from Cyclic
Voltammetric i-E Curves for a Boron-Doped Microcrystalline
Diamond Thin Film in 0.1 M HClO4 at a Scan Rate of 100 mV/s ......... 120

Cyclic Voltammetric Peak Potential Separation (AEp) for
1 mM Fe(CN)63'/4” in 1 M KCl at Boron-Doped Microcrystalline
Diamond as a Function of the Surface State ........................................... 121

Cyclic Voltammetric Peak Potential Separation (AEp) for
1 mM Ru(NH3)63+’2* in 1 M KCl at Boron-Doped Microcrystalline
Diamond as a Function of the Surface State ........................................... 123

Cyclic Voltammetric Peak Potential Separation (AEp) for
1 mM F e3 H 2+ in 0.1 M HClO4 at Boron-Doped Microcrystalline
Diamond as a Function of the Surface State ........................................... 124

Comparison of the Real and Geometric Areas of Boron-Doped
Nanocrystalline Diamond Films Before and Afier Anodic
Polarization ............................................................................................. 1 27

Summary of Contact Angle Measurements for Ultrapure Water on
Boron-Doped Nanocrystalline Diamond Thin Film as a Function
of the Surface Chemistry ........................................................................ 130

Potential Window and Background Current Measured from Cyclic
Voltammetric i-E Curves for a Boron-Doped Nanocrystalline
Diamond Thin Film in 0.1 M HCIO4 at 3 Scan Rate of 100 mV/s ......... 132

Cyclic Voltammetric Peak Potential Separation (AEP) for
1 mM Fe(CN)63‘/4' in 1 M KCl at Boron—Doped Nanocrystalline
Diamond as a Function of the Surface State ........................................... 134

Cyclic Voltammetric Peak Potential Separation (AEP) for

1 mM Ru(NH3)53+/2+ in 1 M KCl at Boron-Doped Nanocrystalline
Diamond as a Function of the Surface State ........................................... 134

xi

Table 4.12

Table 5.1

Table 5.2

Table 5.3

Table 5.4

Table 5.5

Table 6.1

Table 6.2

Table 6.3

Table 6.4

Table 6.5

Cyclic Voltammetric Peak Potential Separation (AEp) for
1 mM Few2+ in 0.1 M HClO4 at Boron-Doped Nanocrystalline
Diamond as a Function of the Surface State ........................................... 135

Particle Analysis and Hydrogen Adsorption Data for Pt-Coated
Microcrystalline Diamond Films as a Function of the Pulse Numbera... 152

Electrochemically Active Pt Surface Area, Roughness Factor (RF)
and Specific Pt Surface Area for Pt-Coated Microcrystalline
Diamond Films as a Function of the Pulse Numbera .............................. 153

Particle Analysis and Hydrogen Adsorption Data for Pt—Coated
Microcrystalline Diamond Films as a Function of the Pulse
Current Densitya ...................................................................................... 155

Electrochemically Active Pt Area, Specific Pt Surface Area and
Roughness Factor for Pt-Coated Microcrystalline Diamond
Films Deposited at Different Pulse Current Densitiesa ........................... 157

Particle Analysis and Electrochemically Active Pt Area for Pt-Coated
Microcrystalline Diamond Film Subjected to 20 min Secondary
Diamond Growth Procedure and Extensive Potential Cycling ............... 173

Particle Analysis Data for Pt-Coated Microcrystalline and
Nanocrystalline Diamond Films as a Function of the Surface
Chemistry ................................................................................................ 192

Electrochemically Active Pt Surface Area (ESA) and Roughness
Factor (RF) for Pt-Coated Microcrystalline and Nanocrystalline
Diamond Films as a Function of the Surface Chemistry ........................ 197

Reduction Peak Potential (133”) and Half-Peak Potential (Ered ),

p/2
Cathodic Peak Current (ip°) and Specific Activity (at E = 600 mV)
for the Oxygen Reduction Reaction at Pt-Coated Microcrystalline
Diamond Films as a Function of the Surface Chemistry ........................ 200

Reduction Peak Potential (E?) and Half-Peak Potential (Ered ),

p/2
Cathodic Peak Current (ipc) and Specific Activity (at E = 600 mV)
for the Oxygen Reduction Reaction at Pt-Coated Nanocrystalline
Diamond Films as a Function of the Surface Chemistry ........................ 202

Summary of Values Obtained from Tafel Analysis of the OR at

Pt-Coated Microcrystalline Diamond Films as a Function of the
Surface Chemistry ................................................................................... 205

xii

 

Table 6.6

Table 6.7

Table 6.8

Summary of Values Obtained from Tafel Analysis of the ORR at
Pt-Coated Nanocrystalline Diamond Films as a Function of the
Surface Chemistry ................................................................................... 205

Electrochemical Quantities Determined Using Cyclic Voltammetric
Data Obtained for the OR at Pt—Coated Microcrystalline
Diamond Films as a Function of the Surface Chemistry ........................ 207

Electrochemical Quantities Determined Using Cyclic Voltammetric

Data Obtained for the OR at Pt-Coated Nanocrystalline
Diamond Films as a Function of the Surface Chemistry ........................ 207

xiii

LIST OF FIGURES

Figure 1.1 The microstructure of (A) glassy carbon and (B) sp3—bonded diamond.
Figures adapted from McCreery and Fischer, respectively.8’10 ................... 3

Figure 1.2 Comparative Raman spectra for glassy carbon and single
crystal diamond. .......................................................................................... 5

Figure 1.3 Background voltammetric i—E curves for glassy carbon and
boron-doped microcrystalline diamond thin film in degassed
0.1 M HClO4. Scan rate = 100 mV/s. Geometric area = 0.2 cm2. .............. 6

Figure 1.4 Summary of relative [C], [H], and [0] concentrations required
to achieve diamond growth. Adapted from Bachmann et (11.3738 ............ 10

Figure 1.5 Schematic of polymer electrolyte membrane fuel cell (PEMFC).
The magnified region shows the active cathode — nanometer-sized
catalyst particles (e.g., Pt) highly dispersed on a support material. .......... 11

Figure 1.6 Illustration of metal adlayer formation on an ideal (i.e., atomically
smooth with homogeneously electronically active) substrate.
Adapted from Pletcher.70 .......................................................................... 17

Figure 1.7 Dimensionless i-t transient for instantaneous and progressive
metal nucleation mechanisms. .................................................................. 18

Figure 1.8 Schematic representation of overlapping depletion layers with
respect to time. Adapted from Schariflter and Hills.7 ............................. 20

Figure 1.9 Top view of overlapping depletion layers as a function of deposition
overpotential and time. Adapted from Penner et a1. 69’” .......................... 20

Figure 1.10 Linear sweep voltammetric i-E curves for the oxygen reduction
reaction (OR) in 0.1 M HClO4 at boron-doped microcrystalline
diamond thin film and polycrystalline Pt. Scan rate = 1 mV/s.
Current is normalized to geometric electrode area, 0.2 cm2. .................... 24

Figure 2.1 Schematic of MPCVD reactor used to grow polycrystalline diamond
films. Note Ar was only used for nanocrystalline diamond growth. ....... 39

Figure 2.2 Plot relating the concentration of BZHé in the gas phase to that

incorporated into the microcrystalline diamond thin-film (0)
and the film resistivity (X). ...................................................................... 42

xiv

Figure 2.3

Figure 2.4

Figure 2.5

Scheme 2.1

Figure 3.1

Figure 3.2

Figure 3.3

Figure 3.4

Figure 3.5

Figure 3.6

Figure 3.7

Figure 3.8

Plot relating the concentration of BzH6 in the gas phase to that
incorporated into the nanocrystalline diamond thin-film (0)

and the film resistivity (>K). ...................................................................... 42
Schematic of the single compartment electrochemical cell used

in the experiments. .................................................................................... 44
Experimental set-up used for pulsed galvanostatic deposition of Pt. ....... 46

Diazonium surface modification process. The potential was

cycled once from 600 to either -600 (4-NBD) or -1300 mV

(4-MBD) vs. Ag/Ag+ QRE at a rate of 200 mV/s. The 4-MBD

modified diamond surface was then exposed to concentrated

H2SO4 for 1 h at a temperature of ~80 oC. ................................................ 56

Atomic force micrographs of diamond films deposited from
different CH4/H2 source gas mixtures. ...................................................... 63

Scanning electron micrographs (secondary electron) of diamond
films deposited from different CH4/H2 source gas mixtures. .................... 65

Raman spectra of diamond films deposited from different CH4/H2
source gas mixtures. Inset is enlarged region around diamond peak ........ 67

Powder X-ray diffraction spectra of diamond films deposited
from different CH4/H2 source gas mixtures. ............................................. 70

Powder X-ray diffraction spectra for highly orientated pyrolytic
graphite (HOPG) and glassy carbon (GC). The dotted lines are
the same as those in Figure 3.4, and inserted as reference points ............. 73

Background cyclic voltammetric i-E curves in degassed
0.1 M HClO4 for diamond films deposited from different CH4/H2
source gas mixtures. Scan rate = 100 mV/s. ............................................. 74

Background cyclic voltammetric i-E curves in degassed 0.1 M HCIO4
for 5% CH4/H2 diamond film as-deposited and following an acid-
washing/rehydrogenation procedure. Scan rate = 100 mV/s. ................... 76

SEM images (secondary electron) of the 5% CH4/H2 ratio

diamond film (A) as-deposited and (B) following acid-washing

and rehydrogenation. Insets are high resolution images of the

center region of each figure. ..................................................................... 78

XV

 

Figure 3.9

Figure 3.10

Figure 3.11

Figure 3.12

Figure 4.1

Figure 4.2

Figure 4.3

Figure 4.4

Figure 4.5

Figure 4.6

Figure 4.7

Cyclic voltammetric i-E curves at diamond films deposited from

different CH4/H2 source gas mixtures for 1 mM concentrations of

(A) Fe(CN)63'/4', (B) Ru(NH3),3*’2+, (C) Fey/2" and (D) 4-ch

dissolved in either 1 M KCl (A and B) or 0.1 M HClO4 (C and D).

Scan rate = 100 mV/s. ............................................................................... 81

Linear sweep voltammetric i-E curves for the oxygen reduction
reaction (ORR) in 0.1 M HClO4 at diamond films deposited from
different CH4/H2 source gas mixtures. Scan rate = l mV/s. ..................... 87

Linear sweep voltammetric i-E curves for the oxygen reduction

reaction (OR) in 0.1 M NaOH at diamond films deposited from

different CH4/H2 source gas mixtures. Scan rate = 1 mV/s. ..................... 91
Plot of ip vs. v“2 for ORR in acidic and alkaline media at

5% CH4/H2 film. ....................................................................................... 95

Scanning electron micrograph of Toray carbon paper coated
with boron-doped nanocrystalline diamond. Inset is of similar
grth at higher magnification. .............................................................. 1 10

Scanning electron micrographs of boron-doped microcrystalline
diamond (A) before and (B) after anodic polarization ............................ l 11

Atomic force micrographs (2 x 2 pm) of a boron-doped
microcrystalline diamond thin film (A) before and
(B) after anodic polarization. .................................................................. 113

Raman spectra of heavily boron-doped microcrystalline diamond

initially and after anodic polarization. Dotted lines correspond to

boron doping and diamond signals, respectively. Spectra are off-set

for viewing clarity ................................................................................... 115

Background cyclic voltammetric i-E curves for microcrystalline
diamond films in degassed 0.1 M HClO4 as a function of the
surface condition. Scan Rate = 100 mV/s. .............................................. 119

Cyclic voltammetric i-E curves for boron-doped microcrystalline

diamond films before, after anodic polarization and after

rehydrogenation in the presence of (A) Fe(CN)63"4‘,

(B) Ru(NH3)63+/2+ and (C) Few”. Scan Rate = 100 mV/s. .................... 122

Scanning electron micrograph of boron-doped nanocrystalline
diamond film after anodic polarization. .................................................. 125

xvi

 

Flgm

Figure 4.8 Atomic force micrographs (500 x 500 nm) of boron-doped
nanocrystalline diamond thin film (A) before and
(B) after anodic polarization. .................................................................. 126

Figure 4.9 Raman spectra for a boron-doped nanocrystalline diamond
film before and after anodic polarization. Spectra are off-set
for viewing clarity. .................................................................................. 129

Figure 4.10 Background cyclic voltammetric i-E curves in degassed
0.1 M HClO4 for boron-doped nanocrystalline diamond films
as a function of the surface condition. Scan Rate = 100 mV/s. .............. 131

Figure 4.11 Cyclic voltammetric i—E curves for boron-doped nanocrystalline
diamond films before, after anodic polarization and after
rehydrogenation in the presence of (A) F e(CN)(53 '/ 4',
(B) Ru(NH3)63+/2+ and (C) Fey/2+. Scan Rate = 100 mV/s. .................... 133

Figure 4.12 AFM height mode images (air) (A and C) and simultaneously
recorded conductivity maps (bias voltage, Em, — Esubsmc = +2 V)
of the same regions (B and D) of a boron-doped nanocrystalline
diamond thin film over a 1 x 1 11m area not polarized (A and B)
and anodically polarized (C and D). ....................................................... 137

Figure 5.] Scanning electron micrographs of microcrystalline diamond
electrodes covered with Pt particles that were deposited using
(A) 10, (B) 25 and (C) 50 pulses, respectively. Pulse width = l 3.
Current density = 0.50 mA/cmz. ............................................................. 150

Figure 5.2 Cyclic voltammetric i-E curves in degassed 0.1 M HClO4 for
microcrystalline electrodes coated with Pt using different pulse
numbers. Pulse width = l s. Current density = 0.50 mA/cmz.
Scan rate = 50 mV/s. The shaded region of the curve represents
the charge associated with hydrogen adsorption. The current is
normalized to the diamond geometric area, 0.2 cm2 ............................... 151

Figure 5.3 Scanning electron micrographs of Pt-coated microcrystalline
diamond electrodes deposited at pulse current densities of
(A) 0.50, (B) 0.75, (C) 1.00, (D) 1.25 and (E) 1.50 mA/cmz.
Pulse number = 10. Pulse width = 1 s. .................................................... 154

Figure 5.4 Cyclic voltammetric i—E curves in degassed 0.1 M HClO4 for
microcrystalline electrodes coated with Pt at the different current
densities. Pulse number = 10. Pulse width = 1 5. Scan rate = 50 mV/s.
The current is normalized to the diamond geomenic area, 0.2 cm2. ...... 156

xvii

.
(Er-1

h;

h.

Figure 5.5

Figure 5.6

Figure 5.7

Figure 5.8

Figure 5.9

Figure 5.10

Figure 5.11

Figure 5.12

Figure 5.13

Figure 6.]

Scanning electron micrographs of Pt-coated nanocrystalline diamond
electrodes deposited at pulse current densities of (A) 0.50 and
(B) 1.25 mA/cmz. Pulse number = 10. Pulse width = l s. ...................... 159

Cyclic voltammetric i-E curves in degassed 0.1 M HClO4 for
nanocrystalline diamond electrodes coated with Pt deposited

at 0.50 and 1.25 mA/cmz. Pulse number = 10. Pulse width = 1 5.

Scan rate = 50 mV/s. The current is normalized to the diamond

geometric area, 0.2 cm2 ........................................................................... 161

Scanning electron micrograph of Pt-coated glassy carbon deposited
using 10 pulses of l 5 width at a current density of 1.25 mA/cm2 .......... 162

Cyclic voltammetric i—E curves for 1 mM KthC16 + 0.1 M HClO4
on different carbon materials. Scan rate = 20 mV/s ................................ 164

Background subtracted chronoamperometric i-t curves for Pt
deposition on microcrystalline diamond thin film in 1 mM KthC16
+ 0.1 M HClO4 at the overpotentials indicated in mV ............................ 166

Background subtracted chronoamperometric i—t curves for Pt
deposition on nanocrystalline diamond thin film in 1 mM KthCl6
+ 0.1 M HClO4 at the overpotentials indicated in mV ............................ 166

Comparison of experimental normalized i—t curves for

microcrystalline (0) and nanocrystalline (1:) diamond with

the theoretical curves corresponding to instantaneous and

progressive nucleation at overpotentials of (A) -512, (B) -612,

(C) -712, (D) -812, (E) -912 and (F) -1012 mV ...................................... 168

Scanning electron micrographs of Pt-coated microcrystalline

diamond electrode (A) before, (B) after a 20 min secondary
microcrystalline diamond growth and (C) the film with the

secondary diamond growth after extensive potential cycling

(1000 cycles between ca. i5 mA/cm2 (geometric area)). Pulse

number = 10. Current density = 1.25 mA/cmz. Pulse width = 1 s. ......... 172

AF M height mode images (air) of a boron-doped nanocrystalline

diamond thin film over a (A) 10 x 10 and (C) 1 x 1 11m area.
Simultaneously recorded conductivity maps of the same regions

(B and D) with a bias voltage of +2 V (Etip - Esubsme). .......................... 178

Scanning electron micrographs of as-deposited Pt/BMD

electrodes deposited using ten l-s pulses at 1.25 mA/cmz.

The BMD surfaces were (A) H-terminated, (B) O-terminated,

(C) 4-NBD modified and (D) 4-MBD modified. .................................... 191

xviii

Figure 6.2 Scanning electron micrographs of as-deposited Pt/BND
electrodes deposited using ten 1-s pulses at 1.25 mA/cmz.
The BND surfaces were (A) H-terminated, (B) O—terminated,
(C) 4-NBD modified and (D) 4-MBD modified. .................................... 194

Figure 6.3 Cyclic voltammetric i—E curves in degassed 0.1 M HClO4 for
Pt-coated microcrystalline diamond electrodes as a firnction
of the surface chemistry. Scan rate = 50 mV/s. ...................................... 196

Figure 6.4 Linear sweep voltammetric i-E curves for the OR in 0.1 M HCIO4
at Pt-coated microcrystalline diamond electrodes as a function
of the surface chemistry. Scan rate = l mV/s. The corresponding
Tafel plots are presented in the inset ....................................................... 199

Figure 6.5 Linear sweep voltammetric i-E curves for the OR in 0.1 M HCIO4
at Pt-coated nanocrystalline diamond electrodes as a fimction
of the surface chemistry. Scan rate = l mV/s. The corresponding
Tafel plots are shown in the inset. .......................................................... 201

Figure 6.6 ORR linear sweep voltammetric i-E curves for Pt/BMDH electrode
in 0.1 M HClO4 before (-— — —) and after (———) 2 h cathodic
polarization of -0.1 A/cm2 (geometric area). Scan rate = 1 mV/s ........... 213

Figure A.1 Front panel and schematic code for the potential pulse generator
program created in LabVIEW Version 6.1 in conjunction with a
National Instruments PC1-5401 card. The parameters in the top
window were those used to deposit Pt using 10 pulses of 1 s (delay
1 s for 50% duty cycle) and a pulse current density of 1.25 mA/cm2 ...... 229

xix

 

 

 

1.]

Eli:

(‘2
p;
-._.

in)

m

CHAPTER I

Introduction

1.1 Background

Catalysis is of vital importance in today's chemical industry. In fact, few
chemical processes are conducted without the use of a catalyst.1 A subsection of
catalysis is electrocatalysis, defined as the study of heterogeneous redox reactions that
involve reactant and product species transferring electrons through an electrolyte/catalyst
interface.2 Of particular interest to our group are supported electrocatalysts due to their
high surface-to-volume ratio, high dispersion over the support surface and stability.
Electrocatalysis is prominent in several applications such as electrochemical chemical

synthesis, chemical sensing and energy storage and conversion.

Growing concern over the environmental impacts from burning fossil fuels,
limited fuel supplies and reliance on foreign oil has lead to considerable interest in fuel
cells as a clean renewable energy alternative for stationary and portable power
applications. The fuel cell is not a new concept as it was invented by Sir William Grove
in 1839. For example, fuel cells have been used by NASA in aerospace applications for
several decades.3 A fuel cell converts chemical energy into electrical energy through
electrocatalytic reactions at two electrodes known as the anode and cathode. Each
electrode consists of a porous gas diffusion material and some form of metallic

electrocatalyst particles supported on carbon.

The US. Department of Energy (DOE) seeks to develop hydrogen, fuel cell and
infrastructure technologies, in order to give the public a practical and cost-effective fuel
cell vehicle by 2020.4 However, there are several technological hurdles that must be
overcome before fuel cells can become a viable, clean energy alternative for
transportation: (i) supplying, dispersing, and storing alternate fuels (e.g., 112),5 (ii)
reducing cost and improving efficiency through reduced electrocatalyst loading or newly
developed non-noble metal electrocatalysts,6 (iii) development of reliable physical and
chemical sensors7 and (iv) preventing activity loss caused by carbon support degradation
and bipolar plate corrosion. This work is a continuation of our group's efforts to develop
an advanced state-of—the-art electrocatalyst support material that can withstand high
temperatures and current density for extended periods of time without structural
degradation or decreased catalytic activity. Specifically, the investigation of

electrocatalyst metal (e. g., Pt) phase formation on diamond will be presented.

1.2 Carbon Material Properties

Various forms of carbon have been used as electrochemical electrodes over the
years. Carbon electrodes are commonly composed of spz-hybridized atoms as found in
highly ordered pyrolytic graphite (HOPG), glassy carbon (GC), as well as carbon
nanotubes (CNTs) and nanofibers (CNFs). However, electrically conducting diamond, a
sp3-hybridized form of carbon, has been a focus of study in recent years as an electrode.
It is also possible for the carbon to contain a mixed sp3 and sp2 microstructure, and such

materials are referred to as diamond-like carbon (DLC).

 

 

Carbon materials differ from metals for electrochemical purposes in several ways,
as discussed by McCreery:8 (i) electronic properties ranging from metallic to
semimetallic depending on the carbon microstructure, (ii) the case at which the carbon
surface can be modified through adsorption or chemical modification, (iii) the very
complex surface chemistry of carbon which includes various types of functional groups
and surface oxides, and (iv) the wider potential window of carbon materials compared to
metal electrodes, thus making possible the study of electrochemical processes that were
previously unattainable.8 There are two types of carbon electrodes relevant to this
research: glassy carbon (a microstructurally disordered spZ-bonded material) and boron-
doped diamond (a polycrystalline sp3-bonded material). The structure of these materials
is shown in Figure 1.1 and the properties are compared in Table 1.1.8‘9 The terms La and
Lc in Figure 1A describe the ordered domain width and graphene sheet stacking

thickness, respectively.

 

Figure 1.1. The microstructure of (A) glassy carbon and (B) sp3-bonded diamond.
Figures adapted from McCreery and Fischer, respectively.8’10

 

Table 1.1. A Physical and Electrochemical Property Comparison of Glassy Carbon and
Boron-Doped Diamond

 

 

. L. Resistivity Density Capacitance
Ma‘ma' (A) (n-cm) (g/cm’) (uF/cm’)
Glassy Carbon ~50 4 x 10'3 1.5 ~33
Boron-Doped Diamond N.A. ~0.1 3.51 3-7

 

 

 

 

Glassy carbon (GC) is the most popular electrode material for electroanalytical
measurements because it is active and yields a reproducible response.11 It has
electrochemical characteristics similar to those of the edge plane of HOPG, and is
mechanically strong and nonporous. GC has metallic-like electronic properties. For
example, freshly polished GC exhibits a capacitance of ca. 30 11F /cm2 and a resistivity of
ca. 4 x 10'3 Q—cm.8 Raman spectroscopy is often used to characterize the microstructures
of carbon materials, including those of glassy carbon and diamond. Figure 1.2 shows
representative spectra for GC (top) and single crystal diamond (bottom). The GC
spectrum yields two recognizable peaks, one at ca. 1350 cm'1 corresponding to the
"disorder" band (commonly referred to as the D or Alg band), and the other at ca. 1590
cm'1 corresponding to the graphite band (commonly referred to as the G or E2g band).12
A ratio of the respective peak intensities (D/G ratio) gives the fractional amount of

exposed edge plane.13

This is usually denoted as Gedge, and provides a measure of the
microstructural disorder in the material. The spectrum for single crystal diamond is quite
different, exhibiting a single peak at 1332 cm'1 which is associated with the first order

diamond phonon. The position and width of this peak is sensitive to factors such as the

crystallite size, optical density, boron doping level (synthetic diamond), net stress, defect

densi

for a

This

density, and film quality (synthetic diamond)”16 Raman spectroscopy is a useful tool

for assessing the relative levels of sp2- and sp3-bonded carbon in a CVD diamond film.

This will be discussed fiirther in Chapters 2 and 3.

 

Diamond

  
     

Intensity (A.U.)

Glassy Carboni
/

 

L

I 1 l n l

1200 1300 1400 1500 1600 1700 1800

 

 

 

Raman Shift (cm-l)

Figure 1.2. Comparative Raman spectra for glassy carbon and single crystal diamond.

Glassy carbon also exhibits characteristic electrochemical properties especially
with regard to the background cyclic voltammetric i-E curves. Figure 1.3 compares the
background cyclic voltammetric i-E curves for GC and a boron-doped microcrystalline
diamond thin film (discussed below). The i-E curve for GC is characterized by a wide
potential window, at least as compared to metal electrodes, a large background current

and a highly visible reversible redox couple associated with electroactive surface quinone

. . . 7-
funct1ona11t1es.l 2°

 

Boron-doped diamond has emerged as an attractive alternate electrode to
commonly used spz-bonded carbon, such as GC, for several electrochemical applications
including electroanalysis and electrolytic water purification.9’2"27 High-quality diamond
(i.e., minimal sp2 nondiamond carbon impurity) possesses several electrochemical

23‘28'30 (i) a low and stable

properties that make it an attractive material including:
background current and capacitance in aqueous media (ca. 5 11F/cm2)3 1, (ii) a wide
working potential window in many aqueous electrolytes (~3-4 V), (iii) rapid electron
transfer kinetics for several redox systems without any conventional pretreatment, (iv)
long-term response stability and (v) morphological and microstructural stability during
anodic and cathodic polarization (~10 A/cmz) in both acidic and alkaline media. A

representative background voltammetric i-E curve for a boron-doped microcrystalline

diamond thin film is shown in Figure 1.3.

 

 

 

 

 

 

 

- I 10 uA
A ' Glassy Carbon
< i
:3. b
v
H
r:
o
a... -
3 Diamond
U -
r
l
-l000 -500 0 500 1000 1500 2000

Potential (mV vs. Ag/AgCl)

Figure 1.3. Background voltammetric i—E curves for glassy carbon and boron-doped
microcrystalline diamond thin film in degassed 0.1 M HClO4. Scan rate = 100 mV/s.
Geometric area = 0.2 cm2.

 

reg

mic

11101
proj

€31:

The diamond microstructure consists of tetrahedrally-bonded sp3-hybridized
carbon atoms with each bonded to four others, as shown in Figure 1.1. The surface is
usually hydrogen-terminated as shown in the figure; however, the surface chemistry can
be easily altered (e.g., oxygen-terminated or aryl molecule modified), as is discussed in
Chapter 6 of this dissertation. Diamond can be synthetically grown using several
methods including a process employed in our group, microwave-assisted plasma
chemical vapor deposition (MPCVD). A typical CVD reactor used for diamond growth
(see Figure 2.2) consists of source gases (CH4 and H2 for microcrystalline growth and
CH4, H2 and Ar for nanocrystalline growth), a dopant gas (e.g., B2H5 to make the
diamond, a natural insulator, electrically conductive), mass flow controllers (M.F.C.) to
regulate the gas flows into the reaction chamber, a plasma activation source (e.g.,
microwave generator) and mechanical pumps and throttle exhaust valve for controlling
the system pressure.9 Two of the most critical parameters for diamond growth are the
substrate temperature and system pressure. Such a reactor can be used to grow two types
of electrically conducting diamond: microcrystalline (>1 11m crystallite size) and
nanocrystalline (~10 to 20 nm crystallite size) thin fihns. Despite the different
morphologies, both types of diamond are highly conductive and possess semi-metallic
properties. The nominal boron dopant concentration is ca. 1020 B/cm3 and the film
resistivity is less than 0.05 O cm. The most common architectures are single crystal and

polycrystalline thin film.

Microcrystalline diamond is deposited using CH4 as the primary carbon source.

The selected grth conditions convert CH4 to methyl radicals, CH3', which serve as the

pn

the
on

p3.

C31
Ckl

lit

25:}
mic
111))

surf

‘0 ac

 

primary growth species.32 The added H2 in the form of atomic hydrogen, H', has three
major functions.9 First, it abstracts hydrogen from methane to produce methyl radicals in
the gas phase (i.e., primary reactant species) and from chemisorbed methyl substituents
on the surface to produce radical sites for further carbon attachment. Second, it
passivates dangling bonds on the surface, which inhibits reconstruction to a spz-bonded
phase on the grth surface. Third, it rapidly gasifies any of the spz-bonded nondiamond

carbon impurity that forms.9

This is why the cool down in atomic hydrogen is an
extremely important part of the diamond growth process in order to achieve high quality

(i. e., no spz-bonded nondiamond carbon impurity) hydrogen-terminated films.

Nanocrystalline diamond is also grown using CH4 as the primary carbon source,
but in an H2-deficient, Ar-rich atmosphere. The CH4/(H2 + Ar) source gas composition
affects the film morphology, surface roughness and grain size.”34 Unlike
microcrystalline diamond, the primary grth mechanism for nanocrystalline diamond
involves the insertion of the C2 carbon dimer into C—C and C—H bonds on the grth
surface.34 The C2 dimer also appears to be the primary nucleation species when the Ar

composition in the gas phase is >90%.3S

The lack of H2 in the source gas as well as a
higher grth pressure results in a higher rate of renucleation. This leads to a continuous
and relatively smooth film at a low nominal film thickness. The nucleation and grth of

nanocrystalline diamond thin film has been reviewed extensively by Gruen.3’5'36

Usually the grth and cool down conditions for both diamond types are selected

to achieve high quality sp3 carbon and to minimize any sp2 nondiamond carbon impurity

formation. A key factor in achieving this involves selecting the proper source gas ratio.
Bachmann et a1. created a diagram showing the relationship between the C/H/O ratios
and diamond growth.“38 An adaptation of the Bachmann triangle is shown in Figure 1.4.
The lower shaded region represents C/H/O ratios where no diamond growth occurs, while
the top shaded region represents C/H/O ratios where spz-bonded nondiamond carbon
growth occurs. The small white region in the middle represents the narrow range of
source gas composition that will result in diamond growth. The high-quality diamond
films produced in our group are grown using only carbon and hydrogen source gases,
without added oxygen. This is expressed by the bottom left comer of the triangle where
the concentration of H is high compared to the concentration of C (and the concentration
of O is zero). As the amount of C increases with respect to H, the quality of the film
decreases as less diamond is grown in favor of nondiamond sp2 carbon growth. This is
the basis for the study presented in Chapter 3 where the physical and electrochemical
properties of diamond films are assessed as a function of increasing C/H ratio in the
diamond growth and hence an increased incorporation of nondiamond carbon within the

film microstructure.

 
   

[H] [C]
[H] + [C] [C—] + [0]

  

0.9

[0] + 111]

Figure 1.4. Summary of relative [C], [H], and [0] concentrations required to achieve
diamond growth. Adapted from Bachmann et al.3 ’38

1.3 Electrocatalyst Support Materials

Electrocatalysts are used in both the cathode and anode of PEM fuel cells. Fuel
cells are gaining popularity for transportation because they represent a form of clean
energy (H20 is the only waste product) and could potentially lessen our society's reliance
on fossil fuels. Currently, Pt and Pt-based alloys are the most common catalysts used in
phosphoric acid and polymer electrolyte membrane fuel cells (PAFC and PEMFC,
respectively). However, these metals are costly and in somewhat limited supply. For
example, in 1987, the lowest loading to give a continuous layer was estimated to cost
~$1000/kW for the catalyst alone!39 The current cost for a PEM fuel cell stack is

~$110/kW, and DOE targets are to reduce this to $30/kW by the year 2015.4

10

There is strong interest in reducing the metal loading and designing more active
and less expensive catalysts to reduce the overall fuel cell cost. Supported
electrocatalysts have the surface area and tum-over number needed for fuel cells but a
suitable material is required on which to highly disperse the nanometer-sized catalyst
particles. Figure 1.5 shows a schematic of a PEMFC with a magnified view of the
cathode showing the electrocatalyst support material and the supported electrocatalyst

particles. Graphitic carbon powder is a commonly used support material that has several

desired properties including high surface area, low cost and electrical conductivity.

 

 

 

 

Figure 1.5. Schematic of polymer electrolyte membrane fuel cell (PEMF C). The
magnified region shows the active cathode — nanometer-sized catalyst particles (e.g., Pt)
highly dispersed on a support material.

Graphitic carbon also has some limitations as a support, most notably its
microporosity, which limits access of reactants to some of the electrocatalyst and
susceptibility to microstructural degradation due to oxidation and/or corrosion processes.

It is important to note that the overall efficiency of a low temperature fuel cell (e.g.,

ll

PAF C or PEMF C) is limited by reactions at the cathode where oxygen reduction occurs.
The thermodynamic cell voltage, Ecen, for this reaction, and hence the optimal fuel cell
operating voltage, is 1.23 v.39~4° However, Ecell values of ~700 mV are more commonly
observed (using Pt loadings of ca. 0.7 mgm/cmz, H2/air PEMFC at 80 °C)6 due to sluggish
kinetics for the oxygen reduction reaction at the cathode as well as carbon support
degradation. The carbon surface can also undergo oxidation and corrosion, the latter by a

reaction written as follows:39

C + 21120 —> C02 +4H+ +4e' (1.3.1)

Since this reaction is electrochemical (E° = 0.118 V vs. SHE), its rate increases with
increasing cathode potential. Additionally, the corrosion rate also increases at a given
potential with increasing temperature, vapor pressure, oxygen content and catalyst
loading.41 Furthermore, Pt, currently the most active catalyst for the oxygen reduction
reaction in acidic media, significantly accelerates the carbon corrosion rate at lower
temperatures (~50 oC) and potentials (~0.55 V vs. SHE) compared to the rate for the base

carbon.41

Carbon support corrosion causes increased ohmic resistance, catalyst particle
agglomeration and particle detachment. All these result in a loss in catalytic activity. If
the degradation is severe enough, complete electrode failure will result. Though carbon
corrosion under PEMF C cathode potentials used in automotive applications (e.g., 0.65 —
0.8 V) was previously thought not to be a serious problem, it was recently determined

that cathode potential ranges that far exceed the stability region of standard carbon

supports even at 80 0C were reached.42 For example, PEMFCs in automotive

12

applications undergo an estimated 30,000 startup/shutdown cycles over the life of a
vehicle, which can lead to short term potential excursions of the cathode to potentials as
high as 1.5 V vs. SHE due to H2/air fronts in the anode compartment. At potentials this

high, the corrosion rate of current sp2 carbon supports is high.42

It is widely recognized that advanced support materials are needed before fuel
cells can become commercially viable, particularly for transportation. These materials
must have the same properties as the currently used spz-bonded carbons but not the short-
comings of microporosity and corrosion susceptibility. Several types of materials are
currently being investigated as alternates. One receiving considerable attention is carbon
nanotubes (CNTs).43‘4S While these materials do have the advantage of being electrically
conductive and mechanically stable, they are currently limited by their inherent
impurities (e.g., Ni, Co, Fe, etc.) and their expense on the industrial scalef”6 Another
alternative is carbon nanofibers (CNFs).““l‘l‘47 While CNFs are electrically conductive
and have a lower impurity content compared to CNTs, they are still quite expensive to
produce and their long-term stability has yet to be determined. Other alternate support

“8'50 and carbon aerogels.“

materials receiving interest are mesoporous carbons
Mesoporous carbons have the advantage of possessing a very high surface area and a
large pore diameter. However, it is currently difficult to control the pore diameter and a
post-synthesis treatment is required. Carbon aerogels show promise for better control of
the pore size but the stability may be an issue. All these materials have some promise as

catalyst supports, however they are still composed of spz-bonded carbon and are likely to

be susceptible to microstructural degradation. To our knowledge, there has been no long-

13

term investigation into the morphological stability of these materials under fuel cell

conditions.

Over the last ten years, electrically conducting diamond has emerged as an
alternative to commonly used spz-bonded carbon materials for a variety of applications
largely because of its large potential operating range, low background current (i.e.,
increased signal to background ratio) and resistance to both molecular adsorption and
fouling.2"24'26’52’53 Boron-doped diamond also possesses several properties desired for an
electrocatalyst support material including (i) high electrical conductivity, (ii) dimensional
stability at high temperature, potential and current density, (iii) resistance to corrosion
and (iv) with advances in diamond synthesis, a low cost.24‘54’55 The material is limited at
present because of its low inherent surface area (maximum in powders is 1-10 mz/g);
although detonation diamond is a form of powderous diamond (electrically insulating)
with a high surface area (1000 mz/g).56 We seek to develop a high surface area diamond
support material that is corrosion resistant, electrically conductive and able to support a
high dispersion of electrocatalyst particles. Our most promising results, so far, have
come from coating sp2 carbon paper or diamond powders with a thin film of electrically
conducting microcrystalline or nanocrystalline diamondm‘57 In addition to developing an
electrically conducting, high surface area form of diamond, we also seek to better
understand, control, and stabilize the electrocatalyst particles on the diamond support
surface. Therefore, the fundamentals of metal phase formation are of paramount

importance.

14

1.4 Introduction to Metal Phase Formation

Electrochemically active metal particles are deposited onto carbon support
materials by one of several methods including (i) the collection of metal colloidal

particles from rnicroemulsions,58 (ii) physical vapor deposition,59 “‘1 ”fs' them" (iii)

“‘3 and (v) electrodeposition.64'66 Electrodeposition is

impregnation,60 (iv) DC sputtering
the preferred method because, unlike the other four, it allows for good control of particle
properties (i.e., size, density and size dispersion based on chosen deposition parameters)

and places the particles in direct contact with areas of the support material that are

electrically conducting and electrochemically active.

Ideal supported catalysts should (i) have a high activity towards the reaction (e. g.,
H2 oxidation or 02 reduction in the case of a fuel cell), (ii) be evenly distributed
nanometer-sized (~5-10 nm) particles, (iii) have a narrow size distribution, (iv) have a
high distribution density, (v) have electronic communication with the supporting material
and (vi) have long-terrn stability with the support material (i.e., not detach from the
support). The concept of using supported catalysts instead of metal electrodes comes
from the fact that only the surface of the metal is involved in electrochemical reactions.
Therefore, reducing the catalyst size increases the electrochemically active surface area,
while decreasing catalyst loading. This is extremely important in commercial
applications where large agglomerates of Pt do not provide a significant increase in

catalyst activity but significantly increase the cost of production.

15

Electrochemically-driven metal phase formation and adlayer growth on foreign
substrates has been extensively studied.“70 Pletcher gives examples of such reactions
including (i) the cathodic deposition of metals, (ii) the anodic formation of oxide layers
either by metal corrosion or the oxidation of metallic ions in solution, (iii) the formation
of insoluble salts on an anode (e. g., AgCl) and (iv) the conversion of one solid phase on a
surface to another with a different lattice structure.70 The focus of the work described in
this dissertation is on the cathodic deposition of metals onto a carbon support material.
Figure 1.6 shows an adaptation from Pletcher, which illustrates the process of metal
phase formation on an ideal support that is atomically smooth, defect free, and

0 First, the solvated ion

homogeneously electrochemically active over the surface.7
diffuses from the bulk solution to the electrode surface where it is reduced to an ad-atom.
These ad-atoms are thermodynamically unstable and should redissolve. This can be seen

by equation 1.4.1 where AGSurface is positive and proportional to the surface area and

AGBulk is negative and its magnitude is proportional to the phase volume.70

AGPhase = AGSurface + AGBulk “'4'”

Small centers are thus unstable because they have too high a ratio of surface area/volume
(AG1)},ase is positive). Therefore, in order for the nucleus to become stable, it must reach a
critical size and this can be achieved by the ad-atoms rapidly diffusing along the surface
until they cluster into a nucleus of sufficient size to be stable. In electrodeposition, a high
concentration of ad-atoms is favored by using a highly concentrated metal salt solution
and by driving the electron transfer process by applying a sufficient overpotential. As

time progresses, the metallic nucleus grows due to the incorporation of more ad-atoms.

16

Therefore, the length of time in which the overpotential is applied must not be too long in
order to keep as high a surface area/volume ratio as possible. At extended times, the
growing particles begin to grow together or agglomerate, and eventually result in the

formation of a continuous layer over the whole cathode surface.

a

linear diffusion

 

     
  

surface

0 —'
0diffusion _ f _ .

electrode I e“ ad-atom growth center

solution

 
 

Figure 1.6. Illustration of metal adlayer formation on an ideal (i.e., atomically smooth
with homogeneously electronically active) substrate. Adapted fiom Pletcher.70

An important aspect of studying metal phase formation is investigating how an
individual metal particle nucleates on the substrate surface. There are two limiting cases
commonly used to describe metal phase formation: (i) instantaneous nucleation and
grth where all of the metal nuclei form at the same time and subsequent growth occurs
at a particular rate and (ii) progressive nucleation and growth where new nuclei form
over time and subsequent grth occurs from each. Scharifker and Hills presented a way
to determine the metal nucleation mechanism by observing the shapes of

chronoamperometric i-t profiles.“ A dimensionless i-t curve can be obtained for each of

17

these limiting mechanisms by plotting the square of the ratio of the current to the current
max (Iz/Imz) vs. the ratio of time and the time at the maximum current (mm).71 This is

shown in Figure 1.7.

 

1.0 ‘

0-8 " Instantaneous

0.6 _

rzlrm2

0.4 r
Progressive

0.2

 

 

 

0.0 1 l n l 4 I 1
0.0 1.0 2.0 3.0 4.0

t/tm

Figure 1.7. Dimensionless i-t transient for instantaneous and progressive metal
nucleation mechanisms.

These dimensionless i-t curves were theoretically derived for the instantaneous
and progressive mechanisms using equations 1.4.2 and 1.4.3, respectively, where I is the
resulting background subtracted current at time t, Im is the current maximum and tm is the
time at which the current reaches its maximum.“ These equations are useful in studying
metal nucleation processes because they do not contain experimental variables, such as

the diffusion coefficient, nucleation density, or nucleation rate.

 

1212 m_19/5:2{1 exp[_ L2564(/m)]}2 (1.4.2)

18

 

12 12254 2 2

= ' l—exp —2.3367(/ ) (1.4.3)
I 2 t tm
m /m

Assuming a potentiostatic deposition was carried out at a sufficiently cathodic
potential, the electrodeposition of metal particles onto an electrode surface would be
limited by the diffusion of the solvated metal ions from the bulk to the electrode surface.
When the ions in the immediate vicinity of the electrode surface are reduced, 3 volume
void of metal ions known as the depletion layer is formed around the metal nucleus. This
results in radial diffusion of ions from the bulk to the metal surface. The depletion layer
around each individual metal particle grows in thickness with time and eventually
neighboring layers begin to overlap and transition the process from 3-dimensional radial
diffusion to 2-dimensional diffusion to a planar electrode. This leads to particle
agglomeration, which results in the formation of large particles of widely—dispersed size
and low density. This model was first demonstrated by Schariflter and Hills and has since
become the model of electrodeposition and is illustrated in Figure 1.8.71 Figure 1.9
shows the same phenomena, but from a top view. Additionally, this illustrates how the
rate of deposition increases with increasing deposition overpotential, and hence quicker

overlapping between neighboring depletion layers.

19

 

Figure 1.8. Schematic representation of overlapping depletion layers with respect to
time. Adapted from Scharifker and Hills.71

metal depletion
nanopamcle la r

  

Overpotential

gati‘oi

Figure 1.9. Top view of overlapping depletion layers as a function of deposition
overpotential and time. Adapted from Penner et (11.69‘7

Over the past several years, our group has been investigating the formation of

64,65,73,74 F 01’

metal adlayers on boron-doped microcrystalline diamond thin films.
example, Wang et a1. investigated the electrodeposition of Pt using constant current
deposition (0.50 mA/cmz) as a function of time (100-400 s). This resulted in particles 50-
300 nm in size and dispersed over the entire surface at an approximate density of 108 —
109 cm'z.65 While it was shown that Pt could be stably deposited on the diamond surface
using this method, the particles produced were too large to be of practical electrocatalytic
use. In addition, several groups including ours have investigated non-

electrochemical63’75’76

and electrochemical metal phase formation on diamond for such
purposes as metal ion analysis (i.e., Ag”, Pb2+, Cup”, Cd2+, Zn”, As3+/5+ etc.),21’27’77'81
wastewater remediation,82 and various electro-catalytic reactions (e.g., H2 or methanol

oxidation and 02 reduction)”85

A possible alternative to constant potential/current deposition is the application of
a pulsed deposition sequence in which the potential/current is applied for short periods of
time with dead periods between each pulse. During this dead period, ions diffuse from
the bulk to replenish the depletion layer surrounding a metal particle before another pulse
is applied. This helps to minimize the overlap of depletion layers from individual
particles, thus limiting particle agglomeration and resulting in particles of smaller
nominal size and more size-monodispersity. Also, the particle density is increased.
Pulsed electrodeposition of metals has been employed for several years to prepare
nanometer-sized particles of metals and alloys.59‘66‘69'72‘86'93 However, metal deposition

on diamond is an extremely complicated process that is not yet completely understood.

21

Unlike most electrode materials used in metal deposition investigations, diamond is quite
heterogeneous electronically and structurally.94 This creates significant challenges in
determining the proper deposition parameters (i.e., number of pulses, pulse
potential/current magnitude, pulse width, delay time between pulses). It was our
supposition at the outset of this research that changing the deposition froma constant
current to a pulsed current approach would result in particles with more desirable
properties in terms of size, size dispersion, and density while significantly decreasing the

metal loading on the electrode surface.

1.5 The Oxygen Reduction Reaction

As mentioned above, electrodes consisting of supported metal catalysts are used
in a variety of electrochemical technologies, including electrosynthesis and
electrochemical energy conversion (e.g., fuel cells). An example of this is the oxygen
reduction reaction (ORR), an important electrochemical reaction because of its relevance
in biological systems and energy conversion.”106 The OR can proceed in aqueous
electrolytes via two overall pathways: a direct, 4-electron pathway or an indirect, 2-
electron pathway.

101.102.107
(1,

Carbon electrodes, particularly diamon exhibit low activity for ORR as

compared to metal electrodes, like Pt. The OR at carbon electrodes proceeds via the 2-
1 103

electron pathway in acidic media, as reported by Yeager and shown in equation 1.5.

producing hydrogen peroxide.

22

02 + 2H+ + 2e' 2 H202 (E° = 0.498 v vs. Ag/AgCl)40 (1.5.1)

The ORR at Pt, on the other hand, proceeds with much lower overpotential via the direct,

4-e1ectron pathway in acidic media, as shown in equation 1.5.2, to produce water.

02 + 411* + 4e' .-:> 21120 (E° = 1.032 v vs. Ag/AgCl)4° (1.5.2)

The direct, 4-electron pathway is most desirable, particularly in fuel cells, as it
does not involve the formation of peroxide species in solution and the Faradaic efficiency
of the reaction is greater (more electrons per molecule). Figure 1.10 compares the ORR
in acidic media at diamond to that at Pt. The voltammetric i-E curve for the diamond
electrode shows a very negative reduction potential of ca. -800 mV vs. Ag/AgCl. This
corresponds to an overpotential of over 1 V. The i-E curve for the polycrystalline Pt
electrode however has a much more positive reduction peak potential of ca. 450 mV vs.
Ag/AgCl. This corresponds to a lower overpotential (ca. 500 mV). It is obvious from
this that Pt is a much more efficient catalyst for the reaction than diamond. However, due
to its considerable expense, there is a desire to achieve the highest catalytic activity at the
lowest Pt mass loading. Ideally, a non-Pt-based alternative electrocatalyst for the ORR is
desired. There have been numerous investigations of the OR at solid metal and metal
electrocatalyst-supported electrodes,97'99"03‘1”" '2 but Pt and Pt-based alloys are currently

the best electrocatalysts for this reaction."99

As an example, Adzic and coworkers
recently developed a procedure which reduces the Pt content to monolayer coverage

using a technique which involves the spontaneous replacement of predeposited Cu (upd)

23

with Pt on Pd nanoparticles. They are also investigating several Pt alloys using the same

procedure to firrther reduce the Pt contentm“ ‘5

 

 

    
 

 

 

 

o .—

A
N Diamond

8

i

:i.

V

3‘

'53 -100

=

o

e Polycrystalline Pt

5

r.

r...

:r

U

-200 r
n l A l 1 I L
-800 -400 0 400 800

Potential (mV vs. Ag/AgCl)

Figure 1.10. Linear sweep voltammetric i-E curves for the oxygen reduction reaction
(OR) in 0.1 M HClO4 at boron-doped microcrystalline diamond thin film and

polycrystalline Pt. Scan rate = l mV/s. Current is normalized to geometric electrode area,
0.2 cm .

Kinoshita previously reviewed various oxygen reduction reaction catalysts.116
However, surging interest in this area for purposes of fuel cells has produced several new
potential electrocatalysts. A comparison of several recently published studies on Pt- and
non-Pt-based electrocatalysts is presented in Table 1.2.1'3'1'7'12' It is sometimes difficult
to directly compare electrocatalyst performance in different reports due to the nature of
individual studies, and for this reason, Gasteiger and coworkers recently presented

activity benchmarks and requirements for Pt, Pt—alloy, and non-Pt oxygen reduction

24

catalysts.6 While Pt and Pt-based alloys are the most active catalysts, state-of-the-art
membrane electrode assemblies (MBA) in fuel cells possess Pt loadings of 0.6-0.8
mgpt/cszgA (Membrane Electrode Assembly) and exhibit power densities of 0.7
W/cszEA at 0.68 V vs. SHE and 80 °C. This translates into a Pt specific power density
of 0.85-1.l gm/kW (or 72-94 g of Pt for an ~85 kW stack).6 In order to satisfy large-scale
automotive applications because of cost and supply limitations, the Pt loading must
decrease 5-fold and the Pt specific power density must drop below 0.2 gpt/kW at
potentials greater than 0.65 V vs. SHE. Additionally, current specific and mass Pt
activities (at 0.9 v vs. SHE) are ca. 170-210 pA/cmzp. and 0.064-0.16 A/mgpt,
respectively. Gasteiger and coworkers point out that if a catalyst was essentially
"costless", it could occupy ~20x the volume of the 47% Pt/C cathode catalyst, the activity
of which needs to be doubled at a minimum. Therefore, activity of a "costless" cathode
needs to be no less than 1/10th of the current Pt activity. Table 6 in this reference

provides a comparison of several "costless" catalysts.6

25

 

m=m .m> 08 80:80.3— flmzaoaoq =< I a

 

 

 

 

 

 

 

 

 

cm :05 .832 A.\ mmzv U ow «a 80:58 oi u seem a . .
E-.." E .oweSaNS T. u 8% 3.5 . Noumea Asmsfié smxé .3: 5
God M iv > 3.o-w.o u moo . Us 8-34 UmEmE 0003-:m
~20 .8 0380 $08 I couficoto 253 E on >06 5 o N N a
menace oases s 0% £208 :02 .538 Sam . C m“. 4%“me 80 a on u s: 2-5 oNC
2 no 803300 AS: w-mv 0N6 2233 o o
8N- u 0?: .832: E- u 82m Ewe .
9.3 n 0?: cm? wdm u > mod :0 33:24 0002 o
Omomnm E E $.03 S o Us ca .406: 2 mo Om0m¢m a:
ANEQEmE
2: Ohm $.03 cm .30 9 083800 33:00 «MO 00.3083 o
N883: com S g cm C0>o 0380 0mg? :00 0
CE 3 034.888 .50 82mm . Noam 23
N u 0?: > ho ® N52 2. see 55:8 £08m . o 8 .822 a: ”Snot 84-8-2 w:
mam—Newman CD 8 3829 0.82 o o
02 55 33:8 55 532 2505 .
038200-50 088 030009 0529 E E 55 809:8 2:00 o
maoumozqam 0368230 Com N N a
$3800 038-8: 0 mo Combs 323 052 I @3300 3:00am 0 0 mm Wmmmm 0093.?“ \i:
Ohm :05 0032 x831 .0380 £003 0 o
33:00 360% 8033 03% $.03 0m N6 0
Ohm :05 03:08 088 >8 flimm 0B Sm o
. a
usage moi 3325 . 0. N .66: 2 3 .sshwwwwhmmwe a N:
ovszamme mo .830 5 .3 Ohm .85 No.33 ©3300 0002 0
3:55:00
.323350 13:25.33”.— .mbSaueboofi 00:383—

 

Bfiwfigfl mfl0m 382.50 flax—50080005 UmEmm 000095-52 van @0395 3:5qu mo Saga .~.— 030,—.

 

26

 

01

D2

pr

IO

SP

512

Ca

pn-

C0

mi

1.6 Dissertation Overview

This dissertation is divided into seven chapters with the research results focused
on metal phase formation on diamond and the potential use of diamond as an
electrocatalyst support material. In this work, I sought (i) to evaluate the potential of
microcrystalline and nanocrystalline diamond as an electrocatalyst support material by
assessing the physical and electrochemical properties of the material and (ii) to
understand metal phase formation, specifically Pt, on both microcrystalline and
nanocrystalline diamond in an attempt to better control and stabilize electrocatalyst metal
particles on the surface. All of the experimental details are reported in Chapter 2, while
Chapters 3-6 present the results, discussions and conclusions of the individual
investigations. Finally, the complete findings of this work as well as additional areas

worth investigating are summarized in Chapter 7.

Chapters 3 and 4 describe some of the basic physical and electrochemical
properties of diamond as well as the dimensional stability of the material during exposure
to aggressive electrochemical conditions, as these are important for a new support
material. As mentioned above, current electrocatalyst support materials are composed of
sp2 carbon. While these materials possess several benefits, they are lacking in physical
stability. Chapter 3 reports on the effect of controlled incorporation of sp2 nondiamond
carbon into the diamond microstructure in terms of physical and electrochemical
properties. The sp2 nondiamond carbon is incorporated through increasing the carbon
concentration in the growth gas. Chapter 4 assesses the potential use of high quality

microcrystalline and nanocrystalline diamond (i.e., low sp2 nondiamond carbon impurity

27

 

Ulll

of

pm

C01

C01

mi.

EX:

content) as an electrocatalyst support material. The microstructural and morphological
stability of the material is monitored under conditions typically observed in a PAFC.
Specifically, the extent of surface oxidation and morphological degradation was
determined to assess whether the material could stably support electrocatalyst particles

under harsh conditions.

As discussed previously, our group has published several papers on the deposition
of Pt nanoparticles onto the surface of microcrystalline diamond using both DC
magnetron sputtering and constant current electrodeposition.“’62‘64'65 However, particles
produced by both of these methods were too large for practical application. A
comprehensive study of the deposition of nanometer-sized Pt particles using a new
deposition approach, pulse galvanostatic deposition, is described in Chapter 5. It was our
supposition that this method would better control the metal particle properties in terms of
size, size dispersion, density and electrocatalytic activity. Additionally, the first
comparison of metal phase formation on two morphologically different types of diamond,

microcrystalline and nanocrystalline, is reported.

In an attempt to better understand the particle-diamond interaction, work to
examine the effect of the diamond surface chemistry on metal phase formation and
stability at both diamond types is described in Chapter 6. Understanding the nature of
catalyst loss should aid in our attempt to stabilize the electrocatalyst particles to the

diamond surface.

28

 

10.

11.

REFERENCES

E. Auer, A. Freund, J. Pietsch, and T. Tacke, Appl. Catal. A: Gen, 173, 259
(1998)

A. J. Appleby, in Comprehensive Treatise of Electrochemistry, B. E. Conway, J.
O. M. Bockris, E. Yeager, S. U. M. Khan, and R. E. White, Editors, 7, p. 173,
Plenum Press, New York, (1983).

M. L. Perry and T. F. Fuller, J. Electrochem. Soc., 149, SS9 (2002).

U. S. Department of Energy, in DOE Hydrogen Program - 2005 Annual Progress
Report, (2005).

U. S. Department of Energy, in Basic Research Needs for the Hydrogen
Economy: Report of the Basic Energy Sciences Workshop on Hydrogen
Production, Storage, and Use, Argonne National Laboratory, Chicago, 1L,150
pages (2003).

H. A. Gasteiger, S. S. Kocha, B. Sompalli, and F. T. Wagner, Appl. Catal. B:
Environ, 56, 9 (2005).

U. S. Department of Energy - Energy Efficiency and Renewable Energy, in
Hydrogen, Fuel Cells and Infrastructure Technologies Program: Multi- Year
Research, Development and Demonstration Plan, (2003).

R. L. McCreery, in Interfacial Electrochemistry: Theory, Experiment, and
Applications, A. Wieckowski, Editor, p. 631, Marcel Dekker, Inc., New York,
(1999)

G. M. Swain, in Electroanalytical Chemistry, A. J. Bard and I. Rubinstein,
Editors, 22, p. 181, Marcel Dekker, Inc., New York, (2004).

A. E. Fischer, Ph. D. Dissertation, Degradation Mechanisms of Carbon-Based
Electrocatalyst Support Materials and Development of an Advanced Support

Based on Electrically Conducting Diamond, Michigan State University, East
Lansing, MI, (2005).

G. M. Jenkins and K. Kawamura, Nature (London), 231, 175 (1971).

29

12.

13.

14.

15.

l6.

17.

18.

19.

20.

21.

22.

23.

24.

25.

R. L. McCreery, in Electroanalytical Chemistry, A. J. Bard and l. Rubinstein,
Editors, 17, p. 221, Marcel Dekker, Inc., New York, (1990).

F. Tuinstra and J. L. Koenig, J. Chem. Phys, 53, 1126 (1970).

W. A. Yarbrough and R. Messier, Science, 247, 688 (1990).

K. Ushizawa, K. Watanabe, T. Ando, I. Sakagucki, M. Nishitani-Gamo, Y. Sato,
and H. Kanada, Diamond and Rel. Mater., 7, 1719 (1998).

D. s. Knight and w. B. White, J. Mater. Res., 4, 385 (1989).

D. H. Karwik, I. F. Hu, S. Weng, and T. Kuwana, ACS Symp. Ser., 288, 582
(1985)

G. W. Hance and T. Kuwana, Anal. Chem, 59, 131 (1987).

J. F. Evans and T. Kuwana, Anal. Chem, 49, 1632 (1977).

J. F. Evans, T. Kuwana, M. T. Henne, and G. P. Royer, J. Electroanal. Chem, 80,
409 (1977).

E. A. McGaw and G. M. Swain, Anal. Chim. Acta, 575, 180 (2006).

J. Park, Y. Show, V. Quaiserova, J. J. Galligan, G. D. Fink, and G. M. Swain, J.
Electroanal. Chem, 583, 56 (2005).

G. M. Swain, A. B. Anderson, and J. C. Angus, MSBull., 23, 56 (1998).

M. Hupert, A. Muck, J. Wang, J. Stotter, Z. Cvackova, S. Haymond, Y. Show,
and G. M. Swain, Diamond and Rel. Mater., 12, 1940 (2003).

P. Sonthalia, E. McGaw, Y. Show, and G. M. Swain, Anal. Chim. Acta, 522, 35
(2004)

30

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

G. W. Muna, V. Quaiserova-Mocko, and G. M. Swain, Anal. Chem, 77, 6542
(2005).

Y. Song, X. Zhao, T. Voice, and G. M. Swain, Manuscript in preparation.

J. Xu, M. C. Granger, Q. Chen, J. W. Strojek, T. E. Lister, and G. M. Swain, Anal.
Chem, 69, 591A (1997).

M. C. Granger and G. M. Swain, J. Electrochem. Soc., 146, 4551 (1999).

J. Wang, G. M. Swain, M. Merrnoux, G. Lucazeau, J. Zak, and J. W. Strojek, New
Diamond Front. Carbon T echnol., 9, 317 (1999).

S. Alehashem, F. Chambers, J. W. Strojek, G. M. Swain, and R. Ramesham, Anal.
Chem, 67, 2812 (1995).

J. Asmussen and T. A. Grotjohn, in Diamond Films Handbook, J. Asmussen and
D. K. Reinhard, Editors, p. 211, Marcel Dekker, Inc., New York, (2002).

D. Zhou, D. M. Gruen, L. C. Qin, T. G. McCauley, and A. R. Krauss, J. Appl.
Phys, 84, 1981 (1998).

S. Jiao, A. Sumant, M. A. Kirk, D. M. Gruen, A. R. Krauss, and O. Auciello, J.
Appl. Phys, 90, 118 (2001).

D. M. Gruen, MRS Bull, 23, 32 (1998).

D. M. Gruen, Annu. Rev. Mater. Sci, 29, 211 (1999).

P. K. Bachmann, D. Leers, and H. Lydtin, Diamond and Rel. Mater., 1, 1 (1991).

P. K. Bachmann, H. J. Hagemann, H. Lade, D. Leers, F. Picht, and D. U.
Wiechert, Diamond, SiC, and Nitride Wide Band Gap Semiconductors,
Pittsburgh, PA, C. H. Carter, G. Gildenblatt, S. Nakamura, and R. J. Nemanich,
Editors, 339, p. 267, MS Symp Proc., Pittsburgh, PA Materials Research
Society, (1994).

31

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

A. J. Appleby, Corrosion (Houston), 43, 398 (1987).

A. J. Bard and L. R. Faulkner, Electrochemical Methods: Fundamentals and
Applications, John Wiley & Sons, Inc., New York (2001).

L. M. Roen, C. H. Paik, and T. D. Jarvi, Electrochem. Solid-State Lett., 7, A19
(2004).

M. F. Mathias, R. Makharia, H. A. Gasteiger, J. J. Conley, T. J. Fuller, C. J.
Gittleman, S. S. Kocha, D. P. Miller, C. K. Mittelsteadt, T. Xie, S. G. Yan, and P.
T. Yu, Interface, 14(3), 24 (2005).

F. Yuan and H. Ryu, Nanotechnology, 15, 8596 (2004).

F. Yuan, G. Sasikumar, and H. Ryu, J. New Mater. Electrochem. Sys. , 7, 311
(2004)

F. Liu, S.—H. Lee, Y. Yan, T. Gennett, B. J. Landi, A. C. Dillon, and M. J. Heben,
Electrochem. Solid-State Lett., 7, A421 (2004).

P. Serp, M. Corrias, and P. Kalck, Appl. Catal. A: Gen, 253, 337 (2003).

J. H. Bitter, M. K. van der Lee, A. G. T. Slotboom, A. J. van Dillen, and K. P. de
Jong, Catal. Lett., 89, 139 (2003).

V. Raghuveer and A. Manthiram, Electrochem. Solid-State Lett., 7, A336 (2004).

G. Wang, G. Sun, Z. Zhou, J. Liu, Q. Wang, S. Wang, J. Guo, S. Yang, Q. Xin,
and B. Yi, Electrochem. Solid-State Lett., 8, A12 (2005).

J. Ding, K.-Y. Chan, J. Ren, and F.-S. Xiao, Electrochim. Acta, 50, 3131 (2005).

A. Smimova, X. Dong, H. Hara, A. Vasiliev, and N. Sammes, Int. J. Hydrogen
Energy, 30, 149 (2005).

M. C. Granger, J. Xu, J. W. Strojek, and G. M. Swain, Anal. Chim. Acta, 397, 145
(1999)

32

53.

54.

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

M. D. Koppang, M. Witek, J. Blau, and G. M. Swain, Anal. Chem, 71, 1188
(1999).

Q. Chen, M. C. Granger, T. E. Lister, and G. M. Swain, J. Electrochem. Soc., 144,
3806 (1997).

G. M. Swain, J. Electrochem. Soc., 141, 3382 (1994).

V. Y. Dolmatov, Russ. Chem. Rev., 70, 607 (2001).

A. E. Fischer and G. M. Swain, J. Electrochem. Soc., 152, B369 (2005).

M. Boutonnet, J. Kizling, P. Stenius, and G. Maire, Colloids Surf, 5, 209 (1982).

J. V. Zoval, J. Lee, S. Gorer, and R. M. Penner, J. Phys. Chem. B, 102, 1166
(1998).

Y. Takasu, T. Fujiwara, Y. Murakami, K. Sasaki, M. Oguri, T. Asaki, and W.
Sugimoto, J. Electrochem. Soc., 147, 4421 (2000).

J. Wang, G. M. Swain, T. Tachibana, and K. Kobashi, Electrochem. Solid-State
Lett., 3, 286 (2000).

J. Wang, G. M. Swain, T. Tachibana, and K. Kobashi, J. New Mater.
Electrochem. Sys. , 3, 75 (2000).

B. E] Roustom, G. Féti, and C. Comninellis, Electrochem. Commun, 7, 398
(2005).

J. Wang and G. M. Swain, Electrochem. Solid-State Lett., 5, E4.(2002).

J. Wang and G. M. Swain, J. Electrochem. Soc., 150, E24 (2003).

M.-S. Ldffler, B. GroB, H. Natter, R. Hempelmann, T. Krajewski, and J. Divisek,
Phys. Chem. Chem. Phys, 3, 333 (2001).

33

“J.

7i

67.

68.

69.

70.

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

E. Budevski, G. Staikov, and W. J. Lorenz, Electrochemical Phase Formation
and Growth, VCH, New York (1996).

E. Budevski, G. Staikov, and W. J. Lorenz, Electrochim. Acta, 45, 2559 (2000).

R. M. Penner, J. Phys. Chem. B, 105, 8672 (2001).

D. Pletcher, A First Course in Electrode Processes, Alresford Press Ltd., UK.
(1991)

B. Scharifker and G. Hills, Electrochim. Acta, 28, 879 (1983).

H. Liu, F. Favier, K. Ng, M. P. Zach, and R. M. Penner, Electrochim. Acta, 47,
671 (2001).

M. Awada, J. W. Strojek, and G. M. Swain, J. Electrochem. Soc., 142, L42
(1995)

J. Wang, Ph.D. Dissertation, Fabrication, Characterization, And Electrocatalytic
Activity of Metal/Diamond Composite Electrodes, Michigan State University, East
Lansing, MI, (2002).

G. Siné and C. Comninellis, Electrochim. Acta, 50, 2249 (2005).

J.-S. Gao, T. Arunagiri, J.-J. Chen, P. Goodwill, and O. Chyan, Chem. Mater., 12,
3495 (2000).

J. Zak and M. Kolodziej-Sadlok, Electrochim. Acta, 45, 2803 (2000).

C. Prado, S. J. Wilkins, P. Grfindler, F. Marken, and R. G. Compton,
Electroanalysis, 15, 1011 (2003).

L. C. Hian, K. J. Grehan, C. H. Goeting, R. G. Compton, J. S. Foord, and F.
Marken, Electroanalysis, 15, 169 (2003).

M. E. Hyde, R. M. J. Jacobs, and R. G. Compton, J. Electroanal. Chem, 562, 61
(2004)

34

 

81.

82.

83.

84.

85.

86.

87.

88.

89.

90.

9].

92.

93.

94.

95.

N. Vinokur, B. Miller, Y. Avyigal, and R. Kalish, J. Electrochem. Soc., 146, 125
(1999)

Y. Zhang, S. Yoshihara, and T. Shirakashi, Electrochim. Acta, 51, 1008 (2005).

B. Riedo, G. Dietler, and O. Enea, Thin Solid Films, 488, 82 (2005).

O. Enea, B. Riedo, and G. Dietler, Nano Letters, 2, 241 (2002).

F. Montilla, E. Morallon, 1. Duo, C. Comninellis, and J. L. Vazquez, Electrochim.
Acta, 48, 3891 (2003).

J. V. Zoval, R. M. Stiger, P. R. Biemacki, and R. M. Penner, J. Phys. Chem, 100,
837 (1996).

J. L. Fransaer and R. M. Penner, J. Phys. Chem. B, 103, 7643 (1999).

S. Gorer, H. Liu, R. M. Stiger, M. P. Zach, J. V. Zoval, and R. M. Penner, in
Meta] Nanoparticles, D. L. Feldheim and C. A. Foss Jr., Editors, p. 237, Marcel
Dekker, Inc., New York, New York, (2002).

R. M. Penner, J. Phys. Chem. B, 106, 3339 (2002).

H. Liu and R. M. Penner, J. Phys. Chem. B, 104, 9131 (2000).

G. Gunawardena, G. Hills, 1. Montenegro, and B. Scharifl(er, J. Electroanal.
Chem, 138, 225 (1982).

P. Allongue and E. Souteyrand, J. Electroanal. Chem, 286, 217 (1990).

H. Kim, N. P. Subramanian, and B. N. Popov, J. Power Sources, 138, 14 (2004).

S. Wang and G. M. Swain, Manuscript in preparation.

L. Carrette, K. A. Friedrich, and U. Stimming, Fuel Cells, 1, 5 (2001).

35

96.

97.

98.

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

J. Y. Lucisano, J. C. Armour, and D. A. Gough, Anal. Chem, 59, 736 (1987).
A. Damjanovic and V. Brusic, Electrochim. Acta, 12, 615 (1967).

N. Giordano, E. Passalacqua, L. Pino, A. S. Arico, V. Antonucci, M. Vivaldi, and
K. Kinoshita, Electrochim. Acta, 36, 1979 (1991).

N. M. Markovic, T. J. Schmidt, V. Stamenkovic, and P. N. Ross, Fuel Cells, 1,
105 (2001).

H.-H. Yang and R. L. McCreery, J. Electrochem. Soc., 147, 3420 (2000).

T. Yano, E. Popa, D. A. Tryk, K. Hashimoto, and A. Fujishima, J. Electrochem.
Soc., 146, 1081 (1999).

T. Yano, D. A. Tryk, K. Hashimoto, and A. Fujishima, J. Electrochem. Soc., 145,
1870 (1998).

E. Yeager, Electrochim. Acta, 29, 1527 (1984).
I. Morcos and E. Yeager, Electrochim. Acta, 15, 953 (1970).

E. Yeager, J. A. Molla, and S. Gupta, in Electrochemistry of Carbon, S.
Sarangapani, J. R. Akridge, and B. Schumm, Editors, PV 84-5, p. 123, The
Electrochemical Society Proceedings Series, Pennington, NJ, (1984).

E. Yeager and D. Tryk, in Proceedings of the 33rd International Power Sources
Symposium, The Electrochemical Society, Inc., Cherry Hill, NJ, (1988).

J. A. Bennett, J. Wang, Y. Show, and G. M. Swain, J. Electrochem. Soc., 151,
E306 (2004).

B. C. H. Steele and A. Heinzel, Nature, 414, 345 (2001).

F. El Kadin', R. Faure, and R. Durand, J. Electroanal. Chem, 301, 177 (1991).

36

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

H. Kita, H.-W. Lei, and Y. Gao, J. Electroanal. Chem, 379, 407 (1994).

T. Abe, G. M. Swain, K. Sashikata, and K. Itaya, J. Electroanal. Chem, 382, 73
(1995)

Z. W. Zhang, D. Tryk, and E. Yeager, in The Electrochemical Society Extended
Abstracts, 83-2, p. Abstract 394, Washington, DC, (October, 1983).

J. Zhang, F. H. B. Lima, M. H. Shao, K. Sasaki, J. X. Wang, J. Hanson, and R. R.
Adzic, J. Phys. Chem. B, 109, 22701 (2005).

J. Zhang, M. B. Vukmirovic, K. Sasaki, A. U. Nilekar, M. Mavrikakis, and R. R.
Adzic, J. Am. Chem. Soc., 127, 12480 (2005).

J. Zhang, M. B. Vukmirovic, K. Sasaki, F. Uribe, and R. R. Adzic, J. Serb. Chem.
Soc., 70, 513 (2005).

K. Kinoshita, Electrochemical Oxygen Technology, John Wiley & Sons, Inc.,
New York (1992).

F. Jaouen, F. Charreteur, and J. P. Dodelet, J. Electrochem. Soc., 153, A689
(2006)

V. Raghuveer, A. Manthiram, and A. J. Bard, J. Phys. Chem. B, 109, 22909
(2005)

P. J. Bouwman, W. Dmowski, J. Stanley, G. B. Cotten, and K. E. Swider-Lyons,
J. Electrochem. Soc., 151, A1989 (2004).

L. Xiong and A. Manthiram, J. Electrochem. Soc., 152, A697 (2005).

R. G. Gonzalez-Huerta, J. A. Chavez-Carvayar, and O. Solorza-Feria, J. Power
Sources, 153, 11 (2006).

37

t

CHAPTER 2

Experimental Methods

2.] Diamond Thin-Film Deposition

Boron-doped microcrystalline (BMD) and nanocrystalline (BND) diamond thin-
film electrodes were deposited using microwave-assisted plasma chemical vapor
deposition (CVD; 1.5 kW, 2.54 GHz, Model AX3120, Astex, Inc., Lowell, MA) on
highly conducting p-Si(100) substrates (< 0.001 Q-cm, Virginia Semiconductor, Inc.)
The substrates were pretreated by hand polishing on a felt pad for 5 min using a 0-2 nm
diameter diamond powder (GE Superabrasives, Worthington, OH) slurried with ultrapure
water (18 M9). The scratched substrates were then rinsed with ultrapure water and 2-
propanol and sequentially cleaned by ultrasonication in acetone, 2-propanol and ultrapure
water (5 min each). The clean substrates were dried and examined under an Olympus
BX60M optical microscope (Olympus America, Inc.) to check for surface cleanliness. If
needed, the substrates were resonicated in acetone until the surface was visibly devoid of
polishing debris. This procedure served to microscopically scratch the substrate and
embed small diamond particles; both of which served as the initial diamond nucleation

sites.

The microcrystalline films were deposited for 10 h using a methane-to-hydrogen
(CH4 in H2) source gas ratio of 0.50%, except those discussed in Chapter 3, which were

deposited using CH4 in H2 source gas ratios of 0.3, 0.5, 1, 2, 3, and 5%. The total gas

38

 

 

Fig
Not

flow was 200 sccm with the microwave power, pressure and temperature held constant at
1 kW, 45 torr and ~800 °C, respectively. The nanocrystalline films were deposited using
gas flow rates of l, 94 and 5 seem for CH4, Ar, and H2, respectively. These films were
grown for 2 h using a microwave power of 800 W, a deposition pressure of 140 torr, and
a temperature of ~800 °C. All gases were ultrahigh purity (99.999%) grade (CH4 and H2
from AGA Specialty Gas, Cleveland, OH; Ar from BOC Group, Inc., Murray Hill, NJ).
The thickness was 3-7 pm for both film types. Figure 2.1 presents a schematic of the
CVD reactor used to grow the diamond. Note that the Ar was only used when growing
the nanocrystalline film types. The growth conditions for the two types of diamond films

and some of the film properties are summarized in Tables 2.1 and 2.2, respectively.

 

Microwave
Generator

 

 

 

 

Mass Flow
Controllers Antenna

W . . _ C(liggbzer

 

 

 

 

 

 

       
 
 
 

Substrate
Holder
(M0)

    

 

 

 

 

 

 

 

 

    
 

 

 

 

Gas Inlet

 

 

BzHa Cylinculer)-l><‘r
Ar Cylinder ){X}

H2 Cylinder
CH4 Cylinder

 

Rotary Pump TEL/2m

 

 

 

 

 

 

 

 

 

 

 

Figure 2.1. Schematic of MPCVD reactor used to grow polycrystalline diamond films.
Note Ar was only used for nanocrystalline diamond growth.

39

 

 

F
D
—TI-

I
E
1|-
-

the

Inc

 

Table 2.1. Conditions Used for Microcrystalline and Nanocrystalline Boron-Doped
Diamond Thin Film Deposition on Si

 

 

Parameter Microcrystalline N anocrystalline
Source Gases (sccm) CH4/H2 (1/ 199) CH4/H2/Ar (1/5/94)
Methane Concentration 0.5 % l %
Dopant 10 ppm B2H6 a 10 ppm B2H6
Microwave Power 1000 W 800 W

Growth Pressure 45 Torr 140 Torr

Substrate Temperature ~800 oC ~800 °C
Growth Time 10 h 2 h

 

 

 

' Microcrystalline diamond films discussed in Chapter 3 were doped using a B203 solid boron diffusion
source

 

Table 2.2. Summary of Physical Properties of Microcrystalline and Nanocrystalline
Boron-Doped Diamond Thin Films Deposited on Si

 

 

Film Resistivity Carrier Carrier
Diamond Thickness Q ) Concentration Mobility
(um) ( -cm (Blcm’) (cm2/V s)
Microcrystalline ~5-7 ~0.00l ~5 x 1019 ca. 5
Nanocrystalline ~4-6 ~o.002 ~5 x 1020 ca. 0.5

 

 

 

 

The microcrystalline diamond films, discussed in Chapter 3, that were deposited
using varying CH4/H2 source gas ratios were doped using a solid boron diffusion source,
B203 (GS-126, BoronPlus, Techniglas, Inc., Penysburg, OH).1 After these films were
grown, our group switched to a gaseous source of boron, B2H6. Both the microcrystalline
and nanocrystalline diamond films discussed in the subsequent chapters were doped from
the gas phase by adding 10 ppm B2H6 (0.1% B2H6 diluted in H2, Matheson Gas Products,

Inc.) to the source gas mixture. The actual flow of B2H6 into the growth chamber was 2.0

40

 

sccm for microcrystalline and 1.0 sccm for nanocrystalline. However, the actual boron
concentration incorporated into the film is much higher. Figures 2.2 and 2.3 show plots
relating the B2H6 concentration in the source gas to the actual boron concentration in the
thin-film determined using boron nuclear reaction analysis (0) and the film resistivity
determined using 4-pt. probe measurements OK) for microcrystalline and nanocrystalline
diamond thin films, respectively. The film resistivity was calculated using equation 2.1.1
where R is the resistivity (in Q-cm), t is the film thickness in cm (estimated to be 0.0006
and 0.0005 cm for microcrystalline and nanocrystalline, respectively) and r is the
resistance measured using a 4-pt. probe. The reactor was thoroughly cleaned prior to any
diamond growth via acetone-soaked KimWipes and having the quartz chamber dipped in

an HP bath and baked to minimize residual boron in the reactor.

R = 4.532tr (2.1.1)

It is obvious from the plots that the actual amount of boron incorporated into each
type of diamond film is larger than the gas phase concentration despite growing inside a
clean reactor. We also measured the thin-film boron concentration and resistivity for 10
ppm films grown several months later afier many diamond growths in order to have a
better understanding of the types of films used in our studies. The microcrystalline film

exhibited a [B] of 2036 ppm and a 4-pt. resistivity of ~0.001 Q-cm. The nanocrystalline

film exhibited a [B] of 906 ppm and a resistivity of ~0.002 Q-cm.

41

 

 

 

 

 

800 _ , 0.8
I ,A :
A 700 _‘ x /” -: 0,7
E : I” : A
a 600 _- ,x’ 1 0.6 E
T: : 1 T
a 500 ,- /’/ y2=73.0x +24.6 .- 05 g
E : / R =0.9986 ; 5.
:1 400 :- ;” 1 0.4 '5‘
In . ,, ‘ .5
a ’ ’1’ . .a
E 300 : l/ i 0.3 g
[-1 ’ /’ ‘ m
200 f /’ 1 0.2
h ,”’ X —-> i
100 _- f, -: 0.1
+’ x .
0 ' ' ‘ ' ‘ ‘ ‘ ' ‘ ‘ ‘ 0.0
0 2 4 6 8 10 12

Gas Phase [B2H6] (ppm)

Figure 2.2. Plot relating the concentration of B2H6 in the gas phase to that incorporated
into the microcrystalline diamond thin-film (0) and the film resistivity OK).

 

 

 

 

1400 ,' 1.0
E 1200 - x x/
a. - ll’ - 0.8 A
E 1000 - 5
a . X —’ II, c
_ x’ - 06 V
E 800 ,/ y=25.7x+65.0 b
. ,, R —0.9929 .3
= 600 '- <— ,/ {n
E I/ E 0.4 'a
E- ’ " °
400 *- l 0‘
h ’z”. X ‘ 0.2
200 "O I,"
1 1 1 I 1 1 1 l 1 1 1 J 1 1 1 l 1 1 1 l 1 L
0‘r 0.0
0 10 20 30 40 50 60

Gas Phase [BzH6] (ppm)

Figure 2.3. Plot relating the concentration of B2H6 in the gas phase to that incorporated
into the nanocrystalline diamond thin-film (0) and the film resistivity OK).

42

Following the deposition, the CH4 and B2H6 flows were stopped, and the films
remained exposed to the H2 (microcrystalline) or H2/Ar (nanocrystalline) plasma for an
additional 10 min. They were then gradually cooled over several minutes to less than 400
°C in the presence of atomic hydrogen by reducing the microwave power and system
pressure. This post-growth annealing in atomic hydrogen served to etch away
adventitious nondiamond sp2 carbon impurity, minimize dangling bonds, and ensure
maximum hydrogen termination. Both film types had a nominal boron dopant

concentration of ca. 1020 B/cm3 and a film resistivity of < 0.005 Q-cm.

2.2 Electrochemical Methods

2.2.1 Electrochemical Measurements

All electrochemical measurements were performed at room temperature in a
single compartment glass cell using either a CH Instruments Model 650A or 6508
Electrochemical Workstation (CH Instruments, Inc., Austin, TX). The design of the
electrochemical cell was reported previously and is shown in Figure 2.4.2 The diamond
electrodes were clamped against the bottom of a glass cell. A VitonTM O-ring was used to
define the exposed geometric area, 0.2 cm2. All currents were normalized to this
geometric area unless otherwise noted. In aqueous media, a Ag/AgCl reference electrode
was positioned in close proximity to the working electrode using a cracked-glass
capillary filled with 4 M KCl. All potentials reported herein are referenced to this
electrode. A large-area carbon rod served as the auxiliary electrode and was positioned
normal to the working electrode. In non-aqueous media, a Ag/Ag+ quasi-reference

electrode (QRE) was used in conjunction with a Pt foil auxiliary electrode. Electrical

43

 

f‘)

contact was made to the diamond working electrode by scratching the backside of the Si
substrate with a diamond scribe and then cleaning and coating the area with graphite from
a rod. The graphite-coated substrate was then placed in contact with a Cu plate current
collector. Once mounted in the cell, the electrode surface was typically cleaned by

soaking in distilled 2-propanol for 20 min.3

Carbon Rod
Ag/AgCl Reference Auxiliary Electrode
/ Electrode l

   
  
  
  

N,/o2

 

Gas Inlet
Diamond Working
Electrode
Clamp 7/////////////////////////////////2 /
, ----------- —— Cu Current
. m . /////////Ill/llllillllll/Illl/Ilflfl
Viton o-rmg Collector

Figure 2.4. Schematic of the single compartment electrochemical cell used in the
experiments.

2.2.2 Pulsed Galvanostatic Deposition of Pt

After evaluating an electrode's response for several well known redox systems,
the cell was thoroughly rinsed with ultrapure water and filled with 1 mM potassium
hexachloroplatinate (IV) (98%, Aldrich) dissolved in 0.1 M HClO4. The solution was
purged with N2 for 20 min prior to deposition. Nanometer-sized Pt particles were

deposited under an N2 blanket using the pulsed galvanostatic method. The hardware

 

fl

30

$01
In
0. l
pur
not

bell

consisted of a homemade galvanostat, configured from an OMNI 90 potentiostat
(Cypress Systems, Inc., Lawrence, KS) that was interfaced with a PCI-5401 signal
generating controller board (National Instruments, Austin, TX). A block diagram of the
instrument is shown in Figure 2.5. Note that the electrochemical cell only has the
diamond working electrode and carbon rod auxiliary electrode exposed to the solution.
The pulse generation was computer-controlled using National Instruments LabVIEW
software (version 6.1). The LabVIEW code for the potential pulsing program can be
found in Appendix A. All potential pulses were 1 s in length with a 1 s delay time
between each, yielding a 50% duty cycle (pulse time/period). A constant current (pulsed
format) was delivered to the cell by applying the voltage (outputted fi'om the PCI-5401
card) between the working and reference OMNI 90 leads across a precision resistor. The
pulse number and current density were systematically investigated to determine the

optimum deposition conditions in terms of nominal metal particle size and density.

After the Pt was deposited, the galvanostat was disconnected and the platinate
solution was removed. This was followed by a gentle rinse of the cell with ultrapure
water that was applied by flowing down the sides of the glass cell. Metal particles are
sometimes weakly attached to the hydrogen-terminated diamond surface and vigorous
rinsing can result in significant metal phase loss.4’5 The cell was then filled with a fresh
0.1 M HClO4 solution, again by pouring the solution down the sides of the cell, and
purged with N2 for 30 min. The gas was introduced far enough from the surface so as to
not disturb the catalyst particles. Under the N2 blanket, the potential was cycled 10 times

between -200 and 1300 mV vs. Ag/AgCl at a rate of 50 mV/s to clean the Pt deposits.

45

The potential was kept within this window to avoid significant gas evolution, which can
physically dislodge the particles.“’5 The voltammetric i-E curve usually stabilized within
5 cycles. The voltammetric charge for H+ adsorption was used to determine the
electrochemically-active Pt surface area, assuming a lelPt surface interaction and a
coulombic charge of 210 11C/cm2 for hydrogen adsorption on an atomically flat Pt(111)

surfacef"8

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

CPU PC1-
5401
A W R 3:
Omni—90
Potentiostat
/_ Carbon rod
W— ? N,/02

gas inlet

R=le

    
  

O O
Vitonm o-rmg
W/ll/lllllllllllllll

—)

  

 

Cu Contact Diamond

working electrode

U/I/I/I/I/I/I/I/I/I/A

 

 

Figure 2.5. Experimental set-up used for pulsed galvanostatic deposition of Pt.

46

2.2.3 Reagents

2.2.3.1 Aqueous Solutions

All chemicals were reagent-grade quality or better and used without additional
purification. The electrochemical response of the electrodes was assessed using some or
all of the following: 1 mM solutions of potassium ferrocyanide (Fe(CN)63’/4', Aldrich),
hexaammineruthenium (III) chloride (Ru(NH3)63+/2+, Aldrich), 4-tert-butylcatechol (4-
tBC, Aldrich), and ferric sulfate (Fey/2+, Fisher) prepared fresh daily. The supporting
electrolyte was either 1 M KC1 (Spectrum) or 0.1 M HClO4 (redistilled, 99.999%,
Aldrich). The oxygen reduction reaction measurements were performed in either 0.1 M
HClO4 (C and Pt/C) or 0.1 M NaOH (C only) afier purging the solution with 02 for 30

min.

2.2.3.2 Non-aqueous Solutions

The films were exposed to either 1 mM 4-nitrobenzenediazonium
tetrafluoroborate (4-NBD) (97%, Aldrich) or 4-methoxybenzediazonium
tetrafluoroborate (4-MBD) (98%, Aldrich) dissolved in 0.1 M tetrabutylammonium
tetrafluoroborate (NBmBF4) (99%, Aldrich) plus acetonitn'le (ACN). The 4-NBD salt
was recrystallized following a procedure presented previously.9 The salt was dissolved in
cold ACN (0 0C) and recrystallized by adding cold (0 0C) diethyl ether. The crystals
were then collected via vacuum filtration, allowed to dry, placed in a glass vial and
purged with N2. The vial was then sealed and placed in a freezer for storage. The 4-
MBD salt was used as received and, after opening, was also stored in a freezer.10 The

NBmBF4 salt was purified by dissolving in ethyl acetate and precipitating one time using

47

t9

'9
2

P}:

Sp)

N11
in 1.

H00

pentane. The salt was collected and dried at ca. 120 °C overnight (>10 h) under
vaccum.11 The ACN was dried using 3 A molecular sieves that were activated by heating
at 300 0C for 15 h followed by cooling in a dessicator. The ACN was treated with the
activated sieves (5-10% w/v) for several days and then transferred to a clean amber bottle

and allowed to stand over a second batch of molecular sieves.”

2.3 Physical Characterization Methods

The physical properties of the microcrystalline and nanocrystalline diamond films
were routinely monitored using atomic force microscopy (AFM), scanning electron

microscopy (SEM), Raman spectroscopy, and powder X-ray diffraction spectroscopy.

2.3.1 Atomic Force Microscopy

2.3.1.1 Contact Mode AF M

The film morphology was investigated by AF M using a Nanoscope III instrument
(Veeco Metrology Group, Inc., Santa Barbara, CA) operated in the contact mode.
Pyramidal-shaped Si3N4 tips, mounted on gold cantilevers (100 mm legs, 0.38 N/m

spring constant), were used in air to acquire the topographical images.

2.3.1.2 Conductive Probe-AFM (CP-AF M)

CP-AFM measurements were performed by Shiua Wang in our group using a
NanoScope Illa instrument (Veeco Instruments, CA). A Au-coated silicon tip was used
in the contact mode. A homemade electrical circuit was used to measure the current

flowing between the conductive tip and the electrode surface as the tip was rastered over

48

an area. The diamond samples were placed on a metal stage and electrical contact was
made to the bottom of the conducting Si substrate using carbon tape. A bias voltage of
either i0.2 or 12 V was applied between the tip and the substrate. Conductivity maps

were generated simultaneously with the topographical images.

2.3.2 Scanning Electron Microscopy

The films were also physically characterized using either a J SM-64OOV (LaB6) or
JSM-6300F (field emission) scanning electron microscope (JEOL LTD., Tokyo, Japan,
Center for Advanced Microscopy at MSU). Micrographs were recorded using both
secondary and backseattered electrons generated with an accelerating voltage of 10 or 20
kV. Particle analysis was performed with the AnalySIS imaging software (Software
Imaging System Corp., Lakewood, CO). In the cases where Pt deposition was
investigated, the presence of Pt was verified by energy dispersive x-ray (EDX) analysis

(Noran Instruments Inc., Middleton, WI) in conjunction with the J SM-6400V.

2.3.3 Raman Spectroscopy

The film microstructure was studied using Raman spectroscopy. Spectra were
acquired at room temperature using a Chromex RAMAN 2000 spectrograph (Chromex,
Inc., Albuquerque, NM) consisting of a diode-pumped, frequency-doubled CW
Ndzyttin'um aluminum garnet (YAG) laser (500 mW at 532 nm, COHERENT), a
Chromex 500is spectrometer (f/4, 600 grooves/mm holographic grating, 50 um slit
width), and a therrnoelectrically cooled, 1024 x 256 element charge-coupled device

(CCD) detector (ANDOR Tech., Ltd.) Spectra were collected with an incident power

49

density of ca. 500 kW/cm2 (100 mW at the sample and a 5 pm diameter spot size), a 10 8
integration time and an approximate spectral resolution of 4 cm]. A white-light spectrum

was collected and used as the background for spectral correction.

2.3.4 X-ray Diffraction Spectroscopy (XRD)

The crystallinity of diamond films grown using different CH4/H2 source gas ratios
was investigated using powder XRD. Spectra were obtained by scanning 20 from 20 to
100° on a Rikagu Rotaflex RTP300 RC instrument. X-rays were produced at 1.540 A

from a Cu anode.

2.4 Chapter-Specific Experimental Methods

2.4.1 Chapter 3

This chapter describes an investigation into the effect of nondiamond sp2 carbon
impurity on the physical and electrochemical response of diamond thin-film electrodes.
The electrochemical measurements were performed in a single compartment glass cell as
described above and shown in Figure 2.2. Once mounted in the cell, the electrode surface
was cleaned by soaking in distilled 2-propanol for 10 min. All measurements were
performed at room temperature. Prior to the oxygen reduction reaction (ORR)
measurements, the solution was saturated with 02 (99.5% purity, BOC Group, Inc.,
Murray Hill, NJ) by bubbling the gas through the solution for 20 min. All other cyclic
voltammograms (CVs) were obtained after solution degassing ’with bubbling N2 for 15

min and then maintaining a blanket of N2 over the solution during the measurements.

50

2.4.2 Chapter 4

This chapter presents an investigation of the microstructural stability of both
microcrystalline and nanocrystalline diamond thin films during exposure to harsh
conditions similar to those of a phosphoric acid fuel cell (PAFC). The electrode was
mounted in the single compartment electrochemical cell, as described above. The
electrode surface was first rinsed copiously with ultrapure H20 and then the cell was
filled with 85 wt% H3PO4 (99.999+%, Aldrich) and heated to ca. 180-185 °C using
heating tape connected to a variac. The heating was performed in a ventilated hood, and
the instrumentation was covered with cling wrap to keep the corrosive vapor out of the
lab and instrumental circuitry. The set-up was insulated with cotton and aluminum foil to
maintain temperature consistency. Upon reaching the desired temperature, +20 mA of
current (or 0.1 A/cmz) was applied for 2 h. After 2 h, the solution was allowed to cool to
room temperature. The cell was then emptied and rinsed thoroughly with ultrapure

water.

In addition to probing the cyclic voltammetric background current in N2-saturated
0.1 M HClO4, the electrode response toward several well known redox systems
(Fe(CN)6'3/4, Ru(NH3)6+3/+2, and Pea/+2) was evaluated before and after the anodic

polarization.2"2’l3

The extent of surface oxidation was assessed from contact angle measurements

performed with a FTAZOO (fn‘st ten angstroms) contact angle analyzer. Ultrapure H2O

with a drop volume of 10 11L was used and both the left and right contact angles between

51

the drop and the diamond surface were measured in triplicate. The reported value is the

average of these 6 values.

2.4.3 Chapter 5

This chapter presents an in-depth investigation into the nucleation mechanism of
Pt on different diamond surfaces. The electrodes were mounted in the single
compartment electrochemical cell shown in Figure 2.2. Chonoamperometric i—t curves
were obtained at applied overpotentials (r1) ranging from -312 to -1012 mV based on the

redox reaction shown in 2.4.1.‘4

PtCl62‘ + 2:»: (:2 PtCl42' + 2Cl' (13° = 530 mV vs. Ag/AgCl) (2.4.1)

This value is used as the reduction potential even though, as shown, the reduction does
not proceed to Pto. However, as shown in equation 2.4.2, the next step in the reduction
process occurs at a more positive reduction potential.14 Therefore, we assume that once

the initial 2 electrons are transferred, the reduction quickly proceeds to completion.

PtCl42' + 2e' .—_> Pto + 40‘ (13° = 561mV vs. Ag/AgCl) (2.4.2)

The resulting chronoamperometric i-t curves were then compared to those predicted by

Schariflcer and Hills for instantaneous and progressive nucleation. '5

52

Tel

c0;

Sol

This chapter also contains a discussion of the effect of the pulse parameters (i.e.,
pulse number and current density) and diamond film morphology on the particle size,
shape and dispersion. Glassy carbon was used for comparison measurements. The
electrode (GC-30, Tokai, Ltd., Japan) was manually polished on separate felt polishing
cloths using aqueous slurries of 1.0, 0.3 and 0.05 pm diameter deagglomerated alumina
powder (Buehler, Ltd., Lake Bluff, 1]). After each polishing step, the electrode was
copiously rinsed with ultrapure water and ultrasonically cleaned in this medium for 10-15
min. After the final polishing and sonication step, the electrode was dried in a stream of
N2 and placed in the electrochemical cell where it was exposed to distilled 2-propanol for

10 min as a final cleaning step.

Atomic absorption spectroscopy (AAS) was used to determine the catalyst
loading for a given set of deposition conditions. For these measurements, the Pt particles
were deposited onto a separate set of diamond films as a function of the current density
and pulse number. After deposition, the Pt metal was then dissolved from the electrode
by soaking for 1 h in 4 mL of warm aqua-regia, which consisted of 3 mL HCl (36.5 —
38.0%, Columbus Chemical Industries, Inc., Columbus, WI) and 1 mL HNO3 (70%,
Columbus Chemical Industries, Inc., Columbus, WI). Pt is known to dissolve in warm
aqua-regia, while the diamond lattice is quite stable in this solution and undergoes no

morphological alteration. I 6" 7

Upon cooling to room temperature, the electrode was
removed from the solution and rinsed thoroughly with ultrapure water. The rinses were
collected and mixed with the original solution to collect all the dissolved metal. The total

solution volume was 50 mL. AAS was performed directly on this solution using a

53

Hitachi Z-9000 Atomic Absorption Spectrophotometer (Hitachi, Ltd., Tokyo, Japan) with
a graphite tube fumace. The metal ion concentration in each solution was determined

from multiple measurements in order to estimate the metal loading.

2.4.4 Chapter 6

This chapter describes the effect of surface chemistry on the formation of Pt metal
phases on microcrystalline and nanocrystalline diamond. Four different surface
chemistries were investigated: H-terminated, chemically oxidized and modified with aryl-

adlayers using either 4-nitrobenzene or 4-methoxybenzene diazonium salts.

The oxygen-terminated microcrystalline (BMDox) and nanocrystalline (BNDOX)
films were prepared by gentle refluxing in warm aqua-regia (3:1 HCleNO3) for 30 min
followed by refluxing in 30% H202 for 30 min and cooling in air to room temperature.
This procedure introduces oxygen functionalities on the surface while leaving the bulk
microstructure of the diamond unaltered. X-ray photoelectron spectroscopy was
performed by Wayne Jennings at Case Western Reserve University and used to quantify
the O/C ratio of the diamond. The H-terminated microcrystalline and nanocrystalline
diamond surfaces both exhibited an atomic O/C ratio of ca. 0.05 as some oxygen is
slowly picked up during air exposure. Assuming that 5 layers of carbon were sampled
(ca. 5 x 10‘5 atoms/layer), this ratio corresponds to a surface coverage of approximately
25% of a monolayer.18 However, atomic O/C ratios of ca. 0.13 and 0.10 were observed
for the oxidized microcrystalline and nanocrystalline diamond surfaces, respectively.

This corresponds to a surface coverage of 65 and 48% of a monolayer, respectively, for

54

the two diamond surfaces. Therefore, the diamond surfaces possess approximately 2 — 3

times more surface oxygen following the chemical oxidation treatment.

The diazonium modified surfaces were prepared using a well known
electrochemical reduction procedure.”’20 The overall modification process for the two
diazonium salts used is depicted in Scheme 2.1. H-terminated diamond films were
mounted in the single compartment electrochemical cell presented in Figure 2.2. A
CHEMRAZ® o-ring (Greene, Tweed & Co., Inc., Kulpsville, PA) was used to confine the

solution and define the electrode area of 0.2 cm2

. The films were exposed to the non-
aqueous solutions described above and the surface modification was performed by
cycling the potential once from 600 mV to a value slightly negative of the diazonium
reduction peak (-600 or -1300 mV for 4-NBD and 4-MBD, respectively). A Ag/Ag+
quasi-reference (QRE) and Pt foil counter electrode were used, and the potential sweep
rate was 200 mV/s. After modification, the cell was thoroughly rinsed with clean ACN
and a cyclic voltammogram was obtained after filling the cell with the NBu4BF4/ACN
electrolyte solution. The 4-NBD modified surface exhibited a reversible couple for the
le' redox reaction to form N02” at ca. -1200 mV indicating that the diamond surface had
indeed been modified. These films will be denoted as BMD4.NBD and BNDsNBD. The 4-
MBD modified electrode showed a single reduction peak at ca. -900 mV, again
signifying that the diamond surface had indeed been modified. In either case, the cell

was then sequentially rinsed with ACN, a 50/50 ACN/ultrapure H2O mixture and

ultrapure H20.

55

‘\ f

I

The 4-MBD modified surface underwent an additional modification step to
change the charge of the diamond surface. This procedure was presented by Yan and
coworkers and is illustrated in Scheme 2.1.10 The cell containing the 4-MBD modified
electrode was filled with cone. H2804 (98%, Columbus Chemical Industries, Inc.,
Columbus, WI) and heated to ~80 oC. Once 80 °C was reached, the temperature was held
constant for 1 hr. According to Yan and coworkers, this results in the addition of sulfonic
acid groups to the ring at the ortho positions to the methoxy group.10 After an hour, the
solution was allowed to cool to room temperature. The cell was then emptied, rinsed
copiously with ultrapure water, and filled with the Pt salt solution for the metal phase

formed as described above.

 

R = N02, OCH3
R ~———>
+1 e' Pt4+ Pt4+
‘ I OCH3 I
+ H035 SO3H
I N2 0 Conc. HZSO,
W —’ —'

 

 

Scheme 2.1. Diazonium surface modification process. The potential was cycled once
from 600 to either -600 (4-NBD) or -1300 mV (4-MBD) vs. Ag/Ag+ QRE at a rate of 200
mV/s. The 4-MBD modified diamond surface was then exposed to concentrated H2SO4
for 1 h at a temperature of ~ 80 0C.

56

10.

11.

12.

REFERENCES

J. A. Bennett, J. Wang, Y. Show, and G. M. Swain, J. Electrochem. Soc., 151,
E306 (2004).

M. C. Granger, M. Witek, J. Xu, J. Wang, M. Hupert, A. Hanks, M. D. Koppang,
J. E. Butler, G. Lucazeau, M. Mermoux, J. W. Strojek, and G. M. Swain, Anal.
Chem, 72, 3793 (2000).

S. Ranganathan, T.-C. Kuo, and R. L. McCreery, Anal. Chem, 71, 3574 (1999).

N. Vinokur, B. Miller, Y. Avyigal, and R. Kalish, J. Electrochem. Soc., 146, 125
(1999)

S. Nakabayashi, D. A. Tryk, A. Fujishima, and N. Ohta, Chem. Phys. Lett., 300,
409 (1999).

J. Perez, A. A. Tanaka, E. R. Gonzalez, and E. A. Ticianelli, J. Electrochem. Soc.,
141,431(l994)

A. Essalik, K. Amouzegar, and O. Savadogo, J. Appl. Electrochem., 25, 404
(1995).

J. M. D. Rodriguez, J. A. H. Melian, and J. P. Pena, J. Chem. Ed., 77, 1195
(2000)

A. O. Solak, L. R. Eichorst, W. J. Clark, and R. L. McCreery, Anal. Chem, 75,
296(2003)

M. M. Waje, X. Wang, W. Li, and Y. Yan, Nanotechnology, 16, S395 (2005).

A. J. Fry, in Laboratory Techniques in Electroanalytical Chemistry: Second
Edition, Revised and Expanded, P. T. Kissinger and W. R. Heineman, Editors,
Marcel Dekker, Inc., New York, (1996).

Y. Show, M. A. Witek, P. Sonthalia, and G. M. Swain, Chem. Mater., 15, 879
(2003)

57

13.

14.

15.

16.

17.

18.

19.

20.

A. E. Fischer, Y. Show, and G. M. Swain, Anal. Chem, 76, 2553 (2004).

A. J. Bard and L. R. Faulkner, Electrochemical Methods: Fundamentals and
Applications, John Wiley & Sons, Inc., New York (2001).

B. Scharifker and G. Hills, Electrochim. Acta, 28, 879 (1983).

J. Wang and G. M. Swain, Electrochem. Solid-State Lett., 5, E4 (2002).

J. Wang and G. M. Swain, J. Electrochem. Soc., 150, E24 (2003).

G. M. Swain, in Electroanalytical Chemistry, A. J. Bard and I. Rubinstein,
Editors, 22, p. 181, Marcel Dekker, Inc., New York, (2004).

M. Delamar, R. Hitmi, J. Pinson, and J. M. Savéant, J. Am. Chem. Soc., 114, 5883
(1992)

J. Pinson and F. Podvorica, Chem. Soc. Rev., 34, 429 (2005).

58

CHAPTER 3

Effect of spz-Bonded Nondiamond Carbon
Impurity on Boron-Doped Polycrystalline
Diamond Thin-Film Electrodes

3.1 Introduction

Research in our group is focused on understanding how the physical, chemical
and electronic properties of boron-doped diamond thin films affect the electrode response
in both aqueous and nonaqueous media."7 Several factors can influence the electrode
response (i.e., electrode reaction kinetics): potential-dependent density of electronic
states, surface chemistry, morphology and microstructure, defect density, nondiamond

sp2 carbon impurity content and double-layer structure.

Common to all electrodes is the fact that the rate of heterogeneous electron
transfer is influenced, among other things, by the density of electronic states (DOS) at a
given potential. The surface concentration of electrons is important for reduction
reactions, while the concentration of unoccupied states (i.e., acceptors) is important for
oxidations.8 For metals, the number of electronic states available at the surface is greater
than the number of molecules in solution being electrolyzed, therefore, electron transfer
rates being limited by DOS is not an issue. However, the number of charge carriers for
semiconductor electrodes can be of the same order of magnitude as the number of

molecules in solution, and therefore, is a factor that will limit the current flow (i.e., rate

59

of reaction). The density of states (DOS) for chemical vapor deposited (CVD) diamond
is influenced in a complex manner by the boron-doping level, lattice hydrogen,
adventitious nitrogen impurity, nondiamond sp2 carbon impurity and the defect density.
Undoped diamond films deposited by CVD, using conventional CH4/H2 source gas
mixtures, are generally electrically insulating, if they are free of significant levels of
nondiamond carbon and lattice hydrogen. However, incorporated nondiamond carbon
impurity can function electronically by providing charge carriers and pathways of high
canier mobility and/or physically by catalyzing redox reactions, providing adsorption
sites for reactant and reaction intermediates, providing sites for the formation of carbon-

oxygen fiinctionalities and modifying the electrode double-layer structure.

The electrochemical properties of polycrystalline CVD diamond have been
investigated in some detail in recent years.“"‘5'7'9'l3 These electrodes can be viewed as
having a composite architecture consisting of micrometer-sized grains, or crystallites, of
sp3-bonded carbon (nanometer-sized grains for nanocrystalline diamond), incorporated
dopant atoms such as boron, extended and point defects and grain boundaries where spz-
bonded carbon can form. The deposition conditions are usually selected to minimize the
incorporation of spz-bonded nondiamond carbon impurity; however, this impurity can be
introduced in a controlled manner by adjusting the deposition conditions. Presently, the
spz-bonded carbon was introduced by increasing the CH4/H2 source gas mixture used to
deposit the films. Increasing this ratio generally leads to more impurity incorporation. It
has been known for some time that the cyclic voltammetric background current increases

and the working potential window for aqueous electrolytes decreases with increasing spz-

60

ban:

ell

inc

inf

of

0.

1i

bonded nondiamond carbon content.2'3'7’9 However, there has been no systematic study
of how the electrode response toward various solution redox analytes is affected by the
incorporation of the impurity. One study by Scherson and coworkers showed the positive
influence of nondiamond carbon impurity on solvent reduction processes in lithium-ion
battery solvents.l4 In another more recent work, Duo et a]. attempted to study the effect
of spz-carbon impurity by mechanically abrading diamond with graphite particles.” Our
films have the nondiamond carbon impurity incorporated into the diamond matrix, and
this is far different from simply placing graphite particles on the surface. In related work,
Shi and coworkers have deposited N-incorporated amorphous carbon films that have a
mixed sp3/sp2 microstructure.16 The sp2 carbon domains play a key role in the electrode

activity.

The effect of nondiamond spZ-bonded carbon impurity on the physical and
electrochemical properties of boron-doped diamond electrodes fabricated by CVD is
reported on herein. Several aqueous-based redox analytes [Fe(CN6)3’/4', Ru(NH3)63+’2+,
Fey/2+, 4-tert—butylcatechol (4-tBC), and O2/H2O2] were used to evaluate the electrode
response. The electrodes were comprehensively characterized by ex situ atomic force
microscopy (AFM), scanning electron microscopy (SEM), Raman and powder X-ray
diffraction spectroscopy, and voltammetric methods of analysis. The results demonstrate
that for some redox analytes, nondiamond carbon impurity has a relatively minor effect,
while for others, like the oxygen reduction reaction, the impurity type and relative

amount exhibit a strong influence on the electrode-reaction kinetics.

61

3.2 Results & Discussion

3.2.1 Effect of CH4/H2 Ratio on the Physical Properties

Diamond films were deposited on p-Si(100) substrates using CH4/H2 source gas
ratios of 0.3, 0.5, 1, 2, 3 and 5% (% = (sccm CH4/sccm H2) x 100). Figure 3.1 shows
representative atomic force microscopic (AFM) images (ex situ) of films deposited from
different CWH2 source gas ratios. Clearly, the grain size decreases with increasing
CH4/H2 ratio resulting in macroscopically smoother films. The extent of secondary
nucleation increases with increasing CH4/H2 ratio, and this leads to a decrease in the
nominal grain size (i. e., smoother films) and an increase in the fraction of grain boundary.
It is at these grain boundaries that much of the nondiamond sp2 carbon impurity exists.
Raman imaging has revealed that the nondiamond carbon impurity in polycrystalline
films often exists locally at the grain boundaries.17 These trends are reflected in the data
presented in Table 3.1. At the high CH4/H2 ratios, the rate of renucleation exceeds the

rate of crystal growth, and this leads to the smooth texture.

 

Table 3.1. Nominal Diamond Crystallite Diameter and Height for Films Deposited from
Different Methane-to-Hydrogen Source Gas Ratios

 

 

CH4/H2 Ratio Nominal Crystallite Diameter Maximum Grain Height
(%) (um) (mm)
0.3 2.7 :l: 0.6 1.0 i 0.2
0.5 2.0 i 0.2 1.3 i 0.2
1 1.8:t0.3 1.0i 0.3
2 1.5 :1: 0.1 0.72 i 0.08
3 1.4:0.1 10:01
5 0.5 i 0.1 0.47 i 0.08

 

 

 

 

62

E E
= 1:
c c
c c
S 8
10pm
s a
c 9
§ §
10 pm
E E
= 8
c
8 8
o
10pm

 

Figure 3.1. Atomic force micrographs of diamond films deposited from different CH4/H2
source gas mixtures.

63

Scanning electron microscopy (SEM) was also used to assess the physical
features of the diamond films deposited using the different CH4/H2 ratios. These are
presented in Figure 3.2. There is little difference between the 0.3 and 0.5% films
consistent with the AFM images. Both possess large crystals with extremely smooth
facets and little defect density. The 1% film also possesses large crystals; however, the
facets of the crystals are notably rougher. This ratio may be where the deposition of
nondiamond spz-bonded carbon begins. This is in agreement with the Raman spectrum
for this film presented in Figure 3.3. The 2% CH4/H2 ratio film has slightly smaller
crystals than the 1% film. Additionally, the facets are increasingly rougher, possessing a
large number of defects. Interestingly, the Raman spectrum indicates minimal
nondiamond carbon impurity residing within this film. Increasing the CH4/H2 ratio to 3%
results in the deposition of nondiamond spz-bonded carbon all over the surface, especially
within the grain boundaries. As with the AFM images, diamond crystals are still visible
on the surface. When the CH4/H2 ratio is increased to 5%, there are no visible diamond
crystals in the film. Instead, the surface is macroscopically smooth with dendritic-like
features and likely possesses mostly nondiamond spZ-bonded carbon. This is consistent

with the Raman spectrum for this film presented in Figure 3.3.

64

 

Figure 3.2. Scanning electron micrographs (secondary electron) of diamond films
deposited from different CH4/H2 source gas mixtures.

65

Raman spectrosc0py was used to probe the film microstructure. This technique is
routinely used to characterize the microstructure of both sp2- and sp3-bonded carbon
materials.'8'22 Spectra for films deposited from different CH4/H2 ratios are displayed in
Figure 3.3. The first-order phonon modes for Si at 514 cm’1 and diamond at 1332 cm'I
are denoted by the dotted lines. The first-order diamond line intensity, width and position
are sensitive to such factors as the crystallite size, optical density, boron doping level, net
stress, defect density and film quality.'8°'9‘23 One evident trend in Figure 3.3 is the
decrease in peak amplitude and increase in peak width of the diamond phonon line with
increasing CH4/H2 ratio. These phenomena are better illustrated by the inset in Figure 3.3
which shows an enlargement of the spectral region surrounding the diamond phonon
peak. Table 3.2 summarizes the line position and linewidth data for films deposited from

the different source gas ratios.

 

Table 3.2. Raman Spectroscopic Data for Films Deposited from Different Methane-to-
Hydrogen Source Gas Ratios

 

CH4/H2 Ratio Average Diamond Peak Position Average Diamond Linewidth

 

(%) (cm'l) (FWHM, cm")
0.3 1333.7 1 0.2 12.7 1 0.3
0.5 1332.4 1 0.5 15 1 2

1 13326103 1612

2 13330108 1712

3 1332.8 1 0.1 40 1 20

5 NA NA

 

 

 

 

66

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Diamond
5%
% .
Diamond 5 sp2 carbon
2%
1% ,
0.5% 1% 1 _ 5%’
0.3% g
1250 1300 1350 1400
CE ‘
o , l 3%
E r wow“ .
2%
"A: ‘ r ~~ 1%
0.5%
w0.3V
200 500 800 l 1 00 1400 l 700 2000
Raman Shift (cm-1)

Figure 3.3. Raman spectra of diamond films deposited from different CH4/H2 source gas
mixtures. Inset is enlarged region around diamond peak.

The diamond linewidth is a qualitative indicator of the crystalline film quality.3’l7

For comparison, the linewidth (FWHM) for a high pressure high temperature piece of

diamond used to calibrate the instrument was ca. 5-7 cm'1 (a perfect single crystal sample

should be ca. 2-3 cm"). To a first approximation, the linewidth is inversely related to the

67

phonon lifetime and is a measure of the defect density. The Si line intensity decreased as
the CH4/H2 ratio was increased, due to increasing diamond film thickness as the growth
rate scales with carbon concentration in the gas phase. The diamond line intensity
decreased because the films became more opaque due to the increasing levels of spz-
bonded nondiamond carbon impurity. The elevated level of spz-bonded carbon is
evidenced by the scattering intensity between 1500 and 1590 cm", which increased with
increasing CH4/H2 ratio.3’9"7'1“”24'29 The nondiamond carbon impurity most likely
consists of a mixture of sp2- and sp3-bonded carbon. Scattering intensity for the
nondiamond carbon is most pronounced for the 3 and 5% CH4/H2 films but is also
observed for the 2% film and even minimally for the 1% film. This is in agreement with
the SEM data above. Interestingly, the 2 and 3% films exhibit a peak maximum at ca.
1500 cm'1 (1487 and 1495 cm", respectively). Scattering at this wavenumber is
characteristic of amorphous carbon (i. e., a mixture of sp3- and spZ-hybridized

' 7,1 , 4, .
b0nd1ng).l 82 2728

As previously mentioned, the first-order diamond phonon line was not detected
for the 5% CH4/H2 film. Rather, two peaks were observed at 1350 and 1589 cm"; peaks
that are characteristic of glassy carbonm'zz’zs‘zg’30 The peak at 1350 cm'1 is due to
disorder in the material (i.e., "D" band), while the peak at 1589 cm’1 is attributed to the
graphitic sp2 nondiamond carbon (i.e., "G" band) in the material.'8"9’2' Though this
spectrum closely resembles that for glassy carbon, Marcus et a]. noted that the presence
of these spectral features does not necessarily indicate the complete absence of

diamond.28 For example, the authors observed a Raman spectrum for a sample in which

68

the first-order diamond line was absent even though X-ray diffraction (XRD) showed
reflections characteristic of cubic diamond. The Raman scattering by nondiamond
carbon can mask the scattering intensity from diamond because the cross-sectional
scattering coefficient for the former (graphite as a model microstructure) is ~50 times
larger than that for diamondlg’29 Therefore, a small amount of nondiamond carbon near
the surface can mask the spectral response for diamond. Interestingly, the D/G ratio,
discussed in Chapter 1 for GC (Bedgc = 0.1-0.18), for the 5% CH4/H2 ratio film was ca. 1.3
indicating the nondiamond carbon is highly disordered with significant fraction of edge

plane exposed.

The carbon microstructure was also evaluated using powder X-ray diffraction
analysis (XRD). The XRD spectra were obtained using a Cu primary target (1.5407 A)
and scanning 2-theta from 20 to 100°. XRD spectra of the diamond films deposited from
the different CH4/H2 ratios, shown in Figure 3.4, can be compared to those of HOPG and
GC, model spz-bonded carbon materials, shown in Figure 3.5. Table 3.3 compares
measured peaks with standards from the American Society for Testing and Materials
(ASTM) to identify resulting peaks. Reflections for the diamond crystallite orientations
of (111), (220), and (311) are observed for each film and are marked along with the
Si(400) substrate peak, the graphite (002) (G(002)) peak, and a peak due to non-
orientated carbon (C), by the dotted lines, respectively. Note that the Si(400) peak is not
observed for the 0.5% CH4/H2 film because there is a signal at a 20 value of 28.54
corresponding to Si(111) (i.e., Si substrate has different orientation). The broadness

observed for the Si peak is likely due to Si-C formation at the beginning of the growth.

69

Perhaps the thickness of this layer influences the relative intensity and width of this peak
as both generally increase with increasing CH4/H2. One would assume that the Si signal
should decrease with increasing film thickness. Therefore it is unlikely that the film

thickness increases appreciably with increasing CH4/H2 ratio.

 

 

 

 

 

 

 

 

 

 

 

 

50" G(002) é: min) 51(400) 11(220) 0(311)

’ 5%
0 A :LQ
800

. 3%
, -1 L31 D 1 1
400 2%
,_ 4L 4

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1600 A A A 1‘
l I 1%»
500 ‘ ‘ G T ‘
. 0.5%
0 .1 _ AK
6001
. 0.3%
0
20 ‘ 30 ‘ 40 50 60 70 80 90 100

20 (degrees)

Figure 3.4. Powder X-ray diffraction spectra of diamond films deposited from different
CH4/H2 source gas mixtures.

70

 

Table 3.3. Comparison Between Measured Powder X-ray Diffraction Data and Standard
Values from ASTM 6-0675, 23-0064, and 46-0943

 

 

 

 

Measured Values ASTM
1.22... 8.1.1" “'81:?“ “'81:?“ 1:11:13
26.50 Graphite (002) 3.36 3.36 100
33.96 Carbon None 2.72 2.70 50
43.92 Diamond (1 1 1) 2.06 2.06 100
75.30 Diamond (220) 1.26 1.26 25
91.50 Diamond (311) 1.08 1.08 16

 

 

 

 

The intensity of the peak associated with the diamond (111) phase decreases and
the relative intensity of the peak for the (220) phase increases with increasing CH4/H2
ratio suggesting that the incorporation of the nondiamond impurity also changes the
crystalline orientation of the diamond. Table 3.4 shows the change in the full width at
half maximum (F WHM) as well as the intensity ratio between the two peaks associated
with the diamond (111) and (220) phases. The general increase in FWHM is consistent
with a decrease in diamond quality. Interestingly, films that exhibited considerable
nondiamond carbon impurity in the Raman spectra have a (111)/(220) intensity ratio
value of 8-9, suggesting a change in the nature of the diamond growth with increasing
CH4/H2 ratio. This value however decreased for the 5% film, presumably due to the large

graphite presence observed in both the Raman and XRD spectra.

71

 

Table 3.4. Full Width at Half Maximum (FWHM) and Peak Intensity Ratio for the
Diamond (111) and (220) Peaks

 

 

 

 

FWHM
CH4/H2 Ratio (degrees) 1012;121:331)
D(lll) D(220)
0.3 0.181 0.182 0.241
0.5 0.190 0.203 0.1 19
1 0.226 0.254 0.134
2 0.196 0.192 8.13
3 0.307 0.249 8.92
5 0.546 0.156 1.26

 

 

 

The powder XRD spectra for the different diamond films can be compared to the
spectra of two model spz-bonded carbon materials, HOPG and GC, which are shown in
Figure 3.5. The dotted lines are the same as those in Figure 3.4 and are inserted as
reference points. Only the 5% CW2 film possesses an appreciable amount of graphite
of the (002) orientation. Interestingly, despite not exhibiting a Raman signal
corresponding to diamond, the XRD spectrum does show some diamond signal for both
the (111) and (220) orientation. This is in agreement with the work of Marcus et al. and

discussed above.28

The signal at a 20 value of 33° becomes more prominent with increasing CH4/H2
ratio, suggesting that it is due to a nondiamond form of carbon. The corresponding d-
spacing value correlates well with that of non-orientated (possibly amorphous) carbon
(ASTM 46-0943). However, interestingly this signal was not observed for HOPG, GC,

or tetrahedral amorphous carbon (TAC, spectrum not shown because the only peak

72

observed was due to the Si substrate). Perhaps the signal is due to a mixed sp3/sp2 carbon

incorporation resulting in a signal located between that for graphite and diamond,

 

 

 

 

 

 

 

 

 

 

respectively.
600
CC
C(100)
i: G (ilo)
0 b JMMM
150000 A A I A AAAAAAAAAAAAA 4
HOPG
(”i0“) G(006‘l

l

0LLJL ------ Li-

20 30 40 50 60 70 80 90 100
20 (degrees)

Figure 3.5. Powder X-ray diffraction spectra for highly orientated pyrolytic graphite
(HOPG) and glassy carbon (GC). The dotted lines are the same as those in Figure 3.4,
and inserted as reference points.

3.2.2 Effect of CH4/H2 Ratio on the Electrochemical Properties

The diamond electrodes were mounted in a single compartment glass cell, as
described in Chapter 2 and exposed to potential cycling in N2-saturated 0.1 M HClO4
solution at a scan rate of 100 mV/s. Background CV current-potential (i-E) curves for
films deposited from different CH4/H2 ratios are shown in Figure 3.6. The potential
window clearly decreases and the background current magnitude increases with
increasing CH4/H2 ratio. The potential window, defined as the difference between the

anodic and cathodic potentials at which the current reaches i50 11A (or 1250 uA/cmz), is

73

presented in Table 3.5. The values range from 3.3 V for the 0.3% CH4/H2 film to 2.3 V

for the glassy carbon-like 5% CH4/H2 film.

 

 

 

 

 

 

 

 

 

 

5%
o ‘J
/ 0 ~
g g 2/0 I
v 0
id
5 0.5% J
h
s I .
U ! 03/0
,/
11011191
-1000 A A A 0 A A A 1000 A A A 2000 A

Potential (mV vs. Ag/AgCl)

Figure 3.6. Background cyclic voltammetric i-E curves in degassed 0.1 M HClO4 for
diamond films deposited fi'om different CH4/H2 source gas mixtures. Scan rate = 100
mV/s.

 

Table 3.5. CV Working Potential Window (i = i50 11A) for Films Deposited from
Different Methane-to-Hydrogen Source Gas Ratios

 

 

 

Methane-to-Hydrogen Ratio (%) Potential Window (V)
0.3 3.3
0.5 3.0
1 2.9
2 2.7
3 2.6
5 2.33

 

 

“ Anodic limit was only +200 pA/cm2

74

The voltammograms for the 0.3-3% films are largely featureless between -500
and 1000 mV, 3 potential region where surface redox processes associated with
electroactive carbon-oxygen functionalities occur.]’3’4’3 1'35 Although the Raman spectrum
for the 3% CH4/H2 film indicates a large presence of spz-bonded nondiamond carbon
impurity, the background voltammogram shows no evidence for any redox-active carbon-
oxygen functional groups that tend to populate the edge-plane sites of spz-bonded carbon.
This suggests that the nondiamond carbon contains a mixed spz/sp3 microstructure with a

limited number of sites where redox-active carbon-oxygen functionalities can exist.

One or more oxidation peaks can be seen prior to the onset current for oxygen
evolution, and these have been linked to the oxidation of spz-bonded carbon impurity at
the diamond surface.ll Again, this impurity is likely localized at the surface and consists
of a mixture of sp2- and sp3-carbon bonding. There is a general trend of increasing peak
charge (anodic) with increasing CH4/H2 ratio, consistent with the increased presence of

oxidizable nondiamond carbon impurity.

A significant change is seen for the 5% film, which exhibits a reduced potential
window and an enlarged background current. A reversible redox couple is observed near
400 mV (more visible in Figure 3.7), which is consistent with the presence of redox-
active carbon—oxygen functionalities at the nondiamond carbon impurity sites (e. g.,

quinone/hydroquinone type of couple)”36

Based on the Raman spectral data, the
nondiamond carbon impurity in this film is largely spz-bonded, much like the turbostratic

microstructure of glassy carbon. In fact, this CV greatly resembles one for freshly

75

polished glassy carbon. It would appear that the edge-plane carbon impurity sites are
highly susceptible to oxidation during exposure to the laboratory O2/H2O environment

and to these electrochemical polarization conditions.

Figure 3.7 compares the background voltammetric i-E curve for a 5% film as-
deposited with that for the same film following an acid-washing/rehydrogenation
procedure. This treatment involved cleaning the diamond surface in warm aqua-regia
(3:1 HCleNO3) and then in warm 30% H202 for 30 min each. The film was then
exposed to a H2 plasma for 30 min at a pressure of 35 torr and a power of 1 kW. This
procedure serves to clean the diamond of both inorganic and organic impurities and to

chemically etch away the surface spz-bonded nondiamond carbon impurity.

 

 
    

 

 

 

 

 

 

 

 

 

20
—— As-deposited
--- Rehydrogenated
,\ 10 '
<
:1
V
H
E
2 0 -
l...
5
U
-10 '
-20

-500 0 500 1000 1500
Potential (mV vs. Ag/AgCI)

Figure 3.7. Background cyclic voltammetric i-E curves in degassed 0.1 M HClO4 for 5%
CH4/H2 diamond film as—deposited and following an acid-washing/rehydrogenation
procedure. Scan rate = 100 mV/s.

76

There are two things worth noting here. One is the slight increase in background
current upon rehydrogenation. This is possibly due to the removal of the spz-bonded
carbon, mainly within the grain boundaries, resulting in a slightly rougher surface and
larger exposed surface area. This large background current also suggests that the large
nonfaradaic current observed in the background is not due solely to surface
functionalities produced by the nondiamond carbon. It is likely that the large double-
layer capacitance observed (Cu) is due to the n-bonding possessed by the spz-bonded
nondiamond carbon incorporated within the film microstructure. It is also worth
mentioning that it is also possible that the lattice hydrogen introduced via the H2 plasma
increases the surface conductivity resulting in an increase in electrode capacitance. The
second observation to note is the decrease in the signal of the redox couple near 400 mV.
This is consistent with the signal observed in the original i-E curve being due to carbon-
oxygen functionalities on the surface, which would be expected to decrease in

concentration after rehydrogenation.

SEM images were taken of the 5% CH4/H2 ratio film following the acid-
washing/rehydrogenation procedure and compared to the as-deposited film. These are
shown in Figure 3.8A and B. The large images were obtained at a magnification of 30K,
while the insets for both were obtained at a magnification of 70K. The as-deposited film
(A) is the same as that shown in Figure 3.2, just enlarged for comparison. There is no
significant difference observed between the two images. Therefore, it appears that an

increased surface area is not the reason for the increased background current.

77

‘11'

8‘: I
f

4.

 

Figure 3.8. SEM images (secondary electron) of the 5% CH4/H2 ratio diamond film (A)
as-deposited and (B) following acid-washing and rehydrogenation. Insets are high
resolution images of the center region of each figure.

78

The effect of the nondiamond impurity on the electrochemical response towards
several well known redox analytes was also investigated. Figure 3.9A-D shows cyclic
voltammetric i-E curves (scan rate = 100 mV/s) for several redox test analytes at films
deposited from different CH4/H2 ratios. A summary of the peak potential separations,
AEp, at a scan rate of 100 mV/s (n = 4 electrodes) is presented in Table 3.6. The values
are not iR corrected. There is a rationale for studying these particular redox systems, one
that originates largely from the work of McCreery and coworkers with glassy carbon.”40
Redox reactions generally fall under two classifications. One type includes reactions that
proceed by simple diffusion of the analyte to the electrode-solution interface, with the
electrode serving solely as a source or sink for electrons. In this case, the electrode-
reaction kinetics is relatively insensitive to factors such as the surface chemistry and
microstructure but is very sensitive to the density of electronic states near the formal
potential (so—called outer-sphere reaction). The other type includes reactions that occur
via some specific interaction with the electrode surface; for example, reactions that
proceed through an adsorbed state. In this case, the electrode-reaction kinetics is
sensitive to surface chemistry and microstructure as well as the density of electronic
states near the formal potential (so-called inner-sphere reaction). The redox systems,
Ru(NH3)63+/2+, methyl viologen, IrC162'/3 ', Ru(bpy)33"A/2+ and chlorpromazine, all undergo
outer-sphere electron transfer and the electrode kinetics are relatively insensitive to the
physicochemical properties of diamond. Apparent heterogeneous electron-transfer rate
constants, koapp, between 0.01 and 0.2 cm/s, are commonly observed for conducting
polycrystalline films (both microcrystalline or nanocrystalline) without extensive

3.4.6.7941

pretreatment. Normally, carbon and metal electrodes must be pretreated (e.g.,

79

polishing) to achieve such a large rate constant. Fe(CN)63'/4' proceeds through a more
inner-sphere electron—transfer pathway, and the electrode-reaction kinetics are highly
sensitive to the surface termination of diamond.“42 Presumably, the reaction proceeds
through a specific surface site that is blocked by surface oxygen, although this needs
further study. k0app values ranging from 0.01 to 0.1 cm/s are commonly observed for
clean, hydrogen-terminated films, but the rate constants decrease by over two orders of
magnitude for oxygen-terminated films.“'5 '41'43 For spz-carbon electrodes, the electrode
kinetics are strongly influenced by the fraction of exposed, clean edge plane.374° It
should be emphasized that Fe(CN)63'/4' is not a simple, outer-sphere redox system on
carbon (both diamond and glassy carbon) as researchers often supposes. Rather, it is a

surface sensitive redox system that is quite useful for probing the electronic properties,

surface cleanliness and surface chemistry of electrodes, particularly of diamond.

 

Table 3.6. Summary of the Cyclic Voltammetric Peak Potential Separations (AEp) for
Different Redox Analytes at Films Deposited from Different Methane-to-Hydrogen
Source Gas Ratios

 

 

 

 

CPL/Hz Ratio Peak Potential Separation, AEp (mV)

(%) F e(CN)63"" Ru(NH3)63+’2+ F e3+72+ 4-tBC
0.3 8016 90110 890160 7501150
0.5 69 1 8 76 1 8 780 1 50 500 1100
1 7011 7311 730160 5301100
2 7117 7313 760150 520130
3 6612 6812 690150 470150
5 6011 6213 120120 6819

 

Scan rate = 100 mV/s

 

80

 

 

_ 0.3 0/0 an..." 0.50/0 “0. 10/0

 

 

 

 

 

 

 

 

 

 

 

N
S

 

 

Current Density (uA/cmz)

vg‘

 

 

 

1.
8'

 

 

 

Figure 3.9. Cyclic voltammetric i-E curves at diamond films deposited fiom different
CH4/H2 source gas mixtures for 1 mM concentrations of (A) Fe(CN)63'/4', (B)
Ru(NH3)63+’2*, (C) Fe“2+ and (D) 4-tBC dissolved in either 1 M KCl (A and B) or 0.1 M
HClO4 (C and D). Scan rate = 100 mV/s.

Well-defined symmetric curves are observed for both Fe(CN)63"’4' and
Ru(NH3)63+/2+ (Figure 3.9A and B), with AEp values ranging from 60 to 90 mV,
depending on the CHs/H2 ratio. It is clear from the voltammetric i-E curves that AEp for
either redox system is not significantly affected by the presence of spz-bonded
nondiamond carbon impurity, regardless of the level or type. There is a general trend of
decreasing AEp with increasing CH4/H2 ratio; however, only over a relatively small range

of ca. 20-30 mV. The low AEp (theoretical value of 59/n mV) for some films reflects

81

relatively rapid electrode-reaction kinetics. The fact that the kinetics are rapid, even
though the electrodes were exposed to the laboratory atmosphere for an extended period
of time and no conventional pretreatment was applied prior to use, reflects the material's
activity and resistance to fouling. The electrodes were only soaked in clean IPA for 10
min to clean the surface prior to use. There is some random variation in Ep/2 for
F e(CN)63'/ 4” but nothing that follows a predictable trend. The peak currents, if" and ip'“,

are nearly equal for all the electrodes, with an ip°"/ipred

ratio near unity for both analytes.
The peak currents for the 5% CH4/H2 film do appear slightly larger than the others, but
this is due to the larger background current, as the curves shown were not background

corrected. When background corrected, the ip°" and ipred values are all nominally the

same.

The AEp for Fey/2+, shown in Figure 3.9C, decreases from 890 to 690 mV as the
CH4/H2 ratio increases from 0.3 to 3%. There is, however, a more marked decrease for
the 5% film, as AEp drops to 120 mV. Clearly, this redox reaction undergoes electron
transfer that is site dependent and faster at a more spz-bonded microstructure; presumably
sites with more surface oxygen. The question is, how does this microstructure influence

”2+ is of the inner-sphere type, with k0app being very sensitive to the

the reaction rate? Fe
presence of surface carbon-oxygen functionalities, specifically carbonyl types, for both
glassy carbon and diamond.4"42’44 Carbonyl functional groups have been shown to
catalyze this redox reaction, thereby improving the electrode kinetics.44 We suppose the

nondiamond carbon impurity, particularly that with more sp2 bonding character, contains

dangling bonds at the surface that react with 02 and/or H2O in the air or solution. Based

82

on the great reduction in AEp for the 5% CH4/H2 film, it appears that the mixed spz/sp3
carbon microstructure up to the 3% level has a limited number of oxidizable sites so that
the oxygen incorporation is relatively low. However, the glassy carbon-like
microstructure of the nondiamond carbon in the 5% film has a much higher number of
oxidizable sites. This surface is the most extensively oxidized and probably contains
carbonyl functionalities. Therefore, we suppose that the nondiamond carbon impurity
influences this particular redox system by providing sp2 sites where carbon-oxygen
functional groups can readily form, specifically catalyzing carbonyls.“4 The catalysis
may involve lowering of the reorganization energy of the H20 solvation sphere

- /
surrounding the Fe3+ 2+

couple. Interestingly, acid-washing and rehydrogenating one of
the 5% films did not significantly alter the electrode response towards this analyte. The
AEp was 136 mV at a scan rate of 100 mV/s. While this could be due to incomplete
removal of the oxygen functional groups or quick oxidation of the oxygen-free surface
upon air exposure, it is possible that the remaining n-rr interactions are enough to
catalyze the reaction through promoting molecular adsorption. All these would be

consistent with the narrow potential window observed in the background voltammetric i-

E curve for this film.

Figure 3.9D shows cyclic voltammetric i—E curves for 4-tBC at the various films.
Clearly, nondiamond carbon impurity significantly influences the electrode response for
this analyte as there is a progressive decrease in AEp from 750 to 70 mV with increasing
CH4/H2 ratio. The more complicated organic redox analytes, dopamine, 4-

methylcatechol, and 4-tBC, are inner-sphere systems with k0app values of 10“1 to 10'6

83

 

cm/s, common for hydrogen-terminated diamond.3’4’6’7‘9 It is postulated that the slow
kinetics for these polar aromatic analytes result because of weak adsorption on the
diamond surface. It is possible that n-n interactions, through the extended electronic
conjugation, are perhaps more influential than surface oxides in promoting molecular
adsorption for these analytes. This conclusion comes, in part, from our past work with
low-oxygen, hydrogenated glassy carbon.“46 We suppose that the spz-bonded carbon
impurity provides adsorption sites for 4-tBC, with the adsorption lowering the activation
barrier for charge transfer. The nondiamond carbon impurity content grows with
increasing CHx/H2, as determined by Raman spectroscopy and this leads to an increased
number of adsorption sites. The greater peak current seen for the 5% CH4/H2 film, as
compared to the others, is consistent with increased adsorption. Interestingly, the AEp
observed for a rehydrogenated 5% film only increased slightly compared to an as-
deposited film. The AEp was forlnd to be 94 mV at a scan rate of 100 mV/s for the
rehydrogenated film compared to ca. 70 mV for the as-deposited film. This again could
be attributed to incomplete removal of the oxygen functional groups, quick oxidation of
the oxygen-free surface or, if the oxide content is minimal, significant influence from the

n-lr interactions in the film which promote molecular adsorption of the analyte.

3.2.3 Effect of CH4/H2 Ratio on the Oxygen Reduction Reaction (ORR)

The ORR in 0.1 M HClO4 (Figure 3.10 and Table 3.7) and 0.1 M NaOH (Figure
3.11 and Table 3.8) for select CH4/H2 ratios was also used to probe the electrochemical
effects of the incorporated nondiamond carbon impurity. The carbon surface chemistry

and microstructure are known to play key roles in the electrode kinetics for this surface-

84

47-51

sensitive redox reaction. At carbon electrodes, the reduction reaction product (2e'

/2H+) is usually H202 or HO2', depending on the pH, and the kinetics are generally much

44.50.51

slower than at metals. These products are formed via reactions 3.2.1 and 3.2.2

below, respectively.50

02 + 2H+ + 26- :2 H202

(3.2.1)
13° (pH 0) = 498 mV vs. Ag/AgCl (3 M NaCl)50
02 + H20 + 2e' :2 110‘2 + 011'
(3.2.2)
13° (pH 13) = -262 mV vs. Ag/AgCl (3 M NaCl)50
52,53

Diamond is known to possess a large kinetic overpotential for this reaction.

One possible cause for this is weak adsorption of the reactant, 02, or the primary reaction
intermediate, 05— (super oxide), on the hydrogen-terminated surface. Yang and
McCreery found that when there are available adsorption sites (i.e., clean GC), the

electrogenerated 03— adsorbs to the electrode surface. However, when these adsorption

sites are blocked or unavailable, the 05’ species cannot adsorb and remains in aqueous

solution.47 The mechanistic aspects of this reaction on diamond are not fully understood
at present. The linear sweep voltammetric (LSV) i-E curves shown in Figure 3.10 reveal
a positive shift in the onset potential for the reduction current with increasing CH4/H2.

The maximum current increases with CH4/H2 from 0.3 to 2% but occurs within the same

85

 

potential region from -600 to -800 mV. For example, the maximum current increases
from ca. -60 uA/cm2 for the 0.3% CH4/H2 film to ca. -137 )uA/cm2 for the 2% film. The
most positive onset potential is seen for the 3 and 5% CH4/H2 films. The current for both
is approximately the same, -90 uA/cmz, and is lower than that for the 2% film. The most
positive onset and peak potentials are seen for the 5% film-the "glassy carbon-like"
electrode. The positive-shifting onset potential indicates that the nondiamond carbon
impurity catalyzes this reaction in some manner. The sp2 carbon sites involved in this
reaction are clearly more active for ORR than sites on the diamond surface. The number
of active sites increases proportionally with CH4/H2. Interestingly, the change in
microstructure for the 2 and 3% CH4/H2 films, as seen in the Raman spectroscopic data,
does not cause much of a shift in the reduction peak position but does lead to a sizeable

decrease in the peak current. The reason for the current decrease is unknown.

One possibility is that the reaction mechanism could be changing, depending on
the type and fraction of nondiamond carbon exposed. For example, the reaction might be
proceeding via the 46’ pathway to water for the 0.3-2% CH4/H2 films, given the large
cathodic overpotential. In this case, the current increase observed might simply be due to
an increase in the number of available reaction sites. The microstructure changes,
particularly for the 3 and 5% CH4/H2 films, as seen in the Raman data. Even though the
relative fraction of nondiamond carbon increases with CHs/I-I2, this change in
microstructure might lead to a reduction in the number of active sites. A change in the
reaction mechanism from a 4e’ (higher 11) to a 2e‘ (lower 1]) pathway is also possible.

This is pure speculation at this point, because we did not test for the presence of H202.

86

One final point that should be made is that for all the films, particularly those with more
nondiamond carbon impurity, the kinetically controlled current (i. e., rising portion of the
voltammogram) increases gradually rather than sharply over a wide potential range. We
suppose that this gradual slope reflects the heterogeneity of the composite electrode
architecture, as regions of higher activity at the nondiamond carbon sites are isolated

from regions of lesser activity at the diamond sites.

 

 

('11
c

596

-100

 

 

-150

Current Density (uA/cmz)

 

l

I

 

-200 4 1 1 l 1 1 1 1 1 1 l 1 1 1 1 1 1
-800 -600 -400 -200 0 200

Potential (mV vs. Ag/AgCl)

Figure 3.10. Linear sweep voltammetric i-E curves for the oxygen reduction reaction
(OR) in 0.1 M HClO4 at diamond films deposited from different CH4/H2 source gas
mixtures. Scan rate = 1 mV/s.

87

 

Table 3.7. Data for ORR in 0.1 M HClO4 for Films Deposited fi'om Different Methane-
to-Hydrogen Source Gas Mixtures

 

 

 

(31:49:: Reduction Onset Potential Tafel Slope an jo 2
(%) (E (V) at -25 uA/cm ) (V/dec) (A/cm )
0.3 -0.63 -010 0.60 4.15 x 10"°
0.5 -0.60 -013 0.46 2.43 x 10‘13

1 -0.58 -014 0.44 1.31 x 10'12
2 -053 -019 0.31 1.80 x 10"°
3 -0.48 -0.26 0.23 6.76 x 10'9
5 -0.01 -0.28 0.28 7.23 x 10'8

 

 

 

Table 3.7 presents a summary of the Tafel analysis data for the ORR in 0.1 M

HClO4. Tafel analysis can be performed for irreversible reactions at large overpotentials
(n > 118 mV) (i. e., slow reaction kinetics) and in the absence of mass transfer effects. A
plot of 11 vs. log i is known as a Tafel plot where the slope equals -2.3RT/naF and the y-
intercept equals (2.3RT/naF) log i0. The Tafel slope provides information about the
mechanism of the reaction as an can be found, while the y-intercept provides information

about the rate of the reaction as the exchange current or "idling" current can be found.

A trend of a more positive onset potential with increasing CH4/H2 is evident in the
data. The onset potentials for the 0.3-2% CH4/H2 films are within a 100 mV range (-0.63
to -0.53 V) of each other. The most positive shifts are seen for the 3 and 5% films.
Clearly, the sp2 microstructure present in the 3 and 5% films is the most active. The

Tafel slopes, uncorrected for mass-transport effects, generally increase with increasing

88

CH4/H2. Good linearity was observed over at least one order of magnitude in current,
with least-squares correlation coefficients of 0.985, or better. The Tafel slopes for the
0.3, 0.5 and 1% CH4/H2 films are approximately the same with values ranging from —0.10
to -0.14 V/dec. These values are close to the -0.12 V/dec value expected for n = 1 and a
transfer coefficient, a, of 0.5. The slopes for the 2, 3, and 5% CH4/H2 films increase with
values of -0.19, -O.26, and -0.22 V/dec, respectively. The higher slopes reflect a decrease
in the an product. The calculated an values are shown in the table. There is a
progressive decrease from 0.60 to 0.28 with increasing CH4/H2. The transfer coefficient
decreases with increasing CH4/H2, if n = 1 is assumed. Fujishima et a]. previously
reported large Tafel slopes and small a values for the OR at diamond electrodes.”53
The transfer coefficients obtained for the 3 and 5% CH4/H2 films are similar to those
reported by Yang and McCreery for polished glassy carbon (0.29) at a pH below 9.47
Because these Tafel slopes were not corrected for mass-transfer effects, we cannot rule
out the possibility that the values are influenced from this to some extent. We do not
think that the electrode-reaction kinetics are influenced by the electronic properties of the
electrode (i. e., potential-dependent density of electronic states), because prior work in our
group has revealed that the diamond electrodes, boron-doped at the 1020 to 102‘ cm'3
level, behave as a semimetal over a wide potential range.4'6‘7 Outer-sphere redox systems

in this same potential region (e.g., Ru(NH3)(53”"2+

and methyl viologen) have relatively
large heterogeneous electron-transfer rate constants, 0.05-0.2 cm/s, and a values near 0.5.
We suppose the increasing Tafel slopes (decreasing a values) are caused by

heterogeneity in the electrode reactivity (variable reaction sites). The OR is a surface-

sensitive reaction and the kinetic overpotential is dependent on the sites available.

89

Multiple reaction sites, with different activation energies, are probably present, which
vary with CH4/H2. The exchange current density, 15, increases considerably with CH4/H2,
as the values range from 10'16 to 10'8 A/cm2 (geometric area) for the 0.3-5% CH4/H2

films.

The OR reaction kinetics at carbon tend to be faster in alkaline than in acidic
media (less 11 as can be seen by comparing the onset potentials with the standard
reduction potentials for Reactions 3.2.1 and 3.2.2). Figure 3.11 shows linear sweep
voltammetric i-E curves for the OR in 0.1 M NaOH. The onset potential shifts positive
and the maximum cathodic current increases with increasing CH4/H2. Again, these shifts
occur presumably because of a larger fraction of exposed low-energy reaction sites,
which are associated with the spZ-bonded carbon. The response for the 5% CH4/H2 film,
in terms of the peak current and potential, is nearly identical to that in 0.1 M HClO4, even
though the solution pH is quite different. The absence of a significant pH effect for this
reaction at diamond has also been reported by Fujishima et al.52’53 This may mean that
the rate-determining step is the le’ reduction of O2 to form superoxide, a pH-independent

reaction shown below in equation 3.2.3.

02 +e_ 3;) O52.-(ads) (3-2-3)

This electrochemical reaction could then be followed by subsequent chemical and

electrochemical reaction such as 3.2.4 and 3.2.5 below.52'53

90

o§‘(,d,) + H20 a H0§(,d,) + OH" (3.2.4)

Ho:(,ds) +e‘ : Hogm) (3.2.5)

 

 

    

 

 

 

0
51‘
E
i
v -30
a?
m
:1
5
1.. -60
1::
3 5°/
h 0
:1
U
-90 1 1 1 L 1 1 1 l 1 1 1 I 1 1 1
-800 -600 -400 -200 0

Potential (mV vs. Ag/AgCI)

Figure 3.11. Linear sweep voltammetric i-E curves for the oxygen reduction reaction
(ORR) in 0.1 M NaOH at diamond films deposited from different CH4/H2 source gas
mixtures. Scan rate = l mV/s.

Table 3.8 presents a summary of the Tafel analysis data for ORR in 0.1 M NaOH.
The trend of more positive onset potential with increasing CH4/H2 is evident. The Tafel
slopes range from -0.064 to -0.11 V/dec, but there is no clear trend with CH4/H2. These
slopes are certainly in the range of those reported for other types of carbon electrodes in

alkaline media.47‘49‘5 ' In contrast with the data for HClO4, the lowest slope in NaOH is

91

seen for the 5% CH4/H2 film. The an values range from 0.55 to 0.70, except for the 5%
film for which the value increases to 0.93. This value is similar to that reported for clean
glassy carbon in alkaline media.47 Importantly, the exchange current density, 1;), is
several orders of magnitude larger in alkaline than in acidic media. The exchange current

density progressively increases from 10'7 to 10'5 A/cm2 with an increase in CH4/H2 from

0.3 to 5%.

 

Table 3.8. Data for OR in 0.1 M NaOH for Films Deposited from Different Methane-
to-Hydrogen Source Gas Mixtures

 

 

(31:14:32 Reduction Onset Potential Tafel Slope an jo 2
(%) (E (V) at -25 uA/cm ) (V/dec) (A/cm )
0.3 -073 -0.086 0.69 4.77 x 10‘7

2 047 -011 0.54 8.86 x 10'7
3 -043 -0.086 0.69 1.26 x 10‘
5 .024 -0.064 0.93 8.05 x 10'5

 

 

 

 

Rigorous kinetic and mechanistic analysis of the OR is impossible from the data
at hand. Further work is needed to understand this complicated redox reaction on
diamond. However, there are several points to be made regarding the role of nondiamond
carbon impurity on the reaction.

1. There appears to be a direct correlation between the amount and type of
nondiamond carbon impurity present at the surface and the most positive onset potential
and largest exchange current density for the reaction. The nondiamond carbon

microstructure, s ecificall , the " lass carbon-like" structure, is the most active for
P y g y

92

ORR. Presumably, the diamond surface exhibits sluggish electrode kinetics for this

reaction because of weak adsorption of 02 or the primary reaction intermediate 03—

2. There appears to be a minimal pH effect on the OR at the most active

electrode, the 5% film. This may suggest that the rate-determining step for the reaction,

in both media, is the 1e" reduction to superoxide, 05'.

3. Large Tafel slopes are observed for all the diamond electrodes in acidic media,
while the slopes in alkaline media are more similar to those reported for other types of
carbon. The large Tafel slopes, and correspondingly small values of an, at least in acid,
may reflect a very heterogeneous electrode surface possessing multiple reaction sites with

different activities.

3.2.4 Probing the Oxygen Reduction Reaction Mechanism on Carbon

In an attempt to further investigate the OR mechanism on the thin-film
electrodes, the reaction was investigated voltammetrically in both 0.1 M H001 and 0.1
M NaOH at scan rates ranging from 1 mV/s to 10 V/s, and amperometrically. All data
were background corrected. Unfortunately, only the 5% film was able to be investigated
as the OR overpotential was too large on the other films and no peak was able to be

observed at scan rates above 1 mV/s despite background subtracting.

Investigating the ORR by increasing the potential scan rate allows one to assess
several parameters. As expected for an irreversible reaction, the reduction peak potential,
Ep, shifted negative in a linear fashion with the natural log of the square root of the scan

rate, In v10, as shown by equation 3.2.6 below.

93

 

 

1/2 1/2
13p: 13° — RT 0.780+ln D0 +m(EF—") (3.2.6)
anaF k0 RT

Assuming the reaction is irreversible, the peak current, ip, can be shown to be

proportional to the square root of the scan rate, V”, as shown in equation 3.2.7:54
1,, = 2.99 x 105n(ana)1/2AD01/ 2C(,"‘v“2 (3.2.7)

where n is the stoichiometric number of electrons involved in the reaction, 01 is the
transfer coefficient, n, is the number of electrons in the rate-determining step, A is the
electrode area (assumed here to be 0.2 cm2, the geometric area), Do is the diffusion
coefficient (1.82 and 2.0 x 10'5 cmz/s m 0.1 M 11004 and 0.1 M NaOH at 25 °C,
respectively),55’56 and Co. is the bulk oxygen concentration (1.3 and 1.15 x 1045 mol/cm3
in 0.1 M HC104 and 0.1 M NaOH at 25 °C, respectively).55’56 Figure 3.12 shows a plot of
1', vs. v“2 for the OR at the 5% CHa/I-I2 film in both acidic and alkaline media. Both
plots are relatively linear over the entire range (R2 = 0.969). Solving equation 3.2.7 to
find n(oln,,)1/2 results in values of 0.419 and 0.416 for the acidic and alkaline media,

respectively.

94

 

 

 

 

50
[ 0 Acid
-50 I Alkaline
'4“
. "\F
-150 _ ' 11‘ y = 415.01: - 50.367
A ‘ 3‘. f 2, R2 = 0.9685
~ . \
3 -250 - - ‘~.‘\
3. ’ ~ . \ x
N A . \ x
s. I x
'350 _ . ~ ~ \ \
y=-139.06x-50.572 - \\\.
2 = ‘ ~ ~
450 _ R 0.9686 . . ~ . .
-550 1 1 n 1 P 1 1 1 1 i 1 L 1 1 l r 1 4
0 1 2 3 4

Figure 3.12. Plot of ip vs. v”2 for OR in acidic and alkaline media at 5% CH4/H2 film.

For an irreversible reaction, the peak current is also known to depend on the

standard heterogeneous rate constant, k0, as shown in equation 3.2.8.
, 4- an F 1
1p = 0.227nFACo koexp[—(fi—)(Ep — E0 ):| (3.2.8)

This equation can then be rearranged to arrive at equation 3.2.9 below.

ln 1,, = 1n 0.227nFAC0"k0 —(%T£)(Ep - 13°) (3.2.9)

Therefore plotting 1n ip vs. (Ep - Ed) results in a linear trend with a slope equal to —
(anaF/RT) and a y-intercept proportional to the rate constant k0. Assuming the activity

coefficient of the two solutions is unity, the E° values given in equations 3.2.1 and 3.2.2

95

 

can be used in place of Ed, respectively. The oma value can then be calculated using the
slope and substituted into the n(ana)”2 solution from equation 3.2.7 to find n.

Additionally, using the y-intercept, and the calculated n value, k0 can also be found. All

of these quantities are listed in Table 3.9.

 

Table 3.9. ORR Parameters Obtained from Linear Sweep Voltammetric i-E Curves in
Acidic and Alkaline Media at a 5% CH4/H2 Film

 

 

Parameter 0.1 M HC104 0.1 M N 3011
an, 0.24 0.33
n 0.85 0.73
k° (cm/s) 1.44 x 10" 2.64 x 10'3

 

 

 

 

The experimental 71 value determined for each solution is nearly 1 suggesting that
l-electron is transferred in the irreversible step in the ORR process. The question is why
a value of l is observed if the literature indicates that n = 2 on carbon surfaces? One

possibility is that the reaction does not go to completion under these conditions. Perhaps

only the initial l-electron reduction to OE" occurs. Another possibility is a surface area

less than the assumed area of 0.2 cm2 participates in the reaction. For example, an active
surface area of ca. 0.05 cm2 (or 25% of the geometric area) would yield an an, value of
~0.5 assuming n = 2. This does not appear likely since n values calculated using
chronoamperometric i—t curves (discussed below) was found to be ~2 using an assumed

area of 0.2 cm2.

A final source of error is that theory does not adequately describe the ORR

process. Equation 3.2.7 is intended for an irreversible l-electron transfer under semi-

96

infinite linear diffusion limitations. It is obvious that this reaction is much more
complicated and perhaps equations 3.2.7 — 3.2.9 oversimplify the situation. According to
the proposed mechanism above, equation 3.2.3 is the only equation limited by semi-
infmite linear diffusion. Equations 3.2.4 and 3.2.5 both assume an adsorbed species.
Since only the initial l-electron transfer is limited by semi-infmite linear diffusion, an 71
value of 1 is perhaps reasonable if the reaction proceeds via reaction 3.2.3. Obviously it
is difficult to fit a mixed diffusion/adsorption reaction to a theoretical situation of only

semi-infinite linear diffusion.

Additionally, if one assumes n, = 1, then or for the acidic and alkaline media is
0.24 and 0.33, respectively. Since these equations are intended for an oxidation reaction,
the value (1 — a) would be more appropriate in this instance. This value for the two
media would then be 0.76 and 0.67, respectively. The 01 in alkaline agrees well with that
found by Yang and McCreery for polished GC at a similar pH.47 The rate constant in the
alkaline solution agrees well with that previously determined in solutions of similar
pH.‘“"52 The reaction rate is several orders slower in the acid solution compared to that of

53

the alkaline solution in agreement with previous reports. Yang and McCreery also

noted a decrease in k0 with decreasing pH. They concluded that at high pH, the rate-

deterrnining step (rds) in the reaction was equation 3.2.4. However, at pH < 9, the

adsorbed 03— protonates to form a relatively stable hydroperoxide analog of ROOH.

This species appears to occupy adsorption sites, thus preventing further 05" adsorption.

97

This means that reaction 3.2.3 becomes the rds and proceeds via an outer sphere route.47

The inhibition for 03— to adsorb to the surface results in a decrease in rate constant.

Alternatively, the total number of electrons transferred in the OR reaction can be
determined using a large amplitude potential step technique and fitting the resulting i—t

profile to the well known Cottrell equation shown in equation 3.2.10 below.

i(t)_ 11:11),“ch
nl/Ztl/Z

 

(3.2.10)

If the reaction rate is limited by semi-infinite linear diffusion, plotting 1' vs. t’“2 results in
a curve with a linear region whose slope is proportional to n, and a y-intercept ideally of
zero. Using this technique, the value of n in both the acidic and alkaline media was
found to be 1.8, and 2.2, respectively. These values were found assuming A is the
geometric area and are close to the predicted value of 2. This suggests that
overestimating the surface area was not the cause of n = 1 determined using voltammetric
sweeps. These results are consistent with the reaction mechanism proposed in equations
3.2.3 — 3.2.5, with an overall 2e" reduction proceeding through an initial irreversible 1e"

step independent of pH.
Exposing the 5% CHa/H2 film to the cleaning and rehydrogenating procedure did

not significantly affect most of these quantities. The quantities determined from linear

sweep voltammetric i-E curves in both media after acid-washing and rehydrogenation are

98

shown in Table 3.10. Cleaning the spz-bonded nondiamond carbon off the surface did
not alter any of the values in the table to a great extent. This means that the nondiamond
carbon removed from the surface does not affect the reaction in either solution. This
possibly suggests that the n—n interactions likely promote O2 adsorption and not the
surface nondiamond carbon. This is similar to what was found with the 4-tBC redox
analyte above. It should be mentioned however, that it is possible that the procedure
changes the surface termination but not the bonded carbon. Therefore one can only say
that the surface oxygen is definitely removed while possibly some or all of the sp2 carbon

remains.

 

Table 3.10. ORR Parameters Obtained from Linear Sweep Voltammetric i-E Curves in
Acidic and Alkaline Media at a 5% CHa/H2 Film After Acid-Washing and
Rehydrogenating

 

 

 

Parameter 0.1 M HCIO4 0.1 M NaOH
on, 0.22 0.42
n 1.1 0.75
k° (cm/s) 2.93 x 10" 1.26 x 10'3

 

 

 

3.3 Conclusions

The physical and electrochemical properties of boron-doped polycrystalline
diamond thin-film electrodes, prepared with varying levels of spz-bonded nondiamond
carbon impurity, were systematically investigated. This impurity was introduced through
adjustment of the CH4/H2 source gas ratio used during the deposition. Variations in the
morphology and microstructure were monitored using AFM, SEM and Raman

spectroscopy. Proportional increases in the fraction of grain boundary, the extent of

99

 

secondary nucleation and the spZ-bonded carbon impurity content occurred with
increasing CHa/H2 ratio. Additionally, the increased rate of secondary nucleation

resulted in progressively smoother films with increasing CHa/H2 ratio.

The electrode response was assessed using Fe(CN)63"/4’, Ru(NH3)63+/2A, Fe3+/2+,
and 4-tert-butycatechol (4-tBC). While the increasing spZ-bonded carbon content had
little effect on the CV peak separation (AEp) or peak current for the first two redox
systems, the impurity had a significant impact on the latter two, as AEp decreased
proportionally with the increased spZ-bonded carbon content. Specifically, the 5%
CHa/H2 film exhibited a marked improvement in the electrochemical response towards
both Few2+ and 4-tBC, as compared to the other films. The AEp for this film towards
these analytes was severa_l hundred mV less than the AEp for the other films towards these
redox couples. It is obvious that the spz-bonded nondiamond carbon in this film is quite
different from the nondiamond carbon present in the other films, consistent with the
Raman spectral data. In fact, supposable removal of the surface nondiamond carbon
from the 5% CHa/H2 film did not significantly alter the electrochemical response of the

electrode towards either F e3 H 2+

or 4-tBC. This means that the electrochemical response
is a constituent of the microstructure (sp3/sp2 carbon mixture) and not just a surface
impurity. It is also possible that the hydrogen plasma either failed to completely remove
the nondiamond carbon or the oxygen-free surface completely oxidized upon air
exposure. It is clear that the nondiamond carbon affects the reaction kinetics, likely by

providing sites for strong adsorption and sites where catalyzing carbon-oxygen functional

groups can form.

100

The effect of the nondiamond carbon impurity on the OR in 0.1 M HClO4 and
0.1 M NaOH was also investigated. A direct correlation was found between the relative
amount of the impurity, as estimated from Raman spectroscopy, and the overpotential for
the reaction. The greater the nondiamond content, the lower the kinetic overpotential for
the reduction reaction and the higher the exchange current density were. The composite
electrodes exhibited greater activity in alkaline solutions than in acid. The microstructure
of the nondiamond carbon appears to influence the reaction, particularly in acid, as
removing nondiamond carbon from the surface of the 5% CH4/H2 film did not
significantly alter the OR mechanism. Tafel plots yielded an apparent exchange current
density that increased and a transfer coefficient that decreased with the increased
nondiamond carbon content. The results indicate that the grain boundaries, and the sp2
carbon impurity presumably residing there, can have a profound impact on the electrode-
reaction kinetics for certain redox systems, while for others this impurity has little

influence on the diamond electrode response.

101

10.

ll.

12.

13.

REFERENCES

J. Xu, M. C. Granger, Q. Chen, J. W. Strojek, T. E. Lister, and G. M. Swain, Anal.
Chem, 69, 591A (1997).

G. M. Swain, A. B. Anderson, and J. C. Angus, WSBulL 23, 56 (1998).

M. C. Granger, J. Xu, J. W. Strojek, and G. M. Swain, Anal. Chim. Acta, 397, 145
(1999)

M. C. Granger, M. Witek, J. Xu, J. Wang, M. Hupert, A. Hanks, M. D. Koppang,
J. E. Butler, G. Lucazeau, M. Mermoux, J. W. Strojek, and G. M. Swain, Anal.
Chem, 72, 3793 (2000).

A. E. Fischer, Y. Show, and G. M. Swain, Anal. Chem, 76, 2553 (2004).

Y. Show, M. A. Witek, P. Sonthalia, and G. M. Swain, Chem. Mater., 15, 879
(2003)

M. Hupert, A. Muck, J. Wang, J. Stotter, Z. Cvackova, S. Haymond, Y. Show,
and G. M. Swain, Diamond and Re]. Mater., 12, 1940 (2003).

H. Gerischer, Electrochim. Acta, 35, 1677 (1990).

J. Wang, G. M. Swain, M. Mermoux, G. Lucazeau, J. Zak, and J. W. Strojek, New
Diamond Front. Carbon T echno]., 9, 317 (1999).

T. N. Rao, T. A. lvandini, C. Terashima, B. V. Sarada, and A. Fujishima, New
Diamond Front. Carbon T echno]., 13, 79 (2003).

H. B. Martin, A. Argoitia, U. Landau, A. B. Anderson, and J. C. Angus, J.
Electrochem. Soc., 143, L133 (1996).

J. Wang and G. M. Swain, J. Electrochem. Soc., 150, E24 (2003).

Y. V. Pleskov, Russ. Chem. Rev., 68, 381 (1999).

102

14.

15.

l6.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

L.-F. Li, D. A. Totir, N. Vinokur, B. Miller, G. Chottiner, E. A. Evans, J. C.
Angus, and D. A. Scherson, J. Electrochem. Soc., 145, L85 (1998).

1. Duo, A. F ujishima, and C. Comninellis, Electrochem. Commun, 5, 695 (2003).

K. Yoo, B. Miller, K. Rafi, and X. Shi, Electrochem. Solid-State Lett., 2, 233
(1999)

M. Mermoux, B. Marcus, G. M. Swain, and J. E. Butler, J. Phys. Chem. B, 106,
10816 (2002).

W. A. Yarbrough and R. Messier, Science, 247, 688 (1990).

D. S. Knight and W. B. White, J. Mater. Res, 4, 385 (1989).

R. E. Shroder, R. J. Nemanich, and J. T. Glass, Phys. Rev. B, 41, 3738 (1990).

M. Silveira, M. Becucci, E. Castellucci, F. Polla Mattiot, V. Barbarossa, R.
Tomaciello, and F. Galluzzi, Diamond and Re]. Mater., 2, 1257 (1993).

K. G. Ray 111 and R. L. McCreery, J. Electroanal. Chem, 469, 150 (1999).

K. Ushizawa, K. Watanabe, T. Ando, I. Sakagucki, M. Nishitani-Gamo, Y. Sato,
and H. Kanada, Diamond and Re]. Mater., 7, 1719 (1998).

J. R. Dennison, M. Holta, and G. M. Swain, Spectroscopy (Amsterdam), 11, 38
(1996).

M. C. Rossi, Appl. Phys. Lett., 73, 1203 (1998).

C. S. Feigerle and R. W. Shaw, Opt. Photonics News, 7, 24 (1996).

L. H. Robins, E. N. Farabaugh, and A. F eldman, J. Mater. Res, 5, 2456 (1990).

B. Marcus, L. Fayette, M. Mermoux, L. Abello, and G. Lucazeau, J. Appl. Phys,
76, 3463 (1994).

103

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

43.

N. Wada, P. J. Gaczi, and S. A. Solin, J. Non-Cryst. Solids, 35 & 36, 543 (1980).
R. L. McCreery and R. T. Packard, Anal. Chem, 61, 775A (1989).

R. L. McCreery, in Neuromethods, A. Boulton, G. Baker, and R. N. Adams,
Editors, 27, p. 1, Humana Press, Inc., Totowa, NJ, (1995).

I.-F. Hu, D. Karweik, and T. Kuwana, J. Electroanal. Chem. Interfacial
Electrochem., 188, 59 (1985).

J. F. Evans and T. Kuwana, Anal. Chem, 49, 1632 (1977).

J. F. Evans and T. Kuwana, Anal. Chem, 51, 358 (1979).

C. W. Miller, D. H. Karwelk, and T. Kuwana, Anal. Chem, 53, 2319 (1981).
D. T. Fagan, I. F. Hu, and T. Kuwana, Anal. Chem, 57, 2759 (1985).

R. J. Rice, N. M. Pontikos, and R. L. McCreery, J. Am. Chem. Soc., 112, 4617
(1990)

K. R. Kneten and R. L. McCreery, Anal. Chem, 64, 2518 (1992).

P. Chen, M. A. Fryling, and R. L. McCreery, Anal. Chem, 67, 3115 (1995).
P. Chen and R. L. McCreery, Anal. Chem, 68, 3958 (1996).

H.-H. Yang and R. L. McCreery, Anal. Chem, 71, 4081 (1999).

S. H. DuVall and R. L. McCreery, J. Am. Chem. Soc., 122, 6759 (2000).
M. C. Granger and G. M. Swain, J. Electrochem. Soc., 146, 4551 (1999).

C. A. McDermott, K. R. Kneten, and R. L. McCreery, J. Electrochem. Soc., 140,
2593 (1993).

104

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

56.

Q. Chen and G. M. Swain, Langmuir, 14, 7017 (1998).

J. Xu, Q. Chen, and G. M. Swain, Anal. Chem, 70, 3146 (1998).

H.-H. Yang and R. L. McCreery, J. Electrochem. Soc., 147, 3420 (2000).

J. Xu, W. Huang, and R. L. McCreery, J. Electroanal. Chem, 410, 235 (1996).
I. Morcos and E. Yeager, Electrochim. Acta, 15, 953 (1970).

E. Yeager, Electrochim. Acta, 29, 1527 (1984).

A. J. Appleby and J. Marie, Electrochim. Acta, 24, 195 (1979).

T. Yano, D. A. Tryk, K. Hashimoto, and A. Fujishima, J. Electrochem. Soc., 145,
1870 (1998).

T. Yano, E. Popa, D. A. Tryk, K. Hashimoto, and A. Fujishima, J. Electrochem.
Soc., 146, 1081 (1999).

A. J. Bard and L. R. Faulkner, Electrochemical Methods: Fundamentals and
Applications, John Wiley & Sons, Inc., New York (2001).

N. Wakabayashi, M. Takeichi, M. Itagaki, H. Uchida, and M. Watanabe, J.
Electroanal. Chem, 574, 339 (2005).

B. Case, Electrochim. Acta, 18, 293 (1973).

105

CHAPTER 4

Investigating Microcrystalline and
Nanocrystalline Boron-Doped Diamond as a
Corrosion Resistant Advanced Electrocatalyst
Support Material

4.1 Introduction

Electrocatalytic electrodes are used in industrial processes (e. g., electrosynthesis)
and energy conversion devices (e.g., fuel cells)"2 Such electrodes consist of nanometer-
sized metallic (e.g., Pt) or metal alloy particles supported on a chemically inert,
electrically conductive support material. The current support material, at least for fuel
cells, is sp2 bonded carbon (e.g., physically or chemically activated carbon or carbon
black).3 An ideal electrocatalyst support material possesses several properties including
(i) electrical conductivity (>10 S/cm), (ii) high surface area (>100 mZ/g), (iii) chemical
stability in aqueous environments, (iv) thermal stability up to 200 °C and (v)
morphological and microstructural stability at high potentials and current densities (e. g.,
>1 V vs. the standard hydrogen electrode (SHE) and several A/cmz, respectively).4
Additionally, the materials used should not be overly expensive. Current sp2 carbon
support materials possess the first four properties; however, they are limited by both their
microporosity and susceptibility to microstructural degradation (i.e., oxidation and
corrosion) at high temperatures, potentials and currents?”5 The microporosity of some
supports limits the accessibility of the electrochemical reactant gases to all of the surface.

Support degradation leads to decreased catalytic activity because of catalyst detachment,

106

aggregation, and/or fouling by gasification products. If the degradation is severe enough,

complete electrode failure can result.

Ongoing research in our laboratory is focused in the long-term on developing
electrically conducting polycrystalline diamond as an advanced electrocatalyst support

1.6'lo Diamond possesses several properties desirable for a support material

materia
including: (i) electrical conductivity when highly doped with boron atoms, (ii) low cost
due to advances in diamond production, (iii) thermal stability at elevated temperatures,
(iv) corrosion resistance and (v) dimensional stability at high potentials and current
densities (e.g., >1 V and 1-10 A/cmz, respectively)”14 In order to make diamond a
viable support material, two technical challenges remain to be overcome: the formation
of electrically conducting diamond in a high surface area architecture and establishment

of a protocol for depositing and stabilizing highly dispersed electrocatalyst particles on

the diamond surface.

There are currently two types of polycrystalline diamond being investigated in our
laboratory: microcrystalline (>1 um crystallite size) and nanocrystalline (~10 to 20 nm
crystallite size). Each possesses a different morphology and microstructure but both
exhibit similar electrical conductivity and electrochemical activity when highly doped
with boron. Our group has previously investigated metal phase formation on
microcrystalline diamond using both dc sputtering and electrochemical deposition at
constant potential or current.6'9 Electrodeposition was found to be the superior method

because it allowed for better control of the deposition process (i.e., particle properties)

107

and placed particles solely in areas of the diamond that were electrically conductive. The
particles were then physically anchored to the surface via a short secondary
microcrystalline diamond growth procedure. These films were tested during exposure to
conditions commonly experienced by the carbon electrodes in phosphoric acid fuel cells
(PAFCs) and were found to exhibit stability far superior to that of commonly used sp2

carbon support material.8

Additionally, very recent findings in our laboratory suggest that nanocrystalline
diamond may be a more promising support material then microcrystalline diamond for
several reasons. First, depositing metal using a pulsed galvanostatic approach (see
Chapters 5 and 6) produced particles with more desirable properties in terms of size,
density and electrocatalytic activity towards the oxygen reduction reaction (the cathode
reaction in fuel cells) in acidic media compared to similarly deposited particles on the
microcrystalline diamondls’lé Additionally, metal phase formation on nanocrystalline

diamond has been shown to be better for anodic stripping voltammetric applications.'7-‘18

Second, while a short secondary microcrystalline diamond growth procedure
proved effective in stabilizing metal particles that were 50-300 nm in size, a similar
secondary microcrystalline diamond growth procedure used to stabilize particles that
were an order of magnitude smaller resulted in complete coverage of several particles
(i.e., lost catalyst activity) as well as some particle agglomeration.” This will be
presented in detail in section 5.2.5 of this dissertation. Nanocrystalline diamond is

composed of diamond crystallites (i. e., grains) that are much smaller than those possessed

108

by microcrystalline diamond. If the nanocrystalline diamond is indeed stable, than
perhaps it could also serve to physically anchor the smaller particles to the diamond
surface without completely covering them. Additionally, the lower growth temperatures
and higher nucleation rate for nanocrystalline diamond deposition could also inhibit

particle agglomeration during the secondary diamond growth.

Finally, we have recently taken a step toward our quest to develop electrically
conductive diamond possessing a high surface area. We coated sp2 carbon paper with
boron-doped nanodiamond to produce a diamond electrode with considerably higher
surface area.19 This material, shown in Figure 4.1, was stable during 1 h potentiostatic
polarization in 0.1 M HClO4 at 1.6 v vs. Ag/AgCl and 25 °C.19 The low 112 content in
the gas phase of the nanocrystalline diamond growth made it possible to coat the sp2
carbon paper due to the rate of diamond deposition being higher than the rate of carbon
etching due to H2. It is likely that with smaller carbon fibers, the diamond surface area
could be further increased. While the surface area of this material is still too low for

practical applications, it is a step in the right direction.

We report herein on a comprehensive investigation of the morphological and
microstructural stability of both microcrystalline and nanocrystalline boron-doped
diamond thin films (11 = 3 different films for each diamond type). Each diamond
electrode was exposed to polarization conditions commonly observed in PAP Cs (85 wt%
H3P04, 180 °C, 0.1 A/cmz) for 2 h. The films were characterized using atomic force

(AF M) and scanning electron (SEM) microscopy and Raman spectroscopy. Changes in

109

 

 

the diamond surface chemistry were indirectly tracked using contact angle measurements
with ultrapure H2O. Changes in the electrochemical properties of the films were
investigated using cyclic voltammetry in supporting electrolyte (0.1 M HClO4) and in the

presence of several common redox systems (e.g., Fe(CN)63'/4', Ru(NH3)63 1,21, and Fey/2+).

 

Figure 4.1. Scanning electron micrograph of Toray carbon paper coated with boron-
doped nanocrystalline diamond. Inset is of similar grth at higher magnification.

4.2 Microcrystalline Diamond Results & Discussion

The stability of the physical, chemical, and electrochemical properties of
microcrystalline diamond thin film was assessed. Specifically, the electrodes were
characterized initially, after a 2 h anodic polarization treatment at 0.1 A/cm2 and ca. 180

°C in 85 wt% H3PO4, and after a 30 min H2 plasma treatment.

4.2.1 Physical Properties

The surface morphology of microcrystalline diamond was investigated before and

after anodic polarization using scanning electron and atomic force microscopy. Figure

110

4.2 shows scanning electron micrographs of a microcrystalline diamond film before (A)
and after (B) the anodic polarization treatment. As expected, there is no catastrophic
degradation of the diamond microstructure; however there is some toughening of the
facets and the grain edges as can be seen in the slightly higher magnification image

(Figure 4.2B).

 

Figure 4.2. Scanning electron micrographs of boron—doped microcrystalline diamond (A)
before and (B) after anodic polarization.

lll

Figure 4.3 shows 2 x 2 pm atomic force micrographs obtained (A) before and (B)
after anodic polarization. Clearly, the exposure to the corrosive conditions results in
some minor film roughening of the facet surfaces, but there is no gross morphological
damage or film debonding. This agrees well with the SEM images. Two parameters that
quantitate the surface roughness are the real image surface area (as opposed to the
geometric surface area) and the image RMS roughness which is defined as the absolute
value of the average height and depth away from the zero line (set point ca. -2 V). The
average value for the roughness factor (RF = real area/geometric area) as a function of
image size before and after anodic polarization, is summarized in Table 4.1. On average,
the real surface area was 15 1 2% larger than the geometric area (RFavg = 1.15). This
did not significantly increase after anodic polarization as a value of 18 1 3% was
observed (RFavg = 1.18). The RMS roughness also showed little change before and after
polarization. The RMS roughness value was initially 110 1 16 nm and after anodic
polarization changed little with a value of 105 1 23 nm. These microscopic data suggest
that the diamond is dimensionally stable under these conditions as there was no film

debonding, pitting or cracking.

112

 

0 1 2pm

Figure 4.3. Atomic force micrographs (2 x 2 pm) of a boron-doped microcrystalline
diamond thin film (A) before and (B) after anodic polarization.

113

 

Table 4.1. Comparison of the Real and Geometric Areas of Boron-Doped
Microcrystalline Diamond Films Before and After Anodic Polarization

 

 

 

 

 

Geometric Initial Anodically Polarized
Surface Area
(umz) Real SA Roughness Real SA Roughness
(umz) F actor'I (umz) Factor'

4 4610.1 1.15 4.810.] 1.20

25 29.0 1 0.6 1.16 29.5 1 0.3 1.18

100 11511 1.15 116.7108 1.17

”avg - 1.15 - 1.18

 

 

“ Roughness Factor = Real SA/Geometric SA

 

The microcrystalline diamond microstructure was investigated using Raman
spectroscopy. Figure 4.4 shows representative spectra for a film before and after anodic
polarization. As discussed in Chapter 3, Raman spectroscopy is sensitive to changes in
the diamond microstructure including the relative levels of sp2- and sp3-bonded carbon,
the atomic ordering and residual internal stress. These spectra are quite different from
the one presented in Figure 3.2 for high quality (i.e., low nondiamond sp2 carbon
impurity content) diamond. This is due to the significantly higher doping level in this
particular film, as gaseous B2H6 was used as the dopant source instead of the solid B203
diffusion source. The high boron-doping level increases the tensile stress in the film as
evidenced by the down-shift in the diamond phonon peak from 1332 to ca. 1323 cm".20
Wang et a]. attributed this increase in tensile stress to a high defect density introduced by
the blocking of grain boundary sliding during polycrystalline diamond growth.20 The

asymmetry of the diamond phonon band is possibly due to a Fano interference induced

114

 

by quantum mechanical interference between the discrete phonon state and the electronic

continuum.”

 

 

Boron Doping Diamond
’1‘
:1
55.
a?
U)
5
E Anodically Polarized
1—1
Initial
1 1 l m I 1 i l 1 I 1 l 1

 

 

 

 

800 1000 1200 1400 1600 1800 2000
Raman Shift (cmAl)

Figure 4.4. Raman spectra of heavily boron-doped microcrystalline diamond initially and
after anodic polarization. Dotted lines correspond to boron doping and diamond signals,
respectively. Spectra are off-set for viewing clarity.

The spectrum for a highly doped diamond film shown in Figure 4.4 also displays
a broad band at ~1220 cm]. This band has been previously observed along with a down-
shift of the diamond phonon peak for films possessing boron concentrations >1020
B/cm3.22 Gonon et a]. attributed this band to a maxima in the phonon density of states
originating from disordered regions or regions of very small crystallite size.22 The dotted
lines in Figure 4.4 denote the down-shifted diamond phonon band and high boron

concentration band at 1323 and 1220 cm], respectively.

115

There is no observable difference in the spectral features before and after anodic
polarization, specifically in terms of the diamond peak position or the presence of spz-
bonded nondiamond carbon. This indicates that the diamond microstructure and optical
properties were largely unaffected by the polarization. The initial average diamond peak
position for the film was 1320 1 3 cm". Anodic polarization did not significantly alter
this value as the average position was 1324 1 3 cm". The down-shifted peak is
consistent with the region of the film probed being under tensile stress. The line width
(FWHM) of the diamond phonon peak is a qualitative indicator of the crystalline film
quality.”24 However, the presence of the boron signal in the higher doped films makes it
difficult to measure. Regardless, efforts were made to estimate this value as best as
possible. Values of 31 1 6 cm’1 and 30 1 5 cm’1 were obtained before and after anodic
polarization, respectively. Comparatively, the FWHM measured for a piece of single
crystal diamond under the same settings was 5-7 cm". According to Gonon et al.,
FWHM values of 16 and 40 cm’1 were measured for CVD diamond/Si samples

possessing boron concentrations of 1.8 x 1020 and 1.1 x 102] B/cm3, respectively.22

We recently measured the boron concentration in a microcrystalline diamond film
grown with 10 ppm B2H6 in the source gas using boron nuclear reaction analysis (BNR).
The B/C concentration was determined to be 2036 ppm. Assuming an atomic density of
1.76 x 1023 atoms/cm3 for diamond, this equates to a value of 3.6 x 1020 B/cm3. This
value agrees well with that deduced from the FWHM analysis discussed above.
Additionally, powder x-ray diffraction spectra identified diamond crystallite orientations

of (111), (220) and (311) for films grown in our laboratory. This means our value also

116

 

agrees well with that of Ushizawa et a]. who found the boron concentration in the (111)
facet of a 2000 ppm CVD diamond/Si sample to be 2.3 x 102' B/cm3 using secondary ion

mass spectrometry (SIMS).21

4.2.2 Surface Chemistry

The surface chemistry (i.e., oxide content) was investigated by measuring the
instantaneous contact angle for a 10 uL drop of ultrapure water interacting with the
diamond surface. Initially, the hydrogen-terminated diamond surface was hydrophobic
and, as expected, exhibited a high instantaneous contact angle of 69 1 4°. The diamond
surface was expected to become more hydrophilic after anodic polarization. Such a
change should be manifested in a decreased water contact angle. However, the contact
angle did not decrease significantly, only to a value of 61 1 4°. These values agree well
with those of Ostrovskaya et al.25 However, following the rehydrogenation procedure,
the contact angle substantially increased to 101 1 7°, suggesting the original surface
contained some carbon-oxygen functionalities. Again this is in agreement with
Ostrovskaya eta]. and Pinzari et al.25’2° Based on this indirect measurement, the surface
oxygen content increased some, but not significantly, after polarization. The values are
summarized in Table 4.2. The difference in the contact angles for the film before and
afier rehydrogenation is likely due to exposure of the electrode to the laboratory
atmosphere. As described in Chapter 2, our diamond deposition procedure involves a
slow cool down in an H2 plasma (i.e., atomic hydrogen) in order to etch away
adventitious nondiamond carbon impurity, minimize dangling bonds, and terminate the

diamond surface with hydrogen atoms. Prior to the initial contact angle measurements,

117

 

the diamond films had been exposed to the laboratory atmosphere for a considerable
period of time (i. e., days or even weeks) and slowly oxidized. Recent XPS data revealed
that films exposed to atmospheric conditions for a similar amount of time exhibited an
O/C ratio of ~0.05. Assuming a sampling depth of 5 layers, this is consistent with a
surface coverage of 0.25. Additionally, Ostrovskaya et a]. obtained similar contact

angles for naturally oxidized diamond films.25

The final set of contact angle data was
obtained immediately following the rehydrogenation process, before the diamond surface

had a chance to oxidize.

 

Table 4.2. Summary of Contact Angle Measurements for Ultrapure Water on Boron-
Doped Microcrystalline Diamond Thin Film as a Function of the Surface Chemistry

 

 

Diamond Condition Initially Anodically Polarized Rehydrogenated
Contact Angle 69 1 4° 61 1 4° 101 1 7°
Percent Change — -12% +66%

 

 

 

 

 

4.2.3 Electrochemical Response

Cyclic voltammetric i-E curves in 0.1 M HClO4 (N2 purged) were obtained
before, after anodic polarization and after rehydrogenation. Representative examples are
shown in Figure 4.5. All the curves are flat and featureless between the regions of
solvent breakdown. The potential window data, determined from measurement of three
microcrystalline electrodes at 150 11A or 1250 uA/cmz, are summarized in Table 4.3.
Prior to polarization, the average potential window was 2.9 1 0.1 V. Anodically
polarizing the diamond surface did not affect the potential window as the average

remained 3.0 1 0.2 V. The similarity of these values suggests that the diamond films

118

 

possessed minimal sp2 carbon content on the surface initially. Nondiamond sp2 carbon
phases tend to reduce the potential window.27 Anodic polarization removes such carbon
resulting in an increased window. The potential window was also not altered much

following the rehydrogenation as the nominal value was 3.05 1 0.01 V.

 

L I25lJ-A

K

 

 

Rehydrogenated
r
A
g. ..
j: Anodically Polarized
= -
a)
i-
i- -
:1
U Initial

 

 

 

I
l—
l 1 l 1 l

-1500 -500 500 1500 2500
Potential (mV vs. Ag/AgCl)

 

Figure 4.5. Background cyclic voltammetric i-E curves for microcrystalline diamond
films in degassed 0.1 M HC104 as a firnction of the surface condition. Scan Rate = 100
mV/s.

The anodic background current was also recorded at a potential of 400 mV; a
potential at which the current is mostly nonfaradaic. The background current magnitude
for each film after the different treatment was quite similar. Values of 0.7 1 0.1, 0.8 1 0.1
and 0.72 1 0.03 uA were obtained for the initial, anodically polarized, and

rehydrogenated films, respectively. The similar values indicated that neither the anodic

119

 

 

polarization nor the rehydrogenation roughens the surface, consistent with the

microscopy.

 

Table 4.3. Potential Window and Background Current Measured from Cyclic
Voltammetric i-E Curves for a Boron-Doped Microcrystalline Diamond Thin Film in 0.1
M HC104 at a Scan Rate of 100 mV/s

 

 

 

, , , ' ' Back round Current
Diamond Condition Potentrzr‘ll)Wmdow g (11A)
Initial 2.9 1 0.1 0.7 1 0.1
Anodically Polarized 3.0 1 0.1 0.8 1 0.1
Rehydrogenated 3.05 1 0.01 0.72 1 0.03

 

 

3‘ Potential window is defined as 150 uA or 1250 ).lA/cm2

The electrochemical response for the microcrystalline diamond films towards
several well known redox analytes was also measured before and alter the treatments.
Figure 4.6A-C shows representative cyclic voltammetric i-E curves for Fe(CN)63'/4',
Ru(NH3,)63"A/2+ and Few2+ at a microcrystalline diamond before, after anodic polarization
and after rehydrogenation. The basis for the selection of these analytes as redox probes
was given in section 3.2.2. Figure 4.6A shows voltammetric i—E curves for Fe(CN)53'/4'.
Fe(CN)6°'/4' is an inner-sphere redox system (i.e., specific analyte-surface interaction
required for electron transfer) with a heterogeneous electron—transfer rate constant that is
sensitive to the electronic properties, surface cleanliness and the presence of surface

. . 4, .
oxrdes on diamond electrodes.1 27

28 In other words, AEp increases from the reversible
value of 60 mV if the film has poor electrical conductivity, an unclean surface or surface
oxygen functionalities. The largest rate constant (i.e., smallest AEp) is seen for the clean

hydrogen-terminated surface. The potential peak separation, AEp, at each scan rate is

displayed in Table 4.4. The AEp at 100 mV/s before and after rehydrogenation were both

120

65 1 2 mV, near Nemstian, suggesting that these surfaces were both clean and hydrogen-
terminated. This value increased considerably to 91 1 7 mV after anodic polarization,
during which some surface oxygen functional groups are introduced.28 It appears that
oxides formed on the diamond surface somehow block sites that are active toward this
reaction when H-terminated. Our group has previously determined this blocking effect to
be of electrostatic origin as the AEp for this reaction was directly proportional to the

electrolyte pH for an oxidized diamond surface (i. e., protonation of the surface oxides).28

 

Table 4.4. Cyclic Voltammetric Peak Potential Separation (AEp) for 1 mM F e(CN)63" 4’ in
1 M KCl at Boron-Doped Microcrystalline Diamond as a Function of the Surface State

 

 

 

Scan Rate A13, for Fe(CN),°""(1nV)
Initially Anodically Polarized Rehydrogenated
50 65 1 0 85 1 6 64 1 1
100 6512 9117 6512
300 7112 11217 7311
500 7813 12919 7912

 

 

 

 

Curves for Ru(NH3)63A/2+ at a diamond film before, after anodic polarization and

after rehydrogenation are shown in Figure 4.68. Ru(NH3)63+/2+

is an outer-sphere redox
system (i.e., no specific interaction with the surface) with a rate constant that is sensitive
to the electronic properties but relatively insensitive to the surface cleanliness or oxide
content. The data are summarized in Table 4.5. Clearly, AEp is unchanged for all the

films and is near the Nemstian value of 60 mV at 100 mV/s. This means that the

electronic properties of the diamond are metallic-like and unaltered by the treatments.

12]

 

 

   
   

 

 

 

 

 

 

 

   
    

 

 

 

 

 

 

 

 

60
40
2 .
3 20
H
5
1.1 0
811
:1
U -20
—- Initial
‘40 _._ Anodically Polarized
—*— Rehydrogenated
-60
0 200 400 600
Potential (mV vs. Ag/AgCl)
60
.
40
SE .
3 20
H
:1 .
2 0
h
:1
U -20 ,
’ ' — initial
-40 ' + Anodically Polarized
' -+- Rehydrogenated
.60 .—
-400 -200 0
60 Potential (mV vs. Ag/AgCl)

 

 

 

 

 

 

 

 

 

A
<2
3
H
E
u
1...
l1-
3
0
Initial
- 0 - Anodically Polarized
-60 .. -+- Rehydrogenated
-80 1 l 1 J 1 l 1 l
-400 0 400 800 1200

Potential (mV vs. Ag/AgCl)

Figure 4.6. Cyclic voltammetric i-E curves for boron-doped microcrystalline diamond
films before, after anodic olarization and after rehydrogenation in the presence of (A)
Fe(CN),3"4', (B) Ru(NH3)6 “2* and (C) Few”. Scan Rate = 100 mV/s.

122

 

 

Table 4.5. Cyclic Voltammetric Peak Potential Separation (7113,) for 1 mM Ru(NH3)63+/2+
in l M KCl at Boron-Doped Microcrystalline Diamond as a Function of the Surface State

 

 

 

Scan Rate AEP f" R“(NH3)63+/2+ (mV)
Initially Anodically Polarized Rehydrogenated
50 64 1 1 66 1 1 65 1 1
100 6513 6612 6611
300 7212 7213 7314
500 7612 7412 8013

 

 

 

 

3+/2+ 3+/2+

The voltammetric i-E curves for Fe are shown in Figure 4.6C. The Fe
couple is an inner-sphere redox system with a rate constant that is known to be sensitive
to electronic properties, surface cleanliness and surface oxide content, particularly
carbonyl functional groups (e.g., C=O). In general, the AEp values for this redox system
are much larger than for Fe(CN)63'/4' and Ru(NH3)(;3+/2+ reflecting slower electron transfer
kinetics. There apparently is a large activation energy barrier to adjust the solvation
sphere around the metal when going from the +2 to +3 oxidation state.29’3° The data are
summarized in Table 4.6. The peak separation for the film initially was 380 1 60 mV.
This separation increased some after anodic polarization to 460 1 80 mV. This would be
consistent with few carbonyl functional groups being introduced onto the surface, based
on the work by McCreery.31 The peak separation increased significantly after
rehydrogenation to 700 1 240 mV (note the large variance). Most, if not all, the oxygen
functionalities present on the surface before or after anodic polarization were removed by

the H2 plasma treatment. Even though this surface is highly electrically conductive, the

electron-transfer kinetics is quite sluggish for reasons that remain unclear. It could be

123

that the hydrophobic diamond surface provides an additional barrier to reorganization of

the solvent sphere surrounding the metal center during the redox event.

 

Table 4.6. Cyclic Voltammetric Peak Potential Separation (AEp) for 1 mM F e3” 2+ in 0.1
M HClO4 at Boron-Doped Microcrystalline Diamond as a Function of the Surface State

 

 

 

 

Scan Rate AEP for Few“ (mV)
Initially Anodically Polarized Rehydrogenated
50 330 1 60 390 1 100 660 1 220
100 380 1 60 460 1 80 700 1 240
300 460 1 60 530 1 90 800 1 250
500 500 1 60 580 1 80 840 1 250

 

 

 

The forward reaction peak current for the Fe(CN)63'/4' and Ru(NH3)63+/2+ redox
systems at all the surfaces varied proportionally with scan rate"2 (VI/2) (r2 > 0.999)
consistent with a redox reaction rate limited by semi-infmite diffusional mass transport of
reactant to the electrode surface. Since the reactions are nearly reversible for these two
redox systems (AEp is near 59 mV), the peak current ratio is 1.0, as expected. The
ip°"/ipred ratio for F cw 2+ couple however was less than 1.0, generally ca. 0.7. The forward

1/2

oxidation peak current scaled linearly with v (r2 > 0.999) also expected for a kinetically

irreversible reaction.

4.3 Nanocrystalline Diamond Results & Discussion

4.3.1 Physical Properties

The physical, chemical and electrochemical properties of boron-doped

nanocrystalline diamond thin film were evaluated before, after anodic polarization and

124

 

after rehydrogenation. Figure 4.7 shows a typical SEM image of a nanocrystalline
diamond thin film after anodic polarization. Clearly, there is no large scale bulk
degradation or pitting on the surface. Macroscopically, the film looks the same before

and after polarization.

 

Figure 4.7. Scanning electron micrograph of boron-doped nanocrystalline diamond film
after anodic polarization.

The atomic force micrographs shown in Figure 4.8A and B also do not indicate
any visible morphological change to the film after the polarization. The images appear
slightly grainy which may indicate some minor toughening or pitting. Atomic force
micrographs were obtained over geometric areas of 0.25, 4, and 25 umz. The difference
between the geometric area and actual surface area, summarized in Table 4.7, as well as
the RMS roughness for the initial and polarized nanocrystalline diamond films were used
to probe for changes in the microscopic surface roughness. Comprehensively, the real
area for all the images was 8 1 3% larger than the geometric image area (RFavg = 1.09)
for the film prior to any treatment. This value decreased slightly after anodic polarization

to a value of 4 1 3% (RFavg = 1.05), presumably due to etching of some spZ-bonded

125

carbon in the grain boundaries. The RMS roughness did not change significantly alter
anodic polarization as the value was 20 1 6 before and 18 1 6 nm after anodic

polarization.

 

0 250 500nm

Figure 4.8. Atomic force micrographs (500 x 500 nm) of boron-doped nanocrystalline
diamond thin film (A) before and (B) after anodic polarization.

126

 

 

Table 4.7. Comparison of the Real and Geometric Areas for Boron-Doped
Nanocrystalline Diamond Films Before and After Anodic Polarization

 

 

 

 

 

Projected Initially Anodically Polarized
Surface Area
(umz) Real SA Roughness Real SA Roughness
(umz) F actora (umz) F actor'I

0.25 0.28 1 0.01 1.12 0.274 1 0.001 1.10

4 4.3 1 0.1 1.08 4.13 1 0.01 1.03

25 26.5 1 0.7 1.06 25.6 1 0.1 1.02

RF1vg — 1.09 — 1.05

 

 

3‘ Roughness Factor (RF) = Real-SA/Geometric SA

The microstructure of nanocrystalline diamond was investigated using Raman
spectroscopy. The Raman spectra for nanocrystalline diamond films grown from argon-
rich (Ar-rich) source gas mixtures are often characterized by broad peaks at ~1150, 1332,
1350, ~1480, and 1550 cm'1.'3‘32'33 Nanocrystalline diamond, grown from Ar-rich
mixtures, consists of small clusters of sp3-bonded diamond grains connected through sp2

hybridized carbon bonds in the grain boundaries.l3’3°’35

As with microcrystalline
diamond, the peak at 1332 cm", partially obscured by a large background scattering
signal, is associated with the first-order phonon mode of sp3-bonded diamond. The peak
width (F WHM) for the diamond line is larger in nanocrystalline than in microcrystalline
diamond (140 vs. 10 cm") for two possible reasons. First, is the confinement model
which states that the smaller the crystal size is, the larger the number of phonon modes

(with different q vectors and energies) that can participate in the Raman process.36 The

larger number of modes of slightly different energy leads to the increased line width.

127

 

Second, is the phonon scattering by impurities and defects (i.e., grain boundaries) which

shortens the phonon lifetime and results in an increased line width.

The peak at 1150 cm'1 has often been used as an indicator for high-quality

37,38

nanocrystalline diamond grown from Ar-rich mixtures. However, Ferrari and

Robertson have argued against this suggesting that the peak is attributable to spz-bonded
carbon, specifically trans-polyacetylene segments at grain boundaries.”’33 °“° "fs' them“
This is based on their observations that (i) the peak position changes and the intensity
decreases with decreasing excitation wavelength, something that is not possible for a
density-of-states feature which should remain invariant, and (ii) the peak is always
accompanied by the 1480 cm’1 peak, which behaves in a similar fashion. The former is
due to increasing cross sectional scattering coefficient for diamond at the shorter
excitation wavelengths (decreasing cross section for nondiamond sp2 carbon). This has
recently been confirmed by Birrel et a]. who found a transition from nanocrystalline to
microcrystalline morphology with increasing H2 content (0 — 20%) in a CH4/Ar/H2 source

39 Thus, it was confirmed that despite consisting of mostly sp3-bonded

gas mixture.
carbon (~95%), the Raman signal was due entirely to the spZ-bonded carbon in the grain
boundary.” Additionally, the peaks at 1120 and 1480 cm'1 were loosely attributed to C-
H bonds in the grain boundaries, while the peaks at 1350 and 1560 cm'l were attributed

to the D- and G-bands, respectively, that are routinely observed for disordered spz-

bonded carbon.39

128

Figure 4.9 shows representative Raman spectra for a nanocrystalline diamond
film before and after anodic polarization. Each frequency of interest is denoted by a
dotted line. The first-order diamond peak at 1332 cm", as well as the D and G band
peaks, remained unchanged in terms of peak position, intensity and width after the
polarization. These trends are consistent with an unchanging microstructure. The Raman
peaks at 1150 and 1480 cm'1 are represented by V] and V3, respectively. These peaks
become slightly less pronounced after polarization. This likely indicates a decrease in the
amount of hydrogenated sp2 carbon (e. g., C-C—H) and possibly H atoms within the grain
boundaries. Ferrari et a]. have shown that the v1 and V3 modes of trans-polyacetylene are
connected to the presence of hydrogen. Post-deposition annealing as well as growth of
the films using deuterated and hydrogenated plasmas resulted in either a significant shift

in the peak position or the complete loss of the peaks, depending on the experiment.33 “d

refs. therein

 

Dirnoud i D

\

    
 

V1

Anodically
Polarized

Intensity (A.U.)

Initial

 

 

 

 

 

 

 

800 1000 1200 1400 1600 1800 2000
Raman Shift (cnil)

 

Figure 4.9. Raman spectra for a boron-doped nanocrystalline diamond film before and
after anodic polarization. Spectra are off-set for viewing clarity.

129

4.3.2 Surface Chemistry

The surface chemistry, as assessed from the relative
hydrophobicity/hydrophilicity, of the nanocrystalline diamond thin film was also
evaluated before, after anodic polarization and after rehydrogenation. Again a 10 uL
drop of ultrapure water was used in contact angle measurements. The measured
instantaneous contact angles are summarized in Table 4.8. Initially, the nanocrystalline
diamond film exhibited a contact angle of 82 1 4°. This was slightly higher than the
value initially obtained for the microcrystalline diamond surface and is likely due to the
reduced surface roughness of the film. The contact angle decreased considerably after
the anodic polarization to a value of 43 1 6° consistent with increased surface oxygen
content. The contact angle then increased for the rehydrogenated film to 64 1 4°, but not

quite to the original level.

 

Table 4.8. Summary of Contact Angle Measurements for Ultrapure Water on Boron-
Doped Nanocrystalline Diamond Thin Film as a Function of the Surface Chemistry

 

 

Diamond Condition Initially Anodically Polarized Rehydrogenated
Contact Angle 82 1 4° 43 1 6° 64 1 4°
Percent Change — -48% +49%

 

 

 

 

 

4.3.3 Electrochemical Response

The cyclic voltammetric background i-E curves for a film before, after anodic
polarization and after rehydrogenation was measured in N2 purged 0.1 M HClO4.
Typical curves, over a narrow potential range, are shown in Figure 4.10. The measured

potential window (150 11A or 1 250 uA/cmz) and background current magnitude at 400

130

mV are displayed in Table 4.9. The voltammetric i-E curve for the film before any
treatment was flat and featureless between the regions of gas evolution indicating a clean
diamond surface void of any electroactive carbon. The nanocrystalline films exhibited an
average potential window and background current of 3.0 1 0.2 V and 1.7 1 0.1 uA,
respectively. After anodic polarization, the potential window decreased to 2.1 1 0.2 V
and the background current increased to 41 1 16 11A. The voltammetric i-E curve is still
flat suggesting the surface was largely void of electroactive surface functional groups.
This change in potential window and background current is considerably greater than that
observed for the microcrystalline diamond. These quantities M ngt recover with
rehydrogenation as a potential window and background current were 2.2 1 0.2 V and 39

1 14 uA, respectively.

 

500

      
   
 

IsouA

400 Rehydrogenated

I Y I rj U

300

 

200

'TUfI'T

Anodically Polarized

1...... J

 

100

Current (11A)

IfiIfT

 

/

-200 1 1 1 I 1 1 1 l 1 1 1 l L 1 1 1 m 1 1 I 1 1 1
-1000 -500 0 500 1000 1500 2000

Potential (mV vs. Ag/AgCl)

-100

f7 '1 r f! 1"

 

 

 

Figure 4.10. Background cyclic voltammetric i-E curves in degassed 0.1 M HClO4 for
boron-doped nanocrystalline diamond films as a function of the surface condition. Scan
Rate = 100 mV/s.

131

 

Table 4.9. Potential Window and Background Current Measured from Cyclic
Voltammetric i-E Curves for a Boron-Doped Nanocrystalline Diamond Thin Film in 0.1
M HClO4 at a Scan Rate of 100 mV/s.

 

Background Current

 

Diamond Condition Potential Window (V)' (PA)
Initially 3.0 1 0.2 1.7 1 0.1
Anodically Polarized 2.1 1 0.2 41 1 16
Rehydrogenated 2.2 1 0.2 39 1 l4

 

 

 

a Potential window is defined as 150 uA or 1250 uA/cm2

The electrochemical response of the film towards Fe(CN)63'/4', Ru(NH3)(,3+’2+ and
F e” 2+ was recorded before and after the treatments. Typical voltammetric i-E curves for

each of these analytes are presented in Figure 4.11A-C. The AEp value for each at

various scan rates is displayed in Tables 4.10, 4.11 and 4.12.

The AEp for Fe(CN)63'/4' at the nanocrystalline film was initially 69 mV (100
mV/s), as shown in Table 4.10. Afier anodic polarization, AB], increased to 120 mV
consistent with an increase in surface oxygen content. Also note the increase in
background current in the voltammetric i—E curve in Figure 4.11A. After
rehydrogenation, AEp returned to near the original value of 65 mV. The voltammetric i-E
curve though still exhibits the large background current. The oxidation peak current

1/2

varied linearly with the scan rate (r2 > 0.999) indicating the rate of the redox reaction is

limited by diffusional mass transport to the surface.

132

100

 

 

U!
G
V I T

 

 

 

Current (pA)

  
 

 

— Initial
-0- Anodically Polarized

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

. + Rehydrogenated
_100 1 1 1 l 1 1 1 l 1 1 1 1 1 1 1 1
-200 0 200 400 600
8 0 Potential (mV vs. AgIAgCl)
:5 4o -
5
E
E.“ 0 -
L-
=
U
'40 f --— Initial
_ + Anodically Polarized
--— Rehydrogenated
.80 1 1 1 l 1 1 4 L 1 1 1 L L 1 l
-600 -400 -200 0 200
100 Potential (mV vs. AflgCl)

 

 

 

J

Current (uA)
U.
Q

 

 

 

 

 

 

 

 

_50 j — Initial
+ Anodically Polarized
. “0- Rehydrogenated
-100 l J L L L 1 1
0 400 800 1200

Potential (mV vs. Ag/AgCl)

Figure 4.11. Cyclic voltammetric i-E curves for boron-doped nanocrystalline diamond

films before, after anodic olarization and afier rehydrogenation in the presence of (A)
Fe(CN)63'/*, (B) Ru(NH3)6 ”2* and (C) Fem“. Scan Rate = 100 mV/s.

133

 

Table 4.10. Cyclic Voltammetric Peak Potential Separation (AEp) for 1 mM F e(CN)63‘/4'
in 1 M KCl at Boron-Doped Nanocrystalline Diamond as a Function of the Surface State

 

 

 

 

Sm Rate 112,, for Fe(CN)63"‘ (mV)
Initially Anodically Polarized Rehydrogenated
50 6612 100130 6510
100 6912 120140 6512
300 7416 160170 7814
500 8215 180170 8512

 

 

 

The 113, for Ru(NH3),3+’2*, which is known to be relatively insensitive to the
surface chemistry of diamond, remained relatively unchanged throughout the course of
the experiments. AEp values of 64-82, 69-93, and 66-85 mV were observed at scan rates
of 50-500 mV/s before, after anodic polarization and after rehydrogenation, respectively,
as seen in Table 4.11. This indicates that despite the increase in capacitance, the
electrical conductivity and activity of the film remained unaltered. As with the F e(CN)63 '
oxlipred

W couple, the ip ratio remained approximately unity and each scaled linearly with

v'”, indicating that the increased capacitance did not alter the reversibility of the reaction.

 

Table 4.11. Cyclic Voltammetric Peak Potential Separation (AEp) for 1 mM
Ru(NH3)63+/2+ in 1 M KC1 at Boron-Doped Nanocrystalline Diamond as a Function of the
Surface State

 

 

 

 

Scan Rate 11:, for RulNHslté’”2+ (mV)
Initially Anodically Polarized Rehydrogenated
5° 64 i 2 69 1 2 66 1 2
100 6711 6714 6713
300 7414 8318 75:3
500 8215 9316 8512

 

 

134

 

 

The 112, for Fey/2+ exhibited a wide variation throughout the course of this work
as shown in Table 4.12. The AEp ranged from 480-660 mV before any treatment,
consistent with an oxygen-free surface. As expected, this value decreased to 129-165
mV afier anodic polarization. The relatively low AEp indicates that surface carbon-
oxygen functional groups were introduced, particularly carbonyl groups. After
rehydrogenation, AISp increased to a high value of 390-640 mV. Similar to the
microcrystalline diamond, the ip°"/ipmd ratio for this redox system was less than unity (ca.
0.7-0.8) as expected due to the electrochemical irreversibility. Interestingly, the increase
in film capacitance did not influence the electrochemical response towards any of the

redox analytes.

 

Table 4.12. Cyclic Voltammetric Peak Potential Separation (AEP) for 1 mM F e3 H 2+ in 0.1
M HClO4 at Boron-Doped Nanocrystalline Diamond as a the Function of Surface State

 

 

 

Scan Rate AEP for Few“ (mV)
Initially Anodically Polarized Rehydrogenated
50 480 1 50 130 1 30 390 1 130
100 530 1 50 130 1 30 470 1130
300 620 1 50 170 1 50 570 1 110
500 660 1 50 165 1 9 640 1 120

 

 

 

 

The nanocrystalline diamond could also serve as a dimensionally stable
electrocatalyst support. However, the irreversible increase in background current is a
concern. There is a least some minor damage to the nanocrystalline diamond film that is
not detected in the microscopy. One thing worth noting is that when the nanocrystalline

diamond was removed from the cell, the current collecting Cu plate had turned purple

135

 

iii- _ i |

where the film had been. Additionally, the VitonTM o-ring was stuck to the
nanocrystalline diamond surface and had to be peeled off, leaving a noticeable ring of
possible residue on the surface. This residue could not be removed with cleaning and
rehydrogenating. In fact after rehydrogenation, the nanocrystalline film still possessed a
dark spot where the film was exposed to the solution. None of this reactivity was seen
with the microcrystalline film. It is almost as if the nanocrystalline film (due to
resistance) got hotter during the anodic polarization than the microcrystalline diamond.
Perhaps the VitonTM residue causes the increase in background current when that portion
of the film is exposed to the electrolyte solution when remounted in the electrochemical
cell. A possible source of increased capacitance is degradation of the carbon within the
grain boundaries making the underlying Si substrate susceptible to solution. If this
occurred during the anodic polarization, the Si would have been etched; a reaction which
is exothermic.4043 Perhaps this released heat caused both the discoloration of the Cu
plate and melting of the o-ring. Additionally, exposure of the electrolyte solution to the

Si substrate would increase the observed capacitance in the i-E curves.

4.3.4 Electrical Conductivity Maps

Conductivity probe-atomic force microscopy (CP-AFM) was used to
simultaneously map the film topography and electrical conductivity. Figure 4.12 presents
AFM height mode images over a 1 x 1 pm area as well as conductivity maps of the same
area for a boron—doped nanocrystalline diamond film after the completed experiment (i.e.,
after both polarization and rehydrogenation) and compared to a H—terminated

nanocrystalline diamond film which was previously presented.15 Though these are not of

136

the same area, there is a greater distance between areas of higher conductivity for the
polarized film compared to a normal nanocrystalline diamond film. The anodically
polarized film shows very distinct "hot spots" of high conductivity and regions of almost
complete insulation. It is possible that the electrically conducting regions are
redistributed during the anodic polarization resulting in the increase in electrode
capacitance. It is also possible that the removal of trans-polyacetylene (i.e., H atoms)

from the grain boundaries explains the change in the conductivity maps.

 

Figure 4.12. AFM height mode images (air) (A and C) and simultaneously recorded
conductivity maps (bias voltage, Erip - Esubsm = +2 V) of the same regions (B and D) of
a boron-doped nanocrystalline diamond thin film over a 1 x 1 pm area not polarized (A
and B) and anodically polarized (C and D).

137

4.4 Conclusions

Boron-doped microcrystalline and nanocrystalline diamond thin-film electrodes
were evaluated as possible electrocatalyst support materials by assessing their stability
under conditions commonly encountered in PAFCs. The microcrystalline diamond
exhibited some surface oxidation resulting in slight roughening of the crystallite edges.
Neither atomic force nor scanning electron microscopic images revealed signs of gross
morphological degradation (e.g., pitting or debonding). Additionally, any changes in the
electrochemical properties brought about by the oxidative condition of the polarization
were reversed after rehydrogenation. This is consistent with the changes being brought
about by surface chemistry effects. The microcrystalline form of diamond was
dimensionally stable and maintained a high degree of electrical conductivity and
electrochemical responsiveness after the polarization. Indeed it appears that
microcrystalline diamond would be a physically stable material if employed as an

electrocatalyst support.

The nanocrystalline form of diamond also appears to be a candidate as an
electrocatalyst support. The nanocrystalline film was altered to a greater extent than the
microcrystalline film during anodic polarization as evidenced by the larger change in
contact angle, increased background voltammetric current, and response change for
Fe(CN)63'/4‘ and Fey/2+. It appears that solution can penetrate the grain boundaries and
react with the underlying Si during the polarization. The diamond is stable but the grain
boundary carbon is not. However, the response towards Fe(CN)63'/4' and F e3“2+ returned

to the original values upon rehydrogenation indicating that alterations in the

138

 

 

 

nanocrystalline diamond surface chemistry are reversible. Microscopically, no gross
morphological degradation was seen but some changes occurred that were irreversible.
The electrical conductivity of the material was unchanged after polarization and
rehydrogenation. All of this leads to the conclusion that the nanocrystalline form of

diamond might function as a stable electrocatalyst support material.

139

10.

ll.

12.

13.

REFERENCES

E. Antolini, J. Mater. Sci., 38, 2995 (2003).

F. Rodriguez-Reinoso, Carbon, 36, 159 (1998).

E. Auer, A. Freund, J. Pietsch, and T. Tacke, Appl. Catal. A: Gen, 173, 259
(1998)

M. W. Reed and R. J. Brodd, Carbon, 3, 241 (1965).

B. C. H. Steele and A. Heinzel, Nature, 414, 345 (2001).

J. Wang, G. M. Swain, T. Tachibana, and K. Kobashi, Electrochem. Solid-State
Lett., 3, 286 (2000).

J. Wang, G. M. Swain, T. Tachibana, and K. Kobashi, J. New Mater.
Electrochem. Sys. , 3, 75 (2000).

J. Wang and G. M. Swain, Electrochem. Solid-State Lett., 5, E4 (2002).

J. Wang and G. M. Swain, J. Electrochem. Soc., 150, E24 (2003).

A. E. Fischer and G. M. Swain, J. Electrochem. Soc., 152, B369 (2005).

M. Witek, J. Wang, J. Stotter, M. Hupert, S. Haymond, P. Sonthalia, G. M.
Swain, J. K. Zak, Q. Chen, D. M. Gruen, J. E. Butler, K. Kobashi, and T.
Tachibana, J. Wide Bandgap Mater., 8, 171 (2002).

J. Wang, G. M. Swain, M. Mermoux, G. Lucazeau, J. Zak, and J. W. Strojek, New
Diamond Front. Carbon T echn01., 9, 317 (1999).

Y. Show, M. A. Witek, P. Sonthalia, and G. M. Swain, Chem. Mater., 15, 879
(2003)

140

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

M. C. Granger, M. Witek, J. Xu, J. Wang, M. Hupert, A. Hanks, M. D. Koppang,
J. E. Butler, G. Lucazeau, M. Mermoux, J. W. Strojek, and G. M. Swain, Anal.
Chem, 72, 3793 (2000).

J. A. Bennett, Y. Show, S. Wang, and G. M. Swain, J. Electrochem. Soc., 152,
E184 (2005).

J. A. Bennett and G. M. Swain, Manuscript in preparation.

P. Sonthalia, E. McGaw, Y. Show, and G. M. Swain, Anal. Chim. Acta, 522, 35
(2004)

E. A. McGaw and G. M. Swain, Anal. Chim. Acta, 575, 180 (2006).

A. E. Fischer, Ph. D. Dissertation, Degradation Mechanisms of Carbon-Based
Electrocatalyst Support Materials and Development of an Advanced Support

Based on Electrically Conducting Diamond, Michigan State University, East
Lansing, MI, (2005).

W. L. Wang, M. C. Polo, G. Sénchez, J. Cifre, and J. Esteve, J. Appl. Phys, 80,
1846 (1996).

K. Ushizawa, K. Watanabe, T. Ando, I. Sakagucki, M. Nishitani-Gamo, Y. Sato,
and H. Kanada, Diamond and Rel. Mater., 7, 1719 (1998).

P. Gonon, E. Gheeraert, A. Deneuville, F. Fontaine, L. Abello, and G. Lucazeau,
J. Appl. Phys, 78, 7059 (1995).

M. C. Granger, J. Xu, J. W. Strojek, and G. M. Swain, Anal. Chim. Acta, 397, 145
(1999).

M. Mermoux, B. Marcus, G. M. Swain, and J. E. Butler, J. Phys. Chem. B, 106,
10816 (2002).

L. Ostrovskaya, V. Perevertailo, V. Ralchenko, A. Dementjev, and O. Loginova,
Diamond and Rel. Mater., 11, 845 (2002).

141

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

F. Pinzari, P. Ascarelli, E. Cappelli, G. Mattei, and R. Giorgi, Diamond and Rel.
Mater., 10, 781 (2001).

J. A. Bennett, J. Wang, Y. Show, and G. M. Swain, J. Electrochem. Soc., 151,
E306 (2004).

M. C. Granger and G. M. Swain, J. Electrochem. Soc., 146, 4551 (1999).

Y. Bu and C. Deng, J. Phys. Chem. A, 101, 1198 (1997).

B. S. Brunschwig, J. Logan, M. D. Newton, and N. Sutin, J. Am. Chem. Soc., 102,
5798 (1980).

C. A. McDermott, K. R. Kneten, and R. L. McCreery, J. Electrochem. Soc., 140,
2593 (1993).

A. C. Ferrari and J. Robertson, Phys. Rev. B, 63, 121405 (2001).

A. C. Ferrari and J. Robertson, Philos. Trans. R. Soc. London, A, 362, 2477
(2004)

D. M. Gruen, MRSBulI., 23, 32 (1998).

Q. Chen, D. M. Gruen, A. R. Krauss, T. D. Corrigan, M. Witek, and G. M. Swain,
J. Electrochem. Soc., 148, E44 (2001).

L. Bergman and R. J. Nemanich, J. Appl. Phys, 78, 6709 (1995).

R. E. Shroder, R. J. Nemanich, and J. T. Glass, Phys. Rev. B, 41, 3738 (1990).

S. Prawer, K. W. Nugent, D. N. Jamieson, J. O. Orwa, L. A. Bursill, and J. L.
Peng, Chem. Phys. Lett., 332, 93 (2000).

J. Birrel], J. E. Gerbi, O. Auciello, J. M. Gibson, J. Johnson, and J. A. Carlisle,
Diamond and Rel. Mater., 14, 86 (2005).

142

40.

41.

42.

43.

G. G. Budaguan, A. A. Aivazov, and A. Y. Sazonov, Calorimetric Investigation
of Relaxation Processes in Disordered Semiconductors, 420 (Amorphous Silicon
Technology—1996, p. 635, Materials Research Society Symposium Proceedings,
Materials Research society, (1996).

T. Hoshino and Y. Nishioka, Etching Process of Al Oxide on Si Surface with HF
Treatment, 492(Microsc0pic Simulation of lnterfacial Phenomena in Solids
and Liquids), p. 201, Materials Research Society Symposium Proceedings,
Materials Research Society, (1998).

C. W. Lo, P. R. Varekamp, D. K. Shuh, T. D. Durbin, V. Chakarian, and J. A.
Yarrnoff, Surf Sci., 292, 171 (1993).

C. J. Wu and E. A. Carter, .1. Am. Chem. Soc., 113, 9061 (1991).

143

CHAPTER 5

Pulsed Galvanostatic Deposition of Pt Particles on
Microcrystalline and Nanocrystalline Diamond
Thin-Film Electrodes

5.1 Introduction

Carbon-supported metal catalysts are widely used in electrochemical
technologies. For example, dispersed Pt particles on high-surface-area carbon supports
are used as catalysts for hydrogenation and oxidation reactions, such as for both hydrogen
oxidation and oxygen reduction in phosphoric acid (PAFCs) and proton exchange
membrane fuel cells (PEMFCS).1’2 Platinum is the best electrocatalyst for the oxygen
reduction reaction (ORR); an important reaction in both biological systems and fuel

cells.3'l3

Among the carbon support materials regularly employed, activated carbons are
the most common. Activated carbons possess a turbostratic microstructure of spz-bonded
carbon with a high surface area and microporosity."2 The properties of the
electrocatalyst support material are crucial for the operational performance."3"“"l5 The
primary role of the support is to provide a high surface area over which highly dispersed,
nanometer-sized electrocatalyst particles can be dispersed and stabilized, and electrical
connection to these particles can be made. Highly dispersed supported catalysts are

attractive for electrochemical processes because a high electrochemically active surface

area can be achieved with a low loading of the normally expensive electrocatalyst.

144

As mentioned previously, current electrocatalyst support materials are generally
porous, spZ-bonded carbon materials (e.g., physically or chemically activated carbon or
carbon black). These materials firnction well in many applications but are susceptible to
oxidation and/or corrosion in the presence of reactive oxygen, in aggressive chemical
environments and/or under harsh electrochemical conditionsl’z‘” '16 If the degradation is
severe enough, complete electrode failure can result. However, even minimal
oxidation/corrosion causes morphological and microstructural damage that leads to
decreased catalytic activity due to catalyst detachment, aggregation and/or fouling by

gasification products.

Advanced electrocatalyst support materials are needed for those demanding
applications where conventional spZ-bonded carbon supports degrade. One example of
this is in fuel cells during start-up and shut-down as a local cathode potential as high as
1.2-1.5 V can be reached, which is a potential well positive of the thermodynamic value
for carbon corrosion.17 The rate of carbon corrosion increases considerably at such high
voltages and cumulatively this leads to degradation of the stack voltage. Our group has
been investigating the possibility of using electrically conducting diamond as an
advanced carbon electrocatalyst support material.18'2] There are two technical challenges
to be overcome in this effort: development of an electrically conducting, high surface

area form of diamond (e. g., powder)”23

and establishment of a protocol for depositing
and stabilizing highly dispersed electrocatalyst particles on the diamond surface. Thus
far, our efforts have been focused on the latter task with the goals of (i) determining the

dimensional stability of diamond thin film (low surface area) under a variety of

145

electrochemical conditions with19 and without (see Chapter 4) deposited metal particles
and (ii) studying the formation of electrocatalytic Pt adlayers on the film surface.20
Diamond possesses properties that make it an attractive support material: dimensional
stability, resistance to corrosion and the ability to operate for extended periods of time in
harsh chemical environments at high potential, current density, and temperature without

microstructural or morphological degradation.” '24

Two types of electrically conducting diamond thin film are under investigation:

microcrystalline (>1 mm crystallite size) and nanocrystalline (~10-20 nm crystallite

25-30

size). Each possesses a different morphology and microstructure, but both exhibit

similar electrical conductivity and electrochemical activity when highly doped with

boron. The formation of metal phases on polycrystalline diamond has been reported?“3 1'

35 A common observation, though, is weak attachment of the metal particles to the

hydrogen-terminated diamond surface.19’2°’3’3'34

Consequently, the metal particles are
easily dislodged from the surface by gas evolution, for instance. In an effort to better
stabilize the metal particles, we have adopted an approach whereby a short secondary
diamond growth is applied to the metal-coated surface. This secondary growth
physically entraps the metal particles into the surface microstructure without causing
major loss in the electrocatalytic activity. We refer to this material as a metal/diamond

composite electrode.”21

This sequential diamond growth/Pt deposition/secondary
diamond grth procedure produces a dimensionally stable and electrically conductive
catalytic electrode. The metal catalyst particles are strongly anchored in the

microstructure by the secondary diamond growth. For instance, no morphological or

146

 

microstructural damage and no lost catalyst activity were observed after two l-h periods
of anodic polarization in 85% H3PO4 at 170 °C and 0.1 A/cmz.19 In comparison, a Pt-
impregnated sp2 carbon cloth electrode (E-Tek) failed catastrophically within the first

few minutes of exposure to the same conditions.

We have previously studied the formation of Pt adlayers on diamond by dc
magnetron sputtering18 and electrodeposition, using both constant potential and current
deposition.‘9‘20 Electrodeposition is the preferred method of the two because it affords
better control over the particle size and coverage. Despite the excellent dimensional
stability of the Pt/diamond composite electrode, the nominal Pt particle size produced
was too large and the coverage too low to be practical. For example, the particle size
varied from 50 to 300 nm and the particle coverage was in the 108 cm'2 range for metal
phase formation at constant currentlg’20 Low loadings of highly dispersed metal particles
are desired for a practical supported catalyst. Pulsed galvanostatic deposition can
produce metal particles that are nominally smaller in size and more unifome distributed
over a surface compared to what is achieved using constant current deposition. Pulsed
electrodeposition approaches have been employed to prepare nanometer-sized particles of
metals and alloys.3547 Individual particle growth is favored as the conditions are selected
to minimize overlap of reactant depletion layers fiom neighboring particles on the
surface. Maintaining separate depletion layers around each particle results in similar
growth rates for all and leads to more highly dispersed particles of nominally smaller

size.

147

We report herein on the use of pulsed galvanostatic deposition to form Pt adlayers
on both microcrystalline (BMD) and nanocrystalline (BND) diamond thin films. This
work was a continuation of our effort to control and understand metal phase formation on
diamond. The purpose for the work was to learn how the nucleation and growth of Pt
electrocatalyst particles on both diamond types is affected by the pulse current density
and pulse number. In particular, the spatial distribution of the deposits over the surface,
the nominal particle size and variance, and the particle coverage were all factors of
interest. Another objective was to learn how the diamond film morphology and
microstructure affect the metal phase formation. Polycrystalline diamond is a
heterogeneous electrode material in terms of its microstructural and electronic properties,
which complicates the study of metal phase formation. The microstructural and
electronic properties certainly play a key role in the kinetics of metal electrodeposition
and the time-dependent surface morphology of the deposits. The uncoated and coated
diamond electrodes were characterized by scanning electron microscopy (SEM), cyclic

voltammetry, and conductivity-probe atomic force microscopy (CP-AF M).

5.2 Results

5.2.1 Pt Deposition on Microcrystalline Diamond

Pt metal was deposited onto the diamond surface by pulsed galvanostatic
deposition with the pulse width and current density maintained constant for 1 s at 0.50
mA/cm2 (geometric area), and a pulse number of 10, 25 or 50. This current was selected
as a starting point based upon previous work in the group .20 Figure 5.1A-C shows

secondary electron micrograph (SEM) images of Pt-coated microcrystalline diamond

148

thin-film electrodes for these three pulse numbers. EDX analysis confirmed that the
bright spots were indeed Pt. The particles decorate both the grains and grain boundaries,
indicating that most regions of the polycrystalline film are electrochemically active and
support charge transfer. This is somewhat in contrast to observations made recently
using CP-AFM and SECM.48 The formation of the multilayer islands or particles is
consistent with a Volmer-Weber mechanism of deposition, which incorporates the
postulate that deposition of metal on preexisting metal centers is more favorable than
metal deposition on a foreign substrate that exhibits a low surface free energy.”41
Increasing the number of pulses lead to an increase in the metal loading as more overall
deposition charge was passed; however, the nominal particle size increased and the
particle coverage decreased. These trends are evident in the Table 5.1 data. Depositing
Pt with 10 pulses yielded a nominal particle diameter of 50 1 30 nm [relative standard
deviation (RSD) = 60%] and a nominal particle coverage of 1.1 (1 0.6) x 109 cm'z.
Increasing the number of pulses to 25 approximately doubled the average particle
diameter to 100 1 90 nm (RSD = 90%) and reduced the particle coverage to 1.0 (1 0.4) x
109 cm'z. Increasing the number of pulses to 50 slightly increased the average particle
diameter to 110 1 70 nm (RSD = 64%) and significantly reduced the particle coverage to

0.38 (1 0.01) x 109 cm’z. Careful examination of the images, in particular Fig. 5.1C,

reveals that the particles are not smooth but rather are rough with a dendritic morphology.

149

 

Figure 5.]. Scanning electron micrographs of microcrystalline diamond electrodes
covered with Pt particles that were deposited using (A) 10, (B) 25 and (C) 50 pulses,
respectively. Pulse width = l 5. Current density = 0.50 mA/cmz.

150

Figure 5.2 shows background cyclic voltammetric i-E curves in 0.1 M HClO4 for
the Pt-coated microcrystalline thin-film electrodes formed with different pulse numbers.
The scans were initiated at 200 mV with the forward sweep in the positive direction at a
rate of 50 mV/s. All the voltammetric features characteristic of clean polycrystalline Pt
was observed. The total current response increased with pulse number, which is
consistent with an increase in the Pt loading; more specifically an increase in the
electrochemically active Pt. Even though the nominal particle diameter increased and the
coverage decreased, the electrochemically active Pt increased because of a highly rough,

dendritic particle morphology, as is evident in Figure 5.1C.49

 

 

     

 

 

 

 

 

 

150
"A
E 100 ~
3
:1.
V 50 "
2?
m
a 0 _
D
H
= l
3 -so
'5 F -- 10 pulses
U i m 25 pulses
-100 ' '''' 50 pulses
’150 l 1 I 1 l 1 L 1
-200 200 600 1000 1400

Potential (mV vs. Ag/AgCl)

Figure 5.2. Cyclic voltammetric i-E curves in degassed 0.1 M HC104 for
microcrystalline electrodes coated with Pt using different pulse numbers. Pulse width = 1
s. Current density = 0.50 mA/cmz. Scan rate = 50 mV/s. The shaded region of the curve
represents the charge associated with hydrogen adsorption. The current is normalized to
the diamond geometric area, 0.2 cm2.

151

 

Table 5.1. Particle Analysis and Hydrogen Adsorption Data for Pt—Coated
Microcrystalline Diamond Films as a Function of the Pulse Numbera

 

 

Mean Particle Size Particle Coverage Hydrogen Adsorptron
N o. of Pulses (nm) (109 cm'2)" Charge
(mC/cmz)b
10 50130 1.1106 0.071001
25 100190 1,010.4 0.131001
50 110 1 70 0.38 1 0.01 0.21 1 0.02

 

 

 

a Pulse width = l 5. Delay time = 1 5. Current density = 0.50 mA/cmfigeometric area).
b Values normalized to the geometric area of the diamond film (0.2 cm )

The well-defined voltammetric signatures for Pt indicate that the metal particles
are in good electrical communication with the current-collecting p-Si substrate through
the boron-doped diamond film. The H+ adsorption charge density, listed in Table 5.1,
was found by integrating the cathodic current for this reaction between 100 and -190 mV
vs. Ag/AgCl and correcting for the double-layer charging current (see the shaded region
of the i-E curve in Figure 5.2). The electrochemically active Pt surface area (ESA),
displayed in Table 5.2, was calculated to be 0.06, 0.13 and 0.20 cmzpt, respectively for
10, 25 and 50 pulses. All electrodes possessed the same deviation of 1 0.02 Cl’l’lzpt.
Consequently, the roughness factor [RF, the electrochemically active Pt area/geometric
area of the diamond electrode (0.2 cm2)] increased with pulse number, as is shown in
Table 5.2. It appears that, once formed, the deposition of Pt on existing Pt clusters occurs
more facilely than on the bare diamond surface and this leads to a greater microscopic
surface roughening with increasing pulse number. The loading also increased with pulse
number and this was confirmed by AAS (described in Chapter 2) with calculated values
of 2.4, 3.3 and 7.4 ugpr/cm2 measured for just one film at each at pulse numbers of 10, 25

and 50. An important performance parameter for electrocatalytic electrodes is the

152

specific Pt surface area, defined as the electrochemically active Pt area normalized to the
Pt loading. The specific surface area was calculated to be 13, 20 and 14 mzpr/gpr,
respectively, and is shown in Table 5.2. This specific surface area, ~15 mzpt/gpr, is about
a factor of ten lower than what is necessary for a fuel cell electrode. This is largely due

to the low surface area of the planar thin-film electrode.

 

Table 5.2. Electrochemically Active Pt Surface Area, Roughness Factor (RF) and
Specific Pt Surface Area for Pt—Coated Microcrystalline Diamond Films as a Function of
the Pulse Numberal

 

 

Electrochemically Active Pt Specific Pt
No. of Pulses Surface Area RF Surface Area
(cmzn) (mth/gPt)
10 0.06 1 0.02 0.31 13
25 0.13 1 0.02 0.64 20
50 0.20 1 0.02 0.99 14

 

 

 

“ Pulse width = l s. Delay time = 1 5. Current density = 0.50 mA/cmY (geometric area).
RF is the electrochemically active Pt area obtained from the hydrogen adsorption charge
normalized to the geometric area of the diamond film (0.2 cmz).

The electrodeposition of Pt on microcrystalline diamond was also investigated as
a function of the pulse current density from 0.50 to 1.50 mA/cm2 (geometric area). The
deposition was canied out using a pulse number of 10 and a pulse width of l s.
Increasing the current density has the effect of increasing the overpotential (n) for the
reduction reaction, although in this case the actual values are not known. SEM images of
metal-coated films are displayed in Figure 5.3A-E. Table 5.3 presents a summary of the
particle analysis data. Statistically, the nominal particle size was the same for all five
current densities, ranging from 40 to 70 nm with an RSD of approximately 50%. The

particle coverage increased from 1.1 (1 0.6) to 7.5 (1 0.9) x 109 cm'2 with current density

153

up to 1.25 mA/cmz. However, the particle coverage for the highest current density, 1.50
mA/cmz, decreased to 1.9 (1 0.9) x 109 cm'z. The film deposited at 1.25 mA/cm2 has
particles that are nominally the same size as those produced at 0.50 mA/cmz, but with a

particle density nearly six times greater.

 

Figure 5.3. Scanning electron micrographs of Pt-coated microcrystalline diamond
electrodes deposited at pulse current densities of (A) 0.50, (B) 0.75, (C) 1.00, (D) 1.25
and (E) 1.50 mA/cmz. Pulse number = 10. Pulse width = l s.

154

 

Table 5.3. Particle Analysis and Hydrogen Adsorption Data for Pt-Coated
Microcrystalline Diamond Films as a Function of the Pulse Current Densitya

 

 

Pulse Current Mean Particle Size Particle Coverage Hydrogen Adsorption
Densr (nm) (109 cm'z)” Charge

(mA/cm )b (mC/cm2)"
0.50 50 1 30 1.1 1 0.6 0.07 1 0.02
0.75 70 1 40 1.5 1 0.4 0.095 1 0.007
1.00 60 1 30 2.2 1 0.6 0.112 1 0.005
1.25 50 1 30 7.5 1 0.9 0.156 1 0.005
1.50 70 1 40 1.9 1 0.9 0.19 1 0.03

 

 

 

’ Pulse width = 1 5. Delay time = l 8. Current density = 0.50 mA/cm2 (geometric area).
b Values normalized to the geometric area of the diamond film (0.2 cm )

Figure 5.4 shows background cyclic voltammetric i-E curves in 0.1 M HC104 for
the Pt-coated microcrystalline thin-film electrodes formed with different pulse current
densities. The scans were initiated at 200 mV with the forward sweep in the positive
direction at a rate of 50 mV/s. Again, all the voltammetric features characteristic of clean
polycrystalline Pt are present. There is a progressive increase in the total voltammetric
current as the deposition current density is increased from 0.50 to 1.50 mA/cmz,
consistent with an increasing electrochemically active metal phase. The charge
associated with the H+ adsorption increases with deposition current density, as shown in
Table 5.3. This is due to an increase in particle coverage, as is observed for the films

coated at current densities from 0.50 to 1.25 mA/cmz, and particle roughening.

Table 5.4 shows the estimated metal loading, the electrochemically active Pt area,
the specific Pt area and the roughness factor for the Pt-diamond electrodes formed using

different pulse current densities. The electrochemically active Pt area and roughness

155

factor both increase with deposition current density, as reflected in the voltammetric i-E
curves. Thisiis due to the increase in particle coverage (at least up to 1.25 mA/cmz) and
microscopic roughness of the particles. The measured metal loading increased from 2 to
4 light/cm2 with deposition current density. The specific Pt surface area ranged from 10

to 25 mzpr/gp, with the values clustered around 20 mzpt/gpt.

17S

 

125 '

75 ”

 

Current Denisty (uA/cmz)

-25 r "Si’fiugrfi‘xx

 

 

 

 

 

 

 

 

-75 - l
-125 1 l 1 l L L
-300 200 700 1200

Potential (mV vs. Ag/AgCl)

Figure 5.4. Cyclic voltammetric i-E curves in degassed 0.1 M HClO4 for
microcrystalline diamond electrodes coated with Pt at the different current densities.

Pulse number = 10. Pulse width = l s. Scan rate 50 = mV/s. The current is normalized to
the diamond geometric area, 0.2 cm2.

156

 

Table 5.4. Electrochemically Active Pt Area, Specific Pt Surface Area and Roughness
Factor for Pt-Coated Microcrystalline Diamond Films Deposited at Different Pulse
Current Densities“

 

 

 

Pulse Current Measured Electrochemically Specific Pt
Densi Loading Active Pt Surface Area RF Surface Area
(mA/cm )" (light/ml)" (cmzpr) (fin/gr.)
0.50 2.4 0.06 1 0.02 0.31 13
0.75 2.3 0.090 1 0.007 0.45 20
1.00 2.2 0.107 1 0.005 0.53 25
1.25 3.4 0.149 1 0.004 0.74 22
1.50 4.3 0.18 1 0.04 0.89 21

 

 

“ Pulse number = 10. Pulse width = l 3. Delay time = 1 8. Loading estimated from AAS
measurements of the dissolved metal deposits.
b Normalized to the geometric area of the diamond film (0.2 cmz).

5.2.2 Pt Deposition on Nanocrystalline Diamond

Pulsed galvanostatic deposition was also used to form Pt adlayers on boron-doped
nanocrystalline diamond electrodes. Metal deposition is a surface "sensitive" charge-
transfer process; therefore, the morphology and microstructure of a substrate electrode
can influence the metal phase nucleation and growth. Comparing the deposition at these
two diamond types provided insight about the influence of the film microstructure and

morphology (e. g., grain boundaries, defects, facets, etc.) on the metal phase formation.

Two deposition conditions were evaluated, ten l-s pulses at 0.50 and 1.25
mA/cmz. SEM images of these two metallized surfaces are shown in Figure 5.5A (0.50
mA/cmz) and B (1.25 mA/cmz). Nanocrystalline diamond has a much smoother surface
morphology than does microcrystalline diamond and a larger fraction of exposed grain

boundary due to the smaller aggregate size.28‘29 Though it is currently not as corrosion

157

resistant as microcrystalline diamond, it still is an improvement over sp2 carbon
materials. The morphology consists of ca. 50-100 mm aggregates of diamond crystallites
that are 10-20 nm in size. We supposed that the higher fraction of grain boundary might
lead to a higher nucleation site density for Pt and, in turn, to smaller particles and higher
coverages. It has been observed in other work in our group that this is in fact the case, as
nanocrystalline outperforms microcrystalline diamond for the anodic stripping
voltammetric determination of dissolved metal ions, in part, because metal phase
formation on the former occurs with lower particle size and higher coverage.50 It can be
seen in the images that many of the metal deposits are near the grain boundaries. Particle
size analysis of the film in Figure 5.5A revealed a nominal diameter of 70 1 40 nm (RSD
= 57%), which is statistically similar to the particle size seen for microcrystalline
diamond at the same current density. The particles are not as unifome distributed over
the surface as they are on the microcrystalline diamond surface, at least under these
deposition conditions. The uneven distribution is reflected in a lower particle coverage of
6 (1 5) x 108 cm’z, which is an order of magnitude less than that seen for the

microcrystalline film.

The metal particles formed at 1.25 mA/cm2 however, shown in Figure 5.58, have
a smaller nominal size of 30 1 30 nm (RSD = 100%). The Pt particles are also more
unifome distributed over the surface and this is reflected in a larger particle coverage of
1.9 (1 1.0) x 1010 cm'z. Many of the metal particles appear to reside at the grain

boundaries, which is consistent with these regions being the primary nucleation sites.

158

This particle coverage is the highest observed for any of the diamond films under study,

microcrystalline or nanocrystalline.

 

Figure 5.5. Scanning electron micrographs of Pt-coated nanocrystalline diamond
electrodes deposited at pulse current densities of (A) 0.50 and (B) 1.25 mA/cmz. Pulse
number = 10. Pulse width = 1 s.

159

...i ..... 3...... : il.‘...,. 1 H .. .r...r...: ,,V. .....
. 12,... . .12.... , , .. 5.11:: . 1.1.3... .:. :1. .; . .1 .. :.. . . .o.....,._:. .1. . .. .1... 1.... t ..

:2.

11:13.1. . 1 .3», 1.3.0.9.. . n

 

en... . J. . .11,‘fq..a...m..o.. 1.1.11.3?— J

l.........c.

 

.
o'er:’109.‘
r #1... .

.

 

Figure 5.6 shows background cyclic voltammetric i-E curves in 0.1 M HClO4 for
nanocrystalline diamond thin-film electrodes decorated with Pt that was deposited at 0.50
and 1.25 mA/cmz. The scans were initiated at 200 mV with the forward sweep in the
positive direction at a rate of 50 mV/s. Again, all the voltammetric features characteristic
of clean polycrystalline Pt are present. There is an increase in the total voltammetric
current for increased deposition current density. This is consistent with the SEM data
and results from an increased Pt loading. It is worth noting that the nanocrystalline
diamond electrode decorated with Pt at 1.25 mA/cm2 shows much larger voltammetric
currents than does the microcrystalline diamond electrode coated at the same current
density. This is due to the smaller nominal particle size and larger particle coverage on
the former. Interestingly, the i-E curve for the film coated at 1.25 mA/cm2 exhibits a pre-
peak prior to Pt oxide formation at ca. 700 mV. This pre-peak is associated with a PtxOH
species formed prior to the Pt0 monolayer in the plateau of the curve.5 1‘52 This peak may
be due to Pt40H formation on the (100)—oriented surfaces and/or Pt3OH formation on the

(l l 1)-oriented surfaces.52

The H+ adsorption charge for the 0.50 and 1.25 mA/cm2 films is 0.071 1 0.010
and 0.25 1 0.027 mC/cmz, respectively. The electrochemically active Pt area for the
films is 0.07 1 0.01 (RSD = 15%) and 0.24 1 0.03 (RSD = 13%) cmzpr, respectively.
These values result in roughness factors of 0.34 and 1.2. Calculations, based on AAS
data revealed approximate Pt loadings of 1.4 and 2.4 pgpr/sz for the 0.50 and 1.25

mA/cm2 deposition currents, respectively. This translates into specific Pt surface areas of

160

24 and 50 mzpr/gpt, respectively, which are the highest values we have recorded. It
appears from these preliminary studies that the diamond film morphology and

microstructure do indeed influence the nucleation and grth of Pt, and the highest

specific areas are seen for the nanocrystalline films

 

 

 

 

 

 

 

 

 

 

 

150
NA 0.

E
i 75 '- .,’.“‘ '0' s. ........... .o' '0’

:r. i i. .-'
v ‘\ '0‘ 0
£3

8 0 ”

o

G .

a ,' s

= .0 ‘0‘

2 '75 '- 0". o. ‘. '0'

'5 t I. ‘ ‘1 " — '2

."o r. t' 0.50 mA cm
U i t. o: "" 1.25 mA out2
-150 - X,"
l 1 I 1 l 1 l 1
-200 200 600 1000 1400

Potential (mV vs. Ag/AgCl)

Figure 5.6. Cyclic voltammetric i-E curves in degassed 0.1 M HC104 for nanocrystalline
diamond electrodes coated with Pt deposited at 0.50 and 1.25 mA/cmz. Pulse number =

10. Pulse width = l 8. Scan rate = 50 mV/s. The current is normalized to the diamond
geometric area, 0.2 cm2.

5.2.3 Comparison Studies with Glassy Carbon

In order to assess how metal deposition on diamond stacks up against deposition
on sp2 carbon electrodes, tests were conducted with glassy carbon. Based on the results
discussed previously, the following conditions were selected for Pt deposition: Ten

pulses of 1 s duration (delay time of l s between pulses) at 1.25 mA/cmz. SEM was used

161

to evaluate the metal deposits. The micrograph is shown in Figure 5.7. Particle size
analysis revealed a particle diameter of 70 1 60 nm (RSD = 85%) and a particle coverage
of 5 (1 l) x 109 cm'z. Cyclic voltammetry was employed, in the same manner as with the
Pt/diamond electrodes, to study the Pt activity for H" adsorption in 0.1 M HClO4.
Integration of the i-E curves revealed an H+ adsorption charge of 0.25 1 0.02 mC/cm2 (n
= 3). This charge is similar to that seen for the nanocrystalline film and is a little larger
than the value for microcrystalline diamond. The electrochemically active Pt area was
calculated to be 0.24 1 0.02 cmzm. In summary, the metal phase formation on glassy
carbon under these deposition conditions appears similar to what is seen for both
diamond types in terms of the nominal particle size, coverage, and electrochemically

active Pt area.

 

 

Figure 5.7. Scanning electron micrograph of Pt-coated glassy carbon deposited using 10
pulses of 1 3 width at a current density of 1.25 mA/cmz.

162

The specific Pt surface areas on diamond were in the range of 10-50 mzpt/gpr (see
Tables 5.2 and 5.4). These values are about a factor of 210 less than that necessary for a
typical fuel cell electrode, which has an area of 100-200 mzpr/gpr. Gloaguen et al.
reported values of 15-73 mzpr/gpr for adlayers on three different types of carbon
powders.53 It is important to remember that the specific Pt surface area depends on the
surface area of the carbon support. Kucemak et al. reported values of 20-50 mzpr/gpr for
electrodeposited adlayers on highly ordered pyrolytic graphite (HOPG).54 Mo et al.
reported on the immobilization of high-area Pt particles on glassy carbon with specific
surface areas of ca. 30 mzpt/gpr.55 It is clear that the deposition on diamond yields
electrocatalyst structures and activities that are similar to those for other sp2 carbon

support materials.

5.2.4 Investigating Pt Nucleation and Growth Mechanism

In order to control the metal particle properties one must gain a better
understanding of the nucleation and growth mechanism of Pt on the diamond substrate.
Unlike most materials used to study metal nucleation, diamond is a heterogeneous
material both morphologically and electronically. Figure 5.8 shows cyclic voltammetric
i-E curves for the deposition of Pt on four different carbon surfaces: microcrystalline and
nanocrystalline diamond, GC and HOPG. The solution consisted of 1 mM KthCls
dissolved in 0.1 M H00; and the potential was scanned at a rate of 20 mV/s. The scans
were initiated at 1 V and scanned negative with a Pt nucleation peak appearing between -

100 and 100 mV (peak I), followed by the onset of hydrogen adsorption at ca. -200 mV

163

(peak 11) on the deposited particles. On the subsequent scan, the onset of peak 111 is

observed at ca. 900 mV. This is attributed to the formation of oxides on the Pt layer.

 

 

 

 

 

"I
Q_—_
4o~ l BMD
~80’ u
or ‘—
l
.40- I BND
A
3..-
E l1[!1l l 1111111 l 1111
q, , II
In
t- 0‘
a . *—
U

e
at
n

 

 

-80 r
II
III
4.—
1 /
«to . 1 HOPG
-80 ~

 

 

1111 1 1 1 l n 1 A

-400 -200 o 200‘ i 4460“ 666 i 800‘ 4 i000
Potential (mV vs. Ag/AgCl)

 

Figure 5.8. Cyclic voltammetric i-E curves for 1 mM K2PtCl6 + 0.1 M HClO4 on
different carbon materials. Scan rate = 20 mV/s.

164

The i-E curve for HOPG, the most ordered material of the four, is quite different
from the other curves. This material had the most negative reduction curve for the
nucleation of Pt and is the only curve for which the reverse scan was at a lower current
(see region between 0 and 200 mV) than the forward scan. This is consistent with the
small amount of edge plane exposed on the freshly cleaved HOPG. The edge plane is

known to be more active towards the formation of metal adlayer than the basal plane.39

The other three carbon materials exhibited a nicely resolved reduction peak
positive of 0 mV indicating metal adlayer formation on the surface. There was however,
more variability in the reduction peak potential for these materials than observed for the
HOPG. The GC exhibited a deviation of ca. 50%, while the deviation for both diamond
types was >100%. This increase in deviation is likely due to an increase in the disorder
in the carbon with the diamond materials being the most heterogeneous. It appears from
the curves that an overpotential of approximately -500 mV is required to deposit Pt on the

diamond surfaces.

A series of chronoamperometric steps were performed with both diamond types in
an attempt to investigate the Pt nucleation mechanism as discussed in Chapter 1. The
potential was held at 800 mV, a potential at which no Pt deposition occurs, and stepped to
different final potentials at which deposition occurred. The background subtracted i-t
curves for the microcrystalline and nanocrystalline diamond is presented in Figures 5.9

and 5.10, respectively. Note that the potentials indicate the overpotential for the reaction

(11 = Eeq — EFinal)-

165

 

 

 

 

 

 

 

 

 

 

 

2.0
L -1012
1.5 r
. -912
2 .
E - -812
E 1.0 _
r... -712
h
= .
U 0.5 - -612
i , '512-412 W _
0.0 / -312
_0.5 1 l J I 1 I 1 I 1
0 2 4 6 8 10
Time (ms)

Figure 5.9. Background subtracted chronoamperometric i-t curves for Pt deposition on

microcrystalline diamond thin film in 1 mM KthCl6 + 0.1 M HClO4 at the overpotentials
indicated in mV.

2.0

 

-1012

I"
U!

i"
c

P
u-

 

Current (mA)

9
:

 

.'=
Ur
r

 

 

J1 1 1l 1
-100 l l

 

Time (ms)
Figure 5.10. Background subtracted chronoamperometric i-t curves for Pt deposition on

nanocrystalline diamond thin film in 1 mM K2PtC16 + 0.1 M HClO4 at the overpotentials
indicated in mV.

166

As expected, overpotentials less than -500 mV did not cause significant
deposition current to flow as seen in Figure 5.9 for overpotentials of ~312 and —412 mV.
This is consistent with that predicted from the cyclic voltammetric i-E curves in Figure
5.8. As the overpotential increases, the resulting current maximum (1m) increases and is
reached at shorter times (tm) consistent with an increase in the amount of metal deposited
and the rate at which it is formed. Following the maximum, the current decreased as
depletion layers around each particle start to overlap resulting in a planar diffusion

pattern to the electrode.

The i-t transients were analyzed according to the theory described by Schariflrer
and Hills presented in section 1.4. Theoretical curves representing the limiting cases for
instantaneous and progressive nucleation mechanisms were constructed based on
equations 1.4.] and 1.4.2, and are the same as those presented in Figure 1.7. The
normalized i-t transients at the different overpotentials for both diamond types are
presented in Figure 5.11A-F. As the overpotential increases, the nucleation on both
diamond surfaces transitions fiom an instantaneous nucleation mechanism towards a
more progressive one. This is similar to what was found for Cu deposition on
microcrystalline diamond.56 Interestingly, the curve for the nanocrystalline diamond is
always closer to the theoretical curve for progressive nucleation than is the curve for
microcrystalline diamond. This perhaps explains the greater dispersity in particle size
observed for the Pt/BND films compared to the Pt/BMD films as progressive nucleation

generally leads to an increase in particle size dispersion.

167

 

0 Microcrystalline

El Nanocrystalline

 

 

 

 

 

 

 

 

 

 

 

 

1.2
l 0 Instantaneous
N? 0.8 f. "a
S 0.6 ‘ p i S
V rogress ve V
0.4 ~'
0.2 A
0..%_1111111.L111111
0 0. 5 1.0 1.5 2. 0 2. 5 3. 0 3. 5 4. 0
t/tm
1.2
’ Instantaneous
1.0 ¥ J
N 0.8 i” hon“ N
’E L 5?: “'0 ”g
E 0.6 . é Mb%:“‘.‘ 5
5 ’ Progressive V
0.4 5:1
"2 1" C
0.15 ‘ ‘ J 4 ’ L *
0 0. 5 1.0 I .5 2. 0 2 .5 3. 0 3. 5 4. 0
at...
1.2
I 0 L Instantaneous
N 0.8 ’ N
’2 * ’E
a 0.6 ’ E

0.4 r
0.2

021.0 051.0 1.5 2.0 2.5 3.0 3.5 4.0

mm

 

 

 

 

Figure 5.1]. Comparison of experimental normalized i-t curves for microcrystalline (a)
and nanocrystalline (El) diamond with the theoretical curves corresponding to
instantaneous and progressive nucleation at overpotentials of (A) -512, (B) -612, (C) -

712, (D) -812, (E) -912 and (F) -1012 mV.

168

1.2

g
0.8 ’
r. ;
0.4

0.2

0.1),.

 

Instantaneous

 

 

 

 

o 0.5 A 1.0 I 1.5 A 2.0 A 2.5 3.13.3411

 

 

 

 

 

 

 

 

t/tm
1.2
1.0 _ Instantaneous
0.8L 4%?“
’ “‘9. at...
0.6 - 1‘ h. _“r
0.4 L PTWSNC “‘0‘.“ 9'01“:
0.2 g D
0.1)) " AAAAAA
0 0.51.015 2.0 2.5 3.0 3.5 4.0
mm
1.2
1.0 _ Instantaneous
0.8 7 .
' \‘Ira...
0.6 ' n”- T-‘roh
0.4 _ Progressive "
p “‘0‘" l
0.2 F ‘
" :1
0.% i 11111111111111
.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
mm

As shown in Figure 5.11, the normalized i-t curve rarely follows one of the
limiting cases perfectly. These limiting cases are for primary nucleation and deviation
from these curves is generally attributed to secondary nucleation, however this could also
be attributed to a mixture of mechanisms across a heterogeneous surface.57 However, in
Scharifl<er and Mostany's approach, it is not necessary to classify the nucleation by one of
the limiting cases.58 Instead, the values of 1m and tm can be used to calculate two
parameters: the nucleation rate per active site on the surface, A, and the number density
of active sites for nucleation, No. We attempted to calculate these values, however,

according to this approach, the value for Imtml/Z/a (where a = szI/ZC/rt”2

) needs to fall
between 0.7153 (instantaneous) and 0.9034 (progressive). Only two of our
chronoamperometric i-t curves fit this criterion. They were for overpotentials of -912 and
-1012 mV at the nanocrystalline diamond. For these transients, the values for A were
931 and 28.7 s'1 and for No were 3.5 x 109 and 1.3 x 1011 cm'z, respectively. All the other
curves exhibited Imtml/z/a values that were too low, consistent with low values of 1m
and/or tm. It appears that while the rate of Pt nucleation on diamond is quite fast, there is
not enough metal deposited to achieve considerable current to satisfy this approach.
However, the difference between the values of A and No found for the 2 nanocrystalline

films is consistent with an increase in both the deposition rate and number of active sites

with increasing overpotential as would be expected.

5.2.5 Secondary Diamond Growth

The effect of a secondary diamond growth on the activity and stability of the Pt
nanoparticles on microcrystalline diamond was also investigated, and is illustrated in

Figure 5.12A-C. Figure 5.12A shows an SEM image of a microcrystalline diamond thin-

I69

 

film electrode freshly coated with Pt using ten l-s pulses at a current density of 1.25
mA/cmz. The SEM image reveals good metal coverage over the entire diamond surface.
A nominal particle size of 30 1 30 nm and particle coverage of 6 (1 2) x 109 cm'2 were
determined from particle size analysis. No loss of Pt activity was observed after long-
term solution exposure. The initial hydrogen adsorption charge in 0.1 M HC104 was 0.21

mC/cmz. This corresponds to an electrochemically active Pt area of 0.20 cmzpt.

The metal-coated film was then subjected to a 20 min diamond deposition using a
0.5% CH4/H2 source gas mixture at 1 kW, 200 sccm of total gas flow, 40 Torr, and ca.
800 °C. These conditions are not necessarily optimal but rather were selected as a
starting point. Figure 5.12B shows an SEM image of the Pt-coated electrode
immediately following the secondary diamond growth. A nominal particle size of 40 1
40 nm and particle coverage of 4 (1 l) x 109 cm"2 was observed. The slightly larger
nominal particle size and lower particle coverage are consistent with some coalescence of
neighboring particles as the deposition temperature is ca. 800 °C. More importantly, the
particle shape changed from a rough, dendritic morphology to a smoother, more rounded
morphology during the plasma treatment as a result of annealing effects. This is

important as smoothening of the particle surface results in a decrease in surface area.

After the secondary diamond growth, background cyclic voltammetric i-E curves
were re-recorded in 0.1 M HClO4. The hydrogen adsorption charge decreased to 0.015
mC/cm2 (a factor of 10 loss), and this corresponds to an electrochemically active Pt area

of 0.014 cmzpr. There is clearly some loss in Pt activity due to particle coalescence,

170

annealing of the surface roughness and blockage by the diamond deposition. We also
suspect that some percentage of the loss seen here resulted not from "buried" particles but
rather from poor alignment of the metallized electrode area inside the O-ring of the cell.
The Pt exposed to the solution was anchored strongly and exhibited good stability with
the surface area during extended potential cycling (1000 cycles) between -0.4 and 1.9 V
vs. Ag/AgCl in 0.1 M HClO4. The maximum currents were near 1 5 mA/cm2 at the two
potential extremes. The hydrogen adsorption charge actually increased slightly,
exhibiting a value of 0.020 mC/cm2 afier cycling likely due to removal of carbon
contaminants from the surface and particle roughening. This correlates to a calculated
electrochemically active surface area of 0.019 cmzpt, or an increase of 31%. In
comparison, over 63% of the electrochemically active Pt was lost from a metal-coated
diamond surface exposed to the same cycling conditions when no secondary diamond
growth was applied. The SEM image of the Pt/diamond composite electrode following
the extensive electrochemical treatment is shown in Figure 5.12C. There is no visible
change compared to the image taken prior to potential cycling. A nominal particle size of
50 1 40 nm and a particle density of 3 (1 l) x 109 cm“2 was determined. Both of these
values are comparable to those prior to potential cycling supporting our assertion of
particle stability. The particle size, particle density, and calculated electrochemically
active area for the Pt/diamond film as-deposited, after secondary diamond growth, and
after extensive potential cycling are summarized in Table 5.5. Research is ongoing to
optimize the secondary diamond grth conditions in order to maximize the particle
stability while minimizing the extent of Pt activity loss. One possible alternative is a very

short nanocrystalline diamond growth procedure to anchor the metal particles to either a

171

 

.1... o q .. u.«..

1|:

:21... .

¢-.1—.

1.. u o 1:.—

.. . 1. a.

11- .1

 

microcrystalline or nanocrystalline diamond surface without significantly covering the

smaller catalyst particles.

 

Figure 5.12. Scanning electron micrographs of Pt—coated microcrystalline diamond
electrode (A) before, (B) after a 20 min secondary microcrystalline diamond growth and
(C) the film with the secondary diamond grth after extensive potential cycling (1000
cycles between ca. 15 mA/cm2 (geometric area)). Pulse number = 10. Current density =
1.25 mA/cmz. Pulse width = l s.

172

 

Table 5.5. Particle Analysis and Electrochemically Active Pt Area for Pt-Coated
Microcrystalline Diamond Film Subjected to 20 min Secondary Diamond Growth
Procedure and Extensive Potential Cycling

 

 

Mean Particle Particle Electrochemically
Pt/diamond Size coverage Active Pt Area
(um) (10’ cm") (cm’ro
As-deposited 30 1 30 6 1 2 0.20
After 2° BMD Growth 40 1 40 4 1 1 0.014
After Potential Cycling 50 1 40 3 1 1 0.019

 

 

 

 

5.3 Discussion

The results presented herein demonstrate that pulsed galvanostatic deposition is a
useful method for depositing nanometer-sized particles of Pt on diamond electrode
surfaces. In terms of the nominal particle size, particle coverage, and electrochemically
active Pt, pulsed deposition is superior to constant current deposition.19'2‘ The optimum
conditions for depositing Pt on both types of diamond were ten l-s pulses at 1.25 mA/cm2
with a 1 s delay between each pulse. The nominal particle size of ca. 30 nm and particle
coverage of ca. 10lo cm'2 seen for the nanocrystalline film are good values for a
supported electrocatalyst. These values are also comparable to what was achieved with
glassy carbon, indicating that metal deposition on diamond occurs in a similar manner to

that on 3 sp2 carbon electrode.
The basic objective of pulsed deposition is to reduce the particle grth rate,

which is expected to reduce the interactions between closely space ones. While we did

not monitor the deposition and growth of the metal particles in situ, there are some trends

173

in the data that shed light on the mechanistic aspects of the deposition. The nominal
particle size increased and the particle coverage decreased with increasing pulse number
at a fixed deposition current density (i. e., constant 11). It appears from our results that the
initial nucleation occurs at a fixed number of sites and that new Pt particles nucleate and
grow primarily on existing Pt rather than forming at new sites on the diamond surface.
This explains the rough, dendritic morphology of the deposits. With longer pulse times,
and therefore more charge passed, the growth of neighboring nuclei eventually becomes
coupled. This results in particle coalescence, reduced particle coverage and an increased

nominal particle size.

Increasing the deposition current density (i.e., increasing 1]) produced increased
particle coverage up to the 1.25 mA/cm2 level, at which point particle coalescence began
to occur. It is possible that the increased current density activates areas of the electrode
surface that were not active for metal phase formation at lower current density. The
observation is consistent with nucleation via a more progressive mechanism, in
agreement with the chronoamperometric i-t curves. Growth of new Pt particles occurred
on both existing Pt sites and at sites on the diamond surface. We have previously found
using EC-AFM and i-t transient measurements, that the nucleation of Cu on
microcrystalline diamond follows an instantaneous mechanism at low overpotentials (n <
500 mV) and shifts to a progressive mechanism at higher overpotentials.“ This is in
agreement with that found here for Pt deposition on microcrystalline and nanocrystalline
diamond thin film. Additionally, Vinokur et al. studied the electrodeposition of Ag and

Hg on diamond using i-t transients and found that the mechanism of nucleation depended

174

on the overpotential and the metal ion concentration}4 At low overpotentials,
instantaneous nucleation was observed and at high overpotentials or metal ion
concentrations the process became progressive. Preliminary results using EC-AFM and
i-t transients have revealed that the electrodeposition of the more electronegative Zn on
microcrystalline diamond follows a more progressive nucleation mechanism at both low

and high overpotentials.56

Finally, it is useful to consider what challenges exist for electrochemically
producing a dimensionally uniform metal adlayer on polycrystalline diamond. The
condition of the deposition method certainly will impact the dimensional uniformity of
the metal deposits, as does the heterogeneity of the electrode's microstructural and
electronic properties. Electrically conducting diamond is a heterogeneous electrode
material, both microstructurally and electronically. The polycrystalline films possess
diamond crystallites or grains, grain boundaries, and a multitude of other defects (e.g.,
dislocations, step edges, etc.). The microcrystalline films are rough with grain
boundaries between the crystallites that could serve as sites of higher coordination for the
metal phase, whereas the nanocrystalline films are smoother but possess a high fraction
of exposed grain boundary due to the small crystallite size. We suppose that the grain
boundaries are the primary nucleation sites initially and, therefore, the high grain

boundary density leads to a smaller nominal particle size and larger particle coverage.

The electronic properties of diamond also play a key role on the dispersion and

uniformity of metal adlayers. The electronic properties of diamond are controlled in a

175

complex manner by (i) the boron doping level and distribution throughout the film,”64
(ii) incorporated hydrogen which can act both as an acceptor and provide a charge carrier
or as a donor and passivate a boron acceptorf’s’69 and (iii) defect density.64’7°’71 Of these,
the boron doping level and distribution are probably the most influential. Furthermore, it
is known that boron incorporates into polycrystalline diamond with different
concentrations depending on the growth sector.72‘73 The boron concentration in the {111}
growth sector is five or more times greater than that found in the {100} sector. In other
words, polycrystalline diamond, particularly the microcrystalline films, is composed of
sites of faster (reversible) kinetics where the dopant level is high and sites of slower
kinetics (less reversible) where there is less dopant. We have recently observed such
electrical and electrochemical heterogeneity using CP-AFM mapping and scanning

48’” Similar

electrochemical microscopy (SECM) of microcrystalline diamond films.
heterogeneity is seen for the nanocrystalline films as well. An example of this is shown
in Figure 5.13A-D. Both height mode images and conductivity maps are presented over a
10 x 10 (A) and l x 1 pm (C ) area. Nodular features are seen in the images that are ~50-
100 nm in dimension. These are actually a collection of diamond grains that are 10-20
nm in size.”29 The intersection of the nodules constitutes the grain boundaries. The
figure also shows conductivity maps over the same areas (B and D). In these
measurements, a Au-coated Si tip contacting the surface was rastered over an area and
the current was measured using a bias voltage of +2 V (Erip — Esubstm). The bright areas
correspond to regions of higher conductivity and the darker features to regions of lower

conductivity. Clearly, there are conductivity variations across the surface. The more

conductive regions do not appear to be exclusively at the grain boundaries. The higher

176

 

 

resolution image in Figure 5.13D reveals that the distance between the more conducting
zones is on the order of 200-300 nm in some cases. These data are important and have
implications for the dimensional uniformity of the metal adlayers. The take-home
message is that inherently there is going to be spatial variation in the metal adlayer
formation on polycrystalline diamond due to this heterogeneity in electrical conductivity.
It remains to be proven but it is logical to suppose that the conductive regions are the
sites where initial nucleation occurs. Therefore, the fraction of these conductive sites on
a film will control the initial nucleation density. A similar conclusion was recently
reached by Enea et al. in their AFM study of the electrodeposition of Pt on

polycrystalline diamond.35

177

1000 nm

1000 nm

0. 75 500 nm

   

0.50 01""

0.25

 

0 ' . __
0 0. 25 0. 50 0.75 1.00pm 0 0- 25 0-50 0-75 1-0011m

Figure 5.13. AFM height mode images (air) of a boron-doped nanocrystalline diamond
thin film over a (A) 10 x 10 and (C) 1 x 1 pm area. Simultaneously recorded conductivity
maps of the same regions (B and D) with a bias voltage of +2 V (E..,, — Esubmtc).

5.4 Conclusions

The pulsed galvanostatic deposition of Pt electrocatalyst particles on boron-doped
microcrystalline and nanocrystalline diamond thin film was investigated. The deposition
was studied as a firnction of the pulse number and current density, while maintaining a
pulse length of l-s and a duty cycle of 50%. Pulsed galvanostatic deposition produced a
nominal particle size that was an order of magnitude smaller than that previously
observed for constant current deposition.”20 Deposition using ten pulses and a current

density of 1.25 mA/cm2 yielded the most ideal electrocatalyst on both diamond types, in

178

terms of the nominal particle size, particle coverage and the highest specific Pt surface
area. For both electrode types, the nominal particle size was ca. 30 nm with an RSD of

50% or less, and the particle coverage was ca. 109-1010 cm'2

under the optimum
deposition conditions. The typical specific surface areas were in the 10-50 mzpt/gp. range.
The deposition of Pt on diamond compared favorably to deposition on glassy carbon.
The microstructure and morphology of diamond strongly influenced the nucleation and
grth of Pt. The highest specific surface area of 50 mzpt/gpt was observed for the
nanocrystalline film, which possesses a higher fraction of exposed grain boundaries. The
grain boundaries appear to serve as the primary nucleation sites on this film. Finally, the
influence of the film's electronic properties on the metal phase formation was
investigated. The polycrystalline diamond film is quite heterogeneous in terms of the
electrical conductivity, as evidenced by CP-AF M. The film is composed of sites of faster
(reversible) kinetics where the dopant level is high and sites of slower kinetics (less

reversible) where there is less dopant. It is supposed that the regions of high electrical

conductivity serve as the initial nucleation sites.

179

10.

ll.

12.

13.

14.

REFERENCES

E. Antolini, J. Mater. Sci., 38, 2995 (2003).

F. Rodriguez-Reinoso, Carbon, 36, 159 (1998).

L. Carrette, K. A. Friedrich, and U. Stimming, Fuel Cells, 1, 5 (2001).

J. Y. Lucisano, J. C. Armour, and D. A. Gough, Anal. Chem, 59, 736 (1987).
A. Damjanovic and V. Brusic, Electrochim. Acta, 12, 615 (1967).

N. Giordano, E. Passalacqua, L. Pino, A. S. Arico, V. Antonucci, M. Vivaldi, and
K. Kinoshita, Electrochim. Acta, 36, 1979 (1991).

N. M. Markovic, T. J. Schmidt, V. Stamenkovic, and P. N. Ross, Fuel Cells, 1,
105 (2001).

H.-H. Yang and R. L. McCreery, J. Electrochem. Soc., 147, 3420 (2000).

T. Yano, E. Popa, D. A. Tryk, K. Hashimoto, and A. Fujishima, J. Electrochem.
Soc., 146, 1081 (1999).

E. Yeager, Electrochim. Acta, 29, 1527 (1984).
l. Morcos and E. Yeager, Electrochim. Acta, 15, 953 (1970).

E. Yeager, J. A. Molla, and S. Gupta, in Electrochemistty of Carbon, S.
Sarangapani, J. R. Akridge, and B. Schumm, Editors, PV 84-5, p. 123, The
Electrochemical Society Proceedings Series, Pennington, NJ, (1984).

E. Yeager and D. Tryk, in Proceedings of the 33rd International Power Sources
Symposium, The Electrochemical Society, Inc., Cherry Hill, NJ, (1988).

B. C. H. Steele and A. Heinzel, Nature, 414, 345 (2001).

180

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

E. Auer, A. Freund, J. Pietsch, and T. Tacke, Appl. Catal. A: Gen, 173, 259
(1998)

A. J. Appleby, Corrosion (Houston), 43, 398 (1987).

R. Makharia, P. Yu, J. Pisco, S. Kocha, and H. Gasteiger, Paper 1888 presented at
The 206th Meeting of the Electrochemical Society, Honolulu, HI, Oct. 3-8, 2004.

J. Wang, G. M. Swain, T. Tachibana, and K. Kobashi, Electrochem. Solid-State
Lett., 3, 286 (2000).

J. Wang and G. M. Swain, Electrochem. Solid-State Lett., 5, E4 (2002).
J. Wang and G. M. Swain, J. Electrochem. Soc., 150, E24 (2003).

M. Hupert, A. Muck, J. Wang, J. Stotter, Z. Cvackova, S. Haymond, Y. Show,
and G. M. Swain, Diamond and Rel. Mater., 12, 1940 (2003).

A. E. Fischer and G. M. Swain, J. Electrochem. Soc., 152, B369 (2005).

A. E. Fischer, Degradation Mechanisms of Carbon—Based Electrocatalyst Support
Materials and Development of an Advanced Support Based on Electrically
Conducting Diamond, Michigan State University, East Lansing, MI, (2005).

G. M. Swain, J. Electrochem. Soc., 141, 3382 (1994).

J. Xu, M. C. Granger, Q. Chen, J. W. Strojek, T. E. Lister, and G. M. Swain, Anal.
Chem, 69, 591A (1997).

M. C. Granger, J. Xu, J. W. Strojek, and G. M. Swain, Anal. Chim. Acta, 397, 145
(1999)

G. M. Swain, A. B. Anderson, and J. C. Angus, MRS Bull, 23, 56 (1998).

Q. Chen, D. M. Gruen, A. R. Krauss, T. D. Corrigan, M. Witek, and G. M. Swain,
J. Electrochem. Soc., 148, E44 (2001).

181

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

Y. Show, M. A. Witek, P. Sonthalia, and G. M. Swain, Chem. Mater., 15, 879
(2003)

A. E. Fischer, Y. Show, and G. M. Swain, Anal. Chem, 76, 2553 (2004).

M. Awada, J. W. Strojek, and G. M. Swain, J. Electrochem. Soc., 142, L42
(1995)

Y. Zhang, S. Asahina, S. Yoshihara, and T. Shirakashi, Electrochim. Acta, 48,
741 (2003).

S. Nakabayashi, D. A. Tryk, A. Fujishima, and N. Ohta, Chem. Phys. Lett., 300,
409 (1999).

N. Vinokur, B. Miller, Y. Avyigal, and R. Kalish, J. Electrochem. Soc., 146, 125
(1999)

O. Enea, B. Riedo, and G. Dietler, Nano Letters, 2, 241 (2002).

E. Budevski, G. Staikov, and W. J. Lorenz, Electrochemical Phase Formation
and Growth, VCH, New York (1996).

M.-S. Ldffler, B. GroB, H. Natter, R. Hempelmann, T. Krajewski, and J. Divisek,
Phys. Chem. Chem. Phys, 3, 333 (2001).

J. V. Zoval, R. M. Stiger, P. R. Biemacki, and R. M. Penner, J. Phys. Chem, 100,
837 (1996).

J. V. Zoval, J. Lee, S. Gorer, and R. M. Penner, J. Phys. Chem. B, 102, 1166
(1998)

J. L. Fransaer and R. M. Penner, J. Phys. Chem. B, 103, 7643 (1999).

S. Gorer, H. Liu, R. M. Stiger, M. P. Zach, J. V. Zoval, and R. M. Penner, in
Metal Nanoparticles, D. L. Feldheim and C. A. Foss Jr., Editors, p. 237, Marcel
Dekker, Inc., New York, New York, (2002).

182

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

R. M. Penner, J. Phys. Chem. B, 105, 8672 (2001).
R. M. Penner, J. Phys. Chem. B, 106, 3339 (2002).

H. Liu, F. Favier, K. Ng, M. P. Zach, and R. M. Penner, Electrochim. Acta, 47,
671(2001)

H. Liu and R. M. Penner, J. Phys. Chem. B, 104, 9131 (2000).

G. Gunawardena, G. Hills, 1. Montenegro, and B. Schariflrer, J. Electroanal.
Chem, 138, 225 (1982).

P. Allongue and E. Souteyrand, J. Electroanal. Chem, 286, 217 (1990).
S. Wang and G. M. Swain, Manuscript in preparation.

A. Essalik, K. Amouzegar, and O. Savadogo, J. Appl. Electrochem., 25, 404
(1995)

P. Sonthalia, E. McGaw, Y. Show, and G. M. Swain, Anal. Chim. Acta, 522, 35
(2004)

H. Angerstein-Kozlowska, B. E. Conway, and W. B. A. Sharp, J. Electroanal.
Chem, 43, 9 (1973).

B. V. Tilak, B. E. Conway, and H. Angerstein-Kozlowska, J. Electroanal. Chem,
48, l (1973).

F. Gloaguen, J .-M. Léger, and C. Lamy, J. Appl. Electrochem., 27, 1052 (1.997).

A. R. Kucemak, P. B. Chowdhury, C. P. Wilde, G. H. Kelsall, Y. Y. Zhu, and D.
E. Williams, Electrochim. Acta, 45, 4483 (2000).

Y. Mo, S. Sarangapani, A. Le, and D. A. Scherson, J. Electroanal. Chem, 538-
539,35(2002)

183

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

J. Wang, Ph.D. Dissertation, Fabrication, Characterization, And Electrocatalytic
Activity of Metal/Diamond Composite Electrodes, Michigan State University, East
Lansing, MI, (2002).

L. M. Plyasova, I. Y. Molina, A. N. Gavrilov, S. V. Cherepanova, O. V.
Cherstiouk, N. A. Rudina, E. R. Savinova, and G. A. Tsirlina, Electrochim. Acta,
51, 4477 (2006).

B. R. Schariflter and J. Mostany, J. Electroanal. Chem, 177, 13 (1984).

J. T. Huang, C. S. Hu, J. Hwang, H. Chang, and L. J. Lee, Appl. Phys. Lett., 67,
2382 (1995).

K. Nishimura, K. Das, and J. T. Glass, J. Appl. Phys, 69, 3142 (1991).

N. Vinokur, B. Miller, Y. Avyigal, and R. Kalish, J Electrochem. Soc., 143, L238
(1996)

M. N. Latto, G. Pastor-Moreno, and D. J. Riley, Electroanalysis, 16, 434 (2004).

R. F. Mamin and T. Inushima, Phys. Rev. B, 63, 033201 (2001).

W. Zhu, in Diamond: Electronic Properties and Applications, L. S. Pan and D.
R. Kania, Editors, p. Chapter 5, Kluwer Academic Publishers, Boston, (1995).

M. I. Landstrass and K. V. Ravi, Appl. Phys. Lett., 55, 1391 (1989).

K. Hayashi, S. Yamanaka, H. Okushi, and K. Kajimura, Appl. Phys. Lett., 68, 376
(1996)

H. J. Looi, R. B. Jackman, and J. S. Foord, Appl. Phys. Lett., 72, 353 (1998).

H. J. Looi, L. Y. S. Pang, A. B. Molloy, F. Jones, J. S. Foord, and R. B. Jackman,
Diamond and Rel. Mater., 7, 550 (1998).

H. J. Looi, M. D. Whitfield, J. S. Foord, and R. B. Jackman, Thin Solid Films,
343-344, 623 (1999).

184

70.

71.

72.

73.

74.

B. Fiegl, R. Kuhnelt, M. Ben-Chorin, and F. Koch, Appl. Phys. Lett., 65, 371
(1994)

D. Stenbach, A. F loter, H. Gl'ittler, R. Zachai, and P. Ziemann, Diamond and Rel.
Mater., 8, 273 (1999).

T. Kolber, K. Piplits, R. Haubner, and H. Hutter, Fresenius. J. Anal. Chem, 365,
636 (1999).

K. Ushizawa, K. Watanabe, T. Ando, I. Sakagucki, M. Nishitani-Gamo, Y. Sato,
and H. Kanada, Diamond and Rel. Mater., 7, 1719 (1998).

K. B. Holt, A. 1. Bard, Y. Show, and G. M. Swain, J. Phys. Chem. B, 108, 15117
(2004)

185

 

 

CHAPTER 6

The Effect of Surface Chemistry on Metal Phase
Formation and Particle Stability

6.1 Introduction

Our effort developing diamond as an advanced electrocatalyst support material
involves learning how to form and stabilize Pt catalyst particles on the surface.M Most
of the metal electrodeposition research reported to date has been performed with H-
terrrrinated diamond (microcrystalline and nanocrystalline) thin-film electrodes.
Therefore, little is known about how the surface chemistry influences metal phase
formation and stability. Recently, some work on the formation of metal catalyst particles
on oxidized diamond film was reported.7 In this work, Zhang et a1. observed that Au
particles electrodeposited on an oxidized diamond surface form as 40-400 nm clusters
located principally near the grain boundaries. In comparison, Au particles
electrodeposited on hydrogen-terminated diamond were randomly dispersed over the
whole surface with a nominal diameter of ca. 60 nm. This suggests that the oxidized

grains of diamond are less favorable sites for metal phase formation.

Understanding how these metal phases form and interact with diamond is
necessary because there is considerable research showing the importance of metal-
support interactions in electrocatalytic reactions; specifically, the effect of carbon surface

oxides on supported catalysts.8'18 In general, an increase in carbon-oxygen functionality

186

on various spz-bonded carbon supports has little influence on the catalyst morphology
(i.e., real Pt area); however, the particles often exhibit increased catalytic activity for
certain reactions. For example, Jovanovié et al. found that electrochemically oxidizing
glassy carbon lead to both a better distribution of Pt over the surface and an increase in
activity towards methanol oxidation, but did not significantly affect the real Pt surface
area.18 They attributed the increased activity to (i) a higher percentage of acidic groups
on the oxidized GC promoting a higher fraction of oxygen containing species in the Pt
catalyst and (ii) OH-like functional groups present on the oxidized GC promoting the
oxidation of absorbed intermediates in the oxidation reaction.18 Similar results were
found by Gloaguen et al. for Pt potentiostatically deposited on graphite powder and
carbon black.14 They determined that neither the structure of the carbon support nor the
amount of surface oxide significantly affected the particle morphology. However, the
increased surface oxide content did lead to an increase in both the Pt mass and specific

activity for methanol oxidation.

Chapter 5 of this dissertation presented a comparison of Pt metallization of both
microcrystalline (BMD) and nanocrystalline (BND) diamond thin-film electrodes by
pulsed galvanostatic deposition. The effect of the morphology and microstructure on the
nucleation and growth of Pt electrocatalyst particles was studied as a function of the pulse
current density and pulse number. In this chapter, how the surface chemistry of
microcrystalline and nanocrystalline diamond surfaces affects the formation and
stabilization of Pt particles is reported on. The surface chemistries studied were

hydrogen (H-terminated), oxygen formed by chemical oxidation (O-terminated), 4-

187

 

 

nitrobenzene diazonium~modified (4-NBD) and 4-methoxybenzene diazonium-modified
(4-MBD). The comparison surface is the H-terminated one as such films emerge from
the CVD reactor afier growth. The O-terminated surfaces were studied because oxygen
is slowly incorporated during air exposure. The effect of surface oxides on metal
formation and stability are particularly important for fuel cell applications as the
electrocatalyst support material slowly oxidizes. The aryl molecule modification was
used to see if the molecular adlayer would act as an extension of the electrode surface or
simply as a blocking layer. Specifically, 4-NBD was selected because it had already been
used by our group in other work, which revealed that diamond surfaces can be modified
with this molecule.l9’2' Finally, 4-MBD (derivatized with SO3- groups) was chosen to
see if excess surface charge can influence metal phase formation.22 The Pt adlayers were
deposited by pulsed galvanostatic deposition using the parameters described in Chapter 5
(ten I-s pulses with a 50% duty cycle at a pulse current density of 1.25 mA/cmz). This
work is part of our long-term goal of understanding metal phase formation on diamond
and gaining fundamental insight into what electrode properties affect the particle
formation and stability. The Pt-coated films were characterized by scanning electron
microscopy (SEM) and cyclic voltammetry. The catalytic activity of the electrode

toward the oxygen reduction reaction in acid was investigated at room temperature.

188

 

6.2 Results & Discussion

6.2.1 Physical Characterization

The target particle size and density for the catalytic Pt phase is ca. 10 nm and
1010 cm'z, respectively. Scanning electron micrographs of Pt adlayers formed on BMD
surfaces terrrrinated by (A) H, (B) O, (C) 4-NBD and (D) 4-MBD are shown in Figure
6.1A-D. A summary of the particle analysis data is presented in Table 6.1. The particles
on BMDH are spherical, moderately rough and evenly distributed over the entire surface
including the facets and grain boundaries. The nominal particle diameter and density
were 40 1 30 nm and 8 (1 1) x 109 cm'z.6 A lower density of particles with a larger
nominal diameter is seen on BMDO,‘ (B). This is consistent with the observation made by
Zhang et al. for oxidized diamond.7 The particles are fairly spherical, but rougher than
those on BMDH and unevenly distributed over the surface indicating that certain regions
are more favorable for metal adlayer formation than others. The nominal size was 100 1
50 nm and the particle density was 1 (1 l) x 109 cm’z. The increase in particle roughness
and size as well as the decrease in density, as compared with BMDH, is consistent with a
decrease in the number of active sites initially available for metal deposition and
consequent metal deposition on pre-existing growth centers. It appears that the surface

carbon-oxygen functionalities block sites for metal phase formation possibly by lowering

. . . 2
the surface electrical conductrvrty. 3

Figure 6.1C and D show Pt deposited on BMD4-NBD and BMD4-MBD, respectively.

The nominal particle size on both of these surfaces was similar, 80 1 50 nm and 70 1 30

-2
nm, respectively. Similar particle coverages of 4 (1 2) and 6 (1 3) x 109 cm were also

189

 

observed. There is a larger deviation in the particle density on the aryl-modified surfaces
compared to the hydrogenated surface. This is likely due to variation in the aryl-layer
formation on the heterogeneous diamond surface. We suspect that while the aryl-

admolecule coverage, determined by integrating the faradaic charge associated with the

l-electron reduction of the nitro functional group (Ar-NO2 +e‘ —> Ar-NOi), is the

equivalent of a theoretical monolayer, it is likely that the aryl-adlayer formed as a
multilayer in isolated regions over the surface. Such was a conclusion from tapping
mode AF M measurements.” The similar particle size and density possessed by each film
suggests that the 4-MBD modification procedure does not inhibit Pt formation. Recall
that the 4-MBD contains SO3‘ functional groups and the reactant is the anionic PtC164’
species. Additionally, the particles on these surfaces were not as spherical as those seen
on the H- and O-terminated films, but rather more elongated. This is possibly due to
blocking caused by the aryl-adlayer. Particle roughness was difficult to determine for
BMD4.NBD due to the tendency for this film to charge in the SEM, however BMD4-MBD

exhibited particles that were quite dendritic in nature.

All three surface chemistries examined provided no benefit in terms of achieving
the target nominal particle size and coverage. In fact, it appears that all three chemistries
act as partial blocking layers, either because of chemical or electronic effects. The H-
terrninated surface yielded the best particle characteristics. Certainly, the particle size
was increased, the particle density was reduced and the particle shape was affected by the

three surface chemistries. The particle shape for the Pt/BMDH film appeared to be

190

 

 

spherical but the shapes for the Pt/BMD4.NBD and Pt/BMD4_MBD films were more

elongated and dendritic with high surface area.

 

Figure 6.1. Scanning electron micrographs of as-deposited Pt/BMD electrodes deposited
using ten l-s pulses at 1.25 mA/cmz. The BMD surfaces were (A) H-terminated, (B) O-
terminated, (C) 4—NBD modified and (D) 4-MBD modified.

191

 

 

 

 

 

Table 6.1. Particle Analysis Data for Pt-Coated Microcrystalline and Nanocrystalline
Diamond Films as a Function of the Surface Chemistry

 

 

 

Microcrystalline Diamond N anocrystalline Diamond
Surface . .
Chemistry Particle Size (Particle Particle Size (Particle
(nm) o’era. e. (nm) ostera ell
(10 cm ) (10 cm" )
I-I-terminated 40 1 30 8 1 1 3O 1 30 20 1 10
O-terminated 100 1 50 1 1 1 130 1 60 0.5 1 0.3
4-NBD modified 80 1 50 4 1 2 100 1 50 1 1 l
4-MBD modified 70 1 30 6 1 3 50 1 3o 5 1 3

 

 

 

 

' Area normalized to geometric area of diamond, 0.2 cm2.

Pt deposition was also performed on BND thin films modified by the same
chemistries. The high resolution SEM images and particle analysis data are shown in
Figure 6.2A-D and Table 6.1, respectively. BNDH (A) possessed particles that were
mostly spherical, quite rough, and distributed evenly over the surface. The particles were
nominally 30 1 30 nm in size at a density of 20 (1 10) x 109 cm'z.6 Of all the films tested,
this one most closely meets the desired targets for particle size and density. This
particular image shows many particles on the order of 30 nm, plus a few that are sizably
larger. The oxidized microcrystalline diamond, BNDox (B) possessed spherical particles
that were significantly (i.e., an order of magnitude) larger and less dense than those seen
for BNDH. Additionally, these larger agglomerates possess an increased roughness
compared to Pt/BNDH. The average particle diameter was 130 1 60 nm and the particle
density was 5 (1 3) x 108 cm'z. The H-terminated sample had particles deposited over the
entire surface. The O-terminated sample had a much more localized distribution of metal

particles. BND4-NBD (C) exhibited a nominal particle size of 100 1 50 nm and a particle

192

 

density of l (1 1) x 109 cm'z. The larger size and lower density, as compared to the H-
terminated surface, is consistent with a blocking effect produced by the aryl-adlayer that
results in a decrease in the number of active sites for metal phase formation. While there

are a significant number of particles that are quite spherical, there are a large number of

agglomerates that are elongated and quite dendritic. Finally, BNDaMBD exhibited a

nominal particle size and density of 50 1 30 nm and 5 (1 3) x 109 cm'z, respectively.
Again, the aryl-adlayer modification produced non-spherical particles that were quite
rough. Interestingly, the type of aryl molecule used to modify the surface had a more
significant impact on particle size and density on the nanocrystalline diamond than on the
microcrystalline diamond. Perhaps the aryl molecules are more evenly distributed on the
smoother nanocrystalline surface, whereas modifying the rough microcrystalline surface

results in an uneven distribution of the adlayer. This results in voids in the modified

layer and less impact on the metal deposition process.

193

 

Figure 6.2. Scanning electron micrographs of as-deposited Pt/BND electrodes deposited
using ten l-s pulses at 1.25 mA/cmz. The BND surfaces were (A) H-terminated, (B) O-
terminated, (C) 4-NBD modified and (D) 4-MBD modified.

6.2.2 Electrochemical Response

Cyclic voltammetric i-E curves were obtained in deoxygenated 0.1 M HClO4 for
Pt/BMD and Pt/BND films possessing the different surface chemistries. The curves for
the Pt/BMD films are shown in Figure 6.3. All scans were initiated at 200 mV and
scanned in the positive direction at a scan rate of 50 mV/s. It can be seen that the curves
for the H-terminated and O-terminated surfaces possess voltammetric features

characteristic of clean polycrystalline Pt. These include a flat Pt oxide formation region

194

from 800-1100 mV, Pt oxide reduction at ca. 450 mV and four clearly resolved peaks
between 0 and -200 mV due to hydrogen adsorption and desorption. However, both the
diazonium modified electrodes had curves with some noticeable differences. First, the Pt
oxide formation current is not flat but rather possesses a peak shape character that is most
pronounced for the 4-NBD modified surface. Also, the charge in the hydrogen
adsorption/desorption region is suppressed with no resolvable peaks. It is supposed that
aryl admolecules on the diamond surface near the metal deposits interact with the surface
to increase the "Pt oxide" current due to their oxidation and suppress the hydrogen
adsorption/desorption charge. Surprisingly, potential cycling 100 times between -300
and 1500 mV did not improve the Pt voltammetric features consistent with strong
contaminant adsorption on the surface. The voltammetric i-E curves for Pt deposited on
the BND surfaces (not shown) exhibited similar features to those described for the

Pt/BMD films (i. e., well resolved Pt features for H- and O-terminated surfaces and poorly

resolved Pt features for diazonium modified surfaces).

195

 

 

  
   

 

 

 

 

 

 

30
20 _' ,.
:5 .
1 10 r
v .
H r-
5 .
r... 0 r
h t-
: .
U .
'10 . — H—terminated
_ ~' . "'0'" O-terminated
-20 — ' , -----x4-NBD modified
I 9‘ --<>-- 4-MBD modified
_30 g 1 I 1 1 1 J 1 1 L I 1 1 1 I 1
0 400 800 1200

Potential (mV vs. Ag/AgCI)

Figure 6.3. Cyclic voltammetric i-E curves in degassed 0.1 M HC104 for Pt-coated
microcrystalline diamond electrodes as a function of the surface chemistry. Scan rate =
50 mV/s.

The poorly resolved hydrogen adsorption/desorption region for the diazonium
modified diamond surfaces made determining the electrochemically active surface area
(ESA) problematic. The hydrogen adsorption/desorption charge is often used to
calculate this value and can be found by integrating either the anodic or cathodic current
for the respective reaction between ca. 100 and -190 mV vs. Ag/AgCl and correcting for
the double-layer charging current, as discussed previously.6 While the charges associated
with hydrogen adsorption and desorption on the clean Pt surfaces were approximately
equal, the values for the diazonium modified surfaces were not. The charge associated
with hydrogen adsorption was always several times larger than that for desorption.

Therefore, we suppose that a parasitic reaction(s) was adding to the hydrogen adsorption

196

charge. One possibility is that the admolecule oxidation product formed during Pt-oxide
formation gets reduced at potentials where hydrogen adsorption occurs. In any case, the
hydrogen desorption charge was used to calculate the Pt BSA. The ESA obtained from
the hydrogen adsorption/desorption charge and the roughness factor (RF = ESA
normalized to the geometric area of the diamond, 0.2 cm2) are summarized for the two
diamond morphologies in Table 6.2. Surprisingly, the ESA for the H— and O-terminated
BMD surfaces were quite similar despite the large difference in particle size and density.
This is attributed to an increased surface roughness of the particles on the oxidized
surface (i.e., metal depositing on pre-existing growth centers). As expected, the 4-NBD
modified surface exhibited the lowest ESA of the four. Interestingly, BMD4-MBD had a
higher Pt BSA and RF despite possessing particles of similar size and density. This is

attributed to the increased surface roughness due to the more irregularly shaped particles.

 

Table 6.2. Electrochemically Active Pt Surface Area (BSA) and Roughness Factor (RF)
for Pt-Coated Microcrystalline and Nanocrystalline Diamond Films as a Function of the

Surface Chemistry

 

 

 

 

Microcrystalline Diamond N anocrystalline Diamond
Surface
Chemistry Pt ESA Roughness Pt ESA Roughness
(cmz) Factor (cmz) Factor
H-terminated 0.15 1 0.01 0.74 0.24 1 0.03 1.2
O—terminated 0.18 1 0.02 0.90 0.18 1 0.03 0.90
4—NBD modified 0.013 1 0.001 0.065 0.014 1 0.003 0.070
4-MBD modified 0.06 1 0.02 0.28 0.02 1 0.01 0.11

 

 

 

RF = Electrochemically Active Pt Surface Area Normalized to Geometric Area of the
Diamond, 0.2 cm2.

197

The ESA values for the Pt/BND surfaces follow a more expected trend. The
clean H-terminated surface exhibited the largest ESA, possessing the smallest particle
size and largest particle density, followed by the clean O-terminated surface, and finally
the 4-MBD and 4-NBD modified surfaces. Interestingly, the larger particle size and
rougher particles on the O-terminated surface still do not result in a higher ESA than was

observed for the large coverage of much smaller particles on the H-terminated surface.

6.2.3 Probing the Oxygen Reduction Reaction

The oxygen reduction reaction in 0.1 M HCIO4 at room temperature was used to
investigate the effect of the diamond surface chemistry on the catalytic activity of the
fi'eshly deposited metal particles. Pt and Pt-based alloys are currently considered the
most active catalysts for the OR under acidic conditions, promoting the reaction via a 4-

124

electron pathway as shown in equation 6.2. This pathway is preferred from an energy

density standpoint over the 2-electron pathway presented in equation 3.2.1.

02 + 4H+ + 4e” .—_+ 21120 (E° =1.03 v vs. Ag/AgCl) (6.2.1)

The linear sweep voltammetric i-E curves for the Pt-containing microcrystalline
and nanocrystalline diamond films that were first modified by the four various surface
chemistries are shown in Figures 6.4 and 6.5. The scans were recorded in the presence of
dissolved Oz, initiated at 1000 mV and swept negative at a rate of l mV/s. For the

microcrystalline diamond surfaces, it can be seen that both the reduction peak potential

(Bred) and half-peak potential (Ered ), displayed in Table 6.3, are invariant with the
p p/2

198

 

 

surface chemistry. However, the curves vary significantly at potentials negative of the
reduction peak suggesting that the mass transport of 02 is different for each. The clean
Pt/BMDH and Pt/BMDox films reach a near steady state condition at ca. 200 mV. The
Pt/BMD4-MBD film appears to approach a steady state condition at more negative
potentials while the Pt/BMD4-NBD film does not. Instead, the mass transfer for this film

1/2

appears to be a mixture of semi-infinite linear diffusion (t' ) and steady-state diffusion

(time independent decay), based on a qualitative analysis of the i-t curve shape. The

 

differences in mass transfer are attributed to differences in the particle shape, size and

spacing in relation to the time scale of the potential sweep rate.

 

l0

 

-0.20

' at
i
i :
' I
EJJO’ ?
0 ,__ . . ........................................
- '0.”- .."

’ -oso»
'10 : r

'0'6973 :65 16.3 A ls.s ‘ -5.0 4.5 :41)

 

 

 

 

log I (A)

 

Current (uA)

— H-terminated
---o--- O-terminated
1 ------x 4-NBD modified

 

 

 

 

 

 

 

 

.40 :- 1 . "Om 4-MBD modified
L
1
_50 I 1 l 1 I 1 I 1 I 1
0 200 400 600 800 1000

Potential (mV vs. Ag/AgCl)

Figure 6.4. Linear sweep voltammetric i-E curves for the OR in 0.1 M HClO4 at Pt-
coated microcrystalline diamond electrodes as a function of the surface chemistry. Scan
rate = 1 mV/s. The corresponding Tafel plots are presented in the inset.

199

 

Table 6.3 Reduction Peak Potential (Eifd) and Half-Peak Potential (EEC/(21), Cathodic

Peak Current (ipc) and Specific Activity (at E = 600 mV) for the Oxygen Reduction
Reaction at Pt-Coated Microcrystalline Diamond Films as a Function of the Surface

Chemistry

 

 

Surface Ered 12",“; inc Specific Actrvrty
Ch ' t p p (mA/cm2 )" (E — 600 mV,
emrs ry (mV) (mV) Pt ILA/cmzpt)‘.
H-terminated 397 1 7 479 1 6 0.28 1 0.01 4 1 l
O-terminated 390 :l: 30 460 1 20 0.22 1 0.01 3 1 l
4-NBD modified 380 1 20 480 1 10 2.6 1 0.2 32 1 8
4-MBD modified 390 1 20 470 1 10 0.7 :l: 0.3 8 1 2

 

 

 

" Values Normalized to the Calculated Electrochemically Active Pt Surface Area (BSA)
and the potential is reported on the Ag/AgCl reference scale

In contrast to the trends for the microcrystalline diamond electrodes, the

voltammetric i-E curves for the Pt-coated nanocrystalline diamond electrodes, shown in
Figure 6.5, exhibit significant variation in the Eged and 13373 values (summarized in

Table 6.4) as a function of surface chemistry. Interestingly, these two values were
significantly more positive for Pt/BNDH than for Pt/BNDox and even Pt/BMDH
suggesting that these particles were more electrochemically active towards the reaction.
This is somewhat contrary to that observed by others for oxidized GC towards methanol
oxidation.”’18 The activity differences might also be related to structural differences
(Pt(hkl)) in the metal catalyst particles. For example, it is well known that the exchange
current density for hydrogen evolution increases in the order io(1 l l) > 15000) > io(110),
which gives a measure of the kinetic activity of the three surfaces. In contrast, the

opposite is seen for the ORR, (100) < (110) z (111). The 4-NBD modified BND surface,

however, showed significantly more negative Eged and E333 values than the same

200

 

modified BMD film. This may be due to a more evenly distributed aryl-adlayer over the

smoother nanocrystalline surface and suggests the adlayer inhibits the reaction as

opposed to acting as an extension of the electrode surface. All the curves reach a steady

state mass transfer at potentials negative of 200 mV. It appears that the 4-NBD adlayer

produces more of a blocking effect than the 4-MBD layer.

10

Current (uA)

 

 

 

 

 

 

 

 

 

 

 

 

 

_-Il.r
I .02
E 413 ‘ . '
T: 40.4 ........................................
- o3
'_ 41.6
' 4.77.0 -6.5 -6.0
L — H-terminated
L "'0"- O-terminated
; "X"- 4-NBD modifed
_ "0"- 4-MBD modified
b l 1 . l . l . 4 .
0 200 400 600 800 1000

Potential (mV vs. Ag/AgCl)

Figure 6.5. Linear sweep voltammetric i-E curves for the ORR in 0.1 M HClO4 at Pt-
coated nanocrystalline diamond electrodes as a function of the surface chemistry. Scan
rate = 1 mV/s. The corresponding Tafel plots are shown in the inset.

The normalized peak current density was greatest for the 4-NBD and 4-MBD

modified surfaces and smallest for the hydrogen— and oxygen-terminated surfaces, with

the values differing by about an order of magnitude. We suppose though that these

201

values are high because of an underestimate of the active Pt area for the reaction (e. g.,
ESA) based on the hydrogen adsorption/desorption charge measurement. In other words,
the active area for the ORR is larger than that calculated from the voltammetric curves
for hydrogen adsorption/desorption. In support of this, chronoamperometric i-t transients
(discussed in the following section) obtained in the same solution and plotted using the
well known Cottrell equation resulted in an unrealistically high calculated value for the
number of electrons involved in the reaction, ca. 50-80, when using the Pt ESA
determined from the hydrogen adsorption/desorption. However, if a Pt area of 0.2 cm2
(RF = l) was used, values between 3 and 5 were calculated, which is close to the
predicted value of 4. The Pt ESA appears affected by admolecule contamination on the
aryl-modified diamond surfaces. It is possible that the admolecule adsorption inhibits
hydrogen adsorption/desorption but may facilitate the OR. Verification of this would

require further study.

 

Table 6.4. Reduction Peak Potential (Effd) and Half-Peak Potential (13373), Cathodic

Peak Current (if) and Specific Activity (at E = 600 mV) for the Oxygen Reduction
Reaction at Pt-Coated Nanocrystalline Diamond Films as a Function of the Surface

Chemistry

 

 

 

Surface Ered Erfig ipc Sp ecu—fie Actrvrty
Chemistr p p (mA/ m2 )' (E ’ 600 "'V’
y (mV) (mV) ° Pt Mammy
H-terminated 470 :1: 20 540 i 10 0.20 i 0.02 13 i 5
O-terminated 400 :t 10 474 :t 7 0.23 i 0.04 3 i l
4-NBD modified 330 :t 20 430 i 20 2.3 :t 0.4 21 i 3
4-MBD modified 390 i 20 480 i 10 1.6 i 0.5 28 i 3

 

 

“ Values Normalized to the Calculated Electrochemically Active Pt Surface Area (ESA)

202

2.11

__

 

The specific activity (SA = current / cm2 of electrochemically-active Pt) of an
electrocatalyst is ofien used to describe the kinetic activity for the ORR.25 Values
calculated in this work are displayed in Tables 6.3 and 6.4. The SA is obtained by
measuring the current at an overpotential of -330 mV; 3 potential at which the current is
controlled by the electron transfer kinetics.26’27 The equilibrium potential (Eeq) here was
taken to be 930 mV vs. Ag/AgCl, assuming an 0; concentration of 1.3 mM (estimated for
0.1 M HClO4 at 25 °C)”, a 4-electron transfer and accounting for the solution pH.

Therefore, the current was measured at a potential of 600 mV, which corresponds to an

overpotential of -330 mV.

Interestingly, the microcrystalline and nanocrystalline samples do not exhibit the
same trends. The largest SA value for the Pt/BMD samples is observed for the 4-NBD
film followed by the 4-MBD, H-terminated and O-terminated surfaces. The largest SA
value for the Pt/BND samples, however, is observed for the 4-MBD surface, followed by
the 4-NBD, ill-terminated and O-terminated surfaces. The diazonium modified surfaces
give the highest SA values likely due to an underestimated ESA. However, the SA on the
clean diamond surfaces appears to correlate well with what one might expect based on
the particle size and density observed in the SEM images. Pt/BNDH exhibited the largest
SA ( 13 i 5 uA/cmzpt) while possessing rough spherical particles of the smallest size and

largest density. Pt/BMDH exhibited the second largest SA (4 i 1 uA/cmzpt) while
possessing slightly larger, more spherically shaped particles and at a lower density

compared to Pt/BNDH. Finally, both the oxidized films exhibited similar SA values (ca.

203

3 pA/cmzp‘) as a result of comparably large Pt agglomerates and a much lower particle

density compared to the H-terminated diamond surfaces.

Tafel analysis of the i-E data was performed in the potential region where the
reaction rate is limited by electron transfer. The corresponding plots are shown in the
insets of Figures 6.4 and 6.5. The Tafel slope, uncorrected for mass transfer effects, the
transfer coefficient and the exchange current density are presented for the two diamond
electrode types in Tables 6.5 and 6.6. It can be seen that the Tafel slope and transfer
coefficient depend little on the diamond film morphology or surface chemistry. These
values range between ca. -70 — -75 mV/dec. and ca. 0.75 —— 0.87, respectively. This
indicates that mechanistically, the ORR is the same for all the surfaces and depends on
the properties of Pt and not the support. In general, the exchange current density ranged
from 10’10 to 10'” A/szpt (normalized to the Pt ESA), with the exception of the
Pt/BND4-MBD sample. This film exhibited the highest exchange current density of 2 (2t 1)
x 10'9 Ncmzn. This could be connected to the high specific activity observed for the
sample. The exchange current density depends on the crystallographic orientation, the Pt
BSA and the solution conditions. The orientation and ESA are indirectly affected by the

surface chemistry by virtue of its effect on Pt nucleation and growth.

204

 

 

Table 6.5. Summary of Values Obtained from Tafel Analysis of the ORR at Pt-Coated
Microcrystalline Diamond Films as a Function of the Surface Chemistry

 

 

Surface Chemistry 15:3 “81:2,; (1 (10-10 1.0/c1112“)
H-terminated -75 i 2 0.79 i 0.02 1.5 _+. 0.2
O-terminated -74 i 4 0.80 i 0.04 0.7 i 0.2

4-NBD modified -70 i 1 0.85 i 0.01 5 i 2
4-MBD modified -71 i 9 0.8 i 0.1 2 i 1

 

 

 

 

 

Table 6.6. Summary of Values Obtained from Tafel Analysis of the ORR at Pt-Coated
Nanocrystalline Diamond Films as a Function of the Surface Chemistry

 

 

Surface Chemistry {:3 [3:136 a (10-10 lie/cmzpt)
H-terminated -68 :t 2 0.87 :t 0.02 1.7 i 0.4
O-terminated -73 i 2 0.81 :t 0.02 0.9 :t 0.4

4-NBD modified -77 i 3 0.77 i 0.03 8 i 2
4-MBD modified -80 i 4 0.74 i: 0.03 20 i 10

 

 

 

 

6.2.4 Investigating the Oxygen Reduction Mechanism

The oxygen reduction mechanism was investigated at the different Pt—diamond
electrodes. The conclusions reached are based on one set of measurements using each
technique. Cyclic voltammetric i-E curves were recorded at scan rates from 1 mV/s to 10
V/s, and the data were evaluated in the manner described in Chapter 3. Since the OR is
an irreversible reaction, the i-E curve is expected to follow the trends given by equations

6.2.2 to 6.2.4.

205

 

 

1/2 1/2
Ep= Eo - RT 0.780+ In D0 + ln(9-F—v] (6.2.2)
anaF k0 RT

)1/2

ip = 2.99 x 105n(oma .Atr)<,1/2Co"‘v“2 (6.2.3)

 

- _ * 0 _ anal: _ 0'
11,—0.227nrnatc0 k expl: [RT )(Ep E )] (6.2.4)

Indeed plots of ip vs. v“2 and 1n ip vs. (Ep — E°) (assuming E° = E") exhibited linear trends
and were used to calculate own, the stoichiometric number of electrons transferred (n) and
the rate constant (k0). These values along with the individually determined Pt ESA from
the hydrogen adsorption/desorption charge measurements are summarized for the two
diamond types in Tables 6.7 and 6.8. Note that there are two k0 values presented in each
table that correspond to a value calculated using the n value determined experimentally

(referred to in the table as new.) and the other to a value found using the assumed value n

=4.

There are two points worth noting here. First, Pt on hydrogen- and oxygen-
terminated diamond behaves in a similar manner independent of the diamond type as n
values of 2-4 and k0 values in the mid-10'5 cm/s range are seen. In comparison, Markovié
et al. found n values of 4 and k0 values of ca. 10'2 cm/s for single crystal Pt electrodes in
0.1 M HClO4 at 22 °C using rotating-disk measurements.29 Second, there is no n value

listed for any of the aryl-modified films because of its unrealistic value. It appears that

206

 

there must be a larger surface area participating in the ORR than estimated from the
hydrogen adsorption/desorption charge for the aryl-molecule modified films because the
an, values are similar to that obtained for the clean films and there is no possible way to
observe n values of 18-25. Additionally, considering the voltammetric i—E curves in
Figures 6.4 and 6.5, there is no way that the modified surfaces could exhibit a higher rate

constant for the reaction by nearly an order of magnitude.

 

Table 6.7. Electrochemical Quantities Determined Using Cyclic Voltammetric Data
Obtained for the ORR at Pt-Coated Microcrystalline Diamond Films as a Function of the
Surface Chemistry

 

 

Parameter H—terminated O-terminated 4-NBD mod. 4-MBD mod.
Pt ESA (cmz) 0.14 0.12 0.016 0.014
cm. 0.231 0.187 0.187 0.182
n 3.5 2.4 17?? 7??
k” (cm/s, n = new.) 4.8 x 10'5 3.4 x 10'5 _ —
k0(cm/s,n=4) 4.2x 10'5 2.1x10'5 1.2x10'4 1.5x 10“

 

 

 

 

 

Table 6.8. Electrochemical Quantities Determined Using Cyclic Voltammetric Data
Obtained for the OR at Pt-Coated Nanocrystalline Diamond Films as a Function of the
Surface Chemistry

 

 

 

Parameter H-terminated O-terminated 4-NBD mod. 4—MBD mod.
Pt ESA (cmz) 0.16 0.20 0.015 0.011
cm, 0.185 0.175 0.190 0.212
n 2.9 2.2 7?? 7??
k” (cm/s, n = new.) 5.0 x 10'5 7.0 x 10'5 — .—
k" (cm/s, n = 4) 3.7 x 10‘5 4.0 x 10'5 2.2 x 10“ 1.3 x 10“

 

 

207

 

It appears that the calculated Pt ESA values for the aryl-modified surfaces are an
underestimate for the OR active area. By comparison, if it is assumed that the active Pt
area is equal to the geometric electrode area (0.2 cm2) and the values for the aryl
molecule-modified surfaces in Tables 6.7 and 6.8 are recalculated, then n values of ca.

1.5 and k0 values of ca. 10'5 are obtained, which are certainly more reasonable.

The same set of measurements was also preformed using a polished
polycrystalline Pt electrode. A Pt ESA value of 0.31 cm2 was determined based on the
hydrogen adsorption charge. This electrode exhibited the following values: em, = 0.188,
n = 1.5 and k0 = 4.0 x 10'5 (n = new) or 1.5 x 10'5 (n = 4) cm/s. The Pt nanoparticles
deposited on the diamond surfaces exhibited similar values suggesting that indeed the
reaction proceeds via the same mechanism on the Pt nanoparticles as it does on the bulk
Pt electrode. As observed in Chapter 3, n values determined using this method is lower

than the predicted value.

Comparatively, the number of electrons involved in the reaction (n) was also

determined using a large amplitude potential step technique and corresponding i-t'w

plots
with the Cottrell equation (3.2.10). As expected, an 71 value of 4.2 was observed for the
polycrystalline Pt electrode, consistent with the 4-electron reaction. Pt deposited on the
oxidized diamond surfaces exhibited similar values, ranging from 4-5. However, n
values calculated for the Pt-coated H-terminated diamond surfaces were ca. 8. This also

suggests that the Pt ESA calculated from hydrogen adsorption is less than that

participating in the ORR. These results were able to be repeated on another H—terminated

208

film. However, if we assume the literature value for n of 4, the Pt ESA calculated for the
Pt-coated H-terminated diamond films using the Cottrell equation is ca. 0.33 cmz.
Additional evidence that the Pt BSA was underestimated for the aryl molecule-modified
diamond films is the fact that n values calculated using the Pt ESA values determined
from the hydrogen desorption process ranged from 47-79. However, if we again assume
a Pt surface area of 0.2 cm2, n values of 3.7-4.5 were determined. It is obvious from the
data at hand that the Pt ESA calculated using the hydrogen adsorption/desorption process
is not the same area that participates in the OR. It is possible that the reduction of Pt-
oxides or oxygen concentrated within the Pt lattice contributes to the ORR current
therefore resulting in errant measurements. This significantly complicates matters in
regards to determining the reaction mechanism. Much more work is needed to
understand how structural effects influence the reaction mechanism on Pt-diamond

electrodes.

It appears that the OR on the H- and O-terminated films proceeds via a 4-
electron pathway. The question is does the mechanism proceed through a direct 4-
electron pathway to H20, as shown in equation 6.2.1, or through an indirect mechanism

as shown in equations 6.2.5 — 6.2.7 and proposed by Yeager?3°’3'

02 + 2H+ + 2e' —+ 11202 (E° = 473 mV vs. Ag/AgCl) (6.2.5)

followed by either

209

11202 + 2H+ + 2e' —> 21120 (E° = 1.57 v vs. Ag/AgCl) (6.2.6)

01'

2H202 —> 2H20 + 02 (6..27)

Oxygen reduction on Pt and Pt-alloys supported on carbon is known to proceed primarily
through the direct pathway with minor contributions from the indirect mechanism.”’32
This would be consistent with the initial step shown in equation 6.2.8 followed by the 4-

electron reduction shown in equation 6.2.1.

02 —> 02 (ads) (6.28)

Examining the standard reduction potential for equation 6.2.5, it is clear that H202
can not be produced except at potentials past the reduction peak potentials observed in
the voltammetric i—E curves (i.e., E < 473 mV). Additionally, if H202 is produced, this
species is also likely adsorbed on Pt since it appears that the reaction goes to completion
with a final n value of 4. It is suspected that the ORR on H-terminated Pt-diamond
proceeds almost entirely though the direct 4-electron pathway as with each experiment,
the H-terminated surfaces gave a more positive reduction potential, higher n value and
slightly higher k0 compared to the oxidized sample. It is likely that the OR on the
oxidized diamond films also proceeds primarily through the direct pathway; however, it

appears that the indirect pathway contributes more to the overall reaction based on the

210

calculated n value from the i-E curves. This could be due to oxygen-containing
functional groups forcing an end-on interaction between the 02 and Pt (as opposed to a
bridged interaction) by blocking certain regions of the metal and promoting H202
formation. This is similar to the theory proposed by Jovanovié et al. whereby oxygen
groups on the carbon surface promote a higher fraction of oxygen containing species in

the Pt catalyst.’8

6.2.5 Investigating Particle Stability

There have been numerous studies devoted to evaluating catalyst surface area loss
for carbon-supported platinum.”42 It is well known that metal particles, while in good
electrical contact, can be unstable on the diamond surface.2"'3‘44 In addition to examining
surface chemistry affects on metal deposition, we also investigated the causes for lost
catalytic activity with the goal of learning how to better stabilize the particles on the
diamond surface without the use of a binding agent or secondary diamond growth.
Several experiments were performed, all in 0.1 M HClO4 at room temperature. The

individual experiments and key findings are summarized below.

6 The stability of the Pt particles was assessed under open circuit conditions by exposing

the Pt/diamond film to a quiescent 0.1 M HClO4 solution in the presence and absence
of 02. In general, there was no change in the Pt ESA in the deoxygenated solution (2
50 h) while the Pt ESA decreased by 2-15% in the presence of 02, independent of the

surface chemistry or the exposure time (S 18 h). This means that catalyst detachment

211

during solution exposure is not a problem, but does suggest some slight oxidation of

the metal upon exposure to O2 saturated solutions.

Cycling the potential of the Pt/diamond (H- and O-terminated) films 1000 times

between -0.35 and 1.7 V (ca. i3 mA/cm2 (geometric area)) resulted in a Pt ESA

decrease of ca. 40-50% (11 = 2 — 4). The Pt/BMDox film lost the least amount of Pt

ESA of all the films tested; however, the loss was still significant (ca. 30%). Exposure

to these extreme potentials caused a significant reduction in the available Pt area. Two

possible causes for this are oxidation (corrosion) of the Pt at anodic potentials and

detachment of the particles during hydrogen evolution.

The particle stability on H-terminated diamond was also investigated by cycling the

potential 1000 times in deoxygenated 0.1 M HClO4 within increasingly expanded

potential limits. Overall, the Pt ESA decreased only ~l4% if the potential limits were

kept below the values where significant gas evolution occurs (-0.2 and 1.2 V).

The effect of the different gas evolution reactions was studied at freshly deposited

Pt/BMDH films (11 = 2) by holding the current at either -0.1 (H2 evolution) or +0.1 (02

evolution) A/cm2 (geometric area) for 2 h.

> Holding the current at -0.l A/cm2 for 2 h initially resulted in a 25-30% decrease in
the Pt ESA as determined using the hydrogen adsorption process, but returned to
the initial value (within 3%) after allowing the film to soak overnight in ultrapure
water. It would appear that some of the area of the rough, porous Pt particles was
initially blocked by H2 gas.
> Additionally, the response towards the OR (recorded afier full recovery of the Pt

ESA) changed as shown in Figure 6.6. Instead of exhibiting a reduction peak as

212

usual, both electrodes exhibited "microelectrode" behavior, reaching a steady-
state current (constant flux to the active sites). The different mass transfer
indicates there must be some redistribution of the particles such that they are
further spaced (i. e., individual hemispherical diffusion zones).

> Holding the current at +0.1 A/cm2 for 2 h; however, resulted in a Pt ESA decrease
of 84%. This suggests that Pt loss observed from potential cycling is primarily

due not to gas evolution but rather to metal oxidation and corrosion.

 

 

 

   

 

 

 

 

 

 

0:. n
l
2 i
.1 -10 r
v L
H
= n
o
,_ .
‘5 -20 r,--~\
U ' ‘\
x
\
-30 --- As-deposited
' V 1— After 2 h at -0.1 A/cm2
40 h r . . l r I . I n 4 L l 1 1 A l 4— 1 n
0 200 400 600 800 1000

Potential (mV vs. Ag/AgCl)

Figure 6.6. ORR linear sweep voltammetric i-E curves for Pt/BMDH electrode in 0.1 M

HClO4 before (— — —) and after (—) 2 h cathodic polarization of -0.1 A/cmz. Scan rate =
l mV/s.

6.3 Conclusions

It was previously concluded from studies of microcrystalline and nanocrystalline

diamond that the film morphology plays a role in metal phase formation by influencing

213

particle size and density.6 Additionally, it has been shown here that the surface chemistry
plays only a minor role in metal phase formation on both types of diamond. The most
relevant surface terminations to compare are hydrogen and oxygen. Rough spherical
particles were formed on both diamond film types possessing H- and O-termination but
the roughness and nominal diameter increased and the density decreased with increasing
surface oxide content, based on XPS measurements. We previously suggested that metal
deposits form preferentially on active sites and that these sites possess a differential
potential dependent activity. Oxidizing or modifying the diamond surface further
enhances this effect by decreasing the number of active sites through blockage effects.
This produces particles of lower density, larger diameter and rougher surface
morphology. Of all the films studied, the Pt/BNDH film possessed particles closest to our
target in terms of particle size and density, and exhibited the best electrocatalytic activity

towards the OR.

Interestingly, oxidizing the surface had minimal effect on the electrochemically
active surface area of the Pt, based on the recorded hydrogen adsorption/desorption
charges and the catalytic activity of the particles towards the oxygen reduction reaction.
The particles deposited on this surface possess an increase in microscopic surface
roughness. It is likely that a given set of pulse conditions results in a similar relative
mass loading and the properties of the diamond (i.e., surface chemistry, morphology,

conductivity) affect where and how the particles are deposited.

214

In addition to the effects of H- and O-terrnination, a brief look was given to the
effect of modifying the diamond surface with aryl—adlayers using the reduction of
corresponding diazonium salts. In short, the presence of the aryl-adlayer did not appear
to influence the particle size or density to a great extent. There was a slight increase in
size and decrease in coverage with the presence of the adlayer suggesting that it acts
more as a blocking layer rather than an extension of the electrode surface. The presence
of an aryl-adlayer did appear to influence the particle shape as it transitioned from a
spherical shape on the non-modified diamond surfaces to a more elongated shape with a
dendritic morphology on the modified surfaces. Interestingly, the presence of the adlayer
had a profound effect on the electrochemical behavior of the deposited metal. Despite a
similar particle size and density, the calculated ESA based on hydrogen desorption was
an order of magnitude less than on the clean diamond surfaces consistent with the adlayer
acting as a contaminant for the reaction. Additionally, the structure of the aryl-adlayer
appears to be a function of diamond morphology as particles deposited on aryl-modified
diamond in the same manner possessed very different catalytic activities. We suppose
that modifying the smoother more homogeneous nanocrystalline diamond surface likely
results in a more continuous coverage while modifying the rougher more heterogeneous
microcrystalline diamond likely results in unevenly dispersed clumps of the aryl

molecules, though this has yet to be confirmed.

It appears that putting a negative charge on the diamond surface did not promote

nor inhibit Pt complex-ion adsorption. While it is possible that this adsorption would

result in more stable particles at lower pulse numbers and at 10 pulses this advantage is

215

 

not seen, it is likely that the procedure proposed by Yan and coworkers does not work
because the Pt cation remains complexed to the Cl anions in solution making the overall
complex anionic (PtC164‘). Both these possibilities can be investigated further by (1')
putting a positive charged molecule on the diamond surface and (ii) investigating how the
deposition proceeds with each individual pulse number using electrochemical atomic

force microscopy (EC-AF M).

In addition to developing a new carbon support material to mitigate catalyst
activity loss due to support oxidation and/or corrosion, research was performed to learn
how to better stabilize the metal deposits on the diamond surface. Potential cycling
experiments suggested that decreases in the active Pt area appear to be relatively
independent of the applied potential as long as the value is within the working potential
window. This means that simple metal phase detachment in solution is not a major cause
for lost catalytic activity. This also suggests that the increase in Pt crystallite size is not
due to dissolution and re-deposition of platinum ions.”42 Hydrogen gas evolution did
not cause severe loss in activity so the turbulence of gas evolution is not problematic.
The main decrease in catalyst activity occurs under anodic conditions where the Pt can

oxidize or corrode via potential cycling beyond the limits of excessive gas evolution.

The SEM image after the H2 evolution exposure showed that the majority of the
particles were large rough agglomerates residing primarily in defect sites (i.e., grain
boundaries or intersections between two crystallite faces) on the diamond surface. One

possible explanation for this is the Ostwald ripening model in which metal crystallites

216

dissociate to metal atoms which then diffuse to and associate with larger particles.42 Bett
et al. point out that this model is frequently dismissed due to the large energy difference
between the adatom on the support and the adatom on the metal surface.42 This energy
difference would make the activation energy for an atom transfer process too large.
However, they also go on to say that Geus has argued that the presence of defect sites has
the possibility of significantly increasing the bonding strength of the adatom to the
support and thereby reducing the activation energy required for an atom transfer
process.45 The large particles in the SEM image mainly reside at defect sites suggesting
the possibility of Ostwald ripening as the major cause of particle sintering on diamond.
Additionally, the presence of mostly large particles as well as some very small particles is

also consistent with an Ostwald ripening model.”36

Significant Pt loss was only observed upon 02 evolution suggesting that the major
cause of Pt loss from diamond was due to Pt oxidation/corrosion and not physical
detachment. This Ostwald ripening mechanism has been observed by others with regards
to specific fuel cell testing.42’46’47 Since a weak Pt-diamond interaction is not the reason
for Pt loss, it seems logical to assume that the surface chemistry of the diamond would
not significantly influence Pt stability. This is consistent with what we observed here
with regards to H- vs. O-termination. Eliminating both the support material and the
metal-support interaction as the cause for catalyst degradation further supports our
conclusion that the properties of the metal are the most influential cause for decreased
activity. Therefore, research is necessary in developing a more corrosion resistant

catalyst material to further advance the use of fuel cells in real life applications.

217

10.

11.

12.

l3.

14.

REFERENCES

A. E. Fischer and G. M. Swain, J. Electrochem. Soc., 152, B369 (2005).

J. Wang, G. M. Swain, T. Tachibana, and K. Kobashi, J. New Mater.
Electrochem. Sys. , 3, 75 (2000).

J. Wang, G. M. Swain, T. Tachibana, and K. Kobashi, Electrochem. Solid-State
Lett., 3, 286 (2000).

J. Wang and G. M. Swain, Electrochem. Solid-State Lett., 5, E4 (2002).

J. Wang and G. M. Swain, J. Electrochem. Soc., 150, E24 (2003).

J. A. Bennett, Y. Show, S. Wang, and G. M. Swain, .1. Electrochem. Soc., 152,
E184 (2005).

Y. Zhang, V. Suryanarayanan, I. Nakazawa, S. Yoshihara, and T. Shirakashi,
Electrochim. Acta, 49, 5235 (2004).

C. G. Vayenas, D. Archonta, and D. Tsiplakides, J. Electroanal. Chem, 554-555,
301(2003)

J. s. Noh and J. A. Schwarz, Carbon, 28, 675 (1990).

F. Coloma, A. Sept'rlveda-Escribano, J. L. G. F ierro, and F. Rodriguez-Reinoso,
Langmuir, 10, 750 (1994).

H. E. van Dam and H. van Bekkum, J. Catal., 13], 335 (1991).

C. Prado-Burguete, A. Linares-Solano, F. Rodriguez-Reinoso, and C. Salinas-
Martiez De Lecea, J. Catal., 115, 98 (1989).

D. J. Suh, T.-J. Park, and S.-K. Ihm, Carbon, 31, 427 (1993).

F. Gloaguen, J .-M. Léger, and C. Lamy, J. Appl. Electrochem., 27, 1052 (1997).

218

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

Y. Zhang, M. L. Toebes, A. van der Eerden, W. E. O'Grady, K. P. de Jong, and D.
C. Koningsberger, J. Phys. Chem. B, 108, 18509 (2004).

M. Kang, H. K. Yoon, B. H. Kim, M. W. Song, and C. H. Lee, React. Kinet.
Catal. Lett., 80, 139 (2003).

M. C. Roman-Martinez, D. Cazorla-Amoros, A. Linares-Solano, C. Salinas-
Martiez De Lecea, H. Yamashita, and M. Anpo, Carbon, 33, 3 (1995).

V. M. Jovanovié, S. Terzié, A. V. Tripkovic', K. D. Popovic', and J. D. Lovic’,
Electrochem. Commun, 6, 1254 (2004).

T.-C. Kuo, R. L. McCreery, and G. M. Swain, Electrochem. Solid-State Lett., 2,
288 (1999).

J. Wang, M. A. Firestone, O. Auciello, and J. A. Carlisle, Langmuir, 20, 11450
(2004).

D. Knigge, P. Kaur, and G. M. Swain, Unpublished results.

M. M. Waje, X. Wang, W. Li, and Y. Yan, Nanotechnology, 16, S395 (2005).

L. Zhang, T. Sakai, H. Yoshida, S. Yamanaka, and H. Okushi, J. Appl. Phys., 91,
4585(2002)

A. J. Bard and L. R. Faulkner, Electrochemical Methods: Fundamentals and
Applications, John Wiley & Sons, Inc., New York (2001).

K. Kinoshita, Electrochemical Oxygen Technology, John Wiley & Sons, Inc.,
New York (1992).

J. A. Poirier and G. E. Stoner, J. Electrochem. Soc., 141, 425 (1994).

U. A. Paulus, T. J. Schmidt, H. A. Gasteiger, and R. J. Behm, J. Electroanal.
Chem, 495, 134 (2001).

219

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

N. Wakabayashi, M. Takeichi, M. Itagaki, H. Uchida, and M. Watanabe, J.
Electroanal. Chem, 574, 339 (2005).

N. M. Markovic, R. R. Adzic’, B. D. Cahan, and E. B. Yeager, J. Electroanal.
Chem, 377, 249 (1994).

E. Yeager, Electrochim. Acta, 29, 1527 (1984).

E. Yeager, J. Mol. Catal., 38, 5 (1986).

M. Inaba, H. Yamada, J. Tokunaga, and A. Tasaka, Electrochem. Solid—State
Lett., 7, A474 (2004).

E. Antolini, J. Mater. Sci., 38, 2995 (2003).

M. S. Wilson, F. H. Garzon, K. E. Sickafirs, and S. Gottesfeld, J. Electrochem.
Soc., 140, 2872 (1993).

Y. Mo, S. Sarangapani, A. Le, and D. A. Scherson, J. Electroanal. Chem, 538-
539, 35 (2002).

G. A. Gruver, R. F. Pascoe, and H. R. Kunz, J. Electrochem. Soc., 127, 1219
(1980).

P. Bindra, S. J. Clouser, and E. Yeager, J. Electrochem. Soc., 126, 1631 (1979).

A. C. C. Tseung and S. C. Dhara, Electrochim. Acta, 20, 681 (1975).

L. J. Bregoli, Electrochim. Acta, 23, 489 (1978).

K. F. Blurton, H. R. Kunz, and D. R. Rutt, Electrochim. Acta, 23, 183 (1978).

A. Honji, T. Mori, K. Tamura, and Y. Hishinuma, J. Electrochem. Soc., 135, 355
(1988)

J. A. S. Bett, K. Kinoshita, and P. Stonehart, J. Catal., 41, 124 (1976).

220

43.

44.

45.

46.

47.

S. Nakabayashi, D. A. Tryk, A. Fujishima, and N. Ohta, Chem. Phys. Lett., 300,
409 (1999).

N. Vinokur, B. Miller, Y. Avyigal, and R. Kalish, J. Electrochem. Soc., 146, 125
(1999)

J. W. Geus, in Chemisorption and Reactions on Metallic Films, J. R. Anderson,
Editor, 1, p. 129, Academic Press, New York, (1971).

P. J. Ferreira, G. J. la 0', Y. Shao-Horn, D. Morgan, R. Makharia, S. Kocha, and
H. A. Gasteiger, J. Electrochem. Soc., 152, A2256 (2005).

R. M. Darling and J. P. Meyers, J. Electrochem. Soc., 150, A1523 (2003).

221

CHAPTER 7

Key Findings and Future Research Directions

7.1 Key Findings

Growing concern over environmental impacts, limited fuel supplies and reliance
on foreign oil has lead to considerable interest in fuel cells as a clean, renewable energy
alternative for stationary and portable power applications. While there have been
considerable advances made in fuel cell materials research in the last decade, there are
still several technological hurdles that need to be overcome before fuel cells can become
commercially viable. Two hurdles are (1') catalyst cost and activity loss which must be
reduced through decreased metal loading and development of non-noble metal catalysts
and (ii) prevention of carbon support degradation through the development of more

dimensionally stable and corrosion resistant electrocatalyst supports.

The research conducted for this dissertation had the goal of advancing diamond as
a state-of-the—art electrocatalyst support material through a better understanding of metal
phase formation, electrocatalytic activity and stabilization. The key findings of this

research are summarized as follows:
1. The physical and electrochemical properties of diamond can be easily altered by

adjusting the CH4/H2 deposition source gas ratio. Increasing the CH4/H2 ratio increases

the rate of secondary nucleation leading to an increase in the fraction of grain boundary

222

exposed, a decrease in both the diamond crystallite size and overall film roughness, and
an increasing amount of spz-bonded nondiamond carbon impurity incorporated into the
film microstructure. The presence of this spz-bonded nondiamond carbon strongly

F e3+/2+, 4-tert-

influences the electron-transfer kinetics for some reactions (e. g.,
butylcatechol, and the OR in acid and alkaline solution) and exerts little influence on

others (e. g., Fe(CN)63’/4' and Ru(NH3)63+/2+).

2. Boron-doped microcrystalline and nanocrystalline diamond are both structurally
stable under polarization conditions commonly encountered in PAFCs. Both diamond
types exhibit some surface oxidation resulting in slight roughening of the crystallite
edges; however, no gross morphological degradation occurs. Additionally, any alteration
in the electrochemical response of the material towards several well known redox couples
is reversed by rehydrogenation. The only irreversible change seen is an increase in
background voltammetric current exhibited by the nanocrystalline diamond film. It
appears that for these thin, nanograined films, solution can penetrate the grain boundaries
and react with the underlying Si. This penetration is enabled by the minor structural

changes that occur during polarization.

3. The use of pulsed galvanostatic deposition produces metal particles with a small
nominal particle diameter and a high particle density. This method is superior to constant
current/potential deposition methods reported on by our group in the past. Deposition
using ten l-s pulses at 1.25 mA/cm2 yields the most ideal electrocatalyst on both diamond

types in terms of the nominal particle size (ca. 30-40 nm), particle coverage (109-1010 cm'

223

2) and the highest specific Pt surface area (ca. 25-50 mz/gpt). The heterogeneous diamond
surface possesses sites that exhibit a differential potential dependent activity towards
metal phase formation. Additionally, the Pt nucleation mechanism on diamond
transitions from an instantaneous mechanism at low overpotentials to a more progressive

one at high overpotentials.

4. The diamond surface chemistry plays a minor role in metal phase formation or
stability. The roughness and nominal particle diameter increases and the density
decreases with increasing surface oxide content. Oxidizing or modifying the diamond
surface decreases the number of active sites available for deposition through blocking
effects. Pt/BNDH exhibits the highest catalytic activity for the oxygen reduction reaction.
Finally, the loss of Pt catalyst from the diamond surface is not due to degradation of the
diamond or physical displacement of the particles, but rather due to oxidation and

corrosion of the metal. The results are consistent with an Ostwald ripening mechanism.

7 .2 Future Research Needs

The research advanced diamond as an advanced electrocatalyst support material;
however, a number of issues remain to be determined. Some future research topics are

given below.

1. While diamond shows promise as a dimensionally stable electrocatalyst support
material, further work is required to produce the material in a high surface area form

(~100 mz/g). A step towards this goal may require the incorporation of spz-bonded

224

nondiamond carbon impurity through increases in the CH4/H2 source gas ratio. A mixed
sp3/sp2 material may be easier to deposit on high surface area materials. Additionally,
diamond is a hydrophobic material and increasing the hydrophilicity of the material
through spz-bonded nondiamond carbon incorporation may result in better contact with
the Nafion membrane and thus reduce the resistance in the membrane electrode assembly

while still maintaining satisfactory electrical conductivity and dimensional stability.

2. Additional optimization of the metal phase formation by studying the effect of
pulse width and duty cycle. An informative way to study metal adlayer formation on
diamond is to image the surface in situ using electrochemical-atomic force microscopy
(EC-AFM). This method will (i) aide in determining the exact nucleation and growth
mechanism of Pt on diamond, (ii) allow for the real-time monitoring of particle growth
with each pulse so the pulse parameters can be more easily optimized and (iii) allow for
the monitoring of any changes in the particles with electrochemical cycling (i. e., particles
moving during cyclic voltammetric cleaning). Perhaps the particles change when we do
the potential cycling to activate the Pt and we only see the end result when viewing using

SEM.

3. Additionally, a procedure for electrodepositing metal particles on high surface
area forms of carbon (e.g., diamond) needs to be developed. Two approaches are being
followed by the group: overcoating conducting diamond on diamond polishing grit and
on carbon paper. The architecture of the high surface area carbon will likely affect the

metal deposition process, especially in terms of wettability and current density.

225

However, an increased support area will likely make it easier to deposit particles with the
desired properties. We have designed electrochemical cells that could house powderous
diamond for electrochemical testing. Two ideas that appear promising are based on the

works of Tran and Langerl or Porter and coworkers;3

4. Work is needed to reduce the metal loading while still maintaining high
electrocatalytic activity. Also, alternate electrocatalysts need to be discovered. Perhaps
the deposition of Pt-alloys or Pt on non-noble catalysts will help to reduce both Pt
particle size and loading while increasing the catalyst activity. Additionally, if the
electrodeposition process can not be optimized, alternative deposition methods must be
investigated. Possible alternatives include impregnation, microemulsion or colloid. A
key factor in optimizing the catalyst layer is understanding the factors that influence the

activity of the deposited catalyst towards H2 oxidation and 02 reduction.

5. Understanding the interaction between the catalyst particle and the carbon support
material is important for particle stabilization on the diamond surface. So far, a
secondary diamond grth appears to be the most promising method, however as
mentioned previously, microcrystalline diamond has crystallites that are too large for our
target particles. Therefore a secondary nanocrystalline diamond grth procedure should
be investigated as a method for stabilizing particles < 30 nm as well as particles on a high

surface area form of diamond.

226

6. Finally, membrane electrode assemblies (MEAs) must be developed and tested in
actual PEMFCs and PAFCs. The stability and activity of the electrocatalyst particles

must be monitored and compared to that for commonly used fuel cell electrodes.

227

REFERENCES

T. D. Tran and S. H. Langer, Anal. Chem, 65, 1805 (1983).

R. S. Deinhammer, E.-Y. Ting, and M. D. Porter, Anal. Chem, 67, 237 (1995).

J. A. Hamisch, D. B. Gazda, J. W. Anderegg, and M. D. Porter, Anal. Chem, 73,
3954 (2001).

228

APPENDIX A

 

 

 

 

 

 

 

 

 

Figure A.1. Front panel and schematic code for the potential pulse generator program
created in LabVIEW Version 6.1 in conjunction with a National Instruments PCI—5401
card. The parameters in the top window were those used to deposit Pt using 10 pulses of
l s (delay 1 s for 50% duty cycle) and a pulse current density of 1.25 rnA/cmz.

229

MICHIGAN STA TE UNI

111111111111 111111111131

3 1293 02845 5