gawk" g, A

.1... .2

 

    

       

 

x
,me..an «

v1. 5 .
agau. 35.2
:1”.

.45 :u., »
FL.

  

.v
5;.
I :5

     

 

 

        

..,?x.r..za......r.
:t. .9. A fl.

‘7
Z

 

 

 

 

 

 

.. V -‘.p..-¢..n~san

 

 

 

.. : , m . ‘
u . , : ‘

 

 

This is to certify that the

dissertation entitled

UVN IS AND IR SPECTROELECTROCHEMICAL INVESTIGATIONS OF
SMALL MOLECULES AND REDOX-ACT IVE BIOMOLECULES
USING OPTICALLY TRANSPARENT DIAMOND ELECTRODES

presented by

Shannon {Marie Haymonzf

has been accepted towards fulfillment
of the requirements for

PhD. degree in Chemistry

 

 

Jug/14%

V a Major pr'ofessor

Date 13 November 2002

 

MS U is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

W;

MiChiQan 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

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 cJCIFiC/DateDuepssosz

-w-HW

UVNIS AND IR SPECTROELECTROCHEMICAL INVESTIGATIONS OF SMALL
MOLECULES AND REDOX-ACTIVE BIOMOLECULES USING OPTICALLY
TRANSPARENT DIAMOND ELECTRODES
By

Shannon Marie Haymond

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

2002

ABSTRACT

UVNIS AND IR SPECTROELECTROCHEMICAL INVESTIGATIONS OF SMALL
MOLECULES AND REDOX-ACTIVE BIOMOLECULES USING OPTICALLY
TRANSPARENT DIAMOND ELECTRODES

By

Shannon Marie Haymond

Spectroelectrochemistry combines two very different techniques—
spectroscopy and electrochemistry—providing more information about the redox
mechanism of an electroactive analyte than can be acquired using traditional
electrochemical methods alone. The oxidation state of an analyte is changed
electrochemically by adding or removing electrons at an optically transparent
electrode (OTE) while the spectroscopic response of the species in solution is
simultaneously measured. Spectroelectrochemistry is useful for obtaining
spectra and redox potentials for electrogenerated species as well as important
information in the investigation of electrode reaction mechanisms, in
electroanalytical applications, and, as demonstrated herein, in elucidating
structural information about redox-active protein active sites.

The use of electrically conductive diamond as an OTE is a new area of

research. Boron-doped diamond (BDD) possesses attractive qualities as both an

electrode and an optically transparent material, making it an obvious choice for
utilization as an OTE in transmission spectroelectrochemical measurements.
Diamond OTES offer several advantages over other materials including: (i) the
possibility of optical measurements from the near-UV into the far-IR (0.25—
100 pm), (ii) a low background current,‘(iii) a wide working potential window,
(iv) stability in aqueous and nonaqueous solution environments during both
cathodic and anodic polarization, and (v) resistance to fouling.

An added advantage of BDD OTES is the versatility of fabrication, as BDD
thin-films can be grown on a variety of substrates (e.g., quartz, undoped Si, and
white diamond grown by chemical vapor deposition), and can also exist as
freestanding disks. Each type of OTE possesses characteristics that can be
exploited for a specific application. This work is a product of our long-term goal
to develop BDD as an OTE for electroanalytical applications and the study of
biological electron transfer mechanisms. Development of the OTE comprised the
design and testing of thin-layer spectroelectrochemical cells and the
electrochemical and spectroelectrochemical characterization of two systems: (i) a
small molecule, ferrocene, in nonaqueous solvent/electrolyte systems, and (ii) a
redox-active protein, cytochrome 0. Besides the basic electrochemical and
spectroelectrochemical characterizations of these systems, this work focused on
determining the deposition conditions that produce an optimized balance

between electrical conductivity and optical transparency in this material.

to Jerry

ACKNOWLEDGEMENTS

This journey was in no way a lonely one, and for their various
contributions to my scientific progress or social well being, I would like to
acknowledge the following people:

To the past and present members of the Art’s Softball Team for many fun
times and amusing memories on, and off of the field at a fine establishment
where peaches are served with beer, and people always win at the skill crane
and sing loudly with the jukebox. To my co-hostesses and co-counsel, the
notorious Asphalt Angel Hudy and Sweet Cheeks Lamp: I consider our
transformation of 5077 Wardcliff into the “Hog Wyld Roadhouse” and
subsequent, victorious legal battle to be among my greatest accomplishments!
You two have been great friends and roommates and I will miss you both.

To the current members of the Babcock group for displaying the guts and
composure to make the best of a bad situation and for their input and assistance
throughout my time here. And I can’t forget my former labmates and friends that
have gone on to further greatness, Kristi, Heather, and Jose, although you guys
are out of sight, you will never be out of mind. I think I speak for the entire
Babcock group when I give my heartfelt thanks to our personal grammar lady,
Vada O’Donnell, for keeping us true to Jerry’s values of pristinely correct
grammar usage, and for always bringing out the best in us. Additionally, I’d like

to thank the members of the Swain group for their help and support. I am

fortunate to have gone through this dissertation writing and defending process
with my dear friends Maggie and Matt.

For their scientific collaboration, I am indebted to following people, as their
contributions were significant and integral to the work presented herein:
Professor Jerzy Zak for sharing with me his expertise in spectroelectrochemical
cell design and thin-layer electrochemistry, to Zuzana Cvackova for continuing
the voltammetric study of cytochrome c at boron-doped diamond electrodes, to
Professor Shelagh Ferguson-Miller, Dr. Denise Mills, and Bryan Schmidt for
demonstrating interest in my projects and, thus, providing me with high-caliber
protein samples, and to Glenn Wesley for his ability to turn my crude designs into
functioning pieces. To Dr. Warwick Hillier for his guidance, encouragement, and
contributions to my scientific development. Thanks for answering all of my
mundane questions and teaching me that sometimes turning the “do not touch”
knob is allowed. I am fortunate that Jerry hired you “for my amusement”. Good
luck on your bowling game, Pip.

To Professor Jerry Babcock for instilling in me a desire to be great at what
I enjoy, scientific and otherwise. I have strived to carry on as you wished, and for
that, I leave here more daring and resourceful. To my other great advisor,
Professor Greg Swain, for providing stability and support during a difficult
transition. I would also like to thank you for the guidance and encouragement
you provided as you took over the presidential role of the SH fan club.

To my family for their pride in me, for all that I choose to do. Without you

all, I would have less to my credit, personally and otherwise. Thank you for your

vi

unending support. To my love, Ryan, for indulging my stubbornness and
Independence while always pulling me back up when I called from the basement
of the chemistry building, literally, in the pits of despair. Thanks for always

making me laugh and for showing continuous faith in my pursuit of this degree.

vii

TABLE OF CONTENTS

LIST OF TABLES ................................................................................................. xi
LIST OF FIGURES ............................................................................................. xiii
CHAPTER 1
1 Introduction
1.1 Insight into Metalloprotein Mechanisms .................................................... 1
1.2 Spectroelectrochemical Methods .............................................................. 3
1.3 Optically Transparent Electrodes ............................................................ 10
1.3.1 Conventional Types .......................................................................... 10
1.3.2 A New Material: Boron-Doped Diamond ........................................... 11
1.4 Scientific Relevance ................................................................................ 14
1.5 Dissertation Outline ................................................................................. 16
1.6 References .............................................................................................. 19
CHAPTER 2
2 Materials and Methods
2.1 Boron-Doped Diamond Growth Conditions ............................................. 34
2.1.1 Microcrystalline Boron-Doped Diamond ............................................ 34
2.1.2 Nanocrystalline Boron-Doped Diamond ............................................ 38
2.1.3 Freestanding Boron-Doped Diamond OTE ....................................... 38
2.1.4 Boron-Doped Diamond IR OTE ........................................................ 40
2.2 Sample Preparation ................................................................................ 40
2.2.1 Ferrocene .......................................................................................... 40
2.2.2 Cytochrome c .................................................................................... 41
2.2.3 Cytochrome c Oxidase ...................................................................... 41
2.3 Cyclic Voltammetric Experiments ............................................................ 42
2.4 Spectroelectrochemistry .......................................................................... 44
2.4.1 000 Experiments - Au Grid OTTLE Cell ........................................... 44
2.4.2 BDD UV/vis OTTLE Cell ................................................................... 48
2.4.3 BDD IR OTTLE Cell I ........................................................................ 50
2.4.4 800 IR OTTLE Cell ll ....................................................................... 53
2.5 Temperature-Control Set Up ................................................................... 54
2.6 References .............................................................................................. 58

viii

CHAPTER 3
3 Ferrocene Electron Transfer Kinetics Measured at Boron-Doped Diamond

Electrodes

3.1 Introduction ............................................................................................. 60
3.2 Results .................................................................................................... 64
3.3 Conclusions ............................................................................................ 72
3.4 References .............................................................................................. 73
CHAPTER 4
4 Direct Voltammetry of Cytochrome c at Boron-Doped Diamond Electrodes
4.1 Introduction ............................................................................................. 76
4.2 Results .................................................................................................... 81
4.2.1 Microcrystalline BDD Thin-Film Electrodes ....................................... 82
4.2.2 Nanocrystalline BDD Thin-Film Electrodes ....................................... 94
4.3 Discussion ............................................................................................. 102
4.4 Conclusions .......................................................................................... 1 1 1
4.5 References ............................................................................................ 1 12
CHAPTER 5

5 Low-Frequency Electrochemical Difference FTIR Spectroscopy of
Metalloproteins: Challenges and Solutions

5.1 Introduction ........................................................................................... 1 17
5.2 Structural lnforrnation Possible (Far-IR Region) ................................... 119
5.3 FTIR of Aqueous Samples .................................................................... 122
5.4 Detectors ............................................................................................... 132
5.5 Window Materials .................................................................................. 136
5.6 Temperature Control ............................................................................. 138
5.7 Conclusions .......................................................................................... 140
5.8 References ............................................................................................ 143
CHAPTER 6
6 Electrochemical Difference FTIR Spectroscopy of Cytochrome c Oxidase

6.1 Introduction ........................................................................................... 147
6.2 Results .................................................................................................. 152

6.2.1 Electrochemical Difference Spectra ................................................ 152

6.2.2 Comparison of Bacterial and Mammalian Forms of CcO ................ 159

6.2.3 IsotopicaIIy-Labeled Spectra ........................................................... 166
6.3 Discussion ............................................................................................. 173

6.3.1 Carbonyl Region ............................................................................. 173

6.3.2 Histidine Modes .............................................................................. 176
6.4 Implications ........................................................................................... 185
6.5 Conclusions .......................................................................................... 187
6.6 References ............................................................................................ 190

CHAPTER 7
7 Spectroelectrochemical Studies Using a Novel Optically Transparent
Electrode: Boron-Doped Diamond

7.1 Introduction ........................................................................................... 198
7.2 Results - UV/vis Spectroelectrochemistry ............................................. 206
7.2.1 Optical Properties of BDD OTES — UV/vis ....................................... 206
7.2.2 Electrochemical Characterization of the UV/vis OTTLE Cell ........... 211
7.2.3 Ferrocene ........................................................................................ 212
7.2.4 Cytochrome c .................................................................................. 219
7.3 Results - IR Spectroelectrochemistry .................................................... 223
7.3.1 Optical Properties of BDD OTES — IR ............................................. 223
7.3.2 Ferrocene ........................................................................................ 228
7.3.3 Electrochemical Characterization of the IR OTTLE Cell .................. 230
7.3.4 Ferrocyanide ................................................................................... 233
7.3.5 Cytochrome c .................................................................................. 234
7.4 Conclusions .......................................................................................... 237
7.5 References ............................................................................................ 240

Table 2—1

Table 3—1

Table 4-1

Table 4—2

Table 4—3

Table 5—1

Table 5—2

Table 5-3

Table 6—1

Table 6—2

LIST OF TABLES

Mediators Used in Spectroelectrochemical Measurements of
Cytochrome c Oxidase .................................................................. 47

Apparent Heterogeneous Electron Transfer Rate Constants for
Ferrocene at Untreated Boron-Doped Diamond and Polished
Glassy Carbon Electrodes ............................................................. 69

Effect of Cytochrome 0 Exposure on Test Analyte Cyclic
Voltammetric Peak Splitting for Acid-Washed Rehydrogenated
10 ppm Microcrystalline Boron-Doped Diamond Electrodes ......... 93

Effect of Cytochrome 0 Exposure on Test Analyte Cyclic
Voltammetric Peak Splitting for Acid-Washed Rehydrogenated
10 ppm Nanocrystalline Boron-Doped Diamond Electrodes ........ 102

Apparent Heterogeneous Electron Transfer Rate Constants
Measured from the Direct Voltammetry of Cytochrome c at Solid

Electrodes ................................................................................... 105
Low-Frequency Vibrational Modes of Metalloenzymes ............... 121
Infrared Properties of H20 and 020 ............................................ 123
Properties of Infrared Transmissive Materials ............................. 138

Discrepant Modes and Tentative Assignments for FTIR
Electrochemical Difference Spectra of Cytochrome c Oxidases.. 162

15N Sensitive Peaks in Rhodobacter Sphaeroides Cytochrome c
Oxidase FTIR Spectra ................................................................. 169

xi

Table 6—3

Table 6—4

Table 6-5

Table 6—6

Table 6—7

Table 7—1

CSN1 Stretching Modes for Histidine and 4-Melm ...................... 181

CSN1 Stretching Modes for Metal-Bound Histidine and 4-Melm .181

C4C5 Stretching Modes for N8 Protonated Histidine and 4-Melm
.................................................................................................... 182

C4C5 Stretching Modes for Metal-Bound N8 Protonated Histidine
and 4-Melm ................................................................................. 182

Metal-Histidine Stretching Modes ................................................ 186

Effect of Selected Growth Parameters on the Optical and Electronic
Properties of Boron-Doped Diamond Optically Transparent
Electrodes ................................................................................... 201

xii

Figure 1-1

Figure 2-1

Figure 2-2

Figure 2-3

Figure 2-4

Figure 2-5

Figure 2-6

Figure 2-7

Figure 3-1

Figure 3-2

Figure 4-1

Figure 4-2

LIST OF FIGURES

Spectroelectrochemical Techniques Using Absorption
Spectroscopies ................................................................................ 9

Conventional 3-Electrode Glass Electrochemical Cell ................... 43

Optically Transparent Thin-Layer Electrochemical Cell Design
Employing a Au Grid Optically Transparent Electrode ................... 45

UV/vis Spectroelectrochemical Cell Employing a Freestanding
Boron-Doped Diamond Optically Transparent Electrode ............... 49

IR Spectroelectrochemical Cell Employing a Boron-Doped Diamond
Optically Transparent Electrode -— Model I ..................................... 51

IR Spectroelectrochemical Cell Employing a Boron-Doped Diamond

Optically Transparent Electrode - Model II .................................... 53
Temperature-Control Set Up ......................................................... 55
Cryostat Box Diagrams .................................................................. 56
Simulated Cyclic Voltammogram — Diffusion-Limited Case ........... 63

Ferrocene Cyclic Voltammetric Response at a Boron-Doped
Diamond Electrode ........................................................................ 66

Cytochrome c Heme Redox Center ............................................... 78

Cyclic Voltammetry of Cytochrome c at “As Deposited” 1 ppm and
10 ppm Microcrystalline Boron-Doped Diamond Electrodes ......... 83

xiii

Figure 4-3

Figure 4-4

Figure 4-5

Figure 4-6

Figure 4-7

Figure 4-8

Figure 4-9

Figure 4-10

Figure 4-11

Figure 4-12

Figure 4-13

Cyclic Voltammetry of Cytochrome c at “As Deposited”
Microcrystalline Boron-Doped Diamond Electrodes ...................... 84

Cyclic Voltammetry of Cytochrome c at “Acid-Washed,
Rehydrogenated” Microcrystalline Boron-Doped Diamond
Electrodes ..................................................................................... 85

Demonstration of the Cyclic Voltammetric Response Stability for
Cytochrome c at a Microcrystalline Boron-Doped Diamond
Electrode ....................................................................................... 86

Cyclic Voltammetric Response for Cytochrome 0 Measured at a
Microcrystalline Boron-Doped Diamond Electrode ........................ 88

Peak Current Dependence on the Square Root of the Scan Rate
Measured for Cytochrome c at a Microcrystalline Boron-Doped
Diamond Electrode ........................................................................ 90

Peak Current Dependence on the Concentration Measured for
Cytochrome c at a Microcrystalline Boron-Doped Diamond
Electrode ....................................................................................... 91

Cyclic Voltammetric Background Curves Measured for a Boron-
Doped Diamond Electrode after Exposure to Cytochrome c ......... 92

Cyclic Voltammetry of Cytochrome c at “As Deposited" 1 ppm and
10 ppm Nanocrystalline Boron-Doped Diamond Electrodes .......... 94

Cyclic Voltammetry of Cytochrome c at “As Deposited”
Nanocrystalline Boron-Doped Diamond Electrodes ...................... 95

Cyclic Voltammetry of Cytochrome c at “Acid-Washed,
Rehydrogenated" Nanocrystalline Boron-Doped Diamond
Electrodes ..................................................................................... 96

Demonstration of the Cyclic Voltammetric Response Stability for
Cytochrome c at a Nanocrystalline Boron-Doped Diamond
Electrode ....................................................................................... 97

xiv

Figure 4-14

Figure 4-15

Figure 4-16

Figure 5-1

Figure 5-2

Figure 5-3

Figure 5-4

Figure 6-1

Figure 6-2

Figure 6-3

Figure 6-4

Figure 6-5

Figure 6-6

Cyclic Voltammetric Response for Cytochrome 0 Measured at a
Nanocrystalline Boron-Doped Diamond Electrode ........................ 98

Reduction Peak Current Dependence on the Square Root of the
Scan Rate Measured for Cytochrome c at a Nanocrystalline Boron-
Doped Diamond Electrode .......................................................... 100

Reduction Peak Current Dependence on the Concentration
Measured for Cytochrome c at a Nanocrystalline Boron-Doped
Diamond Electrode ...................................................................... 101

Infrared Absorption Spectra of H20 and D20 .............................. 123

Comparison of the H20 Absorption Spectrum and the Noise
Spectrum for an Aqueous Sample ............................................... 124

Theoretical Construction of an Electrochemical Difference FTIR

Spectrum ..................................................................................... 129
Effect of Temperature Drift on Baseline Linearity ........................ 139
Cytochrome c Oxidase from Rhodobacter Sphaeroides .............. 148

UV/vis Electrochemical Difference Spectra of Rhodobacter
Sphaeroides Cytochrome c Oxidase ........................................... 153

FTIR Electrochemical Difference Spectra of Rhodobacter
Sphaeroides Cytochrome c Oxidase ........................................... 156

Illustration of Characteristic Mid-IR Bands ................................... 158

Comparison of the FTIR Difference Spectra of Bacterial and
Mammalian Cytochrome c Oxidases ........................................... 161

Comparison of the FTIR Difference Spectra of Cytochrome c
Oxidase Measured in D20 and H20 ............................................ 167

XV

Figure 6-7

Figure 6-8

Figure 6-9

Figure 6-10

Figure 6-11

Figure 7-1

Figure 7-2

Figure 7-3

Figure 7-4

Figure 7-5

Figure 7-6

Figure 7—7

Figure 7-8

Effect of Global 15N Labeling on Mid-IR FTIR Spectra of

Cytochrome c Oxidase ................................................................ 168
Isotope Exchange Effects in the Amide II Region ........................ 170
Deuterium Exchange Effects in the Carbonyl Region .................. 174
Closer Examination of a Possible Histidine Mode ....................... 177
Structures of the Two Neutral Histidine Tautomers ..................... 179
Simulated Cyclic Voltammogram - Thin-Layer Case ................... 204

UV/vis Transmission Spectra of Freestanding Boron-Doped and
CVD White Diamond Disks: Effect of Boron Doping Level .......... 208

UV/vis Transmission Spectra of Optically Transparent Electrodes:
Boron-Doped Diamond compared to Indium-Doped Tin Oxide ...209

Ferrocene Cyclic Voltammetric Response Measured in the UV/vis
Optically Transparent Thin-Layer Electrochemical Cell ............... 211

UV/vis Spectroelectrochemical Absorbance Spectra of Ferrocene
Oxidation at a Freestanding Boron-Doped Diamond Optically
Transparent Electrode ................................................................. 213

Nernst Plot for the UV/vis Spectroelectrochemistry of Ferrocene at
a Freestanding Boron-Doped Diamond Optically Transparent
Electrode ..................................................................................... 215

Absorbance-Potential (Voltabsorptometry) Profiles of Ferrocene
UV/vis Spectroelectrochemistry at a Freestanding Boron-Doped
Diamond Optically Transparent Electrode ................................... 216

Concentration Dependence for the Electrogenerated Ferricenium
UV/vis Absorbance Peaks ........................................................... 218

xvi

Figure 7-9

Figure 7-10

Figure 7-11

Figure 7-12

Figure 7-13

Figure 7-14

Figure 7-15

Figure 7-16

Figure 7-17

UV/vis Spectroelectrochemical Absorbance Spectra of
Cytochrome c at a Freestanding Boron-Doped Diamond Optically
Transparent Electrode ................................................................. 220

Direct Electrochemical Redox Titration of Cytochrome c at a
Freestanding Boron-Doped Diamond Optically Transparent
Electrode ..................................................................................... 222

The Effect of Boron Doping Level on the IR Transmission of
Diamond Optically Transparent Electrodes ................................. 223

IR Transmission Spectra of Different Types of Boron-Doped
Diamond Optically Transparent Electrodes ................................. 226

IR Spectroelectrochemical Absorbance Spectra of
Ferrocene/Ferricenium Measured at a Boron-Doped Diamond
Optically Transparent Electrode — Mid-IR .................................... 229

IR Spectroelectrochemical Absorbance Spectra of
Ferrocene/Ferricenium Measured at a Boron-Doped Diamond
Optically Transparent Electrode - Far-IR .................................... 230

Ferrocyanide Cyclic Voltammetric Response Measured in the IR
Optically Transparent Thin-Layer Electrochemical Cell (Model II)
.................................................................................................... 231

IR Spectroelectrochemical Absorbance Spectra of
Ferri-lFerrocyanide Measured at a Boron-Doped Diamond Optically
Transparent Electrode ................................................................. 233

Comparison of the Mid-IR Spectra of Cytochrome 0 Measured at a
Au Grid and at a Boron-Doped Diamond Optically Transparent
Electrode ..................................................................................... 235

xvii

CHAPTER 1

1 Introduction

1.1 Insight into Metalloprotein Mechanisms

Metals in biology are divided into two classes. The first class consists of
highly concentrated ions, such as K, Na”, M92+, and Ca2+, which are important

in maintaining the structure of proteins and in regulating cell membrane function.
The second class consists of the so-called metalloproteins, in which low levels of
ionic forms of Mn, Fe, Co, Cu, Zn, Mo, and so on, are incorporated Into the
protein. Depending on function, metalloproteins are further divided into two
categories: (A) transport and storage proteins and (B) enzymes. Type A includes
Fe-containing oxygen-transport proteins such as hemoglobin, myoglobin,
hemerythrin, and the Cu-centered hemocyanine; electron-transfer proteins such
as cytochromes (Fe), iron-sulfur proteins (Fe), and blue-copper proteins (Cu);
and metal storage proteins such as ferritin (Fe) and ceruloplasmin (Cu). The
enzyme category, type B, includes hydrolases such as carboxypeptidase (Zn)
and aminopeptidase (Zn, Mg); oxidoreductases such as the heme-copper

oxidases (Fe, Cu, Mg, Mo), photosynthetic enzymes (Mn, Ca), and nitrogenase
(Mo, Fe); and isomerases such as vitamin B12 coenzyme (Co).

Understanding the role Of metal ions in metalloproteins first requires
knowledge of the coordination chemistry (structure and bonding) at the active

sites. Active sites are the discrete regions of a protein where the chemistry

occurs. Detailed structural information is difficult to obtain because these sites
are usually buried amid a complex protein structure. X-ray crystallography would
be the method of choice for this purpose; however, its application is often
hampered by the difficulties in growing single crystals of high molecular weight
protein molecules and in analyzing diffraction data with high resolution. For a
number of proteins, these difficulties have been overcome, and knowledge of
precise geometries from x-ray crystallographic data has made a great
contribution to our understanding of their biological functions [1, 2]. The
crystallization of membrane proteins appears to be the exception, as little x-ray
structural information is available or definitive [3-5]. In such cases,
complementary information about the structure, bonding and environment of the
metal is acquired via a variety of physiochemical measurement techniques.
These include electronic, infrared (IR), resonance Raman, magnetic, nuclear,
circular dichroism, and Mossbauer spectroscopies, and electrochemical,
thermodynamic and kinetic measurements.

Information on the interaction of participating molecules is required to
understand the molecular mechanisms of enzymes. Infrared spectroscopy has
been essential in addressing such problems [6, 7]. Often, only a small part of the
large enzyme actually participates in the catalytic reactions. Infrared difference
spectra, which result as the difference of two unique states of the enzyme, report
only on the modes that undergo a change upon the transition. For small
proteins, direct subtraction of the absorbance spectra of the two states, or

analysis by band-narrowing techniques [8-10] are successful approaches for

obtaining difference spectra. Alternatively, investigation of larger, more complex
proteins requires in situ methods, whereby a trigger initiates the transition from
one state of the protein to the next, directly in the infrared cell. When combined
with sensitive instrumentation, this approach allows for the accurate detection of
the subtle (0.1% of the total absorbance) changes in the vibrational spectrum that
are associated with enzyme catalysis [6].

Reaction-induced difference spectroscopy is the term given to such a
method, in which a trigger is used to advance a molecule from one state to
another in an effort to obtain a difference spectrum of the unique states. There
are several options for performing reaction-induced difference spectroscopy,
including indirect [11, 12] and direct photochemical reactions [13, 14], and
electron transfer reactions [15]. Metalloproteins are well suited for the
electrochemical difference method, as many contain redox-active metal centers.
One of the advantages of using electrochemistry as the trigger is its generality of
application, as there is no requirement for highly absorbing chromophores or for
a light-initiated reaction cycle—only the need for a redox-active metal. For
example, this method has proven useful for studies of the proteins involved in

photosynthesis [16] and respiration [17-19].

1.2 Spectroelectrochemical Methods

The combination of electrochemistry and spectroscopy—so-called
spectroelectrochemistry—has been in use for almost forty years. Kuwana’s
pioneering approach coupled UV/vis spectroscopy with an electrochemical cell to

monitor solution concentrations of dissolved redox species [20]. These results

were revolutionary, in that, by this method, it was possible to obtain time- and
potential-dependent spectra of reaction components from which one could
elucidate reaction mechanisms and kinetic information. Without the additional,
optical probe, previous experiments of this type were much more difficult. Over
the next several years, Kuwana’s group [21-25], and others [26-38], continued to
report on UV/vis Spectroelectrochemical studies of solution species, including
those of proteins [35, 39-41]. In the mid 1960s, there was a shift in the
experimental interest to include investigations of surface processes occurring at
the electrode. UV/vis radiation was not useful for probing the details of such
surface phenomena; therefore, some focus turned to coupling infrared (IR)
spectroscopy with electrochemistry.

It was this shift in experimental interest, accompanied by the commercial
development and availability of Fourier transform infrared (FTIR)
spectrophotometers that spurred the development of IR spectroelectrochemistry.
The first reports on the use of internal reflection IR spectroscopy to monitor the
spectrum of a species at the electrode surface, during electrolysis, were
published in 1966 [42, 43]. At the outset, IR techniques were primarily focused
on electrode surface characterization [42-47]. There were few reports of the
application of IR to solution studies, due primarily to the large interference of
solvent absorption with the absorption by the analyte(s) of interest. However, in
that it provided a high level of detail regarding the molecular structure of reaction
components, IR offered an advantage over other spectroscopies that had been

coupled to electrochemistry (e.g., UV/vis, electron spin resonance, and

Mossbauer) for solution studies. Throughout the next two decades, the
availability of higher sensitivity instrumentation and novel spectroelectrochemical
cell designs resulted in a steady increase in the number of reports of solution IR
spectroelectrochemical studies in the fields of organic and inorganic chemistry
[48], and resulted in a number of general review articles on the subject [6, 14, 49-
53].

Today, the focus of spectroelectrochemical studies has expanded to
include the development of new types of optically transparent electrodes (OTES)
[54, 55], the incorporation of OTEs as sensors [56, 57], and the investigations of
intricate biological systems [6, 14, 58-63]. In particular, recent advances in
solution IR Spectroelectrochemistry include applications that are capable of
elucidating single amino acid involvement in complex enzyme mechanisms [15,
64]. Examples of such experiments on the metalloprotein, cytochrome c oxidase
(cco), are described in this dissertation. The catalytic mechanism of cco involves
an efficient reduction of molecular oxygen to water, concomitant with the uptake
of eight protons, of which four are used in the chemistry and four are translocated
across the membrane. The way in which cco links the complex oxygen
chemistry at different stages to the proton pumping function, in order to drive
protons about 50 A across the mitochondrial membrane against an
electrochemical gradient, is not understood. Insights into amino acid involvement
in this process can be gained via IR spectroelectrochemistry.

Spectroelectrochemistry combines two very different techniques (i.e.,

spectroscopy and electrochemistry); and because of this, the method provides

more information about the redox mechanism of electroactive analytes than can
be acquired using traditional electrochemical methods alone. Oxidation states
are changed electrochemically, by adding or removing electrons at an OTE. The
spectroscopic response of the species in solution is measured simultaneously
with the electrolysis. Spectroelectrochemical methods of analysis have been
used for over three decades to investigate heterogeneous electron transfer
processes. Spectroelectrochemistry is useful for obtaining spectra and redox
potentials for electrogenerated species in the investigation of electrode reaction
mechanisms, as well as in electroanalytical applications.

In particular, we are interested in using Spectroelectrochemistry to
elucidate structural information about redox-active protein active sites, and,
specifically, that of cytochrome c oxidase. Our goals are to be able to routinely
measure high-quality electrochemical difference FTIR spectra of cytochrome c
oxidase, and then to establish a protocol for performing such measurements in
the low-frequency infrared region of the electromagnetic spectrum. Although
mid-IR (2500 - 1000 cm°1) measurements have contributed substantially to our
understanding of amino acid involvement in catalysis, accessing the low-
frequency region (< 1000 cm'1) is also crucial, as it is this region that reports
directly on metal-ligand modes of protein active sites. In order to directly monitor
substrate binding and intermediate structure via metal center coordination, low-
frequency FTIR difference measurements of proteins are necessary. However,
many technical difficulties, including low Signal-to-background ratios and limited

options of suitable infrared windows, have plagued such measurements, These

problems and corresponding solutions are discussed in detail, in a later part of
this work. We set out to accomplish our goal in two phases: First, we
implemented a conventional transmission method for acquiring mid-IR spectra of
redox-active proteins, developed by Mantele and coworkers [65, 66]. This
approach uses a Au minigrid OTE in a spectroelectrochemical cell that was
specially designed to meet the technical challenges inherent in such
measurements. Preliminary results were obtained for a simple electron transfer
protein, cytochrome 0, upon which we quickly adapted our setup to handle the
more complex enzyme cytochrome c oxidase. The second phase involved
efforts to access the low-frequency IR region by modifying this experimental set-
up to improve the signal-to-noise ratio of such measurements. These changes
included a highly sensitive low-frequency detection scheme (Si bolometer), a
stable background signal achieved via precise temperature control, and
maximized optical throughput using diamond optical windows. However, despite
these efforts, low-frequency measurements were unsuccessful for the modified
set-up. Additionally, as part of such efforts, we have begun developing a new
type of OTE, boron-doped diamond (BDD), for spectroelectrochemical
measurements of proteins, again using cytochrome c as a test system.

Part of the development of this novel OTE material included
electrochemical characterization in nonaqueous solvents, in preparation for
studies of bioinorganic compounds that mimic the structure of protein active sites
and, therefore, serve as simple model systems. The model compound studies

are not discussed in this dissertation. The electrochemical characterization of

cytochrome c was also carried out at this relatively new electrode material. This
was important for understanding the nature of the interaction and electron
transfer kinetics of the protein with the BDD surface. Besides basic
electrochemical characterizations, we aimed to determine the deposition
conditions that lead to an optimized balance between electrical conductivity and
optical transparency in this material. This dissertation reports on the application
of BDD OTES for spectroelectrochemical studies of two test analytes,
ferrocyanide and ferrocene, in the UV/vis and IR regions of the electromagnetic
spectrum. Spectroelectrochemical investigations were also made of
cytochrome c, which enabled us to make a direct comparison of the results
obtained for BDD OTES to those obtained with a traditional Au grid OTE in the
UV/vis and IR regions of the spectrum.

A variety of optical methods have been coupled with electrochemistry,
including absorption, as is shown in Figure 1-1, as well as light scattering
spectroscopies. Transmission UV/vis spectroscopy is perhaps the most
commonly employed method, owing to the ease of cell design and measurement.
In the transmission mode (Figure 1-1A, B), the optical beam is directed normal to
the surface and passes directly through an OTE and the adjacent solution.
Infrared spectroscopy, in both the transmission and reflectance modes, has also
enjoyed widespread use in spectroelectrochemistry [50, 51, 53]. The specular
reflectance mode (Figure 1-1C) involves passing IR radiation through the solution
onto a reflective surface (electrode), at which point the light is reflected back

through the solution to the detector. In attenuated total reflectance (ATR)

infrared spectroscopy, the probe beam of light is introduced into a highly
refractive material, at an angle greater than the critical angle so that the beam is
totally internally reflected (Figure 1-1D). Light attenuation occurs through
interaction between the evanescent wave of the totally reflected light and the

light-absorbing solution species adjacent to the electrode surface.

Figure 1-1 Spectroelectrochemical Techniques Using Absorption
Spectroscopies

Methods employing UV, visible, and IR Spectroscopies can be divided into two classes,
those which measure transmission (A, B), and those measuring reflectance (C, D).
(A) In the thin-layer orientation, the path/ength is less than or equal to 0.2 mm.
(B) Conventional cell orientation, where the solution thickness (~ 1 cm), is greater than
the diffusion-layer thickness. (C) Specular reflectance at a non-transparent electrode.
(D) lntemal reflectance (commonly called A TR) method illustrating the evanescent wave.

Figure adapted from [49].
l——— UV, VISIBLE, IR

 

 

 

 

       

 

 

 

fié’l'RANSMlSSIONii HIREFLEc'rANca-iw
0.2 mm
OTE\ HK
electrode
A. Thin-layer cell C. External reflectance
1 cm

OTE '—"'I
\[Y ,,

B. Conventional cell D. Internal reflectance

 

 

 

 

 

 

 

An assortment of spectroelectrochemical cell types are reported in the

literature [25, 28, 34, 37, 42, 54, 66-84]. Although it is possible, and sometimes

necessary, to use a cell geometry analogous to a conventional electrochemical
cell (Figure 1-1B), in that the solution thickness is much greater than the
diffusion-layer thickness normal to the electrode, an optically transparent thin-
layer electrochemical (OTTLE) cell design (Figure 1-1A) is most commonly
employed. One of the main advantages of this cell type is that the electroactive
species in the thin layer can be completely electrolyzed in a very short time
(typically 20 - 120 S). Additionally, diffusion effects are eliminated, simplifying the
mathematical descriptions of the electrochemical parameters. Often, in the
absorption spectroscopy mode, experimental requirements impose functional
limits on the pathlength or thin-layer distance. An example of such a situation will
be presented later in this dissertation, in which the difficulties of low-frequency

infrared measurements of proteins will be described.

1.3 Optically Transparent Electrodes

1.3.1 Conventional Types

Several types of optically transparent electrodes (OTES) are available for
use in spectroelectrochemical measurements. The conventional types include
thin-films of a conductive material on transparent substrates, and minigrid
electrodes. Thin-films of platinum, gold, indium-doped tin oxide (ITO), or carbon
have been deposited on glass (visible), quartz (UV-visible), or germanium and
silicon (infrared), depending on the region of interest. The transparency
(20 - 85%) of the thin-film OTE depends on several properties, including the film

thickness. The second type of conventional OTE, the minigrid, consists of a

10

metal (Ni, Pt, Au, Ag or Cu) micromesh of 100 - 2000 wires per inch. In this
case, the transparency (20 - 80%) is due to the openings in the mesh. Minigrid
OTES have been used primarily in transmission experiments employing thin-layer
cell designs [32, 66, 78-80, 84-87]. An example of such a design for UV/vis and
IR transmission experiments of redox-active proteins is one of the OTTLE cells

that will be presented later, in Chapter 2.

1.3.2 A New Material: Boron-Doped Diamond

Much of the work detailed herein involves the development and
characterization of a new material for use as an OTE—boron-doped diamond.
Although the application of conductive diamond films as electrode materials is a
relatively new field, with its beginnings in the early 1990s [88-92], a sound
knowledge base has been developed in recent years regarding structure-function
relationships between redox analytes and the electrode surface [93-97]. From
electrochemical studies in aqueous media [94, 98-100], it is clear that high-
quality (i.e., hydrogen-terminated, with minimal nondiamond (spZ-bonded) carbon
impurity content, and low secondary nucleation density) BDD thin-film electrodes
exhibit a number of advantageous electrochemical properties, including: (i) low
and stable background currents and capacitance; (ii) a wide working potential
window (3 - 4 V); (iii) rapid electron transfer kinetics for many redox systems
without the requirement of conventional pretreatment procedures; (iv) long-term
response stability; (v) morphological and microstructural stability during anodic

and cathodic polarization (~10 A/cmz) in acidic and alkaline media; and (vi) high

11

resistance to fouling, due to the generally weak molecular adsorption of polar
molecules. Experimental focus is currently centered on two types of electrically
conducting, chemically vapor deposited (CVD) materials: microcrystalline and
nanocrystalline diamond thin-films. The details of BDD film growth are discussed
in Chapter 2. Generally, microcrystalline films are grown using a CH4/H2 (~0.5%)
source gas mixture. Under these growth conditions, the rate of crystal growth
generally exceeds the rate of nucleation, producing large (1 -20 pm), well-
faceted diamond grains. The nanocrystalline films are grown using a CH4/Ar
(~1°/o) source gas mixture with little or no H2 added [101]. Under these growth
conditions, the rate of nucleation generally exceeds the rate of crystal growth,
resulting in the formation of small grains (~15 nm). Such films are very smooth,
with a 30-50 nm surface roughness, and have a higher fraction of grain

boundaries than do the microcrystalline films, due to the smaller grain size [95,
102]. The contact angle measured for water (740°) indicates a hydrophobic

surface, and x-ray photoelectron spectrosc0pic data indicates surface oxygen
levels of less than 2 atomic percent. lmparting boron-dopant atoms during film
deposition renders the normally insulating diamond material electrically
conductive. Film conductivity is, therefore, directly dependent upon the level of

boron doping (i.e., electron carrier concentration), with the most conductive films

doped at levels greater than 1020 B/cm3, yielding resistivities less than

0.05 0 cm. As mentioned above, most of the research with diamond electrodes
has involved the use of aqueous redox analytes, with very few studies performed

using nonaqueous redox analytes [97, 103-106]. There are even fewer reports of

12

the application of BDD electrodes toward biological electron transfer reactions
[107, 108]. Both of these applications are described in this work.

Much of the literature on the use of diamond in electrochemistry is in the
field of electroanalysis, as, during the past few years, numerous reports have
been published on the application of this material in chemical measurements [54,
99, 100, 105-157]. The success of diamond in such studies is attributed to the
significant advantages it offers over other carbon and metal electrodes in terms
of stability, limit of detection, robustness, and dynamic range [99, 100, 105, 106,
153, 154, 156, 157]. Another unique property of BDD, which has yet to be
exploited, is the optical throughput. As an optical material, diamond is
mechanically strong, resistant to chemical attack, and, most importantly, it has

the widest optical window of any material, extending from the band gap
absorption edge in the visible (~ 225 nm) deep into the far infrared (~ 30 cm'1).

Although imperfections and impurities in the diamond lattice, such as boron,

decrease the Optical throughput, BDD retains a wide optical window transmitting
(> 50%) in the visible (300 - 900 nm) and infrared (6000 - 2600 and < 1500 cm'1)

regions. Factors that influence the optical throughput of diamond include the
defect density, chemical composition, dOping level, film thickness, and grain size
[158-176]. Thus, it is possible to manipulate and optimize the Optical properties
of this material through adjustments in the CVD growth conditions. Knowledge of
these growth parameters is essential for establishing an optimized balance

between the electrochemical and optical properties of BDD OTES.

13

A wide optical window, coupled with attractive electrochemical properties
makes BDD a viable OTE material. The application of electrically conducting
diamond for this purpose is in its infancy [54, 104, 106, 177, 178]. Boron-doped
diamond offers several advantages over other traditionally used OTES, such as
indium-doped tin oxide (ITO) including: (i) transparency extending from the UV
into the far IR, (ii) remarkable stability in aqueous (acidic and alkaline) and
nonaqueous (e.g., chlorinated) media, (iii) a wide working potential window in
aqueous and nonaqueous media, (iv) a reasonably well-defined and stable
surface chemistry, and (v) the ability to withstand cathodic polarization.
Additionally, aside from thin-films coated onto a transparent substrate, BDD
exists in a freestanding form. The ability to manufacture BDD OTES based on
deposited thin-films, as well as in freestanding forms, offers the option for
application specificity. Our group's research with the optically transparent BDD
electrodes is in the early stages; therefore, we have only begun to test the extent
of applications for this material. For instance, our preliminary
spectroelectrochemical studies of cytochrome 0 show great promise for future

applications of BDD OTES in the field of metalloprotein spectroelectrochemistry.

1.4 Scientific Relevance

Spectroelectrochemical studies of metalloproteins have attracted
widespread interest and attention. Such studies yield important information
about not only the intrinsic thermodynamic and kinetic properties of the redox-
active biomolecules, but also the structural properties, which are important for

understanding the molecular mechanisms of enzyme function. Beyond the basic

14

research goals to elucidate mechanistic insights into metalloprotein catalysis,
applications of a more applied nature do exist. These include the development of
spectroelectrochemical biosensors based on redox-active metalloproteins.
Currently, such methods involve sol-gel encapsulation of metalloproteins with
optical detection schemes [179-185]. An alternate approach, in which both
optical and electrochemical sensing are possible, involves the immobilization of

protein molecules on an OTE. Preliminary studies of this type have been

conducted at nanoporous titanium dioxide (T102) for cytochrome c and

hemoglobin [186].

Hemoglobin is of particular interest for optical biosensing methods due to
its capability for high-affinity binding of Oz, CO, and NO, which drastically change

its absorption spectrum when bound to the heme iron. However, hemoglobin
only binds these molecules in the Fe(|l) oxidized state, which, incidentally, is not
its resting state (Fe(lll)). Sol-gel methods, therefore, require repeated additions
of chemical oxidants and reductants to cycle this enzyme between the Fe(ll) and
Fe(lll) states [180, 185]. The ability to perform electrochemistry directly at a
transparent electrode would alleviate this problem, which complicates
technological development of such sensors. Topoglidis and coworkers have
demonstrated the viability of such an approach for sensing CO in aqueous

solution by optical detection at a hemoglobin/I102 film [186]. BDD has Shown

great promise as an optically transparent electrode and in its application for direct

electrochemistry of cytochrome 0. Therefore, it is conceivable to expect that this

15

material may be exploited in the future developments of spectroelectrochemical

sensing using biomolecules.

1.5 Dissertation Outline

The following chapter, Chapter 2, of this dissertation is referred to
throughout as the experimental section, as it details all experimental parameters
and procedures used for the measurements discussed in later chapters.
Chapters 3 and 4 demonstrate the usefulness of BDD as an electrode material
for investigating electron transfer kinetics of two very different systems.
Chapter 3 reports on the first system, ferrocene, which is routinely used as a
model system in nonaqueous kinetic studies because it demonstrates “ideal”
electrochemical behavior in a variety of solvents. Cyclic voltammetry was used
to study the ferrocene/ferricenium redox reaction at microcrystalline BDD films as
a function of the scan rate and the electrolyte/solvent composition. The response
at BDD was compared to a well-characterized carbon electrode, glassy carbon.
This study demonstrated that the electrode microstructure does not significantly
affect the rate of ferrocene oxidation at these two carbon electrodes, and that
BDD is a viable electrode for performing electrochemistry in nonaqueous
solvents. In Chapter 4, a more challenging system is investigated by cyclic
voltammetry at BDD electrodes. This system, cytochrome c, is a well-studied,
electron transfer protein. Until recently, it was believed that, in order to observe

facile electron transfer of cytochrome c at a bare electrode, the surface must be
terminated with negatively charged, hydrophilic, oxygen groups. The effects of

protein concentration, boron-doping level, and microstructure were investigated.

16

This study demonstrated that quasi-reversible, stable electron transfer could be
performed at hydrophobic, hydrogen-terminated, BDD electrodes, thus raising
important questions of the conventional wisdom regarding the necessary surface
structure for cytochrome c electron transfer.

In Chapters 5, 6, and 7, the focus turns to spectroelectrochemistry. The
technical difficulties associated with performing low-frequency FTIR spectroscopy
of biological samples are discussed in Chapter 5, and the concept of reaction-
induced infrared difference spectroscopy is described in detail. Chapter 6
discusses applications of electrochemical difference infrared spectroscopy to the
metalloprotein cytochrome c oxidase. This technique has been extended to its
technical limits, in an effort to elucidate structural and mechanistic aspects of
oxygen reduction to water by cco. Isotope-labeling experiments of the protein
samples were performed in an effort to simplify mode assignments. Additionally,
cco from two different organisms, a bacterial and a mammalian source, were
compared. Although the oxidase structure and mechanistic function is highly
conserved between these two organisms, there were observable deviations in
the IR spectra. These experiments identified a potential histidine amino acid side
chain that is linked to the electron transfer reaction of cco and may be important

in the proton pumping mechanism. Questions remain regarding whether this
histidine is a ligand to the Cue of the active site. These studies are

representative of some of the most challenging FTIR experiments currently being

conducted, as they are attempting to identify single amino acid residues

(AA ~ 10"4 AD.) in complex protein FTIR spectra.

17

The last chapter, Chapter 7, describes BDD OTE spectroelectrochemical
studies in the UV/vis and IR regions of the spectrum. The examination of the
optical properties of BDD is in its infancy. Of the several types of BDD OTES
available, freestanding disks were used for studies in the UV/vis region and thin-
films of BDD on undoped silicon substrates for the IR. Novel
spectroelectrochemical cells were designed to utilize these two types of OTES.
The cell designs were tested with ferrocyanide, ferrocene, and cytochrome c, and
key insights into the relationship between the diamond growth conditions and
OTE optical properties emerged. For the first time, the spectroelectrochemical
response for cytochrome c was measured at BDD OTES. The observed UV/vis
and IR spectra were compared to those obtained using a conventional OTE, a Au
minigrid. This study indicated the viability of using BDD thin-films as OTES but
also demonstrated that further work is required to optimize the growth conditions
to yield a BDD OTE that possessed both the conductivity and transparency to
compete with 3 Au minigrid OTE. Subsequent studies may determine that a
different type of BDD OTE, one with better conductivity and optical transparency,
is better suited for this purpose. However, these studies of the protein system
are among the most advanced experiments currently being conducted with BDD
OTES. The ability to combine direct electrochemistry of biomolecules with UV/vis
and IR spectroscopies on one electrode may have further implications in the
applied fields of bioanalytical devices and enzyme-based spectroelectrochemical
sensors. Such applications could benefit from the resistance to fouling and

extreme potential ranges that are possible with BDD OTES.

18

1.6

[1]

[2-1

[3-1

[4-1

[51

[6.]

[7-1

[8-1

[91

[10.]

[11.]
[12.]

[13.]

References

L. Esposito, L. Vitagliano and L. Mazzarella Protein Pept. Left. 2002, 9,
95-105.

C. M. Wilmot and A. R. Pearson Curr. Opin. Chem. Biol. 2002, 6, 202-207.

A. Zouni, H. T. Witt, J. Kern, P. Fromme, N. Krauss, W. Saenger and P.
Orth Nature 2001, 409, 739-743.

S. Yoshikawa, K. Shinzawa-Itoh and T. Tsukihara J. Bioenerg. Biomembr.
1998, 30, 7-14.

T. Tsukihara, H. Aoyama, E. Yamashita, T. Tomizaki, H. Yamaguchi, K.
Shinzawaltoh, R. Nakashima, R. Yaono and S. Yoshikawa Science 1996,
272, 1136-1144.

C. Zscherp and A. Barth Biochemistry 2001, 40, 1875—1883.

J. Heberle and T. Gensch Nat. Struct. Biol. 2001, 8, 195-197.

A. Barth Spectrochim. Acta A 2000, 56A, 1223-1232.

J. K. Kauppinen, D. J. Moffatt, H. H. Mantsch and D. G. Cameron Appl.
Spectrosc. 1981, 35, 271-276.

H. H. Mantsch, D. J. Moffatt and H. L. Casal J. Mol. Struct. 1988, 173,
285-298.

P. R. Rich and J. Breton Biochemistry 2001, 40, 6441-6449.

A. Barth and C. Zscherp FEBS Lett. 2000, 477, 151-156.

H. A. Chu, W. Hillier, N. A. Law, H. Sackett, S. Haymond and G. T.
Babcock Biochim. Biophys. Acta 2000, 1459, 528-532.

19

[14.]

[15.]

[16.]

[17.]

[18.]

[19.]

[20.]

[21 .1
[22.]

[23.]

[24.]
[25.]

[26.]

W. Mantele In Biophysical Techniques in Photosynthesis; J. Amesz and A.
J. Hoff, Eds.; Kluwer Academic Publishers, Dordrecht, 1996; Vol. 3, 137-
160.

P. Hellwig, U. Pfitzner, J. Behr, B. Rost, R. P. Pesavento, W. von Donk, R.
B. Gennis, H. Michel, B. Ludwig and W. Mantele Biochemistry 2002, 41,
9116-9125.

W. Mantele, A. Wollenweber, F. Rashwan, J. Heinze, E. Nabedryk, G.
Berger and J. Breton Photochem. Photobiol. 1988, 47, 451-455.

J. Zhang, W. Oettmeier, R. B. Gennis and P. Hellwig Biochemistry 2002,
41, 4612-4617.

B. Barquera, P. Hellwig, W. Zhou, J. E. Morgan, C. C. Haese, K. K.
Gosink, M. Nilges, P. J. Bruesehoff, A. Roth, C. R. D. Lancaster and R. B.
Gennis Biochemistry 2002, 41 , 3781-3789.

P. Hellwig, B. Rost and W. Mantele Spectrochim. Acta A 2001, 57A, 1123-
1 131.

T. Kuwana, R. K. Darlington and D. W. Leedy Anal. Chem. 1964, 36,
2023-2025.

T. Kuwana and W. R. Heineman Acct. Chem. Res. 1976, 9, 241-248.

J. Zak, M. D. Porter and T. Kuwana Anal. Chem. 1983, 55, 2219-2222.

T. Kuwana and N. Winograd In Electroanalytical Chemistry: A Series of
Advances; A. J. Bard, Ed.; Marcel Dekker, Inc., New York, 1974; Vol. 7, 1-
78.

T. Osa and T. Kuwana J. Electroanal. Chem. 1969, 22, 389-406.

J. W. Strojek and T. Kuwana J. Electroanal. Chem. 1968, 16, 471—483.

W. R. Heineman, T. P. DeAngeIis and J. F. Goelz Anal. Chem. 1975, 47,
1364-1369.

20

[27.]

[28.]

[29.]

[30.]

[31.]

[32.]

[33.]

I341

[35.]

[36.]

[37.]

[38.]

[39.]

[40.]

N. R. Armstrong, R. K. Quinn and N. E. Vanderborgh J. Electrochem. Soc.
1976, 123, 646-649.

F. M. Hawkridge, J. E. Pemberton and H. N. Blount Anal. Chem. 1977, 49,
1646-1647.

D. E. Albertson, H. N. Blount and F. M. Hawkridge Anal. Chem. 1979, 51,
556-560.

J. L. Anderson Anal. Chem. 1979, 51, 2312-2315.

E. A. Blubaugh, A. M. Yacynych and W. R. Heineman Anal. Chem. 1979,
51, 561-565.

R. K. Rhodes and K. M. Kadish Anal. Chem. 1981, 53, 1539-1541.

E. E. Bancroft, H. N. Blount and F. M. Hawkridge Adv. Chem. Ser. 1982,
23-49.

E. F. Bowden, D. J. Cohen and F. M. Hawkridge Anal. Chem. 1982, 54,
1005-1008.

K. A. Rubinson and H. B. Mark, Jr. Anal. Chem. 1982, 54, 1204-1206.

H. O. Finklea, R. K. Boggess, J. W. Trogdon and F. A. Schultz Anal.
Chem. 1983, 55, 1177-1179.

X. Q. Lin and K. M. Kadish Anal. Chem. 1985, 57, 1498-1501.

D. A. Scherson, S. Sarangapani and F. L. Urbach Anal. Chem. 1985, 57,
1501-1503.

L. N. Mackey and T. Kuwana Bioe/ectrochem. Bioenerg. 1976, 3, 596-613.

F. R. Shu and G. S. Wilson Anal. Chem. 1976, 48, 1676-1679.

21

[41.]

[42.]

[43.]

[44.]

[45.]

[46.]

[47.]

[48.]

[49.]

[50.]

[51.]

[52.]

[53.]

[54.]

[55.]

C. W. Anderson, H. B. Halsall and W. R. Heineman Anal. Biochem. 1979,
93, 366-372.

H. B. Mark, Jr. and B. S. Pons Anal. Chem. 1966, 38, 119-121.

W. N. Hansen, T. Kuwana and R. A. Osteryoung Anal. Chem. 1966, 38,
1810-1821.

W. N. Hansen, R. A. Osteryoung and T. Kuwana J. Am. Chem. Soc. 1966,
88, 1062-1063.

J. D. E. McIntyre and D. M. Kolb Symp. Faraday Soc. 1970, NO. 4, 99-
113.

W. J. Plieth Symp. Faraday Soc. 1970, No. 4, 137-144.

H. B. Mark, Jr. and E. N. Randall Symp. Faraday Soc. 1970, No. 4, 157-
172.

H. E. Toma and K. Araki Curr. Org. Chem. 2002, 6, 21-34.

W. R. Heineman Anal. Chem. 1978, 50, 390A-402A.

J. K. Foley and B. S. Pons Anal. Chem. 1985, 57, 945A-956A.

K. Ashley Talanta 1991, 38, 1209-1218.

F. Siebert In Biochemical Spectroscopy, 1995; Vol. 246, 501-526.

J. J. Niu and S. J. Dong Rev. Anal. Chem. 1996, 15, 1-171.

J. K. Zak, J. E. Butler and G. M. Swain Anal. Chem. 2001, 73, 908-914.

A. Neudeck and L. Kress J. Electroanal. Chem. 1997, 437, 141-156.

22

[56.]

[57.]

[58.]

[59.]

[60.]

[61.]

[62.]

[63.]

[64.]

[65.]

[66.]

J. N. Richardson, A. L. Dyer, M. L. Stegemiller, l. Zudans, C. J. Seliskar
and W. R. Heineman Abstracts of Papers, 223rd ACS National Meeting,
Orlando, FL, United States, April 7-11, 2002.

M. Maizels, M. Stegemiller, S. Ross, A. Slaterbeck, Y. Shi, T. H. Ridgway,
W. R. Heineman, C. J. Seliskar and S. A. Bryan In ACS Symposium
Series 2001, Vol. 778, 364-378.

J. D. Burgess and F. M. Hawkridge In Electroanalytical Methods for
Biological Materials; A. Brajter—Toth and J. O. Chambers, Eds; Marcel
Dekker, New York, 2002, 109-142.

C. H. Taboy, C. Bonaventura and A. L. Crumbliss In Redox Cell Biology
and Genetics, Pt B, 2002; Vol. 353, 187-209.

N. B. Ugulava, B. R. Gibney and J. T. Jarrett Biochemistry 2001, 40, 8343-
8351.

P. Yao, Y. H. Wang, Y. Xie and Z. X. Huang J. Electroanal. Chem. 1998,
445, 197-201.

D. J. Cohen, B. C. King and F. M. Hawkridge J. Electroanal. Chem. 1998,
447, 53-62.

S. J. Dong, J. J. Niu and T. M. Cotton In Biochemical Spectroscopy, 1995;
Vol. 246, 701-732.

P. Hellwig, J. Behr, C. Ostermeier, O.-M. H. Richter, U. Pfitzner, A.
Odenwald, B. Ludwig, H. Michel and W. Mantele Biochemistry 1998, 37,
7390-7399.

F. Baymann, D. A. Moss and W. Ma‘nteIe Anal. Biochem. 1991, 199, 269-
274.

D. Moss, E. Nabedryk, J. Breton and W. Mantele Eur. J. Biochem. 1990,
187, 565-572.

23

[67.]

[68.]

[69.]

[70.]

[71.]

[72.]

[73.]

[74.]

[75.]

[76.]

[77.]

[78.]

[79.]

[80.]

[81.]

P. A. Flowers, M. A. Maynor and D. E. Owens Anal. Chem. 2002, 74, 720-
723.

l. S. Zavarine and C. P. Kubiak J. Electroanal. Chem. 2001, 495, 106-109.

D. Daniel and I. G. R. Gutz Electroanal. 2001, 13, 681-685.

J. 8. Yu, C. Yang and H. O. Fang Anal. Chim. Acta 2000, 420, 45-55.

Y. J. Ma, J. M. Zheng and S. M. Zhu Chin. J. Inorg. Chem. 1999, 15, 61-
67.

D. Kardash, J. M. Huang and C. Korzeniewski J. Electroanal. Chem. 1999,
476, 95-100.

M. E. Rosa-Montanez, H. De Jesus-Cardona and C. R. Cabrera-Martinez
Anal. Chem. 1998, 70, 1007-1011.

P. A. Flowers and S.-A. Callender Anal. Chem. 1996, 68, 199-202.

J. J. Niu and S. J. Dong Electrochim. Acta 1995, 40, 823-828.

J. J. Niu, G. J. Cheng and S. J. Dong Chin. Chem. Lett. 1994, 5, 317-320.

F. Hartl, H. Luyten, H. A. Nieuwenhuis and G. C. SchoemakerAppl.
Spectrosc. 1994, 48, 1522-1528.

F. Yang, X. Lin and E. Wang Fenxi Huaxue 1993, 21, 857-860.

M. Krejcik, M. Danek and F. Hartl J. Electroanal. Chem. 1991, 317, 179-
187.

X. Q. Lin, Y. Liu and E. K. Wang Acta Chim. Sin. 1990, 48, 1080-1084.

P. B. Graham In Univ. Massachusetts,Amherst, MA, USA, 1990, p 299 pp.

24

[82.]

[83.]

[84.]

[85.]

[86.]

[87.]

[88.]

[89.]

[90.]

[91.]

[92.]

[93.]

[94.]

[95.]

X. Lin, D. Liu and E. Wang WuIiHuaxue Xuebao 1989, 5, 719-722.

K. M. Kadish, X. H. Mu and X. Q. Lin Electroanal. 1989, 1, 35-41.

R. W. Murray, W. R. Heineman and G. W. O'Dom Anal. Chem. 1967, 39,
1666-1668.

K. Jiao, L. Q. Ren and T. Yang Chin. J. Anal. Chem. 2001, 29, 1299-1302.

J. P. Bullock, D. C. Boyd and K. R. Mann Inorg. Chem. 1987, 26, 3084-
3086.

W. R. Heineman, J. N. Burnett and R. W. Murray Anal. Chem. 1968, 40,
1974-1978.

M. lwaki, S. Sato, K. Takahashi and H. Sakairi Nucl. Instrum. Methods
Phys. Res. 1983, 209-210, 1129-1133.

Y. V. Pleskov, A. Y. Sakharova, M. D. Krotova, L. L. Builov and B. V.
Spitsyn J. Electroanal. Chem. 1987, 228, 19-27.

K. Patel, K. Hashimoto and A. Fujishima J. Photochem. Photobiol., A
1992, 65, 419-429.

K. Patel, K. Hashimoto and A. Fujishima DenkiKagaku1992, 60, 659-661.

G. M. Swain and R. Ramesham Anal. Chem. 1993, 65, 345-351.

M. C. Granger and G. M. Swain J. Electrochem. Soc. 1999, 146, 4551-
4558.

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. 2000, 72, 3793-3804.

O. Y. Chen, D. M. Gruen, A. R. Krauss, T. D. Corrigan, M. Witek and G.
M. Swain J. Electrochem. Soc. 2001, 148, E44-E51.

25

[96.] Q. Chen and G. M. Swain Langmuir1998, 14, 7017-7026.

[97.] S. Alehashem, F. Chambers, J. W. Strojek, G. M. Swain and R.
Ramesham Anal. Chem. 1995, 67, 2812-2821.

[98.] M. C. Granger, J. Xu, J. W. Strojek and G. M. Swain Anal. Chim. Acta
1999, 397, 145-161.

[99.] J. S. Xu, M. C. Granger, Q. Y. Chen, J. W. Strojek, T. E. Lister and G. M.
Swain Anal. Chem. 1997, 69, A591-A597.

[100.] G. M. Swain, A. B. Anderson and J. C. Angus MRS Bulletin 1998, 23, 56-
60.

[101.] D. M. Gruen Annu. Rev. Mater. Sci. 1999, 29, 211-259.

[102.] B. Fausett, M. C. Granger, M. L. Hupert, J. Wang, G. M. Swain and D. M.
Gruen Electroanal. 2000, 12, 7-15.

[103.] S. Haymond, G. T. Babcock and G. M. Swain Electroanal. 2002, in press.

[104.] S. Haymond, J. K. Zak, J. E. Butler, G. T. Babcock and G. M. Swain in
preparation 2002.

[105.] G. Pastor-Moreno and D. J. Riley Electrochim. Acta 2002, 47, 2589-2595.

[106.] 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 Materials 2002, 8, 171-188.

[107.] S. Haymond, G. T. Babcock and G. M. Swain J. Am. Chem. Soc. 2002,
124, 10634-10635.

[108.] F. Marken, C. A. Paddon and D. Asogan Electrochem. Commun. 2002, 4,
62-66.

26

[109.] N. Wangfuengkanagul and O. Chailapakul J. Phann. Biomed. Anal. 2002,
28, 841-847.

[110.] K. Van Hege, M. Verhaege and W. Verstraete Electrochem. Commun.
2002, 4, 296-300.

[111.] K. Tsunozaki, Y. Einaga, T. N. Rao and A. Fujishima Chem. Lett. 2002,
502-503.

[112.] l. Troster, M. Fryda, D. Herrmann, L. Schafer, W. Hanni, A. Perret, M.
Blaschke, A. Kraft and M. Stadelmann Diam. Relat. Mat. 2002, 11, 640-
645.

[113.] C. Terashima, T. N. Rao, B. V. Sarada, D. A. Tryk and A. Fujishima Anal.
Chem. 2002, 74, 895-902.

[114.] N. Spataru, B. V. Sarada, D. A. Tryk and A. Fujishima Electroanal. 2002,
14, 721-728.

[115.] D. Sopchak, B. Miller, R. Kalish, Y. Avyigal and X. Shi Electroanal. 2002,
14, 473-478.

[116.] T. N. Rao, B. H. Loo, B. V. Sarada, C. Terashima and A. Fujishima Anal.
Chem. 2002, 74, 1578-1583.

[117.] C. Prado, S. J. Wilkins, F. Marken and R. G. Compton Electroanal. 2002,
14, 262-272.

[118.] C. Prado, G.-U. Flechsig, P. Gruendler, J. S. Foord, F. Marken and R. G.
Compton Analyst 2002, 127, 329-332.

[119.] K. Ohnishi, Y. Einaga, H. Notsu, C. Terashima, T. N. Rao, S.-G. Park and
A. F ujishima Electrochem. Solid-State Lett. 2002, 5, D1-D3.

[120.] F. Montilla, P. A. Michaud, E. Morallon, J. L. Vazquez and C. Comninellis
Electrochim. Acta 2002, 47, 3509-3513.

27

[121.] K. J. McKenzie and F. Marken Electrochem. Solid-State Left. 2002, 5,
E47-E50.

[122.] A. Manivannan, M. S. Seehra, D. A. Tryk and A. Fujishima Anal. Lett.
2002, 35, 355-368.

[123.] T. A. Ivandini, B. V. Sarada, C. Terashima, T. N. Rao, D. A. Tryk, H.
Ishiguro, Y. Kubota and A. Fujishima J. Electroanal. Chem. 2002, 521,
117-126.

[124.] W. Haenni, J. Gobet, A. Perret, L. Pupunat, P. Rychen, C. Comninellis
and B. Correa New Diamond and Frontier Carbon Technology 2002, 12,
83-88.

[125.] A. Fujishima, C. Terashima, K. Honda, B. V. Sarada and T. N. Rao New
Diamond and Frontier Carbon Technology 2002, 12, 73-81.

[126.] S. Ferro and A. De Battisti J. Phys. Chem. B 2002, 106, 2249-2254.

[127.] S. Ferro and A. De Battisti Physical Chemistry Chemical Physics 2002, 4,
1915-1920.

[128.] S. Ferro and A. De Battisti Electrochim. Acta 2002, 47, 1641-1649.

[129.] N. G. Ferreira, L. L. G. Silva and E. J. Corat Diam. Relat. Mat. 2002, 11,
657-661.

[130.] A. Chatterjee, R. Wiltshire, K. B. Holt, R. G. Compton, J. S. Foord and F.
Marken Diam. Relat. Mat. 2002, 11 , 646-650.

[131.] P. Canizares, F. Martinez, M. Diaz, J. Garcia-Gomez and M. A. Rodrigo J.
Electrochem. Soc. 2002, 149, 0118-0124.

[132.] B. Boye, P.-A. Michaud, B. Marselli, M. M. Dieng, E. Brillas and C.
Comninellis New Diamond and Frontier Carbon Technology 2002, 12, 63-
72.

[133.] M. A. Witek and G. M. Swain Anal. Chim. Acta 2001, 440, 119-129.

28

[134.] N. Wangfuengkanagul and O. Chailapakul Anal. Sci. 2001, 17, a349-a352.

[135.] R. Uchikado, T. N. Rao, D. A. Tryk and A. Fujishima Chem. Lett. 2001,
144-145.

[136.] Y.-C. Tsai, B. A. Coles, K. Holt, J. S. Foord, F. Marken and R. G. Compton
Electroanal. 2001, 13, 831-835.

[137.] C. Terashima, T. N. Rao, B. V. Sarada and A. Fujishima Chemical
Sensors 2001, 17, 124-126.

[138.] A. J. Saterlay, S. J. Wilkins, K. B. Holt, J. S. Foord, R. G. Compton and F.
Marken J. Electrochem. Soc. 2001, 148, E66-E72.

[139.] A. J. Saterlay, J. S. Foord and R. G. Compton Electroanal. 2001, 13,
1065-1070.

[140.] M. A. Rodrigo, P. A. Michaud, I. Duo, M. Panizza, G. Cerisola and C.
Comninellis J. Electrochem. Soc. 2001, 148, D60-DG4.

[141.] T. N. Rao, B. H. Loo, D. A. Tryk and A. Fujishima Chemical Sensors 2001,
17, 160-162.

[142.] M. Panizza, P. A. Michaud, G. Cerisola and C. Comninellis Electrochem.
Commun. 2001, 3, 336-339.

[143.] M. Panizza, P. A. Michaud, G. Cerisola and C. Comninellis J. Electroanal.
Chem. 2001, 507, 206—214.

[144.] F. Montilla, P. A. Michaud, E. Morallon, J. L. Vazquez and C. Comninellis
Port. Electrochim. Acta 2001, 19, 221-226.

[145.] F. Marken, Y.-C. Tsai, A. J. Saterlay, B. A. Coles, D. Tibbetts, K. Holt, C.
H. Goeting, J. S. Foord and R. G. Compton J. Solid State Electrochem.
2001, 5, 313-318.

[146.] J.-J. Lee, B. Miller, X. Shi, R. Kalish and K. A. Wheeler J. Electrochem.
Soc. 2001, 148, C183-C190.

29

[147.] T. A. Ivandini, B. V. Sarada, C. Terashima, T. N. Rao, D. A. Tryk and A.
Fujishima Chemical Sensors 2001, 17, 163-165.

[148.] J. Iniesta, P. A. Michaud, M. Panizza and C. Comninellis Electrochem.
Commun. 2001, 3, 346-351.

[149.] J. Iniesta, P. A. Michaud, M. Panizza, G. Cerisola, A. Aldaz and C.
Comninellis Electrochim. Acta 2001, 46, 3573-3578.

[150.] P. L. Hagans, P. M. Natishan, B. R. Stoner and W. E. O'Grady J.
Electrochem. Soc. 2001, 148, E298-E301.

[151.] L. Gherardini, P. A. Michaud, M. Panizza, C. Comninellis and N. Vatistas
J. Electrochem. Soc. 2001, 148, 078-082.

[152.] L. Gherardini, C. Comninellis and N. Vatistas Anna/i di Chimica (Rome,
Italy) 2001, 91, 161-168.

[153.] A. Deneuville and E. Gheeraert Vide-Sci. Techn. Appl. 2001, 56, 427-+.

[154.] A. Fujishima NASA Conference Publication 2001, 210948, 150-157.

[155.] O. Chailapakul, P. Aksharanandana, T. Frelink, Y. Einaga and A.
Fujishima Sensors and Actuators, B: Chemical 2001, B80, 193-201.

[156.] J. C. Angus, H. B. Martin, U. Landau, Y. E. Evstefeeva, B. Miller and N.
Vinokur New Diamond and Frontier Carbon Technology 1999, 9, 175-187.

[157.] A. Fujishima, T. N. Rao and D. A. Tryk Proc. - Electrochem. Soc. 2000,
99-32, 383-388.

[158.] A. M. Zaitsev Optical Properties of Diamond: A Data Handbook; Springer-
Verlag, Berlin, 2001.

[159.] E. Bustarret, F. Pruvost, M. Bernard, C. Cytermann and C. Uzan—Saguy
Phys. Status Solidi A 2001, 186, 303-307.

30

[160.] J. Wang, G. M. Swain, M. Mermoux, G. Lucazeau, J. Zak and J. W.
Strojek New Diamond and Frontier Carbon Technology 1999, 9, 317-343.

[161.] E. Gheeraert, A. Deneuville and J. Mambou Diam. Relat. Mat. 1998, 7,
1509-1512.

[162.] Z. Yin, Z. Akkerman, B. X. Yang and F. W. Smith Diam. Relat. Mat. 1997,
6, 153-158.

[163.] H. Opyrchal, K. K. Chin, E. Kohn and W. Ebert Diam. Relat. Mat. 1997, 6,
940-943.

[164.] A. Ogasawara, T. Inushima, T. Shiraishi, S. Ohya, S. Karasawa and H.
Shiomi Diam. Relat. Mat. 1997, 6, 835-838.

[165.] E. Gheeraert, A. Deneuville, J. Mambou and P. Ribot Shallow-Level Cent.
Semicond, Int. Conf., 7th 1997, 159-164.

[166.] F. Zhang, X. Erqing, B. Yang, Y. Cal and G. Chen BandaotiXuebao1995,
16, 468-472.

[167.] R. Locher, J. Wagner, F. Fuchs, C. Wild, P. Hiesinger, P. Gonon and P.
Koidl Mater. Sci. Eng, B 1995, B29, 211-215.

[168.] J. I. Pankove and C.-H. Qui In Synthetic Diamond: Emerging CVD
Science and Technology; K. E. D. Spear, J.P. Dismukes, Ed.; John Wiley
& Sons, Inc., New York, 1994, 401-407.

[169.] K. M. McNamara, B. E. Williams, K. K. Gleason and B. E. Scruggs
Electrochem. Solid-State Lett. 1994, 76, 2466-2472.

[170.] F. Zhang, B. Yang, E. Xie, Y. Cai, X. Jiang and G. Chen Int. Conf. Appl.
Diamond Films Relat. Mater., 2nd 1993, 451-456.

[171.] A. Khomich, V. Polyakov, P. Perov, M. Ermakov, l. Teremetskaya, V.
Varnin and G. Sokolina Int. Conf. Appl. Diamond Films Relat. Mater., 2nd
1993, 387-392.

31

[172.] G. Zou, S. Yu, Z. Jin, X. Lu and Z. Yang Proc. SP/E-lnt. Soc. Opt. Eng.
1992, 1759, 224-228.

[173.] S. R. P. Silva, G. A. J. Amaratunga and C. P. Constantinou J. Appl. Phys.
1992, 72, 1149-1153.

[174.] K. Okano, T. Kurosu, M. lida, T. Eickhoff, H. Wilhelm and D. R. T. Zahn
Vacuum 1990, 41, 1387-1389.

[175.] K. Okano, T. Kurosu, M. lida, T. Eickhoff, H. Wilhelm and D. R. T. Zahn
Proc. SPlE-lnt. Soc. Opt. Eng. 1990, 1275, 25-35.

[176.] C. Wild, N. Herres, J. Wagner, P. Koidl and T. R. Anthony Proc. -
Electrochem. Soc. 1989, 89-12, 283-295.

[177.] H. B. Martin and P. W. Morrison Electrochem. Solid-State Lett. 2001, 4,
E17-E20.

[178.] J. M. Stotter, J. K. Zak, Z. Behler, Y. Show and G. M. Swain submitted for
publication Anal. Chem. 2002.

[179.] S. Bharathi and O. Lev Appl. Biochem. Biotechnol. 2000, 89, 209-216.

[180.] D. J. Blyth, J. W. Aylott, D. J. Richardson and D. A. Russell Analyst 1995,
120, 2725-2730.

[181.] D. J. Blyth, S. J. Poynter and D. A. Russell Analyst 1996, 121, 1975-1978.

[182.] D. J. Blyth, J. W. Aylott, J. W. B. Moir, D. J. Richardson and D. A. Russell
Analyst 1999, 124, 129-134.

[183.] K. E. Chung, E. H. Lan, M. S. Davidson, B. S. Dunn, J. S. Valentine and J.
I. Zink Anal. Chem. 1995, 67, 1505-1509.

l184-] E. H. Lan, B. Dunn, J. S. Valentine and J. I. Zink J. Sol-Gel Sci. Technol.
1996, 7, 109-116.

32

[185.] E. H. Lan, B. C. Dave, J. M. Fukuto, B. Dunn, J. I. Zink and J. S. Valentine
J. Mater. Chem. 1999, 9, 45-53.

[186.] E. Topoglidis, T. Lutz, R. L. Willis, C. J. Barnett, A. E. G. Cass and J. R.
Durrant Faraday Discuss. 2000, 35-46.

33

CHAPTER 2

2 Materials and Methods
2.1 Boron-Doped Diamond Growth Conditions

2.1.1 Microcrystalline Boron-Doped Diamond

The polycrystalline boron-doped diamond (BDD) thin films were deposited
on Si using microwave-assisted chemical vapor deposition (CVD). The reactor
was a 1.5 kW, 2.54 GHz commercial system (ASTex, Inc., Lowell, MA) with a 1 L
quartz bell jar. The Si substrates (largely undoped, gift from Texas Instruments,
Inc. (spectroscopy measurements) or p-Si(100), < 0.001 0 cm, Virginia
Semiconductor, Inc. (electrochemistry measurements» were pretreated in the
following manner: (i) organic solvent cleaning for 10 s each in toluene,
dichloromethane, acetone, methanol, and isopropanol, (ii) a 60 s soak in
concentrated hydrofluoric acid, followed by a rinse with ultrapure water and air
drying, (iii) a 20 min ultrasonication in a diamond powder (0.1 pm diameter, GE
Superabrasives)-acetone mixture followed by a rinse with clean acetone. The
scratched substrates were then placed in the CVD reactor on top of a boron
diffusion source (GS-126, BoronPlus, Techniglas, Inc., Perrysburg, OH), which,
itself, rests on top of the molybdenum substrate stage. This diffusion source
provides the boron dopant atoms during film growth. In some depositions, the
substrates were also placed around a small block of boron nitride, which also

served as a source for boron dopant atoms. The films were deposited for 4 - 20 h

34

using the following parameters: a 0.35% CH4/H2 volumetric ratio, a total gas flow
of approximately 200 sccm, a system pressure of 45 torr, and a microwave
power of 1000 W. The substrate temperature was estimated to be 850 °C, as

measured with a disappearing filament optical pyrometer. At the end of the
growth period, the methane flow was stopped, and the films were hydrogen-
plasma treated for an additional 15 min at the deposition conditions listed above.

Following this treatment, the plasma power and pressure were slowly reduced to
400 W and 20 torr over a 30 min period to cool the sample to less than 400 °C
(estimated) in the presence of atomic hydrogen. The plasma was then

extinguished and the films were further cooled under a flow of hydrogen gas for

1 h. Recently, a gas-handling system was incorporated onto one of the CVD
reactors, allowing for well-controlled gas-phase doping with B2H5. This set-up
enabled a substantial decrease in the film growth times, and therefore, thickness,

without sacrificing film quality or B content.

Post-deposition, the diamond films were chemically treated by a two-part

treatment to remove adventitious metal and Sp2-bonded nondiamond carbon
impurities from the surface: (i) exposure to warm 3:1 HNOg/HCI (v/v) for 30 min

followed by a rinse with ultrapure water and (ii) exposure to warm 30% H202

(w/v) (Aldrich Chemical) for 30 min followed by a rinse with ultrapure water [1].
The films were dried and then placed into a microwave-assisted CVD system for
hydrogen-plasma treatment. The wet chemical treatments introduce significant

levels of oxygen, ca. 12- 15 atomic percent, so the final hydrogen-plasma

35

treatment is required to rehydrogenate the surface. The treatment conditions
were as follows, 45 min at a total flow of 200 sccm, a system pressure of 35 torr,
and a microwave power of 1000 W. The films were slowly cooled over a 15 min
period in atomic hydrogen, as described above. The methane and hydrogen
gases were of ultrahigh purity (99.999%).

Films were characterized by Optical and atomic force microscopy
(morphology), Raman spectroscopy (microstructure), cyclic voltammetry, and
4-point probe resistivity measurements (electrical conductivity). The estimated
film thickness was between 2 and 5 pm, as measured by scanning electron
microscopy, and the apparent in-plane resistivity was, generally, 0.1 0 cm or
less. This resistivity is referred to as “apparent" because the measurement was
made with the diamond attached to the electrically conducting Si substrate.
Electrochemical characterization involved measuring the cyclic voltammetric

responses for two redox couples, ruthenium hexaamine ([Ru(NH3)5] +3M) and

—3/-4

ferrocyanide ([Fe(CN)5] ). If the peak currents varied linearly with the square

root of the scan rate, and the peak splitting was near 0.070 V at 0.1 V/s, the

electrodes were considered high quality. The heterogeneous electron transfer
rate constant for [Ru(NH3)5] +3” is sensitive to the density of electronic states at
the formal potential, while the rate constant for [Fe(CN)5] "3"4 is sensitive to both

the density of states and the diamond electrode surface chemistry. Previous

studies of BDD electrodes showed that the highest rate constants for

[Fe(CN)5] "3"4 are observed at the hydrogen-terminated diamond surface [2].

The doping level was estimated to be in the mid 10’19 to low 10'20 range, based

36

on boron nuclear reaction analysis measurements performed on other films
deposited using similar conditions. Optical microscopy revealed that the
polycrystalline films completely covered the substrate, showing no cracks or
voids. Atomic force microscopy showed a well-faceted, polycrystalline
morphology with a nominal grain size of 1 — 2 pm. The crystallites were primarily
of octahedral and cubo-octahedral shape with a relatively low fraction of

secondary growths [1]. Raman spectroscopy revealed a sharp, one-phonon

diamond line centered near 1332 cm'1. The line width was 6 - 8 cm’1, which is

slightly larger than the 2 cm'1 width observed for a single-crystal diamond

reference. The broader line width is due, in part, to scattering by grain
boundaries, which reduces the phonon lifetime. The one—phonon diamond line
exhibited some asymmetry, due to the Fano-effect, and was superimposed on a

very low photoluminescence background [3]. There was also negligible

scattering intensity between 1500 and 1600 cm'1, due to spz-bonded

nondiamond carbon impurities.

The Fano interaction is attributed to a quantum mechanical interference
between the Raman phonon and transitions from the broadened impurity band to
continuum states composed of excited acceptors and valence band states [3, 4].
The Fano line shape is asymmetric, with an intensity enhancement of the high
wavenumber side and an intensity reduction on the other flank (low wavenumber

side). Fano-like Iineshapes are observed as the boron-dopant concentration

3

reaches 1019 cm' , and above. The doping level may be qualitatively estimated,

based on the level of asymmetry. For the observed asymmetry, the doping level

37

was estimated to be in the high 1019 cm'3 range [4-6]. It Should be mentioned,

however, that the dOping level threshold for the appearance of the diamond line

asymmetry varies somewhat from report to report [7].

2.1.2 Nanocrystalline Boron-Doped Diamond

The nanocrystalline BDD thin films were deposited using microwave-
assisted CVD, as described above. The films were deposited on scratched and

cleaned Si (p-Si (100), < 0.001 0 cm), using the following source gas mixture:
1 sccm CH4, 4 sccm H2, 1 sccm B2H5 (0.1%) diluted in H2, and 94 sccm Ar. The
microwave power was 800 W, the system pressure 140 torr and the estimated
substrate temperature was 800 °C. The growth time was 2 h, resulting in a film

thickness of ca. 4 pm, as determined from weight change measurements. The
resulting electrodes were highly conductive (< 0.05 0 cm) being prepared from a
gas phase BZHG concentration of approximately 10 ppm. The electrodes were
used as prepared, with no additional pretreatment (e.g., polishing) or potential

cycling required for activation.

2.1.3 Freestanding Boron-Doped Diamond OTE

A freestanding, mechanically polished, BDD disk (380 pm thick and 8 mm
diameter) was used as the optically transparent electrode (OTE) for the UV/vis
measurements of ferrocene and cytochrome c, which are discussed in Chapter 7.
This disk was grown by microwave plasma CVD (2.54 GHz) In a customized

reactor at the Naval Research Lab, based on a 5 kW commercial system (model

38

5400, ASTeX, Inc). The ultrahigh-purity reactants (CH4 and BzHe prediluted

in H2) were hydrogen purified, using a Pd diffusion cell. A computerized system
was used to control and monitor the reactor pressure and reactant gas flow rates.
Samples were grown on 2 in diameter refractory metal substrates, which
released freestanding plates of CVD diamond upon rapid cooling from the
deposition temperature. The diamond plate was then laser machined to the
required sample size and cleaned as follows: aqua regia followed by a boiling
mixture of concentrated sulfuric and nitric acids. The sample was then
mechanically polished using a resin-bonded diamond wheel (Coburn Eng. Co.
Ltd.). Further details of this growth process are reported In [8]. The other
freestanding films discussed in Chapter 7 have the following thicknesses:
(I) freestanding CVD white diamond disk, 230 um; (ii) polished BDD on CVD
white diamond disk, 210 pm; and (iii) unpolished BDD on CVD white diamond
disk, 220 pm.

Specifically, the growth conditions used for the freestanding disk consisted

of a tungsten substrate at 810 °C, reactor pressure of 117 torr, 5.0 kW of
microwave power, and gas flow rates of 115 sccm for H2, 8 sccm for CH4, and

0.4 sccm of 0.1% BZHG diluted in H2. After polishing, AFM (Nanoscope Ill, Veeco
Instruments, Santa Barbara, CA) measurements in the contact mode revealed a

smooth surface, with a root-mean-square roughness of 7 nm over a 10 um2 area.

39

2.1.4 Boron-Doped Diamond IR OTE

The OTE used for the IR measurements reported in Chapter 7 consisted
of a microcrystalline BDD thin film (~ 2 pm thick) deposited onto a highly resistive
(i.e., very low doped) Si substrate (650 pm thick, Texas Instruments, Inc.) via
microwave-assisted CVD. The deposition procedure was the same as is detailed

above, in Section 2.1.1. A nonconducting Si substrate was used to maximize the
optical throughput. Typically, films were doped with 1 ppm BZHS diluted in H2

using growth times ranging from 4 - 10 h. The effect of the growth time on the

optical throughput will be discussed in the last chapter.

2.2 Sample Preparation

All aqueous solutions were prepared using water purified using either a
Barnstead E-Pure or Millipore system (resistivity > 17 MO cm). Ferrocyanide
(Aldrich) was recrystallized from ethanol. All other analyte and solution

preparation procedures are described below.

2.2.1 Ferrocene

Ferrocene and sodium perchlorate (NaCIO4) were purchased from Aldrich
and used as received, without further purification. Tetrabutyl ammonium
perchlorate (TBACIO4) salt was purchased from Fluka and used without further

processing. Dichloromethane (Spectrum, spectrophotometric grade) was

distilled over calcium hydride. Acetonitrile (Merck, HPLC grade) was dried by

40

distillation over calcium hydride, followed by twice passing over a column of

activated alumina.

2.2.2 Cytochrome c

Cytochrome c stock solutions were obtained from the Ferguson-Miller lab
(MSU Biochemistry), courtesy of Dr. Denise Mills. For the experiments described
herein, the stock solution (3.5 mM cytochrome c, 0.3 M NaCI, 10 mM Tris HCI
pH 7) was diluted with purified water to the desired protein concentration. Stock
solutions of cytochrome c were prepared by chromatographically purifying and
concentrating horse heart cytochrome 0 purchased from Sigma Chemical
Company, following a published procedure [9]. For ionic strength studies,
measured amounts of NaCI were added to the protein solution, described above,
to yield the desired ionic strength. Cytochrome 0 concentrations were

determined optically from the amplitude of the reduced minus oxidized difference

band at 550 nm using 17 mM'1 as the extinction coefficient [10].

2.2.3 Cytochrome c Oxidase

The cytochrome c oxidase (cco) protein samples were also received as
part Of a collaboration with the Ferguson-Miller lab. The bovine heart oxidase
was isolated from bovine heart muscle, as previously described by Yoshikawa
[11]. The enzyme was concentrated to 0.7 - 1.0 mM in a solution of 40 mM
HEPES buffer pH 7.4, by centrifugal filtration. The Rhodobacter Sphaeroides
oxidase was isolated from the YZ-300 strain [12] engineered to over-express a

histidine-tagged form of the enzyme [13] and was cultured aerobically in

41

Sistrom’s growth media at 30 °C with rapid swirling. Global 15N labeling was

achieved by substituting 15NH4CI (99% 15N; Isotec, Miamisburg, OH) into the

growth media which constituted ~95% of the cell’s nitrogen. The isolated
enzyme was concentrated In a 40 mM HEPES pH 7.4, 1 mM EDTA and 0.1%
Iauryl maltoside buffer medium with a 50 kDa centrifugal filter (Ultrafree filter,

Millipore, MA). Deuterium exchange of both oxidases was performed by
repetitive concentration and incubation of the enzyme in 99.5% DZO enriched

buffer. To ensure maximal exchange, four concentration steps and four
incubation steps of 6 - 8 h were performed (and the enzyme was turned over).

Enzyme concentrations were determined from UV/vis absorbance spectra, using

26°6— 3630 = 24 mM'1 cm‘1 [14].

2.3 Cyclic Voltammetric Experiments

Figure 2-1 shows a drawing of the standard 3-electrode glass cell that was

employed for the cyclic voltammetric (CV) experiments described in Chapters 3

and 4 [1]. The cell volume was approximately 5 mL. In this cell, a ChemRaz®

(Ace Glass) o-ring defined the electrode area (0.39 cm2). This type of o-ring was

selected due to its chemical resistance toward organic solvents. Contact to the
working electrode was made through a conductive lnGa alloy applied to the
backside of the cleaned silicon substrate. This Side of the substrate was pressed

onto a copper current-collecting plate, and electrical connection was made to this

plate. In the ferrocene measurements, a Ag wire, encased in Teflon®, served as

42

the quasi-reference electrode, and 3 Pt mesh was used as the counter electrode.
For the CV measurements of cytochrome c, a similar, smaller-scale cell was
employed, requiring a much smaller volume (ca. 0.5 mL), the counter was a Pt
wire coil, and the reference was changed to a commercial (Cypress Systems,

Lawrence, KS) Ag/AgCl (1.0 M KCI) electrode. The exposed electrode area in
this cell was 0.2 cm2, as determined by a Viton o-ring. All solutions were

deaerated by gently bubbling with nitrogen, which was then kept flowing over the

solution during measurements.

Figure 2-1 Conventional 3-Electrode Glass Electrochemical Cell

A diagram of the glass electrochemical cell used in cyclic voltammetric experiments.
Cell volume ca. 5 mL. A conventional three-electrode configuration is employed,
including (a) a copper current collector plate, (b) a Ag wire quasi-reference electrode,
(c) 3 Pt mesh counter electrode, (d) a nitrogen gas inlet, (a) an o-ring, and (t) a BDD
thin-film on Si substrate as the working electrode.

 

 

 

 

t
t
I
i

a:
it
:3:
15%

b\
s
it, .
a?
a '-" I!
\ Eagwmyamg f

Im|
mi]

43

Cyclic voltammetric experiments were conducted using either an EG&G
Princeton Applied Research Model 263A (cytochrome c) or a CH Instruments
Model 650 (ferrocene) computer-controlled potentiostat. Low analyte
concentration (0.1 mM) and iR compensation were used for the CV
measurements of ferrocene, in an effort to minimize ohmic distortion. Digital
compensation of the uncompensated solution resistance, via a positive-feedback
method, was applied at 70% to generate this data, which was used for the rate
constant calculations.

Glassy carbon electrodes (GO-30), obtained from Tokai Ltd. (Japan), were
used for some CV measurements of ferrocene described in Chapter 3. These
electrodes were polished smooth on felt pads, using slurries of 1.0, 0.5 and
0.3 pm alumina (Buehler) in ultrapure water. The electrodes were rinsed with

deionized and purified water between each polishing step.

2.4 Spectroelectrochemistry

2.4.1 CcO Experiments - Au Grid OTTLE Cell

Figure 2-2 shows a schematic of the optically transparent thin-layer
electrochemical (OTTLE) cell used for the spectroelectrochemistry of cytochrome
c oxidase, described in Chapter 6. This cell was custom built, according to a
design proposed by Mantele and coworkers [15, 16]. The protein solution formed

a thin layer on a 70% transparent, 6 pm thick gold grid working electrode (200

wires per inch; Buckbee-Mears, St Paul, MN) between two Ban windows. The

gold mesh electrode was in electrical contact with a platinum wire counter

44

electrode and a AglAgCl reference electrode (0.197 V vs. NHE; BAS

Instruments, West Lafayette, IN).

Figure 2-2 Optically Transparent Thin-Layer Electrochemical Cell Design
Employing a Au Grid Optically Transparent Electrode

Design schematic of the optically transparent thin-layer electrochemical (O TTLE) cell
used for U V/vis and F TIR spectaoelectrochemical studies of cytochrome c oxidase.

Components of the cell: (a) Delron plastic body (b) window (25 mm) mounting piece (0)
o-ring (d) Au mesh working electrode (e) Au ring contact for working electrode (0
solution inlet channel (g) Pt wire counter electrode (h) reference electrode port/solution
outlet channel and (i) lR windows (top 20 x 2 mm, bottom 25 x 2 mm). Drawing not to
scale.

 

 

 

t -
e \ ‘
h I
"\i‘: . . "if"? x
\ .:.'. .‘ \ \
;\‘ . \\;\\ \\_\";V
:36:- I .~\. ‘\.\" \ N“
If: \ ‘
d A :‘LL'EtE‘LI'i \
"'1'“. ”1 a
c \

 

 

The cell was filled with the 25 mm window, and the working and counter
electrodes in place, in the DelronTM plastic body, as is shown in Figure 2-2. A

10 pL drop of protein solution was placed onto the gold grid and the two windows
were progressively pressed together by screwing the top window into place
against the resistance of rubber o-rings, until the sample drop completely filled

the 20 mm diameter. Contact to the counter and reference electrodes was

45

established by filling the space around the periphery of the windows with
supporting electrolyte solution via a tangential port. Although the sample solution
was in direct contact with the surrounding electrolyte solution, the rate of dilution
was minimal (< 5% in 24 h).

The sample solution consisted of 0.7 - 1 mM cco in 40 mM HEPES buffer
(pH 7.4), 200 mM KCI as the supporting electrolyte, and a cocktail of mediators
(Table 2-1) at 40 DM each. This series of mediators was necessary to accelerate
the electrochemistry. At such a relatively low concentration, the mediator
solution did not interfere with the spectra of the cco enzyme.

Prior to use, the electrode grid was cleaned in an acid solution (3:1 (v/v)
H2804 and 30% H202) by soaking for 30 min followed by copious rinsing with

ultrapure water. The cleaned grid was then chemically modified by immersion in
a 2 mM cysteamine (Aldrich) aqueous solution, according to procedures
established for obtaining direct electrochemistry of cytochrome c [17, 18]. After
2 h, excess modifier was removed by rinsing with ultrapure water, and the mesh
was dried under a stream of nitrogen gas. The cysteamine formed a
monomolecular layer on the gold electrode by self-assembly, due to strong Au-S
bonds. This monolayer prevents proteins from adhering and denaturing on the
surface.

Proteins may interact directly with the modifier to undergo electron transfer
at the electrode. However, when large proteins contain deeply buried metal
centers that result in sluggish electron transfer between the protein and the

modified electrode, additional efforts are required to facilitate rapid electron

46

transfer. Therefore, a cocktail of mediators with midpoint reduction potentials,
spaced ~0.06 - 0.10 V, was added to the sample solution to exchange electrons
between the electrode and the enzyme. Table 2-1 lists the mediators and their

redox midpoint potentials.

Table 2—1 Mediators Used in Spectroelectrochemical Measurements of
Cytochrome c Oxidase

 

 

Em (V) vs.

Mediator AgIAgCl Solvent
Ferricyanide 0.216 "'20
Dimethylparaphenylendiamine (DMPPD) 0.163 DMSO
1,1-dimethylferrocene 0.133 DMSO
Tetrachlorobenzoquinone 0.072 DMSO
Tetramethylparaphenylendiamine (TMPPD) 0.062 DMSO
2,6-dichlorophenolindophenol (DCPIP) 0.009 H2O
Hexaammineruthenium (III) chloride -0.008 H2O
1,2-napthoquinone -0.063 DMSO
Trimethylhydroquinone -0.108 DMSO
Menadione -0.220 DMSO
2-hydroxy-1,4-naphthoquinone -0.333 DMSO
Anthraquinone-Z-sulfonate -0.433 DMSO

Benzyl viologen -0.568 H2O
Methyl viologen -0.654 H2O

 

DMSO, dimethylsulfoxide

47

Spectroscopy was performed in the OTTLE cell, as a function of applied
potential. An EG&G Princeton Applied Research 263M potentiostat was used to
initiate the electrochemical reactions. This instrument was operated at a scan
rate of 0.010 V/s and under no iR compensation. The UV/vis spectroscopy was
recorded with a Hewlett Packard 8453 diode array spectrometer; and FTIR
spectroscopy was performed, separately, with a Bruker Equinox 55 instrument,
using a MCT detector. The IR optics consisted of a KBr beamsplitter, used in
conjunction with a Ge filter (OCLI L02547-9, Santa Ross, CA), which reduced
spectral bandwidth. Temperature control was maintained at 10 1 0.01 °C as
described below (section 2.5). A stable background spectrum could be achieved
after temperature equilibration of ~ 60 min. At this point, the potentiostat was
used to electrochemically poise cytochrome c oxidase in a reduced or oxidized
form. The enzyme was held at 0.50 V (vs. Ag/AgCl) to generate the oxidized
form and —0.50 V (vs. Ag/AgCI) for the reduced form. Electrolysis was complete
within 2 min and was recognized by a steady-state current reading and identically
shaped successive spectra. Typically, the oxidized and reduced spectra were a
ratio of 800 scans for a single-beam spectrum. Each 800-scan acquisition lasted

-1 , . .
~ 80 5. All spectra were collected at 4 cm resolutIon. No baseline corrections

were made to any of the spectra.

2.4.2 BDD UVlvis OTTLE Cell

The freestanding diamond OTE (growth described in Section 2.1.3) was

mounted into a specially designed thin-layer cell for the transmission UV/vis

48

spectroelectrochemical measurements detailed in Chapter 7. A schematic of this

design is shown in Figure 2-3. This is a very Simple and user-friendly design, as

the entire cell is conveniently housed in a 1 cm2 quartz cuvette.

Figure 2-3 UV/vis Spectroelectrochemical Cell Employing a Freestanding
Boron-Doped Diamond Optically Transparent Electrode

Side and front views of the UV/vis OTTLE cell designed for freestanding BDD disk
OTES. As presented, the cell consists of (a) the BDD freestanding disk working

electrode, (b) a mica spacer, (c) the Kel-FTM body, (d) a Pt foil contact to the working
electrode, (a) a Ag wire reference electrode, (0 8 Pt coil counter electrode, (9) a

standard 1 cm2 quartz cuvette and (h) springs. Drawing not to scale.

 

light—>

 

 

   

SIDE FRONT

The thin-layer cavity was formed between the diamond OTE and the wall

of the cuvette, using an 80 — 150 pm mica spacer (cell volume ~ 3 — 6 [.IL). An
o-ring, contained within the KeI-fm body, combined with spring rods, pressed the
electrode against the spacer. Electrical contact with the diamond disk was made
with a piece of platinum foil, also contained within the KeI-fTM body. A small

channel was cut in the spacer to connect the thin-layer cavity with the solution

49

contacting the reference and counter electrodes. A Ag wire, coated with Teflonm,

was used as the reference electrode. The potential of this reference electrode
was —0.01 V relative to a saturated calomel reference electrode, as measured in
1 M KCI. A coil of Pt wire was used as the counter electrode. The solution was
introduced into the cell, using a syringe, taking care to avoid trapping air in the
thin-layer cavity or in the solution compartment hosting the reference and counter

electrodes. The active area of the electrode was defined by the opening of the

mica spacer (0.39 cm2).

UV/vis spectrophotometric measurements were made, using a Hewlett
Packard Model 8453 spectrometer with a diode array detector. An EG&G
Princeton Applied Research 263M potentiostat was used to control the applied
potential. A cyclic voltammetric waveform was used with a 0.002 V/s scan rate
and no iR compensation when the current response was measured
simultaneously with the spectra. Othenrvise, a potential step platform was
employed, in which the potential was stepped to a constant value while acquiring
an Optical Spectrum. A steady-state current reading and identical successive

spectra indicated that complete electrolysis had been achieved.

2.4.3 BDD IR OTTLE Cell I

The diamond OTE for FTIR measurements was a diamond thin-film
deposited on undoped Si, described above In Section 2.1.4. For the ferrocene

measurements reported in Section 7.3.2, this OTE was mounted onto the metal

body of a custom designed cell shown in Figure 2-4. A TeflonTM spacer defined

50

the thickness (25 to 200 pm) of the thin-layer cavity, which was formed between
the AgCI window and diamond electrode. The cell volume is ca. 3 pL using a
25 um spacer. The cell was filled by filling the solution port containing the
reference and counter electrodes. An lnGa alloy was used to make a contact

between the metal body and diamond electrode. A Pt coil counter electrode was
wrapped around a Ag wire reference electrode that was encased in Teflonm.

This cell was the prototype for an IR spectroelectrochemical cell designed for

BDD thin-film on Si OTES.

Figure 2-4 IR Spectroelectrochemical Cell Employing a Boron-Doped
Diamond Optically Transparent Electrode - Model I

The spectroelectrochemical cell consists of the following components (a) a TeflonTM
spacer, (b) a AgCI window (25 x 2 mm), (c) the metal body, (d) 3 Pt wire counter

electrode, (a) 3 Ag wire encased in Teflon (reference electrode), (0 the BDD OTE
working electrode, and (g) the lnGa contact to the working electrode.

 

 

 

 

 

 

 

 

 

 

 

51

A major drawback of this design was the method in which the OTE was
mounted in the cell. One difficulty encountered in working with the BDD thin-
films on nonconductive Si was the necessity to make electrical contact to the
front of the electrode, the same side, which is exposed to the analyte solution.
Therefore, the electrode was placed into the metal body and an epoxy mask was
used to prevent the electrical contact point from encountering the solution. The
epoxy mask did Shield the lnGa alloy from the solution; however, it did not enable
easy exchange of OTES, as OTES were sacrificed upon removal from this cell. In
our next design, we aimed to find a different method for shielding the working
electrode contact from the solution without permanently fixing the OTE to the cell
body.

No electrochemical characterization was carried out for this cell. At the
time this cell was used, there was no capability to produce films using the gas
phase dopant source. Instead, films were doped using a solid B source,
requiring longer growth times, and therefore, yielding either highly conductive,
thicker films with poor transparency, or lightly doped, thinner films with sufficient
optical transparency. The thinner films were conductive enough to sustain
electron transfer reactions and were used in the potential step experiments.
However, attempts to measure CVs at such films were unsuccessful. Potential
step spectroelectrochemical experiments for ferrocene demonstrated that BDD
thin-films on Si were useful as OTES in the IR region of the spectrum. Upon this

finding, we quickly changed our OTTLE cell to the design described below.

52

2.4.4 BDD IR OTTLE Cell II

 

Figure 2-5 IR Spec; --' ‘ L ' ‘ Cell Employing a Boron-Doped
Diamond Optically Transparent Electrode — Model II
Schematic of the three-piece IR OTTLE cell designed to incorporate a BDD thin film on

Si OTE. The cell was constructed from Kel-F plastic and consists of the following
components (a) a solution inlet port, (b) the reference electrode/solution outlet port,
(c) the BDD thin-film on Si OTE, (d) a Pt wire counter electrode, (9) the Pt ring contact
for the working electrode, (f) o-rings, and (9) an IR window (9 x 5 mm).

9

 

IR WINDOW BDD OTE
MOUNTING PIECE CENTER BODY MOUNTING PIECE

The spectroelectrochemical cell used for the FTIR measurements of
ferrocyanide and cytochrome 0 (Sections 7.3.4 and 7.3.5) consisted of a three-
piece assembly constructed from KeI-fm. For such measurements, a diamond
thin film deposited on undoped Si, described above in Section 2.1.4, was used as
the OTE. The diamond OTE was mounted into the thin-layer cell, as shown in
Figure 2-5. This piece was held onto the center body by four screws. A 1 — 3 ].IL
drop was placed on the diamond OTE and pressed to a thickness of ca. 10 pm
by screwing the NaCl window piece into place. The volume around the NaCI

window, containing the Pt wire counter electrode, was then filled with electrolyte

53

via an external port. Electrical contact to the diamond electrode was achieved by
pressing a flat Pt ring onto its surface. An o-ring with a smaller diameter than the

Pt ring ensured that the solution was isolated from this contact. The reference
electrode, 3 Ag wire encased in Teflonm, fit into a small cavity in the center body.

FTIR spectra were measured, using a Bruker Equinox 55 Spectrometer
with a MCT detector; the potential was applied by an EG&G Princeton Applied

Research 263M potentiostat in potential step mode. Typically, difference spectra

were an average of 200 scans acquired at 15 °C, using the temperature control

apparatus detailed later in Section 2.5. All spectra were collected at 4 cm'1

resolution, and no baseline corrections were made. As was mentioned for the
cco studies, temperature equilibration required ~60 min. At this point, the
potentiostat was used to electrochemically poise the analyte in a reduced or
oxidized form. Within 2 min, the electrolysis was complete and was indicated by

a steady-state current reading and identical successive spectra.

2.5 Temperature-Control Set Up

The necessity for precise temperature control during IR
spectroelectrochemical measurements will be discussed in Chapter 5. The

system designed to meet such experimental needs is shown in Figure 2-6. This
system maintained a very precise temperature (i 0.005 °C), yet allowed for facile

sample changing and high optical throughput.

54

Figure 2-6 Temperature-Control Set Up

Schematic of the experimental setup used for maintaining precise temperature control
during the spectroelectrochemical experiments. The components include (a) copper
coil, (b) liquid nitrogen dewar, (0) copper transfer line insulation, (d) dewar transfer arm,
(e) nichrome heater element, (0 Si diode temperature sensor, and (g) cryostat box. As
depicted, the nitrogen gas flow originates in the copper coil and is vented from the
cryostat box.

 

Lake Shore 340
T control unit

 

 

 

 

\ —1
gas flow
inlet fl

/

t~
\(D

 

gas flow
F outlet

 

 

C

 

 

 

A

 

::
69

 

 

 

 

 

 

 

 

 

 

 

 

9

 

 

 

The sample temperature was controlled by surrounding the sample with a
stream of gas at a precisely controlled temperature. To achieve this, nitrogen
gas, at room temperature, was flowed into a copper coil, which was submerged
into a dewar of liquid nitrogen. The gas was then passed through a glass-dewar

transfer arm that contained a nichrome heating element, used to heat the gas to

55

the desired temperature in a controlled manner. The heated gas flowed into the
cryostat box that was located In the sample compartment of the FTIR

spectrometer.

Figure 2-7 Cryostat Box Diagrams

Top and side views of the cryostat box sample holder employed in the IR
spectroelectrochemical measurements. The components include: (a) Spectroelectro-
chemical cell mount, (b) sample compartment adapter, (0) Germanium filter,
(d) temperature sensor, (e) temperature-controller connection, (f) electrochemical
connections, and (g) KBr infrared window.

gas flow

inlet gas flow
a \ F outlet

 

 

 

 

 

 

 

 

 

 

 

 

TOP VIEW SIDE VIEW

The cryostat box, shown in Figure 2-7, was constructed from DelronTM
plastic and contained a removable top. This style was well suited for the IR
spectroelectrochemical measurements described in the following chapters
(Chapters 6 and 7), as it allowed for quick purging (i.e., within 20 S) of the sample
compartment, precise temperature control, and easy sample changing. The box

was designed to fit entirely inside of the sample compartment of the Bruker

56

Equinox 55 spectrophotometer. A mounting screw was used to ensure that the
position Of the box was reproducibly and properly aligned in the IR beam.

The front side of the box was equipped with an outlet for the temperature
controller and for electrochemical connections between the potentiostat and the
electrodes in the sample cell. A Si diode temperature sensor (Lake Shore
Cryotronics, Inc., DT-471-BR), mounted onto an aluminum-alloy heat-Sink
material, was attached to the inside wall of the cryostat box. The Si diode was
important in this design because it had a very fast response time, relative to other
types of temperature sensors (e.g., thermocouples or platinum wires). Two
sample compartment adaptors were used to seal the cryostat box to the
instrument, maintaining an air-free purge. Purging was important to suppress
spectral contributions from the air, such as those due to carbon dioxide and
water vapor. The box was sealed from the instrument by attaching a KBr window

and a Ge window. The Ge window acted as a 3.3 pm filter, cutting out high-
frequency radiation greater than 4500 cm'1. Decreasing the bandwidth, in this

way, served to improve the Signal-to-noise ratio and decreased required
scanning time. The spectroelectrochemical cell was mounted onto a support that
ensured maximum optical throughput. The temperature was controlled to
i 0.005 °C by a control unit (Model 340, Lake Shore Cryotronics, Inc.). The
thermal contributions to the spectra were largely eliminated with this level of

control.

57

2.6

I1]

I21

I31

[41

[5.]

[51

[7-1

[8-1

[9-1

[10.]

[11.]

[12.]

References

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. 2000, 72, 3793-3804.

M. C. Granger and G. M. Swain J. Electrochem. Soc. 1999, 146, 4551-
4558.

U. Fano Phys. Rev. 1961, 124, 1866-1878.

J. W. Ager, W. Walukiewicz, M. McCluskey, M. A. Plano and M. l.
Landstrass Appl. Phys. Lett. 1995, 66, 616-618.

K. Ushizawa, K. Watanabe, T. Ando, l. Sakaguchi, M. Nishitani-Gamo, Y.
Sato and H. Kanda Diam. Relat. Mat. 1998, 7, 1719-1722.

P. Gonon, E. Gheeraert, A. Deneuville, F. Fontaine, L. Abello and G.
Lucazeau J. Appl. Phys. 1995, 78, 7059-7062.

M. Mermoux, B. Marcus, G. M. Swain and J. E. Butler J. Phys. Chem. B
2002, 106, 10816-10827.

J. K. Zak, J. E. Butler and G. M. Swain Anal. Chem. 2001, 73, 908-914.

D. L. Brautigan, S. Ferguson-Miller and E. Margoliash J. Biol. Chem.
1978, 253, 130-139.

M. Brunori, P. Sarti, A. Closimo, G. Antonini, F. Malatesta, M. G. Jones
and M. T. Wilson EMBO J. 1985, 4, 2365-2368.

S. Yoshikawa, T. Tera, Y. Takahashi, T. Tsukihara and W. S. Caughey
Proc. Natl. Acad. Sci. U. S. A. 1988, 85, 1354-1358.

Y. Zhen, J. Qian, K. Follmann, T. Hayward, T. Nilsson, M. Dahn, Y. Hilmi,
A. G. Hamer, J. P. Hosler and S. Ferguson-Miller Protein Expression and
Purification 1998, 13, 326-336.

58

[13.] D. M. Mitchell and R. B. Gennis FEBS Lett. 1995, 368, 148-150.

[14.] W. Vanneste Biochemistry 1966, 5, 838-848.

[15.] D. Moss, E. Nabedryk, J. Breton and W. Mantele Eur. J. Biochem. 1990,
187, 565-572.

[16.] F. Baymann, D. A. Moss and W. Mantele Anal. Biochem. 1991, 199, 269-
274.

[17.] P. M. Allen, H. A. 0. Hill and N. J. Walton J. Electroanal. Chem. 1984,
178, 69-86.

[18.] l. Taniguchi, K. Toyosawa, H. Yamaguchi and K. Yasukouchi J.
Electroanal. Chem. 1982, 140, 187-193.

59

CHAPTER 3

3 Ferrocene Electron Transfer Kinetics Measured
at Boron-Doped Diamond Electrodes

3.1 Introduction

Good quality, hydrogen-terminated diamond thin-film electrodes exhibit
properties superior to other carbon electrodes: (i) a low and stable background
current over a wide potential range, (ii) a resistance to fouling, due to weak
adsorption of polar molecules on the nonpolar, H-terminated surface, (iii) a wide
working potential window, (iv) excellent morphological and microstructural
stability at extreme anodic and cathodic potentials, and current densities, and
(v) good electrochemical responsiveness for several redox systems without any
conventional pretreatment [1]. There have been a number of papers reviewing
the electrochemical response of boron-doped diamond (BDD) thin films toward a
variety of aqueous redox systems [1-5]. However, there have been very few
reports of the electrode’s responsiveness toward nonaqueous redox systems [6-
9]. A question that needs to be answered is, do diamond electrodes possess the
same attractive features in nonaqueous media that they do in aqueous media?

Using the ferrocene/ferricenium redox couple as a test system, one can
begin to answer this question. Ferrocene is routinely used as a model system in
electrochemical kinetic studies, because it demonstrates “ideal” behavior in a
variety of solvents. Many researchers rely on the well-behaved kinetics of

ferrocene as an internal standard in nonaqueous electrochemistry [10-12]. The

60

one-electron oxidation of ferrocene to the stable ferricenium cation is a simple
electron transfer reaction, in that there are no mechanistically complicated steps
arising from adsorption or associated chemical reactions. Generally, this redox
couple exhibits highly reversible, well-behaved voltammetric responses at a
variety of electrode materials (Pt, Au, glassy carbon (GC)) [11, 13-21]. A
schematic for the electron transfer reaction of the ferrocene/ferricenium couple is
Shown below, in which the formal oxidation state of the metal changes from +2 to

+3.

4? SR“

 

 

+2 +3 ..
Fe ' Fe + e E0 = 0.307 V VS SCE
FERROCENE FERRICENIUM

A cyclic voltammetric study of the ferrocene/ferricenium redox reaction at
boron-doped, microcrystalline diamond thin-film electrodes as a function of the

scan rate, the composition of the electrolyte solution is reported on herein.

0

app , were calculated

Apparent heterogeneous electron transfer rate constants, k

from the dependence of the peak potential separation, AEp, on the scan rate,

according to the theory developed by Nicholson [22, 23].

Cyclic voltammetry (CV) is one of the most versatile electroanalytical
methods for the study of redox-active species. The technique has been used
extensively in electroanalytical studies and for the characterization of organic and
inorganic compounds, as well as for probing the properties of electrode surfaces,

and, more recently, for studies of biological molecules (see [24] for a recent,

61

comprehensive review of dynamic electrochemical methods). The effectiveness
of this technique lies in its ease of measurement combined with the wealth of
information contained in the resulting voltammogram. From cyclic voltammetric
studies, one can obtain information about the thermodynamic and kinetic
properties of an electroactive analyte, as well as information on the nature of its
interaction with the electrode surface. Often, CV is the first experiment
performed In the study of a new compound or electrode material.

A CV experiment is conducted by cycling the potential of an electrode,
which is immersed, Often unstirred, in a solution containing the electroactive
analyte, and measuring the resulting current. A potentiostat is used to apply the
potential and measure the resulting current. For optimal performance, three
electrodes are used in a standard CV experiment. A conventional three-
eIectrode glass cell employed for CV measurements was presented as
Figure 2-1 in the previous chapter. The electrolysis takes place at the working
electrode, to which a potential is applied relative to the reference electrode. The
current required for electrolysis flows through the counter electrode. In an
unstirred solution, diffusion is responsible for transporting the redox analyte to
and from the electrode/solution interface. The voltage applied to the working
electrode is linearly scanned from an initial value to a “switching” potential, at
which the direction of the scan is reversed. This results in a triangular waveform
of voltage as a function of time. The slope of the waveform gives the scan rate in
units of potential/time. The resulting voltammogram is a plot of the current

response as a function of the applied potential.

62

A reversible cyclic voltammogram, with the current limited by semi-infinite

linear diffusion, has a very characteristic shape, as is shown in Figure 3-1.

Important features of the voltammogram include the initial (and final) potential, E],

the switching potential, E), the reduction and oxidation peak potentials, Egedand

~red

on respectively, the reduction peak current, Ip

p , , and the oxidation peak

current, ig" . To a first approximation, the formal reduction potential, E", can be
determined as the average of the peak potentials.

Figure 3-1 Simulated Cyclic Voltammogram - Diffusion-Limited Case

Simulated cyclic voltammogram for a reversible, one electron transfer reaction at a

planar electrode (A = 0.39 cmz) for the diffusion-limited case at 0.1 V/s. The electrolyte
solution initial/y contained the reduced form but not the oxidized form of the redox

analyte. Other parameters used in the simulation include: E0 = 0.025 V, a = 0. 5, ke) =
0.1 cm/s, c0 = 1x10‘3 M, and D0 = 5x105 cmZ/S.

 

 

 

 

 

 

-20
40
-60 - red
. A I L I A lEp A 1 + I A I
-0.3 -0.2 -0.1 0.0 0.1 0.2 0.3
Potential (V)

63

3.2 Results

The oxidation of ferrocene is a one-electron process leading to a

chemically stable ferricenium cation. Figure 3-2A shows a cyclic voltammetric i-E
curve for 0.1 mM ferrocene in acetonitrile containing 0.1 M TBACIO4 at a

diamond film electrode. The scan rate was 0.2 V/s. The background response
at the same scan rate is also presented, for comparison. The scan was initiated
at 0.70 V and was swept in a negative direction, at a constant rate of 0.2 V/S, to
the switching potential, 0.10 V. Scanning in the negative direction results in the
reduction of electrogenerated ferricenium to ferrocene, as indicated by the peak
at approximately 0.45 V. At 0.10 V, the direction of the scan is reversed, and the
peak due to the oxidation from ferrocene to ferricenium is observed at
approximately 0.50 V.

According to theory, for a redox reaction to be kinetically reversible, as

dictated by the Nernst equation,

EZE”'_flln{m ,
NF [0] PO

a reaction must occur as fast as the potential sweep rate, so that the
concentrations of the reduced, [R], and oxidized, [0], forms are in equilibrium
with each other at the surface. In this expression, R is the molar gas constant
(8.314 J/mol K), T is temperature (K), n is the number of electrons transferred per

mole, F is Faraday’s constant (96,485 C/mol) and E” is the formal reduction

potential of the couple. Reversible systems are characterized by three criteria:

64

(1) a peak potential separation, AEp, of 59m mV

A5,, = E12" — 13;?" = 59/n m V

(2) an oxidation and reduction peak current ratio of 1

-(),\l
Il’ _

-r't~'t/

/)
(3) a peak potential separation, AEp, and therefore, an E1,2 value that is
independent of scan rate

Ejsx _ E;)Ud — AE

[)
2 ‘2'

 

151/2 =
A redox couple whose AEp increases with increasing scan rate is considered to

be quasi-reversible (10'2 >k° >10'6 cm/s) or irreversible (k0 < 10'6 cm/S). For

app app
a reversible system under control by semi-infinite linear diffusion, the peak

current, ip, is given by the Randles-Sevcik equation,

ip = (2.69x105)n3"2 A DJ,” v“ (7;.
The electrode area, A (cmz), the diffusion coefficient for the oxidized species,

Do (cm/s), the scan rate, v (V/s), and the bulk concentration, C; (moI/cms), yield

current in amperes [25]. From this relationship, it is seen that the current varies

linearly with the square root of the scan rate and the concentration.

65

Figure 3-2 Ferrocene Cyclic Voltammetric Response at a Boron-Doped
Diamond Electrode
(A) Cyclic voltammetric i-E curve for 0.1 mM ferrocene in acetonitrile containing 0.1 M

TBAC/O4 at 0.2 V/s. The featureless response for the blank electrolyte solution is also
Shown. (B) Oxidation peak current versus the square root of the scan rate plot for

0.1 mM ferrocene in acetonitrile containing 0.1 M TBAC/O4. Data are Shown for scan
rates from 0.05 to 1.0 V/s.

 

 

 

 

 

 

 

 

 

 

15 40 j
I A I —LtnearfIt
35» B

10» 3.: ,
.. 30.
<3 5? E *
P x L
E 0 § 20.
t3 . 5 15+»
'5' E 10.
'10)- E 5”

0T 0230.3 A 0.4‘0.5T 0.64 0.7 93.0 ‘ d2 ‘ 0T4 ‘ 0.6 ‘ ole L 11.0

Potential (V vs. Ag wire QRE) V10 ((V/s)”2)

A well-defined voltammetric response was observed for ferrocene at BDD

(Figure 3-2A) with a AEp of 0.074 V, an E1/2 of ca. 0.440 V, and a peak current

ratio of 1. The oxidation peak current varied linearly with the square root of the
scan rate between 0.05 and 1.0 V/s, as shown in Figure 3-28. The correlation
coefficient was 0.999 and the y—axis intercept was near zero. This indicates the
ferrocene oxidation reaction is limited by semi-infinite linear diffusion of the
reactant to the interfacial reaction zone. The calculated rate constant is
consistent with reversible electrochemical reaction kinetics. It can be seen that
the background current is low and featureless over the potential range—a

characteristic feature of diamond electrodes [1].

66

Cyclic voltammetric AEp values were recorded at different scan rates, to

determine the apparent heterogeneous electron transfer rate constant, kgpp, in

the three different electrolyte/solvent systems. The AEp can be used to

determine kgpp, using the classical treatment developed by Nicholson [22, 23].
This method relates AEp to the dimensionless kinetic parameter, lIt, and is useful

for determining rate constants for redox systems with peak separations of
0.065 — 0.20 V.
Nicholson generated a working curve demonstrating the variation of peak

potential separation with the dimensionless kinetic parameter [22]. From the

experimental AEp, the corresponding value of ti! is determined from the working

curve. If the scan rate and diffusion coefficients are known, the rate constant can

[.00] k,
DR

— ([0,, III/(itF/RT) '

be calculated from

’2”

 

 

0,22
In most cases, the assumption that [DO] is near unity is valid, with the
R

exception being cases of large differences between Do and DR, regardless of the
value of the transfer coefficient, a.
Table 3-1 lists the kgpp values calculated for both untreated diamond and

freshly polished GC. One of the properties of high-quality diamond electrodes is

the inherently active response for many redox systems, and long-term

67

maintenance of this response, without any conventional pretreatment. Triplicate

results were obtained and averaged for five scan rates between 0.05 and 1.0 V/s.

The k°

app values were calculated using Nicholson’s method, assuming that

O = 0.5. The diffusion coefficient for ferrocene in acetonitrile was taken as

DR: 2.4 x 10'5 cm2 s'1 [26]. The value used for ferrocene in dichloromethane,

DR = 2.0 x 10'5 cm2 S1, was calculated from the Stokes-Einstein relation,

6nr;r}=K—T,
I)

applying the Walden rule,

I), r] = constant .

It is assumed that the radius of the solvated ion, ri, does not change drastically in

acetonitrile relative to dichloromethane [27]. The viscosities, n, used for
acetonitrile and dichloromethane were 0.345 and 0.390 cP, respectively.

0

app values for diamond

There are no striking differences between the k

and those for GO in any of the three electrolyte/solvent systems. This suggests

that there is no significant affect Of the electrode microstructure (spz- vs.

sp3-bonding) on the electrode kinetics for this particular redox system.

Therefore, the most important factor influencing the kinetics appears to be the
density of electronic states, which is related to the number of charge carriers, at
the formal potential. Clearly, both electrodes possess a sufficient density of

states to support relatively rapid electron transfer kinetics.

68

Table 3—1 Apparent Heterogeneous Electron Transfer Rate Constants for
Ferrocene at Untreated Boron-Doped Diamond and Polished Glassy
Carbon Electrodes

 

 

0.1 mM Ferrocene BDD k3,, (cm/s) GC-30 kgpp (cm/s)
0.1M NaClO4/CH3CN 0.042 2‘. 0.015 0.059 : 0.014
0.1M TBAC|O4ICH3CN 0.048 i 0.015 0.051 i 0.003
0.1M TBACIO4/CH2CI2 0.008 t 0.002 0.006 i 0.001

 

BDD, Boron-Doped Diamond Electrode; GC-30, Tokai Glassy Carbon Electrode

The rate constants reported herein can be compared to values reported in

O

app values reported herein for

the literature for a variety of solid electrodes. The k

diamond and glassy carbon are 5-10 times larger than some values reported in

the literature [14, 17]. However, Panzer and Elving reported similar values for

AEp at 0.1 V/s in 0.1 M LiCIO4/CH3CN on glassy carbon [18]. The k°

app values

reported for platinum are generally 2 -4 times larger than those observed for
carbon electrodes. For example, the values reported in acetonitrile range from
0.02 to 0.22 cm/s [11, 13, 15, 16, 20, 21]. Much larger rate constants, up to
1 cm/s, have been observed for gold and platinum microelectrodes, using fast-

scan cyclic voltammetry [28].

0

app at either electrode

There appears to be no electrolyte cation effect on k

0

app observed for both

in CH3CN. There is, however, an effect of the solvent on k

electrodes; with the kgppin CH2CI2 being a factor of 4-6 lower than in CH3CN.

69

The difference is attributed to a dielectric constant for CH2C|2 that is a factor of

4.3 less than that for CH3CN (£(20 CC) = 9.08 vs. 38.8). The difference in

dielectric constant is slightly less than the difference in rate constants, which is a

factor of 5 lower for diamond and a factor of 8.5 lower for glassy carbon in

CH2CI2. The lower solvent dielectric constant can affect the CV response, from

which k°

app is determined, in at least two ways.

First, the solution will have a lower ionic conductance because of less
effective charge separation. This results in reduced ionic strength and, therefore,

less ion conductivity. The lower ionic conductivity Increases the magnitude of
iRu. In the experiment, the cell current, i, is measured as a function of Eappl,

which is the applied potential difference between the working and reference

electrodes. However, the true potential across the working electrode/solution

interface, EW, is related to Eapp. by
Ew = Euppl _ l.Ru _ (pref I
where RU is the uncompensated portion of the solution resistance and (Ilref is the

0

reference electrode potential. The evaluation of AED, and its relation to kapp,

becomes problematic if iRu becomes significant (2 mV or greater). In the present
data analysis, 70% of the iRu was electronically compensated for, so solution

resistance is expected to have little affect on the measured AEp values, or the

related kgpp values.

70

Second, changing the dielectric constant of the solvent can strongly
influence the electric double-layer structure—in particular, the magnitude of the
potential drop at the plane of closest approach, the outer Helmholtz plane, OHP
(Gouy-Chapman-Stern model) [25]. It is supposed that this is the main effect in
the present data. A lower dielectric constant means there will be less
accumulated ions at any given potential in the compact layer to compensate for

the excess surface charge on the electrode. Thus, for the excess charge on the
surface to be balanced by the charge in the solution, qm = -q5, there will have to

be a significant diffuse-layer thickness. A larger diffuse-layer thickness means
that the magnitude of the electric field at OHP, t’iCID/t5xx:x2 , will be less. This is
the so-called Frumkin effect [25]. The potential at the OHP, (1)2, will not be equal

to the potential in bulk solution, tits, because of the potential drop through the

diffuse layer. Thus, the effective potential is Eapp. — (1)2. The kgpp values can be

corrected for such a double-layer effect according to

k” k” (an—z)F¢2
. = . . ex
ti/rp true p RT

 

where kf’me is the true heterogeneous electron transfer rate constant, a is the

transfer coefficient, n is the number of electrons transferred per molecule, 2 is the
charge on the molecule, and F, R, and T have their usual meanings. This

correction is possible, if one can calculate (>2. However, one must know the

potential at the point of zero charge, in order to calculate this value; and, for

diamond, this potential is unknown. The fact that the relative rate constant

71

difference is larger for glassy carbon than for diamond may indicate that the
overall structure of the electric double-layer of these two carbon electrodes is

different.

3.3 Conclusions

The electron transfer kinetics for ferrocene were investigated at
microcrystalline, thin-film BDD electrodes as a function of the solvent and
electrolyte composition. The background current was low and stable with time.
Well-defined CV responses were observed in all three electrolyte/solvent

systems, consistent with quasi-reversible electron transfer reaction kinetics.

Measured values of k° were in the low to mid 10'2 cm/s range for the untreated

app

BDD thin-film electrodes, very similar to the values for freshly polished glassy
carbon. Hence, there is no apparent effect of the electrode microstructure on the
electrode kinetics for this redox system.

The kinetics are primarily affected by the density of electronic states at the
formal potential. Clearly, these two electrodes have a sufficient density of states
to support relatively rapid kinetics. A comparison of the rates measured in

dichloromethane and acetonitrile reveals the expected trend of decreased rates

0
app

in the lower dielectric solvent, dichloromethane. The relatively large k
observed for ferrocene at untreated BDD demonstrates the viability of using
these electrodes for electrochemical analysis of nonaqueous redox systems and

that the attractive features of this electrode material exhibited for aqueous redox

systems are retained for nonaqueous redox systems.

72

3.4

[I]

I21

I31

[41

[5-1

[5-1

[7-1

[81

[9-1

[10.]

[11.]

References

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. 2000, 72, 3793-3804.

T. A. Ivandini, B. V. Sarada, C. Terashima, T. N. Rao, D. A. Tryk, H.
Ishiguro, Y. Kubota and A. Fujishima J. Electroanal. Chem. 2002, 521 ,
117-126.

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 Materials 2002, 8, 171-188.

G. M. Swain In Thin Film Diamond; C. Nebel and J. Ristein, Eds;
Academic Press, 2002, in press.

J. Wang, G. M. Swain, M. Mermoux, G. Lucazeau, J. Zak and J. W.
Strojek New Diamond and Frontier Carbon Technology 1999, 9, 317-343.

S. Haymond, G. T. Babcock and G. M. Swain Electroanal. 2002, in press.

M. Yoshimura, K. Honda, T. Kondo, R. Uchikado, Y. Einaga, T. N. Rao, D.
A. Tryk and A. Fujishima Diam. Relat. Mat. 2002, 11 , 67-74.

Z. Y. Wu, T. Yano, D. A. Tryk, K. Hashimoto and A. Fujishima Chemistry
Letters 1998, 503-504.

S. Alehashem, F. Chambers, J. W. Strojek, G. M. Swain and R.
Ramesham Anal. Chem. 1995, 67, 2812-2821.

A. M. Bond, E. A. McLennan, R. S. Stojanovic and F. G. Thomas Anal.
Chem. 1987, 59, 2853-2860.

J. W. Diggle and A. J. Parker Electrochim. Acta 1973, 18, 975-979.

73

[12.]

[13.]

[14.]

[15.]

[16.]

[17.]

[18.]

[19.]

[20.]

[21.]

[22.]

[23.]

[24.]

[25.]

R. R. Gagne, C. A. Koval and G. C. Lisensky Inorg. Chem. 1980, 19,
2854-2855.

N. R. Armstrong, R. K. Quinn and N. E. Vanderborgh J. Electrochem. Soc.
1976, 123, 646-649.

L. Bjelica, R. Parsons and R. M. Reeves Proc. - Electrochem. Soc. 1980,
80-3, 190-212.

T. Gennett, W. Geiger, B. Willett and F. C. Anson J. Electroanal. Chem.
1987, 222, 151-160.

K. M. Kadish, J. Q. Ding and T. Malinski Anal. Chem. 1984, 56, 1741-
1744.

M. Noel, V. Suryanarayanan and R. Santhanam Electroanal. 2000, 12,
1039-1045.

R. E. Panzer and P. J. Elving J. Electrochem. Soc. 1972, 119, 864-874.

I. Ruff, V. J. Friedrich, K. Demeter and K. Csillag J. Phys. Chem. 1971, 75,
3303-3309.

M. Sharp, M. Petersson and K. Edstroem J. Electroanal. Chem. 1980,
109, 271-288.

M. Sharp Electrochim. Acta 1983, 28, 301-308.

R. S. Nicholson Anal. Chem. 1965, 37, 1351.

R. S. Nicholson and I. Shain Anal. Chem. 1964, 36, 706.

J. L. Anderson, L. A. Coury and J. Leddy Anal. Chem. 2000, 72, 4497-
4520.

A. J. Bard and L. R. Faulkner Electrochemical Methods: Fundamentals
and Applications; 2nd ed.; John Wiley 8 Sons, Inc., New York, 2001.

74

[26.] T. Kuwana, D. E. Bublitz and G. Hoh J. Am. Chem. Soc. 1960, 82, 5811-
5817.

[27.] J. Koryta, J. Dvorak and L. Kavan Principles of Electrochemistry; 2nd ed.;
John Wiley & Sons, New York, 1993.

[28.] D. O. Wipf, E. W. Kristensen, M. R. Deakin and R. M. Wightman Anal.
Chem. 1988, 60, 306-310.

75

CHAPTER 4

4 Direct Voltammetry of Cytochrome c at
Boron-Doped Diamond Electrodes

4.1 Introduction

Electron transfer reactions play an integral role in many biological
processes, including photosynthesis and respiration. Additionally, many
biosensor and biocatalytic devices are based on electron transfer processes
involving biomolecules. Thus, the physiological relevance and exciting
application possibilities have made direct bioelectrochemistry the subject of a
number of recent reviews [1-3]. The practice of directly measuring biological
electron transfer reactions spans over three decades, beginning with early work
aimed at determining the reduction potentials of proteins via potentiometric
methods [4]. From protein electrochemistry, one can Obtain valuable
thermodynamic and kinetic information, which is useful when accompanied by
detailed structural information available from crystallographic and spectroscopic
methods. Biochemists and bioinorganic chemists routinely use cyclic
voltammetry to determine the solution redox potentials of small proteins and to
investigate the influence of site-directed mutations and experimental conditions
on protein redox centers [5-7]. Electrochemists, however, are primarily
concerned with using voltammetry to understand the interfacial interactions
between proteins and electrode surfaces via electron transfer kinetics, adsorption

kinetics, and thermodynamic measurements [1]. From the outset, the difficulties

76

of making direct electrochemical measurements of proteins were realized, as an
early report of the direct reduction Of the electron transfer protein cytochrome c at
a platinum electrode indicated that application of very negative potentials were
necessary (La, -1.2 V) to observe reduction [8]. The field has progressed
beyond the first measurements of cytochrome c to include direct methods that
routinely observe quasi-reversible electron transfer for small electron transfer
proteins, such as cytochrome c, the ferredoxins, and the blue-copper proteins,
and has expanded to include the investigation of increasingly complex enzyme
systems with multiple redox-active metal centers, such as cytochrome c oxidase.
However, cytochrome c continues to be frequently used in the electrochemical
investigation of new electrode materials and chemically modified surfaces and
remains the subject of numerous bioelectrochemistry papers per year.
Cytochrome c is a relatively small (12,400 Da), water-soluble protein that
functions as an electron carrier in many biological reactions. The redox-active
metal center consists of a Fe-containing heme unit, in which the Fe is ligated to a
Met and His residue, as shown in Figure 4-1A. The structure of the heme 0
center is also shown in Figure 4-1, for clarity. The protein contains three
conserved core helices which form a "basket" around the heme group with one
heme edge exposed to the solvent [9]. Located near the heme edge is a
phenylalanine residue, Phe82, which also serves as the exit point for a positive
dipole moment. This may be the region involved in the closest approach of the
heme edge to the electrode surface [10]. The redox reaction of cytochrome 0

occurs through a single, outer-sphere electron transfer step for which the formal

77

reduction potential of the heme Fe(l|l) to Fe(ll) transition is accepted to be
0.260 V (vs. NHE) [3]. Physiologically, electrostatic interactions involving several
lysine residues on the surface of cytochrome c enable it to dock at its reaction
partner protein, promoting the necessary redox center orientations required for
electron transfer [11-15]. Electrochemists have used this knowledge to mimic the
cytochrome c docking site at metal and chemically-modified electrode surfaces,

in an effort to facilitate direct, rapid electron transfer between cytochrome c and

the electrode surface [3, 16-21].

Figure 4-1 Cytochrome c Heme Redox Center

(A) Heme coordination scheme for class l type cytochromes, in which the redox-active
heme Fe is ligated to Methionine and Histidine amino acid residues. (B) Detailed

schematic of the cytochrome c heme, heme c, structure.

A B
Cys\
S
Met SvCYS
\s
|
H
I
N
6
/ HIS
HN O OH

OH O

The cyclic voltammetric response for cytochrome c has been studied at a
variety Of electrodes. It has been found to undergo rapid electron transfer with
clean metal (e.g., Au, Pt, Ag), carbon, and metal oxide electrodes, as well as with
chemically-modified metal electrodes [10, 22-25]. As a well-characterized,

electron transfer protein, cytochrome c has been used extensively as a test

78

system for redox reactions of proteins. The electrode kinetics of this system are
known to be strongly dependent upon a combination of the interfacial
electrostatic and chemical interactions, which are derived from the enzyme
structure and the nature of the electrode surface. Fouling of the electrode
surface caused by protein denaturation and aggregation plagued early work
focused on direct voltammetry Of cytochrome c at metal electrodes. However, it
has been Since been shown that two factors are necessary to Observe relatively
rapid electrode reaction kinetics for cytochrome c at bare electrodes: (i) a
chromatographically purified cytochrome c solution (is, free from oligomeric
contaminants) and (ii) electrode pretreatment to produce a clean, hydrophilic
surface [10]. Armstrong and coworkers observed a marked difference in the
electron transfer rate of cytochrome c at the edge vs. basal plane of pyrolytic
graphite. The authors report a drastic increase in the kinetics at the polished
“edge” plane, which has higher surface O/C ratio, than at the freshly cleaved,
hydrophobic “basal” plane [26]. Thus, these results suggest that cytochrome c
electron transfer at an uncharged, hydrophobic surface should be sluggish, if a
response is observed at all.

Herein, we report a somewhat different finding, in that we observe
reversible, stable, diffusion-controlled kinetics at an uncharged, hydrogen-
terminated electrode surface. We observed this behavior as part of our research
work to develop boron-doped diamond (BDD) optically transparent electrodes
(OTES) for use in infrared (IR) spectroelectrochemical studies to probe protein

structure-function relationships. We began our protein study with cytochrome c,

79

as it is a structurally and electrochemically well-characterized heme protein that
serves as a simple model for the more complicated heme-containing systems we
intend to study. One of the first steps in our multi-tasked project was to
characterize the electrochemical kinetics Of cytochrome c at an untreated BDD
electrode. The spectroelectrochemistry of cytochrome c in the UV/vis and IR at
BDD OTES will be described in a later chapter (Sections 7.2.4 and 7.3.5).

BDD is a new electrode material, as discussed in Chapter 3 [27-30]. Good
quality, hydrogen-terminated diamond thin-film electrodes exhibit properties well
suited for protein electrochemical studies such as: (i) a low and stable
background current over a wide potential range and (ii) a resistance to fouling,
due to weak adsorption of polar molecules on the nonpolar surface [31]. This
chapter will discuss investigations of the effects of the different surface
morphology (micro- and nanocrystalline films), the effect of boron dopant level
(1 and 10 ppm films), and the effect of surface pretreatment (acid washing and
rehydrogenation) on cytochrome c electron transfer.

BDD thin-film electrodes can be produced synthetically by microwave-
assisted chemical vapor deposition. Alterations in the growth conditions result in

different BDD thin-film morphologies, namely, micro- and nanocrystalline.
Microcrystalline films are grown using a CH4/H2 (~ 0.5%) source gas mixture,

with growth conditions such that the rate of crystal growth exceeds the rate of

nucleation. This produces large (1 -20 um), well-faceted diamond grains. In

contrast, the nanocrystalline films are grown using a CH4/Ar (~ 1%) source gas

mixture with little or no H2 added [32]. Under these growth conditions, the rate of

80

nucleation generally exceeds the rate of crystal growth, resulting in the formation
of small grains (~ 15 nm). Nanocrystalline films are very smooth, with a
30 - 50 nm surface roughness, and have a higher fraction of grain boundaries
than do the microcrystalline films due to the smaller grain size [33, 34]. The

instantaneous contact angle measured for water is routinely larger than 70.00,

indicating a mildly hydrophobic surface. As deposited, the films are H-terminated
with XPS data indicating surface oxygen levels Of less than 2 atomic percent.

Electrical conductivity is imparted to both types of films by boron doping
(~ 1019 - 1021 B cm'3) and both types of films exhibit excellent electrochemical

responsiveness. We have exploited these properties and demonstrated that
BDD is a viable material for studying bioelectrochemical redox reactions—in
particular, the reduction and oxidation of cytochrome c [35]. Our preliminary
results indicated that quasi-reversible electron transfer could be easily observed
for cytochrome c at BDD thin-film electrodes. We then turned our focus towards
determining the surface properties that would yield optimal and reproducible

heterogeneous electron transfer rates for cytochrome c at these electrodes.

4.2 Results

All electrochemical measurements were performed using a conventional
three-electrode cell, as described in the experimental chapter, Section 2.3.
Details of the BDD electrode fabrication (Sections 2.1.1-2.1.2) and the
cytochrome 0 sample preparation procedure (Section 2.2.2) were also provided

in Chapter 2. The 1 ppm and 10 ppm nomenclature used throughout refers to

81

the growth conditions of BDD thin films, specifically, to the gas phase

concentration (B/C ratio). These values correspond to carrier concentrations of

ca. 1018 B cm'3 (1 ppm) and ca. 1019 B cm'3 (10 ppm).

4.2.1 Microcrystalline BDD Thin-Film Electrodes

The cyclic voltammetric i-E curves measured for cytochrome c at 1 ppm
and 10 ppm microcrystalline BDD thin-film electrodes are shown in Figure 4-2.
Despite the increased carrier concentration present in the 10 ppm film, there is
no observable difference in the cytochrome c electron transfer rate constant at
these electrodes. This was the exception, rather than the general observation for

a variety of redox couples. Comparison measurements using redox analytes like
Fe(CN)53'/4' and Ru(NH3)63+/2+ generally revealed larger cyclic voltammetric AEp

values for the 1 ppm films. We believe this is due to increased film resistance.
This is reflected by the fact that the apparent film resistivities were a factor of 10
larger for the 1 ppm films. These films are called “as deposited” because no
surface pretreatment was applied after film deposition. Later studies revealed
that rate constants measured for the 10 ppm films were indeed larger than those
for 1 ppm films, so the duration of this section describes studies at 10 ppm

microcrystalline BDD electrodes.

82

Figure 4-2 Cyclic Voltammetry of Cytochrome c at “As Deposited” 1 ppm
and 10 ppm Microcrystalline Boron-Doped Diamond Electrodes

Cyclic voltammetric i-E curves for 200 pM cytochrome c with 20 mM NaCl and 1 mM Tris
buffer (pH 7) measured at “as deposited” 1 ppm (dotted line) and 10 ppm (solid line)
microcrystalline BDD electrodes. The potential was initially scanned in the negative

direction from 0.30 to -0. 20 V at 0.010 V/S. Electrode area = 0.2 cm2

 

   

..... 1ppm
—10ppm

 

 

-02 A .01 l 0.0 A 0.1 ‘ 0.2 L 0.3
Potertial (V vs. Ag/AgCI)

 

Cyclic voltammograms (CVs) for cytochrome c measured at four different
10 ppm microcrystalline BDD electrodes, from the same growth batch, are
presented in Figure 4-3. In general, the electrode responsiveness of the “as
deposited” 10 ppm films is very good for the aqueous-based redox test analytes,

3+/2+

Fe(CN)63'/4‘ and Ru(NH3)6 , as the CV peak splitting at 100 mV/s is routinely

75 mV or less. Figure 4-3 demonstrates that the response is sluggish for the
protein. In contrast, Figure 4-4 shows cytochrome c CVs at the same four
electrodes after a surface regeneration procedure. This procedure involves

exposing the electrodes to a 2-step acid washing, followed by rehydrogenation in

83

H2 plasma. The details of the acid washing and rehydrogenation were given in

Section 2.1.1. It is clear that the electron transfer kinetics and reproducibility
greatly improve after the pretreatment. The pretreatment produces a clean,

hydrogen-terminated surface.

Figure 4-3 Cyclic Voltammetry of Cytochrome c at “As Deposited”
Microcrystalline Boron-Doped Diamond Electrodes

Cyclic voltammetric i-E curves for 200 pM cytochrome c with 50 mM NaCI and 1 mM Tris
buffer (pH 7) measured at “as deposited” microcrystalline (10 ppm) BDD electrodes.
Data are presented for four different electrodes from the same batch, as indicated by the
numbers in the figure legend. The potential was initially scanned in the negative

direction from 0. 30 to -0. 20 V at 0.020 V/S. Electrode area = 0.2 cm2.

 

Current (uA)

 

 

 

4L L L

0.2 0.3 0.4

LL L

0. 1
Potential (V vs. Ag/AgCI)

-02 A 0.1 A 00

The deposition Of the microcrystalline films involves placing four of the Si
SLJbStrates in the microwave reactor, all positioned near the center of the
$leStrate stage. The substrate temperature is controlled indirectly by immersion

I“ the plasma. However, with four substrates present, it often results that the

84

temperature of one or more Of the substrates can be inhomogeneous across the
surface. For example, a substrate placed near the edge of the plasma discharge
zone will have the side exposed to the discharge at a higher temperature than
the side further removed from the discharge zone. Such temperature gradients
result in lower boron incorporation and more poorly conducting amorphous
carbon impurity incorporation. This effect is manifested in irreproducible

behavior, as shown in Figure 4-3.

Figure 4-4 Cyclic Voltammetry of Cytochrome c at “Acid-Washed,
Rehydrogenated” Microcrystalline Boron-Doped Diamond Electrodes

Cyclic voltammetric i-E curves for a 200 )JM cytochrome c with 50 mM NaCl and 1 mM
Tris buffer (pH 7) measured at acid-washed, rehydrogenated microcrystalline (10 ppm)
BDD electrodes. Data are presented for four different electrodes, as indicated by the
numbers in the figure legend. The potential was scanned in a negative direction from

0. 30 to —0. 20 V at 0.020 V/s Electrode area = 0.2 cm2.

 

 

 

 

A l A

0.2 0.3 0.4

L

 

-02 A -01 A 0.0 I 0.1
Potential (V vs. Ag/AgCI)

85

Similar improvements were observed for CVs measured using acid
washed, rehydrogenated 1 ppm microcrystalline BDD thin-film electrodes over
those measured at as deposited surfaces (data not shown). However, in
general, the rate constant values for the 10 ppm films were larger than the 1 ppm
films even after pretreatment. The reason for the response improvement is
unclear. It is supposed that the pretreatment removes poorly conducting,
amorphous carbon impurities from the surface, leading to larger electron transfer
rate constants for this surface-sensitive analyte. For all of the remaining data

presented in this section, the electrodes were acid washed and rehydrogenated.

Figure 4-5 Demonstration of the Cyclic Voltammetric Response Stability for
Cytochrome c at a Microcrystalline Boron-Doped Diamond Electrode

Cyclic voltammetric i-E curves for 100 pM solution of cytochrome c containing 50 mM
NaCl and 1 mM Tris at pH 7. The potential was initially scanned negative/y from 0. 30 to
-0. 20 V at 0. 020 V/s. The numbers in the legend represent the scan number. Data are
presented for scans between 1 and 15 at an acid-washed, rehydrogenated
microcrystalline (10 ppm) BDD electrode (referred to as #4 in Figures 4-3 and 4-4).

Electrode area = 0.2 cm .

 

CUTent (tIA)

 

 

 

 

.01 A 0.0 A 0.1 A 0.2 A 0.3 A 0.4
Potential (v vs. Ag/AgCI)

86

The reproducibility and responsiveness of the electrodes was improved
after pretreatment. For the voltammograms presented in Figure 4-4, the applied
potential was swept from 0.30 V to —0.20 V at a scan rate of 0.020 V/s. The
reduction peak at ca. 0.026 V is indicative of the heme Fe(lll) to Fe(ll) transition.
Beyond —0.10 V the reduction current is attributed to the reduction of residual
dissolved oxygen. The catalysis of oxygen reduction at the electrode surface by
metal macrocycles, such as Fe porphyrins, is well known [36]. Furthermore,
experiments involving added oxygen to the solution resulted in similar currents.
On the reverse scan, the peak near 0.133 V is due to the oxidation of heme Fe(ll)
to Fe(lll).

The short-term response stability for cytochrome c at bare electrodes has
been shown to be a sensitive test of protein sample purity [10]. The CV signals
of cytochrome c samples that are not chromatographically purified and, therefore,
contain oligomers, typically decrease over time with increasing scan number.
Niki and coworkers have concluded that a decreased signal amplitude over time
indicates fouling by denatured protein blocking the electroactive area of the
electrode [37]. Therefore, Figure 4-5 demonstrates an important point for

cytochrome c electrochemistry at BDD electrodes, and that is there is no
observed decrease in signal amplitude or increase in AEp over several minutes

with repeated scans.
The protein used in all CV experiments was chromatographically purified.
This suggests that the interaction of cytochrome c with the H-terminated BDD

electrode surface does not result in irreversible adsorption and denaturation of

87

the protein molecules. The fact that the redox potential is in good agreement
with that observed for cytochrome c in solution further supports the interpretation
that the protein is In its native form. Generally, electrode fouling at metal and

metal oxide electrodes is evident within 10 scans at 0.02 V/S (~ 10 min) [2].

Figure 4-6 Cyclic Voltammetric Response for Cytochrome c Measured at a
Microcrystalline Boron-Doped Diamond Electrode

(A) Background response for 50 mM NaCl, 1 mM Tris HCI at 0. 020 V/s. (B) Cyclic
voltammograms for 50 pM horse heart cytochrome c in 50 mM NaCl, 1 mM Tris HCI
pH 7 buffer at an acid-washed, rehydrogenated microcrystalline (10 ppm) BDD electrode
(referred to as #4 in Figures 4-3 and 4-4). Data are presented for scan rates
0. 002 - 0.100 V/s. Electrode area = 0.2 cm2.

 

 

 

 

 

. A
*___.J
l B
A 1.0 -
530.5.
2 ’ A
5 0.0 I
.N
—0.5‘
-1.0 -
-1.5 . L

 

on A 0.0 0.1 A 0.2 A 0.3
Potertial (V vs. Ag/AgCl)

Figure 468 shows a series of cyclic voltammetric i-E curves for
cytochrome c at a 10 ppm microcrystalline BDD electrode at different scan rates
from 0.002 to 0.100 V/s. The solution concentration was 200 [M in 1 mM

Tris HCI buffer (pH 7) containing 50 mM NaCI. A well-defined voltammetric

88

response, characteristic of a diffusion-controlled reaction, is observed in the
potential range of 0.30 to —0.20 V. This response was observed immediately
upon introducing cytochrome c solution into the electrochemical cell with no

conditioning cycles required. The AEp values ranged from 0.058 to 0.091 V over

this scan rate range, indicating quasi-reversible behavior.
The background scan (50 mM NaCI solution) collected after the CV
measurements of cytochrome c shown in Figure 4-6A was featureless,

suggesting no protein adsorption. The apparent heterogeneous rate constant,

k was calculated from the scan rate dependence of the peak splitting via the

0
399'
Nicholson method, using the data presented in Figure 4-6B [38]. The value

obtained from this method, 7.2 x 10'3(: 2) cm/s, compares well with the rate

constant reported for other bare electrodes [24]. The standard reduction
potential, measured, as the average of the anodic and cathodic peak potentials,
remained independent of scan rate over this range. The value of 0.073 V vs.
Ag/AgCl (0.270 V vs. NHE) is in very good agreement (1 10 mV) with that
reported in the literature for cytochrome c in solution at this pH [39].

The anodic and cathodic peak currents are shown to vary linearly with the
square root of the scan rate (Figure 4-7) and the concentration (Figure 4-8), as
expected for a one-electron diffusion-controlled process. The diffusion coefficient

for the oxidized form of cytochrome c was calculated from the slopes of the plots

in Figure 4-8 and found to be .7 x 10’7 cm2/s. The value measured here is in

agreement with that reported by Armstrong and coworkers (5.0x 10‘7 cm2/s)

from CV studies of cytochrome c at the edge plane of pyrolytic graphite (PG)

89

electrodes [26]. These rates are low compared to the solution value of
11 x 10'7 cmzls, owing to the quasi-reversible nature of the cytochrome c electron

transfer reaction at solid electrodes.

Figure 4-7 Peak Current Dependence on the Square Root of the Scan Rate
Measured for Cytochrome c at a Microcrystalline Boron-Doped Diamond
Electrode

Plot of the reduction (squares) and oxidation (circles) peak currents versus the square
root of the scan rate. Data points were obtained from the CVS presented in Figure 4-6.
Data shown for scan rates from 0. 002 to 0.100 V/s. Linear correlation constants = 0. 999
(oxidation) and 0.998 (reduction) and y-intercepts = 0.010 pA (oxidation) and - 0.010 trA

(reduction). Electrode area = 0.2 cm .

0.9

 

 

0.6 -

 

 

 

0.3 -

 

0.0 -

-0.3 -

Peak Current (HA)

-06 _

I

 

 

0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
San Rate"2 ((V/s)‘0)

I l

- J

 

90

Figure 4-8 Peak Current Dependence on the Concentration Measured for
Cytochrome c at a Microcrystalline Boron-Doped Diamond Electrode

Plot of the peak current versus the cytochrome 0 solution (50 mM NaCl and 1 mM Tris at
pH 7) concentration. Data points obtained from cyclic voltammograms measured at
0.020 V/s at an acid-washed, rehydrogenated microcrystalline (10 ppm) BDD electrode
(referred to as #4 in Figures 4-3 and 4-4). Data is shown for concentrations from 25 to
200 pM. Linear correlation constants = 0.999 (oxidation) and 0.994 (reduction) and
y-intercepts = 0.085 pA (oxidation) and - 0.069 pA (reduction). Electrode area =

0.2 cm2.

1.2

 

 

I red
09- . 0x

 

 

 

0.6 -
0.3 -

0.0 -

ak Current (IA)

-0.3 -

Pe
95
a:

-0.9 -

 

 

 

-12 -

0 25 50 75 100 125 150 175 200 225
Concentration (ttIVI)

 

Although no evidence of strong adsorption was Observed for cytochrome c
electron transfer at BDD thin-film electrodes, changes in the electrode surface
termination were evident after exposure to cytochrome c. In particular, the
background CV after exposure to cytochrome 0 showed a marked reduction
current beyond -0.10 V, as is shown in Figure 4-9. As mentioned earlier, surface
Oxygen reduction catalysis by porphyrins is well documented. Additionally,
Similar current profiles were observed during experiments involving added

Oxygen to the electrolyte solution.

91

Figure 4-9 Cyclic Voltammetric Background Curves Measured for a Boron-
Doped Diamond Electrode after Exposure to Cytochrome c

Background cyclic voltammograms of the electrolyte solution (50 mM NaCl and 1 mM
Tris at pH 7) before (solid line) and after (dashed line) cytochrome 0 measurements.
Cyclic voltammograms measured at 0.020 V/s at an acid-washed, rehydrogenated

microcrystalline (10 ppm) BDD electrode.

 

 

 

-0.3 A -02 A .01 A 0.0 A 0.1 A 0.2 A 0.3
Potential (v vs. AglAgCI)

 

The indication of catalytic oxygen reduction coupled with increased
surface N levels, as measured by XPS, after the cytochrome 0 measurements
suggests the presence of heme molecules on the surface. As shown in
Table 4-1, the peak splitting observed for the surface-sensitive test analyte,
Fe(CN)53’/4‘ increased after exposure to the protein solution. Granger et al
demonstrated that the electron transfer kinetics for this analyte at BDD
electrodes is highly sensitive to the surface termination [31]. The lowest peak

splitting, (i.e., fastest rates) were Observed for clean, hydrogen-terminated

92

surfaces. Thus, we would expect to see a decrease in the electron transfer rate

of Fe(CN)53’/4’ if residual heme molecules, or oxygen groups are present at the

2+

electrode surface. In contrast, Ru(NH3)53+/ is not sensitive to the surface

conditions, thus the peak splitting values measured before and after exposure to

the protein solution are relatively constant.

Table 4—1 Effect of Cytochrome c Exposure on Test Analyte Cyclic
Voltammetric Peak Splitting for Acid-Washed Rehydrogenated 10 ppm
Microcrystalline Boron-Doped Diamond Electrodes

 

 

RU(NI‘I3)53+/2+ F8(CN)63A/4A
electrode before after before after
1 69 71 80 129
2 69 70 73 135
3 69 73 74 145
4 67 71 73 151

 

Values measured for 0.1 mM analyte In 1 M KCI at 0.05 V/s

In general, the electrochemical and spectroelectrochemical results
obtained at BDD thin-film electrodes do not indicate denaturation of the
cytochrome c protein upon exposure to this surface. Surface Fe was not
detected by XPS, although if present, the levels may be well below that of the
detection limit. However, it is well known that cytochrome c adsorbs strongly to
glass and can only be removed by acid. The measurements following the
cytochrome c electrochemistry were not performed after washing the glassware

in acid, thus, the possibility for residual protein contaminating these experiments

93

can not be ruled out, at this time. Further experiments to investigate this issue

and the role Of adsorption are required.

4.2.2 Nanocrystalline BDD Thin-Film Electrodes

Figure 4-10 Cyclic Voltammetry of Cytochrome c at “As Deposited” 1 ppm
and 10 ppm Nanocrystalline Boron-Doped Diamond Electrodes

Cyclic voltammetric i—E curves for 100 pM cytochrome c with 50 mM NaCl and 1 mM Tris
buffer (pH 7) measured at “as deposited” 1 ppm (dotted line) and 10 ppm (solid line)
nanocrystalline BDD electrodes. The potential was scanned in a negative direction from

0. 30 to —-0. 20 V at 0.010 V/s. Electrode area = 0.2 cm2.

 

   
  

..... 1ppm
—-10F>Pm

 

oz -0.1 A 0.0 0.1 A 0.2 A 0.3
Potential(V vs. Ag/AgCI)

41%

 

 

 

A comparison of the cyclic voltammetric i-E curves measured for
cytochrome c at 1 ppm and 10 ppm as deposited nanocrystalline BDD thin-film
electrodes is shown in Figure 4-10. In contrast to the cytochrome c voltammetry
observed at the microcrystalline BDD electrodes represented in Figure 4-2, the

rate constant measured for the 10 ppm nanocrystalline film is clearly larger than

94

that for the 1 ppm film. Again, the 10 ppm films have a higher doping level, a
greater electron carrier concentration, and, thus, a higher conductivity. A
comparison of Figures 4-2 and 4-10 suggests that the as deposited
nanocrystalline BDD surface is better suited for promoting relatively rapid
cytochrome c electron transfer kinetics. Owing to the increased rate Observed for

the 10 ppm films, all subsequent studies were carried out using these electrodes.

Figure 4-11 Cyclic Voltammetry of Cytochrome c at “As Deposited”
Nanocrystalline Boron-Doped Diamond Electrodes

Cyclic voltammetric i-E curves for 100 pM cytochrome c with 50 mM NaCl and 1 mM Tris
buffer (pH 7) measured at “as deposited” nanocrystalline (10 ppm) BDD electrodes.
Data presented for three different electrodes, as indicated by the numbers in the figure
legend. The potential was initially scanned in the negative direction from 0. 30 to —0. 20 V

at 0.020 V/S. Electrode area = 0.2 cm2.

 

Current (uA)

 

 

 

 

-0.1 0.0 A 0.1 A 0.2 A 0.3
Potential (V vs. Ag/AgCI)

In the cyclic voltammograms presented in Figures 4-11 through 4-13, the
applied potential was swept from 0.30 V to —0.20 V. Figure 4-11 shows a set of

CVS that were collected using three different, “as deposited,” electrodes. The

95

el

('1‘! ‘4 h

I:

electrodes were grown under identical conditions in three separate runs.
Therefore, the surface temperature gradient problem observed for the
microcrystalline electrodes was nonexistent. The electron transfer kinetics
observed for electrodes 2 and 3 are faster than those observed for electrode 1.
In comparison, the cytochrome c CVs presented in Figure 4-12 were measured
at the same three electrodes after acid-washing and rehydrogenating each
electrode. The pretreatment again improved the reproducibility and decreased

the CV peak splitting. The improvement was most dramatic for electrode 1.

Figure 4-12 Cyclic Voltammetry of Cytochrome c at “Acid-Washed,
Rehydrogenated” Nanocrystalline Boron-Doped Diamond Electrodes

Cyclic voltammetric i-E curves for a 100 pM cytochrome c with 50 mM NaCl and 1 mM
Tris buffer (pH 7) measured at acid-washed, rehydrogenated nanocrystalline (10 ppm)
BDD electrodes. Data presented for three different electrodes, as indicated by the
numbers in the figure legend. The potential was initially scanned in the negative

direction from 0. 30 to —0. 20 V at 0. 020 V/s. Electrode area = 0.2 cm2.

 

A

 

 

 

A

 

01 A 00 0.1 A 0.2 A 0.3
Potential (V vs. Ag/AgCl)

96

I.A
"it.

Successive cytochrome c CV scans are presented in Figure 4-13 for one
of the acid-washed, rehydrogenated nanocrystalline BDD thin-film electrodes,
referred to as electrode 3 in Figures 4-11 and 4-12. As noted above for
microcrystalline BDD thin-film electrodes, Figure 4-13 demonstrates short-term
signal stability over several minutes with repeated scans. Therefore, it is
concluded that like the microcrystalline surface, the nanocrystalline surface
morphology does not promote cytochrome c denaturation, such that response

attenuation or increased peak splitting were observed.

Figure 4-13 Demonstration of the Cyclic Voltammetric Response Stability
for Cytochrome c at a Nanocrystalline Boron-Doped Diamond Electrode

Cyclic voltammetric i-E curves for 100 pM solution of cytochrome 0 containing 50 mM
NaCl and 1 mM Tris at pH 7. The potential was initially scanned negatively from 0. 30 to
-0. 20 V at 0. 020 V/s. The numbers in the legend represent the scan number; data are
presented for scan numbers between 1and 10 at an acid-washed, rehydrogenated
nanocrystalline (10 ppm) BDD electrode (referred to as #3 in Figures 4-10 and 4-11).

Electrode area = 0.2 cm .

 

 

 

 

L

 

0.2 0.3
Potential (V vs. Ag/AgCI)

0.1 A 0.0 A 0.1

97

Figure 4-14 Cyclic Voltammetric Response for Cytochrome c Measured at a
Nanocrystalline Boron-Doped Diamond Electrode

(A) Background CV at 0.020 V/s for a 50 mM NaCl, 1 mM Tris HCI aqueous solution and
(B) CVS for 50 pM horse heart cytochrome c in 50 mM NaCl, 1 mM Tris HCI pH 7 buffer
at a 10 ppm nanocrystalline BDD electrode. Data are presented for scan rates from
0. 002 - 0.100 V/s. Electrode area = 0.2 cm2.

 

 

 

 

 

 

 

 

-0.1 0.0 0.1 0.2 0.3
Potential (V vs. AglAgCl)

Figure 4-14B shows a series of cyclic voltammetric i-E curves for
cytochrome c at a 10 ppm nanocrystalline BDD electrode at different scan rates
from 0.002 to 0.050 V/s. The solution concentration was 50 [M in 1 mM Tris HCI
buffer (pH 7) containing 50 mM NaCl. The CV peak splitting ranged from 0.066
to 0.120 V over this scan rate range, indicating reversible electron transfer
kinetics. A well-defined voltammetric response, characteristic of a diffusion-
controlled reaction, is observed in the potential range from 0.30 to —0.20 V. This
response was observed immediately upon introducing cytochrome 0 solution into

the electrochemical cell, and no time-dependent changes in the response were

98

observed. Although the data presented herein were Obtained at acid-washed,
rehydrogenated electrodes, we have previously demonstrated that it is possible
to measure quasi-reversible kinetics for cytochrome c at as deposited 10 ppm
nanocrystalline electrodes, with no extensive cycling required to activate or
precondition the electrode surface [35].

A background scan for a 50 mM NaCI solution at 0.020 V/s is also shown
in Figure 4-14A. This scan was obtained after the cyclic voltammetric
measurements of cytochrome c were performed. It is important to note that,
even after exposure to the protein solution, the background current profile is
featureless and the current magnitude remains the same as that prior to
exposure to the protein. This suggests that irreversible adsorption of the protein,

either in the active or denatured state, is not occurring to any quantifiable extent.

The apparent heterogeneous rate constant, kgpp, was calculated from the scan

rate dependence of the peak splitting via the Nicholson method, using the data
presented in Figure 4-14B [38]. The value obtained from this method,
4.2 x 10’3 (i 1) cm/s, is in accordance with the rate constant reported for other

bare electrodes [24]. The standard reduction potential, measured, as the
average of the anodic and cathodic peak potentials, remained independent of
scan rate over this range. The value of 0.080 V vs. Ag/AgCI (0.277 V vs. NHE) is
in very good agreement (i 10 mV) with that reported in the literature for

cytochrome c in solution [39].

99

Figure 4-15 Reduction Peak Current Dependence on the Square Root of the
Scan Rate Measured for Cytochrome c at a Nanocrystalline Boron-Doped
Diamond Electrode

Plot of the reduction (squares) and oxidation (circles) peak currents versus the square
root of the scan rate. Data points were obtained from the CVS presented in Figure 4-14.
Data are Shown for scan rates from 0. 002 to 0.100 V/s. Linear correlation constants =
0.997 (oxidation) and 0.999 (reduction) and y-intercepts = 0.011 pA (oxidation) and

- 0.013 pA (reduction). Electrode area = 0.2 cm2.

.0
(0

 

 

,Ired
Cox

 

 

.0
05
r

 

.0
O.)
T

Peak Current 04A)
$5 0
OJ 0

9'3
O)
r

 

 

'09 4 I L L l . l . 1 . L L
0.00 0.05 0.10 0.15 0.20 0.25 0.30 0.35
Scan Rate” ((V/s)"2)

4L

 

The anodic and cathodic peak currents are shown to vary linearly with the
square root of the scan rate (Figure 4-15) and the concentration (Figure 4-16), as
expected for a diffusion-controlled process. The ratio of the reduction and
oxidation peak currents was 1. The diffusion coefficient for oxidized form of

cytochrome c was calculated from the slopes of the plots in Figure 4-16 and

found to be 4.5 x 10'7 cm2/s.

100

Figure 4-16 Reduction Peak Current Dependence on the Concentration
Measured for Cytochrome c at a Nanocrystalline Boron-Doped Diamond
Electrode

Plot of the peak current versus the cytochrome 0 solution (50 mM NaCl and 1 mM Tris at
pH 7) concentration. Data points obtained from cyclic voltammograms measured at
0. 020 V/s for an acid-washed, rehydrogenated nanocrystalline (10 ppm) BDD electrode
(referred to as #3 in Figures 4-11 and 4-12). Data are shown for concentrations from
25 to 200 pM. Linear correlation constants = 0. 999 (oxidation) and 0. 998 (reduction) and
y-inztercepts = 0.045 pA (oxidation) and - 0.057 pA (reduction). Electrode area = 0.2

cm.

1.2

 

 

 

 

0-9 - C ox

 

0.6 b
0.3 -
I
0.0 -
-0.3 -

Peak Cunent

-0.6 b
-0.9 E

l
0 A25 A so A 75 A100A125A150A175A200A225
[wtochromec](uM)

 

 

 

-1.2

As was presented for the microcrystalline case, similar increases in
background current and test analyte peak splitting were observed after the
cytochrome c electrochemistry measured for BDD nanocrystalline thin-film

electrodes. Table 4-2 lists the effects of the protein solution on the CV peak

splitting of Fe(CN)53'/4' and Ru(NH3)63+/2+. As discussed above, protein

contamination can not be ruled out at this time.

101

Table 4—2 Effect of Cytochrome c Exposure on Test Analyte Cyclic
Voltammetric Peak Splitting for Acid-Washed Rehydrogenated 10 ppm
Nanocrystalline Boron-Doped Diamond Electrodes

 

 

Ru(NH3)63*’2” Fe(CN)63"4‘
electrode before after before after
1 72 81 81 168
2 92 141 93 156
3 81 1 19 88 185

 

Values measured for 0.1 mM analyte in 1 M KCI at 0.05 V/s

4.3 Discussion

The increased electrode kinetics for the 10 ppm BDD thin-film electrodes
compared to that of the 1 ppm electrodes is not unexpected. The higher gas
phase diborane concentration leads to a higher film doping level, or more

specifically, a higher carrier concentration. The carrier concentrations at room

3, respectively, for the

temperature have been found to be ~ 1019 and ~1020 B cm'
1 and 10 ppm films. Thus, a higher B doping level yields more conductive films,
less deformation of the cyclic voltammetric AEp values, and higher calculated

electron transfer rate constants.

The improvement in the response observed after acid-washing and
rehydrogenation of both the microcrystalline and nanocrystalline BDD electrodes
suggests that there are low levels of contaminants on the surface that are
removed by the treatment. The identity of the contaminants is unknown, but they

are likely poorly conducting, amorphous nondiamond (spZ-bonded) carbon. The

surface contaminants can increase the ohmic resistance at the electrode surface

102

and/or alter the protein-electrode interaction by increasing the electron tunneling
distance, which decreases the electrode kinetics.

It is important to point out that for some as deposited 10 ppm
nanocrystalline BDD electrodes relatively rapid kinetics were observed, with no
extensive cycling required to activate or precondition the surface, and no need
for extended equilibrium with the protein solution. It has been shown that
reversible kinetics can be achieved for chromatographically purified cytochrome 0
solutions at bare metal electrodes only after pretreatment including polishing
followed by sonication or annealing in a hydrogen flame [24]. Without special
cleaning procedures, quasi-reversible cytochrome c voltammetry has been
measured at polished gold and glassy carbon electrodes only after long
equilibration times [40]. The cleaning procedures seem to serve two purposes:
(i) to remove surface contaminants that may promote protein adsorption and
denaturation, and (ii) to produce a hydrophilic surface [10, 23].

In this work, we focused on two dopant levels, 1 and 10 ppm, which

3

correspond to ~ 1019 and 1020 B cm’ , respectively. At these levels, the BDD thin

films act like a semi-metal rather than a semiconductor, enabling rapid electron
transfer kinetics over a wide (~ 3 V) potential range [27, 41, 42]. Therefore, it is
possible to compare the rates observed for BDD thin-film electrodes to those
measured for other solid metal electrodes, such as glassy carbon (GC), Au, Ag,
and Pt.

Table 4-3 presents a comparison of the heterogeneous electron transfer

rate constants measured by direct voltammetry of cytochrome c. All reports of

103

cytochrome c kinetics demonstrate a strong dependence on cytochromec
concentration and scan rate range, thus these are included in Table 4-3. The
rate constant for the 10 ppm microcrystalline BDD thin-film electrodes increases
by an order of magnitude after acid washing and rehydrogenation. As deposited,
10 ppm nanocrystalline films demonstrate the same rate constants as observed
for the acid-washed and rehydrogenated 10 ppm microcrystalline electrodes.
Thus, there appears to be little influence of the surface morphology and
microstructure on the responsiveness of diamond electrodes as long as the
electrode is conductive and the surface is clean. Upon acid washing and
rehydrogenation, the rate constant for the films shows a marked increase over
that for the 1 ppm films. Finally, it is clear that the rates measured for the best
BDD electrodes (10 ppm acid washed and rehydrogenated micro- and
nanocrystalline BDD) compare well with the observed rates for Ag, Au, and GC.
The values measured for the PG edge plane are among the fastest of all solid
electrodes. Although the rate constants for BDD are somewhat less than those
observed at the hydrophilic edge plane of PG, it is crucial to note that they are
faster than the rate constants observed for the hydrophobic basal plane of PG.
This is surprising, considering that the conventional wisdom suggests a
negatively charged, oxygen-rich, hydrophilic surface is necessary to observe
facile electron transfer kinetics of cytochrome 0. Thus, the results presented
herein may have implications for understanding the nature of the required

surface composition for direct cytochrome c electron transfer.

104

Table 4—3 Apparent Heterogeneous Electron Transfer Rate Constants
Measured from the Direct Voltammetry of Cytochrome c at Solid Electrodes

 

 

 

 

kgpp x 103 [cm] 3:er
Electrode (cm/s) (pM) (mVIs) Source
Au 1.6 (:l: 0.9) 134 4 - 100 [24]
Pt 2.8 (i 0.5) 85 8 - 200 [24]
Ag 1.5 (.t 0.4) 200 1 - 10 [24]
ln02 6.0 52 10 - 500 [22]
GO 6.5 40 1 - 400 [43]
PG (edge) 4.5 150 2 - 100 [26]
PG (basal) 0.25 150 2 — 100 [26]
1 ppm micro BDD (as dep) 0.24 (:t 0.03) 200 2 - 20 this work
1 ppm nano BDD (as dep) 0.23 (.t 0.08) 100 2 - 20 this work
10 ppm micro BDD (as dep) 0.27 (i 0.06) 200 2 - 20 this work
10 ppm nano BDD (as dep) 0.86 (i 0.1) 100 2 - 100 this work
10 ppm nano BDD (as dep) 1.1 (.t 0.2) 200 2 - 50 [35]
10 ppm micro BDD (awrh) 7.2 (i 2) 50 2 - 100 this work
10 ppm micro BDD (awrh) 1.4 (:t 0.3) 200 2 - 100 this work
10 ppm nano BDD (awrh) 4.2 (1: 1) 50 2 - 100 this work
10 ppm nano BDD (awrh) 1.3 (i 0.4) 200 2 - 100 this work

 

cyt c, cytochrome c; anz, indium oxide; GC, glassy carbon; PG, pyrolytic graphite; micro,
microcrystalline; nano, nanocrystalline; as dep, as deposited; awrh, acid-washed and

rehydrogenated.

105

In all cases, both the microcrystalline and nanocrystalline films exhibited a
CV peak current response that increased linearly with (i) the square root of the
scan rate, suggesting that the electrode reaction kinetics are limited by semi-
infinite diffusion of the protein to the interfacial reaction zone, and
(ii) concentration. In the case of cytochrome c heterogeneous electron transfer
at bare electrodes, researchers have found that two parameters are necessary
for measuring facile electron transfer kinetics at carbon, metal, and metal oxide
electrodes: (i) use of chromatographically purified, non-Iyophilized cytochrome c
(i.e., free from oligomeric contaminants) and (ii) electrode treatment to produce a
hydrophilic surface [23, 26]. However, the factors that determine electron
transfer rates at such protein/electrode interfaces are not fully understood.

Comparisons of the rates observed for differently-sized hemeproteins
(cytochrome c [23, 26], cytochrome c3 [44], cytochrome c553 [45], and myoglobin

[46]) suggest that the size of the biological molecule, thus, the distance of closest
approach between the heme edge and the electrode surface, is responsible for
controlling the rate of heterogeneous electron transfer at bare electrodes.

This idea is further supported by direct CV studies of cytochrome c on
SAM/electrode surfaces, which show that rates depend exponentially on the
chain length (longer chain length, slower rates), consistent with a through-bond
tunneling mechanism of electron transfer [47]. It has been proposed that this
state of closest approach is achieved by a combination of electrostatic and

chemical interactions at the electrode/solution interface [10, 23]. Whether the

106

prevailing effects are chemical or electrostatic in nature depends upon the
particular electrode/solution system under consideration.

The fact that facile cytochrome c electron transfer kinetics have been
observed for a number of different electrode surfaces raises questions regarding
the existence of or requirement for a particular electrostatic orientation of
cytochrome c at an electrode. Instead, it has been suggested that the point of
zero charge (pzc) of an electrode in a given solution may be a key parameter for
describing the varying voltammetric responses observed at different metal
electrodes [10, 23]. Ignoring purely chemical interactions, at electrode potentials
negative of the pzc, excess negative surface charge should interact attractively
with the positively-charge exposed heme region of cytochrome 0, thus enhancing
electron transfer. Too strong of an interaction would inhibit electron transfer due
to slow removal of the protein molecules following electron transfer. Excess
positive charge will be present at electrode potentials positive of the pzc, which
should interact repulsively with the heme site, decreasing the electron transfer
rate.

Arguments for dominant electrostatic effects are based on the well-
established idea that electrostatic interactions, via lysine residues on the
cytochrome 0 surface, play an integral role for the binding of cytochrome c to its
natural reaction partners, cytochrome c oxidase, cytochrome c reductase, and
cytochrome c perioxidase [9]. These strategically-placed lysine residues form a
ring-like pattern around the exposed heme edge and match the arrangements of

the negative charges of the partner proteins. Chemical modification or site-

107

directed mutagenesis efforts to neutralize the positions of these lysines showed
substantial reduction in the enzymatic activity of the partner proteins toward
cytochrome c [48-50]. Thus, it is generally accepted that a specific arrangement
of complementary charges in the protein-protein interface is essential for the
proteins in the reactive complex to be properly aligned. This alignment ensures
that the redox centers are oriented in positions most favorable for intermolecular
electron transfer. In addition, studies have shown that the involvement of these
lysine residues in the interactions with cytochrome c oxidase are also important
for inducing the functionally relevant structural changes in cytochrome c [51].
Cytochrome c undergoes the same structural changes when bound to anionic
surfaces exhibiting a largely uniform distribution of negative charges, such as
phospholipid vesicles, micelles, polyanions, and electrodes (bare and chemically
modified) [51-54]. These studies, combined with the findings from studies of
cytochrome c/lipid interactions, clearly suggest that Coulombic forces represent
the dominant mode of interaction between cytochrome c and its reaction partners
as well as with biological membranes. However, there also exists a body of
literature that indicates the necessity of hydrophobic interactions in the formation
of the cytochrome c/cytochrome c oxidase electron transfer complex.
Site-directed mutagenesis studies of Paracoccus denitrificans
cytochrome c oxidase show that the Trp residue at position 121 is crucial for
cytochrome c docking, implying that hydrophobic residues of cytochrome 0 must
also be involved in the complex formation [55]. Hildebrandt’s group has carried

out surface enhanced resonance Raman (SERR) spectroscopy of cytochrome c

108

immobilized on Ag/SAM electrodes in an effort to monitor the heme
conformational changes accompanying the electron transfer. They studied the
electrode response for both carboxyl— [56] and alkane-terminated SAM
electrodes [57]. The SERR results indicate that cytochrome c can be
immobilized onto hydrophobic surfaces and that the observed heme structure, in
this case, is the same as for the electrostatically-bound case [57]. The authors
conclude that the driving force for cytochrome 0 binding is likely the reduction of
the solvent-exposed hydrophobic area.

Additionally, hydrophobic interactions have been deemed important in
computational studies of the cytochrome c/cytochrome c oxidase docking
interface. Combining computational technology with information from static
protein crystallographic structures, Roberts and Pique were able to model the
orientation of this electron transfer complex [13]. Their work resulted in an
optimized configuration that placed a cytochrome c oxidase Trp residue (Trp 143,
Rhodobacter Sphaeroides; Trp 121, Paracoccus denitn'ficans) within 4 A of the
cytochrome c heme edge, suggesting a possible electron transfer pathway,
further supporting the idea that this Trp is the point of electron entry [11, 12, 55].
This interface consisted of a relatively hydrophobic contact region, which, in turn,
was surrounded by hydrophilic, electrostatic interactions between complementary
Lys and Asp residues.

Bowden’s group has shown that cytochrome c electron transfer kinetics

measured at mixed-monolayer SAM/electrodes, consisting of equimolar amounts

of SH(CH2)1OCOOH and SH(CH2)30H, resulted in dramatically increased (e.g.,

109

5-fold for horse heart cytochrome 0) rates over those measured at pure
HS(CH2)1OCOOH SAM/electrodes [20, 47]. The authors concluded that the more

irregular, protein-like surfaces, provided by mixed component SAMs, appear to
be better suited for binding electron transfer proteins with optimal electronic
coupling.

The pzc for BDD is not currently known. BDD is known to possess the
following surface characteristics: (i) hydrogen surface termination,
(ii) hydrophobicity, (iii) absence of negatively charged (acidic) surface oxygen
groups, and (iv) no extended rr-bonded electrons. Therefore, at this point we can
only conclude that interactions other than electrostatic coupling between
positively charged residues on cytochrome c and negatively charged electrode
surface oxygen species are involved in promoting facile electron transfer at BDD
electrode surfaces. Interactions between positively charged cytochrome c and a
negative excess surface charge at BDD cannot be ruled out. Hydrophobic
interactions likely play a key role in the observed voltammetric response for
cytochrome c electron transfer at BDD electrodes.

It seems that cytochrome 0 adsorption cannot be ruled out, because there
is an observed effect of surface structure on the voltammetric response at BDD
thin-film electrodes. However, if adsorption is occurring, it likely involves weakly
bound cytochrome 0 molecules, as rinsing and successive measurement results
in a featureless background signal. This is in contrast to the findings of 82003
and Novak, who present evidence for strongly adsorbed cytochrome c

monolayers being involved in the electron transfer mechanism at glassy carbon

110

and Au electrodes [40]. Accordingly, further studies, namely, potential step
chronocoulometry and long optical pathlength thin-layer spectroelectrochemical
experiments, are required to determine if a similar process occurs at BDD

thin-film electrodes.

4.4 Conclusions

The reported results demonstrate that reversible, diffusion-controlled
electron transfer kinetics for cytochrome c are observed at microcrystalline and
nanocrystalline BDD thin-film electrodes without the requirement of long
equilibration times or potential cycling. The response for both microcrystalline
and nanocrystalline BDD electrodes was shown to improve after acid washing
and rehydrogenation, presumably due to a surface cleaning effect. In both
cases, the H-terminated electrode response is stable over time with no indication
of fouling. This is an interesting result, considering that previous measurements
of cytochrome c electrochemical kinetics at bare and modified electrodes have
indicated the necessity of a hydrophilic, negatively charged, and oxygen-rich
electrode surface. The fastest heterogeneous electron transfer rates were
measured for 10 ppm BDD thin-film electrodes and are in accordance with that
published for other solid electrodes. Further studies are required to determine if
surface adsorption is involved in the heterogeneous electron transfer of
cytochrome c at BDD, and also to investigate the effect of exposed surface

oxygen groups via anodic polarization.

111

4.5

[1]

[2-1

[3]

[4-1

[5-1

[5-1

[7-1

[8-1

[9-1

[10.]

[11.]

References

F. A. Armstrong and G. S. Wilson Electrochim. Acta 2000, 45, 2623-2645.

F. M. Hawkridge and I. Taniguchi In The Porphyrin Handbook; K. M.
Kadish, K. M. Smith and R. Guilard, Eds; Academic Press, Inc., New
York, 2000; Vol. 8, 191-202.

J. D. Burgess and F. M. Hawkridge ln Electroanalytical Methods for
Biological Materials; A. Brajter-Toth and J. O. Chambers, Eds; Marcel
Dekker, New York, 2002, 109-142.

W. M. Clark Oxidation-Reduction Potentials of Organic Systems; Williams
and Watkins Publishing Company, Baltimore, 1960.

G. J. Tilley, R. Camba, B. K. Burgess and F. A. Armstrong Biochem. J.
2001, 360, 717-726.

K. S. Chen, C. A. Bonagura, G. J. Tilley, J. P. McEvoy, Y. S. Jung, F. A.
Armstrong, C. D. Stout and B. K. Burgess Nat. Struct. Biol. 2002, 9, 188-
192.

K. S. Chen, Y. S. Jung, C. A. Bonagura, G. J. Tilley, G. S. Prasad, V.
Sridhar, F. A. Armstrong, C. D. Stout and B. K. Burgess J. Biol. Chem.
2002, 277, 5603-5610.

T. Kono and S. Nakamura Bull. Agr. Chem. Soc. Japan 1958, 22, 399-
403.

G. W. Pettigrew and G. R. Moore Cytochrome c - Biological Aspects;
Springer-Verlag, Berlin, 1987.

F. M. Hawkridge and l. Taniguchi Comments Inorg. Chem. 1995, 17, 163-
187.

Y. J. Zhen, C. W. Hoganson, G. T. Babcock and S. Ferguson-Miller J.
Biol. Chem. 1999, 274, 38032-38041.

112

[12.]

[13.]

[14.]

[15.]

[16.]

[17.]

[18.]

[19.]

[20.]

[21.]

[22.]

[23.]

[24.]

K. F. Wang, Y. J. Zhen, R. Sadoski, S. Grinnell, L. Geren, S. Ferguson-
Miller, B. Durham and F. Millett J. Biol. Chem. 1999, 274, 38042-38050.

V. A. Roberts and M. E. Pique Journal of Biological Chemistry 1999, 274,
38051-38060.

D. Flock and V. Helms Proteins 2002, 47, 75-85.

V. Drosou, B. Reincke, M. Schneider and B. Ludwig Biochemistry 2002,
41, 10629-10634.

K. Niki, K. R. Pressler, J. R. Sprinkle, H. Li and E. Margoliash Russ. J.
Electrochem. 2002, 38, 63-67.

A. El Kasmi, M. C. Leopold, R. Galligan, R. T. Robertson, S. S. Saavedra,
K. El Kacemi and E. F. Bowden Electrochem. Commun. 2002, 4, 177-181.

I. Taniguchi, S. Yoshimoto, M. Yoshida, S. Kobayashi, T. Miyawaki, Y.
Aono, Y. Sunatsuki and H. Taira Electrochim. Acta 2000, 45, 2843-2853.

E. Lojou and P. Bianco J. Electroanal. Chem. 2000, 485, 71-80.

A. El Kasmi, J. M. Wallace, E. F. Bowden, S. M. Binet and R. J.
Linderman J. Am. Chem. Soc. 1998, 120, 225-226.

S. Terrettaz, J. Cheng and C. J. Miller J. Am. Chem. Soc. 1996, 118,
7857-7858.

P. Yeh and T. Kuwana Chem. Lett. 1977, 1145-1148.

E. F. Bowden, F. M. Hawkridge and H. N. Blount J. Electroanal. Chem.
1984, 161, 355-376.

S. C. Sun, D. E. Reed, J. K. Cullison, L. H. Rickard and F. M. Hawkridge
Mikrochim. Acta 1988, 3, 97-104.

113

[25.]

[26.]

[27.]

[28.]

[29.]

[30.]

[31.]

[32.]

[33.]

[34.]

[35.]

[36.]

[37.]

F. A. Armstrong, H. A. 0. Hill, B. N. Oliver and N. J. Walton J. Am. Chem.
Soc. 1984, 106, 921-923.

F. A. Armstrong, H. A. 0. Hill and B. N. Oliver J. Chem. Soc, Chem.
Commun. 1984, 976-977.

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. 2000, 72, 3793-3804.

J. S. Xu, M. C. Granger, Q. Y. Chen, J. W. Strojek, T. E. Lister and G. M.
Swain Anal. Chem. 1997, 69, A591-A597.

G. M. Swain In Thin Film Diamond; C. Nebel and J. Ristein, Eds;
Academic Press, 2002, submitted.

G. M. Swain, A. B. Anderson and J. C. Angus MRS Bulletin 1998, 23, 56-
60.

J. S. Xu, O. Y. Chen and G. M. Swain Anal. Chem. 1998, 70, 3146-3154.

D. M. Gruen Annu. Rev. Mater. Sci. 1999, 29, 211-259.

B. Fausett, M. C. Granger, M. L. Hupert, J. Wang, G. M. Swain and D. M.
Gruen Electroanal. 2000, 12, 7-15.

Q. Y. Chen, D. M. Gruen, A. R. Krauss, T. D. Corrigan, M. Witek and G.
M. Swain J. Electrochem. Soc. 2001, 148, E44-E51.

S. Haymond, G. T. Babcock and G. M. Swain J. Am. Chem. Soc. 2002,
124, 10634-10635.

Y. 0. Su, T. Kuwana and S. M. Chen J. Electroanal. Chem. 1990, 288,
177-195.

T. Sagara, K. Niwa, A. Sone, C. Hinnen and K. Niki Langmuir1990, 6,
254-262.

114

[38.]

[39.]

[40.]

[41.]

[42.]

[43.]

[44.]

[45.]

[46.]

[47.]

[48.]

[49.]

[50.]

[51.]

R. S. Nicholson Anal. Chem. 1965, 37, 1351.

F. M. Hawkridge and T. Kuwana Anal. Chem. 1973, 45, 1021-1027.

A. Sziics and M. Novak J. Electroanal. Chem. 1995, 383, 75-84.

M. Hupert, A. Muck, J. Wang, J. Stotter, Z. Cvackova, S. Haymond, Y.
Show and G. M. Swain Diam. Relat. Mat. 2002, in press.

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 Materials 2002, 8, 171-188.

W. R. Hagen Eur. J. Biochem. 1989, 182, 523-530.

W. F. Sokol, D. H. Evans, K. Niki and T. Yagi J. Electroanal. Chem. 1980,
108, 107-115.

K. B. Koller, F. M. Hawkridge, G. Faque and J. LeGall Biochem. Biophys.
Res. Commun. 1987, 145, 619-625.

B. C. King and F. M. Hawkridge J. Electroanal. Chem. 1987, 237, 81 -85.

8. Song, R. A. Clark, E. F. Bowden and M. J. Tarlov J. Phys. Chem. 1993,
97, 6564-6572.

S. H. Speck, S. Ferguson-Miller, N. Osheroff and E. Margoliash Proc. Natl.
Acad. Sci. U. S. A. 1979, 76, 55-59.

H. T. Smith, N. Staudenmayer and F. Millett Biochemistry 1977, 16, 4971-
4974.

R. Rieder and H. R. Bosshard J. Biol. Chem. 1980, 255, 4732-4739.

H. Wackerbarth, D. H. Murgida, S. Oellerich, S. Dopner, L. Rivas and P.
Hildebrandt J. Mol. Struct. 2001, 563, 51-59.

115

[52.] S. Dopner, P. Hildebrandt, F. I. Rosell, A. G. Mauk, M. von Walter, G.
Buse and T. Soulimane Eur. J. Biochem. 1999, 261, 379-391.

[53.] P. Hildebrandt Biochim. Biophys. Acta1990, 1040, 175-186.

[54.] J. Peschke and H. Mohwald Colloids and Surfaces 1987, 27, 305-323.

[55.] H. Witt, F. Malatesta, F. Nicoletti, M. Brunori and B. Ludwig J. Biol. Chem.
1998, 273, 5132-5136.

[56.] D. H. Murgida and P. Hildebrandt J. Phys. Chem. B 2001, 105, 1578-
1586.

[57.] L. Rivas, D. H. Murgida and P. Hildebrandt J. Phys. Chem. B 2002, 106,
4823-4830.

116

CHAPTER 5

5 Low-Frequency Electrochemical Difference
FTIR Spectroscopy of Metalloproteins:
Challenges and Solutions

5.1 Introduction

Vibrational spectroscopy is a vital tool in the process of elucidating
structural or mechanistic information about metalloproteins. Of particular interest

is information about bonding interactions between the metal site and its ligands,
which is found in the low-frequency (<1000 cm'1) region of the infrared (IR)

spectrum. Information on metal-ligand vibrations provides insight into the
molecular structure and reaction intermediates of the enzymatic mechanism.
Examples of structures in metalloenzymes that would give rise to modes in the
low-frequency region of the spectrum are shown in Table 5-1. Identification of
low-frequency modes would allow specific questions regarding the detailed
chemistry of the mechanism to be addressed.

Until recently, experimental access to low-frequency vibrations of
metalloproteins has been limited to resonance Raman spectroscopy. This
technique is useful for studying many biological systems, especially those
containing highly absorbing chromophores. However, the technique is not useful
for studying all metal sites in metalloproteins, particularly those without a Raman-
excitable pigment, or those with multiple chromophores. The alternate technique

for obtaining vibrational information is IR spectroscopy. This technique provides

117

information on all lR-active vibrational modes of a protein. According to theory, a
nonlinear molecule, with N atoms, will exhibit 3N - 6 modes. Considering that
even a small (~ 10,000 Da) protein has 1000’s of atoms, the IR spectrum of such
a molecule is information rich. For example, the bacterial form of cytochrome c
oxidase (Rhodobacter Sphaeroides), a 125,000 Da protein, is comprised of
90,000 atoms. The IR spectrum of this protein then contains over 250,000
vibrational modes.

Therefore, difficulty arises not in the measurement of protein spectra, but
in the detection of the small amplitude signals arising from the active site
structural changes. The detection of such single modes is difficult because they
are superimposed on a large background due to water and global protein mode
absorbance. To achieve adequate levels of signal-to-noise ratio, accurate
background subtraction, stable background absorbance, minimization of the
background absorbance, and maximization of the signal absorbance are
required.

In this chapter, methods are described, in which reaction-induced
difference infrared spectroscopy (i.e., electrochemical difference FTIR

spectroscopy) is used. Accurate solvent subtraction is achieved, resulting in
peak-to-peak noise levels of 105-10'6 absorbance units, thus, rendering the

acquisition of low-frequency vibrational information possible. In addition to the
problem of the large water absorbance, other technical difficulties associated with

accessing the low-frequency region of the spectrum will be discussed.

118

5.2 Structural Information Possible (Far-IR Region)

The acquisition of low-frequency infrared spectra of proteins is a relatively

new field of IR spectrosc0pic research. While there has been much work
performed to assign protein vibrational modes in the mid-IR region (> 1000 cm'1),

there is little information to use as a basis for the assignment of metal-ligand
modes in the low-frequency region. Predictions about several possible
vibrational modes of polypeptides have been made based upon studies of model
peptides, such as N-methyl acetamide [1]. Table 5-1 summarizes these modes.
This table predicts that the protein secondary structure should display several
low-frequency vibrations, arising from primarily out-of-plane peptide backbone
vibrations. More interesting information, necessary for understanding
mechanistic details, lies in those modes that originate not from the backbone, but
from the metal-ligand interactions of the protein active site. Table 5-1 lists a
number of metal-ligand vibrations from synthetic model compounds and proteins,
such as heme-ligand complexes and the photosystem M protein (PSII). Several
of these have been detected by FTIR, whereas others are reported from
resonance Raman studies [2-4].

Assignment of vibrational modes is necessary to extract structural data
from an information-rich protein spectrum. Isotope exchange plays a critical role
in the assignment of vibrational modes. This is especially important when there
are many peaks in a spectrum, and it is desired to assign one to a specific
molecular species. Substitution of a heavier atom will affect the frequency of the

mode according to the following relation,

119

where Vis the frequency of the mode, [1 is the reduced mass, k is the force
constant and c is the speed of light. From this equation, it is clear that
introducing a heavier atom will result in a decrease in the frequency of the
affected mode.

Isotope exchange can be accomplished by solvent exchange (e.g., D20
buffers vs. H20 buffers), metal exchange (e.g., 63Cu vs. SSCu), atom exchange
(e.g., 15M vs. 14N or 180 vs. 160) and specifically labeled amino acids

(e.g., 15N His, 13C Ala). Metal-oxygen species, such as those proposed to exist

in the catalytic cycles of cytochrome c oxidase (cco) and PS“, will exhibit infrared
absorptions sensitive to isotope exchange. Site-directed mutagenesis is another
important tool in the mode assignment of protein spectra. Identification of the
mid-IR mode due to a mechanistically important amino acid residue in cco, Glu
286, has been achieved through experiments involving site-directed mutants of
this residue [5]. It is clear from Table 5-1 that a method, in which low-frequency
modes of metalloenzymes can be measured, will be integral in elucidating

mechanistic details of these proteins.

120

Table 5—1 Low-Frequency Vibrational Modes of Metalloenzymes

 

 

 

 

 

Vibrational

Frequency
Structure (cm'1) Reference
Protein Backbone Modes
Amide V (NHob, CN() 725 [6, 7]
Amide IV (Cob, CCS, CNCd) 625 [6, 7]
Amide VI (COob, CN[) 600 [6, 7]
Amide Vll (NHob, CNt, com) 200 [6, 7]
Metal-Ligand Modes
Cu—OEC 400 [8-10]
Cu—N(His) 260 [11, 12]
Cu—OH 550-500 [13]
CU-OHz 440 [14]
Fe—O(Tyr) 600 [1 5]
Fe—N(His) 220-250 [16, 17]
Fe—OEC 475-575 [8-10]
Fe—OH 500 [13]
Fe—OH2 390 [14]
Fe=O 775-850 [18-23]
Mn—OH 400-500 [13]
Mn—OH2 395 [14]
Mn=O 750 [24]

 

s, stretch; d, deformation; t, torsion; ib, in-plane-bend; ob, out-of-plane-bend; His, Histidine;
Tyr, Tyrosine

121

5.3 FTIR of Aqueous Samples

Biology uses water as the solvent for a myriad of biochemical reactions.
However, water does pose certain difficulties for obtaining protein IR spectra.
The vibrational modes of water give rise to very intense absorption throughout
regions of the mid-IR and they overlap with many protein modes of interest.

Rotational and vibrational modes of water also effectively obscure the low-
frequency IR region (below 1000 cm'1). Figure 5-1 demonstrates the four
regions of spectral influence by water. A series of O-H stretching modes at 3920,

3490 and 3280 cm’1 give rise to an intense and broad absorbance band centered
at 3400 cm'1. The peak at 1645 cm"1 arises from the H-O-H bending mode and
the absorbance at 2125 cm'1 is attributed to the association of hydrogen bonds

The broad feature below 1000 cm'1 is due to the strongly coupled lattice

vibrations of aggregates of water molecules [1, 25]. In order to observe
vibrational modes that are obscured by the water absorbance, ways must be
found to either minimize this background absorbance or to stabilize it so that it
can be accurately subtracted out of the spectrum. In particular, methodologies
are discussed, through which one can observe high-quality spectra for

metalloenzymes in the mid-IR and possibly, the far-IR regions of the spectrum.

122

Figure 5-1 Infrared Absorption Spectra of H20 and D20
A water spectrum (solid line) and a deuterium oxide (99.96% D) spectrum (dashed line)

shown for comparison.

The spectra were acquired at 15 °C using AgBr windows

separated by a 6 pm spacer. Each spectrum is an average of 32 scans collected versus
an air background using a narrow band MCT detector.

Absorbance (Abs. Units)

 

.N
O
r

.3
01
r

_L
O
r

.0
(11
I

 

*Mww-w: "1

  

a"

 

. I ‘
-. ‘T' R...‘ .

 

 

4000 3500 3000‘2500‘2000

Frequency (cm'1)

Table 5—2 Infrared Properties of H20 and DZO

 

 

Mode H20 (cm'T £(M'1cm'1) ozo (cm'1) e(M“cm")
O-X stretch 3920 0.83 2900 0.60
3490 62.7 2540 59.8
3280 54.5 2450 55.2
Association 2125 3.23 1555 1.74
X-O-X Bending 1645 20.8 1215 16.1
Libration Broad absorbance below 1000 cm'1

 

123

Figure 5-2 Comparison of the H20 Absorption Spectrum and the Noise
Spectrum for an Aqueous Sample
Comparison of (A) a water absorbance spectrum and (B) a resulting noise spectrum.

The noise spectrum was obtained by subtracting two successive spectra. Spectra
collected using a MCT detector, AgBr windows, a 6 pm pathlength, and averaging 32

scans at 15 OC. Note the different absorbance scales.

 

 

4000'35'00'3000'2500 2000'15'00'1000'

Frequency (cm.1)

The vibrational modes associated with the prominent water absorbances
are well characterized [1, 25]. A list of these vibrational modes and
corresponding extinction coefficients are presented in Table 5-2. In general, in

order for a solvent spectrum to be accurately subtracted from a protein sample,

124

the overall absorbance of a given peak must be less than one absorbance unit

[26]. Therefore, the pathlength of an aqueous solution should not exceed 10 um,

based on the extinction coefficient of 20.8 (M'1 cm'1) for the water bending mode

at 1645 cm'1. Protein mid-IR spectra, showing high-quality mathematical

subtraction of the overlapping water signal, are routinely acquired by using highly
concentrated samples and optical pathlengths of 10 pm or less [27]. Highly
concentrated protein samples serve to increase the signal amplitude and
decrease the effective water concentration. While this method is successful for

the mid-IR region, it is not applicable to measurements in the low-frequency
region. The broad water lattice vibrations below 1000 cm"1 are extremely

sensitive to minute changes in pathlength or sampling conditions (i.e.,
temperature, ionic strength) [25]. Large spectral artifacts remain even after
careful subtraction when measured at pathlengths as short as 6 pm, as shown in
Figure 5-2. This Figure illustrates that regions of prominent water absorbance
correspond to regions of increased noise.

Reaction-induced difference infrared spectroscopy is an essential
technique for the precise subtraction of solvent peaks, such as those due to
water, especially in protein samples. The principle of this method involves the
subtraction of a protein absorbance spectrum in state A from that of the protein in
state B. The resulting difference spectrum (B minus A) reports on the small and
subtle vibrational changes occurring in the protein as it undergoes the chemical
reaction. A trigger initiates a reaction to advance the protein from state A to

state B. Several triggers are applicable, including intrinsic photochemical

125

reactions, electron transfer reactions, and indirect photochemical reactions [28,
29].

Reaction-induced difference spectroscopy is very powerful in that it
introduces selectivity by subtracting out the background absorbance due to
functional groups that are not involved in or influenced by the chemical reaction.
It also allows the issue of solvent subtraction to be finessed, as no physical
perturbation of the sample is involved, resulting in a constant optical pathlength
and no gross perturbation of the solvent concentration or ionic strength. Thus,
the relatively large baseline distortions that are common in solvent subtracted
data are eliminated.

Accurate background subtraction results in spectra without baseline
distortions, which greatly improves the signal-to—noise ratio and enables the
detection of weakly intense signals obscured by the background absorbance.
Global protein structural modes, particularly the so—called Amide l mode,
commonly exhibit infrared absorption approaching one absorbance unit in
magnitude. However, the reaction of interest will result in only slight
perturbations, affecting a few amino acids or just the active site. These single
mode changes, arising from the reaction-induced perturbation, will constitute a

small fraction of a percent of the overall signal. These single modes have
intensities of approximately 10'4 absorbance units, and are commonly detected

using this methodology, with sufficient signal-to-noise ratios. It is these single
mode signals, which may reflect changes in the enzyme structure that are

directly related to the catalytic cycle we are seeking to acquire.

126

Reaction-induced difference FTIR spectroscopy has been successfully
applied using a direct photochemical approach for light-activated photosynthetic
proteins in the mid- and far-IR [2, 3, 30, 31]. Generalizing the method by using
electrochemical reactions instead of actinic light reactions will enable access to a
wider class of metalloenzymes. Therefore, any metalloprotein with a redox-
active metal would be responsive to this technique. This electrochemically-
initiated method has proven successful for mid-IR measurements of several
proteins, from simplistic, single redox center proteins (e.g., cytochrome c,
cytochrome f) [32] to more complex enzymes, which undergo multiple electron
transfer steps at several redox metal centers (e.g., cytochrome c oxidase) [33]. It
has not, as yet, been utilized in far-IR measurements, owing to a number of
complications, as discussed in this chapter. The major drawback of using the
electrochemicalIy-initiated method in the transmission mode is the requirement
for a supporting electrolyte solution. Therefore, the measurements suffer from
water absorption interference in the low-frequency IR region.

This reaction modulated difference technique provides a very detailed look
at changes that occur as an enzyme system undergoes its catalytic cycle.
Figure 5-3 shows a theoretical construction of such a difference spectrum. A
possible step in the catalytic cycle of cytochrome c oxidase, the mitochondrial
respiratory enzyme that is responsible for the four-electron reduction of oxygen to
water, is shown in Figure 5-3. The Figure depicts the affect of the copper
oxidation state on the Cu-N(His) stretching vibration. In the initial state

(Figure 5-3A), the copper ion is in the oxidized (+2) state and a positive peak

127

attributed to the metal-nitrogen vibrational mode is shown in spectrum 5-3A. The
electron transfer reaction is initiated and the copper is reduced (+1), resulting in a
downshift in the frequency of the Cu-N(His) mode (Figure 5-38). Two other
peaks are present at the same position in both states of the enzyme.
Figure 5-3C shows the reduced minus oxidized electrochemical difference
spectrum, which is the result of subtracting the initial (oxidized) spectrum from
that of the final (reduced). This spectrum reveals a derivative feature, consisting
of both positive and negative bands, which are characteristic of a difference
spectrum. The positive modes are indicative of vibrational modes present in the
final (reduced) state and the negative bands represent modes for the initial
(oxidized) state. The peaks that were unaffected by the oxidation state change
of the metal are subtracted out, leaving a flat baseline feature in the difference
spectrum. In theory, modes due to ligands (e.g., amino acids and substrate
molecules) in the vicinity of the active site and those coordinated to the active
site will undergo a frequency shift upon a change in the oxidation state of the
metal. Any vibrational modes that are unaffected by the reaction will display
exactly the same vibrational frequency in the initial and final spectra. This
includes modes that are not redox-sensitive, or are distant from the active site
where the chemistry is occurring. These modes will be subtracted out and,
therefore, do not contribute to the difference spectrum. Any mode that is slightly
perturbed (e.g., an amino acid residue undergoing a H-bonding change, a
solvent molecule that is in the active site, or a subtle change in the composition

of the vibrational mode) by the reaction will appear as difference peaks.

128

Figure 5-3 Theoretical Construction of an Electrochemical Difference FTIR
Spectrum

Schematic showing the components of an electrochemical FTIR difference spectrum.
(A) The absolute (solvent subtracted) spectrum for the initial (oxidized) form. (8) The
absolute (solvent subtracted) spectrum for the final (reduced) form after an electron
transfer step. (C) The resultant difference spectrum generated by subtracting the initial
from the final spectrum (reduced minus oxidized). Peaks have different amplitudes due
to differences in extinction coefficient.

 

 

-1
< Frequency (cm )

 

The importance of this type of difference spectrum lies in its ability to
provide evidence that could validate (or invalidate) a proposed model of the
enzyme’s catalytic cycle. Using the example in Figure 5-3, it is has been
suggested that one of the histidine ligands on the Cu is involved in the proton

pumping process—an important step in the enzymatic reaction of cytochrome c

oxidase [34, 35]. This particular copper ion, CUB, is “invisible” to other types of

129

spectroscopy, such as EPR and Resonance Raman, due to interference by the
surrounding hemes. Therefore, low-frequency data (for this metal center)
obtained by IR would provide valuable information regarding mechanistic

questions involving the role of Cue in the catalytic cycle.

Reaction-induced difference spectroscopy is a useful method for
simultaneously introducing selectivity and minimizing solvent contribution to
complex protein infrared spectra. However, in order to obtain an accurate
solvent subtraction, it is necessary to use sample absorbances that do not
exceed the limitation for a linear detector response. Systematic errors will be
present in difference spectra acquired under conditions where the detector
response is unreliable. As already mentioned, the use of short pathlengths can
ensure that fully hydrated samples do not exceed the absorbance limitations of
the instrumentation. This method has worked well for studies of highly
concentrated protein samples in the mid-IR region [32, 36-40]. However,

additional considerations are required for handling water absorbance below
1000 cm“.

A commonly applied approach for circumventing problems associated with
the presence of strongly absorbing water bands in the mid-IR region of the
spectrum involves substituting deuterium oxide for water. This approach is
particularly suited for cases in which the observation of sample absorption is
compromised by overlapping solvent absorption. Studies of hydrated lipid
dispersions [41], as well as those aimed at determining protein secondary

structure [42], have been successfully performed using this method. As shown in

130

Figure 5-1, substituting DZO for H20 will shift the solvent absorption to lower

frequencies, thereby enabling observation of the underlying sample bands. This
approach is not entirely free of potential problems, as shifting the solvent
absorption to lower frequencies may compromise the detection of other sample
absorption bands of interest. However, access to the low-frequency region can

be considerably extended through the use of deuterated buffers. Examination

and comparison of spectra acquired in H20 and D20 would enable one to access

the entire spectrum of an aqueous protein sample.

Acquiring spectra in deuterated solvents can pose a more serious
problem—the H/D exchange of protons within the protein. This will result in loss
of the absorption band of a protonated species and its replacement with an
absorption band of the deuterated species. The latter band will likely be at a
lower frequency, and in some cases, not observable (i.e., obscured by other
modes). If the H/D exchange rates are slow, the disappearance of the Amide l
band and subsequent replacement by its deuterated counterpart (Amide I‘) can
be exploited to monitor the rates and extent of exchange. As will be discussed in
Chapter 6, such information provides insight into the exposure of exchangeable
proton sites to the bulk solvent. Although H/D exchange may introduce further
complexities to the protein difference spectrum, major problems in mode
assignment can be overcome by appropriately considering such occurrences

when interpreting the data.

131

5.4 Detectors

Detector choice is a critical factor in determining the quality of FTIR
spectra of biological systems, as IR detectors have different sensitivities, linear
dynamic ranges, and response times. This is particularly true for measurements
made in the low-frequency region because of the inherently low signal-to-noise
levels. The drop in signal-to-noise ratio observed with decreasing frequency is
due to decreased light intensity from the source in this region, and smaller
vibrational mode extinction coefficients. Absorption of low-frequency IR radiation
by optical materials also contributes to the decreased throughput observed in the
far-IR region. The useful range and sensitivity of the IR detector greatly affects
the signal-to-noise ratio.

There are many types of lR-sensitive detectors available for use, however,
the two most common are mercury-cadmium-telluride (MCT) and deuterated
triglycine sulfate (DTGS). These detectors differ in their sensitivity, but are
similar in that they enable signal collection across a broad region of the infrared
spectrum (i.e., mid-IR with some access into the low-frequency region). In
general, the MCT is 5 to 50 times more sensitive than the DTGS detector,
depending on the bandwidth (accessible range) of the MCT detector [43, 44].
Another type of infrared-sensitive detector, the liquid helium-cooled bolometer, is
superior to the MCT for low-frequency IR detection [44].

Infrared detectors are evaluated in terms of their noise-equivalent power

(NEP) and compared using their specific detectivity, or D* value [26, 43, 44]. The

132

sensitivity of an infrared detector is usually expressed in terms of the

NEP (w Hz‘”),

NEP = 117—
R"

In this expression, Vn is the root mean square (rms) of the detector voltage in
V Hz‘y’, and RV is the voltage responsivity of the detector to signals in V W'1 [26].

The specific detectivity, D* (cm szz W4), of a detector is related to the NEP by

In
*: (1‘1/)) -

NEP
where AD is the detector element area. The D* value is convenient for
comparing the sensitivity of different types of infrared detectors, as to a first
approximation; it is independent of the area of the element. When comparing
infrared detectors, a higher D* value is indicative of a more sensitive detector.
The D* expression is related to the overall signal to noise ratio (SNR) of the FTIR

instrument by the following

U,—,(T)-®-M-t“2 -.§
NEP

SEN’R :

 

U,—,(T).o-.\v.i"'2-§-D*
A52

 

U ,7 (T) represents the spectral energy density, O is the optical throughput of the

system, 017 is the spectral resolution, t is the acquisition time, and § is the
efficiency factor of the instrument [26]. From this expression, it is clear that the

signal-to-noise ratio of a given spectrum is linearly dependent upon the D* value

133

of the detector, the resolution, and the throughput. Therefore, it is integral to
optimize these values in order to acquire the highest quality spectra in the
shortest amount of time.

The choice of detector will affect the type of FTIR measurements
undertaken. Thermal type detectors are the most common, as they have the
most linear response function. These detectors operate by sensing the change
in temperature of an absorbing material as it experiences a change in the
infrared radiation striking it. The pyroelectric bolometer is the most common
class of thermal-type detectors. The sensing element in a pyroelectric bolometer
is a material that changes its electrical resistance when its temperature changes.
The most commonly used material is DTGS. DTGS detectors have the
advantage that they can be operated at room temperature while displaying highly
linear responses for varying intensities of incident IR light. The main drawback of

this detector, which renders it unsuitable for biological spectroscopy, is its
relatively poor sensitivity (0* = ~ 107 - 108 ). This level of sensitivity makes the

detection of minute signals, obscured by large background signals, very difficult.
The advent of semiconductor-type quantum detectors over the last decade

has revolutionized the field of biological infrared spectroscopy. This class of

highly sensitive detectors has made it possible for investigators to detect small

amplitude signals in highly absorbing samples. The MCT detectors are the most
common of this class. The D* values for MCT detectors range from 109 - 1011,

and higher. The principle of operation is based on a ternary semiconductor

detector element whose bandgap corresponds to the energies present in IR

134

radiation. These detectors must be cooled, usually to liquid nitrogen temperature
(77 K), to suppress the contributions of thermally excited electrons to the
detected signal. The specific stoichiometric ratio of the constituent elements
determines the useful range of these detectors. Although, an MCT detector can

be optimized for response at low-frequency wavelengths (peak response at
~ 500 cm'1) the D* value is approximately an order of magnitude less than one

optimized for the mid-IR. MCT detectors are often classified as high-sensitivity
(narrow-band) or low sensitivity (broadband, or wide-band) detectors. Therefore,
it follows that an MCT with an intermediate bandwidth will possess an
intermediate relative detectivity [26].

An alternate choice over MCT detectors for low-frequency IR detection is
the liquid helium-cooled bolometer. This quantum detector utilizes a material,
such as silicon or germanium, as the detector element, and operates under the
same technology as the MCT detector. However, in this type of detector, the
detector element temperature must be maintained below that attainable by liquid
nitrogen cooling. Therefore, liquid helium is used, necessitating the need for
liquid helium cryostat hardware. The additional complexity associated with the
cryostat is the key difference that results in a distinct cost differential between

this type of detector and the simpler nitrogen-cooled MCT. Helium-cooled
bolometers have D* values larger than MCT detectors (D* ~1012 -1014), and

their response to radiation in the far-IR exceeds that of any MCT [26, 44]. Chu
and coworkers have utilized a helium-cooled Si bolometer for low-frequency

FTIR measurements of the Photosystem Il protein to detect single modes due to

135

the active site metal-ligand interactions [3, 45]. Their results demonstrate the
potential utility of helium-cooled bolometers in measurement of high quality, low-

frequency FTIR spectra of biological samples.

5.5 Window Materials

One difficulty encountered with transmission spectroscopy involves the
bandpass characteristics of the window materials. The material must be
transparent to the incident IR radiation, and must also possess a high enough
optical throughput to achieve sufficient signal-to-noise in the resulting spectra.
Numerous materials are available for use as IR windows; however, the number

of these substantially decreases when one is interested in utilizing aqueous
samples and observing the low-frequency region (< 1000 cm'1) of the spectrum.

Additionally, due to the large amount of useful information in the mid-IR region of
a protein difference spectrum, it would be beneficial to use a material that has
sufficient optical throughput from 2000 crn’1 down to less than 400 cm'1, ensuring
that all of the vibrational modes of interest are observable.

Table 5-3 lists common IR window materials and corresponding
frequencies of transmission, indices of refraction, and water solubility. There are
several materials that transmit into the low-frequency region and are insoluble in

water. The most commonly used are the silver halide salts. Silver bromide and
silver chloride transmit from the visible region to 285 and 435 cm'1, respectively.

Both have a tendency to physically distort (cold flow), and prolonged exposure to

light leads to photoreduction of the window (e.g., Ag+ to Ag) and gradual

136

darkening, resulting in decreased transmission properties over time. The silver
halide salts have been successfully used in light-induced difference FTIR
experiments of PSII [2, 3, 30, 31]. However, in tests of the electrochemical
difference method, they were found to undergo electrochemical attack by the
gold contact wire and gold mesh used as electrodes in this setup. Zinc selenide
and cadmium telluride have not seen widespread use for biological FTIR
experiments due to their higher cost, and the availability of cheaper materials that
exhibit similar optical properties. Germanium is not well suited for transmission

experiments, due to its high refractive index. An added disadvantage is its low
frequency cutoff of 600 cm‘1, as many metal-ligand vibrations fall below this

cutoff. Silicon is most often used as a reflection element, owing to its high
refractive index. Transmission measurements are possible using thin windows of
Si; however, transmission levels are relatively low due to high reflection losses.
Polyethylene suffers from a limited transmission range, but is a good option for
studies only concerned with low-frequency vibrations. Of all the materials listed,
diamond has the widest optical window, transmitting from the UV down to the far-
IR region. This material possesses added advantages of being mechanically
strong, chemically inert, and when doped with boron, useful as an optically
transparent electrode. One of the most significant improvements over the light-
induced method is the incorporation of diamond windows into our design. With
these windows in place, the limitations on the signal-to—noise ratio due to light

throughput are eliminated.

137

Table 5—3 Properties of Infrared Transmissive Materials

 

 

Material Transmission Refractive Water Solubility
Range (cm'1) Index (at 10W“) (gl100 9 H20)
NaCI 40,000 - 625 1.49 Soluble
KBr 40,000 - 400 1.52 Soluble
Csl 10,000 - 200 1.74 Soluble
AgCI 25,000 - 435 1.98 0.00015
AgBr 20,000 - 285 2.2 0.000012
ZnSe 10,000 - 500 2.4 Insoluble
CdTe 5,000 - 320 2.67 Insoluble
Ge 5,000 - 600 4.0 Insoluble
Si 10,000 — 1540 3.4 Insoluble
500 - 30
Polyethylene 625 - 30 1.54 Insoluble
Diamond 40,000 - 33 2.42 Insoluble

 

5.6 Temperature Control

In the case of reaction-induced difference FTIR experiments, where one is
requiring very stable background subtraction, temperature control is a crucial
element in the experimental setup. Fluctuations in sample temperature result in
large baseline distortions, which map to regions of strong absorption in the
infrared spectra of aqueous systems [25]. Figure 5-4 displays this effect on the

baseline, as a series of noise spectra are shown for temperature drift from 5 to

0.005 °C. The spectra were collected for an aqueous sample with a 10 pm

138

pathlength at different temperatures. To acquire such a noise spectrum, a

spectrum was measured at one temperature, and after decreasing the

temperature a second spectrum was collected (i.e., Ti = 15 °C, Tf = 10 °C, so

AT: 5°C). The difference of these two successive spectra yields a noise

spectrum. It is clear that only with precisely controlled temperature does the
baseline become flat and featureless, as the spectra for the more extreme
temperature drifts contain peak-like features. The appearance of these “thermal
bands” can easily be mistaken for real peaks and have undoubtedly contributed

to many imagined peaks in the published literature.

Figure 5-4 Effect of Temperature Drift on Baseline Linearity

A series of noise spectra acquired by subtracting successive spectra at different
temperatures. The presented spectra are representative of the following temperature
drifts (0 C): (A) 5, (B) 0. 5, (C) 0.05, and (D) 0. 005. Spectra were collected for a 10 pm
thick aqueous solution pressed between two BaFg windows. Each spectrum is an
average of 200 scans and the final temperature was set to 10 DC in all cases.

 

0.0003

g

a
O
8
(a)

 

AAbsorbance

00006

I
—

 

 

 

2000.1800‘1600‘1400‘12100 1000

Frequency (cm-1)

139

The spectrometer set-up designed to maintain the required temperature

control precision was detailed in Section 2.5 of the experimental section.
Employment of this set-up enables temperature control to :t 0.005 °C.

Figure 5-40 illustrates that at this level of precision, undesired thermal

contributions to the baseline are largely eliminated, resulting in a baseline that is

linear with an extremely low peak-to-peak noise level (AA ~ 1 x 10'5).

5.7 Conclusions

There are a number of factors that contribute to the technical difficulty of
obtaining high quality, IR difference spectra of proteins. A protocol has been
developed using advanced materials and instrumentation that alleviates many of
the experimental difficulties with electrochemical difference FTIR. The detection
of weak absorption signals, in an othenivise highly absorbing sample is possible
at high signal-to-noise ratio using this experimental set-up. Implementing
reaction-induced difference spectroscopy with a precise background subtraction
and limited path lengths of less than 10 um eliminates the large background
interference from water in the mid-IR region. A highly sensitive quantum detector
(MCT) is employed to improve the signal-to-noise ratio. Lastly, precise
temperature control of the system is utilized to diminish baseline distortions and
to eliminate thermal contributions to the difference spectra, further improving the
signal-to-noise ratio.

To date, attempts to measure low-frequency electrochemical difference

transmission FTIR spectra have been plagued by large water absorbance, even

140

at pathlengths less than 10 pm. In the light-initiated method used to measure
low-frequency IR spectra of the photosystem II protein, samples were dried
directly onto the IR windows, successfully decreasing the amount of water
present [2, 3]. However, as previously mentioned, the electrochemical approach
requires some amount of aqueous solution for maintaining electrical contact
between the electrodes. An alternate approach for performing electrochemical
difference transmission IR measurements, relies on the method of protein film
voltammetry [4649] Protein film voltammetry is a general concept that is based
on the fact that direct electron transfer has been observed for protein films
immobilized at electrode surfaces. There are a variety of ways to achieve such
electroactive protein films, including self-assembled monolayer technology and
cast biomembrane-like films [47]. In the latter approach, the protein is embedded
into a membrane-like film comprised of surfactant molecules that allows for direct
electron transfer between the protein and electrode surface. This method looks
promising for decreasing the water content required for electrochemistry, as
these protein films are essentially dried, yet they retain conductivity, as well as
native protein structure and catalytic activity [48, 50, 51]. Protein film
voltammetry has proven successful for a number of heme proteins at a variety of
electrode surfaces. In particular, Hawkridge and coworkers have successfully
immobilized cco in a lipid bilayer membrane on a Au electrode. They
demonstrated that the cco undenivent direct electron transfer with the electrode
and mediated the electron transfer from cytochrome c in solution to the electrode

[52].

141

Alternatively, a reflection method, for example, attenuated total reflection
(ATR), could be employed, instead of the transmission mode. The advantages of
ATR include less interference from water absorption, the possibility for increased
signal amplitude due to multiple reflections, and ease of spectroelectrochemical
cell design. Additionally, as shown in Table 5-3, many of the materials that are

transmissive below 1000 cm’1 and are not soluble in water possess high indices

of refraction, and are, therefore, commonly employed as ATR crystals
Chapter 7 details the viability of using a boron-doped diamond thin-film on Si as
an optically transparent electrode for IR transmission spectroelectrochemical
measurements of a protein (cytochrome 0). Owing to the high refractive index
properties of both BDD and Si, this OTE could also be implemented in reflection

mode.

Aside from attempts to minimize the water absorption below 1000 cm'1, a

highly sensitive detector is required for measuring the weakly absorbing modes
in the low-frequency IR region. Attempts to interface the He-cooled bolometer
with the Bruker FTIR spectrophotometer using a homemade external detector
compartment were unsuccessful. As discussed earlier, the MCT detector does
not possess adequate sensitivity in the low-frequency IR region, which is
required for protein measurements in aqueous solution. It does not seem
possible to measure high-quality low-frequency data using this FTIR

spectrophotometer until this detector problem is solved.

142

5.8

[1-1

[2-1

[3-1

[4-1

[5.]

[6.]

[7.]

[3-1

[9-1

[10.]

[11.]

[12.]

References

J. L. Arrondo, A. Muga, J. Castresana and F. M. Goni Prog. Biophys. Mol.
Biol. 1993, 59, 23-56.

H.-A. Chu, M. T. Gardner, J. P. O'Brien and G. T. Babcock Biochemistry
1999, 38, 4533-4541.

H.-A. Chu, M. T. Gardner, W. Hillier and G. T. Babcock Photosynth. Res.
2001, 66, 57-63.

M. T. Gardner, G. Deinum, Y. Kim, G. T. Babcock, M. J. Scott and R. H.
Holm Inorg. Chem. 1996, 35, 6878-6884.

P. Hellwig, B. Rost, U. Kaiser, C. Ostermeier, H. Michel and W. Mantele
FEBS Lett. 1996, 385, 53-57.

K. ltoh, T. Shimanouchi and M. Oya Biopo/ymers 1969, 7, 649-658.

T. Miyazawa, T. Shimanouchi and S. Mizushima J. Chem. Phys. 1958, 29,
611-616.

J. Ramsden and T. G. Spiro Biochemistry 1989, 28, 3125-3128.

S. 2. Hu, K. M. Vogel and T. G. Spiro J. Am. Chem. Soc. 1994, 116,
11187-11188.

M. Tsubaki, R. B. Srivastava and N. T. Yu Biochemistry 1982, 21 , 1132-
1140.

L. Nestor, J. A. Larrabee, G. Woolery, B. Reinhammar and T. G. Spiro
Biochemistry 1984, 23, 1084-1093.

T. J. Thamann, P. Frank, L. J. Willis and T. M. Loehr Proc. Natl. Acad. Sci.
U. S. A. 1982, 79, 6396-6400.

143

[13.]

[14.]

[15.]

[16]

[17.]

[18.]

[19.]

[20.]

[21.]

[22.]

[23.]

[24.]

[25.]

[26.]

S. A. Asher, L. E. Vickery, T. M. Schuster and K. Sauer Biochemistry
1977, 16, 5849-5856.

I. Nakagawa and T. Shimanouchi Spectrochim. Acta 1964, 20, 429-439.

K. Nagai, T. Kagimoto, A. Hayashi, F. Taketa and T. Kitagawa
Biochemistry 1983, 22, 1305-1311.

K. Nagai, T. Kitagawa and H. Morimoto J. Mol. Biol. 1980, 136, 271-289.

J. Teraoka and T. Kitagawa J. Biol. Chem. 1981, 256, 3969-3977.

S. Hashimoto, Y. Tatsuno and T. Kitagawa Proc. Jpn. Acad. Ser. B-Phys.
Biol. Sci. 1984, 60, 345-348.

S. Hashimoto, Y. Tatsuno and T. Kitagawa Proc. Natl. Acad. Sci. U. S. A.
1986, 83, 2417-2421.

R. T. Kean, W. A. Oertling and G. T. Babcock J. Am. Chem. Soc. 1987,
109, 2185-2187.

W. A. Oertling, R. T. Kean, R. Wever and G. T. Babcock Inorg. Chem.
1990, 29, 2633-2645.

A. J. Sitter, C. M. Reczek and J. Terner J. Biol. Chem. 1985, 260, 7515-
7522.

A. J. Sitter, C. M. Reczek and J. Terner Biochim. Biophys. Acta 1985, 828,
229-235.

J. T. Groves and M. K. Stern J. Am. Chem. Soc. 1988, 110, 8628-8638.

K. Rahmelow and W. Huebner Appl. Spectrosc. 1997, 51, 160-170.

P. R. Griffiths and J. A. de Haseth Fourier Transform Infrared
Spectrometry; Wiley-Interscience, New York, 1986.

144

[27.]

[28.]

[29.]

[30.]

[31 .1

[32.]

[33.]

[34.]

[35.]

[36.]

[37.]

[38.]

[39.]

J. L. Koenig and D. L. Tabb Infrared Spectra of Globular Proteins in
Aqueous Solution; D. Reidel Publishing Co., New York, 1980.

C. Zscherp and A. Barth Biochemistry 2001, 40, 1875-1883.

W. Mantele In Biophysical Techniques in Photosynthesis; J. Amesz and A.
J. Hoff, Eds; Kluwer Academic Publishers, Dordrecht, 1996; Vol. 3, 137-
160.

H. A. Chu, W. Hillier, N. A. Law and G. T. Babcock Biochim. Biophys. Acta
2001, 1503, 69-82.

H. A. Chu, W. Hillier, N. A. Law, H. Sackett, S. Haymond and G. T.
Babcock Biochim. Biophys. Acta 2000, 1459, 528-532.

D. Moss, E. Nabedryk, J. Breton and W. Mantele Eur. J. Biochem. 1990,
187, 565-572.

P. Hellwig, U. Pfitzner, J. Behr, B. Rost, R. P. Pesavento, W. von Donk, R.
B. Gennis, H. Michel, B. Ludwig and W. Mantele Biochemistry 2002, 41 ,
9116-9125.

H. Michel Biochemistry 1999, 38, 15129-15140.

M. Wikstrém Biochim. Biophys. Acta 2000, 1458, 188-198.

F. Baymann, D. A. Moss and W. Ma'ntele Anal. Biochem. 1991, 199, 269-
274.

J. Behr, P. Hellwig, W. Mantele and H. Michel Biochemistry 1998, 37,
7400-7406.

P. Hellwig, S. Grzybek, J. Behr, B. Ludwig, H. Michel and W. Mantele
Biochemistry 1999, 38, 1685-1694.

P. Hellwig, C. Ostermeier, H. Michel, B. Ludwig and W. Mantele Biochim.
Biophys. Acta 1998, 1409, 107-112.

145

[40.]

[41.]

[42.]

[43.]

[44.]

[45.]

[46.]

[47.]

[48.]

[49.]

[50.]

[51.]

[52.]

P. R. Rich and J. Breton Biochemistry 2001, 40, 6441-6449.

R. N. A. H. Lewis and R. N. McElhaney In Infrared Spectroscopy of
Biomo/ecules; H. H. Mantsch and D. Chapman, Eds; Wiley-Liss, Inc.,
New York, NY, 1996, 159-202.

B. Stuart Biological Applications of Infrared Spectroscopy; John Wiley &
Sons, New York, 1997.

J. D. lngle and S. R. Crouch Spectrochemical Analysis; Prentice Hall,
Englewood Cliffs, NJ, 1988.

J. H. Moore, C. C. Davis and M. A. Coplan Building Scientific Apparatus: A
Practical Guide to Design and Construction; 2nd ed.; Perseus Books,
Reading, MA, 1991.

H. A. Chu, H. Sackett and G. T. Babcock Biochemistry 2000, 39, 14371-
14376.

J. F. Rusling Acct. Chem. Res. 1998, 31, 363-369.

F. A. Armstrong and G. S. Wilson Electrochim. Acta 2000, 45, 2623-2645.

A.-E. F. Nassar, W. S. Willis and J. F. Rusling Anal. Chem. 1995, 67,
2386-2392.

J. F. Rusling and A. E. F. Nassar J. Am. Chem. Soc. 1993, 115, 11891-
11897.

J. F. Rusling and Z. Zhang Electroanalytical Methods for Biological
Materials 2002, 195-231.

J. F. Rusling and Z. Zhang Handbook of Surfaces and Interfaces of
Materials 2001, 5, 33-71.

J. K. Cullison, F. M. Hawkridge, N. Nakashima and S. Yoshikawa
Langmuir 1994, 10, 877-882.

146

CHAPTER 6

6 Electrochemical Difference FTIR Spectroscopy
of Cytochrome c Oxidase

6.1 Introduction

Cytochrome c oxidase (cco) is the terminal electron acceptor in the
respiratory chain and is found in the mitochondria of eukaryotes, in many
bacteria, and in the archea. It is the last of the transmembrane proteins involved
in the respiratory electron transfer chain and is one member of the heme-Cu
terminal oxidase protein family [1]. The chemistry catalyzed by cco involves an

efficient reduction of molecular oxygen to water with rates that exceed 2000 s'1.

Concomitant with this reduction chemistry is the uptake of four chemical protons
and four pump protons, which are translocated across the membrane. It is the
net translocation of protons across the membrane that generates a membrane
potential and is subsequently used for ATP synthesis [1-3]. Simply written, the

chemistry catalyzed by cco can be presented by:

4e‘+8H,;,+02 —9 2r120+4H+

()Ul '

The chemistry of 02 reduction and proton pumping centers on the

involvement of a number of cofactors and specific amino acid residues. The
structural framework for these cofactors is comprised of a number of protein
subunits. The bacterial cco contains 3 - 4 subunits and the mammalian oxidase

from bovine heart contains 13 subunits. Three different forms of cco proteins

147

have been structurally characterized. These are the bovine heart [4, 5],
Paracoccus denitn'ficans (P. denitn'ficans) [6, 7], and Rhodobacter Sphaeroides
(lwata unpublished results) proteins. A schematic of subunits l-lll from
Rhodobacter Sphaeroides (R. Sphaeroides) is shown in Figure 6-1. Thus far the
function of the additional 9- 10 subunits found in the mammalian oxidase is

unclear. They do not appear to affect the chemistry of 02 reduction, but may

eventually be found to have regulatory roles or aid in protein stability.

Figure 6-1 Cytochrome c Oxidase from Rhodobacter Sphaeroides

Simplified reaction scheme of membrane-bound cco including subunits I, II and III. The
electron transfer steps are indicated by arrows, starting at cytochrome c, to the binuclear
CuA site, to heme a, and finally to the binuclear center (heme a3 and CUB). Two proton
uptake pathways are shown starting at residues K362 and D132. Other key residues
are listed for reference. Oxygen binding, proton exit and water exit pathways are also
shown.

 

148

The rapid rate of 02 reduction by cco directly involves four cofactors that

participate in a sequence of long-range electron transfer events. Electron

transfer is initiated by electron donation from cytochrome c to the dinuclear

copper A (CuA) site. This is followed by subsequent electron transfer to the

heme a, and then into the heme a3 - CuB binuclear site where 02 reduction and
bond cleavage is catalyzed. The redox potentials (vs. NHE) for the metal centers

are accepted to be 0190-0245 V for CuA, 0.190-0.210 V for heme a,

0350-0385 V for heme 33 and 0.350 V for CuB, as determined by
spectroelectrochemical [8], redox titration [9] and direct electrochemical [10-12]
methods.

The initial binding of 02 to the reduced form of the enzyme (states: CuA',
Fea”, Fea3", CuB') results in formation of the heme Fea3H—Oz state (intermediate

A), which is then converted via a Fe33IV=O intermediate (Pr) to the oxidized state

of the enzyme (CuA', Fea', Feagm-OH, CuB'l-OH). The electron inventory

between the fully oxidized form and the fully reduced form of the enzyme is 4 and
the kinetic steps have been resolved with optical and vibrational spectroscopies
[1, 2, 13, 14]. Other metal ions present that do not participate directly in electron

transfer include Mg, Ca and Zn ions. The requirement for, and, therefore, the
role of these metals in 02 chemistry is unclear.

Coupled with the 02 electron transfer reaction is the catalytic pumping of
four protons across the membrane and the requirement of four protons as the

substrate 02 is reduced to H20. Two proton transfer pathways, termed the K

149

and D channels, have been identified in subunit l for facilitating the movement of
protons from the interior side of the membrane towards the binuclear site. The K
channel pathway is termed for an essential lysine residue K(l-362)
(R. Sphaeroides numbering) and the D channel for an essential aspartate residue
D(I-132) that leads to glutamic acid E(l-286). The elucidation of these pathways
is derived from structural information [4-7] and from mutational and kinetic

measurements [15-17]. Further pathways for the pumped protons have also

been proposed [5, 18], as have pathways for the substrate 02 [19-21] and

product H20 [22-24].

Recent mechanistic emphasis with the heme-Cu oxidase enzymes has
been directed at understanding the important bioenergetic question concerning
the coupling of electron transfer with proton pumping. In particular, questions

remain regarding the proton pathways and which reaction steps of 02 reduction

are coupled to H+ [25-28]. In order to pump protons against the electrochemical

gradient it is commonly held that amino acid groups undergo a pK change in
response to metal ion reduction-oxidation steps.

In cco, a small number of amino acid residues have been specifically
implicated in this process. One experimental approach offering insight into the
involvement of amino acid residues in the catalysis is Fourier Transform Infrared
(FTIR) difference spectroscopy, as it offers the sensitivity to reveal single amino
acid changes in response to protonation, ligation, and redox state changes of an
enzyme. The FTIR approach can be addressed in a number of ways, from a

kinetic to a static perspective. One approach is to follow events initiated by the

150

photolysis of CO from heme a3, and at low temperatures (< 140K), the transfer of

CO to CuB. This approach has been used by a number of groups to study the

binuclear site [29-32] and has also been used successfully to study changes in a
time-resolved manner [33-35]. Another approach examines changes between

the fully oxidized and reduced states of cco and involves the electron transfer
steps of Cu and heme a, in addition to the chemistry of the binuclear site. This

approach lies solely in the physiological temperature regime, and may be
initiated via several means, including the photochemical reduction of cco [36, 37],
the indirect redox titration of the enzyme [38], or additionally, the pioneering
approach of electrochemical titration (accelerated with redox mediators) of cco
[39-42]. In the future, it may be possible to perform these spectroelectrochemical
experiments with cco directly anchored to an electrode [12].

This chapter discusses the application of electrochemical difference FTIR

to examine cco from R. Sphaeroides and bovine heart. Isotope labeling
experiments were performed to simplify the spectral assignments. Global 15N

labeling of cco of R. Sphaeroides, together with deuterium exchange, has

identified a histidine signal in oxidized minus reduced FTIR spectra. A positive

1 upon 15N exchange and to

band at 1105 cm", which shifted to 1095 cm'
1104 cm'1 upon deuteration, was assigned to the C-N stretching mode of a

histidine side chain. The relatively high frequency at 1105 cm'1 and the observed

insensitivity to deuteration indicated that the imidazole ring of this histidine has

an N6 protonated form. Definitive assignment of this mode to a specific histidine

151

side chain can only be achieved by further experiments to measure the FTIR
spectra of R. Sphaeroides samples containing specifically labeled 15N histidine

residues and site-directed mutants replacing histidine.

6.2 Results

6.2.1 Electrochemical Difference Spectra

The oxidized and reduced forms of R. Sphaeroides cco were obtained by
electrochemical poise of the enzyme in the OTTLE cell facilitated by redox
mediators. The UV/vis optical spectra of the fully oxidized and reduced forms of
the enzyme were then recorded at potentials of +0.50 V and —0.50 V

(vs. Ag/AgCl), respectively. The optical properties of cco report on the oxidation
state of the heme a, heme a3, and CuA metal centers, as well as the coordination

state of the heme iron atoms. The UV/vis spectrum of cco contains two major
bands, the Soret and d-bands The most intense of the two peaks is the Soret
band, which occurs between 400 - 450 nm, depending on the nature of the heme
iron and the state of the enzyme [43]. This band is attributed to the rr—>n*
transitions of the hemes, and is therefore a very sensitive indicator of the heme
iron oxidation state. The or-band (~ 606 nm) is of much lower intensity, and is

due to n—m" transitions of heme a.

152

Figure 6-2 UV/vis Electrochemical Difference Spectra of Rhodobacter
Sphaeroides Cytochrome c Oxidase

Optical UV/vis electrochemical difference spectra of the oxidized minus reduced (solid
line) and reduced minus oxidized (dashed line) forms of cco from Rhodobacter
Sphaeroides. The oxidized and reduced spectra were acquired at +0. 50 and —0. 50 V
(vs. Ag/AgCI), respective/y. Potentials were applied externally via a potentiostat and
electrochemistry was accelerated with a series of redox mediators—see Table 2-1 for
details.

 

l0
3 AA 0.05

 

 

 

 

 

Wavelength (nm)

The oxidized minus reduced (solid line) and reduced minus oxidized
(dashed line) difference spectra are shown in Figure 6-2. When fully oxidized,
the Soret maximum occurs at 423 nm, and when fully reduced, the Soret
maximum shifts to 445 nm and a prominent d-band at 606 nm appears. In the
oxidized minus reduced difference spectrum, these features are manifested as a
positive peak at 423 nm representing the oxidized form and negative peaks at
445 and 606 nm corresponding to the reduced state. These features are the
same as reported earlier [44] and are indicative that the direct electrochemical
titration of this enzyme is achievable. It is important to note that the response is

stable in time and reproducible with repeated cycling. The two difference spectra

153

are mirror images of the other, as expected for a reversible and stable electron
transfer process.

After obtaining a UV/vis difference spectrum, the enzyme sample was
placed into the FTIR cryostat, the temperature was equilibrated, and
electrochemical difference spectra were recorded. Details of the cryostat and
temperature control setup were provided in Chapter 2 (section 2.5). Figure 6-3A
shows the noise spectrum acquired for this experimental setup. This spectrum is
obtained by measuring two consecutive spectra at a constant potential. The

average peak-to-peak noise level is below 1 x 10'5 absorbance units. The

oxidized minus reduced (solid line) FTIR spectrum for bovine cco for a potential
step from -0.50 to 0.50 V is shown in Figure 638. The positive bands in this
spectrum correspond to the oxidized state and the negative, to the reduced state.
The difference spectrum for the reverse process, (i.e., for a 0.50 to -0.50 V step),
the reduced minus oxidized spectrum (dashed line), is also shown in
Figure 638, for comparison. As was observed for the optical difference spectra,
these spectra are the mirror image of each other, indicating a quantitative, fully
reversible electrochemical reaction at the electrode surface. Examination of the
background sample containing only the mediators and HEPES buffer, shown in
Figure 6-3C, reveals minimal spectral contribution that would interfere with the

cco FTIR difference spectrum. However, a small region in the cco difference
spectra, shown in Figure 638, between 1700 - 1600 cm'1 and 1040 - 1000 cm'1,

appears to exhibit reduced symmetry. These regions coincide with the strong

underlying protein/buffer absorption and are indicative of a small baseline artifact.

154

This origin of this baseline phenomena is likely derived from a non-linearity in the
response of the MCT detector, and although apparent, is not of any consequence
to the findings of this work.

In the oxidase difference spectra, a number of vibrational bands can be

seen to change in the 1800 - 1000 cm'1 mid-IR regions. Such peaks arise from

protein vibrational changes in response to the oxidation and reduction of the Cu
and heme Fe centers. The changes in the FTIR difference spectra are detailed,

and in many instances correspond to AA < 104. The most intense band in the

absolute absorbance spectrum of a protein is the Amide l peak at ca. 1658 cm'1,

which represents the nature of the peptide backbone structure via C=O
vibrations. Comparing the total absorbance of the protein at this peak, which is

~ 0.7 absorbance units, to that of the strongest band (1662 cm'1) in the difference

spectrum (AA ~ 1 x 10'3), it is apparent that the majority of the protein acts as an

inert scaffolding during the redox process, undergoing no gross conformational
change in the peptide backbone structure. Instead, the FTIR spectrum of the
fully oxidized minus reduced state of cco corresponds to prominent changes in a
small number of amino acid residues that undergo a change between the
oxidized and reduced forms [45]. A significant contribution to the electrochemical

difference spectra is also derived from the two heme groups [39, 42].

155

Figure 6-3 FTIR Electrochemical Difference Spectra of Rhodobacter
Sphaeroides Cytochrome c Oxidase

Electrochemical FTIR difference spectra (1800 - 1000 cm‘1) recorded at 10 0C.
(A) Oxidized minus oxidized noise spectrum. (8) Spectrum of the cco enzyme from
bovine heart as the oxidized minus reduced (solid line) and reduced minus oxidized
(dashed line) difference spectrum (average of 3 samples). (C) Spectrum of the mediator
cocktail and the sample buffer (no enzyme) as an oxidized minus reduced difference

spectrum. All spectra were acquired using BaFg windows and a 10 pm pathlength.

Potentials of +0. 50 V and -0. 50 V (vs. Ag/AgCI) were applied to obtain the oxidized and
reduced states, respectively.

 

 

 

 

 

A L—u—W ‘W-‘A‘ —==~W
E Amide I
g ‘ 6(N H)
v(C=0) 5 1548 v(CO),5(COH)

  

 

 

 

 

 

 

 

 

1537 Tit v(CN) His

 

 

 

 

 

 

 

 

v(CC) ,5(CH)

 

 

 

 

 

 

 

 

 

 

Tyr
-3
M11 x 10
Amide I
v(C=0)
C WWW-MM
1800 1600 1400 1200 1000

Frequency (cm'1)

156

As illustrated in Figures 6-3 and 6-4, the mid-IR spectrum can be broken
into distinct regions of interest, according to structural moieties. Exclusive

contributions from the carbonyl stretching mode of protonated Asp and Glu
amino acid side chains are expected in the spectral region above 1700 cm'1 [46].
These modes are sensitive to H/D exchange, experiencing downshifts of
8 - 10 cm'1, as well as to hydrogen bonding, with shifts up to 50 cm'1 observed
upon formation of strong hydrogen bonds [46]. In the Amide l range
(1680-1620 cm'1), strong signals at 1684, 1662, 1646 and 1634 cm'1 are
attributed to changes in the carbonyl stretching modes of the polypeptide
backbone. These modes occur below 1700 cm'1 due to the strong hydrogen

bonding experienced by the backbone. The relatively small amplitude of these
signals indicates subtle structural rearrangements of individual peptides upon the
redox process. Additionally, contributions from the formyl and propionate groups

of heme a and heme a3 [29, 47-50], and from individual amino acid side chains

(Asn, Gln, and Arg) are expected in this region [39, 46].

A number of peaks are observed in the region from 1580 to 1515 cm'1

(Amide lI range). Signals due to the coupled CN stretching and NH bending
modes of the protein backbone, the Amide Il modes, are expected in this region
[51]. In addition to the Amide Il modes, vibrational modes of heme C=C groups

[29, 47-50] and aromatic amino acid side chains (Tyr, Phe, and Trp). as well as
asymmetric COO' modes of Asp, Glu or heme propionates, also contribute in this

range [39, 46]. In particular, a coupled C=C stretching and C-H bending mode of

157

Tyr occurs near 1517 cm'1 [46]. Additionally, Lys residues contribute a weak

asymmetric deformation vibration of the NH3+ group near 1526 cm'1 [46]. The
deprotonated Asp, Glu and heme acid symmetric carboxyl modes are expected

between 1450 and 1400 cm“.

Figure 6-4 Illustration of Characteristic Mid-IR Bands

A schematic depiction of some characteristic absorption bands observed in the mid-IR

range (1800 - 1000 cm'1) of protein F T/R spectra. Bending modes are indicated by the
symbol 6 and stretching modes by the symbol v. Asymmetric and symmetric modes
indicated by as and s, respectively.

 

Asp. Glu.
heme-COOH

v<C=Oi .
:\IIII(I(‘I ;.(C=o)

Amide ll ]:] §\'(CN). élNH)

Asp” . 5
Glu fl \Jas(COO-z)

licme-COO- E] g
5 ’ iir El vim)

[it 603011)

His - V(CN)

heme ring fl v(CC)

I + E
lysfl 85(NH3 ) g
[If IE v(CC)ring,8(CH)
EPlie,Trpl : :

Asp. 0hr, E] \'5(COO7)
heme-COO- ' é

 

 

 

1800 1700 1600 1500 1400 1300 1200 1100 1000
Frequency (cm'1)

158

The wide range from 1260 to 1170 cm"1 is expected to contain
contributions from the coupled Tyr C-OH bending and C-0 stretching mode. In

most protein studies this mode has been observed from 1230 to 1250 cm"1 [46].

In the spectral region below 1150 cm'1, contributions from the vibrational modes

of Trp, the CN stretch of His, and the C-0 stretching modes of the aliphatic,
hydroxyl-containing amino acid side chains (Thr and Ser) are expected [46].
Tentative mode assignments in the mammalian cco FTIR spectra can be
made based on comparisons with reports of characteristic modes of model
peptides [52, 53], isolated amino acids and other protein infrared studies [46, 54].
Comparison with resonance Raman data provides additional information,
especially for the heme structural elements [29, 55]. The bacterial form of the
enzyme, however, lends itself to additional approaches that simplify and
substantiate band assignment. These provide direct evidence through
biochemical and molecular biological techniques, which introduce global isotopic

, 13C or 15N), specifically labeled amino acid residues (e.g., 13C Ala

labeling (e.g.
or 15N His) [42, 56], and site-directed mutations [40, 57]. Aside from the isotope

methods, researchers studying cco have an added advantage, in that they have
access to a number of different forms of the enzyme. Comparisons of FTIR
difference spectra for mammalian and two bacterial forms provide further insight

into the mode assignments.

6.2.2 Comparison of Bacterial and Mammalian Forms of CcO

Cytochrome c oxidases from R. sphaeroides and bovine heart perform the

159

same 02 chemistry and show high structure homology, and yet are comprised of
a different number of subunits. A comparison of the 3-subunit bacterial enzyme
(solid line) with that of the 13-subunit mammalian enzyme (dashed line) in Figure
6-5, reveals a number of significant spectroscopic differences between the two
enzymes. Slight variances in the mode intensities and frequencies point to
subtle differences in the arrangement of amino acid side chains and cofactor
sites. Additional or absent modes reflect more extreme deviations, despite the
high degree of structural similarity between both enzymes. Some of the more
prominent changes and proposed assignments are listed in Table 6-1. This
finding provides an indication of the subtle interactions the amino acid residues
provide in facilitating electron transfer to the oxidase cofactors (assuming
cofactor placement is the same).

Earlier work, comparing the redox-induced difference FTIR spectra of two
and four subunit cco from P. denitrificans, indicates that the presence of
additional subunits contributes to protein-protein interactions that influence the
absorption and intensity of heme vibrational signals [58]. Heible and coworkers
also observed a difference in the Raman-active heme vibrational modes in a
comparative study of the two and four subunit forms of P. denitrificans and the
13-subunit oxidase from bovine heart [55]. The spectral discrepancies are in line
with the observations made by Hellwig and coworkers, in a comparison of the
FTIR electrochemical difference spectra of cco from P. denitrificans and bovine

hean[54]

160

Figure 6-5 Comparison of the FTIR Difference Spectra of Bacterial and
Mammalian Cytochrome c Oxidases

Electrochemical oxidized minus reduced FTIR difference spectra (1800 - 1000 cm‘1) of
cco from Rhodobacter sphaeroides (solid line) and bovine heart (dashed line). Spectra
were recorded at 10 0C using Ban windows and a 10 pm pathlength. Potentials of

+0. 50 V and —0. 50 V (vs. Ag/AgCl) were applied to obtain the oxidized and reduced
states, respectively.

 

AA 5x10-4

 

 

 

 

 

iiiliiilimiliiL

1800 1600 1400 1 1200 1000
Frequency (cm' )

 

In accordance with Hellwig et al.’s report, a significant spectral change

between the bacterial and mammalian forms occurs in the carbonyl region. A

difference signal with a positive peak at 1746 cm"1 and a corresponding negative
peak at 1735 cm'1 is observed for the R. sphaeroides sample. The bovine

spectrum reveals two positive peaks at 1748 and 1734 cm'1 that are possibly

161

1

related to a broad negative feature at 1721 cm' . A significant variation also

exists between the bovine and bacterial spectra at 1537 cm’1.

Table 6—1 Discrepant Modes and Tentative Assignments for FTIR
Electrochemical Difference Spectra of Cytochrome c Oxidases

 

 

. a . a . . . o Tentative
Bovrne heart R. sphaerordes P. denitrificans Assignment
1748, 1735 (ox) 1746 (ox) 1746 (ox) v(C=O) Glu/Asp

1740 (red)C 1735 (red) 1734 (red) v(C=O) Glu/Asp
1685 (ox) Amide I (B-sheet)
1684 (red) 1684 (red) Amide I (B-sheet)
Amide I (or-helix),
1651 (ox) 1656 (ox)
v(C=O) heme a3
1634 (red) 1632 (red) Amide I (ct-helix)
Amide II, CuHis,
1537(ox) 1538(ox)
vas(COO‘)
1527 (ox) v(CC) heme
5<NH3*>S Lys.
1525 (red) 1522 (red)

Amide II

 

a This work. bFrom reference [54]. CFrom reference [36], not observed in this work.

A mode at 1537 cm'1 is observed in both bacterial spectra (1538 cm”1 in

P. denitrificans), but is absent in bovine electrochemical difference spectra (this
work and [54]). However, this mode has been observed for bovine CO photolysis

difference spectra [29, 33]. It is possible that the different experimental

162

procedures result in structurally different forms of the enzyme, or that the

extinction coefficient of this mode changes. A comparison of the carbonyl
spectral region above 1700 cm"1 for the CO photolysis and electrochemical

difference procedures also reveals differences, suggesting that subtle differences
in enzyme structure result from the different approaches (this work and [33, 54]).
Based on the observed results from this work and those in the literature,

as well as from characteristic group IR frequencies for amino acids and protein

vibrations, several assignments for the mode at 1537 cm'1 are possible. The
asymmetric COO' stretching mode of Asp, Glu, the C-terminus, or heme
propionates, the Amide ll mode, and the symmetric NH3+ deformation mode of
Lys or the N-terminus are possible candidates (Figure 6-4). At first thought, the
large peak amplitude, and the 16 cm'1 downshift observed upon exchange with

15N, suggest it is due to a contribution from the Amide ll mode. However, Hellwig

and coworkers have shown that a peak at 1538 cm'1 in the P. denitrificans cco
enzyme was sensitive to 13C heme labeling and have assigned it to 3 COO”
asymmetric stretch [42]. The presence of a corresponding COO" symmetric

stretching mode near 1416 cm'1 further supported their assignment.
Comparisons of bovine cco resonance Raman data to that of Cu-bound
imidazole complexes, suggested that a mode at 1537(8) cm’1 would arise from a
Cu-His coupled vibration [29, 33]. The observed 15N isotope shift in this work

would support an assignment of this peak to a N-containing His vibration. It is

likely that the true assignment of the peak at 1537 cm'1 is a combination of the

163

possibilities listed above, rather than that of a single mode. On account of the
relatively large amplitude of this mode, it is possible that there are several
smaller modes overlaid at the same frequency. Noguchi and coworkers
observed this situation for light-induced FTIR difference spectra of the
photosystem II enzyme [59]. Isotope labeling experiments (13C and 15N) and the
presence of the corresponding symmetric mode supported the assignment that a

mode near 1550 cm'1, in the Amide Il region, was actually an asymmetric COO“

stretching mode, overlaid by an Amide ll mode [59].

1

A peak at 1517 cm' in the reduced form of both the bovine and

R. sphaeroides is attributed to a tyrosine ring stretching mode, confirming the
previous assignments by Hellwig in P. denitrificans [60] and Rich and Breton in
bovine [36]. An interesting question then arises, as structural data have shown a
conserved tyrosine that is covalently linked to His 333 of CuB located near the
binuclear center [4] that has been implicated in both the binding of intermediates
[61] and in proton transfer [62, 63]. Could the mode at 1517 crn'1 be attributed to

this unique residue? This cross-linked tyrosine would be influenced by the redox
transitions of the binuclear center, and would, therefore, contribute to an
electrochemical difference FTIR spectrum. Measurements in 020 indicate that
this mode is sensitive to H/D exchange, suggesting an exchangeable protein
residue. Comparison of characteristic bands of the tyrosine side chain to those
of a model tyrosine-histidine compound are helpful in determining if peaks

originate from the cross-linked residue.

164

The neutral tyrosine side chain exhibits strong characteristic modes at
1518 cm'1 and at 1249 cm'1 for a coupled ring CC stretching mode, CH bending
mode, and a coupled CO stretching-COH bending mode, respectively [46]. The

CC ring stretching mode has been observed at 1518 cm'1 and the CO stretching

and COH bending mode as a broad peak at 1282 cm'1 in the FTIR spectrum of a

synthetic model of the cross-linked histidine tyrosine side chain, 2-imidazole-1-yl-

4-methylphenol [60]. From these examples, it is clear that the mode at

1517cm’1 is not strongly influenced by the presence of the cross-linked

imidazole, but that by examining the region from 1250-1300 cm'1 one may

distinguish between the traditional and the cross-linked forms of tyrosine. It
seems that the problem is not so easily solved in the complex oxidase difference
spectrum. However, site-directed mutants of the cross-linked tyrosine, combined
with the known IR spectra of tyrosine and the synthetic cross-linked tyrosine
have led Hellwig and coworkers to assign a feature at 1268 cm’1 in the
P. denitrificans cco spectrum to the cross-linked Tyr284 residue [60]. In this work

peaks at 1264 and 1269 cm'1 are observed for the oxidized forms of

R. sphaeroides and bovine cco, respectively. It is likely that these modes
correspond to that observed by Hellwig in the P. denitrificans cco spectrum, but
without further site-directed mutagenesis data, it is not possible to say so
definitively.

There are a few striking discrepancies observed in Figure 6-5, with

respect to Hellwig and coworkers’ study. For instance, the tentative assignment

165

of a reduced mode at 1588 cm"1 to the asymmetric carboxyl stretch of an
exclusive (i.e., not conserved in bacterial sources) mammalian residue, Asp51,
[54] seems unlikely, since this mode is present at 1586 cm'1 in both spectra

shown in Figure 6-5. Rich and Breton also point out that it would be unusual for

this residue to be deprotonated in the reduced form [36].

6.2.3 Isotopically-Labeled Spectra

The complexity of assigning vibrational bands can be reduced with global
isotopic labeling experiments of the cco protein. Figure 6-6 presents the oxidized
minus reduced FTIR difference spectra from R. sphaeroides (Figure 6-6A) and
bovine (Figure 6-6B) oxidase samples following deuterium buffer exchange, and
Figure 6-7 the R. sphaeroides oxidase sample following global 15M labeling. In
both cases the labeled spectra (dashed lines) are overlaid with the corresponding
unlabeled spectra (solid lines). The deuterium labeling reveals a number of small
spectral changes as a consequence of the H/D exchange. One region to change

1

significantly upon deuteration is the C=O region at 1740-1730 cm” for

R. sphaeroides. However, relatively few H/D changes occur within the strong IR

bands that overlay the Amide I and Amide ll regions, 1689-1660 and

1550 - 1500 cm'1, respectively.

166

Figure 6—6 Comparison of the FTIR Difference Spectra of Cytochrome c
Oxidase Measured in 020 and H20

Electrochemical oxidized minus reduced FTIR difference spectra (1800 - 1000 cm'1) of
cco from (A) Rhodobacter sphaeroides and (8) bovine heart. The spectra of the global
deuterium labeled enzyme (dashed line) are overlaid with the unlabeled enzyme (solid
line). To aid the comparison, the spectra from the deuterated samples were normalized

to those from the non-deuterated samples based on the intensity of the 1540 cm'1 band.

Spectra were recorded at 10 0C using BaF2 windows and a 10 pm pathlength.
Potentials of +0. 50 V and -—0. 50 V (vs. Ag/AgCl) were applied to obtain the oxidized and
reduced states, respectively.

 

M15x10'4

 

will

 

 

 

L l l l 1 l l

1800 1700 1600 1500 1400 1300 1200 1100 1000
Frequency (cm-l)

 

 

In another isotope approach, global 15N labeling of the R. sphaeroides

enzyme was performed to elucidate the involvement of N modes. This method
facilitates the assignment of N-coupled vibrations potentially arising from lysine,

arginine, asparagine, glutamine, and histidine amino acid side chain modes, as

well as heme modes. Figure 6-7 shows the 15N (dashed line) and unlabeled

167

(solid line) oxidized minus reduced spectra from R. sphaeroides. A list of
selected 15N sensitive peaks and associated frequency shifts are presented in
Table 6-2. The labeled minus non-labeled (double difference) spectrum

(Figure 6-7 inset) indicates a number of 15N sensitive modes in the
1615 -1600 cm‘1 region. However, the most obvious changes appear in the

1550 - 1500 cm'1 region, which are coincident with the Amide ll bands

Figure 6-7 Effect of Global 15M Labeling on Mid-IR FTIR Spectra of
Cytochrome c Oxidase

Electrochemical oxidized minus reduced FTIR difference spectra of Rhodobacter
sphaeroides cco unlabeled (solid line) and 15N globally labeled (dashed line). The insert
shows the double difference spectrum (15N minus 14N) from 1625 - 1470 cm'1. Spectra

were recorded at 10 OC using BaFg windows and a 10 pm pathlength. Potentials of
+0. 50 V and —0. 50 V (vs. Ag/AgCI) were applied to obtain the oxidized and reduced
states, respectively.

 

 

 

 

 

 

 

 

, Mlsxm"
15100 ‘ A 1151001
I I I l 1 I L I 1 1 1 I 4 I i
1800 1600 1400 1200 1000

-1
Frequency (cm )

168

Table 6—2 15N Sensitive Peaks in Rhodobacter Sphaeroides Cytochrome c
Oxidase FTIR Spectra

 

 

Redox Frequency in H20 Frequency in 15N
State spectrum (cm'1) spectrum (6'04) AV (cm'1)

Ox 1560 ?

Red 1548 1534 14

Ox 1538 1522 16
Red 1525 151 1 14

OX 1498 1482 16

Ox ? 1428

OX 1 105 1095 10

 

Figure 6-8 presents an enlargement of the Amide Il region showing the
effect of 15N labeling in greater detail. However, as discussed above, the HID

exchange experiments shown in Figures 6-6 and 6-8, and reported elsewhere

[54] show relatively little change to the Amide ll region and no change to the

large 1550 cm'1 mode. The Amide lI modes are expected to downshift

~100 cm'1 upon D exchange, while the Amidel modes are less sensitive to

deuteration [51]. The Amide l signals, which are attributed to reorganization of 01

helical and [3 sheet secondary structure elements, are expected to undergo small
shifts (2- 10 cm”) after H/D exchange, largely arising from changes in the

hydrogen bonding status [52]. This finding, from Figures 6-6 and 6-8, that there
is little H/D effect on the Amide bands has also been observed in previous FTIR

difference measurements of cco from a variety of sources [40, 64]. However, the

169

result is somewhat surprising when compared to the FTIR difference spectra of

other membrane proteins [59] and synthetic peptides [65] that have shown more

substantial (2 20 cm'1) H/D effects to the Amide II bands.

Figure 6-8 Isotope Exchange Effects in the Amide lI Region

Electrochemical oxidized minus reduced FTIR difference spectra (1600 - 1400 cm'1) of
Rhodobacter sphaeroides cco unlabeled (solid line), deuterium labeled (dashed-dotted

line) and 15N globally labeled (dashed line). Spectra were recorded at 10 0C using Ban
windows and a 10 pm pathlength. Potentials of +0. 50 V and —0. 50 V (vs. Ag/AgCI) were
applied to obtain the oxidized and reduced states, respectively.

 

AAI2x10‘

 

ii
I i i

 

 

 

1600 1550 1500 1450 1400
Frequency (cm'1)

The virtual absence of shift in the Amide ll bands upon deuteration has
lead earlier work to concede that there was little or no significant Amide

contribution to cco FTIR difference spectra. Hellwig and coworkers have
assigned many of the modes in the region from 1560 to 1520 cm'1 to asymmetric

stretches of carboxyl groups of Asp or Glu residues and heme propionates,

based on the small shifts observed upon deuteration [39]. Alternate explanations

170

for the small observed shifts include a lack of solvent accessibility to the protein
backbone due to a hydrophobic environment or a tightly ordered structure
(a helix or [3 sheet) [51]. This effect is observed for the core region of
bacteriorhodopsin, which, in contrast to the peripheral region, is largely
inaccessible to HID exchange [66-68].

This raises an interesting question as to whether significant deuterium

exchange is experienced in the core of the cco enzyme following lengthy 020

incubation. Based on measurement of the Amide II, and deuterated Amide ll,

Amide Il', bands in the absolute spectra (results not shown), the D20 incubation

significantly reduced the intensity of the Amide II band at ~ 1550 cm'1 and was

replaced with an Amide II' band at ~ 1450 cm‘1. This suggests that there was

significant global deuterium labeling of the cco enzyme. It is estimated that
> 80% of the enzyme was deuterated. However, precise estimation of protein in

these regions because of the closely overlaying H-O-H and D-O-H water modes.
As the 15N data reveal an indisputable 10 cm'1 downshift of the 1550 cm"1 band

(and neighboring bands), it seems likely that in cco, deuterium exchange of the
inner core is kinetically limited beyond 3 days It is probable that this inner core
is responsible for the Amide l and Amide lI changes observed in the difference
spectra. A low level of exchange (~ 10%) occurred for the inner core, despite
lengthy incubation and concentration efforts, which took the best part of 48 h.

An interesting deuterium effect is observed for the peripheral and core
regions of bacteriorhodopsin. The peripheral region undergoes rapid HID

exchange of backbone peptide groups. However, reconstitution and refolding in

171

D20 is required to observe a complete downshift of the Amide II modes from the

core region, which consists of primarily d-helical structure largely buried in the
lipid bilayer [69]. Therefore, it is conceivable that a similar process, in which the
inner core protons are shielded from exchange, is occurring in cco. Such
shielding would result in non-exchange of the core peptide backbone, which

gives rise to the Amide ll modes.
As discussed above, based on shifts observed upon 15N exchange, it is

clear that the Amide ll modes do contribute to the oxidized minus reduced
electrochemical difference spectra of cco. Nitrogen isotope exchange

experiments are novel for this enzyme and they are critical to providing validation

for tentative mode assignments. For example, a marked (14 cm'1) downshift was
observed for the reduced mode around 1525 cm’1 upon 15N labeling, suggesting

that the previous assignment of this mode exclusively to a C00' asymmetric

stretching mode of a heme propionate [54] is unlikely. It is more conceivable that
this mode is instead a combination of an Amide ll vibration of the peptide
backbone and a weakly intense asymmetric stretching mode of the heme of

nearly the same frequency that is overlaid by the large Amide ll mode.
Additionally, a peak around 1634 cm'1 present in the reduced form of all three
cco samples was assigned to a symmetric stretch of an arginine side chain [54]
and in a recent report, it has been assigned to the 8(NH2) mode of asparagine
[60]. However, the 15N labeling experiment revealed no shift of this peak,

indicating that it is not attributable to a nitrogen-containing vibrational mode.

172

Both of these modes would be very sensitive to deuteration, with expected shifts

of greater than 10 cm'1, which are not observed [46]. An alternate explanation is

that the reduced peak around 1634 cm’1 is due to an Amide l backbone mode.

6.3 Discussion

6.3.1 Carbonyl Region

The FTIR carbonyl region (1750 - 1700 cm'1) of cco reveals a number of

changes arising from protonated Asp and Glu carboxylic acid residues.

Figure 6-9 shows the spectral expansion of some of the potential carbonyl

vibrations in cco spectra. The R. sphaeroides spectra obtained in H20 (solid
line) and D20 (dashed line) are presented in Figure 6-9A and the H20 (solid line)
and D20 (dashed line) bovine spectra in Figure 6-98.

A difference signal at 1746/1735 cm‘1 is observed in the bacterial H20
spectrum that shifts to 1740/1730 cm"1 in DZO. A similar, deuterium-sensitive

peak is observed in the oxidized form of the bovine sample at 1748 cm"1 and the

corresponding reduced peak is absent. An additional positive peak at 1734 cm'1

is observed in the bovine spectrum. Small shifts between the two forms of the
oxidase are not surprising, and are indicative of subtle perturbations of the

environment surrounding the carbonyl group. The reversal of a mode, such as
that observed for the peak near 1735 cm'1, can be explained as the contribution

of an additional Asp or Glu residue or a change in the extinction coefficient as a

173

result of a different environment.

Figure 6-9 Deuterium Exchange Effects in the Carbonyl Region

Electrochemical oxidized minus reduced FTIR difference spectra carbonyl region

(1760 - 1715 cm”) of cco from (A) Rhodobacter sphaeroides and (8) bovine heart. The
spectra of the global deuterium labeled enzyme (dashed line) are overlaid with the
unlabeled enzyme (solid line). To aid the comparison the spectra from the deuterated
samples were normalized to that of the non-deuterated samples based on the intensity

of the 1540 cm'1 band. Spectra were recorded at 10 0C using Ban windows and a
10 pm pathlength. Potentials of +0. 50 V and -0. 50 V (vs. Ag/AgCI) were applied to
obtain the oxidized and reduced states, respectively.

 

-1740

~11746

o
’’’’’

 

..1734 - .."".,."'..

...U 1748
1741

 

 

 

 

1755 1745 1735 _1 1725
Frequency (cm )

Hellwig and coworkers made the first tentative assignment of an amino
acid mode in the carbonyl region of a cco FTIR electrochemical difference

spectrum. Based on mutagenesis experiments, this residue, Glu 286, which is

174

situated at the end of the D-channel proton pathway, has been assigned at

1746 cm'1 in the oxidized enzyme and to a neighboring counter band at
1734 cm'1 in the reduced enzyme [40]. A similar difference pattern at

1745/1735 cm'1 (oxidized/reduced) has been identified in a related ubiquinol

oxidase, cytochrome bo3 oxidase, from Escherichia coli by Liibben and

coworkers [38].
The difference bands at 1745/1735 cm’1 were lacking in studies of site-

directed mutant samples in which the glutamic acid residue was replaced by an
aspartic acid or a glutamine at position 286, indicating they were clearly related

to Glu 286 [38]. In photolysis-induced difference FTIR studies of bovine cco,

Rich and coworkers observed a 1749/1741 cm’1 feature in the CO photolysis
spectrum and a related 1752/1740 cm'1 feature in the CN photolysis spectrum,

which also contained features at 1725/1715 and 1708/1698 cm’1 [36]. As

mentioned earlier, it is conceivable that the two different approaches (photolysis
and electrochemistry) result in slightly different enzyme structures, which are
manifested as variations in the amino acid environment, thus yielding different
peak frequencies.

The question of assignment seems simple, as all three oxidase forms
contain a conserved glutamic acid residue (Glu 286), which has been proposed
to play an integral role in the proton pumping process and is, therefore, linked to
the electron transfer steps. However, the task of definitively assigning these

features to Glu 286 has yet to be resolved. The reported deuterium-induced

175

shifts are consistent with assignment to carboxylic acid species, although, in
many cases, the magnitudes of the changes are small, if observed at all [36, 40,
54, 64]. The complexities observed in this region may be attributed to multiple
carboxylic acids, or the same residue in different states. However, all authors
confidently maintain that the conserved Glu 286 is a major contributor in both
ligand and redox difference FTIR spectra. Questions remain regarding whether
Glu 286 exists in several conformations that are affected differently by H/D
exchange, and in identifying other residues that would contribute carbonyl
stretching modes in this region. Likely candidates include heme propionic acids
and carboxylic acids associated with the magnesium site above heme a, all of
which are in close proximity to the redox-active metal centers. However, more
remote residues may be considered, especially those implicated in the proton
pumping function, such as those associated with the D channel [4, 70] or the

conforrnationally flexible Asp 51 [5].

6.3.2 Histidine Modes

Two technical improvements over previously published methods have
extended the accessible frequency range to include the region below 1200 cm'1.
First, using protein solutions prepared in HEPES instead of phosphate buffer

removed the large phosphate bands around 1100 cm'1 that were obscuring the
spectrum in this region. Secondly, BaF2 windows were implemented instead of
Can, which increased the accessible range from 1100 to 800 cm'1. Amino acid

modes expected to be present in the 1200 - 1000 cm‘1 region include the C-N

176

stretch of histidine, which has been proposed to be an important residue in the
catalytic function of the enzyme. Bioinorganic complexes with histidine ligands

have been reported to exhibit characteristic FTIR signals of the imidazole ring

around 1100 om‘1 [71-78].

Figure 6-10 Closer Examination of a Possible Histidine Mode

Electrochemical oxidized minus reduced FTIR difference spectra (1150 - 1050 cm'1) of
cco from (A) Rhodobacter sphaeroides and (B) bovine heart. The spectra of the global
15N labeled enzyme (dashed line) and the global deuterium labeled enzyme (dashed-
dotted line) are shown with the unlabeled enzyme (solid line). To aid in the comparison,
the spectra from the deuterated samples were normalized to that of the non-deuterated
samples based on the intensity of the 1540 cm’1 band. Spectra were recorded at 10 0C

using BaFg windows and a 10 pm pathlength. Potentials of +0.50 V and —0.50 V
(vs. Ag/AgCI) were applied to obtain the oxidized and reduced states, respectively.

 

 

 

 

 

1140 1120 1100 1080 1060
Frequency (cm-i)

177

Figure 6-7 depicts a 15N sensitive mode around 1100 cm‘1 in the
R. sphaeroides sample. Figure 6-10 shows a magnified view of the region from

1150 - 1050 cm’1 for the mammalian cco samples (H20 and 020) compared to

the three bacterial samples (H20, D20 and 15N). There is a slight (4 cm'1)

difference in the position of this mode between the bovine and R. sphaeroides
samples. This is not surprising, as similar results (i.e., small frequency
discrepancies) were obtained in this work and others [54, 58], for spectral
comparisons of different types of cco. Therefore, this small difference is
attributed to slight changes in the environment of the structural element giving
rise to this mode. More important are the shifts due to isotope exchange, as they
provide insight into the source of this mode.

Noguchi and coworkers have performed similar isotope exchange
experiments in the photosystem II enzyme and model compounds of the histidine
side chain (4- and 5-methyl imidazoles), in an effort to identify histidine modes

[73, 74]. The results from this work are consistent with their findings, in that FTIR

measurements revealed a mode around 1100 cm"1 that undergoes a 1 cm'1

downshift upon deuteration and a 10 cm'1 downshift upon 15N labeling.

Therefore, this mode is attributed to a C-N stretching mode of a histidine

imidazole ring. In a study of histidine modes, Noguchi’s group determined that

1

this C-N stretching mode around 1100 cm' was a useful IR marker band of the

protonation state of the histidine imidazole ring [73].

178

Figure 6-11 Structures of the Two Neutral Histidine Tautomers

Representative structures and atom numbering for the two neutral forms of 4-methy/
imidazole, a histidine side chain analog. (A) The Na—protonated or 4-methyl imidazole
form. (8) The Nt‘iprotonated or 5—methy/ imidazole form. (C) Structure of a metal-bound
N6-protonated or 5-methyl imidazole form. Frequency data is provided based on
vibrational spectroscopic (resonance Raman and F T/R) studies and OF T calculations for
the HON) mode, and from resonance Raman studies and DFT calculations for the
t(MN) mode.

 

(A) (B)

-

 

HN 5 8
V"
4-Melm 5-Melm
Ns protonated His N6 protonated His

(C) \‘(CC) ~ 1600 cm'1

( v(CN) ~ 1100 om'1

E

HN8 Err.

V

(Metal
v(MN) ~ 250 on1

 

 

 

The imidazole side chain of the histidine residue can exist as either of two
neutral tautomers, depending on the protonation state of imidazole N. The model
compounds 4- and 5-methyl imidazoles correspond to the Ne and N6 protonated
forms of DL-Histidine. The structures of these molecules are shown in
Figure 6-11. The two different forms of the side chain exhibit unique FTIR

spectra. The spectrum of the 4-Melm form consists of a characteristic C-N

stretching mode centered at 1087 - 1093 cm'1 depending on temperature, which

179

upshifts by 5 - 10 cm'1 upon N-deuteration. The 5-Melm form, however, exhibits
a C-N stretching mode at slightly higher frequency (1103 - 1106 cm“) and is

relatively unaffected by deuteration (-1 cm'1) [73]. Studies of FePP(4-Melm)2
[76], cytochrome b559 of PSII [76] and non-heme iron of PSII [75, 77, 78]
demonstrate similar results for cases where the imidazole ring is coordinated to a
metal. The data from these and related studies are presented in Tables 6-3
and 6-4.

The observed N-deuteration effect can be explained by normal mode
analysis. Majoube and coworkers have assigned the bands around 1100 cm’1 of
4- and 5-Melm to the C5-N1 stretching vibration [79]. From the structures
presented in Figure 6-11 it is clear that this vibration will be sensitive to
deuteration at the N1 site in the Ne protonated form, but will be rather unaffected
by deuteration at the N3 in the N6 protonated form. It is proposed that the
upshifting observed in the C5-N1 band upon N1 deuteration results from
decoupling of the mixed N-H bending vibration upon deuteration [79].

The C4C5 stretching vibration has also been used as a vibrational marker
of the protonation state and metal binding status of histidine and 4-Melm
molecules [71, 79-84]. Tables 6-5 and 6-6 list the C405 frequencies found in the
literature for metal-free and metal-bound histidine and 4-Melm, respectively. The
values presented are those corresponding to the N6 protonated neutral forms of

4-MeIm and histidine, as they are most relevant to the present discussion.

180

Table 6—3 C5N1 Stretching Modes for Histidine and 4-Melm

 

 

Frequency (cm'1) Method Reference
L-Hisa 1106 RR [71]
DL-His (H/D)b 1093/1098 FTIR [73]
4-Me|m (H/D)C 1087/1097 FTIR [73]
5-Me.m (H/D)C 1103/1104 FTIR [73]
4-Melm (HID) 1075/1079 DFT calc. [80]
5-Melm (H/D) 1126/1125 DFT calc. [80]
BA.a_Hisd 1 105 RR [72]

 

aL-histidine in water (pH 11.02, 41 OC). bCrystalline DL-histidine measured as a KBr pellet. CIn

aqueous solution at room temperature, 4-Melm and 5-Melm in tautomeric equilibrium. quueous
solution of B-alanyI-L-histidine (pH 8.0). *‘5-Melm is equivalent to N6 protonated 4-Melm.

Table 6—4 C5N1 Stretching Modes for Metal-Bound Histidine and 4-Melm

 

 

Frequency (cm'1) Method Reference
Co(lI)-(4-Melm)GClza 1112 RR [71]
Cu(lI)-(BAla-His)b 1112 RR [72]
Mn (psu OEC)C 1114-1113 FTIR [73]
Fe(II)/Fe(|||) (Ps||)d 1111-1110/1103-1102 FTIR [75, 77, 78]
Fe(lll)PP(4-Melm)29 1103 FTIR [76]
Fe(lll) (cyt b559)’ 1104 FTIR I76]
Zn(II)-(4-Melm)g 1108 DFT calc. [81]

 

aCrystalline hexakis(4-methylimidazole)cobalt(Il)ch|oride. bCrystalline (b-alanyl-L-histidino)
copper(ll)dihydrate. CThe Mn cluster of photosystem II. The non-heme iron of photosystem II.
eBis(4-methylimidazole) complex of iron protoporphyrin in water. fOxidized form of cytochrome

b559 of photosystem II. g(4-methylimidazole) zinc(ll) trihydrate complex with N1 metal
coordination and N3 protonated. *‘5-Melm is equivalent to N6 protonated 4-Melm.

181

Table 6—5 C4C5 Stretching Modes for N6 Protonated Histidine and 4-Melm

 

 

Frequency (cm'1) Method Reference
L-Hisa 1585 RR [71]
DL-Hisb 1588 RR [72]
5-Melmc 1594 FTIR [80]
5-Me|mc 1593 RR [80]
5-Melm 1593 DFT calc. [80]
His (zinc finger)d 1585 RR [72]
liAla-His" 1588 RR [72]

 

aL-histidine in water (pH 11.02, 41 OC). bAqueous solution of DL-histidine (pH 8.0). CIn aqueous
solution at room temperature, 4-Melm and 5-Melm in tautomeric equilibrium. quueous solution

of a 27mer zinc finger peptide (pH 7.8). eAqueous solution of B-alanyl-L-histidine (pH 8.0).
"5 Melm is equivalent to N6 protonated 4-Melm.

Table 6—6 C405 Stretching Modes for Metal-Bound N6 Protonated Histidine
and 4-Melm

 

 

Frequency (cm'1) Method Reference
Co(lI)-(4-Melm)eClza 1593 RR [71]
Zn(ll)-(4-Melm)b 1615 DFT calc. [81]
Zn-ABC 1604 RR [72]
Zn-(zinc finger)d 1505 RR [72]
Cu(II)-(8AIa-His)e 1594 RR [72]

 

aCrystalline hexakis(4-methylimidazole)cobaIt(ll)chloride. 6(4-methylimidazole) zinc(ll) trihydrate
complex with N1 metal coordination and N3 protonated. CInsoluble precipitate from a mixture of
Zn(ll) and the human amyloid B-peptide. d27mer Zn finger peptide in the presence of ZnClz

(6.25 mM). eCrystalline (b-alanyl-L-histidino) copper(ll)dihydrate. **5-Me|m is equivalent to N6
protonated 4-Melm.

182

The N6 protonated form contributes a C4C5 band at 1594 cm'1 to the
aqueous solution FTIR spectrum of 4-MeIm [80, 84]. Free histidine in solution
shows a band in the range of 1588 - 1580 cm’1 that has been attributed to the N6
protonated form [80, 84]. In both cases, the modes due to the N6 protonated
form occur at higher frequency (15 - 20 cm'1) than those due to the Na

protonated form [80, 84]. Raman spectroscopic [82] and DFT computational [81]

studies report that metal binding to histidine upshifts the C4C5 frequency by as
much as 21 cm'1.

The C4C5 stretching mode is coupled to the ring NH bending mode,

suggesting it would be sensitive to global 15N labeling. Our work shows evidence
of 15N sensitivity in the region from 1605 - 1615 cm'1, which is where one would

expect to observe a band due to the C405 stretching mode of a metal-bound N6
protonated histidine. However, this region also potentially contains a number of
N-containing modes from the heme moieties. Once again, the complex nature of

these spectra requires one to obtain additional information before definitively
assigning a mode in the 1605 - 1615 cm'1 region to the C405 stretching vibration
of a metal-bound N6 protonated histidine. Proposed experiments to measure the
FTIR spectra of R. sphaeroides samples containing specifically labeled 15N
histidine residues would assist in such an assignment.

Global 15N labeling of cco of R. sphaeroides together with deuterium
exchange has identified a histidine signal in oxidized minus reduced FTIR

spectra. Examination of Figure 6-10 revealed a positive band at 1105 cm'1,

183

which was downshifted to 1095 cm"1 upon 15N exchange and to 1104 cm'1 upon

deuteration, and was assigned to the C-N stretching mode of a histidine side

chain. The structure of the histidine mode was determined using the above
criteria. The relatively high frequency at 1105 cm'1 and the observed insensitivity
to deuteration supported the assignment that the imidazole ring of this histidine

had a N6 protonated form.

We acknowledge that in the above experiments the N atoms were globally
labeled with 15N, therefore, the possibility of shifted bands arising from structural
moieties including N but not imidazole cannot be excluded. Definitive
assignment of the mode at 1105 cm'1 to a C-N stretching vibration of a histidine
side chain can only be achieved by a further experiment to measure the FTIR
spectra of R. sphaeroides samples containing specifically labeled 15N histidine
residues. For our preliminary analysis to be correct, we would expect to see the

same result observed herein for the 1150-1050 cm"1 region and no shift

observed around 1550 cm'1 (Amide ll). Investigation of the high frequency N-H

stretching modes of the histidine side chain will provide further information

regarding the H bonding status of the imidazole ring. When hydrogen bonded,
the N-H stretching mode occurs as a broad band around 2800 cm'1 of weak
intensity, but without hydrogen bonding, this mode is observed as a sharp band
at 3500 cm'1 [79, 85-88]. Frequency regions above 2850 cm'1 are difficult to

access, as strong water absorbance causes very high noise levels.

184

Studies of site-directed mutants of key histidine residues will also aid in

the assignment of this mode to a specific amino acid. The histidine ligands of

Cue, residues numbered His 333, His 334, and His 284, are believed to be

important in the catalysis. Attempts to obtain site-directed mutants for these
residues have been unsuccessful (i.e., the organism was severely disrupted and,

therefore, did not grow) with the exception of His334Asn.

6.4 Implications

Histidine is one of the most common ligands to metal ions in
metalloproteins [89, 90]. This is due to the characteristics of the imidazole ring,
which comprises the histidine side chain. The two nitrogens of the imidazole ring
participate in metal binding, as well as protonation and deprotonation reactions.
In the neutral form, either the Ne or N6 is protonated, and the other is free to act
as a ligand for metal coordination. The metal-free states and the metal-bound
states provide at least four possible forms that neutral histidine is capable of
adopting. Not discussed here are several other charged states of histidine, both
coordinated and not. With so many possible structures, it is not surprising that
histidine often functions as a versatile ligand that is a key residue in a wide
variety of catalytic reactions. For example, histidine is an integral part of the
copper binding site of human serum albumin [91, 92], it is required for the
formation of the zinc finger peptide nucleic acid-binding structural motif [93], it is

involved in the reactions of enzymes such as the Zn, Cu superoxide dismutase

enzyme [94], the photosystem II enzyme [73], and, as a proposed ligand to Cue,

185

it has been implicated in the proton translocation mechanism of the heme-copper
oxidases [27]. Thus, it is important to know the coordination and protonation
state of a histidine ligand, in order to elucidate the reaction mechanisms of many

metalloenzymes.

Table 6—7 Metal-Histidine Stretching Modes

 

 

Frequency (cm'1) Method Reference
Cu—N(His) 250 RR [95, 96]
Fe-N(His) 220-250 RR [97, 98]
Zn—N(His) 260 DFT calc. [81]

 

As discussed above, experiments in the mid-IR have been successful at
determining the protonation state of histidine. It is clear from Tables 6-3 and 6-4,
however, that the histidine C5N1 vibrational mode does not provide a clear
distinction between metal—free and metal-bound forms of histidine. Tables 6-5
and 6-6 indicated that such information might be obtained from the C4C5
stretching mode of histidine; however, at this time, it is not possible to assign
such a mode in our spectra. Direct information regarding the metal ligation state
of histidine is contained in His-metal modes, which occur at frequencies much
less than 1000 cm'1, in the far-IR region. Table 6-7 lists His-metal modes
observed in resonance Raman studies and normal mode calculations of
bioinorganic model compounds.

As indicated in Table 6-7, experimental access to low-frequency vibrations

of metalloproteins has been limited to resonance Raman spectroscopy methods.

186

Resonance Raman spectroscopy is a powerful technique for studying many
biological systems, especially cco. However, this technique is not capable of

addressing the questions at all active sites in metalloproteins. For example, the

Cue site of cco is “invisible" to resonance Raman methods due to interference

attributed to the nearby hemes. Therefore, low-frequency electrochemical
difference data obtained by IR for this metal center would be used to provide

valuable information for answering mechanistic questions involving the role of

Cue, in the catalytic cycle. As indicated in Chapter 5 (Table 5-1), this method

would be useful to determine and directly monitor substrate or metal-ligand
modes such as His-Cue or HO-CuB, that can provide insight into the chemical

details of cco’s catalytic function.

6.5 Conclusions

High quality, electrochemical difference mid-frequency (1800 - 1000 cm'1)

FTIR spectra were obtained for several cco samples. A comparison of a

bacterial and a mammalian form of the enzyme, as well as isotope exchange
experiments involving D20 and 15N were carried out in an effort to simplify mode

assignment in the complex spectra. Evaluation of the measured spectra for
oxidases from bovine heart and R. sphaeroides sources revealed variations, due
to both subtle and significant changes in enzyme active site structure. Many of
the observed variations were consistent with a published comparison of bovine
heart enzyme with that of a bacterial form from the P. denitrificans organism [54].

Although the oxidase structure and mechanistic function is highly conserved

187

between the bovine heart and R. sphaeroides forms of the enzyme, the observed
spectroscopic deviations provide an indication of the subtle interactions the
amino acid residues provide in facilitating electron transfer to the oxidase
cofactors.

Two important insights were gained through isotope global-exchange

studies via solvent exchange with DZO buffers and organism (R. sphaeroides)

growth in 15N media. First, these experiments revealed that the H/D exchange of

the R. sphaeroides enzyme core was kinetically limited beyond three days, thus,

it was very shielded from exchange; and, second, the results aided in the
assignment of a mode at 1105 cm'1 to a histidine, which may be a residue that is

integral in the cco proton pumping mechanism.

The fact that the H/D exchange data show a lack of shifts observed for the
Amide ll region (1580 -1520 cm"), despite expected shifts of up to 100 cm'1

upon deuteration, has led others to believe that the Amide II modes do not

contribute to electrochemical difference FTIR spectra of cco [39]. However, this
work clearly shows that upon 15N labeling, significant changes occur, as

expected, in the Amide Il region, indicating strong contributions from the

N-containing Amide ll modes to this region. Therefore, it was concluded that the
Amide modes do contribute to the spectra in the region from 1580 to 1520 cm'1,

and that the small observed shifts are instead due to a lack of solvent
accessibility to the protein backbone caused by a hydrophobic environment or a

tightly ordered structure (a helix or 8 sheet). Without further isotope data for the

188

bovine sample, it was assumed that the similarly small shifts observed upon H/D
exchange are due to a similar shielding process of the exchangeable core
protons of the bovine oxidase.

The novel 15N global labeling experiments of this enzyme provided further
insight into several previously assigned modes. Most importantly, it also
facilitated in the identification of an unidentified histidine amino acid side chain.
This particular residue is one that is linked to the electron transfer reaction of cco,

and is potentially important in the proton pumping mechanism. Questions remain
regarding whether or not this histidine is a direct ligand to the CUB of the active
site. Further experiments are required involving 15His specifically labeled

R. sphaeroides enzyme, in order to unequivocally assign this mode to a histidine
side chain. Site-directed mutagenesis will aid in assigning this mode to a specific

His of the protein structure, and low-frequency FTIR measurements of the cco
enzyme will provide answers to questions involving its ligation to the Cue, and its

involvement in the enzymatic mechanism.

189

6.6

[1.]

[2.]

[3-1

[4-1

[5-1

[6-1

[7-1

[8-1

[9-1

[10.]

[11.]

[12.]

References

S. Ferguson-Miller and G. T. Babcock Chem. Rev. 1996, 96, 2889-2907.

H. Michel, J. Behr, A. Harrenga and A. Kannt Annual Review of
Biophysics and Biomolecular Structure 1998, 27, 329-356.

M. Wikstrdm Nature 1989, 338, 776-778.

T. Tsukihara, H. Aoyama, E. Yamashita, T. Tomizaki, H. Yamaguchi, K.
Shinzawaltoh, R. Nakashima, R. Yaono and S. Yoshikawa Science 1996,
272, 1136-1144.

S. Yoshikawa, K. Shinzawa-Itoh, R. Nakashima, R. Yaono, E. Yamashita,
N. Inoue, M. Yao, M. J. Fei, C. P. Libeu, T. Mizushima, H. Yamaguchi, T.
Tomizaki and T. Tsukihara Science 1998, 280, 1723-1729.

S. lwata, C. Ostermeier, B. Ludwig and H. Michel Nature 1995, 376, 660-
669.

C. Ostermeier, A. Harrenga, U. Ermler and H. Michel Proc. Natl. Acad.
Sci. U. S. A. 1997, 94, 10547-10553.

J. L. Anderson, T. Kuwana and C. R. Hartzell Biochemistry 1976, 15,
3847-3855.

N. A. Schroedl and C. R. Hartzell Biochemistry 1977, 16, 4961-4965.

J. K. Cullison, F. M. Hawkridge, N. Nakashima and S. Yoshikawa
Langmuir 1994, 10, 877-882.

J. H. Li, G. J. Cheng and S. J. Dong J. Electroanal. Chem. 1996, 416, 97-
104.

A. S. Haas, D. L. Pilloud, K. S. Reddy, G. T. Babcock, C. C. Moser, J. K.
Blasie and P. L. Dutton J. Phys. Chem. B 2001, 105, 11351-11362.

190

[13.]

[14.]

[15.]

[16.]

[17.]

[18.]

[19.]

[20.]

[21.]

[22.]

[23.]

[24.]

[25.]

[26.]

T. Kitagawa and T. Ogura In Progress in Inorganic Chemistry, Vol45,
1997; Vol. 45, 431-479.

M. Wikstrém Biochim. Biophys. Acta 1998, 1365, 185-192.

P. Adelroth, H. Sigurdson, S. Hallén and P. Brzezinski Proc. Natl. Acad.
Sci. U. S. A. 1996, 93, 12292-12297.

A. A. Konstantinov, S. Siletsky, D. Mitchell, A. Kaulen and R. B. Gennis
Proc. Natl. Acad. Sci. U. S. A. 1997, 94, 9085-9090.

T. V. Vygodina, C. Pecoraro, D. Mitchell, R. Gennis and A. A.
Konstantinov Biochemistry 1998, 37, 3053-3061.

D. A. Mills, L. Florens, C. Hiser, J. Qian and S. Ferguson-Miller Biochim.
Biophys. Acta 2000, 1458, 180-187.

S. Riistama, A. Puustinen, A. Garcia-Horsman, S. lwata, H. Michel and M.
Wikstrbm Biochim. Biophys. Acta 1996, 1275, 1-4.

I. Hofacker and K. Schulten Proteins 1998, 30, 100-107.

S. Riistama, A. Puustinen, M. l. Verkhovsky, J. E. Morgan and M.
Wikstrom Biochemistry 2000, 39, 6365-6372.

B. Schmidt, L. Florens, L. Qin, J. McCracken and S. Ferguson-Miller
Biophys. J. 2002, 82, 1391.

L. Florens, B. Schmidt, J. McCracken and S. Ferguson-Miller Biochemistry
2001, 40, 7491-7497.

Y. Georgievskii, E. S. Medvedev and A. A. Stuchebrukhov Biophys. J.
2002, 82, 2833-2846.

H. Michel Proc. Natl. Acad. Sci. U. S. A. 1998, 95, 12819-12824.

H. Michel Biochemistry 1999, 38, 15129—15140.

191

 

[27.] M. Wikstrbm Biochim. Biophys. Acta 2000, 1458, 188-198.
[28.] M. Wikstrbm Biochemistry 2000, 39, 3515-3519.

[29.] S. Park, L. P. Pan, S. I. Chan and J. O. Alben Biophys. J. 1996, 71, 1036-
1047.

[30.] A. Puustinen, J. A. Bailey, R. B. Dyer, S. L. Mecklenburg, M. Wikstrom _
and W. H. Woodruff Biochemistry 1997, 36, 13195-13200. E

[31.] E. D. Dodson, X. J. Zhao, W. S. Caughey and C. M. Elliott Biochemistry
1996, 35, 444-452.

 

[32.] Y. J. Zhen, B. Schmidt, U. G. Kang, W. Antholine and S. Ferguson-Miller i
Biochemistry 2002, 41, 2288-2297.

[33.] D. Heitbrink, H. Sigurdson, C. Bolwien, P. Brzezinski and J. Heberle
Biophys. J. 2002, 82, 1-10.

[34.] J. A. Bailey, F. L. Tomson, S. L. Mecklenburg, G. M. MacDonald, A.
Katsonouri, A. Puustinen, R. B. Gennis, W. H. Woodruff and R. B. Dyer
Biochemistry 2002, 41, 2675-2683.

[35.] B. Rost, J. Behr, P. Hellwig, O.-M. H. Richter, B. Ludwig, H. Michel and W.
Mantele Biochemistry 1999, 38, 7565-7571.

[36.] P. R. Rich and J. Breton Biochemistry 2001, 40, 6441-6449.
[37.] P. R. Rich and J. Breton Biochemistry 2002, 41 , 967-973.
[38.] M. Li'lbben and K. GenNert FEBS Lett. 1996, 397, 303-307.

[39.] P. Hellwig, S. Grzybek, J. Behr, B. Ludwig, H. Michel and W. Mantele
Biochemistry 1999, 38, 1685-1694.

192

[40.]

[41.]

[42.]

[43.]

[44.]

[45.]

[46.]

[47.]

[48.]

[49.]

[50.]

[51.]

[52.]

P. Hellwig, J. Behr, C. Ostermeier, O.-M. H. Richter, U. Pfitzner, A.
Odenwald, B. Ludwig, H. Michel and W. Mantele Biochemistry 1998, 37,
7390-7399.

P. Hellwig, B. Rost, U. Kaiser, C. Ostermeier, H. Michel and W. Mantele
FEBS Lett. 1996, 385, 53-57.

J. Behr, P. Hellwig, W. Mantele and H. Michel Biochemistry 1998, 37,
7400-7406.

S. Horie and M. Morrison J. Biol. Chem. 1963, 238, 2859-2865.

J. P. Hosler, J. Fetter, M. M. J. Tecklenburg, M. Espe, C. Lerma and S.
Ferguson-Miller J. Biol. Chem. 1992, 267, 24264-24272.

C. Zscherp and A. Barth Biochemistry 2001, 40, 1875-1883.

A. Barth Prog. Biophys. Mol. Biol. 2000, 74, 141-173.

P. V. Argade, Y. C. Ching and D. L. Rousseau Biophys. J. 1986, 50, 613-
620.

S. Choi, T. G. Spiro, K. C. Langry, K. M. Smith, D. L. Budd and G. N.
Lamar J. Am. Chem. Soc. 1982, 104, 4345-4351.

S. Choi, J. J. Lee, Y. H. Wei and T. G. Spiro J. Am. Chem. Soc. 1983,
105, 3692-3707.

S. Choi, T. G. Spiro, K. C. Langry and K. M. Smith J. Am. Chem. Soc.
1982, 104, 4337-4344.

B. Stuart Biological Applications of Infrared Spectroscopy; John Wiley &
Sons, New York, 1997.

J. L. Arrondo, A. Muga, J. Castresana and F. M. Goni Prog. Biophys. Mol.
Biol. 1993, 59, 23-56.

193

[53.]

[54.]

[55.]

[56.]

[57.]

[58.]

[59.]

[60.]

[61.]

[62.]

[63.]

[54.]

[65.]

J. L. R. Arrondo and F. M. Goni Prog. Biophys. Mol. Biol. 1999, 72, 367-
405.

P. Hellwig, T. Soulimane, G. Buse and W. Mantele FEBS Lett. 1999, 458,
83-86.

G. E. Heibel, P. Hildebrandt, B. Ludwig, P. Steinrucke, T. Soulimane and
G. Buse Biochemistry 1993, 32, 10866-10877.

P. Hellwig, T. Mogi, F. L. Tomson, R. B. Gennis, J. lwata, H. Miyoshi and
W. Mantele Biochemistry 1999, 38, 14683-14689.

A. Kannt, U. Pfitzner, M. Ruitenberg, P. Hellwig, B. Ludwig, W. Mantele, K.
Fendler and H. Michel J. Biol. Chem. 1999, 274, 37974-37981.

P. Hellwig, C. Ostermeier, H. Michel, B. Ludwig and W. Mantele Biochim.
Biophys. Acta 1998, 1409, 107-112.

T. Noguchi, T. A. Ono and Y. Inoue Biochim. Biophys. Acta 1995, 1232,
59-66.

P. Hellwig, U. Pfitzner, J. Behr, B. Rost, R. P. Pesavento, W. von Donk, R.
B. Gennis, H. Michel, B. Ludwig and W. Mantele Biochemistry 2002, 41,
9116-9125.

R. B. Gennis Biochim. Biophys. Acta 1998, 1365, 241-248.

D. A. Proshlyakov, M. A. Pressler, C. DeMaso, J. F. Leykam, D. L. DeWitt
and G. T. Babcock Science 2000, 290, 1588-1591.

D. A. Proshlyakov, M. A. Pressler and G. T. Babcock Proc. Natl. Acad.
Sci. U. S. A. 1998, 95, 8020-8025.

M. Linben, A. Prutsch, B. Mamat and K. Genrvert Biochemistry 1999, 38,
2048-2056.

Y. Maeda, T. Nakamura and I. lkeda Macromolecules 2001, 34, 8246-
8251.

194

[66.]

[67.]

[68.]

[69.]

[70.]

[71.]

[72.]

[73.]

[74.]

[75.]

[76.]

[77.]

[78.]

[79.]

T. N. Earnest, J. Herzfeld and K. J. Rothschild Biophys. J. 1990, 58, 1539—
1546.

N. W. Downer, T. J. Bruchman and J. H. Hazzard Biophys. J. 1985, 47,
A105—A105.

N. W. Downer, T. J. Bruchman and J. H. Hazzard J. Biol. Chem. 1986,
261 , 3640-3647.

T. Kluge, J. Olejnik, L. Smilowitz and K. J. Rothschild Biochemistry 1998,
37, 10279-10285.

J. R. Fetter, J. Qian, J. Shapleigh, J. W. Thomas, A. Garcia-Horsman, E.
Schmidt, J. Hosler, G. T. Babcock, R. B. Gennis and S. Ferguson-Miller
Proc. Natl. Acad. Sci. U. S. A. 1995, 92, 1604-1608.

M. Itabashi and K. ltoh Chem. Lett. 1979, 1331-1334.

T. Miura, T. Satoh, A. Hori-i and H. J. Takeuchi J. Raman Spectrosc.
1998, 29, 41-47.

T. Noguchi, Y. Inoue and X. S. Tang Biochemistry 1999, 38, 10187-10195.

T. Noguchi, Y. Inoue and X. S. Tang Biochemistry 1999, 38, 399-403.

T. Noguchi and Y. Inoue J. Biochem. 1995, 118, 9-12.

C. Berthomieu, A. Boussac, W. Mantele, J. Breton and E. Nabedryk
Biochemistry 1992, 31, 11460-11471.

R. Hienerwadel and C. Berthomieu Biochemistry 1995, 34, 16288-16297.

C. Berthomieu and R. Hienerwadel Biochemistry 2001, 40, 4044-4052.

M. Majoube, P. Millie and G. Vergoten J. Mol. Struct. 1995, 344, 21-36.

195

[80.]

[81.]

[82.]

[83.]

[84.]

[85.]

[86.]

[87.]

[88.]

[891

[90.]

[91.]

[92.]

[93.]

[94.]

K. Hasegawa, T. Ono and T. Noguchi J. Phys. Chem. B 2000, 104, 4253-
4265.

K. Hasegawa, T. Ono and T. Noguchi J. Phys. Chem. A 2002, 106, 3377-
3390.

S. Hashimoto, K. Ono and H. Takeuchi J. Raman Spectrosc. 1998, 29,
969-975.

A. Toyama, K. Ono, S. Hashimoto and H. Takeuchi J. Phys. Chem. A
2002, 106, 3403-3412.

I. Ashikawa and K. ltoh Biopo/ymers 1979, 18, 1859-1876.

H. Wolff and E. Wolff Spectrochim. Acta A 1971, 27, 2109-2118.

C. Perchard and A. Novak J. Chem. Phys. 1968, 48, 3079-3084.

A. M. Bellocq, C. Perchard, A. Novak and M. L. Josien J. Chim. Phys.
1965, 62, 1334-1343.

H. Zimmermann Z. E/ektrochem. 1961, 65, 821-840.

R. J. Sundberg and R. B. Martin Chem. Rev. 1974, 74, 471-517.

L. Regan Annu. Rev. Biophys. Biomolec. Struct. 1993, 22, 257-281.

T. Peters, Jr. and F. A. Blumenstock J. Biol. Chem. 1967, 242, 1574-1578.

N. Camerman, J. K. Fawcett, T. P. A. Kruck, B. Sarkar and A. Camerman
J. Am. Chem. Soc. 1978, 100, 2690-2693.

J. M. Berg and Y. (3. Shi Science 1996, 271, 1081-1085.

D. J. Wang, X. J. Zhao, M. Vargek and T. G. Spiro J. Am. Chem. Soc.
2000, 122, 2193-2199.

196

 

[95.] Thamann, P. Frank, L. J. Willis and T. M. Loehr Proc. Natl. Acad. Sci.

T. J.
U. S. A. 1982, 79, 6396-6400.

[96.] L. Nestor, J. A. Larrabee, G. Woolery, B. Reinhammar and T. G. Spiro
Biochemistry 1984, 23, 1084-1093.

[97.] K. Nagai, T. Kagimoto, A. Hayashi, F. Taketa and T. Kitagawa
Biochemistry 1983, 22, 1305-1311.

[98.] J. Teraoka and T. Kitagawa J. Biol. Chem. 1981, 256, 3969-3977.

197

 

CHAPTER 7

7 Spectroelectrochemical Studies Using a Novel
Optically Transparent Electrode: Boron-Doped
Diamond

7.1 Introduction

Diamond is something magnificent, the peerless "king of gems" that
glitters, dazzles, and symbolizes purity and strength. It is the optical properties of
diamond that have generated the awe surrounding this material. These
properties, present in both high-quality synthetic and naturally-occurring
diamond, include the highest values of reflectivity (0.17), index of
refraction (2.42), and dispersion (0.057) of any mineral, giving a well-cut diamond
stone a marvelous brilliance and fire [1, 2]. The exceptional scintillation
properties of this hard-to-find mineral have caught the eyes of the wealthy, the
powerful, and the general public for centuries. Aside from imparting beauty to a
gemstone, the Optical properties of diamond have been exploited in more
functional roles.

Gemologists routinely evaluate the quality of a diamond gemstone based
on its optical properties. Spectroscopists have coveted diamond for its superb
optical properties, finding it suitable for application as optical windows for IR and
UV/vis spectroscopic measurements and as optical components of high-power
lasers Additionally, its electronic, thermal, structural, and acoustic properties

have captured the imagination of electrochemists, physicists, and engineers.

198

Chapters 3 and 4 described applications of the electronic properties of boron-
doped diamond (BDD), a synthetic diamond material that is doped with boron
atoms to render it highly conductive. The focus of this chapter will be the
investigation of the combined electronic and optical properties of BDD for
spectroelectrochemical measurements.

The use of electrically conductive diamond as an optically transparent
electrode (OTE) is a new area of research [3-5]. BDD possesses attractive
qualities as both an electrode and an optically transparent material, making it an
obvious choice for utilization as an OTE in spectroelectrochemical
measurements. Diamond OTEs offer several advantages over other materials
including: (i) the possibility of transmission measurements from the near-UV into
the far-IR (0.25 -100 pm), (ii) a low background current, (iii) a wide working
potential window, (iv) stability in both aqueous and nonaqueous solution
environments during both cathodic and anodic polarization, and (v) resistance to
foufing.

An added advantage of BDD OTEs relies on the ease and versatility of
fabrication, as BDD thin-films have been grown on a variety of substrates
(e.g., CVD white diamond, quartz and undoped Si), and also exist as
freestanding disks. Each type of OTE possesses characteristics that can be
exploited for a specific application. This work is a product of our long-term goal
to develop BDD as an optically transparent electrode for electroanalytical
applications and to study biological electron transfer mechanisms. At present,

our group has investigated the detection capabilities of BDD OTEs in the UV/vis

199

 

for analytes such as chlorpromazine [5]. We are also working to develop BDD
OTEs for use in the far-IR detection of single metal-ligand modes (e.g. Cu-N) in
complex metalloenzymes, such as cytochrome c oxidase. Such
spectroelectrochemical studies of biomolecules will provide experimental
evidence for the various models that exist to describe their mechanism of action.

An important part of the material development involves manipulation and
control of the optical properties through adjustments in the deposition conditions.
The first-generation diamond OTEs developed by our multi-investigator team
demonstrated 40 - 80% transmissivity from the UV into the far infrared.
Optimization of the diamond OTEs is achieved through transmission
spectroelectrochemical measurements of both aqueous and nonaqueous redox
systems in the UV/vis, and IR regions of the electromagnetic spectrum.

There are many factors that influence the transmissivity of BDD. Many of

these can be controlled, to some extent, by manipulation of the diamond growth
conditions. Disruption of the lattice symmetry by spz-bonded nondiamond carbon

or introduction of chemical impurities (intentional or not), such as boron and
nitrogen, will decrease transmission, as each imperfection gives rise to its own
characteristic effect [1, 6-9]. This creates an interesting problem, as boron
doping is required to make the diamond films conductive. Scattering losses
contribute to the overall decreased light throughput via structural defects and
large grain size. Losses due to scattering become very important when the

defect or grain size approaches that of the incident wavelength [7]. The

200

thickness of the film affects the transmission, as transmission decreases with
increasing film thickness.

Work remains to completely understand the relationship between the
growth conditions and optical throughput. We continue to investigate the effect

of the boron doping level on electrode conductivity and transmission. Highly

 

doped films (B concentration~1020 cm'3) have been used for measuring P
apparent rate constants for ferrocene [10] and for cyclic voltammetric
measurements of a variety of aqueous systems [11, 12]. Such films are highly
conductive but not transparent. To achieve maximum light throughput, less L

conductive films, with lower doping levels (B concentration ~ 1018 cm'3), are used

for spectroelectrochemical measurements. Table 7-1 demonstrates the general
effect of several controllable growth parameters on the electrical and optical

properties of BDD OTEs.

Table 7—1 Effect of Selected Growth Parameters on the Optical and
Electronic Properties of Boron-Doped Diamond Optically Transparent
Electrodes

 

 

. . Conductivity
Growth Parameter OTE Property Transmrssron -1 -1
(0 cm )
micro vs. nano BDD T surface roughness i no change
1 growth time 1 film thickness I no change
1‘ conc B2H6 gas 1 B doping level i i

 

micro, microcrystalline; nano, nanocrystalline

We previously reported on the development of freestanding BDD for

transmission UV/vis spectroelectrochemical measurements. The electrode and

201

optical responses toward ferri/ferrocyanide and methyl viologen were used to
demonstrate the potential usefulness of this type of OTE [4]. We expand upon
that work, presently, by reporting the first measurements of the
spectroelectrochemical responsiveness of optically transparent diamond
electrodes toward a nonaqueous redox analyte, ferrocene. We also exploit the
IR optical transparency of BDD. Martin and Morrison recently demonstrated the
application of a BDD OTE for attenuated total reflection (ATR) IR measurements
[3]. Our work, however, demonstrates the utility of BDD for IR transmission
spectroelectrochemistry. Fabrication of the BDD OTEs was previously described
in Sections 2.1.3 (freestanding BDD disk) and 2.1.4 (BDD film on undoped Si).
The novel optically transparent thin-layer electrochemical (OTTLE) cell designs
incorporating these BDD OTEs for use in UV/vis and IR were discussed in
Sections 2.4.2 (UV/vis OTTLE cell), 2.4.3 (IR OTTLE cell -— Model I), and 2.4.4
(IR OTTLE cell - Model II). The electrochemical characterization of these thin-
layer cells, as well as their utility in spectroelectrochemical measurements of
ferrocene, ferrocyanide, and cytochrome c are reported in this chapter.

The main advantage of using a thin-layer cell design for
spectroelectrochemical measurements lies in the very short times required
(typically 20 - 120 s) for complete electrolysis. In order to exhaustively
electrolyze the bulk solution, as is desired in spectroelectrochemistry, a large
area (A) to volume (V) ratio is required. One method for achieving a large AN
ratio without convective mass transfer involves confining a very small volume

(i.e., a few uL) to a thin layer (2 - 250 pm) at the electrode surface.

202

According to the random-walk model of diffusion, the diffusion layer
thickness, x, is defined by

x: 20L

in which D is the diffusion coefficient (cmzls) and t is the experimental time (s)

[13]. Mass transfer within the cell can be neglected when the solution thickness,

I, is much less than the diffusion layer thickness, or

I<<JZDI.

Such thin-layer situations are characteristic of diffusionless electron transfer, and
special bulk electrolysis equations apply.

Thin-layer spectroelectrochemical studies are commonly employed using
potential sweep (e.g., cyclic voltammetry) and potential step methods for
controlling the applied potential. As was reported in Chapter 3, for a reversible
system, the ratio of the concentrations of the reduced, [R], and oxidized, [O],

forms in the thin-layer solution is controlled by the applied potential, E, according

to the Nernst equation,

5:5” —flln[yfl] .
nF [0] 4:0

In this expression, R is the molar gas constant (8.314 J/mol K), T is
temperature (K), n is the number of electrons transferred per mole, F is

Faraday’s constant (96,485 C/mol) and E°' is the formal reduction potential of

the couple.

Because the system is free from semi-infinite diffusional limitations,

different equations from those used for the diffusion-limited cases, presented in

203

Chapters 3 and 4, describe the current as a function of the applied potential.
Therefore, the resulting thin-layer cyclic voltammogram possesses marked

differences from that presented for the diffusion-limited case (Figure 3-1).

Figure 7-1 Simulated Cyclic Voltammogram - Thin-Layer Case

Simulated cyclic voltammogram for a reversible, one-electron transfer reaction at a

planar electrode (A = 0. 39 cm‘?) for the thin-layer (diffusionless) case at 0. 002 V/s. The
electrolyte solution initial/y contained the reduced form but not the oxidized form of the

redox analyte. Other parameters used in the simulation include: E0 = 0. 025 V, a = 0. 5,
k9, = 0.1 cm/s, co = 1x103 M, and 0,, = 5x10"6 cmZ/s.

 

1.0
0.8[
0.6-
0.4;

g 0.2_ Ei——>

g 0.0

0.2
0.4
-05
-0.8

 

l

 

r V

V I V

 

 

11*1

 

 

red
L Ep 1 l
-0.2 -O.1 0.0 0.1 0.2
Potential (V)

 

Figure 7-1 shows a reversible cyclic voltammogram for the thin-layer case,
in which diffusion is neglected, and the current is instead limited by the

exhaustive electrolysis of the thin-layer solution. The experiment begins at the

initial potential (E1) of —0.20 V and is scanned toward the switching potential (EA)

204

of 0.20 V, at which point the direction of the scan is reversed. The current

remains minimal until ~-0.10 V, beyond which, the current rises rapidly with

increasingly positive values of Eapp. as the oxidized form is generated. The

oxidative current reaches a maximum (igx) at the oxidation peak potential (ng ).
Beyond the E8", the current drops rapidly to the baseline until the scan reaches

the EA. At this point, the direction of the scan is reversed and the shape of the

reverse scan mirrors that of the fonlvard scan with the reduction peak current

( ‘fed

lp ) having equal magnitude but opposite sign to that of the 10", and the

reduction peak potential (Eged) at the same value as that observed for the

oxidative process.

The shape of the i-E curve presented in Figure 7-1 infers exhaustive
transformation of the reduced form to the oxidized form upon the forward scan
and the complete regeneration of the reduced form during the reverse scan [13-

16]. Theoretically, if the rate of electron transfer is faster than the scan rate, then
the peaks should be symmetric about EO', and due to the exponential

dependence of the current with the potential, Gaussian in shape. It is important

to note that the peak currents, and therefore, peak potentials occur at the formal

reduction potential, E". The peak current is given by

_r12F2vI»'(';
P 4RT ’

 

i

205

in which the scan rate, v (V/s), the cell volume, V (cm3), and the bulk

Ir
concentration, (2,, (moI/cm3), yields current in amperes [13].

From this relationship, it is seen that the current varies linearly with the
scan rate and the concentration. The integrated area under the cyclic
voltammetric peak yields a value of the faradaic charge, Q, passed during the

anodic and cathodic processes. For exhaustive electrolysis, the charge can be

calculated from the number of moles of the oxidized form, NO, or from the cell

volume and concentration according to

Q = nFNU = nFl'CZ.
Examination of the above equation reveals that the total charge passed is
independent of the scan rate [13].

The electrochemical performance of thin-layer spectroelectrochemical
cells is evaluated based on the above criteria. Knowledge of the charge also
enables one to compare the computed cell volume with that expected from the
cell dimensions or injected volume. Thin-layer spectroelectrochemical
measurements are effective, because, in a relatively short time, one can obtain

electrochemical and spectral information from a single experiment.

7.2 Results - UVIvis Spectroelectrochemistry

7.2.1 Optical Properties of BDD OTEs - UVIvis

Optically pure white diamond (i.e., chemically pure and low in defects) has

one of the widest optical windows of any material, extending from the band gap

206

absorption edge at 225 nm well into the far-IR region [7]. Imperfections
(impurities and defects) in diamond, like those that exist in boron-doped
polycrystalline films, give rise to multiple absorption and scattering phenomena
that reduce the light throughput [1, 6, 7]. Such defects can be structural or
chemical in nature. Structural defects include grain boundaries, stacking faults,
dislocations, twin boundaries and point defects, such as vacancies or
divacancies. Nitrogen and boron are common chemical impurities that may be
intentionally introduced, or naturally occurring. These impurities give rise to point
defects or defect complexes [1, 6, 7, 17-19].

Figure 7-2 shows transmission spectra in the UV/visible region of the
electromagnetic spectrum for two BDD OTEs with different levels of boron
doping, compared to the spectrum for a piece of optically-pure white diamond. It
is clear from Figure 7-2 that boron doping results in an absorption continuum
beginning at about 500 nm and extending into the near-IR. Some of the
decreased transmittance is due to increased reflectance with increasing doping
level. This continuum (absorption from 0.37 -2 eV) is due to the creation of
holes as electrons populate the degenerate boron acceptor levels. High
concentrations of boron result in strong absorption in the red, making heavily
doped films appear blue [1]. Nitrogen introduces yellow coloration due to
electronic absorption in the UV and blue regions of the visible spectrum [1, 6, 17-
19]. In general, boron-doped films are more opaque than their non-doped

counterparts. Increased absorption due to nitrogen impurities or non-diamond

207

 

carbon and increased reflectance or scattering may also contribute to the overall

decrease in the transmission of BDD films.

Figure 7-2 UV/vis Transmission Spectra of Freestanding Boron-Doped and
CVD White Diamond Disks: Effect of Boron Doping Level

Transmission spectra demonstrating the optical transparency of a CVD white (type Ila)
diamond window (solid line) and a freestandirztg BDD disk OTE from 200 - 800 nm.

Spectra are presented for a heavily doped (~ 10 B/cm3) disk (dashed line) and a lightly

doped (~ 1018 B/cm3) disk (dashed-dotted line).

% Transm'ssion

 

 

 

 

 

‘
Q
Q
.
~
.-
Q
in-

 

Figure 7-3 compares the optical throughput in the UV/vis region for a

traditional OTE, ITO deposited on quartz, to that of different types of BDD OTEs.

Some of the overall decreased transmittance observed for the BDD OTEs,

compared to the ITO film on quartz, is likely a result of the differing thicknesses.

The BDD freestanding OTEs were ca. 380 pm thick, and the BDD film on quartz

was ca. 1 pm, compared to the 20 nm nominal thickness of the ITO film.

208

Although the BDD films offer a small (~ 75 nm) gain in the accessible range of
the UV/vis region, the most distinct advantages over ITO lie in the extended
transparency of BDD into the IR region [3, 20], as well as in the superior

electrochemical properties and mechanical stability of diamond [5].

Figure 7-3 UV/vis Transmission Spectra of Optically Transparent
Electrodes: Boron-Doped Diamond compared to Indium-Doped Tin Oxide

A comparison of the UV/vis transmission spectrum of (A) an ITO thin-film on quartz to
the spectra of several BDD OTES: (B) a BDD thin-film on quartz, (C) a polished BDD
thin-film on CVD white diamond, and (D) a freestanding BDD disk.

 

 

 

 

80
. .- A
.8
E B
L
'— .
o\° W'RTr-s
C
D
l

 

 

2004 300R400 L500; 500‘7001900
Wavelength(nm)

The spectrum of the BDD film deposited on quartz (Figure 7-3B) does not
show the sharp cut-off at 226 nm possessed by the other BDD films (Figure 7-3C
and D), this is likely due to absorption by aggregates of unintentionally-added

nitrogen atoms, which decrease the transmittance below 300 nm [5, 21]. The

209

adventitious impurity comes mainly from atmospheric leaks into the reactor
during deposition. The freestanding disks were grown by our collaborator, Prof.
James E. Butler at the Naval Research Laboratory (NRL), under conditions that
are pristinely shielded from commonly encountered nitrogen impurities arising
from source gas contamination or atmospheric leaks into the CVD reactor.

An ideal BDD OTE for UV/vis spectroelectrochemical measurements
would consist of an ultrathin film with a nominal grain size less than 500 nm, in
an effort to minimize light absorption and scattering losses, respectively.
Additionally, such films would possess a minimal electrical resistivity
(< 0.05 0 cm) and high optical throughput (50 - 80%). The nanocrystalline BDD
film deposited on quartz approaches these figures of merit. These films can be
reproducibly prepared to possess low electrical resistivity (< 0.03 0 cm) and high
optical throughput (40 - 50% from 300 to 800 nm) [5]. However, at the time the
work reported in this chapter was performed, only the freestanding BDD disks
were available for use in the UV/vis region. The BDD freestanding disk was
chosen over the BDD on CVD diamond disk due to higher conductivity, better
electrode performance characteristics, and the fact that its usefulness had been
demonstrated for aqueous spectroelectrochemical measurements [4]. It is
important to note that the properties of these new BDD OTEs do not represent
the limiting case; but, instead, a starting point from which an understanding of the
required deposition conditions for the optimal balance of electrical conductivity

and optical transparency can be established.

210

7.2.2 Electrochemical Characterization of the UVIvis OTTLE Cell

The cuvette-based OTTLE cell used for spectroelectrochemical
experiments in the UVIvis region was previously described in Chapter 2.4.2 and
shown in Figure 2-3. An important step in OTTLE cell design is the
characterization of the thin-layer electrochemical properties of the cell. As
discussed above, there are a number of criteria that a cell must meet in order to
demonstrate thin-layer behavior. Only when the cell behaves as a thin-layer
case can one be assured that complete electrolysis is occurring. A series of CV
measurements were performed to test the linearity of the peak current variance,

and the relationship of the faradaic charge with the scan rate.

Figure 7-4 Ferrocene Cyclic Voltammetric Response Measured in the
UV/vis Optically Transparent Thin-Layer Electrochemical Cell

(A) Cyclic voltammetric i-E curve for 0.1 mM ferrocene in acetonitrile containing 0.2 M
TBAC/O4 at 0. 002 V/s. (B) Reduction (squares) and oxidation (circles) peak currents

versus the scan rate plot for 0.1 mM ferrocene in acetonitrile containing 0.2 M TBA ClO4.
Data are shown for scan rates from 0. 002 to 0.010 V/s.

 

 

 

 

 

 

Current (uA)

 

 

 

 

 

A

 

 

4”0.2 ‘ 03 ‘ 074 05 ‘ 06 0mo ‘ 0.005 0310
Paertial (V vs. Ag wire QRE) Scan rate (V/s)

211

 

Figure 7-4A shows a cyclic voltammetric i-E curve for the diamond OTE
exposed to 0.1 mM ferrocene in 0.2 M TBACIO4/CH3CN. The measurement was

made in the thin-layer spectroelectrochemical cell. The potential sweep rate was

0.002 V/s, in an effort to minimize ohmic distortion. The E0 is 0.394 V, as

calculated from the average of the oxidation and reduction peak potentials. AEp

is 0.033 V and Qox/Qred is 1. The oxidation and reduction current peaks are
nearly symmetric with no evidence for any diffusion effects, as expected for a
thin-layer cell. The ferrocene reduction (R2 = 0.998) and oxidation (R2 = 0.999)
peak currents increase linearly with the scan rate in the range from 0.002 to

0.020 V/s, as shown in Figure 7-4B. The charge (Q())( = 25.7 (:1: 2.0) uC,

Qred = 24.6 (i' 1.7) uC) passed at the electrode due to the ferrocene faradaic

process is independent of the scan rate over the same range. From the average
charge, the value calculated for the cell volume is 2.52 uL. The expected cell
volume was ca. 3 uL. The voltammogram is slightly distorted with an upward
shift to the right. This is likely due to the changes in the background capacitive
current due to slow solution leakage around the electrode/mica spacer seal. For
ideal thin-layer behavior, the theoretical peak splitting is zero when the electron
transfer kinetics are rapid compared to the potential sweep rate. Some peak

splitting is observed, due to ohmic cell resistance.

7.2.3 Ferrocene

Figure 7-5 shows a series of ferrocene absorbance spectra between 225

and 350 nm collected at increasingly positive potentials. UVIvis absorbance

212

 

spectra for 1 mM ferrocene in 0.5 M TBACIO4/CH30N were obtained with the

freestanding diamond OTE in the thin-layer cell. The potential was stepped
intermittently from 0.20 to 0.60 V and spectra were collected after a one-minute
equilibration period at each potential. Each spectrum was corrected by

subtracting the reference spectrum obtained at 0.10 V.

Figure 7-5 UV/vis Spectroelectrochemical Absorbance Spectra of
Ferrocene Oxidation at a Freestanding Boron-Doped Diamond Optically
Transparent Electrode

A series of U V/vis absorbance spectra for the generation of ferricenium via
electrooxidation of ferrocene. The potential was stepped intermittent/y from 0. 20 V to
0. 60 V. In all cases, the analyte solution consisted of 1 mM ferrocene in acetonitrile

containing 0.5 M TBA CIO4 as the supporting electrolyte. Each spectrum was corrected
for background absorption by subtracting a reference spectrum acquired at 0.10 V. The
series includes spectra acquired after a 1 min equilibration period at (A) 0. 20 V,
(B) 0.34 V, (C) 0.38 V, (D) 0.42 V, (E) 0.46 V, and (F) 0.60 V. The optical pathlength
was 80 pm.

 

 

 

 

 

 

 

0.14 -
» 252 nm
0.12 -
. F
0.10 e
08 » E
0.08 -
. D
OTB e
0.04 -
i C
0.02
f B
0.00 -
’ 1 . 1A A 1 t r . 1 g
225 250 275 SC!) 325 350
Wavelength (nm)

213

 

As the potential is made more positive, the oxidation of ferrocene to
ferricenium occurs and the absorbance at 252 and 285 nm increases. These two
absorbance maxima are due to ligand to metal charge transfer processes of the
ferricenium ion [22]. The absorbance was constant beyond 0.55 V, indicating
complete electrolysis of ferrocene in the thin-layer cell. The spectral features
were completely reversible with repeated anodic and cathodic steps. In addition
to obtaining the absorption spectrum of the oxidized ferricenium species, one can
extract the values of n and E°' from such data via the slope and x-intercept of a
Nernst plot.

As discussed in the introduction to this chapter, the Nernst equation
defines the relationship between the applied potential and the ratio of the
oxidized-to-reduced concentrations in the thin solution layer. Application of
Beer’s Law,

A zabC ,
relates concentration to absorbance, in which, e is the molar
absorptivity(M'1cm'1), b is the optical pathlength (cm), and C is the analyte
concentration (M). Therefore, one can construct a so-called Nernst plot from a
series of absorbance spectra acquired at different potentials, as for each value of
Eappl, the corresponding concentration ratio can be calculated from the observed
absorbance. Assuming that that the pathlength and molar absorptivity remain
constant allows one to construct a Nernst plot directly from the measured

absorbance changes.

214

Figure 7-6 Nernst Plot for the UV/vis Spectroelectrochemistry of Ferrocene
at a Freestanding Boron-Doped Diamond Optically Transparent Electrode

A Nernstian plot overlaid on Figure 7-4A demonstrating the linear relationship between
the log of the ratio of the oxidized and reduced product absorption and potential. Data
were acquired from the potential step experiments presented in Figure 7-5. In all cases
the analyte solution consisted of 1 mM ferrocene in acetonitrile containing

0.5 M TBA C/O4 as the supporting electrolyte.

 

 

 

 

 

-0. - . - 1 - 1 . 1 -1.5
13.2 0.3 , 0.4 0,5 0.6
Potential (V vs. Ag Wire QRE)

Figure 7-6 shows a linear relationship in the Nernstian plot, which

indicates the system reaches equilibrium and the absorbance increase correlates

with a one-electron reaction. The slope of the linear fit (R2 = 0.999) yields a

value of 1.1 for n. The x-intercept results in a E°' value of 0.397 V, which is in
excellent agreement with that calculated from the CV (0.394 V), as emphasized
by the double plot in Figure 7-6. For the data presented in Figure 7-4, the ratio of
Aon/AAred was obtained using the following equation

AA A

AA

0."

A A

red max _

215

where A represents the absorbance at 252 nm at a given potential, and Amax is
the maximum absorbance at 252 nm observed when Eapp. >> EO'

Figure 7-7 Absorbance-Potential (Voltabsorptometry) Profiles of Ferrocene
UV/vis Spectroelectrochemistry at a Freestanding Boron-Doped Diamond
Optically Transparent Electrode

Absorbance-potential profiles for electrogenerated ferricenium absorbance peaks at 252
and 285 nm. In both cases, the data were acquired simultaneously with a cyclic
voltammogram at 0. 002 V/s. The measurement was made in the cuvette OTTLE cell for

a solution of 1 mM ferrocene in acetonitrile containing 0.5 M TBACIO4 at the

freestanding BDD OTE. The absorbance presented is an absolute value, uncorrected
for the background absorbance. The optical pathlength was 80 pm.

1.95

 

 

 

 

Absorbance

1.75 ~

170 " 285 nm

160 4 4 1 1 1 1
0.4 0.5

 

 

 

 

0.1 0.2 A 0.3 0.6 1 0.7
Potential (v vs. Ag wire QRE)

Figure 7-7 shows absorbance-potential (voltabsorptometry) profiles
measured during one complete cyclic voltammetric cycle through the oxidation

and reduction peaks. The scan was initiated at 0.10 V at a rate of 0.002 V/s and

216

reversed at 0.70 V. The absorbance profiles at 252 and 285 nm were collected
as a function of time. For presentation purposes, conversion from time to
potential was accomplished via knowledge of the experimental time over which
the cyclic voltammogram was performed and the scan rate. The shape of the
Abs-E profiles are diagnostic of exhaustive electrolysis of the thin-layer solution
[14-16].

Upon initiation of the scan, the absorbance remains constant until
~ 0.30 V, at which point the absorbance begins to increase as the potential
becomes more positive, due to the electrogeneration of the ferricenium ion.
Beyond 0.55 V, the absorbance again becomes constant despite the increasingly
positive potentials. This constancy demonstrates both the completion of the
oxidation of ferrocene to the ferricenium ion, as well as the retention of the redox
species within the restricted volume of the thin-layer. The reverse sweep yields a
mirror image of the forward scan. The absorbance should theoretically return to
its original value.

The magnitude of the absorbance increase (AA252 = 0.094, AA235 = 0.062)

of both peaks compares well with that measured during the potential step
measurements shown in Figure 7-5 (AA252 = 0.110, AA285 = 0.072). At both
wavelengths, the absorbance increases and decreases with the same rate of
change, and the magnitudes of the initial and final absorbance are nearly the

same. The relative absorbance at 252 to that 285 nm is 1.54, in good agreement

with that expected from the ratio of the extinction coefficients (1.49). From the

217

 

Abs-E data, the reduction potential for the ferrocene/ferricenium was found to be

~ 0.40 V.

Figure 7-8 Concentration Dependence for the Electrogenerated Ferricenium
UV/vis Absorbance Peaks

(A) U V/vis spectra obtained for varying concentrations of ferrocene in acetonitrile

containing 0.2 M TBA CIO4. Spectra are presented for 2 mM (dashed-dotted line),
3 mM (dotted line), and 5 mM (solid line) ferrocene. The measurements were performed
in the UV/vis OTTLE cuvette cell using a freestanding BDD OTE with a 150 pm spacer.
For each measurement, the potential was stepped from 0. 20 V to 0. 60 V and a spectrum
was acquired after a 1 min equilibration period. Each spectrum was corrected for the
background absorbance by subtracting a reference spectrum obtained at 0.10 V. (B)
Linear dependence of the absorbance at 252 (squares) and 285 (circles) nm on the
concentration for the data presented in (A).

 

 

 

 

 

 

1.2 -
B
0.9 ~
0.6 L
O
I Abs 252 nm
0-3 ' o Ab5285 nm

 

 

1.0 ‘ 2.0 3.0 4.0 5.0 A 6.0
Concentration(rrM)

 

Absorption spectra for three different ferrocene concentrations are shown

in Figure 7-8A. The spectra were collected using a 150 pm spacer at 5 mM,
3 mM, and 2 mM ferrocene in acetonitrile containing 0.2 M TBACIO4 supporting

electrolyte. The potential was stepped from 0.20 to 0.60 V and spectra were
collected after a 1 min equilibration period at each potential. Each spectrum was

corrected by subtracting the reference spectrum obtained at 0.10 V. Figure 7-8B

218

 

demonstrates that the linear dependence of absorbance on concentration of
ferrocene for both peaks. The linear correlation coefficients are 0.999 and 0.990
for the peaks at 252 and 285 nm, respectively. From the slope of the

absorbance versus concentration plots, the extinction coefficients are calculated

1

as 15,110 111'1 cm“1 at 252 nm and 10,845 ltli‘1 cm‘ at 285 nm, which is in

agreement with independently reported values [22].

In summary, the results from the diagnostic tests discussed above
establish that the OTTLE cell design permits rapid, accurate, and reproducible
control of the reduction state of the molecules in the thin layer, which are the
necessary conditions for obtaining high-quality UV/vis electrochemical difference
spectra. It is also important to note that despite the use of nonpolar solvents,
voltammetry was not plagued by large values of inherent cell resistance, further

confirming the choice of cell geometry used in these designs.

7.2.4 Cytochrome c

A series of reduced minus oxidized UVIvis difference spectra of
cytochrome c are shown in Figure 7-9. The insert shows the absolute UVIvis
absorption spectra of the fully oxidized (black line) and fully reduced (grey line)
states for comparison. When fully oxidized, the Soret maximum occurs at
410 nm, and when fully reduced, the Soret maximum shifts to 414 nm and a
prominent d-band at 550 nm appears. The largest amplitude difference signal is
due to the spectrum collected at —0.20 V. The other spectra represent
intermediate states of the protein collected at 0.05 V intervals from 0.30 to

-0.20 V. The arrows in the figure depict that positive peaks, corresponding to the

219

reduced state, increase in amplitude with the application of increasingly negative

potentials.

Figure 7-9 UV/vis Spectroelectrochemical Absorbance Spectra of
Cytochrome c at a Freestanding Boron-Doped Diamond Optically
Transparent Electrode

U V/vis reduced minus oxidized (at 0. 40 V) difference absorbance spectra for horse heart
cytochrome c collected at 0.050 V intervals from 0. 30 to —0. 20 V. The inset depicts the
absolute reduced (grey line) and oxidized (black line) spectra, which were recorded after
one-minute equilibration at -0. 30 V and 0.40 V, respectively. The protein solution
contained 3.5 mM cytochrome c in 100 mM phosphate buffer at pH 7.4. Measurements
were performed in the UV/vis OTTLE cell using a freestanding BDD OTE. The optical
pathlength was 40 pm.

 

 

0.10 - 414 _o,

 

 

 

 

 

 

300 350‘4m‘450 500 550‘600
Wavelength(nm)

The UVIvis spectra of heme-containing proteins such as cytochrome 0
contain two major bands, the Soret and o-bands. The most intense of these two
is the Soret band, which occurs between 400 - 450 nm, depending on the nature

of the heme iron and the state of the enzyme. Both bands are attributed to the

220

 

n—m“ transitions of the heme, and are, therefore, a very sensitive indicator of the
heme iron oxidation state [23]. The spectra presented in Figure 7-9 (from
400 - 600 nm) are the same as reported earlier [24] and are indicative that the
direct electrochemical titration of this enzyme is achievable. It is important to
note that the response is stable in time and reproducible with repeated cycling.
The stable isobestic points further indicate that no protein denaturation occurred.

One additional feature that should be pointed out from Figure 7-9 is the
accessible wavelength range possible with a BDD OTE. Conventional methods
for spectroelectrochemical analysis of cytochrome 0, including indirect methods
relying on electrochemical titrant compounds [25] have not demonstrated access
below 400 nm due to interferent absorption. Figure 7-9 shows that the region
between 200 - 400 nm contains two broad absorption bands, the N and L bands,
which, like the Soret and d-bands, are due to rr—->rr* transitions of heme Fe [27].
Additionally, many characteristic amino acid absorption phenomena occur below
400 nm, such as tyrosine absorption at 280 nm [25].

Although the large Soret bands report on the oxidation state of the heme,
the absorbance of the d-band is routinely used to calculate cytochrome 0
concentration and to monitor heme reduction [25]. This is due to the fact that
absorption of 550 nm is characteristic to reduced cytochrome 0, therefore, it is
not prone to overlapping bands, as is the case in the Soret region where the
oxidized and reduced forms give rise to strong, broad absorption bands at nearly
equal energies. The absorbance data at 550 nm obtained from the direct redox

titration of cytochrome c is shown in Figure 7-10. The solid line through the data

221

points in Figure 7-10 shows a near perfect fit to the Nernst equation. The formal
reduction potential of cytochrome c from this plot is 0.072 V, which is In
agreement (3: 0.01 V) with the accepted value of 0.06 V (vs. Ag/AgCI) [28, 29].
The peak amplitudes remained constant negative of —0.10 V and positive of

0.30 V, as indicated by the plateaus in Figure 7-10.

Figure 7-10 Direct Electrochemical Redox Titration of Cytochrome c at a
Freestanding Boron-Doped Diamond Optically Transparent Electrode

Plot of the amplitude of the a band at 550 nm versus the applied potential. Spectra
acquired at discrete steps from 0.40 to —0. 30 V after a 1 min equilibration in the cuvette
OTTLE cell. The line represents the sigmoidal fit of the Nernst equation (R2 = 0. 998).
The protein solution contained 0.5 mM cytochrome c in 100 mM phosphate buffer at
pH 7. 4. The optical pathlength was 40 pm.

 

0. 080

I

 

[I Ab5550nm]

 

0.025 —

0.020

I

AAbsorbanoe
.0
E?
01
I

0.010

I

0. (1)5 e

 

 

 

0.1 l 0.0 A 0.1 L 0.2 A 0.3 A 0.4
Potential (V vs. Ag/AgCI)

AL

0.000 ‘ ' ‘
—0.3 —0.2

222

 

7.3 Results - IR Spectroelectrochemistry

7.3.1 Optical Properties of BDD OTEs — IR

Figure 7-11 The Effect of Boron Doping Level on the IR Transmission of
Diamond Optically Transparent Electrodes

IR transmission spectra from 3000 - 700 crnA1 of (A) a BDD freestanding disk with low

dopant concentration (~ 1018 B/cm3), (B) a highly boron doped (~ 1020 B/cm3)
freestanding BDD disk, and (C) a CVD white (type Ila) diamond disk. Distinct regions of
the diamond IR spectrum are represented by the numbers across the top axis:
( 1) one-phonon vibronic transitions, (2) two-phonon electronic transitions, and

(3 CH) three-phonon electronic and CH” stretching vibronic transitions. All spectra are
the average of 32 scans collected at 4 cm1 resolution.

 

 

3
40H-.. 2 “‘1 k
100»
l o
80..
.6 . r C
3’)
EGO-
8
‘8
1..
5°40-
20-

 

 

l L

3000 2500 2000 A 1500 A 1000
Frequenqr(cm“)

 

Figure 7-11 demonstrates the effect of boron doping on the transmission
properties of diamond in the IR region. The IR absorption spectrum of diamond

is unique, in that, with the exception of the one-phonon region, the transitions

223

involved are electronic rather than vibronic in nature [1]. The CVD white
diamond spectrum (Figure 7-11C) is characteristic for type Ila diamond, which is
relatively free of structural defects and chemical impurities [30]. The IR

absorption of optically pure diamond is comprised of intrinsic multi-phonon

(i.e., lattice vibration) modes and CHn symmetric and asymmetric stretching
vibrations. Studies of diamond-like carbon films attribute modes at < 3000 cm'1
to spa-bonded CH groups and those at > 2950 cm'1 to spz-bonded CH [1, 31, 32].

The absorbance bands in the 2700 - 1400 cm'1 region are due to the tvvo-phonon

absorption [1 , 2, 7, 30]. These modes are intrinsic to all diamond [8]. Due to the
high degree of lattice symmetry in type Ila diamond, the one-phonon mode is
prohibited; therefore, it is not observed in Figure 7-11C. When boron is
introduced into the lattice, creating a conductive OTE material (type llb diamond),
the transmittance decreases, as shown in Figures 7-11A and B.

Comparison of these figures reveals that the IR absorption is proportional
to the impurity content of the diamond film. Introduction of boron breaks the
lattice symmetry, resulting in increased absorption due to impurities and

one-phonon modes [1]. Specifically, boron gives rise to absorption bands at
2460 and 2790 cm’1 due to the electronic transitions from the ground to the first

and second excited states of the dopant atoms, respectively [8, 9, 33].

Figure 7-11B shows that in highly doped films a broad continuum develops

beyond 2000 cm'1 as a result of the interaction of the boron centers, resulting in

224

boron acceptor level degeneracy. The mode at 1290 cm'1 is the one-phonon

vibronic absorption induced by the presence of boron acceptor centers [8, 9, 33].
The spectra in Figure 7-11 have not been normalized for B doping level or
optical pathlength. Therefore, to make comparisons of the relative transmittance
is difficult. Although the optical pathlengths for BDD freestanding disks are larger
than the CVD white diamond disk, they possess increased transmittance below

~1800 cm‘1. This suggests that processes that decrease the transmission

properties of the material, are less influential in the BDD freestanding films. It is
likely that the refractive indices of the BDD films are lower than that of the CVD
white diamond, as the transmissivity in this region is limited by reflectance
losses. Despite the marked decrease in transmittance observed for the region

above 2000 cm'1 due to boron incorporation, it is important to note that the BDD

films retain high levels of throughput in the region below 1000 cm'1.

Ideally, a BDD OTE for IR spectroelectrochemical measurements would
consist of an ultrathin film deposited on a highly transparent substrate to aid in
the mechanical stability. Such OTEs would also possess a minimal electrical
resistivity (< 0.05 0 cm) and high optical throughput (50 - 80%) over the IR range
of the spectrum. It is our opinion that a BDD thin-film on CVD white diamond or a
BDD freestanding disk of minimal thickness would represent the best-case
scenario for BDD OTEs. However, the unavailability and the high cost
associated with the fabrication of such electrodes resulted in the decision to use
BDD thin-films deposited on Si for the studies reported herein. Figure 7-12

shows the IR absorption spectra for three different types of BDD OTEs: (A) a

225

freestanding BDD disk, (B) a BDD thin-film on Si, and (C) a BDD thin-film on
CVD white diamond. It is important to note that these spectra have not been

normalized for boron doping level or thickness.

Figure 7-12 IR Transmission Spectra of Different Types of Boron-Doped
Diamond Optically Transparent Electrodes

IR absorption spectra for three different types of BDD OTEs: (A) a freestanding BDD
disk, (B) a BDD thin film on Si (4 h growth), (C) a BDD thin-film on CVD white diamond, '
and (D) a BDD thin film on Si (10 h growth). All spectra are the average of 32 scans

collected at 4 cmA1 resolution.

 

 

% Transm'ssion _.
8 8 8

8

B

 

 

 

 

 

Figure 7-12A shows that below 2300 cm"1 the freestanding disk transmits

greater than 50% of the IR radiation. This film also possesses a sufficient level
of conductivity for transferring electrons during potential step measurements.

However, it is clear from Figure 7-12A that this film is of a critical thickness that

226

gives rise to interference fringes below 2300 cm'A, rendering it unsuitable for our

purposes. The IR spectrum of the BDD thin-film on CVD white diamond
(Figure 7-12C) shows very low throughput over the entire range; this is likely due
to the high boron content of this film. As mentioned earlier, these freestanding
films were fabricated by our collaborators at the NRL; therefore, it was not ideal
nor time-efficient to try to understand the relationship between growth conditions
and optimal OTE properties via these films. Instead, we chose to focus our
preliminary studies on OTEs that could be fabricated in-house. Figures 7-128
and 7-12D show the IR absorption for such BDD OTEs. The OTE represented in

Figure 7-12B possessed adequate conductivity and ample transparency below

-1 . . . .
2300 cm for use in spectroelectrochemical measurements. Therefore, it is

important to note that the properties of these new BDD OTEs do not represent
the limiting case, but the current status in our quest to understand the necessary
deposition conditions for obtaining optimal OTE properties.

Investigation of the BDD thin-film OTE spectra, presented in
Figures 7-12B and D, reveals the effect of the growth time on the throughput.

The growth conditions for the films were identical except for the growth time. As

discussed above, boron doping contributes modes at 1285, 2460 and 2790 cm'1.

The boron features above 2400 cm'1 drastically affect the transmission properties

of the boron-doped films. In Figures 7-128 and D, these features absorb
intensely, leading to a reduction in throughput to < 30% in the 4h film
(Figure 7-1ZB) and < 1% in the 10 h film (Figure 7-1ZD). The resistivities were

0.4 and 0.01 0 cm for the 4 h and 10 h films, respectively. Thus, the 10 h film

227

was highly conductive, although, not viable for use as an OTE, due to low Optical
throughput. The 4 h film possessed sufficient conductivity and optical
throughput, rendering it useful for the potential step, spectroelectrochemical
studies of ferrocene, ferrocyanide, and cytochrome 0 described later. The highly
conductive, 10 h film was used for the electrochemical characterization of the

thin-layer spectroelectrochemical cell.

7.3.2 Ferrocene

The spectroelectrochemical cell used to measure FTIR difference spectra
of ferrocene was described in Section 2.4.3 and shown in Figure 2-4. In the
mid-IR region, the spectrum for ferrocene has been measured using a number of

thin-layer cells [34-39]. Figure 7-13 shows the IR electrochemical difference
(oxidized minus reduced) spectrum for 20 mM ferrocene and 0.1 M TBAPFe in

dichloromethane. In good agreement with published data, a ferrocene C-H

bending mode at 1004 cm'1 shifts to 1011 cm'1 upon oxidation to ferricenium ion

[34-38]. A cyclopentadienyl ring-breathing mode at 1107 cm'1 is also observed

for ferrocene. The oxidized and reduced spectra were collected after a 2 min

equilibration time at 0.60 V and 0.20 V, respectively.

228

Figure 7-13 IR Spectroelectrochemical Absorbance Spectra of
Ferrocene/Ferricenium Measured at a Boron-Doped Diamond Optically
Transparent Electrode - Mid-IR

Oxidized minus reduced electrochemical FTIR difference spectrum for the

electrooxidation of 20 mM ferrocene in dichloromethane containing 0.1 M TBAPFG. The
spectrum was acquired in the OTTLE cell (Model I) using a BDD thin-film on Si OTE and
a AgCI window. The optical pathlength was 250 pm. Spectra were collected after a
2 min equilibration at 0.20 V (reduced) and 0. 60 V (oxidized). Each spectrum is the

average of 200 scans measured at 15 0C with 4 cm1 resolution.

 

 

 

 

-3
AA]1.5x10
1014
Q)
U
C
(U
'9
O
(D
Q
<
<3
1004
1107
1150 A 1100 A 1050 A 1000 A 950

Frequency (cm'1)

The low-frequency FTIR difference spectrum (oxidized minus reduced) for
10 mM ferrocene plus 0.1M TBACIO4 in acetonitrile is shown in Figure 7-14. The
shift from 823 to 854 cm’1 upon oxidation is clearly observed. This result is in

agreement with a previous report of ferrocene oxidation in this region of the
spectrum [34]. The spectra were collected at 0.20 V and 0.60 V after a 2 min

equilibration. This experiment was an important test for our design, as we

229

eventually intend to use the method to measure low-frequency (<1000 cmA1)

metal-ligand vibrations in protein active sites and inorganic model complexes.

Figure 7-14 IR Spectroelectrochemical Absorbance Spectra of
Ferrocene/Ferricenium Measured at a Boron-Doped Diamond Optically
Transparent Electrode — Far-IR

Oxidized minus reduced electrochemical F T/R difference spectrum of 10 mM ferrocene

in acetonitrile containing 0.1 M TBACIO4. The spectra were acquired in the OTTLE cell
(Model I) using a BDD thin-film on Si OTE and a AgCI window. The optical pathlength
was 250 pm. Spectra were collected after a 2 min equilibration at 0. 20 V (reduced) and

0. 60 V (oxidized). Each spectrum is the average of 200 scans measured at 15 0C with
4 cm1 resolution.

 

354 740.1

A Absorbance

823

1000 900 8CD 700
Frequency (cm'1)

 

 

 

7.3.3 Electrochemical Characterization of the IR OTTLE Cell

The OTTLE cell that was used for the spectroelectrochemical experiments
of the aqueous samples (ferrocyanide and cytochrome c) in the IR region was

previously described in Chapter 2 (Section 2.4.4) and shown in Figure 2-5. A

230

series of CV measurements were performed to test the linearity of the peak
current variance, and the relationship of the faradaic charge with the scan rate, in
an effort to demonstrate thin-layer electrochemical behavior within the OTTLE

cell.

Figure 7-15 Ferrocyanide Cyclic Voltammetric Response Measured in the IR
Optically Transparent Thin-Layer Electrochemical Cell (Model II)

(A) Cyclic voltammetric i—E curves for 1 mM ferrocyanide in 1 M K CI at 0. 002, 0. 005,
0.007, 0.010, and 0. 020 V/s. (B) Plot of the oxidation (circles) and reduction (squares)
peak currents versus the scan rate for the CVs shown in (A).

 

 

 

 

 

 

 

 

 

 

 

 

 

. B
9 .
6 L" I red
3 3 P o ax
75 l
.. 0
5 .
g .3 v
-6 c
-9 _
-12 A i L ‘ . l + i - 1 A _1 A 1 . 1 . J . 1
0.0 0-1 0-2 Q3 0-4 0-5 0-6 02.000 0.005 0.010 0.015 0.020
Potertial (V vs. AgIAgCI) Scat Rate (Vls)

Figure 7-15A shows a series of cyclic voltammetric i-E curves for the
diamond OTE exposed to 1 mM ferrocyanide in 1 M KCI. The measurements
were made in the thin-layer spectroelectrochemical cell with potential sweep
rates from 0.002 to 0.02 V/s. As the potential is swept from 0 to 0.60 V,
ferrocyanide is oxidized to the Fe(III) ferricyanide form. The reduction of
ferricyanide to the Fe(ll) ferrocyanide form takes place upon the reverse scan

from 0.60 to 0 V. The characteristic thin-layer shapes of the oxidation and

231

reduction CV peaks indicate complete conversion of ferro- to ferricyanide and

ferri- to ferrocyanide, respectively.
The Ep/2 of 0.272 V, calculated from the average of the oxidation and
reduction peak potentials, remains constant over the entire scan rate range. The

AEp values range from 0.011 to 0.074 V and Qox/Qred is 1. As expected for a

thin-layer cell, the oxidation and reduction current peaks are nearly symmetric

with no evidence for any diffusion effects. The oxidation (R2 = 0.999) and

reduction (R2 = 0.996) peak currents of the ferri-lferrocyanide couple increase
linearly with the scan rate in the range from 0.002 to 0.020 V/s, as shown in
Figure 7-158. The charge (Q0x = 67.5 (_+_ 4.2) pC, Qred = 68.9 (i 4.0) pC) passed
at the electrode due to the ferrocyanide faradaic process is independent of the
scan rate over the same range. From the average charge, the value calculated
for the cell volume is 0.71 uL. The injected volume was ca. 1 uL. The thickness

of the solution layer was estimated to be 11 pm by dividing the cell volume by the
geometric area of the electrode (0.636 cm2). For ideal thin-layer behavior, the

theoretical peak splitting is zero when the electron transfer kinetics are rapid
compared to the potential sweep rate. Some peak splitting is observed, due to

cell resistance, electrode resistance, or slow electron transfer kinetics.

232

 

7.3.4 Ferrocyanide

Figure 7-16 IR Spectroelectrochemical Absorbance Spectra of
Form/Ferrocyanide Measured at a Boron-Doped Diamond Optically
Transparent Electrode

Oxidized minus reduced (solid line) and reduced minus oxidized (dashed line)
electrochemical difference F TIR spectra for 10 mM ferrocyanide in 1 M KCI. The spectra
were acquired in the OTTLE cell using a BDD thin-film on Si OTE and a CaF2 window.
The optical pathlength was 10 um. Each spectrum is the average of 200 scans
measured at 15 0C with 4 cm'1 resolution. After a 1 min equilibration, oxidized and
reduced spectra were collected at 0. 50 V and 0.10 V, respectively.

 

 

AAbsorbance

 

 

 

 

2400A2200‘2000A1800
Frequency (cm’1)

An experiment in the mid-frequency IR range was performed to evaluate
the utility of BDD OTEs in transmission IR spectroelectrochemistry. In this
region, the spectra for ferrocyanide and ferricyanide are well documented using a
variety of thin-layer cells [34, 36]. Figure 7-16 shows the fonivard (oxidized minus

reduced, solid line) and reverse (reduced minus oxidized, dashed line) IR

233

 

electrochemical difference spectra measured for an aqueous solution of 10 mM

ferrocyanide containing 1 M KCI. In good agreement with published data, the CN

stretching mode at 2115 cm'1 shifts to 2036 cm'1 upon oxidation to ferricyanide

[34, 36, 40]. The oxidized and reduced spectra were collected after a 1 min
equilibration time at 0.50V and 0.10 V, respectively. Figure 7-16 also
demonstrates that the response measured in the OTTLE cell is highly reversible,

in that no signal is lost during the reverse process.

7.3.5 Cytochrome c

As discussed above, the results for ferricyanide and ferrocene
demonstrated the viability of performing electrochemical difference IR
measurements using BDD thin-films on Si as OTEs. One of our ultimate goals
was to incorporate BDD OTES into measurements of redox proteins; therefore,
we began our investigation with a simple, small molecular weight protein,
cytochrome 0. Aside from the simplicity and availability of this protein, the
electrochemical difference IR spectrum had been reported using a Au grid OTE
[24], enabling us to evaluate the performance of our new OTE with respect to

that of a conventional OTE.

234

Figure 7-17 Comparison of the Mid-IR Spectra of Cytochrome c Measured
at a Au Grid and at a Boron-Doped Diamond Optically Transparent
Electrode

Reduced (-0. 20 V) minus oxidized (0. 30 V) electrochemical FTIR difference spectra of

cytochrome c acquired at (A) a Au grid OTE and (B) at a BDD OTE (BDD thin-film on
undoped Si). A noise spectrum (C) is shown for the cell containing the BDD OTE. All

spectra are the average of 800 scans collected at 10 0C with a 10 pm pathlength. The
protein solution consisted of 3.5 mM cytochrome c, 0.3 M NaCI, 10 mM Tris HCI at pH 7.

Spectrum (A) was acquired using two CaFg windows and the OTTLE cell shown in
Figure 2-2. Spectra (B) and (C) were collected in the OTTLE cell shown in Figure 2-5

with a Can window.

 

AAI1X103

 

 

1 A l 1 l 1 l A l n l A l L

1800 1700 1600 1500 1400 1300 1200 1100 1000
Frequency (cm '1)

 

Reduced minus oxidized IR difference spectra of cytochrome c are shown
in Figure 7-17. This figure presents a comparison between spectra obtained
using a conventional Au grid OTE (Figure 7-17A) and a BDD thin-film on
undoped Si OTE (Figure 7-17B). Figure 7-17C shows that the noise spectrum,
produced by application of an oxidizing potential to an oxidized sample, is flat

with weak features in areas of strong background absorbance. Re-oxidation of

235

the reduced sample at the BDD OTE resulted in an oxidized minus reduced
spectrum that was almost the exact mirror image of the spectrum shown in
Figure 7-17B. Both spectra were found to be fully reproducible over a number of
cycles.

Comparison of Figures 7-17A and B reveals excellent agreement between
the spectra obtained at the Au grid and BDD OTEs, in that all spectral features
are retained. Moss and coworkers have reported the same FTIR spectrum for

cytochrome c that was observed herein, as well as its corresponding UV/vis
difference spectrum obtained for a Au grid OTE between Can windows [24].

Although simultaneous measurement of the optical spectrum was not possible in
our case, the agreement of the FTIR spectra with that reported by Moss and
coworkers suggests that the protein is intact and not denatured at the BDD
surface. The many peaks observed in the FTIR electrochemical difference
spectra of proteins are representative of amino acid side chain vibrations, and
peptide backbone vibrations. Chapter 6 contained a thorough discussion of the
assignments of such modes observed for cytochrome c oxidase spectra. A
similar discussion is omitted for the spectrum of cytochrome c because the goal,
in this case, was to evaluate the performance of the new BDD OTEs based on
previously reported results [24], rather than to attempt to assign modes in the

spectrum.

Inspection of the regions above 1700 cm'1 and below 1100 cm‘1,

demonstrates a marked difference in the noise level of the two spectra. This is

primarily attributed to the ~ 20% decrease in transmittance observed for the BDD

236

OTE compared to that of the Au grid OTE. The increased noise level and
decreased throughput are concerning, especially when considering low-
frequency FTIR measurements. It was demonstrated in Chapters 5 and 6 that
measurements of aqueous samples in this region require extremely low noise. in
addition, a significant decrease in conductivity was necessary in order to achieve
adequate throughput in the BDD thin-film on Si OTEs. Therefore, it is important
to investigate the viability of using other types of BDD OTEs, those with higher
throughput and the same or better electrical conductivity. “Membrane" BDD thin-
films fabricated by our collaborators at the NRL provide one attractive option.
These OTEs are based on the same idea used herein, that of depositing a thin-
film of BDD onto a Si substrate; however, in this case, the substrate is carefully
etched away, leaving a very fragile film of minimal thickness. Yet, these films
retain some degree of stability, because only the substrate area in the light path
is removed. Additionally, as discussed in Chapter 5, experiments in reflection
mode, may be further investigated, as Martin and Morrison have demonstrated
the application of a conducting diamond film on Si as a transparent electrode for

ATR FTIR spectroelectrochemical experiments [3].

7.4 Conclusions

The spectroelectrochemical responsiveness of two types of optically
transparent diamond electrodes was investigated. These novel OTEs possess
adequate levels of conductivity and transmissivity for performing
spectroelectrochemical measurements in aqueous and nonaqueous media. A

freestanding, mechanically polished, BDD disk (0.38 mm thick and 8 mm in

237

diameter) served as the OTE for UV/vis spectroelectrochemical measurements.
The OTE for the IR measurements was a BDD thin film (~ 2 pm) deposited on an
undoped Si substrate. Specially designed, thin-layer spectroelectrochemical
cells were constructed for UVIvis and IR transmission measurements.
Transmission spectroelectrochemical measurements were made using
ferrocene in acetonitrile and dichloromethane. The results were in good

agreement with the literature. Ferrocene was electrooxidized to the ferricenium

ion and this product was spectroscopically monitored in the UV/vis (it. = 252,
285 nm) and IR (v = 854, 1014 cm'1) regions of the electromagnetic spectrum.

Additionally, the transmission electrochemical difference FTlR spectrum for the

electrooxidation of ferrocyanide was measured using a BDD OTE. The

conversion of ferrocyanide to ferricyanide was monitored by peaks at 2036 cm'1

and 2115 cm'1 due to the reduced and oxidized forms, respectively. These

results are in agreement with independently reported spectra measured using Au
minigrid OTES.

Lastly, a new application of BDD OTEs was explored. The UVIvis and
FTIR electrochemical difference spectra of cytochrome c were measured using
BDD OTEs. The results demonstrate that BDD is a useful OTE for
spectroelectrochemical measurements of aqueous solutions of proteins.
Excellent agreement was observed for the response at BDD with that from
conventional OTEs such as ITO and Au minigrids. However, work remains to
produce a BDD OTE that can be used for low-frequency FTIR

spectroelectrochemical measurements of proteins. Due to decreased

238

transmission, the noise levels below 1000 cm'1 are greater for BDD OTES than

for Au grid OTEs.

The work described herein, to produce highly conductive and transmissive
BDD OTEs, is in the early stages. Therefore, the optical and electronic
properties of the BDD OTEs presented in this work do not represent the limiting
case. Further attempts to increase conductivity and transmission via
manipulation of the BDD growth conditions will be pursued. Additionally,
alternate methods of IR spectroelectrochemistry, namely ATR, will be

investigated, in efforts to measure high-quality spectra of redox-active proteins.

239

 

7.5

[1-1

[2-1

[3-1

[4-1

[5-1

[6-1

[7-1

[8-1

[9-1

[10.]

[11.]

[12.]

References

J. I. Pankove and C.-H. Qui ln Synthetic Diamond: Emerging CVD
Science and Technology; K. E. D. Spear, J.P., Ed.; John Wiley & Sons,
Inc., New York, 1994, 401-407.

A. M. Zaitsev Optical Properties of Diamond: A Data Handbook; Springer-
Verlag, Berlin, 2001.

H. B. Martin and P. W. Morrison Electrochem. Solid-State Lett. 2001, 4,
E17-E20.

J. K. Zak, J. E. Butler and G. M. Swain Anal. Chem. 2001, 73, 908-914.

J. M. Stotter, J. K. Zak, Z. Behler, Y. Show and G. M. Swain Anal. Chem.
2002, in press.

W. Zhu In DiamondsE/ectronic Properties and Applications; L. S. K. Pan,
D.R., Ed.; Kluwer Academic Publishers, Boston, 1995, 175.

Y. Tzeng Diam. Film Tech. 1991, 1, 31.

K. M. McNamara, B. E. Williams, K. K. Gleason and B. E. Scruggs J. Appl.
Phys. 1994, 76, 2466-2472.

F. Q. Zhang, E. Q. Xie, B. Yang, Y. M. Cal and G. H. Chen Mater. Lett.
1994,19,115-118.

S. Haymond, G. T. Babcock and G. M. Swain Electroanal. 2002, in press.

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 Materials 2002, 8, 171-188.

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. 2000, 72, 3793-3804.

240

[13.]

[14.]

[15.]

[16.]

[17]

[18.]

[19.]

[20.]

[21.]

[22.]

[23.]

[24.]

[25.]

[26.]

A. J. Bard and L. R. Faulkner Electrochemical Methods: Fundamentals
and Applications; 2nd ed.; John Wiley & Sons, Inc., New York, 2001.

J. Niu and S. Dong Electroanal. 1995, 7, 1059-1062.

N. J. Simmons and M. D. Porter Anal. Chem. 1997, 69, 2866-2869.

J. Zak, M. D. Porter and T. Kuwana Anal. Chem. 1983, 55, 2219-2222.

A. Mainwood Phys. Rev. B 1994, 49, 7934-7940.

J. A. Freitas, P. B. Klein and A. T. Collins Appl. Phys. Lett. 1994, 64,
2136-2138.

A. Cox, M. E. Newton and J. M. Baker J. Phys.-Condens. Matter 1994, 6,
551-563.

S. Haymond, J. K. Zak, J. E. Butler, G. T. Babcock and G. M. Swain in
preparation 2002.

A. T. Collins J. Phys.-Condens. Matter 2002, 14, 3743-3750.

Y. S. Sohn, D. N. Hendrickson and H. B. Gray J. Am. Chem. Soc. 1971,
93, 3603-3612.

S. Horie and M. Morrison J. Biol. Chem. 1963, 238, 2859-2865.

D. Moss, E. Nabedryk, J. Breton and W. Mantele Eur. J. Biochem. 1990,
187, 565-572.

R. M. C. Dawson, D. C. Elliott, W. H. Elliott and K. M. Jones Data for
Biochemical Research; 3rd ed .; Oxford University Press, Inc., New York,
1986.

E. F. Bowden, F. M. Hawkridge, J. F. Chlebowski, E. E. Bancroft, C.
Thorpe and H. N. Blount J. Am. Chem. Soc. 1982, 104, 7641-7644.

241

 

[27.]

[28.]

[29.]

[30.]

[31.]

[32.]

[33.]

[34.]

[35.]

[36.]

[37.]

[38.]

[39.]

W. A. Eaton and J. Hofrichter Methods Enzymol. 1981, 76, 175-261.

F. M. Hawkridge and T. Kuwana Anal. Chem. 1973, 45, 1021-1027.

J. D. Burgess and F. M. Hawkridge In Electroanalytical Methods for
Biological Materials; A. Brajter-Toth and J. Q. Chambers, Eds; Marcel
Dekker, New York, 2002, 109-142.

W. Van Enckevort In Synthetic Diamond: Emerging CVD Science and
Technology; K. E. D. Spear, J.P., Ed.; John Wiley & Sons, Inc., New York,
1994, 322-325.

L. J. Bellamy The Infrared Spectra of Complex Molecules; Chapman and
Hall, New York, 1980; Vol. 2.

G. Socrates Infrared Characteristic Group Frequencies; Wiley, New York,
1980.

J. Mort, M. A. Machonkin and K. Okumura Appl. Phys. Lett. 1991, 58,
1908-1910.

Y. J. Ma, J. M. Zheng and S. M. Zhu Chin. J. Inorg. Chem. 1999, 15, 61-
67.

P. A. Flowers and S.-A. Callender Anal. Chem. 1996, 68, 199-202.

J. J. Niu and S. J. Dong Electrochim. Acta 1995, 40, 823-828.

J. P. Bullock, D. C. Boyd and K. R. Mann Inorg. Chem. 1987, 26, 3084-
3086.

M. Krejcik, M. Danek and F. Hartl J. Electroanal. Chem. 1991, 317, 179-
187.

H. Visser, A. E. Curtright, J. K. McCusker and K. Sauer Anal. Chem. 2001,
73, 4374-4378.

242

[40.] T. Noguchi and Y. Inoue J. Biochem. 1995, 118, 9-12.

243