Electroanalytical Chemistry A Series of Advances Volume 22 Electroanalytical Chemistry

ELECTROANALYTICAL
CHEMISTRY
A SERIES OF ADVANCES
edited by
Allen J. Bard
-of chwvllstry
Universily of T m
Audn, T m , USA
Israel Rubinstein
VOLUME 22
M A R C E L
MARCEL DEKKER, INC.
D E K K E R
NEW YORK BASEL
The Library of Congress Cataloged the First Issue of This Title as Follows:
Electroanalytic chemistry: a series of advances, v. 1
New York, M. Dekker, 1966-
v. 23 cm.
Editors: 19661995 A. J. Bard
1966- A. J. Bard and I. Rubinstein
1. Electromechanical analysisAddresses, essays, lectures
1. Bard, Allen J., ed.
QD115E499 545.3 66-11287
Library of Congress
0-8247-4719-4 (v. 22)
This book is printed on acid-free paper.
Headquarters
Marcel Dekker, Inc.
270 Madison Avenue, New York, NY 10016
tel: 212-696-9000; fax: 212-685-4540
Eastern Hemisphere Distribution
Marcel Dekker AG
Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland
tel: 41-61-260-6300; fax: 41-61-260-6333
World Wide Web
http://www.dekker.com
The publisher oers discounts on this book when ordered in bulk quantities. For
more information, write to Special Sales/Professional Marketing at the headquar-
ters address above.
Copyright nnnn 2004 by Marcel Dekker, Inc. All Rights Reserved.
Neither this book nor any part may be reproduced or transmitted in any formor by
any means, electronic or mechanical, including photocopying, microlming, and
recording, or by any information storage and retrieval system, without permission
in writing from the publisher.
Current printing (last digit):
10 9 8 7 6 5 4 3 2 1
PRINTED IN THE UNITED STATES OF AMERICA
This series is designed to provide authoritative reviews in the eld of mod-
ern electroanalytical chemistry dened in its broadest sense. Coverage is
comprehensive and critical. Enough space is devoted to each chapter of
each volume so that derivations of fundamental equations, detailed de-
scriptions of apparatus and techniques, and complete discussions of im-
portant articles can be provided, so that the chapters may be useful without
repeated reference to the periodical literature. Chapters vary in length and
subject area. Some are reviews of recent developments and applications of
well-established techniques, whereas others contain discussion of the
background and problems in areas still being investigated extensively
and in which many statements may still be tentative. Finally, chapters on
techniques generally outside the scope of electroanalytical chemistry, but
which can be applied fruitfully to electrochemical problems, are included.
Electroanalytical chemists and others are concerned not only with
the application of new and classical techniques to analytical problems, but
also with the fundamental theoretical principles upon which these tech-
niques are based. Electroanalytical techniques are proving useful in such
diverse elds as electro-organic synthesis, fuel cell studies, and radical ion
formation, as well as with such problems as the kinetics and mechanisms of
electrode reactions, and the eects of electrode surface phenomena,
adsorption, and the electrical double layer on electrode reactions.
It is hoped that the series is proving useful to the specialist and non-
specialist alikethat it provides a background and a starting point for
graduate students undertaking research in the areas mentioned, and that it
also proves valuable to practicing analytical chemists interested in learning
about and applying electroanalytical techniques. Furthermore, electro-
chemists and industrial chemists with problems of electrosynthesis, electro-
plating, corrosion, and fuel cells, as well as other chemists wishing to apply
electrochemical techniques to chemical problems, may nd useful material
in these volumes.
A. J. B.
I. R.
INTRODUCTION TO THE SERIES
iii
L. DAIKHIN Tel Aviv University, Ramat Aviv, Israel
STEPHEN W. FELDBERG Brookhaven National Laboratory, Upton,
New York, U.S.A.
E. GILEADI Tel Aviv University, Ramat Aviv, Israel
MARSHALL D. NEWTON Brookhaven National Laboratory, Upton,
New York, U.S.A.
JOHN F. SMALLEY Brookhaven National Laboratory, Upton, New
York, U.S.A.
GREG M. SWAIN Michigan State University, East Lansing, Michigan,
U.S.A.
V. TSIONSKY Tel Aviv University, Ramat Aviv, Israel
M. URBAKH Tel Aviv University, Ramat, Israel
v
CONTRIBUTORS TO VOLUME 22
CONTENTS OF VOLUME 22
Introduction to the Series iii
Contributors to Volume 22 v
Contents of Other Volumes xiii
LOOKING AT THE METAL/SOLUTION INTERFACE
WITH THE ELECTROCHEMICAL QUARTZ-CRYSTAL
MICROBALANCE: THEORY AND EXPERIMENT
V. Tsionsky, L. Daikhin, M. Urbakh, and E. Gileadi
I. Introduction 2
A. Is It Really a Microbalance? 3
B. Applications of the Quartz Crystal Microbalance 4
C. The Impedance Spectrum of the EQCM 5
D. Outline of This Chapter 8
II. Theoretical Interpretation of the QCM Response 8
A. Impedance 8
B. The Eect of Thin Surface Films 12
C. The Quartz Crystal Operating in Contact
with a Liquid 16
D. Quartz Crystals with Rough Surfaces 26
III. Electrical Double Layer/Electrostatic Adsorption 33
A. Introduction 33
B. Some Typical Results 34
C. The Potential Dependence of the Frequency 36
vii
IV. Adsorption Studies 43
A. The Adsorption of Organic Substances 43
B. The Adsorption of Inorganic Species 53
V. Metal Deposition 60
A. Deposition on the Same Metal Substrate 60
B. Early Stages of Metal Deposition on a Foreign
Substrate 64
VI. The Inuence of Roughness on the Response of the
QCM in Liquids 70
A. The Nonelectrochemical Case 71
B. The Electrochemical Case 76
VII. Conclusion 83
VIII. Appendix 86
A. Nonuniform Film on the Surface 86
B. Experimental Remarks 86
References 94
THE INDIRECT LASER-INDUCED TEMPERATURE
JUMP METHOD FOR CHARACTERIZING FAST
INTERFACIAL ELECTRON TRANSFER: CONCEPT,
APPLICATION, AND RESULTS
Stephen W. Feldberg, Marshall D. Newton,
and John F. Smalley
I. Introduction 102
A. Why Measure Fast Interfacial Electron Transfer
Rate Constants? And How? 103
B. Background 104
C. The Underlying Principles of the ILIT Method
The Short Version 106
D. Denition of Terms 108
II. The Evolution of the ILIT Method for the Study of Fast
Interfacial Electron Transfer Kinetics 108
A. The Temperature-Jump Approach for Studies of
Homogeneous Kinetics 108
Contents of Volume 22 viii
B. The Temperature-Jump Approach for Studies of
Interfacial Kinetics 108
III. Relevant Electron Transfer Theory: Marcuss
Description of Heterogeneous Nonadiabatic Electron
Transfer Reactions 112
A. Chidseys Approach 112
B. Temperature Dependence 116
C. How Well Does the Butler-Volmer Expression
Approximate Marcuss Formalism? 118
IV. Analysis of the ILIT Response 120
A. Response of the Open-Circuit Electrode Potential to
a Change in the Interfacial Temperature in the
Presence of a Perfectly Reversible Redox Couple
Attached to the Electrode Surface 121
B. The Relaxation of the ILIT Response When the
Rate of Electron Transfer Is Not Innitely Fast 126
C. When Is the ILIT Response Purely Thermal (i.e.,
Devoid of Kinetic Information)? 126
D. The Shape of the Ideal ILIT Perturbation 130
E. Nonidealities of the Shape of the ILIT Perturbation
and ResponseExtracting the Relaxation Rate
Constant, k
m
134
F. Correlating k
m
to Meaningful Physical Parameters 137
V. Experimental Implementation of ILIT 143
A. The Cell 143
B. The Working Electrode: Preparation and Thermal
Diusion Properties 148
C. Preparation of Self-Assembled Monolayers 150
D. The Electronics 151
E. Potential Problems 152
F. Energetic and Timing Considerations for Single and
Multiple Pulse Experiments 156
G. Some Suggested Experimental Protocols 160
VI. A Few Examples of Measurements
of Interfacial Kinetics 161
A. Some Typical Transients 161
B. Determining the Value of k
o
163
C. Arrhenius Plots and Evaluation of DH
p
and DH
k
163
Contents of Volume 22 ix
VII. The Potential of the ILIT Approach 166
VIII. Some Thoughts About Future Experiments 166
IX. Glossary of Terms 170
X. Appendix: One-Dimensional Thermal Diusion into
Two Dierent Phases 173
References 175
ELECTRICALLY CONDUCTING DIAMOND
THIN FILMS: ADVANCED ELECTRODE MATERIALS
FOR ELECTROCHEMICAL TECHNOLOGIES
Greg M. Swain
I. Introduction 182
II. Diamond Thin Film Deposition, Electrode
Architectures, Substrate Materials, and Electrochemical
Cells 185
III. Electrical Conductivity of Diamond Electrodes 194
IV. Characterization of Microcrystalline and Nanocrystalline
Diamond Thin Film Electrodes 195
V. Basic Electrochemical Properties of Microcrystalline and
Nanocrystalline Diamond Thin Film Electrodes 201
VI. Factors Aecting Electron Transfer at Diamond
Electrodes 212
VII. Surface Modication of Diamond Materials and
Electrodes 216
VIII. Electroanalytical Applications 219
A. Azide Detection 219
B. Trace Metal Ion Analysis 221
C. Nitrite Detection 224
D. NADH Detection 225
E. Uric Acid Detection 225
F. Histamine and Serotonin Detection 226
G. Direct Electron Transfer to Heme Peptide and
Peroxidase 227
Contents of Volume 22 x
H. Cytochrome c Analysis 228
I. Carbamate Pesticide Detection 228
J. Ferrocene Analysis 229
K. Aliphatic Polyamine Detection 230
IX. Electrosynthesis and Electrolytic Water Purication 238
X. Optically Transparent Electrodes for
Spectroelectrochemistry 239
XI. Advanced Electrocatalyst Support Materials 251
A. Composite Electrode Fabrication and
Characterization 252
B. Oxygen Reduction Reaction 259
C. Methanol Oxidation Reaction 264
XII. Conclusions 267
References 268
Author Index 279
Subject Index 295
Contents of Volume 22 xi
CONTENTS OF OTHER VOLUMES
VOLUME 1
AC Polarograph and Related Techniques: Theory and Practice,
Donald E. Smith
Applications of Chronopotentiometry to Problems in Analytical
Chemistry, Donald G. Davis
Photoelectrochemistry and Electroluminescence, Theodore Kuwana
The Electrical Double Layer, Part I: Elements of Double-Layer Theory,
David M. Monhilner
VOLUME 2
Electrochemistry of Aromatic Hydrocarbons and Related Substances,
Michael E. Peover
Stripping Voltammetry, Embrecht Barendrecht
The Anodic Film on Platinum Electrodes, S. Gilaman
Oscillographic Polarography at Controlled Alternating Current,
Michael Heyrovksy and Karel Micka
VOLUME 3
Application of Controlled-Current Coulometry to Reaction Kinetics,
Jiri Janata and Harry B. Mark, Jr.
Nonaqueous Solvents for Electrochemical Use, Charles K. Mann
Use of the Radioactive-Tracer Method for the Investigation of the
Electric Double-Layer Structure, N. A. Balashova and
V. E. Kazarinov
Digital Simulation: A General Method for Solving Electrochemical
Diusion-Kinetic Problems, Stephen W. Feldberg
xiii
VOLUME 4
Sine Wave Methods in the Study of Electrode Processes, Margaretha
Sluyters-Rehbach and Jan H. Sluyters
The Theory and Practice of Electrochemistry with Thin Layer Cells,
A. T. Hubbard and F. C. Anson
Application of Controlled Potential Coulometry to the Study of
Electrode Reactions, Allen J. Bard and
K. S. V. Santhanam
VOLUME 5
Hydrated Electrons and Electrochemistry, Geraldine A. Kenney and
David C. Walker
The Fundamentals of Metal Deposition, J. A. Harrison and H. R. Thirsk
Chemical Reactions in Polarography, Rolando Guidelli
VOLUME 6
Electrochemistry of Biological Compounds, A. L. Underwood and
Robert W. Burnett
Electrode Processes in Solid Electrolyte Systems, Douglas O. Raleigh
The Fundamental Principles of Current Distribution and Mass Transport
in Electrochemical Cells, John Newman
VOLUME 7
Spectroelectrochemistry at Optically Transparent Electrodes; I. Electrodes
Under Semi-innite Diusion Conditions, Theodore Kuwana and
Nicholas Winograd
Organometallic Electrochemistry, Michael D. Morris
Faradaic Rectication Method and Its Applications in the Study of
Electrode Processes, H. P. Agarwal
Contents of Other Volumes xiv
VOLUME 8
Techniques, Apparatus, and Analytical Applications of Controlled-
Potential Coulometry, Jackson E. Harrar
Streaming Maxima in Polarography, Henry H. Bauer
Solute Behavior in Solvents and Melts, A Study by Use of Transfer
Activity Coecients, Denise Bauer and Mylene Breant
VOLUME 9
Chemisorption at Electrodes: Hydrogen and Oxygen on Noble Metals
and their Alloys, Ronald Woods
Pulse Radiolysis and Polarography: Electrode Reactions of Short-lived
Free Radicals, Armin Henglein
VOLUME 10
Techniques of Electrogenerated Chemiluminescence, Larry R. Faulkner
and Allen J. Bard
Electron Spin Resonance and Electrochemistry, Ted M. McKinney
VOLUME 11
Charge Transfer Processes at Semiconductor Electrodes,
R. Memming
Methods for Electroanalysis In Vivo, Jir Koryta, Miroslav Brezina,
Jir Prada c , and Jarmila Prada c ova
Polarography and Related Electroanalytical Techniques in
Pharmacy and Pharmacology, G. J. Patriarche,
M. Chateau-Gosselin, J. L. Vandenbalck,
and Petr Zuman
Polarography of Antibiotics and Antibacterial Agents,
Howard Siegerman
Contents of Other Volumes xv
VOLUME 12
Flow Electrolysis with Extended-Surface Electrodes, Roman E. Sioda
and Kenneth B. Keating
Voltammetric Methods for the Studyof AdsorbedSpecies, Etienne Laviron
Coulostatic Pulse Techniques, Herman P. van Leeuwen
VOLUME 13
Spectroelectrochemistry at Optically Transparent Electrodes,
II. Electrodes Under Thin-Layer and Semi-innite Diusion
Conditions and Indirect Coulometric Iterations, William H.
Heineman, Fred M. Hawkridge, and Henry N. Blount
Polynomial Approximation Techniques for Dierential Equations in
Electrochemical Problems, Stanley Pons
Chemically Modied Electrodes, Royce W. Murray
VOLUME 14
Precision in Linear Sweep and Cyclic Voltammetry, Vernon D. Parker
Conformational Change and Isomerization Associated with Electrode
Reactions, Dennis H. Evans and Kathleen M. OConnell
Square-Wave Voltammetry, Janet Osteryoung and John J. ODea
Infrared Vibrational Spectroscopy of the Electron-Solution Interface,
John K. Foley, Carol Korzeniewski, John L. Dashbach, and
Stanley Pons
VOLUME 15
Electrochemistry of Liquid-Liquid Interfaces, H. H. J. Girault and
D. J. Schirin
Ellipsometry: Principles and Recent Applications in Electrochemistry,
Shimson Gottesfeld
Voltammetry at Ultramicroelectrodes, R. Mark Wightman and
David O. Wipf
Contents of Other Volumes xvi
VOLUME 16
Voltammetry Following Nonelectrolytic Preconcentration, Joseph Wang
Hydrodynamic Voltammetry in Continous-Flow Analysis, Hari
Gunasingham and Bernard Fleet
Electrochemical Aspects of Low-Dimensional Molecular Solids, Michael
D. Ward
VOLUME 17
Applications of the Quartz Crystal Microbalance to Electrochemistry,
Daniel A. Buttry
Optical Second Harmonic Generation as an In Situ Probe of
Electrochemical Interfaces, Geraldine L. Richmond
New Developments in Electrochemical Mass Spectroscopy,
Barbara Bittins-Cattaneo, Eduardo Cattaneo, Peter Ko nigshoven,
and Wolf Vielstich
Carbon Electrodes: Structural Eects on Electron Transfer Kinetics,
Richard L. McCreery
VOLUME 18
Electrochemistry in Micelles, Microemulsions, and Related
Microheterogeneous Fluids, James F. Rusling
Mechanism of Charge Transport in Polymer-Modied Electrodes,
Gyo rgy Inzelt
Scanning Electrochemical Microscopy, Allen J. Bard, Fu-Ren F. Fan,
and Michael V. Mirkin
VOLUME 19
Numerical Simulation of Electroanalytical Experiments: Recent Advances
in Methodology, Bernd Speiser
Electrochemistry of Organized Monolayers of Thiols and Related
Molecules on Electrodes, Harry O. Finklea
Contents of Other Volumes xvii
Electrochemistry of High-T
c
Superconductors, John T. McDevitt,
Steven G. Haupt, and Chris E. Jones
VOLUME 20
Voltammetry of Solid Microparticles Immobilized on Electrode Surfaces,
Fritz Scholz and Birgit Meyer
Analysis in Highly Concentrated Solutions: Potentiometric, Conductance,
Evanescent, Densometric, and Spectroscopic Methodologies,
Stuart Licht
Surface Plasmon Resonance Measurements of Ultrathin Organic Films
at Electrode Surfaces, Dennis G. Hanken, Claire E. Jordan, Brian
L. Frey, and Robert M. Corn
Electrochemistry in Neuronal Microenvironments, Rose A. Clark, Susan
E. Zerby, and Andrew G. Ewing
VOLUME 21
Template-Synthesized Nanomaterials in Electrochemistry, Charles R.
Martin and David T. Mitchell
Electrochemical Atomic Layer Epitaxy, John L. Stickney
Scanning Tunneling Microscopy Studies of Metal Electrodes, T. P. Moat
Contents of Other Volumes xviii
LOOKING AT THE METAL/SOLUTION INTERFACE
WITH THE ELECTROCHEMICAL QUARTZ CRYSTAL
MICROBALANCE: THEORY AND EXPERIMENT
V. Tsionsky, L. Daikhin, M. Urbakh, and E. Gileadi
School of Chemistry
Raymond and Beverly Sackler Faculty of Exact Sciences, Tel Aviv University
Ramat Aviv, Israel
I. INTRODUCTION 2
A. Is It Really a Microbalance? 3
B. Applications of the Quartz Crystal Microbalance 4
C. The Impedance Spectrum of the EQCM 5
D. Outline of This Chapter 8
II. THEORETICAL INTERPRETATION OF THE QCM
RESPONSE 8
A. Impedance 8
B. The Eect of Thin Surface Films 12
C. The Quartz Crystal Operating in Contact
with a Liquid 16
D. Quartz Crystals with Rough Surfaces 26
III. ELECTRICAL DOUBLE LAYER/ELECTROSTATIC
ADSORPTION 33
A. Introduction 33
B. Some Typical Results 34
C. The Potential Dependence of the Frequency 36
IV. ADSORPTION STUDIES 43
A. The Adsorption of Organic Substances 43
B. The Adsorption of Inorganic Species 53
V. METAL DEPOSITION 60
A. Deposition on the Same Metal Substrate 60
B. Early Stages of Metal Deposition on a
Foreign Substrate 64
1
VI. THE INFLUENCE OF ROUGHNESS ON THE
RESPONSE OF THE QCM IN LIQUIDS 70
A. The Nonelectrochemical Case 71
B. The Electrochemical Case 76
VII. CONCLUSION 83
VIII. APPENDIX 86
A. Nonuniform Film on the Surface 86
B. Experimental Remarks 86
References 94
I. INTRODUCTION
The literature concerning the quartz crystal microbalance (QCM) and its
electrochemical analogue, the electrochemical crystal microbalance
(EQCM) is wide and diverse. Many reviews are available in the literature,
discussing the fundamental properties of this device and its numerous
applications, including its use in electrochemistry [15]. In this chapter we
concentrate on electrochemical applications, specically in studies of
submonolayer phenomena and the interaction of the vibrating crystal with
the electrolyte in contact with it.
A few examples are treated in detail here, and the advantages and
limitations of the EQCM as a tool for the study of fundamental phenom-
ena at the metal/solution interface are discussed.
When the quartz crystal microbalance was rst introduced in 1959
[6], it represented a major step forward in our ability to weigh matter. Until
then, routine measurements allowed an accuracy of 0.1 mg, and highly
sensitive measurements could be made with an accuracy limit of 0.3 Ag
under well-controlled experimental conditions, (see Ref. 7). The QCM
extended the sensitivity by two or three orders of magnitude, into the sub-
nanogram regime.
Even used in vacuum or in an inert gas atmosphere at ambient
pressure, the QCM acts as a balance only under certain conditions, as
discussed below. Then the change of mass caused by adsorption or de-
position of a substance from the gas phase can be related directly to the
change of frequency by the simple equation derived by Sauerbrey [6]:
Df C
m
Dm 1
Tsionsky et al. 2
where C
m
is a constant, representing the mass sensitivity, which is related to
known properties of quartz and the dimensions of the crystal, and Dmis the
added mass density, in units of g/cm
2
.
A. Is It Really a Microbalance?
Is the quartz crystal microbalance really a microbalance? For one thing, it
should rightly be called a nano-balance, considering that the sensitivity of
modern-day devices is on the order of 12 ng/cm
2
and could be pushed
further, if necessary. More importantly, calling it a balance implies that the
Sauerbrey equation applies strictly, namely that the frequency shift is the
sole result of mass loading. It is well known that this is not the case, and
the frequency shift observed could more appropriately be expressed by a
sum of terms of the form
Df Df
m
Df
g
Df
P
Df
R
Df
sl
Df
T
2
where the dierent terms on the right-hand side (rhs) of this equation
represent the eects of mass loading, viscosity and density of the mediumin
contact with the vibrating crystal, the hydrostatic pressure, the surface
roughness, the slippage eect, and the temperature, respectively, and the
dierent contributions can be interdependent. Even this equation does not
tell the whole story, certainly not when the device is immersed in a liquid or
in gas at high pressure. It does not account for solution occluded between
the ridges of a rough surface or in the pores of a porous substrate. The
nature of the interaction between the liquid and the surface, the type of
roughness, and internal stress or strain could all aect the response of the
quartz crystal resonator. These eects become of major importance
particularly when small changes of frequency, associated with submono-
layer phenomena, are considered. Some of these factors will be discussed in
this chapter.
It should be evident from the above arguments that the term quartz
crystal microbalance is a misnomer, which could (and indeed has) lead to
erroneous interpretation of the results obtained by this useful device. It
would be helpful to rename it the quartz crystal sensor (QCS), which
describes what it really doesit is a sensor that responds to its nearest
environment on the nano-scale. However, it may be too late to change the
widely used name. The QCM or its analogue in electrochemistry, the
EQCM, can each act as a nano-balance under specic conditions, but not
in general.
Electrochemical Quartz Crystal Microbalance 3
B. Applications of the Quartz Crystal Microbalance
The most common commercial use of the QCM is as a thickness gauge in
thin-layer technology. When used to monitor the thickness of a metal lm
during physical or chemical vapor deposition, it acts very closely as a nano-
balance, providing a real-time measurement of the thickness. Indeed,
devices sold for this purpose are usually calibrated in units of thickness
(having a dierent scale for each metal, of course), and claima sensitivity of
less than 0.1 nm, which implies a sensitivity of less than a monolayer.
The other common application of the QCM is as a nano-sensor
proper, made sensitive to one gas or another by suitable surface treatment.
Selecting the suitable coating on the electrodes of the QCM can determine
selectivity and enhance sensitivity. It is not our purpose to discuss sensors
in the present review. It should only be pointed out that any such sensor
would have to be calibrated, since the Sauerbrey equation would not be
expected to apply quantitatively.
1. Applications for Gas-Phase Adsorption
The high sensitivity of the QCM should make it an ideal tool for the study
of adsorption from the gas phase. We note that the number of sites on the
surface of a metal is typically 1.3 10
15
/cm
2
, hence a monolayer of a small
adsorbate, occupying a single site, would be about 2.2 nmol/cm
2
. A
monolayer of water would therefore weigh about 40 ng/cm
2
, while a
monolayer of pyridine would weigh 3060 ng/cm
2
, depending on its
orientation on the surface. Comparing these numbers with the sensitivity
of 2 ng/cm
2
shows that adsorption isotherms could be measured in the gas
phase employing the QCM. This has not been done properly until relatively
recently, mainly because the device was treated as a microbalance, i.e., it
was assumed that the Sauerbrey equation could be applied, and several
important terms in Eq. (2) were ignored. Obtaining adsorption isotherm
one has to change the pressure over a wide range. Therefore, the changes of
properties of the surrounding gas cannot be ignored. This shortcoming was
overcome by the present authors [8], who developed the supporting gas
method. When this method is employed, the overall pressure is kept
constant by a large excess of an inert gas, and the frequency shift of the
QCMis measured as a function of the partial pressure of the material being
investigated. In this manner all terms in Eq. (2), other than Df
m
, are
essentially zero, and the device acts as a true nano-balance. One intriguing
result that was obtained using this method came from a comparison of
the adsorption of benzene and pyridine on a gold surface. It was found
Tsionsky et al. 4
that a monolayer of pyridine weighs roughly twice as much as a monolayer
of benzene. Since the two molecules have almost the same size and
molecular weight, it must be concluded that their conguration in the
adsorbed state is dierent. Benzene is probably adsorbed at on the
surface, while pyridine must be adsorbed perpendicular to it, occupying
only half as many sites.
Although the nominal resolution of 2 ng/cm
2
should be enough to
study the adsorption isotherm if the monolayer weighs around 3060 ng/
cm
2
, it is somewhat marginal, and an increase of sensitivity of about one
order of magnitude would be desirable. Part of this enhancement could be
achieved by increasing the roughness factor on the atomic scale, without
inuencing the roughness on a scale relevant to the resonance frequency
(see Sec. VI).
2. Use of QCM in Liquids
It was not initially obvious that the quartz crystal resonator would operate
in liquids until this was proven experimentally [9,10]. The term associated
with the inuence of the viscosity and density of liquid in Eq. (2) can be
written [11] as
Df
g
C
g
gq
1=2
3
Since the product of

gq
p
in liquids is about two orders of magnitude
higher than in gases at ambient pressure, the crystal is heavily loaded when
transferred from the gas phase into a liquid.
Once the door had been opened to its use in liquids, the potential of
the QCM for interfacial electrochemistry was obvious, and the EQCM
became popular.
When a QCM is placed in contact with a dilute aqueous solution, the
frequency should shift to lower values by about 0.7 kHz according to Eq.
(3). In practice, a shift of 1.02 kH is observed, depending on the surface
roughness. The eect of roughness is also related indirectly to viscosity and
density, since the hydrodynamic ow regime at the surface is altered as a
result of roughness [1214]. Roughness is a major issue in the interpreta-
tion of the response of the QCMin liquids, and it is discussed in some detail
in the following sections.
C. The Impedance Spectrum of the EQCM
In early studies of the QCM and the EQCM, only the resonance frequency
was determined and conclusions were drawn based on the shift of
frequency. Unfortunately, in many cases this shift was attributed to mass
loading alone, and it was used to calculate the weight added or removed
from the surface, disregarding other factors that aect the frequency. In
the past decade, more and more laboratories expanded such studies to
include measurements of the impedance spectrum of the crystal [1525].
This provides an additional experimental variable that can obviously yield
further information and a deeper understanding of the structure of the
interface. For instance, a variation in the resonance width provides
unambiguous proof that mechanisms other than mass loading are also
involved.
A series of typical admittance spectra are presented here. In Fig. 1a
we show a simple case of metal deposition (gold on a gold substrate). The
EQCM acts as a true microbalance in this case. The resonance frequency is
shifted to lower values with increasing load, but the shape of the spectrum
remains unaltered.
In Fig. 1b the eect of viscosity on the admittance spectrumis shown.
Here again the resonance frequency is shifted to lower values with
increasing viscosity, but this has nothing to do with mass loading.
However, the shape of the spectrum is quite dierent, and the width at
half-height (see below) increases dramatically with increasing viscosity and
density of the liquid. Line 1 and the inset in this gure showthe response of
the QCMin H
2
at ambient pressure. The product of viscosity and density is
about four orders of magnitude smaller than in any of the liquids.
Correspondingly, the width of the resonance is only about 20 Hz,
compared to about 2.5 kHz in the liquid corresponding to line 2.
Another aspect of the admittance spectrum is shown in Fig. 1c. Here
the same metal deposition was conducted as in Fig. 1a, but the conditions
were purposely chosen to produce a very rough surface (by plating at a
current density close to the mass-transport limited value). The width of the
resonance is increased and the frequency is shifted to lower values with
increasing roughness.
We chose rather extreme cases of viscosity and roughness in Fig. 1b
and 1c, for the purpose of illustration. The corresponding shift in fre-
quency is very high, in the range of 515 kHz, as compared to changes of
frequency of 540 Hz typically observed in the studies of electrosorption,
double layer, upd, and other submonolayer phenomena. The important
conclusion is that even very small changes of viscosity and/or surface
roughness (produced inadvertently) could lead to a shift of frequency
comparable to that expected for such submonolayer phenomena, and the
Tsionsky et al. 6
FIG. 1. (a) The real part of the admittance versus frequency: during deposition
of gold on a gold-covered EQCM at a current density of 20 AA/cm
2
. (c) The same
at 500 AA/cm
2
. (b) The response of the QCM immersed in dierent media: 1,
hydrogen, 1 atm; 2, dimethyl ether; 3, water; 4 and 5, 40% and 50% aqueous
solutions of sucrose, respectively. (Inset) Admittance for H
2
, on an expanded scale.
Arrow gq shows the increase of product gq. (From Ref. 24.)
change of frequency cannot be generally interpreted to be a result of mass
loading alone.
D. Outline of This Chapter
This chapter contains theoretical and experimental sections. In the theo-
retical section we consider dierent aspects of the behavior of the vibrating
resonator: when it was loaded by additional mass, immersed in viscous
media, has undergone changes in surface roughness, etc. We discuss the
universal perturbation theory of the inuence of slightly rough surfaces on
the QCM response and consider the special model for strong roughness,
noting that a general model does not exist for such surfaces. Special
attention was paid to consideration of the inuence of slippage on the
QCM at the solid/electrolyte interface.
The QCM is now so widely and extensively used that, in the frame-
work of this chapter, it is not possible to review all the available litera-
ture. Hence we limited ourselves here to a review of the experimental data
and ideas concerning the studies of submonolayer adsorption and inter-
actions taking place at the metal/solution interface. In other words, this
review is restricted to the use of the QCM in fundamental electrochem-
istry. Furthermore, we did not include studies of electrochemical kinetics
with the help of the EQCM, which merits a separate review. The problems
of the interpretaion of the EQCM response caused by changes taking
place at the metal/solution interface are obviously of rst priority.
We did not present here a full description of the operation of an
EQCM. This topic is well described in previous reviews (see Refs. 1,2) and
in many articles published in readily accessible electrochemical journals.
However, a few aspects of the experiments with the EQCM are covered in
the Appendix (Sec. VIII.B).
II. THEORETICAL INTERPRETATION OF THE QCM
RESPONSE
A. Impedance
The shear mode resonator consists of a thin disk of AT-cut quartz crystal
with electrodes coated on both sides. The application of a voltage between
these electrodes results in a shear deformation of the crystal due to its
piezoelectric properties. The crystal can be electrically excited into a
Tsionsky et al. 8
number of resonance modes, each corresponding to a unique standing
shear wave across the thickness of the crystal. If a quartz resonator
operates in contact with an outer medium, the oscillating surface interacts
mechanically with the medium and excites motion in it. The mechanical
properties of the medium in contact are reected in the response of the
resonator.
The geometry of the system consisting of a quartz crystal in contact
with the outer medium is schematically shown in Fig. 2. The z-axis is
plotted perpendicular to the plane of contact - the plane z =0 coinciding
with the unconstrained face of the quartz resonator, and the plane z = d is
its constrained face. The thickness of the quartz crystal is d.
When an ac voltage is applied between the electrodes, the motion of
the AT cut quartz crystal can be described by a system of two coupled
dierential equations, which constitute the wave equation for elastic
displacements, u(z,t) = u(z,x) exp(ixt), and the equations that establish
FIG. 2. Schematic sketch of the quartz crystal resonator in contact with a liquid.
The contacting medium is a thin lm rigidly attached to the crystal surface from
one side, at z = d. The opposite surface of the crystal (z =0) is unconstrained. d is
the thickness of the quartz crystal.
the relationship between displacements and the electrostatic poten-
tial,u(z,t) =u(z,x) exp(ixt), [26] are
x
2
q
q
u z; x c
66
d
2
dz
2
u z; x 4
e
22
d
2
dz
2
u z; x e
26
d
2
dz
2
u z; x 5
Here, c
66
l
q
e
2
26
=e
22

ixg
q
; q
q
; l
q
are the density and shear modu-
lus of quartz,e
22
,e
26
are the dielectric constant, and the piezoelectric stress
coecient of quartz, g
q
, is its ctitious viscosity, x=2p f is the angular
frequency, and f is the frequency. Equations (4) and (5) are solved under
the following boundary conditions:
1. At the plane z =0, the potential equals u
0
and the stress is zero.
2. At the plane z = d, the potential equals u
0
and the ratio of the
shear stress, c
66
du(z,x)/dz, acting on the contacting medium to
the surface velocity, ixu(d,x), equals Z
out
.
Here Z
out
is the mechanical impedance of the medium contacting the
quartz surface. Solution of Eqs. (4) and (5) yields the following expression
for the admittance of the quartz resonator [27,28]:
Y ixC
0
Z
1
m
6
where C
0
=e
22
/d is the static capacitance and Z
m
is the motional imped-
ance:
Z
m

1
ixC
0
/
q
K
2
q
2tan /
q
=2
1

/
q
4K
2
q
xC
0
Z
out
=Z
q
1
iZ
out
=Z
q
2tan /
q
=2

7
and K
2
q
e
2
26
=e
22
c
66
; /
q
k
q
d; Z
q
k
q
c
66
=x, and k
q
x

q
q
=c
66
is the
wave number of the shear wave in quartz. The rst term in Eq. (7) de-
scribes the motional resistance of an unloaded quartz resonator. The sec-
ond term arises from surface loading and includes the properties of the
electrode surfaces and the contacting medium through Z
out
.
In QCM experiments the surface loading is relatively small [27], that
is, |Z
out
/Z
q
| << 2tan(/
q
/2). Under these conditions, the shift of the reso-
Tsionsky et al. 10
nant frequency of the quartz-crystal resonator with respect to the resonant
frequency of the unloaded quartz crystal, f
0
, can be written as [12,29]
D

f u Df i
G
2
i
f
0
p
Z
out
Z
q
8
It should be noted that the frequency shift Df
can be a complex number,

and its imaginary part, G, reects the width of the resonance. Equation (8)
shows that the complex frequency shift Df
contains the same information

as the mechanical impedance Z
out
.
The admittance of the quartz resonator can be presented in terms of
an electrical equivalent circuit [24,15,3036]. The equivalent circuit for
the unloaded quartz crystal consists of a motional branch, which reects
the vibration of the quartz, and a static capacitance, which is in parallel
with the motional branch. The motional branch includes a resistance,
capacitance, and inductance connected in series. The relationships between
the electrical elements and the mechanical parameters describing the
crystal motion (mass, compliance, and damping coecient) were consid-
ered in Refs. 3739. When the surface loading is small, |Z
out
/Z
q
| <<
2tan(/
q
/2), the inuence of the contacting medium enters the equivalent
circuit through the mechanical impedance connected in series with the
motional branch (see Fig. 3). This impedance reects the properties of the
contacting medium. In particular, mass loading and energy dissipation in
the medium can be taken into account through inductance and resistance,
respectively. The equivalent circuit representation has been used in a
number of works to simulate the response of the QCM [14,15,38,4047].
This approach is similar to that employed in the analysis of electro-
chemical impedance spectroscopy, and both suer from the same draw-
back: It is necessary to prove that the equivalent circuit is indeed
equivalent. This may be possible in certain simple situations, for example,
the equivalent circuit for a charge-transfer process occurring under purely
activation control, on a homogeneous surface, or the impedance of the
EQCM on an ideally smooth surface. Unfortunately, these are rather rare
cases. In most cases encountered, the response of the interface is much
more complex. It is still possible to use the equivalent circuit approach, but
one must increase the number of circuit elements to t the data to the model
over a suciently large range of frequency. This requires a parameter-
tting procedure involving numerous adjustable parameters, which makes
the validity of the model rather dubious. In the case of the EQCM, for
example, it is not obvious how one would account for a surface roughness
of unspecied nature (slight roughness, strong roughness, or some combi-
nation of both) in terms of an equivalent circuit.
In order to analyze the inuence of the dierent loading mechanisms
on the QCM response, one has to model a dependence of the mechanical
impedance Z
out
or the complex resonance frequency shift on the chemical
and physical properties of the contacting medium. Various models for the
mechanical contact between the oscillating quartz crystal and the outer
medium are discussed below.
B. The Effect of Thin Surface Films
1. Uniform Film Rigidly Attached to the Surface
First we consider the eect of a thin lm, rigidly attached to an ideally at
crystal surface, on the response of the quartz crystal resonator (see Fig. 2).
FIG. 3. The Butterworthvan Dyke equivalent circuit of the loaded quartz
crystal resonator. The parameters R, C, and L describe the behavior of the
unloaded quartz resonator; Z
out
is the impedance of the contacting medium.
Tsionsky et al. 12
For a homogeneous thin lm with a thickness smaller than the
wavelength of the shear oscillations, the shift of the resonance frequency
can be expressed in terms of the change in surface mass density of the lm,
Dm
f
(in g/cm
2
). This was given by Sauerbrey [6] as
Df
2f
2
0
Dm
f
l
q
q
q

1=2
9
This equation coincides with Eq. (1) with C
m
2f
2
0
= l
q
q
q

1=2
. Equation
(9) can be derived by supplementing the wave equation [Eq. (4)] with the
Newtonian equation of motion for the surface lm:
Dm
f
x
2
u
f
x l
q
d
dz
u z; w at z d 10
where u
f
(x) is the displacement of the lm. Here the shear stress,
l
q
du(z,x)/dz, plays the rule of the external force acting on the lm. Here
and everywhere below we use in Eq. (4) l
q
instead of c
66
, neglecting small
values xg
q
and e
2
26
=e
22
. Solving Eqs. (4) and (10) under the standard
boundary condition, namely that (1) at the unrestricted surface, z =0, the
shear stress equals zero and (2) at the surface z =d the quartz surface
displacement is equal to the surface lm displacement, one obtains the
Sauerbrey equation. From Eq. (10) one can see that the frequency shift is
determined by the inertial force of the lm acting on the quartz surface.
Equation (9) shows that the addition of mass rigidly attached to the
surface of the quartz-crystal resonator leads to a decrease of the resonant
frequency, but it does not inuence the width of the resonance.
2. Nonuniform Film Rigidly Attached to the Surface
A natural question arises whether a nonhomogeneous mass distribution
can lead to an additional shift of frequency and/or to a broadening of the
resonance, compared to the result given by the Sauerbrey equation?
In order to answer this question, we consider here the eect of
nonuniform mass loading on the response of the QCM. Lateral displace-
ment of the nonuniform lm can be described by the equation of motion:
Dm
f
Rx
2
u
f
R l
q
d
dz
uz; R for z d 11
where u
f
(R) and u(z,R) represent the displacements of the surface lm and
quartz crystal and Dm
f
(R) is the surface lmdensity, as a function of lateral
coordinates, R. This equation plays the role of a boundary condition for
the wave equation describing the shear-mode oscillations in the quartz
crystal. Equations (4) and (11), with the standard boundary condition, lead
to the following equation for determination of the resonant frequency (see
Appendix):
l
q
k tankd x
2
Dm
f
x
4
Dm
2
l
dK
2p
2
gK
l
q
pKtanpKd x
2
Dm
f
12
Here, k =x(q
q
/l
q
)
1/2
is the wave number of the shear wave in the quartz,
p(K) = k
2
K
2
, and K is the two-dimensional tangential wave vector.
Considering mechanical response of the quartz crystal we use k instead
of k
q
[see text following Eqs. (4) and (5)] because xg
q
and e
2
26
=e
22
are
small compared to l
q
. Also, Dm
l
is the root mean square deviation of the
mass density from the average value Dm
f
, and g(K) is the correlation
function, which describes a nonuniform mass distribution along the
surface. Equation (12) is a general form of the Sauerbrey equation,
applicable for the case of an inhomogeneous surface lm. For uniform
mass distribution (corresponds to Dm
l
=0), it yields the Sauerbrey
equation in its usual form.
Assuming that the correlation function g(K) has a Gaussian form,
with a lateral correlation length, l, Eq. (12) can be solved analytically for
two limiting cases, kl >>1 and kl<<1. When the correlation length of the
mass distribution is larger than the shear mode wavelength, kl >>1,
splitting of the resonant frequency occurs, and the frequency shift can be
estimate as
Df
2f
2
0
q
q
l
q
p
Dm
f
FDm
l

13
In contrast to the case of uniform mass loading, Dm
1
f
R 0, two
values of the resonance frequency appear. This eect can be simulated by
a simple equivalent circuit consisting of two Butterworthvan Dyke [33
35] circuits in series with the inductances corresponding to the two dif-
ferent values of the surface mass densities, Dm
f
Dm
l
and Dm
f
Dm
l
. Due
to overlap of these two resonance states, splitting can manifest itself as a
broadening of the resonance, which will have an eective width of the
Tsionsky et al. 14
order of 2f
2
0
Dm
l
=p l
q
q
q

1=2
. For the 6 MHz quartz resonator this
broadening eect becomes important when the correlation length l is
larger than 0.02 cm.
Inthe secondlimiting case, kl <<1, the eect of the nonuniformmass
loading can be neglected and the standard Sauerbrey equation applies.
3. Slippage at the Interface Between a Thin Film and a Solid
The Sauerbrey equation shows that a thin uniform lm rigidly attached to
the quartz surface does not inuence the width of the mechanical reso-
nance. However, it was experimentally shown for a number of systems
that adsorption on the quartz surface produced both a shift of frequency
and an increase of the width of the resonance [4852]. This phenomenon
can be explained, assuming slippage at the adsorbate/substrate interface.
Slippage occurs as a result of the force of inertia acting on the
adsorbate during the vibrational motion of the crystal. The force of inertia,
F, is extremely weak (f10
13
dyne per atom) [53] and cannot, by itself,
move an adsorbed species over the lateral energy barriers of the adsorbate-
substrate potential [53]. However, this force decreases the barriers in the
direction of F, which leads to a thermally activated drift of the adsorbate in
the direction opposite to the motion of the crystal surface. As a result, the
instantaneous velocity in the adsorbate layer can dier from the velocity of
the surface of the quartz-crystal resonator.
The slippage at the interface between a thin lm of density Dm
a
and
the substrate is usually described in terms of an interfacial friction
coecient (coecient of sliding friction), v. This coecient determines
the stress acting between the lmand the substrate, which move at dierent
velocities. An innite value of v implies that the non-slip (sticking)
boundary condition is applicable. When the interfacial friction coecient
equals zero, the lm is free to slide with no energy dissipation.
The motion of the adsorbed lmon the oscillating quartz surface can
be described by the equation
Dm
a
d
dt
v
f
t v v
f
t v
q
t

14
where v
q
(t) = v
q0
(x)exp(ixt) and v
f
(t) = v
f 0
(x)exp(ixt) are the velocities
of the crystal surface and of the lm. Simultaneous solution of the wave
equation [Eq. (4)] and the equation of motion [Eq. (14)] for the adsorbed
lm yields the following expressions for the changes of the frequency, Df,
and the width of the resonance, G:
Df
2f
2
0
Dm
a
q
q
l
q
p
v
2
v
2
2pf
0
Dm
a

2

15
G
4f
2
0
Dm
a
q
q
l
q
p
2pf
0
Dm
a
v
v
2
2pf
0
Dm
a

2

16
Note that
G
Df
4pf
0
Dm
a
v
17
Thus, the interfacial friction can be evaluated from measurement of G and
Df. This procedure has been applied to a number of systems in which weak
physical adsorption occurs, such as the adsorption of Xe, Kr, N
2
on Au
and of H
2
O and C
6
H
12
on Ag [4852]. In all above cases, slippage was
observed and the ratio of the coecient of sliding friction to the mass
density was of the order v/Dm
a
= (10
8
10
9
)s
1
. As an example, the
frictional stress acting on the monolayer Xe lm sliding on a Ag (111)
surface at a velocity v =1 nm / s, F =vv, equals about 10 N/m
2
[54]. It is
much smaller than typical shear stresses involved in sliding of a steel block
on a steel surface under boundary lubrication condition. The shear stress in
the latter case is of the order c10
8
N/m
2
[53].
In a recent paper [55] the dependence of the slip time, s
s
, on the
amplitude of the crystal surface oscillations, A, and the surface coverages
was investigated. The results refer to the absorption of krypton atoms on
gold at 85jK. The slip time is related to the interfacial friction coecient, v,
as s
s
= Dm
a
/v. It was found that there is a step-like transition between a
low-coverage region, where slippage exists at the solid/lm interface, and a
high-coverage region where the lm is locked to the surface. The transition
occurs at dierent coverages depending on the amplitude, A. Independent
of coverage, the lm is attached rigidly to the surface for A V0.18 nm and
slides for A>0.4 nm. In the region of sliding at small coverages, the values
of the slip time are in the interval 210 nsec, for 0.18 nm < A < 0.4 nm.
C. The Quartz Crystal Operating in Contact with a Liquid
1. General Considerations
When a quartz crystal resonator operates in contact with a liquid, the shear
motion of the surface generates motion in the liquid near the interface. The
Tsionsky et al. 16
velocity eld, v(r,x) related to this motion in a semi-innite Newtonian
liquid is described by the linearized Navier-Stokes equation:
ixqvr; x jPr; x gDvr; x 18
where P(r,x), g, and q are pressure, viscosity, and density of the liquid,
respectively. Under the conditions of the QCM experiments, where the
shear velocities are much smaller than the sound velocity in the liquid, the
displacement of the crystal does not generate compressional waves and a
liquid can be considered as an incompressible one. If the surface is
suciently smooth, the quartz oscillations generate plane-parallel laminar
ow in the liquid, as shown in Fig. 4. The velocity eld obtained as the
solution of Eq. (18) for a at surface has the form
v
x
z v
q0
xexp1 iz=d 19
where v
q0
(x) is the velocity of the liquid at the surface and d
2g=x
0
q
.
Equation (19) represents a damped shear wave radiating into the liquid
from the surface of the oscillating resonator. d is the velocity decay length
of this shear wave, which lies between 250 and 177 nm, for dilute aqueous
solutions at room temperature, for crystals having a fundamental frequen-
cy in the range of 510 MHz. Damping of the shear wave has a number of
important consequences. First, it ensures that the quartz crystal can
operate in liquids, the losses in the liquid being limited by the nite depth
of penetration. Second, a small portion of the liquid is coupled to the
crystal motion and a frequency decrease is observed. Third, the viscous
nature of motion gives rise to energy losses, which are sensed by the
resonator, both as a decrease in frequency and as an increase in the width of
the resonance.
2. The Nonslip Boundary Condition
The response of the QCM at the solid/liquid interface can be found by
matching the stress and the velocity elds in the media in contact. It is
usually assumed that the relative velocity at the boundary between the
liquid and the solid is zero. This corresponds to the nonslip boundary
condition. Strong experimental evidence supports this assumption on the
macroscopic scales [56,57]. In this case the frequency shift, Df
l
, and the
width of the resonance, G
l
, can be written as follows [10,11]:
D f
l

f
3=2
0
qg
p
pq
q
l
q
p
20
G
l
2
f
3=2
0
qg
p
pq
q
l
q
p
21
Equations (20) and (21) show that the generation of a damped
laminar ow in the liquid causes a decrease in the resonance frequency
and an increase in the resonance width, which are both proportional to
qg
p
. In contrast to the case of the mass loading, where Df is proportional
to f
0
2
, the liquid induced response of the QCM is proportional to f
0
3/2
.
FIG. 4. The system geometry and the velocity distribution. Curves 1 and 2
represent the velocity distributions at the liquid/adsorbate interface without and
with slippage, respectively. Curve 3 is the velocity distribution in the quartz. The
thickness of various layers is not drawn to scale.
Tsionsky et al. 18
It is interesting to note that for both a surface lm rigidly attached to
the resonator and a liquid in contact with the surface of the quartz crystal,
the shift of the resonant frequency can be written in the same form, as
Df f
0
q
q
q
kh
eff
22
where k x
0
q
q
=l
q
p
, q is the bulk density of the medium in contact with
the vibrating surface of the solid, lm, or liquid, and h
e
is the thickness of
the layer that responds to the quartz oscillations. In the case of a lm, h
e
coincides with the thickness. For a semi-innite liquid, h
e
presents a
thickness of liquid involved in the motion, and it should be taken equal to
d/2. The dierence in the frequency dependence of the QCM response in
the two cases is a result of the frequency dependency of d. However, in
contrast to the case of pure mass loading, the eect of a liquid results not
only in a frequency shift, but also in a broadening of the resonance.
a. Eect of a Thin Liquid Film at the Interface
The properties (the eective viscosity and density) of the liquid layer in
close vicinity to the interface can dier from their bulk values. There are
various reasons for these phenomena. For example, the properties of a
thin liquid layer conned between solid walls are determined by interac-
tions with the solid walls [58,59]. In electrochemical system the structuring
of a solvent induced by the substrate and a nonuniform ion distribution in
the diuse double layer can signicantly inuence the properties of the
solution at the interface. The nonuniform distribution of species, which
inuences the properties of the liquid near the electrode, also occurs in the
case of diusion kinetics. The latter was considered in Ref. 60, where the
ferro/ferri redox system was studied by the EQCM. This was the case
where the velocity decay length (>25 Am) was much less than the thickness
of the diusion layer (>100 Am), in which the composition of the solution
is dierent from the bulk composition.
Nonuniform distribution of species results in nonuniform distribu-
tion of the properties of liquid near the vibrating surface of the resonator.
The properties change with distance from the interface, until the values
corresponding to the bulk of solution have been reached. In order to
simplify the description of this nonuniformity on the QCM, it is assumed
that a thin lm of liquid, having dierent values of g
f
and q
f
, exists at the
interface [61]. To calculate the eect of this lm on the frequency shift, one
has to solve the wave equation for the elastic displacements in the quartz
crystal [see Eq. (4)] simultaneously with the linearized Navier-Stokes
equation for the velocities in the lm and in the bulk liquid under standard
nonslip boundary conditions.
The shift of the resonant frequency and the width of the resonance
can be written as
Df
f
3=2
0
qg
p
pl
q
q
q
p

2f
2
0
l
q
q
q
p
q
1
g
g
f
q
f
q

L
f
23
G
2f
3=2
0
qg
p
pl
q
q
q
p

4f
2
0
l
q
q
q
p
q
1
g
g
f
q
f
q

L
2
f
d
24
where L
f
and q
f
are the thickness and the density of the lm, respectively.
These equations are valid in a particular case, when L
f
<< d. The general
case for arbitrary L
f
was given in Ref. 61. The rst terms in Eqs. (23) and
(24) yield the liquid-induced frequency shift and width of the resonance in
the absence of a lm. The terms in brackets describe the inuence of the
viscosity and density of a lm of thickness L
f
. According to Eqs. (23) and
(24), the ratio of the lm-induced width to the lm-induced frequency shift
is proportional to L
f
/d. Thus, for L
f
/d << 1, the contribution of the thin
interfacial lm to the width is much smaller than its contribution to the
frequency shift. For g
f
>> g the lm acts as though it were rigidly attached
to the surface: it causes a shift in frequency equal to that caused by its mass.
3. Slip Boundary Conditions
a. Slippage at Solid/Liquid Interface
Although the nonslip boundary condition has been remarkably successful
in reproducing the characteristics of liquid ow on the macroscopic scale,
its application for a description of liquid dynamics in microscopic liquid
layers is questionable. A number of experimental [6264] and theoretical
[65,66] studies suggest the possibiility of slippage at solid/liquid interfaces.
The boundary condition is controlled by the extent to which the
liquid feels a spatial corrugation in the surface energy of the solid. This
depends on a number of interfacial parameters, including the strength of
the liquid-liquid and liquid-solid interactions, the commensurability of the
substrate and the liquid densities, characteristic sizes, and also the rough-
ness of the interface. In order to quantify the slippage eect, the slip length,
Tsionsky et al. 20
k, is usually introduced [65,67,68]. The traditional nonslip boundary
condition is replaced by
dvz; x
dz
zd
1
k
vd; x v
q0
x 25
where v(z,x) is the velocity in the liquid and v
q0
(x) is the velocity of the
quartz crystal surface. Equation (25) expresses the discontinuity of the
velocity across the interface. For k = 0, Eq. (25) is reduced to the usual
nonslip boundary condition: v(d,x) =v
q0
(x). The physical meaning of the
slip length can be claried by comparing velocity proles for the nonslip
and slip boundary conditions. These two proles coincide when the nonslip
boundary condition is imposed at the surface shifted inside the solid on the
distance k with respect to the actual interface.
The slip boundary condition (25) results in the following equations
for the resonant frequency shift and the width of the resonance:
Df
f
2
0
qd
q
q
l
q
p
1
1 k=d
2
k=d
2

26
G
2f
2
0
qd
q
q
l
q
p
1 2k=d
1 k=d
2
k=d
2

27
Equations (26) and (27) show that the inuence of the slippage on the
response of the QCM in liquid is determined by the ratio of the slip length
k to the velocity decay length, d. Even for a small value of k c1 nm, the
slippage-induced correction to the frequency shift, Df
sl
, will be of the order
of 6.5 Hz for the fundamental frequency of f
0
= 5 MHz. This value far
exceeds the resolution of the QCM, but it is dicult to separate it from the
overall QCM signal.
There have been attempts [53] to estimate the slip length at the solid/
liquid interface on the basis of QCM experiments for adsorbed liquid
layers. The slip length can be expressed in terms of the coecient of sliding
friction, v, at the interface
k
g
v
28
Using the sliding friction coecient v = 3 g/cm
2
s, which is obtained for a
monolayer of water on Ag [49] and on Au [69], a surprisingly high slip
length of k = 6 10
4
nm is obtained. Using this value for the interface
between Au and bulk water, Eq. (26) yields for f
0
=5 MHz a value of Df c
710
3
Hz, which turns out to be smaller than that observed experimen-
tally by a factor of 10
5
. This inconsistency is most likely caused by a
roughness of the electrode surface that reduces the eective slip length.
Another reason could be the dierence between friction at the solid/
adsorbed layer and the solid/liquid interfaces. For example, a decrease in
the slip length with increasing lm thickness has been observed recently in
QCM studies of Kr lms on gold electrodes [55].
Recent molecular dynamics simulations [65,70] demonstrated that
the slip length is determined by the ratio of characteristic energies of liquid-
substrate, e
ls
, and liquid-liquid, e
ll
, interactions, k = f(e
ls
/e
ll
). The slip
length is negligible for e
ls
/e
ll
z 1 and grows with the decrease of the
parameter e
ls
/e
ll
. The slip length k may be as large as 15 diameters of liquid
molecules for e
ls
/e
ll
c0.5. It should also be noted that, for a given value of
e
ls
/e
ll
, the slip length is minimal when substrate and liquid molecules are of
the same size and increases with the increase of incommensurability of the
sizes. For smaller coupling between the liquid and the substrate or
incommensurability of their sizes, the spatial corrugation in the interfacial
energy is weaker and interfacial slip can develop.
The latter conditions are satised for partially wetting liquid/solid
interfaces. Wetting is characterized by a contact angle, which can be esti-
mated as [68]
cos h 1 2
q
s
q
e
ls
e
ll
29
where q
s
and q are the density of the solid and the liquid, respectively.
Thus, the contact angle may be interpreted as a measure of the strength
of interaction between the liquid and the solid, e
ls
. One expects a large
value of the slip length for a nonwetting situation (cos(h) !1), when e
ls
becomes much smaller than e
ll
. This conclusion is in agreement with several
experimental observations [62,71] reporting large slip lengths for partially
wetting liquids.
The authors of Refs. 14,72,73 showed that surface treatments
aecting liquid contact angle inuence the response of quartz crystal
resonator: resonant frequency changes caused by liquid loading were
consistently smaller for surfaces having large liquid contact angles. These
results were interpreted as arising from the onset of slippage at the solid/
liquid interface: the solid-liquid interaction becomes suciently weak on a
hydrophobic surface, and shear displacement becomes discontinuous at
Tsionsky et al. 22
the interface. However, this interpretation was called into question by a
series of experiments in which the eect of a hydrophobic monolayer was
examined on devices with various surface roughness [12].
Correlating the wetting properties with the response of the QCM in
contact with liquids seems to be a promising area for future research.
Unfortunately, studies of wetting behavior require ex situ measurements of
the contact angle, which change drastically the properties of the electro-
chemical system at the electrode/adsorbed layer/electrolyte interfaces.
b. Slippage at the Adsorbate/Electrolyte Interface
Slippage is very sensitive to the molecular structure of the interface, as we
have already discussed above. Thus, adsorption can strongly inuence this
phenomenon. In order to describe the eect of adsorption, let it be assumed
that the adsorbed layer is rigidly attached to the surface and slippage
occurs at the adsorbate/liquid interface (see Fig. 4). Then the equation of
motion of the adsorbed layer can be written as [74]
ixDm
a
v
a
x l
q
du z
dz
v v
a
x v
1
x at z d 30
where v
a
(x) is the velocity of the adsorbed layer and Dm
a
is its two-
dimensional density, while v
l
(x) is the velocity of the liquid at the interface.
The rst term on the right-hand side of Eq. (30) describes the driving force
acting on the adsorbed layer from the quartz crystal, while the second term
accounts for the friction at the adsorbate/liquid interface.
The velocity elds in the crystal and the liquid are given by the
solutions of the wave equation [Eq. (4)] and the linearized Navier-Stokes
equation [Eq. (18)], respectively. The solution of Eqs. (4), (18), and (30)
with the boundary conditions for shear stresses and velocities leads to the
following equation for the shift of the resonant frequency, Df, and the
change of the width of the resonance, G:
Df
2f
2
0
Dm
a
q
q
l
q
1=2

f
3=2
0
qg
1=2
pq
q
l
q
1=2
1
1 a
2
a
2

31
G
2f
3=2
0
qg
1=2
pq
q
l
q
1=2
1 2a
1 a
2
a
2

32
Writing Eqs. (31) and (32), we introduced a dimensionless parameter a =
g/ vd = k/d , which is the ratio of the slip length, k = g/v, and the velocity
decay length in the liquid, d. Equations (31) and (32) include both the
interfacial (adsorption) and the bulk solution contributions to the response
of the QCM, given by Eqs. (20) and (21). The latter remains constant in
adsorption studies and can be subtracted from the overall change given by
Eqs. (31) and (32). As a result, the shift of the resonant frequency and the
change of the width due to adsorption, which are measured experimentally,
are given by the equations:
Df Df
l
uDf
m
Df
sl

2f
2
0
Dm
a
q
q
l
q
1=2

f
3=2
0
qg
1=2
pq
q
l
q
1=2
aa 1
1 a
2
a
2

33
G G
l

f
3=2
0
qg
1=2
pq
q
l
q
1=2
4a
2
1 a
2
a
2

34
Equation (33) shows that there are two dierent contributions to the
frequency shift, Df
m
and Df
sl
, which originate from (1) a change of the mass
of the adsorbed layer rigidly coupled to the surface [rst term on the rhs of
Eq. (33)], and (2) partial decoupling between the quartz crystal oscillations
and the solution, caused by slippage at the adsorbate/liquid interface
[second term on the rhs of Eq. (33)]. It should be stressed here that, in
contrast to adsorption from the gas phase, electrosorption can result in
either a decrease or an increase of the resonant frequency, depending on its
eect on the mass of the layer rigidly coupled to the surface and on change
of the coecient of sliding friction, which determines the slip length,
according to Eq. (28).
Consider the eect of adsorption on the parameters Dm
a
and v. The
layer adsorbed at the electrode/electrolyte interface contains two types of
molecules: adsorbate and solvent. In the framework of mean eld approx-
imation, the eective interaction between the liquid and the adsorbed layer
can be characterized by the energy e
lsce
la
&
a
/&
m
+e
ll
(1&
a
/&
m
), where e
la
is the characteristic energy of the adsorbate/liquid interaction and &
m
is the
maximum surface excess of the adsorbate. As a result, the slip length at the
adsorbed layerliquid interface can be expressed as
k f e
la
=e
ll
&
a
=&
m
1 &
a
=&
m
c
fe
la
=e
ll
&
a
=&
m
35
showing an increase of k with &
a
for e
la/
e
ll
< 1. Equation (35) is the
interpolation formula that describes correctly the behavior of k for small
Tsionsky et al. 24
&
a
/&
m
and for &
a
/&
m
=1. We note that when the liquid and adsorbate
molecules are of signicantly dierent size, the incommensurability be-
tween the structures of the adsorbed layer and liquid grows with &
a
, which
may lead to an additional enhancement of the slip length. What is
important here is a relation between scales of corrugations of the poten-
tial energy in the solvent and the adsorbed layer, rather than their physical
sizes of solvent and adsorbed molecules.
The foregoing discussion shows that for e
la/
e
ll
< 1 the parameter a =
k/d in Eqs. (31) and (32), characterizing the eect of slippage on the
response of the QCM, increases with &
a
. For instance, for e
la/
e
l c
0.5, it
may reach values as high as a
c
10
2
for &
a c
&
m
. Correspondingly, the
adsorption-induced slippage leads to a positive frequency shift, which
grows with &
a
. This contribution can be larger than the eect of added
weight. As a result, the overall frequency shift due to electrosorption can
be positive and increases with &
a
[74]. It should be noted that for small
values of the parameter a, the eect of slippage on the resonance frequency
shift is much larger than its eect on the width of the resonance [see Eqs.
(33) and (34)]. Also, slippage will always cause a decrease in the width of
the resonance. Thus, if a positive shift of frequency with adsorption is to be
associated with enhanced slippage, it should also be exhibited as a
reduction of the width of the resonance, although the latter may be hard
to detect experimentally.
Above we discussed the situation where the adsorbed layer is rigidly
attached to the oscillating crystal surface, and there is nite slippage at the
adsorbate/liquid interface. An alternative model based on the assumption
that slippage occurs at the crystal/adsorbed layer interface and nonslip
boundary conditions apply to the adsorbate/liquid interface can also be
considered. For a small slip length, E << d, this model leads to the same
results for the shift of the complex resonance frequency as the model
discussed above [see Eqs. (33) and (34)], and measurements employing the
QCM cannot distinguish between them. However, in the case of specic
adsorption, the assumption of slippage at the crystal/adsorbed layer
interface is hard to justify, since the characteristic energy of adsorbate-
substrate interactions, e
as
, is larger than the energy of adsorbate-adsorbate
interactions, e
aa
, hence the corresponding slip length k = f(e
as
/e
aa
) is
expected to be very small.
It should be noted that the eect of slippage on the response of QCM
could also be interpreted by introduction of a charge-dependent eective
viscosity of a microscopic interfacial layer. In this case, an increase of the
slip length would be equivalent to a decrease of viscosity. Unfortunately,
there is at present no suitable theory to describe the eect of the excess
surface charge density (or the corresponding high electrostatic eld) on the
viscosity of the electrolyte in the double layer. The derivation of such a
model is complicated by the fact that electro-neutrality does not exist on
the solution side of the interface (except at the potential of zero charge),
although the electrostatic energy is reduced by interaction with the image
charges on the metal side of the interface.
D. Quartz Crystals with Rough Surfaces
1. Quartz Crystals with Rough Surfaces Operating in Liquids
When the surface of quartz crystal resonator is rough, the liquid motion
generated by the oscillating surface becomes much more complicated than
for the smooth surface. A variety of additional mechanisms of coupling
between the acoustic waves in the solid and the motion in the liquid can
arise. These may include generation of nonlaminar motion, the conversion
of in-plane surface motion to motion normal to the surface, and trapping
of liquid by cavities and pores. It has been experimentally demonstrated
[12,15,7579] that the roughness-induced response of the QCM includes
both the inertial and viscous contributions. Measurements of the complex
shear mechanical impedance [12] were used to analyze dierent contribu-
tions to the roughness-induced response of the quartz resonator and to
correlate the experimental results with the surface roughness of the quartz
resonator. Nevertheless, this subject is poorly developed, and the inter-
pretation of experimental results can often be ambiguous.
The dependence of the QCM response on the morphology of the
interface is determined by the relation between the characteristic sizes
of roughness and the length scales of the shear modes in the liquid and
the quartz resonator. The length scales in the liquid (the velocity decay
length, d) and in the crystal (wave length of the shear-mode oscillations,
k
q
) are dened by the Navier-Stokes equation and by the wave equation
for elastic displacement, respectively. For typical frequencies used in
QCM experiments, f
0
=510 MHz and the lengths d = (g/pf
0
q)
1/2
and
k
q
= (l
q
/q
q
)
1/2
f
0
1
are of the order 0.1770.25 Am and 0.030.1 cm, re-
spectively.
The surface prole may be specied by a single valued function z =
n(R) of the lateral coordinates R that denes a local height of the surface
with respect to a reference plane (z = 0). The latter is chosen so that the
Tsionsky et al. 26
average value of n(R) will equal zero. Surfaces used in QCM experiments
may have various scales of roughness. In order to clarify this point, let us
consider the two limiting cases: slight and strong roughness structures,
which are schematically shown in Fig. 5. For the slight roughness (Fig. 5a)
the amplitude of deviation from the reference plane z = 0 is much less
than the lateral characteristic length. In the case of strong roughness (Fig.
5b), the amplitude and period of repetitions are of the same order of
magnitude.
In order to stress the multiscale nature of roughness, the prole
function can be written as the sum of the functions that characterize the
prole of the specic scale i:
nR
i
n
i
R 36
For the calculation of the response of the QCM, the height-height pair
correlation function is needed [80]. When rough structures having dierent
FIG. 5. Schematic representation of a slight (a) and a strong (b) roughness. The
prole of slight roughness is described by the function z =n(R). L is the eective
thickness of the porous lm that represents strong roughness. (From Ref. 24.)
scales do not correlate the total correlation function can be written in the
form
< nRVnRVR >
i
< n
i
RVn
i
RVR > 37
where <n
i
(RV)n
i
(RVR)> is the correlation function for the scale i
and <n (RV)n
i
(RVR)> means averaging over the lateral coordinates.
Usually one assumes that the correlation function <n
i
(RV)n
i
(RVR)>
has a Gaussian form <n
i
(RV)n
i
(RVR)>=h
2
i
exp(-|R|
2
/l
i
2
), where h
i
is the
root mean square height of the roughness and l
i
is the lateral correlation
length, which represents the lateral scale. Thus, the morphology of the
rough surface can be characterized by a set of lengths {h
i
, l
i
}.
It is impossible at the present time to provide a unied description of
the response of the QCM for nonuniform solid/liquid interfaces with
arbitrary geometrical structure. Below we summarize results obtained
for the limiting cases of slight and strong roughness.
a. Slight Roughness
For slightly rough surfaces, the problem was solved in the framework of
perturbation theory with respect to the parameters |jn(R)| << l and
h/d << l, where h is the root mean square height of roughness at the
electrode surface [80,81]. The rst condition means that the local slope of
the interface is small, i.e., the height, h, is less than the lateral characteristic
length (i.e., the correlation length, l ) of the roughness.
For roughness described by a one-scale correlation function, the shift
in the resonant frequency and the width of the resonance can be written in
the following form [80,81]:
Df
f
2
0
qd
q
q
l
q
1=2
1
h
2
l
2
Fl=d

38
G
2f
2
0
qd
q
q
l
q
1=2
1
h
2
l
2
Al=d

39
The scaling functions F(l/d) and A(l/d) are expressed through the Fourier
components of the height-height correlation function of the roughness
g(K) [81]. The correlation function g(K) can be dened as
h
2
gK
d RexpiKR < nRVnRVR > 40

The correlation function provides the most detailed characterization of the
surface structure. Sometimes the surface roughness is described by an
Tsionsky et al. 28
integral parameter, the roughness factor, R, which is the ratio between the
true and the apparent (geometrical) surface area. For slight roughness, the
roughness factor is expressed through the correlation function [81] as
R 1
h
2
2
dK
2p
2
gKK
2
41
For the Gaussian random roughness g(K) = pl
2
exp(l
2
K
2
/4) and Eq. (41)
yields R =1+2h
2
/l
2
.
It should be noted that the roughness factor, R, relevant to the
operation of the EQCM is not the same as the roughness factor commonly
referred to in interfacial electrochemistry, because of the dierence in
corresponding length scales. The EQCM roughness factor is mostly
determined by the roughness on the scale of the velocity decay length in
the liquid, d, which assumes values of hundreds of nm, depending on the
frequency of the crystal and the viscosity and density of the liquid. The
interfacial roughness factor is related to charge transfer at the interface
and the double layer structure, and therefore its characteristic scale is
about 1 nm.
The rst terms in braces in Eqs. (38) and (39) dene the shift and the
broadening of the resonance at the interface between an ideally smooth
crystal and the liquid [11]. The surface roughness leads to an additional
decrease of the resonant frequency and a broadening of the width of the
resonance, expressed by the second terms in this equation.
The particular form of the scaling functions F(l/d) and A(l/d) is
determined by the morphology of the surface. However, the asymptotic
behavior of these functions for l/d >> 1 and l/d <<1 is universal [81] and
has the form
Fl=d l=da
1
a
2
d=l at l=d >> 1 42
Fl=d l=db
1
b
2
d=l at l=d << 1 43
Al=d c
1
at l=d >> 1 44
Al=d l=d
2
c
2
at l=d << 1 45
For random Gaussian roughness, the parameters are
a
1
p
1=2
; a
2
2; b
1
3p
1=2
; b
2
2; and c
1
c
2
2 46
It should be noted that for l/d >> 1 the roughness-induced frequency
shift includes a term that does not depend on the viscosity of the liquid,
the rst term in Eq. (42) and Eq. (38). It reects the eect of the non-
uniform pressure distribution, which is developed in the liquid under the
inuence of a rough oscillating surface [80]. The corresponding contri-
bution has the form of the Sauerbrey equation. This eect does not exist
for smooth interfaces. The second term in Eq. (42) and Eq. (44) describes
a viscous contribution to the QCM response. Its contribution to Df has
the form of the QCM response at a smooth liquid/solid interface, but
includes an additional factor R that is a roughness factor of the surface.
The latter is a consequence of the fact that for l/d >> 1 the liquid sees
the interface as being locally at, but with R time its apparent surface
area.
Results obtained in Refs. 80, 81 show that the inuence of slight
surface roughness on the frequency shift cannot be explained in terms of
the mass of liquid trapped by surface cavities, as proposed in Refs. 76,
77. This statement can be illustrated by consideration of the sinusoidal
roughness prole. The mass of the liquid trapped by sinusoidal grooves
does not depend on the slope of the roughness, h/l, and is equal to S-h/p,
where S is the apparent area of the crystal. However, Eq. (38) demonstrates
that the roughness-induced frequency shift increases with increasing slope.
Equation (39) and the asymptotic behavior of the scaling functions
show that in the regions where l/d >> 1 and l/d << 1, the width is
proportional to the factors (qg)
1/2
f
0
3/2
and q
3/2
g
1/2
f
0
5/2
, respectively. In
the high viscosity limit, when g >l
2
pqf
0
, the roughness-induced frequency
shift approaches a constant value and the roughness-induced width tends
to zero.
The results obtained make it possible to estimate the eect of
roughness on the response of the QCM if the surface proles function
n(R) can be found from independent measurements.
b. Strong Roughness
Perturbation theory cannot be applied to describe the eect of the strong
roughness. An approach based on Brinkmans equation has been used
instead to describe the hydrodynamics in the interfacial region [82]. The
ow of a liquid through a nonuniform surface layer has been treated as the
ow of a liquid through a porous medium [8385]. The morphology of
the interfacial layer of thickness, L, has been characterized by a local
permeability, n
H
, that depends on the eective porosity of the layer, /. A
number of equations for the permeability have been suggested. For
instance, the empirical Kozeny-Carman equation [83] yields a relationship
Tsionsky et al. 30
between n
2
H
and the eective porosity n
2
H
f
r
2
/
3
= 1 /
2
, where r is the
characteristic size of inhomogeneities.
The ow of liquid through the interfacial layer can be described by
the following equation [82]:
ixqvz; x g
d
2
dz
2
vz; x gn
2
H
v
q0
vz; x 47
where v
q0
is the amplitude of the quartz surface velocity and v(z,t) =v(z,x)
exp(ixt) is the velocity of the liquid in the layer. In this equation the eect
of the solid phase on the ow of liquid is given by the resistive force, which
has a Darcy-like form, gn
2
H
v
q0
vz; x . In the case of high eective
porosity, the resistive force is small and Eq. (47) is reduced to the Navier-
Stokes equation, describing the motion of the liquid in contact with a
smooth quartz surface. For a given viscosity, the resistive force increases
with decreasing eective porosity and strongly inuences the liquid
motion. At very low eective porosity, all the liquid located in the layer
is trapped by the roughness and moves with a velocity equal to the velocity
of the crystal surface itself.
Brinkmans equation represents a variant of the eective medium
approximation, which does not describe explicitly the generation of non-
laminar liquid motion and conversion of the in-plane surface motion into
the normal-to-interface liquid motion. These eects result in additional
channels of energy dissipation, which are eectively included in the model
by introduction of the Darcy-like resistive force.
The liquid-induced frequency shift and the width of the resonance
have the following form [82]:
Df
2f
2
0
q
l
q
q
q
1=2
Re
1
q
0
L
n
2
H
q
2
1
1
W
1
n
2
H
q
2
1
2q
0
q
1
coshq
1
L 1 sinhq
1
L

48
G
4f
2
0
q
l
q
q
q
1=2
Im
1
q
0
L
n
2
H
q
2
1
1
W
1
n
2
H
q
2
1
2q
0
q
1
coshq
1
L 1 sinhq
1
L

49
where q
0
= (i2p f
0
q/g)
1/2
, q
2
1
= q
2
0
+n
2
H
, and W = q
1
cosh( q
1
L)+q
0
sinh( q
1
L). The rst terms on the right-hand sides of Eqs. (48) and (49)
describe the response of the QCM for the smooth quartz crystal/liquid
interface [11]. The additional terms present the shift and the width of the
QCM response caused by the interaction of the liquid with a non-uniform
interfacial layer.
When the permeability length scale is the shortest length of the
problem,n
H
<< d and n
H
<< L, the layer-induced shift, Df
L
, is propor-
tional to the density of the liquid and does not depend on the viscosity. It
has the formof the Sauerbrey equation for mass loading. This eect results
fromthe inertial motion of the liquid trapped by the inhomogeneities in the
interfacial layer:
Df
L

2f
2
0
q
l
q
q
q
1=2
L n
H
50
The eective thickness of the liquid lm rigidly attached to the
oscillating surface is equal to Ln
H
and is less than the thickness of the
inhomogeneous layer, L. The increase of the permeability n
H
leads to
the enhancement of the velocity gradient in the layer, which results in a
decrease of the shift due to mass loading, and an increase of the width
caused by the energy dissipation. When the layer thickness is the shortest
length of the problem, L<<d, L<<n
H
, and n
H
<<d, the frequency shift is
also proportional to the density of the liquid and does not depend on
viscosity:
Df
L

2f
2
0
qL
3l
q
q
q
1=2
L=n
H
2
51
However, in contrast to the previous case, it cannot be related to the mass
of trapped liquid. The correction to the width of the resonance depends on
the viscosity and is substantially less than the layer-induced shift.
2. Roughness and Slippage
Roughness can inuence the response of the QCM not only through its
direct eect on the hydrodynamics at the interface, but also indirectly
through its eect on various interfacial properties. For instance, roughness
of mesoscopic scale can have a pronounced inuence on slippage at the
solid/liquid interfaces, as discussed above. The authors of Refs. 53, 86
suggest introduction of an eective slip length, k
e
, which takes into
account both slippage and roughness, in order to describe this phenome-
Tsionsky et al. 32
non. In this manner, liquid ow at a rough surface has been simulated as a
ow at a smooth surface with an eective slip length.
Application of this approach to the QCM problem yields the fol-
lowing equation for the eective slip length:
k
eff
k 1
h
0
K
0
2

2
3 4kk
0
1 2kk
0

k
0
h
2
0
2
2 3kk
0
1 kk
0
1 2kk
0

52
Equation (52) was derived for a sinusoidal prole of roughness, z(x) =
d + h
0
sin(k
0
x), with an amplitude h
0
and a period of 2p/k
0
, assuming that
the decay length, d, is the largest characteristic length of the problem,d/k
>>1 and dk
0
>>1. Beyond these conditions the eective slip length is a
complex function. Equation (52) shows that roughness diminishes the
inuence of slippage on the QCM response, namely the eective slip length
becomes smaller than the corresponding length for the smooth interface.
At rough interfaces, the eective slip length decreases with an increase of
the amplitude of the surface corrugation and with a decrease of its period.
It should be noted that an eective slip length is not an intrinsic
property of the surface. Its value depends also on the experimental
conguration, for instance, k
e
found for the Poiseuille owbetween rough
surfaces [86] diers from the corresponding value obtained for QCM
experiments [Eq. (52)].
So far only a few studies [8689] have been devoted to the eect of
roughness on slippage, and this subject requires additional investigation.
III. ELECTRICAL DOUBLE LAYER/ELECTROSTATIC
ADSORPTION
A. Introduction
We shall restrict our consideration here to the simplest electrochemical
case: the electrical double layer, which is not complicated by charge trans-
fer or by specic adsorption. At rst glance it would seemthat, if there is no
change in mass and nothing happens in the bulk of the solution in which the
electrode is immersed, the EQCM response should be zero. However,
essentially all measurements show that in the double-layer region the
frequency of the EQCM depends on potential. The eect is rather
smalla few Hz for crystals with fundamental frequencies of 510 MHz.
In most cases reported in the literature [9096], experiments were
performed employing cyclic voltammetry, with the potential extending to
the region of oxide formation. All data obtained in this manner exhibit
some degree of hysteresis in the double layer region, which increases with
sweep rate. It also increases when the anodic limit of potential is extended,
i.e., when the potential is swept deeper into the oxide formation region.
There is good reason to believe that this hysteresis is due to a memory
eect of the interface, related to residual adsorbed oxygen or to some
traces of dissolved gold remaining near the surface. In only a few papers
[61,9799] has the response of the EQCM to changes in potential in the
double-layer region been studied under experimental conditions, which
excluded hysteresis. In these studies performed ongold and silver, the limits
of potential during cycling were restricted to the double-layer region. Both
metals have suciently extensive potential region where only electrostatic
adsorption can take place. Moreover, it is well known that the surfaces of
these metals do not undergo any changes in their morphology in the course
of cycling in this restricted potential region. Measurements were conducted
in electrolytes that are not specically adsorbed to a signicant extent.
B. Some Typical Results
All attempts to nd rigorous quantitative correlation between data
obtained in dierent laboratories have failed. This is not surprising,
considering that the eects are rather small and the surfaces studied have
dierent histories (technique of producing the EQCM, electrode pretreat-
ment, etc.) and dierent morphologies. The surfaces employed in EQCM
studies are always polycrystalline, even though they often have a preferred
crystal orientation. Nevertheless, some general characteristics of the
behavior of the EQCM response to potential changes in the double layer
region can be summarized as follows:
1. In all cases the change of frequency with potential is quite small,
of the order of a few Hz.
2. The dependence of frequency on potential is typically of a para-
bolic shape around the potential of zero charge (pzc), but the
maximumdoes not necessarily coincide with the potential of zero
charge.
3. The curves are distinctly dierent for dierent solutions.
A few examples of such behavior are shown in Fig. 6. There are
dierent approaches to explain the eects observed:
1. Reorientation of water molecules caused by a change in the
electric eld inside the double layer with potential [100102].
Tsionsky et al. 34
FIG. 6. Dependence of the frequency shift on potential in the double-layer
region on gold in dierent solutions. Lines were calculated for the mass-eect [see
Eq. (54)]. Arrows show the approximate position of the pzc. (From Ref. 98.)
2. Preoxidation of gold at potentials much more negative than the
potentials usually associated with oxide formation [103].
3. Repulsion of water molecules from the electric double layer or
loss of water molecules fromthe hydration shell of ions as a result
of electrostatic adsorption [61,94,99].
4. Anion adsorption (the case of perchlorate solutions) increasing
when the potential is shifted in the positive direction [91,93,99]. It
should be noted in this context that it is not always clear whether
the authors imply specic or electrostatic adsorption.
All these mechanisms could lead to modications of the properties of
the microscopic interfacial layer, changing the interactions of the vibrating
surface with the solution. Taking these mechanisms into account it is
possible, in principle, to consider the EQCMresponse within the models of
thin layer or/and slippage discussed in Sec. II.C. However, below we will
showthat the experimentally observed EQCMresponse in the double layer
region can be described by the thin layer model, taking into account only a
general electrochemical phenomenon: the dependence of the properties of
the diuse double layer on potential, without considering specic phe-
nomena, some of which may be doubtful.
C. The Potential Dependence of the Frequency
1. The Thin-Film Model
An attempt to describe the response of the EQCM in the double-layer
region only on the basis of the properties of the diuse double layer was
undertaken in Ref. 61. In doing so, the specic adsorption of ions and the
potential-dependent specic interactions of the solvent with the metal were
entirely ignored. Under these assumptions one could think of three reasons
leading to the observed dependence of frequency on potential: (1) depen-
dence of the surface tension on potential, (2) electrostatic adsorption of
charged species, and (3) a local change in viscosity in the diuse double
layer.
a. Surface Tension
Shifting the electrode potential from the pzc decreases the surface
tension. This can apply a stress to the surface of the quartz crystal
resonator, decreasing its frequency. The inuence of stress in a thin sur-
face lm of the quartz resonator on its frequency has been discussed in
Tsionsky et al. 36
the literature [104, 105]. In accordance with Ref. 104, the frequency shift
can be written as
Df
P
K f
0
=dDc 53
where f
0
is the fundamental frequency of the EQCM, d is the thickness of
quartz resonator (0.2 mm in case of 6 MHz crystal used in Ref 61), Dc is
the change in surface tension, and the constant K = 2.7510
12
cm
2
/dyn
depends only on the properties of the quartz crystal. The value of Df
P
as
a function of potential can be evaluated using the dependence of the
surface tension on potential for a gold electrode [106]: for a rational
potential of F0.5 V, the decrease in surface tension is about 80 dyn/cm.
Using this value in Eq. (53), we nd Df
P
= 0.07 Hz, which is negligible
compared to the shift in frequency observed experimentally in the same
potential region (see Fig. 6).
b. Electrostatic Adsorption of Charged Species
Let us compare the velocity decay length in solution to the thickness of the
diuse double layer: as already noted in Sec. II.C, the former is given by
d =(g/q pf
0
)
1/2
and for dilute aqueous solutions at roomtemperature has a
value of 0.23 Am. In contrast, the Debye length in 0.1 M solution of a 1-1
electrolyte is j
1
= 1 nm. The relative decrease of the uid velocity over
such a small distance is of the order of 1/230, i.e., less than half a percent.
Thus, it could be argued that the ions pulled into the diuse double layer as
a result of the excess charge density on the metal can be considered to be
vibrating in phase with the surface of the crystal and, in this sense, act as
though they were (almost) rigidly attached to it.
From the Gouy-Chapman theory, one can calculate the dependence
of the positive and negative excess charges (q
and q
+
) on potential. The
change of mass in the diuse double layer, when the potential was shifted
from the pzc, can be expressed as
Dm
a;c
q
=FM
a
M
w
m
a
q
=FM
c
M
w
m
c
54
where M
a
, M
c
, and M
w
are the molecular weights of the anion, the cation,
and of water, respectively. The values of m
a
and m
c
reect the number of
water molecules replaced by the respective ions. The potential dependence
of the added weight is implicit in Eq. (54), through the dependence of the
excess positive and negative charges on potential.
The implications of Eq. (54) are rather more subtle than might
appear at rst sight. Consider the relatively simple case of the perchlorate
ion, which is believed to be unhydrated, because of its relatively large
radius. One might be tempted to calculate the number of water molecules
replaced by this ion [i.e., the value of m
a
in Eq. (54)] by considering the
relative volumes of the ion and of water. This approach tacitly assumes
that the density of water in the diuse double layer in the presence of
C1O
4
ions is equal to its density in the bulk. This is not necessarily true,
since the eld in the vicinity of an ion may be high enough to aect the
structure of water, even if it is not high enough to cause solvation of the ion.
Since the concentration of ions in the diuse double layer can be one or two
orders of magnitude higher than in the bulk of the solution, this eect may
be quite signicant. Considering the partial molar volumes of substances
used in the experiments shown in Fig. 6 (LiClO
4
, HClO
4
, KNO
3
, and
KOH), we nd that two molecules of water are replaced from solution for
each molecule of perchloric acid added, and the corresponding numbers
for KNO3 and KOH are 2.3 and 0.4, respectively. If it is assumed that
replacing a water molecule by an H
3
O
+
ion has no eect on the weight and
one OH
ion replaces one water molecule, we can estimate the contribu-

tion of all the other ions.
*
Using such values, the added weight of anions
adsorbed electrostatically in the diuse double layer can be evaluated. The
solid lines in Fig. 6 show the results of these calculations. It is clearly seen
that these lines do not t the experiments. Moreover, electrostatic adsorp-
tion alone predicts the maximum EQCM frequency always at the pzc,
while the experimentally observed position of the maximum depends on
the composition of the solution.
c. Viscosity Eect in the Diuse Double Layer
A dierence between the viscosity of the uid in the diuse double layer
and in the bulk of the solution could be caused by two factors.
1. The concentration of ions near the electrode surface depends on
charge density and hence on potential. For example, in 0.1 M
solution of a 1-1 electrolyte at a rational potential of +0.5 V, the
concentration of anions at the outer Helmholtz plane, calculated
fromthe Gouy-Chapman theory, is 4.7 M. At this concentration,
*
It must be remembered that electro-neutrality is not maintained on the solution
side of the interface in the diuse double layer, so that the values of v
a
and v
c
obtained in neutral solutions may not apply.
Tsionsky et al. 38
the viscosities of HClO
4
and KOH are higher than in pure water
by 20% and 60%, respectively.
2. According to the same theory, the eld strength in the outer
Helmholtz plane may reach values of 310
6
V/cm[107]. While this
is smaller than the eld in the inner Helmholtz plane or around
small cations, it may be high enough to aect the structure of the
solvent and change its local viscosity.
In order to describe both the eect of the electrostatic adsorption of
ions and the eect of the viscosity inside the diuse double layer on the
response of the EQCM, one can use the thin-layer model described in Sec.
II.C. Since the thickness of the diuse double layer is much less than the
velocity decay length, the corresponding equation of the model is Eqs. (23)
and (24), which can be rewritten in the following form:
Df Df
m
Df
g
55
where
Df
m
C
m
Dm
a;c
56
and
Df
g
C
m
qL
f
1
g
g
f

57
For the lm thickness, as a rst approximation,
*
one can take that
L
f
= j
1
. Another simplifying assumption is that the viscosity changes
abruptly at the boundary between the lm and the solution. Estimation of
the viscosity of the lm as a function of potential is very dicult, since
electro-neutrality is not maintained in the diuse double layer, and it is
dicult to take into account the inuence of the electric eld in the double
layer on the viscosity of the lm. Instead, the viscosity of the lm, g
f
, can be
taken as a parameter, to t the theoretical curve to the experimental results.
To do this one substracts from the observed frequency shift the contribu-
tion of the mass eect caused by electrostatic adsorption of ions [Eq. (56)].
*
While this approximation is ne for a 0.1 M solution, it fails for dilute solutions.
Thus, for a 1 mMsolution j
1
=10 nm, but most of the added mass is concentrated
very close to the surface, so that the lm thickness should be taken as L =12 nm,
independent of the concentration in solution.
The remaining part of the overall eect can then be ascribed to viscosity
eect, g
f
. Using pairs of experimental data Df vs. E, and the dependence of
surface charge density on potential, the concentration of ions (C
a,c
) in the
diuse double layer can be calculated and the dependence of g
f
on C
a,c
can
be obtained. An example of such calculations is shown in Fig. 7 [61,97].
Some specic features of the data shown here merit special atten-
tion:
1. At negative rational potentials, where the diuse double layer is
populated mostly by K
+
ions, the dependence of g
f
on concen-
tration for solutions of KOH and KNO
3
coincided.
2. The dependence of g
f
on concentration is quantitatively dierent,
but follows the same trend as the dependence of the bulk elec-
trolyte viscosity on the concentration of KOH and KNO
3
. This
observation holds even in the unique case of KNO
3
, where a
minimum in the bulk viscosity with increasing concentration is
exhibited.
3. The dependence of the lm viscosity on the concentration is
stronger than that of the viscosity in the bulk. This can be
understood if one takes in to account that g
f
reects the inuence
of one kind of ions on the solvent structure while g is the average
result of the inuence of both cations and anions.
All these observations are characteristic not only of the cases shown
in Fig. 7, but also of all other cases studied: HClO
4
, LiClO
4
, CsOH [61,97].
Thus, the model described above, taking into consideration the variation
of ionic concentration and local viscosity in the diuse double layer, can
account for the potential dependence of the frequency of the EQCM in the
double-layer region on gold electrodes.
It should be noted that in Fig. 7 we ascribed to the viscosity at pzc the
value of the bulk viscosity, suggesting that there is no inuence of the metal
on the solvent in the layer nearest to it. Due to interactions between solvent
molecules and the metal, this may not be the case. Hence one should take
into account that even at the pzc there could be a lm of solvent molecules
having a viscosity that is dierent fromthe value in the bulk. In this case the
inuence of the electric eld and the composition of the solution inside
the diuse double layer has to depend on the metal and on the nature of the
solvent. The latter could lead to a dependence of the EQCM response on
the nature of the metal, which is indeed observed when the results for gold
Tsionsky et al. 40
FIG. 7. Comparison of the eective lm viscosity, calculated for the concen-
trations prevailing in the diuse double layer in 0.1 M solutions of KOH (open
circles) and KNO
3
(open squares, positive surface charge densities; closed squares,
negative charge densities) with the dependence of the viscosity in the bulk, as a
function of concentration [178]. (From Ref. 98.)
FIG. 8. Frequency shift versus surface charge density on gold and silver elec-
trodes in 0.1 M solution of LiClO
4
in water (a) and in butanol (b). (Fig. 8a from
Ref. 98.)
Tsionsky et al. 42
and silver are compared, as shown in Fig. 8. The properties of the diuse
double layer in the Gouy-Chapman model are determined uniquely by the
excess surface charge density, q. Thus, the EQCM response is also
determined by the value of q in the framework of the model described
above. That is the reason for presenting the data in Fig. 8 as Df versus q.
A detailed discussion of the dierent response observed in water (Fig. 8a)
and in butanol (Fig. 8b) is beyond the scope of this review.
IV. ADSORPTION STUDIES
A. The Adsorption of Organic Substances
1. General
Here we restrict our discussion to reversible adsorption of submonolayers
excluding cases of adsorption accompanied by irreversible chemical
changes of the electrode surface and of polymer lm formation. Studies
of adsorption of organic substances frequently show an increase of
resonant frequency [74,98,108111]. This is not entirely unexpected,
considering that electrosorption is a replacement process in which the
adsorbate replaces a certain number of water molecules from the surface
[112,113]. Thus, depending on the number of water molecules displaced for
each organic molecule adsorbed, the total weight could increase, decrease,
or remain essentially constant. The change of mass density resulting from
electrosorption, Dm
ads
, can be expressed in the form
Dm
ads
G
ads
M
ads
vM
w
58
where M
ads
and M
w
are the molecular weights of the adsorbate and of
water, respectively, &
ads
is the surface excess of the adsorbate, and v is the
number of water molecules replaced by each adsorbed organic molecules.
This equation has two unknown parameters, &
ads
and v. For all cases
discussed below, the surface excess was known from measurement by
independent electrochemical techniques. The parameter v can be estimated
from bulk properties of the solvent and the adsorbent (i.e., the variation of
density with concentration). Densities of many organic substances, their
mixtures, and mixtures with water dier little fromthe density of water and
practically always are in the range of 0.81 g/cm
3
. Hence, the frequency
shift associated with mass loading resulting from adsorption is expected to
be quite small. For example, data on the water/pyridine mixtures showthat
one pyridine molecule replaces 4.4 water molecules (v =4.4). This happens
to be exactly equal to the ratio of molecular weights of pyridine and water.
Thus, the mass eect expected for pyridine should be equal to zero at all
potentials or surface charge densities. Similar estimates for pyridine in
n-butanol leads to v =0.85. The molecular weights of these two substances
are very close74 and 79. As a result, the expected frequency shift
associated with mass changes is only 0.25 Hz for the charge densities
where the surface excess on gold assumes its maximum value.
Once the mass of replaced water molecules has been taken into
account (or it was assumed a priori that this was negligible), the response of
the EQCM can be used to compare surface excess &
ads
, obtained from the
frequency shift with that derived from other techniques. However, the
situation is complicated by the fact that the absolute value of the frequency
shift caused by adsorption of organic substances is usually small, similar to
that caused by electrostatic adsorption of ions in the diuse double layer
over the same range of potential, as seen, for example, in Fig. 9. Thus, the
eect of the diuse double layer must be considered as a background, and a
suitable correction should be made.
The properties of the diuse double layer depend directly on the
surface charge density and not on the potential. In order to correct for this
eect quantitatively, one needs to convert the dependence of frequency on
potential Df(E), observed experimentally, to its dependence on charge
density, Df ( q). Having the analogous dependence Df
0
( q) for the support-
ing electrolyte, it is possible to evaluate the real response of the EQCM to
specic adsorption, df( q) = Df( q) Df
0
( q), and use this response for
interpretation of the data obtained.
*
This approach was taken in [74,
108,111] for several systems as seen in Figs. 9 and 10. For all cases studied,
the surface excess was known from independent electrochemical experi-
ments.
Consider rst the inuence of the nature of the metal on the response
of the EQCM to specic adsorption. This can manifest itself in the shapes
of the resulting dependence of df ( q) on q, because of a change of the pzc
*
The function df ( q) can be converted back to a dependence on potential df (E).
The adsorption of organic substances has been discussed in the literature either as
a function of potential (usually referred to as the pzc) or as a function excess
surface charge q. The question regarding which of the two should be preferred has
not been entirely resolved and will not be discussed here.
Tsionsky et al. 44
FIG. 9. The frequency shift versus the surface charge density in four systems:
pyridine (5 mM) in water [111], pyridine (5 mM) in butanol [111], n-butanol (5 mM)
in water [111], and tert-butanol (1 M) in water [108]. Supporting electrolyte: 0.1 M
LiClO
4
. Solid circles represent measurements in the supporting electrolyte. Open
circles represent measurements in the presence of the adsorbate. (Fig. 9ac from
Ref. 111.)
FIG. 10. Frequency shift due to pyridine adsorption, corrected for the shift
of frequency in the supporting electrolyte (0.1 M LiClO
4
), on gold (a) and silver
(b) electrodes versus potential. (From Ref. 98.)
Tsionsky et al. 46
and a dierent dependence of charge on potential with and without
adsorbate in solution. This denes the position of the curves of df ( q)
along the coordinate q. Dierent values of the surface excess &
ads
on
dierent metal dene absolute values df, but not the increase or decrease of
df with q. The latter is determined by the dierence &
ads
(M
ads
- vM
w
) in Eq.
(58) and does not depend on the nature of the metal. Nevertheless, pyridine
adsorption on gold and silver electrodes leads to quite dierent response of
the EQCM (see Fig. 10). Note that this gure shows the shape of the
function of df(E) after corrections for double-layer eects. Considering
these data, it is clearly seen that adsorption of pyridine on gold causes a
positive shift in frequency. For adsorption on silver, the frequency shift is
close to zero or slightly negative. Maximum adsorption of pyridine on
silver is remarkably less than on gold, but this dierence cannot explain the
dierent responses of the EQCM to adsorption in these two cases. It must
be concluded that there are other eects that were not taken into account.
The increase in frequency in the pyridine/water/gold system was
associated in Ref. 98 with the eect of adsorption on the structure of the
solvent at the interface. Adsorbed pyridine molecules destroy the structure
of water and replace the water-water interactions between the rst layer
and further layers at the interface, by pyridine-water interactions, which
must be much weaker. Such a replacement becomes more eective, as the
coverage by pyridine increases. The same explanation was found to be
applicable for the adsorption of n-butanol and t-butanol on gold from
aqueous solutions, as shown in Fig. 9c and d.
In the case of the pyridine-water-silver system, the frequency shift
observed is relatively small, as seen by comparing Fig. 10a and b. To
understand this eect, let us direct our attention to the data presented in
Fig. 8. The dependence of df on q in the case of silver is remarkably sharper
than for gold. It is reasonable to assume that the gold surface interacts with
water molecules, inducing a relatively strong structure, which is dierent
from that of bulk water. The structure near the surface of silver is closer to
that in the bulk of water. The stronger the interaction between the metal
and the solvent, the weaker is the inuence of the electric eld on the
structure of the surface layers. The latter accounts for the dierent
behaviors of df ( q) on gold and on silver in the double-layer region, as
seen in Fig. 8, and in the case of specic adsorption, as seen in Fig. 10a and
b. Adsorption of organic molecules destroys the water structure at the
interface, replacing it with a rather weak interaction of the hydrophobic
parts of adsorbed molecules with water. Since there is no special water
structure at the Ag/water interface, the eect of adding a hydrophobic lm
(adsorbed pyridine) between the metal and the solution has little eect on
the response of the EQCM.
A similar situation is encountered when pyridine is adsorbed on gold
from its solution in n-butanol. The shift in frequency does not exceed 1.5
Hz. Butanol itself is oriented on the gold surface with its alkyl chain toward
the solution. Pyridine adsorption does not seem to cause signicant
changes in the interaction of the vibrating surface with the liquid.
For a general case, the response of the EQCM to adsorption can be
written as
Df
ads
Df
m
Df
g
Df
dl
59
where Df
m
is the frequency change caused by mass loading, taking into
account the replacement of solvent molecules by absorbed species and Df
g
reects the frequency shift caused by changes in interactions of the
vibrating surface with the solution caused by specic adsorption. This
could include viscosity changes in the thin liquid layer attached to the
surface and slippage at the interface discussed in Sec. II.B. The frequency
shift Df
dl
reects the mass and viscosity changes in the diuse double layer
associated with changes of the charge density caused by specic adsorp-
tion. df marked in Fig. 9 is a frequency shift that is equal to Df
m
+ Df
g
,
taken at constant charge density, i.e., the total frequency shift, corrected
for double-layer eects.
The inuence of the structural changes of solvents in the very close
vicinity of the adsorbate/solution interface, mentioned above, was attrib-
uted to development of slippage at the interface [74]. The treatment of the
experimental data using Eq. (33) makes it possible to determine the slip
length as a function of surface exceeses of pyridine. In agreement with the
theoretical prediction, k grows with &
a
. The values of k do not exceed 0.3
and 1.2 nm for adsorption from butanol and water solutions, respectively.
The dependence of slip length on surface excess is essentially linear [see Eq.
(35)] for pyhridine adsorption from butanol solution but deviates from
linearity for pyridine adsorption from water. The deviation is attributed to
a reorientation of adsorbed pyridine molecules at the Au surface.
For evaluation of the surface excess of organic species, it may be
necessary to take into account the dependence of the activity coecient
on the concentration [114,115]. In a recent study of the adsorption of
t-butanol on gold [108], relatively large concentrations of the adsorbate
Tsionsky et al. 48
(up to 1 M) were used. In this particular case, the activity coecients of
both the organic material and of the supporting electrolyte (0.1 M LiClO
4
)
were found to change signicantly with increasing concentration of the
adsorbate. This eect was, of course, quite negligible in the case of pyridine
adsorption, for example, where the concentration in solution did not
exceed a few mM.
Our conclusions concerning the use of the EQCM to study sub-
monolayer or monolayer adsorption of organic substances are as fol-
lows:
1. In many cases the absolute value of the expected frequency shift
associated with mass changes, i.e., the rst term in Eq. (59), is
very small. Hence, the use of the EQCM technique to obtain
parameters such as the surface excess, &
ads
, appears to have
little or no advantage over well-established electrochemical
methods.
2. Diculties in processing the EQCM data are oset by the fact
that the response of the EQCM contains important information
(not accessible by other techniques) regarding the structure of
the liquid at the interface, related to the second term in Eq.
(59), which can be evaluated by careful analysis of the data.
3. Interpretation of the EQCM data related to adsorption requires
knowledge of other electrochemical properties, such as the
surface excess and the surface charge density, in addition to the
data on the EQCM response as a function of surface charge
density in the same system in the absence of the adsorbate.
2. The Adsorption of Uracil and of Thiols
In the context of the above considerations, we discuss here two examples
regarding the adsorption of organic molecules on gold electrodes. In Fig.
11 we present the response of the EQCM for gold electrodes in the pre-
sence of uracil in aqueous solutions.
*
The data are similar to those
*
Uracil presents a case of special interest because it can be adsorbed in dierent
congurations: with the nitrogen atom turned towards the surface, like pyridine,
or with the OH groups attached to the surface. The orientation of adsorbed uracil
can be controlled by potential and by pH of the solution (see Refs. 116, 117).
Thus, the interactions of the electrode with the solution can also be controlled.
published in Ref. 118. The adsorption of uracil increases the fundamental
frequency of the EQCM. This can denitely be started only up to a charge
density of 20 AC/cm
2
, where it is possible to make the comparison between
solutions with and without uracil. At more positive charge densities in the
supporting electrolyte without uracil, the response of the EQCM is
distorted by formation of surface oxides, which is suppressed in the
presence of uracil. The EQCMresponse touracil adsorption has interesting
features (see Fig. 11), which obviously carry important information.
However, it is dicult to interpret these data without having the back-
ground data on bare gold surface at positive charges densities. Moreover,
reliable data on the surface excess of uracil, which are needed to estimate
the mass eect, are also available only at charges density lower than +20
AC/cm
2
for the same reasons. It may be possible to understand the data
presented in Fig. 11 by studying the inuence of dierent homologues of
FIG. 11. The frequency shift versus the surface charge density on gold electrode
in 0.1 M LiClO
4
without (closed circles) and with (open circles) uracil (4 mM).
Tsionsky et al. 50
uracil on the response of the EQCM. However, these experiments have not
yet been reported in the literature.
In a recent paper [119], the adsorption of two alkylthiols (butanethiol
and octanethiol) was studied. The two compounds have the same active
thiol group (-SH) and are probably attached to the surface in the same
manner, by oxidative removal of the proton from the SH group:
RS H Au !RS Au H
e
m
60
The use of the same chemistry but dierent molecular weights could
conceptually, help to verify the validity of the results obtained by the
EQCM. The molecular weight of adsorbed species per mole of electrons
passed during adsorption was calculated using the Sauerbrey equation:
Geq
Df
ads
C
m
F
q
ads
61
where Geq is the equivalent weight of the adsorbed species, Df
ads
is the
observed frequency shift, and q
ads
is the charge passed during adsorption.
The latter was obtained by integration of CV curves and correction for the
capacitive process, which amounted to about 25% of the total oxidative
charge. However, this correction for double-layer charging was based on
measurements performed on a dierent type of electrode in a dierent
experimental set-up and may hence be applicable only as a rst approx-
imation. It is important to emphasize that the gold electrodes employed in
the EQCM measurements may have quite dierent surface properties than
electrodes prepared from bulk metal (e.g., wires, foils). Thus, one would be
well advised to repeat at least some of the electrochemical experiments on
the gold surface of the EQCM itself before such data are used in con-
junction with the measurements of frequency changes of the EQCM.
The calculated equivalent weights of butanethiol and octanethiol
adsorbed on the surface were equal to 84 F 5 and 136 F 4, respectively.
These values are in good agreement with the known values of the molecular
weights of the substances studied: 89 and 145. At rst glance this would
seem to be a very good verication that the EQCM response reects only
mass loading of the electrode during the adsorption process. However, it
should be noted that adsorption of thiols (without charge transfer) can
take place at very negative potentials. In the absence of thiols in solutions,
the currents for hydrogen evolution are large at these potentials. At the
same time, solutions containing thiols at negative potentials, where
oxidative adsorption starts (E
ox
), the CV curves do not show remarkable
reduction currents. This indicates that thiols in solutions inhibit the
hydrogen evolution reaction, even at potentials more negative than E
ox
.
Hence the authors assertion that at these potentials there is no adsorption
of thiols can be questioned. A more realistic picture could imply that thiols
are adsorbed on gold surface over the whole potential range studied. They
are chemically adsorbed and rigidly attached to the surface at potentials
more positive than E
ox
. When E< E
ox
thiols could be physically adsorbed,
perhaps in a manner similar to n-butanol. This kind of adsorption leads to
a certain degree of movement of the alkyl chains, resulting from the
vibration of the resonator surface. Being chemically (i.e., more rmly)
attached to the surface, the adsorbed molecules resist vibration, decreasing
the resonator frequency. The longer the alkyl chains, the bigger the fre-
quency shifts. Most likely this is the eect observed in the above-mentioned
work. To clear the surface of thiol and to observe the process of physical
adsorption, as in the case of n-butanol, one needs to apply very negative
potentials to the gold surface, where hydrogen evolution could interfere
seriously with the measurements.
It should be emphasized again that the common diculty in both
cases discussed above is that it is not possible to obtain the dependence of
charge density on potential in the same potential region with and without
the investigated adsorbate in solution. Thus, it is not possible to determine
what part of the charge passed across the interface should be associated
with the adsorption process proper. In many respects this refers also to the
adsorption of charged particles.
Above we discussed only one paper on the adsorption of thiol de-
rivates. Our choice was to some extent arbitrary, and many other exper-
imental data concerning the adsorption of dierent thiol derivates ob-
tained by employing the EQCM technique can be found in the literature.
Detailed discussion of this topic is outside the framework of this chapter.
The work mentioned above represents a typical approach in which the
EQCM data are interpreted only gravimetrically, i.e. by relating all the
change of frequency to the change in weight. This is an oversimplied
approach. We emphasize again that the EQCM data alone do not yield an
understanding of the processed occurring at the metal/solution interface
when organic substances are adsorbed. One needs in addition the data on
surface charge density, surface excesses, and data of the EQCM in the
supporting electrolyte in the absence of the adsorbate. With these require-
ments the research becomes rather complicated, but it is the only way to
use the EQCM technique correctly. The above remarks do not apply to
Tsionsky et al. 52
EQCM studies of relatively thick organic lms, where the overall eect,
including changes in the mass of the lm and its mechanical properties,
could be much larger than eects due to changes in the interaction at the
organic lm/solution interface.
B. The Adsorption of Inorganic Species
1. The Adsorption of Ions
The response of the EQCM to the adsorption of halide anions on gold
surfaces during CV experiments has been reported in several publications
[120126]. The frequency shift associated with iodide adsorption is about
(57) Hz for a 5 MHz crystal. The absolute value of the frequency shift
decreases in the order I
> Br
> Cl
, reaching zero in the case of F
.
Qualitatively, the eect seems understandable, since there is a decrease of
both surface excesses and molecular weights of the adsorbed species in the
same sequence. Diculties do emerge, however, when one tries to use these
data for a quantitative description of the systems studied. First it should be
noted that the frequency shifts observed are of the same order of magni-
tude as those observed in solutions without specic adsorption (see Figs. 6
and 8). Consequently, a correction for electrostatic adsorption must be
applied, as in the case of adsorption of organic species. It is clearly shown,
for example, in Ref. 123, that Df in the presence of Br
has a maximum
value of -8 Hz, while in the absence of Br
, in the supporting electro-

lyte, the maximum frequency shift equals +1.7 Hz. In order to take into
account the electrostatic adsorption as a background, one needs to know
the charge-potential dependence, with and without adsorbed species. It is
impossible to predict a priori whether the eect of electrostatic adsorp-
tion will be small. The composition of the solution inside the diuse dou-
ble layer could change dramatically by the adsorption of halides, because
the pzc can shift and the rational potential could even change its sign at
constant applied potential as a result of adsorption. Unfortunately, in all
discussions of the data obtained in Refs. 120122, 125, the inuence of the
double-layer eect on the response of the EQCM was ignored. Moreover,
the discussions were based on the assumption that the frequency shift was
determined entirely by the Sauerbrey equation, i.e., that it could be
uniquely associated with mass loading. Under this assumption, the molec-
ular weight of adsorbed species per mole of electrons passed during ad-
sorption was calculated according to Eq. (60). In general, the cited works
[120122,125] present two unrelated approaches to treat the EQCM data.
One is based on an a priori assumption that the valences of adsorbed
halides are always equal to unity [120122] and all deviations of the cal-
culated value of Geq from the atomic weight of the corresponding halide,
M, are ascribed to the mass of co-adsorbed water molecules. Another way
[124,125] to explain the experimental data consists of estimating the va-
lency of adsorbed anion as the ratio M/Geq, without taking into account
any co-adsorption. Both approaches represent gross simplications. As
discussed above, adsorption of neutral organic molecules can induce
changes in the interaction of the solvent with the vibrated metal surface,
and the specic adsorption of the charged species would certainly be ex-
pected to exert similar eects, which have to be taken into account. But in
order to account for this eect quantitatively, one needs to have indepen-
dent data on surface excess and on surface charges densities, as in the cases
of adsorption of organic substances. Lacking that, the only conclusion that
can be reached from these measurements is that the EQCM is sensitive to
adsorption of halides.
2. Surface Oxide Formation and Hydrogen Adsorption
a. Oxide Formation on Platinum and Gold Electrodes
In a number of publications, starting with the seminal work of Brucken-
stein and Shay [93], the molecular weight of adsorbed species has been
calculated using the frequency shift of the EQCM and the charge passed
during surface oxide formation on gold [95,127129] and platinum
[92,130132] electrodes. Figure 12 presents the data obtained in Ref. 93,
as a typical response of the EQCM during cycling of a gold electrode in a
wide range of potentials, and the corresponding cyclic voltammograms,
measured simultaneously. The list of publications cited here is by no means
comprehensive. It does reect, however, the main trends in understanding
and interpreting the data obtained. Studies were conducted in dierent
solutions and with electrodes of dierent pretreatment history, dierent
roughness factor (the parameter R
varies between 1.5 and 60), and

obviously dierent crystallographic structures (metals were electrodepos-
ited, evaporated in vacuum, obtained by vacuum sputtering, etc.). It was
generally observed that the frequency shift, Df
ads
, was essentially a linear
function of the charge passed during adsorption ( q
ads
). The common way
of interpreting this fact was based on the assumption that changes of the
frequency of the EQCM reect only mass changes. Moreover, the com-
petitive nature of electrosorption was ignored. In the framework of these
Tsionsky et al. 54
assumptions, the molecular weight of adsorbed species per mole of
electrons passed during adsorption (which is simply the equivalent weight
of a species formed by charge transfer) was calculated using Eq. (61). The
calculated value of Geq was tantalizingly close to half of the molecular
weight of oxygen, varying typically between 7.5 and 9 [92,130,131,133].
Considering the reactions presented in Table 1, the authors of Ref. 131
concluded that while the present results cannot distinguish between the
FIG. 12. Cycling voltammogram(a) and the frequency shift (b) measured simul-
taneously at gold in 0.2 M HClO
4
. Sweep rate =50 mV/sec [93]. (From Ref. 93.)
Pt(II) and Pt(IV) forms of the a-oxide, they do provide in-situ support of
the assumed anhydrous nature of the oxide lm.
The fact that the plot of Df
ads
versus q
ads
is linear supports the view
that two terms, Df
g
and Df
dl
, in the general equation expressing the
response of the EQCM [see Eq. (59)], must be negligibly small during the
formation of a layer of adsorbed oxygen atoms of Pt, and all the observed
eects are due to mass changes, i.e., that Df
ads cDf
m
.
Consider rst the assumption of Df
dl c
0: this would indicate that in
the potential range between 0.8 and 1.4 versus SHE, where a surface oxide
develops on platinum, there is no change in the excess surface charge
density, q, which determines the composition of the diuse double layer
and its weight. If this is really the case, it could be a very interesting
observation, since it is impossible to evaluate the density of electrostatic
charge on the background of the charge-transfer process of surface oxide
formation by other electrochemical techniques.
Regarding the implication of Df
g c
0: this could take place only if the
solvent structure in the layers adjacent to the surface remains unchanged
during adsorption, unlike the situation commonly observed as a result of
adsorption of organic molecules (see Sec. IV.A). Let us assume that at
potentials less positive than the onset of surface oxide formation on
platinum, the water molecules in the layer nearest to the surface were
oriented with hydrogen atoms toward the metal. Then other water
molecules see one oxygen atom per metal atom. Only if reaction A
(see Table 1) takes place can the solvent see again only one oxygen atom
TABLE 1
Possible Steps in the Formation of a Layer of Surface Oxide on Platinum and Their Eect
on the Response of the EQCM
The overall reaction
Orientation of water molecules
on the surface before oxidation
With oxygen
towards the bulk
With protons
towards the bulk
A Pt + H
2
O V PtO +2H
+
+2e
Df
g
c 0 Df
g
p 0
B Pt +2H
2
O V PtO
2
+4H
+
+4e
Df
g
p 0 Df
g
p 0
C Pt +2H
2
O V Pt(OH)
2
+2H
+
+2e
Df
g
p 0 Df
g
c 0
D Pt +4H
2
O V Pt(OH)
4
+4H
+
+4e
Df
g
p 0 Df
g
p 0
Tsionsky et al. 56
per metal atom. If reaction B had taken place, the solvent would see two
oxygen atoms per atom of Pt. In the case of reactions C and D, the solvent
molecules would have to interact with hydrogen atoms instead of oxygen.
It follows that for reactions, B, C, or D, the eect of viscosity in a thin layer
near the surface, Df
g
, should contribute to the overall measured frequency
shift. It will then be highly improbable that the experimentally observed
dependence of Df
ads
on q
ads
would be linear, and the calculated value of
Geq could be close to half the atomic weight of oxygen. Thus, if before
adsorption the water dipoles were oriented with the positive end toward the
surface, then only reaction Acould take place. If the initial orientation was
opposite, then the same type of argument leads to the conclusion that only
reaction C could account for the fact that the interactions between the
vibrating surface and the solutions do not seem to change while a surface
oxide is formed on Pt.
Thus, the data obtained with the EQCM on Pt are consistent with
either reaction A or C shown in Table 1. In order to be able to distinguish
between these two processes, one needs to know the orientation of water
molecules on platinum at potentials where surface oxide formation starts
(E c0.75 V vs. SHE). These potentials are positive with respect to the pzc,
*
which is in the range of 0.150.30 vs. SHE in acid solutions [134,135] and
it is likely that water molecules are oriented with oxygen toward the metal.
Thus, the most likely process of electrochemical oxidation of a Pt surface is
reaction C in Table 1, which produces Pt(OH)
2
. This is a prime example
showing how the results obtained by the EQCM can complement those
derived from classical electrochemical experiments to gain information
regarding the processes taking place at the metal/solution interface.
In Ref. 130 it was shown that the calculated equivalent weight, Geq
[see Eq. (61)] depends slightly on the concentration of H
2
SO
4
: it was
observed that Geq = 7.5 in 0.1 M solution and increases to 10 in 1 M
solution. There could be many reasons for this dependence. The main one
is probably the inuence of the concentration of SO
2
4
at the electrode
surface and in the diuse double layer on the structure of solvent near the
surface. However, this fact does not point unambiguously to co-adsorption
*
The quantity considered here is the potential of zero free charge, which is not
identical to the potential of zero total charge at the metal/solution interface. It is
the former that determines the composition inside the diuse double layer.
of anions, which would increase the mass of the EQCM but has no other
eects on its frequency.
Studies with gold electrodes did not show a similar reproducibility
and agreement between the observed and the expected values of Geq for
surface oxide formation. In Refs. 93,123 the values of Geq obtained were
almost the same: 8.1 and 7.9, respectively. These data indicate that, as in
the case of platinum, the resulting product of surface oxidation is the
hydrated form Au(OH)
2
. However, the authors of Refs. 91, 92 observed
smaller frequency decrease during surface oxide formation on gold than in
Refs. 93, 123. In Ref. 91 this eect was attributed to desorption of anions
during oxide formation. It is shown in Sec. V.B that the response of the
EQCM strongly depends on the history of the electrode surface, for both
platinum and gold, as well as on their morphology and roughness. Thus,
the dierence in data reported in the literature for gold electrodes could be
caused by a dierence in the state of the electrode.
b. Adsorption of Atomic Hydrogen
A monolayer of adsorbed hydrogen on a smooth Pt electrode is about 2.3
ng/cm
2
. This is close to the resolution of most EQCM devices. Hence, this
technique is inherently unsuitable for the measurement of hydrogen
adsorption, at least to the extent that the shift of frequency due to added
mass is concerned. The sensitivity can be increased to some extent by
increasing the roughness factor, but a signicant eect is expected only for
very rough surfaces (e.g., where R =50 100). Unfortunately, on such
surfaces other eects associated with the roughness of the surface may
overshadow the eect of added weight of adsorbed hydrogen (see Sec.
II.D).
Studies of the response of the EQCM to hydrogen adsorption on
platinum were reported in [92,129131,136138]. In most publications a
positive shift in frequency was observed upon adsorption of hydrogen on
platinumin acid solution. This can obviously not be associated with a mass
loading, which should be too small to be detected. The absolute values of
the eect reported in dierent papers vary, but in most cases they did not
exceed 20 Hz when electrodes having a large roughness factor were not
used. The interpretation of the data has commonly been in the framework
of the concept of specic co-adsorption of ions or adsorption of water
molecules. Interestingly, in alkaline solution the EQCM responds in the
opposite direction, the frequency decreasing with increasing coverage
[92]. The regions of hydrogen adsorption in acid and alkaline solutions
Tsionsky et al. 58
FIG. 13. Frequency shift vs. potential for Pt [136] and Au [98] electrodes in acid
and alkaline solutions. (Data for Pt and Au from Refs. 138 and 98, respectively.)
lie on dierent sides of the pzc. Plotted on the same potential scale, the two
sets of data give rise to a roughly parabolic curve, as shown in Fig. 13a,
which is similar to that in Fig. 13b, obtained in the double-layer region for
gold. It is hence most likely that the eects observed for Pt in the hydrogen
adsorption region can be correlated with eects of the diuse double layer,
particularly since the absolute values of the eects on gold and platinum
are similar. The inuence of adsorbed hydrogen on the composition inside
the diuse double layer and on the water structure near the electrode
surface cannot be ruled out, but it seems that electrostatic adsorption in
the diuse double layer is the most important factor determining the
response of the EQCM in the range of hydrogen adsorption, both in acid
and in alkaline solutions.
V. METAL DEPOSITION
A. Deposition on the Same Metal Substrate
In the early work with the EQCM, metal deposition [10,92,139] or
dissolution [140] was used to study the EQCM technique, in particularly
to calibrate it, rather than using the EQCM to study metal deposition.
Copper or silver deposition were mostly used for this purpose, and the
conditions were chosen such that 100% Faradaic eciency could be
ensured. Changes of the properties of the EQCM and the morphology of
the surface were minimized by limiting deposition to very thin layer and
employing current densities that were lowcompared to the limiting current
density in the same system. The sensitivity of the EQCM in the majority of
the investigations was found to be very close to that calculated from the
properties of the crystal itself, although the sensitivity obtained by silver
deposition was consistently a little higher than that of copper. The
descriptions of the conditions of electrodeposition in those publications
are given, as a rule, in the Experimental section of the papers and are so
short that it is dicult, if at all possible, to discuss them in detail. For this
reason we consider below the experimental data obtained in our own
laboratory.
The Sauerbrey equation for electrochemical deposition can be
rewritten as follows:
Df C
m
j M
nF
t 62
Tsionsky et al. 60
where j is the applied current density, t is the time, M is the molecular
weight of the deposited substance, F is the Faraday number, and n is the
number of electrons participating in the reaction. The three solid lines
shown in Fig. 14 represent the frequency shift predicted by the Sauerbrey
equation for deposition of silver, gold, and copper. The points represent
experimental results. For the purpose of calculation it was assumed
that:
1. The metal was uniformly deposited on the surface of the EQCM,
namely that the current distribution was ideally uniform (see
Appendix).
2. In calculating the slopes of these lines, it was further assumed
that electroplating occurred at a Faradaic eciency of 100% for
all three metals.
3. In order to minimize changes of the properties of the EQCM and
the morphology of the surface, lowcurrent densities were applied
and thin deposits were formed. In all cases presented below, the
deposit thickness was less than 50 nm, and the current density did
not exceed 0.1 mA/cm
2
.
The experimental data obtained for electrodeposition of three met-
als, each on its own substrate (silver on silver, copper on copper, and gold
on gold), shown in Fig. 14 can be summarized as follows:
1. For silver the points t the straight line calculated from Eq. (62)
very well. For the particular experiment shown here the deviation
is only 0.25%. In other experiments under similar conditions, the
experimental slope never deviated from that calculated by more
than 1%.
2. For copper and gold, the experimental points could also be tted
to straight lines. However, in these cases, the measured slopes
were 410%lower than those predicted based on Eq. (62). It may
be concluded that, for the case of silver deposition, the expe-
rimental conditions necessary for this equation to apply have
been met. These results show that the sensitivity of the EQCM
agrees with the theoretical value, C
m
. The small deviations ob-
served in a number of experiments resulted from side eects. The
small deviations observed for copper and gold deposition are
consistent with a model [141,142] that postulates the formation
of lower valency ions (Cu
+
and Au
+
) as soluble intermediates,
FIG. 14. Frequency shift of the EQCM during deposition of Ag on Ag from 1.5
M KCN+0.3 M AgNO
3
[141], Au on Au from 20 mM HAuCl
4
, and Cu on Cu
from 1 M H
2
SO
4
+0.5 M CuSO
4
[141]. Open circles, experimental data; solid lines,
calculated from Eq. (62). (Data for Ag and Cu from Ref. 141.)
Tsionsky et al. 62
and the transport of some of these ions to the bulk of the solution,
without being reduced further to the metallic state. A similar
explanation was also given in Ref. 139.
It should be noted that the charge marked as zero in Fig. 14 does
not correspond to the time at which the deposition current was rst turned
on. It corresponds to a time when the deposition process has reached
steady-state conditions, the potential was constant, and there were no
further changes of the concentration prole near the electrode surface.
Otherwise, depending on conditions at the beginning of deposition, one
might observe deviations from the predicted frequency shift. An example
of such behavior was considered in detail in Refs. 141,142.
In Fig. 15 we show an example of how an intentional modication of
surface morphology during deposition manifests itself in the EQCM
response. It is well known that electroplating under mass-transport
limitation leads to surface roughening. Rough surfaces of gold were
prepared by electroplating at a current density of half the limiting current
( j
L
=1 mA/cm
2
). Curve 1 in Fig. 15a shows the ratio of the observed
frequency shift (Df
exp
) to the shift (Df
calc
) predicted by Eq. (62). At short
times the ratio Df
exp
/Df
calc
is much larger than unity, but it decreases
sharply with time and tends to level o at a value of about 1.25. The
response of the QCM is evidently larger than that predicted for the mass
loading alone. At the same time, the width of the resonance (G) increases
monotonically, and does not seem to approach a constant value, as seen in
curve 1 of Fig. 15b. The results obtained during deposition at much lower
current densities ( j/j
L
<<1) show quite a dierent behavior, as seen in
curves 2 in Fig. 15a and b. The ratio Df
exp
/Df
calc
is very close to unity, and
there are practically no changes in &. It is most likely that the peculiarities
of the behavior of both curves at the initial stage of deposition reect the
fact that steady state has not yet been reached. The fact that curve 2 in Fig.
15a lies below unity is just another way of showing the deviation of the
experimental points for gold deposition from those predicted by Eq. (62),
which are shown in Fig. 14.
When deposition was interrupted, no remarkable changes of either
Df or G were observed at open circuit. Thus, the process described by
curve 1 in Fig. 15 could be terminated at any point during the experiment,
in order to obtain a resonator with the desired properties. This observa-
tion was used in a subsequent publication [24] to prepare gold surfaces of
dierent roughness. The widening of the resonance with increasing
roughness was rst demonstrated experimentally in the work of Yama-
moto et al. [19].
B. Early Stages of Metal Deposition on a Foreign Substrate
One of the popular applications of the EQCM is the study of under-
potential deposition (upd ). This application is so wide that any list of
FIG. 15. Changes of the EQCM response during cathodic deposition of Au on
an Au substrate from 10 mM aqueous solution of HAuCl
4
[24] at current densities:
20 AA/cm
2
, open circles; 500 AA/cm
2
, solid circles. (a) Ratio of the observed shift
of the frequency to the shift predicted for mass loading alone. (b) Width of the
resonance. (From Ref. 24.)
Tsionsky et al. 64
citations is of necessity incomplete. Underpotential deposition on Ag as a
substrate was studied for Cd, Pb, and Tl [143] and for T1 [144], Bi [145],
and Pb [146]. Gold as a substrate was used to study upd of Cd [95,147
149], Bi and Pb [149,150], Cu [91,92,149,151], and T1 [152,153]. On Pt
there are data for upd of Bi [154], Cd [155], Cu [92], Pb [156], and Zn [157].
All above data were obtained in aqueous solutions. In Ref. 158 the upd of
Pb on Au was studied in acetonitrile. It is not our intention here to discuss
the peculiarities of the upd processes studied with the EQCM. We can only
call attention to the ideas used for the interpretation of the EQCM data
and to some experimental details in studying adsorption properties with
the EQCM.
The most common approach to explain the data obtained in the upd
region is based on the assumption that the whole response of the EQCM is
due to the change of mass. Under this assumption, most studies tend to
interpret the frequency shift observed in one of two ways:
1. Obtaining the mass of the deposited metal from the frequency
shift and the charge passed during deposition, it is possible to
calculate the apparent valency of adsorbed metal, which is often
identied with the so-called adsorption valency of the metal ad-
atoms. This approach was used, e.g., in Refs. 143,149.
2. The dierence between the mass calculated from the charge
passed and that calculated from the frequency shift could be in-
terpreted as the mass of anions or solvent molecules co-adsorbed
with the metallic species. Such interpretation of the EQCM data
can be found, e.g., in Refs. 92, 147, 154.
It is interesting to note that these two ways of interpreting the data
have never been applied in the same paper. It is true that both phenomena
discussed above might play a role in determining the total response of the
EQCM. However, there is a third phenomenon that must be also consid-
ered. Usually the shift of frequency associated with the formation of a upd
layer does not exceed 10 Hz, hence, it is of the same order of magnitude as
the eects observed for electrostatic adsorption (see Sec III.C). Among the
papers cited above, not a single one takes into account the simple fact that
formation of a upd layer could change the potential of zero charge (pzc)
signicantly. The latter would change the excess surface charge density (at
constant potential) and hence the composition of the diuse double layer.
Moreover, replacing one metal by another could lead to changes of the
solvent structure in the layers nearest to the electrode surface and to
changes of the local interfacial viscosity and the extent of slippage. This
possibility was discussed in Refs. 143 and 98 for the case of thallim upd on
silver and silver upd and opd (overpotential deposition) on gold, respec-
tively.
In Ref. 98 the frequency shift was measured during plating of silver
on gold at dierent current densities (0.810 AA/cm
2
), well below the mass
transportlimited current density. Results of one of these experiments are
shown in Fig. 16. Similar data were obtained for other current densities.
The expected change in frequency, based on Eq. (62), is given by line 2,
while the observed behavior is shown by line 1. Consider the changes in
potential (line 3) and the EQCM frequency (line 1) with time. In the time
interval ending at t
st
, where the potential has reached a constant value of
about 0.63 V versus SHE, the electrode potential changed from 1.2 V and
the charge in the double layer had to change accordingly. Underpotential
and overpotential deposition of silver up to three to four atomic layers had
occurred. In this region, the frequency shift lags behind the calculated
response of the EQCM. Only after this amount of silver has been
deposited does one observe the same slope for lines 1 and 2, which are,
however, about 50 Hz apart. By this time the gold surface has been
completely covered by silver, and all interactions between the gold surface
and the solution were fully replaced by interactions between silver and the
solution. The delay in the frequency shift reects the dierence in these
interactions.
This frequency shift should be corrected by taking into account
changes in electrostatic adsorption caused by the dierence in the pzc of
the two metals and the dierences in charge density of gold at 1.2 V (at
the beginning of the experiment) and that of silver at 0.63 V, when
steady state has been reached. However, this correction is much less
(about 57 Hz) than that observed experimentally (for details, see Ref.
98).
In experiments with thallium deposition [143], the calculated and
measured slopes became parallel after a charge corresponding to 1.5 at-
omic layers of thallium covered the gold surface has been passed. In this
case the dierence between the calculated and measured frequency, df, is
about 20 Hz.
Qualitatively similar delays of the frequency shift were observed [159]
for the deposition of cobalt, nickel, and iron on the gold surface: in all three
cases the frequency is practically constant until a fewatomic layers of metal
have been deposited on the gold surface. Deposition of these metals is
Tsionsky et al. 66
FIG. 16. Deposition of a thin layer of silver on gold from 0.01 M AgNO
3
+0.5
MHClO
4
at 2.5 AA/cm
2
. Dependence of the frequency shift on time: 1, experiment;
2, calculation; 3, the electrode potential during deposition.
complicated by hydrogen evolution. However, this process occurring in
parallel with metal deposition cannot be the reason for the delay in the
frequency shift. The overvoltage of hydrogen evolution on gold is higher
than on the transition metals forming the upd layer [160], and it is
reasonable to suggest that the eciency of metal deposition decreases as
the metal is being deposited. If hydrogen evolution was the only reason for
the deviation from the straight line predicted by Eq. (62), then the
experimental slope of the frequency-time curves would have to be closer
to the calculated values at the beginning of deposition. We suppose that the
eect observed in Ref. 159 was of the same nature as that shown in Fig. 16
for silver deposition on gold and also observed for thallium deposition on
gold, namely, a dierence in the interactions of the solvent with dierent
metals, including a shift of the pzc associated with the formation of a new
surface.
Thus, there are dierent ways to interpret the EQCM data when
studying upd phenomena. The fact that the frequency shift in the cases of
upd layer formation is of the same order of magnitude as in the double
layer region makes it imperative to take into account the eects of the
latter when discussing upd phenomena. Obviously there are some dicul-
ties, e.g., one can discuss the dierence, Df, between the initial state,
substrate surface before upd, and the nal state, when the surface has been
fully covered by the deposited metal, assuming that the dependences of the
EQCM response on potential are known for both metals. However, it
seems rather problematic to discuss and interpret seriously the intermedi-
ate region, when the substrate is only partially coated by the metal being
deposited. In the latter case one does not know even the charge density of
the surface.
The uncertainty does not stem only from the way the data are
interpreted. It also depends heavily on the experimental method
employed to obtain these data. In Fig. 17 we show an example of data
obtained during the deposition of Ag on Au, in both the upd and the
opd regions, measured galvanostatically ( j = constant) and by cyclic
voltammetry at dierent scan rates. The cathodic parts of the CV curves
were used to construct the plot of Df as a function of the charge passed.
In measurements conducted under galvanostatic conditions, all values
of the frequency shift were above the calculated line, as shown also in
Fig. 16, practically independent of the applied current density. The
response always had the same shape: a delay until a few atomic layers
have been deposited, followed by a line having the theoretical slope, but
shifted upwards (i.e., to higher frequencies). Cyclic voltammetry leads
Tsionsky et al. 68
to a range of dierent results, depending on scan rate. Slow scan rates
show remarkable delay, similar to that observed in galvanostatic
measurements, which however disappears gradually at higher scan rates.
At suciently high scan rates the delay is no longer observed, and the
frequency shift can even be larger than that predicted by mass loading
alone.
FIG. 17. Dependence of the frequency shift on charge passed during deposition
of Ag from 0.01 M AgNO
3
+0.5 M HClO
4
on Au. 1, at constant current density.
Points at dierent potential scan rates as marked. Curve 2 was calculated from
Eq. (63). The inset shows a typical CV.
The dependence of the response of the EQCM on sweep rate is not
surprising, considering that (1) it is well known that the structure and
surface morphology of electrodeposits depends on the applied current
densities and (2) during the scan, the deposition current changes drastically
(see inset, Fig. 17). Thus, cyclic voltammetry does not provide steady-state
conditions, and hence the concentration prole near the electrode surface
changes continuously with time. Clearly, the Sauerbrey equation applies
under steady state only (see Fig. 14 and the relevant discussion). Unfor-
tunately, in the majority of the publications dealing with upd, only cyclic
voltammetry at relatively high scan rates was used. Moreover, in most
publications the dependence of the EQCM response on scan rate was not
discussed. Such results cannot be reliably interpreted in erms of co-
adsorption of ions or the degree of partial charge transfer unless the eect
of sweep rate has been evaluated and discussed in detail or the data have
been compared to results of steady-state measurements such as constant-
current deposition (see Sec. VI.B).
Two other factors should be considered in comparing results
obtained by dierent methods: surface diusion of the metal atoms in
the upd layer and their diusion into the substrate, sometimes referred to
as surface alloy formation [161]. Steady state of these processes is gener-
ally not achieved when deposition is conducted at high current density. A
study of the inuence of the current density on the EQCM response could
help to understand the role of these processes in the early stages of de-
position. In any event, it seems to be preferable to use galvanostatic
conditions in these experiments rather than cyclic voltammetry.
Other systems may behave dierently from that presented in Fig. 17.
Thus, it was found that the upd of Cu on Ag shows characteristics similar to
that of Ag/Au [141,142]. On the other hand, the data for T1/Ag do not
show any dependence on potential scan rate [143]. Signicantly, among
these three systems, only T1 does not form alloys with the substrate upon
which it forms a upd layer.
VI. THE INFLUENCE OF ROUGHNESS ON THE
RESPONSE OF THE QCM IN LIQUIDS
When experimental data on the EQCM response were considered in the
above sections, the mass change of the vibrating surface and its interactions
with the solution were taken into account. It was borne in mind that the
sensitivity of the quartz crystal resonator to change of mass, expressed by
Tsionsky et al. 70
the constant C
m
in Eq. (1), has the dimensions of Hzcm
2
g
1
, where the
unit area refers to the cross section of the resonator rather than the geo-
metrical surface area. The apparent sensitivity of the EQCM is generally
larger than that calculated from its properties, due to the surface rough-
ness, expressed by the roughness factor, R
: Thus, for the homogeneous

deposition on rough surfaces, the apparent sensitivity of the EQCM is
given be R
C
m
, and the Sauerbrey equation takes the form:
Df

RC
m
Dm 63
where Dm refers to the real surface area. This should be taken into account
when comparing data obtained in dierent publications and in the
calculation of some extensive parameter, such as surface excess, from
EQCM data. However, this electrochemical roughness factor, R
,
responds to roughness on the atomic or molecular scale, i.e., on a scale
similar to the thickness of the Helmholtz double layer. In contrast, the
length scale relevant to the EQCM response is the hydrodynamic decay
length (see Sec. II.C). As a result, the roughness factor R
extracted fromthe
EQCM measurements given by Eq. (41) (see Sec. II.C) can dier from the
electrochemically determined factor R
. Analysis of the EQCMresponse on

rough surfaces can yield additional information on the geometry of the
interface and on the hydrodynamics in solution at a vibrating rough
surface.
A. The Nonelectrochemical Case
Before discussing the role of roughness in electrochemical systems, we shall
consider the experimental data on the response of a QCM immersed in
nonconducting liquids, where electrochemical phenomena do not play a
role.
A correlation between the QCM signal and the surface morphology,
as determined by micro-prolometry and STM techniques, has already
been established in early experiments with the QCM in liquids [12,14].
Measurements indicated that the mechanical impedance, Z
out
, increases
with increasing surface roughness. In contrast to smooth surfaces, inter-
actions of rough oscillating surfaces with liquids do not contribute equally
to Re(Z
out
) and Im(Z
out
) [12,162,163]. It was also found that the roughness
leads to new dependencies of the frequency shift on viscosity, which does
not appear for smooth surfaces. For instance, the experimental data
obtained in methanol-water mixtures and in alcohols [14] demonstrated
that the eect of roughness on the QCM is most pronounced for low
viscosities, where the liquid-induced shift of the resonance frequency is
small. This conclusion agrees with the theoretical predictions discussed in
Sec. II.D [see Eqs. (38) and (39)]. Theory shows that at low viscosities, the
QCM response in liquids is mainly determined by the contribution of the
nonuniform pressure distribution which is developed in the liquid under
the inuence of a rough oscillating surface [164].
Experiments in liquids having a wide range of viscosity and density
were performed and the response of the QCM analyzed using the theoret-
ical models described in Sec. II.D [24]. Both parameters characterizing the
resonator, the shift in fundamental frequency and the width of the
resonance were measured simultaneously. The usual form of presenting
the experimental data in liquids is to plot the real and the imaginary
components of the response of the QCM as a function of the density of the
liquid or of the parameter

qg
p
. However, these parameters are the natural
variables only for ideally at interfaces. Equations (38), (39), (48), and (49)
show that for rough surfaces it is more convenient to consider the
quantities G/ f
2
q and Df / f
2
q as a function of the velocity decay length in
the liquid, d, as shown in Fig. 18. The dependence of these two parameters
on d is linear for the ideally smooth surface of the quartz crystal resonator
loaded on one side (see line 1, Fig. 18a,b). Close points in these gures
represent data measured on a relatively smooth surface (obtained by
vacuum sputtering), while open points were taken on a surface with strong
roughness, prepared by electroplating at a current density close to j
L
, as
described in Sec. V.A. The deviation of the data from the straight line 1
calculated for an ideally smooth surface increases with increasing rough-
ness as expected.
The experimental dependence of the quantity G/ f
2
q on the velocity
decay length exhibits a sharp increase at low values of d, followed by a
gentle growth at large value of d. This eect becomes more pronounced
with increasing roughness (open circles).
In Fig. 19a, the theoretical dependence of the function G/ f
2
q on d is
given (lines 24) for dierent values of the correlation length of roughness,
n
H
, and a xed value of the lm thickness parameter, L, in the framework
of the theory developed for strong roughness (see Sec. II.D). At large
values of d, the calculated lines approach line 1 for an ideally smooth
surface. This behavior can be understood since it becomes dicult for the
liquid to move inside pores in the surface lm when d is much larger than
the size of the pores. In the limiting case the liquid moves in phase with the
Tsionsky et al. 72
solid surface, acting only as a mass loading, but adding nothing to the
width of the resonance.
Line 5 in Fig. 19 is calculated for a surface having slight roughness,
according to Eqs. (39) and (44). The hydrodynamic roughness factor R is
chosen so that this line connects the origin with the experimental point for
the highest value of d. This yielded a value of R =1.3.
FIG. 18. Dependence (a) of the parameter G/qf
2
and (b) Df /qf
2
on the velocity
decay length in dierent liquids, for an ideally smooth surface (lines 1), and
experimental data for two real surfaces: vacuum-sputtered gold (open circles) and
electrochemically deposited gold (closed circles). Lines 2 and 3 represent results of
parameter tting (see text). (From Ref. 24.)
FIG. 19. Dependence of the parameter G/qf
2
on the velocity decay length:
Pointsexperimental data. Lines 1a and 1b, ideally smooth surface; 2a4a,
inuence of strong roughness according to Eq. (49) for dierent values of n
H
(2a
69; 3a172; 4a276 nm) and L=506 nm; 2b4b, the same for dierent values of L
(2b460; 3b506; 4b690 nm) and n
H
=172 nm. Lines 5a and 5b calculated for
slight roughness [roughness factor R =1.3, Eqs. (39) and (44)]. (From Ref. 24.)
Tsionsky et al. 74
Curves 24 in Fig. 19b were calculated for dierent values of the lm
thickness, L, and a constant value of the roughness correlation length, n
H
,
according to Eq. (49). Lines 1 and 5 are the same as in Fig. 19a. The width
of the resonance is seen to increase with increasing lm thickness.
Fig. 19 shows that there is no way to t the experimental data as-
suming that only one type of roughness is presented on the surface. We are
thus forced to conclude that in these experiments the surface has a
multiscale roughness shown schematically in Fig. 20. The structure of
this rough surface is a combination of a slight and a strong roughness
shown in Fig. 5a and b. When this is taken into account, it is possible to use
(Eqs. (38), (39), (48), and (49)), to calculate the shift in resonance frequency
and in the width of the resonance and t the experiments to the calculated
curves with properly chosen values of the parameters of strong roughness.
The result of such a t is shown in Fig. 18, curves 2 and 3. For details of the
tting procedure, the limitation associated with the use of a simplied
model, and for a comparison with the STM data, see Ref. 24. Here we
should emphasize only one point of major importance for electrochemical
use of the QCM. The velocity decay length of most solvents of interest for
electrochemical and analytical purposes happens to be at the lower end of
the values of d shown in Figs. 18 and 19. This is the region where the
interplay between the two types of roughness is the strongest, and it is the
FIG. 20. Schematic representation of multiscale roughness. This structure is a
combination of a slight and a strong roughness shown in Fig. 4a,b. (From Ref. 24.)
From Ref. (24), with permisson on the American Chemical Society.
most dicult to t the data to either model. This inherent diculty should
be borne in mind whenever an attempt is made to interpret the impedance
response of the EQCM operating in typical solvents such as water,
alcohols, and many other nonaqueous solvents employed in electrochem-
istry.
The importance of measuring the imaginary component of the quartz
crystal in order to study metal deposition and dissolution processes has
also been noted [163,165]. In this way it was possible to separate contri-
butions of mass loading and roughness to EQCM response and to
characterize the electrode roughness.
B. The Electrochemical Case
There are only few publications where the response of the EQCM was
studied on intentionally roughened surfaces. Platinum, silver, or gold
surfaces were roughened by repeated cycling over a wide range of potential,
including the region of surface oxide formation [23,166]. This is known to
lead to reconstruction of the electrode surface and to its roughening. In
[25,167,168] the roughened surfaces of platinumand gold were prepared by
electrodeposition at current densities close to the mass transport limita-
tion. An example of preparation of rough surfaces by such technique is
shown in Fig. 15.
Fig. 21 shows how the response of the EQCM changes with the
change of surface roughness, induced by extensive cycling into the region
of surface oxide formation. When the surface was not roughened, the loops
describing the shift in frequency with potential associated with surface
oxide formation have a clockwise direction (see curves 1, Fig. 21a,c). On
very rough surfaces (represented by curves 3), the loops are in the oppo-
site direction. The data for platinum electrode (Fig. 21a,b) were taken
from Ref. 166, in which only the shift of resonance frequency was
measured. The data on gold electrode were obtained in our own labora-
tory, and both the shift of frequency and the width of the resonance were
measured (Fig. 21c,d). The latter shows that when the surface is suciently
smooth, there are no changes in the width of resonance with the potential.
The corresponding curves for rough surfaces, when the resonance is wide
(G >3 kHz), show strong potential dependence and remarkable hysteresis.
On the one hand the comparison of voltammograms and dependence of the
responses of the EQCM on potential clearly shows that the hysteresis is
associated with surface oxide formation. On the other hand, the eect
Tsionsky et al. 76
FIG. 21. (a) The inuence of the number of oxidation-reduction cycles on the
frequency response for platinum in 0.2 M H
2
SO
4
at 100 mV/sec (1, 100; 2, 2,000; 3,
10,000 cycles). (b) Stabilized cycling voltammogram for Pt electrode. Frequency
shift (c) and width of resonance (d) for gold electrodes in 0.1 MHClO
4
at 10 mV/sec
(1, 4; 2, 100; 3, 500 cycles). (Fig. 21a,b from Ref. 165.)
cannot be ascribed to mass loading, because the frequency shift on rough
surfaces is not only larger than that on smooth surfacesthe eect has a
reverse sign. Moreover, mass loading alone cannot lead to changes in G.
The loop of frequency shift changes its sign also in the region of hydrogen
adsorption on platinum. It should be noted that the surface of Pt is much
more resistant to roughening than that of gold. Thus, comparing Fig. 21a
and c, it would seem that cycling 2,000 and 10,000 times on Pt has an eect
comparable to that of cycling Au 10 and 500 times, respectively. However,
the experiments on Pt and Au shown here were performed under similar,
but not identical, conditions.
Comparison of Fig. 21c,d for a highly rough surface (line 3) shows
that a decrease in width is associated with a positive shift in resonance
frequency in the region of surface oxide formation. This is consistent with
the notion that both eects result from a weakening of the interaction
between the vibrating surface and the liquid under surface oxidation.
In Fig. 22 we show CV curves on gold surfaces having dierent
degrees of roughness. Curve S1 was taken on a vacuum sputtered (as
received) gold surface, while curves S2S4 were obtained on surfaces of
increasing roughness, achieved in this case by electroplating gold at a
current density close to the limiting current density. Comparison of these
curves shows that plating at currents densities close to j
L
changes both the
real surface area of the electrode and its energetics. The former is
manifested by the increased charge corresponding to the oxidation/reduc-
tion peaks, while the latter shows up as a shift of the various peak
potentials. The ratio of real surface areas of very rough electrode (curve
S4) and untreated gold (curve S1) is in the range of 34. The exact value
depends on the range of potential over which the current is integrated. The
response of the EQCM for the same surfaces in the potential region
including surface oxide formation was very similar to that obtained with
surfaces roughened by repeated cycling (Fig. 21c,d).
The data shown in Fig. 21 are complicated by surface oxide forma-
tion and adsorption of atomic hydrogen. We simplied the situation [25],
restricting the limits of cycling to the double-layer region. In Figs. 23a and
b we showtypical sets of data on the dependence of the resonant frequency
and the width of the admittance on potential in the double-layer region for
gold surfaces of four dierent levels of roughness. The corresponding
cyclic voltammograms are shown in Fig. 22. The dependence Df on E for
an untreated (as received) surface (curve S1, Fig. 23) is given in Fig. 23a.
Tsionsky et al. 78
FIG. 22. Cycling voltammetry (5 mV/sec) for gold electrodes of dierent
roughness [25]. S1, untreated (as received) surface. S2S4 surfaces obtained by
electrodeposition of gold at currents densities close to the limiting current density.
(Inset) Approximate values of the width of resonance. Curves S2 and S3 lie between
curves S1 and S4, in some parts coinciding with them. Arrows P(S1) and P(S4) show
the peak currents for reduction of the surface oxide measured for the surfaces S1
and S4, respectively. (From Ref. 25.)
FIG. 23. Dependence of the frequency shift (a) and the width of the resonance
(b) of the EQCM on potential for dierent gold surfaces, S1S4. (From Ref. 25.)
Tsionsky et al. 80
This is similar to that obtained in our previous work [61,97,98] (see
Fig. 6). The slight decrease of frequency with increasing positive potential
in the range of 0.00.8 V is not seen on the scale of Fig. 23a, compared to
the dramatic increase seen for surfaces with very high roughness, repre-
sented by line S4.
All the data obtained with rough surfaces and the discussion of these
data [23,25,166168] lead to the following conclusions:
1. The roughness of the electrode has a profound inuence on the
response of the EQCM (see Figs. 21 and 23). This may explain
the unusually large discrepancies among data obtained with the
EQCM in dierent laboratories (not necessarily on intentionally
roughened surfaces). A good example is the large discrepancy
among data reported for the region of surface oxide formation on
gold [76,9193,123,140].
2. The response of the EQCM on rough surfaces cannot be treated
in terms of the electrochemically dened roughness factor R
,
which is obtained from adsorption phenomena, e.g., from data
such as presented in Fig. 22. This quantity can be considered as
representing all adsorption sites on the surface, which is
equivalent to the surface roughness on the atomic scale.
However, the response of the EQCM depends on roughness on
a mesoscopic scale, which is comparable to the hydrodynamic
velocity decay length rather than to the double layer thickness.
3. The width of the resonance is an important characteristic of the
surface, as seen in Fig. 23b, and can serve as a semi-quantitative
measure of its roughness on the scale relevant to the response
of the EQCM. Unfortunately, only very few publications so far
contain this information.
4. The peculiarities of the response of the EQCM on rough surfaces
do not depend on the technique employed for roughening the
surface, be it repeated cycling or electrochemical deposition at
current densities close to j
L
.
5. The dierence in behavior of rough and smooth surfaces does not
depend on specic adsorption. It shows up even in the absence of
any specically adsorbed species, as seen in Fig. 23.
In addition to the conclusions drawn above, one is still left with the
need to interpret the dependence of the response of the EQCMon potential
on rough surfaces (see Figs. 21 and 23). Attempts to provide a qualitative
interpretation have been made [167,169]. The authors ascribed the eects
on rough surfaces to formation of a structured region of solvent which
leads to increased viscosity and consequent frequency changes and agreed
that the exact nature of the changes in the surface . . . still has to be
established [167]. Thus, they assumed that the properties of that struc-
tured region of solvent near the electrode depend on adsorption and on
potential.
We [25] tried to describe the eects observed in the double-layer
region (Fig. 23) associated with increased roughness. It was taken into
account that a single potential scan cannot reconstruct the surface mor-
phology, meaning that roughness was independent of potential, and
assumed that the interfacial viscosity depends on electrode potential, but
this dependence itself is independent of the surface morphology, namely it
is the same for rough and smooth electrodes. Under these conditions and
suggestions (Sec. II.D), discussing data in Fig. 18 that every given surface
has only one type of slight roughness and only one type of strong
roughness, we tried to describe the experimetal data shown in Fig. 23 by
corresponding models of roughness. That attempt failed, as discussed in
[25]. However, this conspicuous failure does not seem strange. In reality,
rough surfaces tend to have very complex structures. It is possible in
general to describe the experimental results on the dependencies of the
EQCM response obtained for real surfaces of high roughness on the pro-
perties of liquids with only two types of roughness, as was done for the data
presented in Fig. 18. However, such a schematic approach to describe the
ne eects of inuence of potential on properties of a thin layer close to
the electrode surface is clearly problematic. It is not inconceivable that
the roughened surfaces studied in Ref. 25 and other works [167,169] have
some porous structure or/and structure with cavities. The characteristic
dimensions of these pores/cavities are comparable to the thickness of the
interfacial layer, the properties of which depend in turn on the applied
potential. Qualitatively this possibility has been discussed [25]. However,
one is bound to agree with Wilde et al. [167] that the exact nature of
changes in the surface . . . has still to be established. The mere problem is
very interesting. It seems evident that the large and well-dened eects
observed on rough surfaces are induced by phenomena taking place at the
electrode/solution interface. These phenomena also take place on smooth
surfaces; however, they appear very slightly in the EQCM response. In this
sense, rough surfaces could serve as an amplier to phenomena occur-
ring at the interface.
Tsionsky et al. 82
VII. CONCLUSION
The quartz crystal resonator is a useful device for the study of thin-layer
and interfacial phenomena. The crystals commonly employed have a
fundamental resonance frequency of 510 MHz and a resolution of the
order of 0.10.5 Hz. This high resolution makes the device sensitive to a
myriad of physical phenomena, of which some are interrelated and some
quite independent of each other. It cannot be overemphasized that the
quartz crystal resonator acts as a true microbalance (more appropriately a
nanobalance) only under particular conditions, which have been discussed
in this chapter in some detail.
When acting as a microbalance, it is sometimes stated that the change
in resonant frequency is an absolute measure of the change in mass loading,
namely that this device does not require calibration. This statement is only
partially true, and one is well advised to calibrate the QCM. Admittedly,
the constant C
m
in Eq. (1) can be calculated from the fundamental
properties of the quartz crystal, as given by the Sauerbrey equation [see
Eq. (9)]. However, for this constant to be applicable for the determination
of the added mass, several implicit assumptions must be made. Primarily, it
must be assumed that the mass distribution on the surface is uniform. It
must also be borne in mind that the Sauerbrey equation only applies to thin
lms, such that the thickness of the lm is small compared to the thickness
of the crystal itself. In addition, the EQCM can operate as a true
microbalance only if, in the course of the process being studied, the nature
of the interfaceits roughness, the density and viscosity of the solution
adjacent to it, and the structure of the solvent in contact with itis kept
constant.
In this chapter we have limited our discussion to certain aspects of the
eld, related particularly to submonolayer phenomena, such as the mea-
surement of adsorption (surface excess), double-layer structure, and the
underpotential deposition of metals, with an extension to the study of
deposition of very thin metal lms.
Some of the main conclusions are as follows:
The shift in frequency observed experimentally cannot be interpreted
in terms of a change in mass loading alone unless the conditions
have been carefully chosen to ensure that this is the only factor
aecting the resonance frequency.
It seems to be essential to measure the admittance spectrum and
determine both the resonant frequency shift and the width of the
resonance simultaneously. This yields additional information not
available from measurement of the resonant frequency alone and
can hence provide more detailed interpretation of processes oc-
curring at the solid/liquid interface.
When properly applied, the EQCM can provide invaluable in-
formation regarding the structure of the liquid at the interface and
its dependence on the nature of the surface and the applied
potential.
Adsorption of both organic and inorganic species can change the
nature of the surface (e.g., convert it from hydrophilic to
hydrophobic). The latter can lead to some degree of slippage at
the solid/liquid interface, with the slip length depending on the
surface excesses. Such phenomena can, in principal, be detected as
changes of both the frequency and the width of the resonance.
Electrostatic adsorption of ions or any specic adsorption can result
in changes of the viscosity in a thin liquid layer near the interface.
This eect can be observed by measuring the frequency shift.
However, the changes of the width of resonance could be too small
to be detected, as shown in Eqs. (23) and (24).
Surface roughness is of paramount importance in the use of the
EQCM. The existing theories provide a description of the QCM
response for rough surfaces in two limiting cases of slight and
strong roughness. However, much is left to be developed for a
quantitative interpretation of data obtained for real surfaces. In
order to overcome the gap between existing theory and experi-
ments, measurements on specially prepared surfaces with well-
dened roughness should be performed.
Numerous experimental techniques employed in electrochemistry
(adsorption, double-layer capacitance, optical, rotating disc elec-
trode, and EQCM) are sensitive to roughness of electrode
surfaces. It should be noted that each technique probes roughness
on the particular characteristic scale only, which is the atomic
scale for adsorption and double-layer capacitance measurements,
a wavelength of light for optical measurements, the Nernst dif-
fusionlayer for rotating disc electrode experiments, and the hydro-
dynamic velocity decay length for the EQCM. Thus, the im-
pedance of the EQCMwould be expected to respond to roughness
of about 10 nm and above, ignoring most of the so-called elec-
Tsionsky et al. 84
trochemical (atomic scale) roughness, but detecting roughness
that can usually be ignored in experiments conducted under mass
transport limitations.
It has been observed that the dependence of the response of the
EQCM on potential is strongly enhanced on very rough surfaces.
It would therefore seemthat highly rough surfaces tend to amplify
the eects of potential on the structure of the interface and could
serve as a useful tool to study such eects.
The EQCM should always be used as an addition tool, in com-
bination with common electrochemical techniques (such as EIS,
chrono-coulometry, electromodulated reectance spectroscopy,
FTIR spectroscopy, and in situ STM and AFM techniques, as
may be appropriate), not replacing them.
The results obtained by the EQCM contain information relevant to
the understanding of phenomena in the area of nano-tribology,
where techniques such as surface force apparatus and atomic force
microscopy are used. In both cases the results carry information
regarding the properties of a nano-scale layer of liquid at the
interface.
An important part of modern experimental electrochemistry has
been performed on single-crystal electrodes. In contrast, the metal
deposited on the surface of the quartz resonator has at best a
preferred crystal orientation. Studies with an EQCMhaving a true
single-crystal surface have not yet been reported. Making a thin
(about 1Am) stable single-crystal metal layer on the surface of
quartz seems to be an insurmountable problem.
So far most of the EQCM data were analyzed on a qualitative level
only. The next step in EQCM studies requires a quantitative treatment of
the experimental results. The theoretical basis for the solution of this
problem already exists and has been discussed in this review. Joint
experimental and theoretical eorts to elevate the EQCM technique to a
new level present a challenge for further investigators.
Finally it would seem that, in spite of some shortcomings, the
potential advantages of the EQCM far exceed its limitations. There are
many challenges to overcome, and the EQCM will undoubtedly continue
to be one of the important tools in studies of metal/solution interfaces in
general.
VIII. APPENDIX
A. Nonuniform Film on the Surface
The surface mass-density of the lm can be written in the form
Dm
f
R Dm
f
Dm
1
f
R A1
where Dm
f
is an average mass density of the lm and Dm
1
f
R is the
deviation of the density from the average value.
The shear-mode oscillations in quartz crystal, which are now the
solution of the three-dimensional wave equation, can be written in the form
uz; q Aqcospqz A2
where u(z,q) is the two-dimensional Fourier transform of the displacement
in the bulk of the quartz crystal, q is a tangential wave vector, p(q)
2
= k
2
q
2
and k = x
0
(q
q
/l
q
)
1/2
is the wave number of the shear wave in quartz.
According to the boundary condition [Eq. (11)], the amplitude of the shear
wave A(q) satises the following integral equation:
l
q
pqtanpqd x
2
Dm
f
Aqcospqd
x
2
dK
2p
2
Dm
1
f
q KAKcospKd A3
Then, following the procedure described in Ref. 170, we obtain the equa-
tion for the resonance frequency [Eq. (12)] in the text (see Sec. B.2).
B. Experimental Remarks
The experimental application of the EQCM has been described in a
number of paper [1,2,10,171]. The section of Experimental Methods
in Ref. 3 can be particularly recommended. However, a few additional
remarks may be appropriate.
1. Current Distribution
Uniformity of the current distribution is important, since the current
density in Eq. (62) is calculated per unit of real surface area, while the sen-
sitivity of the EQCM is related to the unit area of cross section of the
piezoelectrically sensitive region of the resonator, and the area of the
Tsionsky et al. 86
working electrode in contact with solution is, in many cases, dierent. This
noncoincidence arises from the geometry of the two electrodes on the
quartz crystal resonators: relatively large ags or nonsymmetrical elec-
trodes, where one side of the resonator is completely covered by the metal
while the other side is covered only in the middle. Crystals that are only
partly covered by metal (which are usual for many types of manufactured
quartz crystal resonators) could lead to highly uneven current distributions
close to the edges of the coating. Moreover, the sensitivity of the resonator
itself near to the edges may be dierent from its value in the center. This
question has been discussed [142,172].
It is better to use the quartz crystal resonators with one electrode that
covers the entire liquid side of the crystal. Besides better current distribu-
tion, there is one additional reason why this resonator construction should
be preferred: It has been shown [173] that the impedance of the EQCM is
inuenced by the conductivity and dielectric constant of the liquid. This
spurious inuence is eliminated if one side of the resonator, which is used as
the working electrode, is completely covered by the metal.
2. Removing the Cr Underlayer
Purchased resonators usually have a thin underlayer of chromium or
titanium between the quartz and the electrodes used to excite the crystal
(usually, but not always, gold). This underlayer serves to improve the
adhesion. In contact with electrolytes and under the inuence of applied
anodic potentials, Cr can diuse through the gold to the metal/solution
interface and be oxidized there anodically, thus distorting the electrochem-
ical response of the system studied. From this point of view, the combina-
tion of chromium underlayer and gold electrode is the most troublesome.
To remove this chromiumlayer, one side of the resonator can be pretreated
chemically (e.g., in concentrated HNO
3
) or electrochemically [111,174,
175]. The procedure used in Ref. 111 allows removal of the chromium
underlayer without deterioration of adhesion. The potential is cycled
initially in a narrow range, between 0.2 and 0.2 V vs. SHE. The potential
range is increased gradually up to 0.4 to 1.4 V vs. SHE. This procedure
must be conducted slowly, so that the oxidation current for Cr dissolution
does not exceed about 0.1 mA/cm
2
. Cycling the potential only in the wider
range can diminish the adhesion of gold to the quartz substrate.
Anodic currents associated with Cr dissolution were absent when
silver, rather than gold, was used as the exiting electrodes. Also, when
Au(Cr)-covered crystals were used in nonaqueous solutions (e.g., butanol
[111] or acetonitrile [158]), anodic oxidation of chromium was not ob-
served. When Ti was used instead of Cr as the adhesion-promoting layer,
dissolution of this metal through gold was very slow. Long-term use of
Au(Ti) coatings shows that Ti also disappears, but the adhesion of gold to
the resonator surface remains as before.
3. Solution Height and Bubbles
There are some requirements pertaining to the construction of electro-
chemical cell in combination with an EQCM. Erratic results can be
observed if the cell is not lled completely with electrolyte, as shown in
Fig. 24. The cell used in these experiments had the working electrode
compartment made in the shape of a vertical tube closed at the bottom by
the disk of quartz resonator. This compartment was of 11 cm height and a
diameter about equal to the diameter of the resonator, 1.5 cm. The regions
marked by A in this gure were taken with the working electrode
compartment full with solution. In regions B and E the distance between
the quartz plate and the meniscus was 10 cm, and for region C it was 3 cm.
The eect of a small gas bubble (about 4 mm in diameter) at the top of the
FIG. 24. Instabilities in frequency resulting from partial lling of the working
electrode compartment: (A) full; (B, C, and E) partially lled; (D) with a small
bubble at the top. (From Ref. 61.)
Tsionsky et al. 88
working electrode compartment is shown in region D. We also noted that
when the meniscus had been left undisturbed for a long time, stable
readings of the frequency could be obtained even in partially lled cells.
However, any disturbance (e.g., low-frequency vibration of the working
table) would cause instabilities similar to those shown in sections A or E of
Fig. 24. These eects seem to be similar to those discussed in Ref. 176.
Instabilities of the type shown in Fig. 24 can be eliminated by carefully
lling the working electrode compartment so that it has no liquid/gas
interfaces. The design of a tube-shaped cell, closed at one side by the
resonator plate, allows using small amounts of electrolyte, down to few
cm
3
, thus oering some experimental advantages. It is evident that any
other constructions where the distance between the vibrating resonator
and gas/solution interface could be enlarged could decrease the under-
sirable side eects shown in Fig. 24.
Many processes are accompanied by gas evolution, which can com-
plicate the use of the EQCM. However, it is possible to use the EQCM if
gas evolution is not accompanied by attachment of bubbles to the surface
and some special precautions are taken (e.g., vertical electrode surface
as used in Ref. 139, or surfaces that are continuously rinsed by ow of
solution).
4. Measurements of the EQCM During Cyclic Voltammetry
In many publications concerning measurement of adsorption or the upd of
metals with the EQCM, the potential is cycled between certain limits and
the response of the EQCMis recorded as a function of potential. While this
method is in principle valid, great care must be exercised in its application
to any specic system for a number of reasons:
1. It must be ensured that the sweep rate does not exceed the rate of
collection of data for the EQCM. There is no well-established
criterion for the highest sweep rate that can be applied in a given
experimental system, but the following simple example might
elucidate the type of reasoning that should be used. Assume that
the frequency of the EQCMis measured four times per second. If
a sweep rate of, say, 100 mV/s is used, then the potential would
change by 25 mV during a single measurement. Considering a
typical cyclic voltammogram, it should be obvious that there are
regions where the current density could change drastically over
25 mV. Hence, the results of the EQCM taken at such a sweep
rate would represent some ill-dened average value.
*
Reducing
the sweep rate to 4 mV/s would probably be satisfactory from
this point of view, since the potential would only change by
1 mV between measurements of the resonance frequency and the
changes occurring on the surface could be considered negligible.
2. The chemistry of the surface process being investigated may be
too slow to follow the rate of change of potential with time. In
such cases, even a low sweep rate of 1 mV/s may show signicant
hysteresis, indicating that equilibrium has not been reached. A
prime example of that is adsorption-desorption processes of
pyridine on gold electrode [110,177]: the cycling voltammetry
technique used for this system shows hysteresis, which depends
on the scan rate. Reliable data could be obtained by potential
step techniques, which show that adsorption-desorption pro-
cesses may take several minutes to reach equilibrium. Another
example of how cycling voltammetry can distort the response of
the EQCMwas shown in Fig. 17 and discussed in the correspond-
ing text (Sec. V.B).
3. Great care must be exercised in the choice of the potential limits
of the CV to avoid causing permanent changes on the surface,
such as roughening, surface alloy formation, or formation of
hard-to-remove reaction products. Moreover, if two dierent
phenomena take place in the chosen potential range and if it is
possible to separate them by dividing the potential range, the
results could be interpreted more clearly. The data presented in
Fig. 23 are a good case in point: in the EQCM response, both Df
and G show no traces of hysteresis. However, if the potential
range is extended up to potentials where surface oxide formation
takes place, hysteresis appears immediately and its shape
depends strongly on the scan rate.
It may be concluded that cyclic voltammetry can be used in conjunc-
tion with the EQCM, but it will be advisable to use low sweep rates. The
* It could be argued that there are other regions in the cyclic voltammogram
where the current hardly changes, so there may not be a problem. This is not a
valid argument, since it is in the region where signicant currents ow that
electrochemistry actually happens.
Tsionsky et al. 90
best approach is to compare the data obtained by cycling voltammetry
with those obtained by potential-step or galvanostatic charging at small
currents, which seem to be more reliable. If one insists on employing cyclic
voltammetry in conjunction with the EQCM, experiments should be
repeated at several sweep rates. The results can be considered reliable only
if a further decrease of sweep rate has no eect on the variation of the
response of the EQCM with potential.
LIST OF SYMBOLS
A(K) amplitude of the nonhomogeneous shear-mode
oscillations
a k/d
C
m
mass sensitivity of the QCM, 2f
2
0
=

l
q
q
q
C
g
sensitivity of the QCM operating in contact with a liquid,
f
3=2
0
=

pl
q
q
q
c
66
l
q
e
2
26
=e
22
ixg
q
C
0
static quartz capacitance
C
a,c
concentrations of anions and cations
d thickness of the quartz crystal
e
26
piezoelectric constant
E potential
E
ox
potential of oxidation
f frequency
f
0
fundamental frequency of the resonator
g(K) correlation function for surface roughness and for non-
homogeneous mass distribution
G
eq
equivalent weight of adsorbed species
h root mean square height of a roughness
j current density
j
L
limiting current density
K
q
electromechanical coupling factor
k
q
x

q
q
=c
66
k wave vector of shear waves in quartz, x

q
q
=l
q
l correlation length of surface roughness and of non-

homogeneous mass distribution
L
f
thickness of the liquid lm
L thickness of interfacial layer
M
a
, M
c
, M
w
molecular weights of anion, cation, and water,
respectively
M
ads
molecular weight of adsorbate
M ionic or molecular weight
n number of electrons taking part in electrochemical
reaction
P(r, x) pressure in a liquid
q surface charge density
q
+
,q
positive and negative excesses of charge

q
0
ix
0
q=g
q
1
q
2
0
n
2
H
q
ads
charge passed during adsorption
R hydrodynamic roughness factor
R
electrochemical roughness factor

r =(z,R) coordinates (normal and lateral)
u(z,x) amplitude of the shear displacements of the quartz
u
f
(x) displacements of the homogeneous surface lm
v
f
(t) lm velocity
v
f 0
(x) amplitude of lm velocity
v
q
(t) velocity of the quartz surface
v
q0
(x) amplitude of the quartz surface velocity
v
a
(x) velocity of absorbed later
v
l
(x) liquid velocity at the interface
v(r,x) velocity of liquid
Y admittance
Z
out
mechanical impedance of the contacting medium
Z
m
motional quartz crystal impedance
Z
q
characteristic mechanical impedance of quartz crystal
G width of the resonance
G
l
width of the resonance for quartz crystal resonator
contacting a semi-innite liquid
&
a
, &
ads
, &
m
surface excesses and maximum surface excess of
adsorbate
d velocity decay length in a liquid, d
g=pqf
0
Dc change of surface tension

Df resonant frequency shift
Df
m
mass-induced resonant frequency shift
Df
g
viscosity-induced resonant frequency shift
Tsionsky et al. 92
Df
P
pressure-induced resonant frequency shift
Df
R
roughness-induced resonant frequency shift
Df
sl
slippage-induced resonant frequency shift
Df
T
resonant frequency shift due to change of temperature
Df
l
frequency shift for quartz resonator contacting a
semi-innite liquid
Df
0
( q) frequency shift due to electrostatic adsorption
Df
ads
frequency shift due to adsorption
Df
dl
frequency shift due to mass and viscosity changes in the
double layer
Dm surface mass density
Dm
f
surface mass density of a lm
Dm
a
average surface density of the adsorbed atoms
Dm
a,c
change of mass density due to electrostatic adsorption of
ions
Dmf (R) nonhomogeneous surface lm density
Dm
f
average surface mass density
Dm
t
root mean square deviation of the mass distribution
e
22
dielectric constant of the quartz crystal
e
ls
, e
ll
energy of liquid-substrate and liquid-liquid interactions
g viscosity of a liquid
g
q
ctitious viscosity of the quartz crystal
g
f
viscosity of the liquid lm
j
1
Debye length
k slip length
k
q
wave length of shear-mode quartz oscillations
l
q
shear modulus of the quartz crystal
v
a
,v
c
,v numbers of water molecules replaced by anion, cation,
and organic molecule, respectively
n(R) surface prole
n
H
permeability of interfacial layer
q density of a liquid
q
q
density of the quartz crystal
q
f
density of the liquid lm
q
s
density of solid
s
s
slip time
u(z,t) potential distribution in the quartz crystal
u(z,x) amplitude of the potential distribution in the quartz
crystal
u
0
potential at the surface z =0
/
q
wave phase shift in quartz, k
q
d
/ porosity of interfacial layer
v coecient of sliding friction
x angular frequency
x
0
2pf
0
ACKNOWLEDGMENTS
Financial support for this work by the Israel Science Foundation (grants
No. 472/98, 573/00 and 2/01) is gratefully acknowledged. We thank Dr.
G. Zilberman, Dr. D. Zagidulin, and Mrs. G. Katz for their contribution
in obtaining experimental data discussed in this chapter.
REFERENCES
1. Schumacher, R. Angew. Chem. Int. Engl. 1990, 29, 329.
2. Buttry, D.A. Electroanalytical Chemistry; Bard, J., Ed.; Marcel Dekker Inc.:
New York, 1991; 1 p.
3. Buttry, D.A. In: Electrochemical InterfacesModern Techniques for In Situ
Interface Characterization; Abruna, D., Ed.; VCH: New York, 1991; p. 531.
4. Buttry, D.A.; Ward, M.D. Chem. Rev. 1992, 92, 1355.
5. Hepel, M. Interfacial Electrochemistry; Wieckowski, A., Ed.; Marcel
Dekker, Inc.: New York, 1999; 599 p.
6. Sauerbrey, G. Z. Phys. 1959, 155, 206.
7. Moret, H.; Louwerix, E. In: Vacuum Microbalance Techniques; Behrndt,
K.H.N., Ed.; Plenum Press: New York, 1966; p. 59.
8. Tsionsky, V.; Gileadi, E. Langmuir 1994, 10, 2830.
9. Nomura, T.; Iijima, M. Anal. Chim. Acta 1981, 131, 97.
10. Bruckenstein, S.; Shay, M. Electrochim. Acta 1985, 30, 1295.
11. Kanazawa, K.K.; Gordon, J.G. Anal. Chim. Acta 1985, 175, 99.
12. Martin, S.J.; Frey, G.; Ricco, A.; Senturia, S. Anal. Chem. 1993, 65, 2910.
13. Tsionsky, V.; Daikhin, L.; Urbakh, M.; Gileadi, E. Langmuir 1995, 11,
674.
14. Yang, M.; Thompson, M.; Duncan-Hewitt, W.C. Langmuir 1993, 9, 802.
15. Beck, R.; Pitterman, U.; Weil, K.G. Ber. Bunsenges, Phys. Chem. 1988, 92,
1363.
16. Frubose, C.; Doblhofer, K.; Soares, D.M. Ber. Bunsenges. Phys. Chem.
1993, 97, 475.
Tsionsky et al. 94
17. Bandey, H.L.; Gonsalves, M.; Hillman, A.R.; Glidle, A.; Bruckenstein, S. J.
Electroanal. Chem. 1996, 410, 219.
18. Soares, D.M.; Tenan, M.A.; Wasle, S. Electrochim. Acta 1998, 44, 263.
19. Yamamoto, N.; Yamane, T.; Tatsuma, T.; Oyama, N. Bull. Chem. Soc. Jpn.
1998, 68, 1641.
20. Bund, A.; Schwitzgebel, G. Anal. Chem. 1998, 70, 2584.
21. Wunsche, M.; Meyer, H.; Schumacher, R. Z. Phys. Chem. 1999, 208, 225.
22. Noel, M.A.M.; Topart, P.A. Anal. Chem. 1994, 66, 484.
23. Plausinaitis, D.; Raudonis, R.; Daujotis, V. Polish. J. Chem. 2000, 74, 559.
24. Daikhin, L.; Gileadi, E.; Katz, G.; Tsionsky, V.; Urbakh, M.; Zagidulin, D.
Anal. Chem. 2002, 74, 554.
25. Tsionsky, V.; Katz, G.; Gileadi, E.; Daikhin, L. J. Electroanal. Chem. 2002,
524, 110.
26. Kanazawa, K.K. Faraday Discuss. 1997, 107, 77.
27. Lucklum, R.; Behling, C.; Cernosek, R.W.; Martin, S.J. J. Phys. D-Appl.
Phys. 1997, 30, 346.
28. Bandey, H.L.; Hillman, A.R.; Brown, M.J.; Martin, S.J. Faraday Discuss.
1997, 107, 105.
29. Martin, S.J.; Frey, G.C. IEEE Ultrasonics Symp. Proc., IEEE: New York,
1991, 393.
30. Auge, J.; Hauptmann, P.; Eichelbaum, F.; Rosler, S. Sensor Actuat. B-
Chem. 1994, 19, 518.
31. Yang, M.; Thompson, M. Anal. Chem. Acta 1993, 282, 505.
32. Muramatsu, H.; Ye, X.; Ataka, T. J. Electroanal. Chem. 1993, 347, 247.
33. Muramatsu, H.; Kimura, K. Anal. Chem. 1992, 64, 2502.
34. Yang, M.S.; Chung, F.L.; Thompson, M. Anal. Chem. 1993, 65, 3713.
35. Glidle, A.; Hillman, A.R.; Bruckenstein, S. J. Electroanal. Chem. 1991, 318,
411.
36. Ward, M.D. Physical Electrochemistry; Rubinstein, I., Ed.; Marcel Dekker
Inc.: New York, 1995; 293 p.
37. Cady, W.G. Piezoelectricity; Cady, W.G., Ed.; Dover Publ. Inc.: NewYork,
1964.
38. Muramatsu, H.; Tamiya, E.; Karube, I. Anal. Chem 1988, 60, 2142.
39. Bottom, V.G. Introduction to Quartz Crystal Unit Design; Van Nostrand
Reinhold: New York, 1982.
40. Barnes, C. Sensor Actuat. 1992, 30, 197.
41. Lin, Z.X.; Yip, C.M.; Joseph, I.S.; Ward, M.D. Anal. Chem. 1993, 65, 1546.
42. Yang, M.; Thompson, M. Anal. Chem. 1993, 65, 1158.
43. Shana, Z.A.; Josse, F. Anal. Chem. 1994, 66, 1955.
44. Okajima, T.; Sakurai, H.; Oyama, N.; Tokuda, K.; Ohsaka, T. Electrochim.
Acta 1993, 38, 747.
45. Soares, D.M. Meas. Sci. Technol. 1993, 4, 549.
46. Topart, P.A.; Noel, M.A.M.; Liess, H.D. Thin Solid Films 1994, 239, 196.
47. Wang, J.; Ward, M.D.; Ebersole, R.C.; Foss, R.P. Anal. Chem. 1993, 65,
2553.
48. Krim, J.; Solina, D.H.; Chiarello, R.P. Phys. Rev. Lett. 1991, 66, 181.
49. Krim, J.; Watts, E.T.; Digel, J. J. Vac. Sci. Technol. A 1990, 8, 3417.
50. Watts, E.T.; Krim, J.; Widom, A. Phys. Rev. B 1990, 41, 3466.
51. Krim, J.; Widom, A. Phys. Rev. B 1988, 38, 12184.
52. Krim, J.; Chiarello, R.P. J. Vac. Sci. Technol. A. 1991, 9, 2566.
53. Persson, B.N.J. Sliding Friction: Physical Priciples and Applications;
Springer: New York, 2000.
54. Daly, C.; Krim, J. Phys. Rev. Lett. 1996, 76, 803.
55. Bruschi, L.; Carlin, A.; Mistura, G. Phys. Rev. Lett. 2002, 88, no046105.
56. Feinman, R. Lectures on Physics; Addison-Wesley, Reading: MA, 1965.
57. Koplik, J.; Banavar, J.R. Annu. Rev. Fluid Mech. 1995, 27, 257.
58. Raviv, U.; Laurat, P.; Klein, J. Nature 2001, 413, 51.
59. Zhu, Y.X.; Granick, S. Phys. Rev. Lett. 2001, 87, 6104.
60. Lee, W.W.; White, H.S.; Ward, M.D. Anal. Chem. 1993, 65, 3232.
61. Tsionsky, V.; Daikhin, L.; Gileadi, E. J. Electrochem. Soc. 1996, 143, 2240.
62. Pit, R.; Hervet, H.; Leger, L. Phys. Rev. Lett. 2000, 85, 980.
63. Churaev, N.N.; Sobolev, V.D.; Somov, A.N. J. Colloid. Int. Sci. 1984, 97,
574.
64. Krim, J. Sci. Am. 1996, 275, 74.
65. Thompson, P.A.; Troian, S.M. Nature 1997, 389, 360.
66. Gupta, S.A.; Cochran, H.D.; Cummings, P.T. J. Chem. Phys. 1997, 107,
10316.
67. de Gennes, J.-P. C. R. Acad. Sci., Paris, B 1979, 288, 219.
68. Barrat, J.L.; Bocquet, L. Faraday Discuss. 1999, 112, 119.
69. Rodahl, M.; Kasemo, B. Sensor Actuat. A-Phys. 1996, 54, 448.
70. Barrat, J.L.; Bocquet, L. Phys. Rev. Lett. 1999, 82, 4671.
71. Pit, R.; Hervet, H.; Leger, L. Tribol. Lett. 1999, 7, 147.
72. Thompson, M.; Kipling, A.L.; Duncanhewitt, W.C.; Rajakovic, L.V.;
Cavicvlasak, B.A. Analyst 1991, 116, 881.
73. Kipling, A.L.; Thompson, M. Anal. Chem. 1990, 62, 1514.
74. Daikhin, L.; Gileadi, E.; Tsionsky, V.; Urbakh, M.; Zilberman, G.
Electrochim. Acta 2000, 45, 3615.
75. Bruckenstein, S.; Fensore, A.; Li, Z.; Hillman, A.R. J. Electroanal. Chem.
1994, 370, 189.
76. Schumacher, R.; Borges, G.; Kanazawa, K.K. Surf. Sci. 1985, 163, L621.
77. Schumacher, R.; Gordon, J.G.; Melory, O.J. J. Electroanal. Chem. 1987,
216, 127.
78. Yang, M.; Thompson, M. Langmuir 1993, 9, 1990.
79. Beck, R.; Pitterman, U.; Weil, K.G. J. Electroanal. Chem. 1992, 139, 453.
Tsionsky et al. 96
80. Urbakh, M.; Daikhin, L. Phys. Rev. B 1994, 49, 4886.
81. Urbakh, M.; Daikhin, L. Langmuir 1994, 10, 2836.
82. Daikhin, L.; Urbakh, M. Langmuir 1996, 12, 6354.
83. Sahimi, M. Rev. Mod. Phys. 1993, 65, 1393.
84. Jones, J.L.; Marques, C.M.; Joanny, J.F. Macromolecules 1995, 28, 136.
85. Brinkman, H.C. Appl. Sci. Res. A 1947, 1, 27.
86. Einzel, D.; Panzer, P.; Liu, M. Phys. Rev. Lett. 1990, 64, 2269.
87. Ponomarev, I.V.; Meyerovick, A. E. Phys. Rev. E 2003, 67, no026302.
88. Panzer, P.; Liu, M.; Einzel, D. Int. J. Mod. Phys. B 1992, 6, 3251.
89. Jabbarzadeh, A.; Atkinson, J.D.; Tanner, R.I. Phys. Rev. E 2000, 61, 690.
90. Jusys, Z.; Bruckenstein, S. Electrochem. Solid-State Lett. 1998, 1, 74.
91. Uchida, H.; Ikeda, N.; Watanabe, M. J. Electroanal. Chem. 1997, 424, 5.
92. Watanabe, M.; Uchida, H.; Ikeda, N. J. Electroanal. Chem. 1995, 380, 255.
93. Bruckenstein, S.; Shay, M. J. Electroanal Chem. 1985, 188, 131.
94. Kautek, W.; Sahre, M.; Soares, D.M. Ber. Bunsenges. Phys. Chem. 1995, 99,
667.
95. Santos, M.C.; Miwa, D.W.; Machado, S.A.S.; Avaca, L.A. Quimica Nova
2001, 24, 465.
96. Jerey, M.; Woods, R. J. Electrochem. Soc. 2001, 148, E79.
97. Tsionsky, V.; Daikhin, L.; Gileadi, E. J. Electrochem. Soc. 1995, 142, L233.
98. Tsionsky, V.; Daikhin, L.; Zilberman, G.; Gileadi, E. Faraday Discuss.
1997, 107, 337.
99. Jusys, Z.; Bruckenstein, S. Electrochem. Commun. 2000, 2, 412.
100. Ataka, K.; Yotsuyanagi, T.; Osawa, M. J. Phys. Chem. 1996, 100, 10664.
101. Kitamura, F.; Nanbu, N.; Ohsaka, T.; Tokuda, K. J. Electroanal. Chem.
1998, 452, 241.
102. Kitamura, F.; Ohsaka, T.; Tokuda, K. Electrochim. Acta 1997, 42, 1235.
103. Gordon, J.S.; Johnson, D.C. J. Electroanal. Chem. 1994, 365, 267.
104. EerNisse, E.P. J. Appl. Phys. 1972, 43, 1330.
105. Thorston, R.N.; Brugger, K. Phys. Rev. 1964, 133, A1604.
106. Stolberg, L.; Richer, J.; Lipkowski, J.; Irish, D.E. J. Electroanal. Chem.
1986, 207, 213.
107. Delahay, P. Double Layer and Electrode Kinetics; Interscience Publ. John
Wiley and Sons., Inc.: New York, 1996; 44 p.
108. Zilberman, G. J. Electroanal. Chem. 2001, 502, 100.
109. Kern, P.; Landolt, D. J. Electrochem. Soc. 2001, 148, B228.
110. Zilberman, G.; Tsionsky, V.; Gileadi, E. Can. J. Chem. 1997, 75, 1674.
111. Zilberman, G.; Tsionsky, V.; Gileadi, E. Electrochim. Acta 2000, 45, 3473.
112. Bockris, J.O.; Devanathan, M.A.V.; Muller, K. Proc. Roy. Soc. 1963, A274,
55.
113. Gileadi, E. Electrosorption; Plenum Press: New York, 1967.
114. Mohilner, D.M.; Theodore, R.B. J. Phys. Chem. 1979, 83, 1160.
115. Mohilner, D.M.; Nandi, N. J. Electroanal. Chem. 1975, 65, 843.
116. Holzle, M.H.; Wandlowski, Th.; Kolb, D.M. Surf. Sci. 2000, 335, 281.
117. Futamata, M.; Diesing, D. Vibrational Spectroscopy 1999, 19, 187.
118. Emov, I.O.; Heusler, K.E. J. Electroanal. Chem. 2000, 490, 1.
119. Qu, D.; Morin, M. J. Electroanal. Chem. 2001, 517, 45.
120. Lei, H.W.; Uchida, H.; Watanabe, M. J. Electroanal. Chem. 1996, 413, 131.
121. Lei, H.W.; Uchida, H.; Watanabe, M. Langmuir 1997, 13, 3523.
122. Deakin, M.R.; Buttry, D.A. Anal. Chem. 1989, 61, 1147.
123. Mao, Y.; Hwang, E.; Scherson, D. Anal. Chem. 1995, 67, 2415.
124. Stockel, W.; Schumacher, R. Ber. Bunsenges. Phys. Chem. 1989, 93, 600.
125. Deakin, M.R.; Tomi, T.L.; Melory, O.J. J. Electroanal. Chem. 1988, 243,
343.
126. Wasberg, M.; Bacskai, J.; Inzelt, G.; Horanyi, G. J. Electroanal. Chem.
1996, 418, 195.
127. Jerey, M.; Woods, R. J. Electrochem. Soc. 2001, 148, E79.
128. Uchida, H.; Ikeda, N.; Watanabe, M. J. Electroanal. Chem. 1997, 424, 5.
129. Watanabe, M.; Uchida, H.; Ikeda, N. J. Electroanal. Chem. 1995, 380, 255.
130. Visscher, W.; Gootzen, J.F.E.; Cox, A.P.; van Veen, J.A.R. Electrochim.
Acta 1998, 43, 533.
131. Birss, V.I.; Chang, M.; Segal, I. J. Electroanal. Chem. 1993, 355, 181.
132. Zolfaghari, A.; Conway, B.E.; Jerkiewicz, G. Electrochim. Acta 2002, 47,
1173.
133. Shimazu, K.; Kita, H. J. Electroanal. Chem. 1992, 341, 361.
134. Frumkin, A.N.; Petrii, O.A.; Damaskin, B.B. Comprehensive Treatise of
Electrochemistry; Bockris, J.O., Conway, B.E., Yeager, E., Eds.; Plenum
Press: New York, 1980; 221 p.
135. Weaver, M.J. Langmuir 1998, 14, 3932.
136. Gloaguen, F.; Leger, J.-M.; Lamy, C. J. Electroanal. Chem. 1999, 467, 186.
137. Santos, M.C.; Miwa, D.W.; Machado, S.A.S. Electrochem. Commun. 2000,
2, 692.
138. Vigier, F.; Gloaguen, F.; Leger, J.M.; Lamy, C. Electrochim. Acta 2001, 46,
4331.
139. Li, F.B.; Hillman, A.R.; Lubetkin, S.D.; Roberts, D.J. J. Electroanal. Chem.
1992, 335, 345.
140. Stockel, W.; Schumacher, R. Ber. Bunsenges. Phys. Chem. 1987, 91, 345.
141. Gileadi, E.; Tsionsky, V. J. Electrochem. Soc. 2000, 147, 567.
142. Tsionsky, V.; Gileadi, E. Mater. Sci. Eng. A-Struct. Mater. 2001, 302, 120.
143. Tang, Y.; Furtak, T.E. Electrochim. Acta 1991, 36 (11/12), 1873.
144. Daujotis, V.; Gaidamauskas, E. J. Electroanal. Chem. 1998, 446, 151.
145. Hepel, M.; Bruckenstein, S.; Kanige, K. J. Chem. Soc. Faraday Trans. 1993,
89, 251.
146. Hepel, M.; Kanige, K.; Bruckenstein, S. J. Electroanal. Chem. 1989, 266, 409.
Tsionsky et al. 98
147. Niece, B.K.; Gewirth, A.A. Langmuir 1997, 13, 6302.
148. Inzelt, G.; Horanyi, G. J. Electroanal. Chem. 2000, 491, 111.
149. Deakin, M.R.; Melory, O.J. J. Electroanal. Chem. 1988, 239, 321.
150. Henderson, M.J.; Bitziou, E.; Hillman, A.R.; Vieil, E. J. Electrochem. Soc.
2001, 148, E105.
151. Watanabe, M.; Uchida, H.; Miura, M.; Ikeda, N. J. Electroanal. Chem.
1995, 384, 191.
152. Hepel, M. J. Electroanal. Chem. 2001, 509, 90.
153. Niece, B.K.; Gewirth, A.A. J. Phys. Chem. B 1998, 102, 818.
154. Wilde, C.P.; Zhang, M.J. Langmuir 1994, 10, 1600.
155. Daujotis, V.; Raudonis, R. J. Electroanal. Chem. 1992, 326, 253.
156. Zhang, M.J.; Wilde, C.P. J. Electroanal. Chem. 1995, 390, 59.
157. Aramata, A.; Terui, S.; Taguchi, S.; Kawaguchi, T.; Shimazu, K. Electro-
chim. Acta 1996, 41, 761.
158. Bruckenstein, S.; Swathirajan, S. Electrochim. Acta 1985, 30, 851.
159. Lachenwitzer, A.; Magnussen, O.M. J. Phys. Chem. B 2000, 104, 7424.
160. Kuhn, A.T.; Mortimer, C.J. J. Electroanal. Chem. 1972, 34, 1.
161. Vaskevich, A.; Gileadi, E. J. Electroanal. Chem. 1998, 442, 147.
162. Beck, R.; Pitterman, U.; Weil, K.G. J. Electrochem. Soc. 1992, 139, 453.
163. Bund, A.; Schneider, M. J. Electrochem. Soc. 2002, 149, E331.
164. Urbakh, M.; Daikhin, L. Phys. Rev. B 1994, 49, 4866.
165. Bund, A.; Schwitzgebel, G. Electrochim. Acta 2000, 45, 3703.
166. Raudonis, R.; Plausinaitis, D.; Daujotis, V. J. Electroanal. Chem. 1993, 358,
351.
167. Wilde, C.P.; De Cli, S.V.; Hui, K.C.; Brett, D.J.L. Electrochim. Acta 2000,
45, 3649.
168. Wilde, C.P.; Zhang, M.J. J. Electroanal. Chem. 1992, 340, 241.
169. Daujotis, V.; Jasaitis, D.; Raudonis, R. Electrochim. Acta 1997, 42, 1337.
170. Farias, G.A.; Maradurin, A.A. Phys. Rev. B 1983, 28, 5675.
171. Jerkiewicz, G.; Vatankhah, G.; Zolfaghari, A.; Lessard, J. Electrochem.
Commun. 1999, 1, 419.
172. Kelly, J.J.; Rahman, K.M.A.; Durning, C.J.; West, A.C. J. Electrochem.
Soc. 1998, 145, 492.
173. Rodahl, M.; Hook, F.; Kasemo, B. Anal. Chem. 1996, 68, 2219.
174. Kim, M.S.; Hwang, T.S.; Kim, K.B. J. Electrochem. Soc. 1997, 144, 1537.
175. Hoogvliet, J.C.; van Bennekom, W.P. Electrochim. Acta 2001, 47, 599.
176. Lin, Z.; Ward, M.D. Anal. Chem. 1995, 67, 685.
177. Scendo, M.; Malyszko, J. Monatshefte Chem. 1997, 128, 123.
178. Nikolsky, P., Ed.; Handbook of Chemistry; Khimia: Moscow-Leningrad,
1962; 716 p.
THE INDIRECT LASER-INDUCED TEMPERATURE-
JUMP METHOD FOR CHARACTERIZING FAST
INTERFACIAL ELECTRON TRANSFER: CONCEPT,
APPLICATION, AND RESULTS
Stephen W. Feldberg, Marshall D. Newton, and John F. Smalley
Brookhaven National Laboratory,
Upton, New York, U.S.A.
I. INTRODUCTION 102
A. Why Measure Fast Interfacial Electron Transfer
Rate Constants? And How? 103
B. Background 104
C. The Underlying Principles of the ILIT
MethodThe Short Version 106
D. Denition of Terms 108
II. THE EVOLUTION OF THE ILIT METHOD
FOR THE STUDY OF FAST INTERFACIAL
ELECTRON TRANSFER KINETICS 108
A. The Temperature-Jump Approach for Studies
of Homogeneous Kinetics 108
B. The Temperature-Jump Approach for Studies
of Interfacial Kinetics 108
III. RELEVANT ELECTRON TRANSFER THEORY:
MARCUSS DESCRIPTION OF HETEROGENEOUS
NONADIABATIC ELECTRON TRANSFER
REACTIONS 112
A. Chidseys Approach 112
B. Temperature Dependence 116
Approximate the Marcus Formalism? 118
101
IV. ANALYSIS OF THE ILIT RESPONSE 120
A. Response of the Open-Circuit Electrode Potential
to a Change in the Interfacial Temperature
in the Presence of a Perfectly Reversible Redox
Couple Attached to the Electrode Surface 121
B. The Relaxation of the ILIT Response When the Rate
of Electron Transfer Is Not Innitely Fast 126
C. When Is the ILIT Response Purely Thermal
(i.e., Devoid of Kinetic Information)? 126
D. The Shape of the Ideal ILIT Perturbation 130
E. Nonidealities of the Shape of the ILIT Perturbation
and ResponseExtracting the Relaxation Rate
Constant, k
m
134
F. Correlating k
m
to Meaningful Physical Parameters 137
V. EXPERIMENTAL IMPLEMENTATION OF ILIT 143
A. The Cell 143
Diusion Properties 148
C. Preparation of Self-Assembled Monolayers 150
D. The Electronics 151
E. Potential Problems 152
F. Energetic and Timing Considerations for Single
and Multiple Pulse Experiments 156
G. Some Suggested Experimental Protocols 160
VI. A FEW EXAMPLES OF MEASUREMENTS
OF INTERFACIAL KINETICS 161
A. Some Typical Transients 161
B. Determining the Value of k
o
163
p
and DH
k
163
VII. THE POTENTIAL OF THE ILIT APPROACH 166
VIII. SOME THOUGHTS ABOUT FUTURE
EXPERIMENTS 166
IX. GLOSSARY OF TERMS 170
X. APPENDIX: ONE-DIMENSIONAL THERMAL
DIFFUSION INTO TWO DIFFERENT PHASES 173
REFERENCES 175
Feldberg et al. 102
I. INTRODUCTION
The objective of this chapter is to describe the indirect laser induced tem-
perature method (ILIT) - an approach for studying heterogeneous electron
transfer; ILIT is particularly useful for studying systems where the redox
species are attached to the electrode. Our focus will be on the fundamen-
tals of the ILIT methodology and some of the questions that might be
answered with its application.
A. Why Measure Fast Interfacial Electron Transfer Rate
Constants? And How?
A fundamental reason for measuring rates of electron transfer (ET) is to
identify the physical-chemical factors that control those rates and to learn
how to control those factors. For the past several years we have focused
our experimental eorts on the measurement of the heterogeneous elec-
tron-transfer rate constant, kj (units: s
1
), for systems where the redox
species are covalently attached to the electrode by any of a variety of
molecular tethers [13].* The indirect laser-induced temperature method,
which we developed [46], has the capability of measuring kj values as
large as
f
10
8
s
1
and has proven to be ideally suited for these types of
studies. We will show some recent development that we anticipate will
allow the measurement of kj values greater than 10
9
s
1
.
The development and application of the ILIT method is the primary
subject of this chapter. However, the driving force for the development of
the ILIT method has been our interest in the mechanisms by which elec-
trons tunnel through organic tethers, occasionally doing so over longer
distances and with greater facility than might intuitively be expected. The
gold-oligophenylenevinylene-ferrocene system is a striking example, with
kj >10
6
s
1
over distances up to 3 nm[3]. The importance of long distance
electron transfer in biological systems has been recognized for some time
(see, e.g., Ref. 7), and similar mechanisms control electron tunneling be-
tween a molecular acceptor and donor and between an electrode and a re-
dox couple [8]. With the current evolving interest in molecular electronics
and computing (see, e.g., Refs. 911), one can argue that electron transfer
*
Heterogeneous electron transfer can also refer to electron transfer between a
solid (electrode) phase and redox species in solution; the units of kj are then cm/s.
ILIT Method 103
between a metal and a redox moiety connected to the metal by a tether is a
model of electron transfer between a metal and a molecular switch
connected by a molecular wire. The relevance of such measurements to
molecular electronics is further expanded by Nitzans recent conjecture
[12] that the kinetics of a donor-bridge-acceptor system (and by implica-
tion the value of kj for a particular metal-tether-redox system) may be
related to the conductivity of metal-tether-metal systems [1319].
B. Background
As of 2002 the time resolution for the study of electrochemical interfacial
processes was approximately 6 orders of magnitude slower than the femto-
second time domain currently accessible for the study of homogeneous
reactions (see, e.g., recent works of Zewail and Barbara [20,21]). The es-
sence of picosecond and subpicosecond measurement methods is the
pump-probe method [20,21], which depends upon the time delay between
a perturbing (pump) laser pulse and an interrogating (probe) laser pulse.
*
The measurement time is dependent upon the production and control of
an ultrashort optical pulse and is no longer limited by amplier speed or by
capacitive and/or inductive distortion of electrical signals as they move
through a wire. Miller and coworkers [22] are applying the pump-probe
approach to systems involving an excited state of a redox species or of the
electrode. In the present work we will show that we can use a laser pulse as
the pump to eect an electrochemical perturbation of the ground state,
and this could be the basis of a pump-probe measurement of ground-state
electron transfer kinetics. The interrogating pulse could probe the dier-
ence in the absorption coecients of the oxidation and reduction species,
but that would not be a generally applicable approach. However, even
without an optical probe the use of a laser pump overcomes some of di-
*
Both the pump and probe pulses emanate from a single laser pulse, which is then
split, with the probe pulse often frequency shifted or polychromaticized to optimize
interrogation. The probe pulse impinges upon the target some time, Dt
delay
, after
the pump pulse has initiated the process of interest. Lengthening the path of the
probe pulse relative to the pump pulse by Dd
delay
eects Dt
delay
= Dd
delay
/c, where
c is the speed of light (3 10
10
cm/s). Thus, to eect Dt
delay
= 1 ps requires a
dierence in path length of 0.03 cm, the distance traversed by light in 1 ps.
Feldberg et al. 104
culties associated with electrical perturbations and oers the capability of
characterizing interfacial kinetics in the subnanosecond time domain.
We describe here the ILIT method, which we developed and applied
to the measurement of interfacial electron transfer rate constants greater
than 10
8
s
1
[16,23]. The ILIT method is best suited for studies of systems
where the redox species are covalently attached to the electrode, thereby
eliminating diusional eects and ensuring that the surface concentration
of redox moieties (typically
f
0.010.1 monolayers) and their distance
from the electrode surface are well controlled.
*
The resultant lm, often
formed by self-assembly from components dissolved in solution, is com-
monly referred to as self-assembled monolayer (SAM). Strictly speaking,
the appellation self-assembled monolayer implies a level of order that
may not obtain [25]. Li and Weaver [26] were the rst to use this approach
to study the distance dependence of the rate constant for electron transfer
between a metal and a covalently attached redox couple [a Co(III/II)
pentamine complex].
y
In 1991 Chidsey [29] introduced the idea of using a
two-component SAM comprising the redox species and a diluent. This
approach eliminates some of the nonidealities associated with a neat layer
comprising only redox active species and is now routinely used by
numerous workers [13,3036]. Finklea and coworkers [37] have intro-
duced a very nice modication whereby the -COOH terminal groups of an
extant single-component thiol SAM are partially reacted with [(4-amino-
methylpyridine)Ru(NH
3
)
5
]
2+
to form a mixed monolayer comprising
covalently attached ruthenium moieties with the remaining, unreacted
thiols-COOH serving as the diluent.
*
When the redox couple is present in the bulk electrolyte solution the electron
transfer reaction necessarily perturbs the concentrations of the donor and acceptor
at the outer Helmholtz plane. The diusion of the redox species becomes a factor
and modies the time dependence of DV
oc
. The details of the analysis and an
application are presented elsewhere [24]. One must also be aware that low levels of
adsorption (considerably less than one monolayer) of the redox species can
dramatically alter the response and complicate the analysis [24].
y The use of a rigid or quasi-rigid linkage between an electrode and a redox center is
analogous to the method used by Miller [27] and Closs et al. [28], who covalently
linked acceptor-donor pairs using a rigid spacer, thereby precisely controlling the
reaction distance as well as the nature of the linkage and eliminating diusion as a
controlling (and often limiting) factor.
ILIT Method 105
C. The Underlying Principles of the ILIT
MethodThe Short Version
The underlying principles of the ILIT method as applied to the study of
electrochemical systems are fundamentally the same as for any tempera-
ture jump method [3840]:
Start with a system at equilibrium, i.e., at a preset potential that
denes the equilibrium concentration of the redox species.
Rapidly induce a small (V5 K) step change in the interfacial tem-
peratureor as close to a step change as possible. Note: ILIT is a
small perturbation technique.
Observe the open-circuit potential, which changes during and after
the perturbation as the system adjusts (relaxes) to the new inter-
facial temperature.
We will show that the ILIT method eliminates some of the problems
associated with an electrical perturbation and, not surprisingly, creates
new, interesting, and challenging problems. Improved electronics devel-
oped at the time of the termination of this program, coupled with
picosecond or subpicosecond laser pulses, should, in principle, allow the
ILIT method to probe interfacial relaxations of the order of 1 ns or less
(our published work has used a slower system with a response function of
the order of
f
15 ns). Of course, really dramatic improvement in response
time will be achieved only with a pump-probe approach. Nevertheless,
even at its present stage of development, ILIT eects signicant improve-
ment in time resolution over methods using conventional electrochemical
perturbations where the time resolution is limited by solution resistance
and interfacial capacitance (see Ref. 41):
The imposition of a rapid, controlled change in an electrodes in-
terfacial potential, a requirement of many electrochemical tech-
niques (e.g., chronoamperometry and cyclic voltammetry [42]),
is limited by the systems RC time constant, aR
u
C
i
, where a
(units: cm
2
) is the electrode area, R
u
(units: ohms) is the uncom-
pensated solution resistance, and C
i
(units: F/cm
2
) is the interfacial
capacitance.
The rapid injection of charge, as required by the coulostatic tech-
nique [4345], is limited by the speed with which a very large cur-
rent can be switched on and o and by how well an amplier,
Feldberg et al. 106
designed to sense millivolt changes in the open-circuit potential,
can be protected from saturation when exposed to the large
potential accompanying charge injection.
Over the years various methods have been devised to diminish the problem
of the RC charging time: the eective value of R
u
(and therefore of aR
u
C
i
) can be decreased electronically by using positive feedback to correct the
applied potential by DE
iR
=iR
u
[42]. Decreasing the area of the electrode is
also eective. For a disk electrode iR
u
is proportional to r
o
; the transient
current, i (in contradistinction to the steady-state current), is proportional
to r
2
o
, and R
u
is proportional to 1/r
o
[46] when the reference and counter are
suciently far away from each other. The limiting challenge may be
designing an amplier that can measure the currents associated with very
small electrodes while retaining the desiredresponse time. Apossible way to
increase the measured current while retaining the advantages of the ultra-
microelectrode (ume) geometry is to construct a ume array, an approach
being pursued by Martin and coworkers [47]. Alternatively, a band
electrode of width w, while not as eective at reducing the uncompensated
resistance as the disk whose radius is w/2 [48], can have a much larger area
than the corresponding disk and therefore operate with a w/2 that is much
smaller than any disk radius that might practically be used, never mind
constructed [49]. Thus, the band can be a very eective tool for fast
transient studies [50] or steady-state experiments with an eective time
resolution of w
2
/4D. [49]. Less eective methods for reducing iR
u
place the
reference electrode as close as possible to the working electrode (the classic
Luggin capillary [42]) ora recently suggested variationplacing the
counterelectrode as close as possible to the working electrode [51]. Bard
and coworkers have used scanning electrochemical microscopy (which
employs small and very closely spaced electrodes operating at the steady
state) to measure kj >1 cm/s for bulk redox species [5254].
The charge-injection method can also be improved by decreasing the
electrode area and thereby the injection current. This reduces the problem
of saturation of the measurement amplier during injection. However, the
corresponding decrease in the electrodes capacitance decreases its ability
to drive the measurement amplier, exacerbating the diculty of achieving
a suciently fast response. Again, placing the reference and working
electrodes or counter and working electrodes as close to each other as
possible will also help.
ILIT Method 107
D. Denition of Terms
The terms used in this chapter will be dened in the text and in Sec. IX.
Please note that notation used in the present work may not be identical to
that used in previous publications.
II. THE EVOLUTION OF THE ILIT METHOD FOR
THE STUDY OF FAST INTERFACIAL ELECTRON
TRANSFER KINETICS
A. The Temperature-Jump Approach for Studies
of Homogeneous Kinetics
The temperature-jump methodology applied to the study of kinetics of
homogeneous chemical reactions took a great leap forward in the 1950s
with the work of Eigen and coworkers, who developed and exploited the
resistive heating-induced temperature-jump perturbation for the study of
homogeneous kinetics [3840]. The time resolution was
f
10
6
s. Micro-
wave heating [55,56] was hampered by ineective transfer of the microwave
energy into the mediumof interest. Eective laser heating methods evolved
in the late 1960s [5760], paralleling the evolution of laser technology, and
achieved nanosecond time resolution; Callender and coworkers have re-
ported 50 ps time resolution [61].
B. The Temperature-Jump Approach for Studies
of Interfacial Kinetics
The objective of a temperature-jump method applied to the study of
interfacial kinetics is to eect, as closely as possible, an instantaneous step
change in the interfacial temperature. The change in the interfacial
temperature will disturb the extant interfacial electronic equilibrium, and
the open circuit potential will readjust as the interface establishes a new
equilibrium at the new temperature (see Sec. IV). In this section we will
focus solely on how best to change the interfacial temperature in a manner
that will be conducive to the study of interfacial kinetics.
Electrical heating methods have been used to change the interfacial
temperature [6265], but that approach is inherently slowat best,
temperature changes occurred on a millisecond time scale [62]. Using a
laser beam to eect a change in the interfacial temperature is an obvious
Feldberg et al. 108
way to eect a rapid change in the interfacial temperature. There are two
ways to do this: (1) direct irradiative heating of the electrode at the
electrode-solution interface, and (2) indirect irradiative heating, whereby
the nonsolution (back) side of a thin metal lm electrode is initially
heatedthe heat then diuses to the electrode/solution interface.
1. The Direct Laser-Induced Temperature-Jump Method
One way to eect a rapid change in the interfacial temperature is by
impinging a pulsed laser beam directly onto the electrode [6671]; we will
refer to this method as the direct laser-induced temperature-jump method
(DLIT). DLIT has several fundamental problems:
The irradiation passes through the solution phase, possibly being
absorbed by the solution and/or inducing complicating photo-
chemistry in the vicinity of the electrode.
Electrons may be photoemitted into the solution, eecting additional
complicating chemical processes. The probability that this will
occur depends upon the metal and increases dramatically with the
energy of the of the impinging light. However, Barker et al. [72]
noted that even for k
hv
as short as 600 nm, the heating eect would
dominate and photo-ejection of electrons would be inconsequen-
tial.
Local overheating of the interfacial region: the nonreected energy of
a laser pulse (fullwidth-half-max (fwhm) <
f
10
12
s)
*
impinging on
a gold surface is absorbed within
f
10
6
cm and the hot electrons
thermalize in
f
10
12
s [73]. If we assume that the energy, q
T
, is
deposited instantaneously as a spatial delta function located at the
electrode surface, the temperature at the surface (assuming an
aqueous solution) is (see Appendix, Sec. X):
DT
t

q
T
1 r
T
H
2
O=Au
q
Au
C
T
Au
pD
T
Au
t
1
*
Suciently energetic laser pulses with fwhm<
f
10
12
s are available.
Brookhaven National Laboratorys Laser Electron Accelerator Facility (LEAF)
uses a titanium-sapphire laser k
hv
= 798 nm; fwhm =
f
100 fs; and energy per
pulse is 0.01 J.
ILIT Method 109
where
r
T
H
2
O=Au

q
H
2
O
C
T
H
2
O
D
T
H
2
O
q
Au
C
T
Au
D
T
Au
2
q
Au
and q
H
2
O
(units: g/cm
3
) are the densities of gold and water, C
Au
T
and
C
H
2O
T
(units: J g
1
K
1
) are the heat capacities of gold and water, D
T
Au
and
D
T
H
2
O
(units: cm
2
/s) are the thermal diffusion coefficients for gold and
water, and t is the time after energy deposition. From the data in Table 1,
we can deduce that r
T
H
2
O=Au
= 0.056, small enough to be ignored for the
purposes of this discussion. Equation (1) then simplifies to
DT
t

q
T
q
Au
C
T
Au
pD
T
Au
t
3
Equation (3) exaggerates DT
t
at very short times since the thermal energy is
initially deposited in a region of nite thickness, l (c10
6
cm) [73]we
still assume that the energy is instantaneously deposited, a reasonable
approximation even if we allow 10
12
s for thermalization; thus, the
interfacial temperature immediately after thermalization, DT
t=0
, will be
DT
t0

q
T
lq
Au
C
T
Au
4
Combining Eqs. (1) and (4) gives
DT
t
DT
t0
pD
T
Au
t
5
Equation (5) is valid when t >> l
2
/D
T
Au
, a condition easily met after
f
10
11
s. If the goal is to eect DT
t
= 2 K after 10
9
s, then, with D
T
Au
=
1.26 cm
2
/s and l = 10
6
cm, we deduce from Eq. (5) that DT
t=0
f
125 K,
TABLE 1
Summary of Parameter Values for Quartz, Gold (and Titanium), and Water
at 25jC
Phase D
T
j
(cm
2
/s) C
T
j
(Jg
1
K
1
) q
j
(g/cm
3
) K
T
j
(J cm
1
s
1
K
1
)
#1 Quartz 0.0093 0.725 2.20 0.0148
#2 Gold 1.26 0.129 19.32 3.14
(titanium) (0.093) (0.523) (4.50) (0.22)
#3 Solution 0.001435 4.1796 0.997 0.00598
Feldberg et al. 110
a large change in the interfacial temperature that might irreversibly
damage the interface.
From the analysis just completed we note that if DT
t
= 2 K after
10
9
s, then DT
t
= 1 K after 4 10
9
s and continues to decrease as 1/
t
p
.
Such a rapid change in DT
t
will dominate the control of the interfacial
potential and will mask the relatively subtler eects produced by the
interfacial relaxations of interest.
Changes in the character of the electrode/solution interface as a
function of voltage and time can modify the eciency of energy absorption
and thereby eect a change in DT even if the impinging laser energy is
perfectly constant.
2. The ILIT Method
Many of the shortcomings of the DLIT method can be eliminated or
greatly diminished by the ILIT method. In ILIT the laser pulse impinges
upon the back (nonsolution) side of a thin lm metal electrode. The ab-
sorbed laser energy is thermalized in
f
10
12
s [73] (see Sec. II.B.1. above),
the heat diuses through the electrode, ultimately eecting a change in the
temperature of the electrode/solution interface in a time that depends in-
versely on the square of the thickness of the electrode. The basic concept
of ILIT was presaged by the work of Miller [74], who used a 1.5 W contin-
uous wave laser beamto thermally modulate an electrode interface, but the
time resolution was tens of milliseconds and limited by the lowpower of the
laser beam. Later, Coufal [75,76] achieved nanosecond time resolution
using the same basic idea, but these were non-electrochemical investiga-
tions. They nevertheless demonstrated that the ILIT concept could work.
In our ILIT application, a thin gold lm (e.g., 10
4
cm) is deposited
on a quartz substrate with a thinner underlayer (
f
5 10
6
cm) of
titanium. The titanium underlayer serves three purposes:
1. It acts as a glue, stabilizing the gold lm on the quartz
substrate.
2. It absorbs the 1064 or 798 nm laser pulse much more eectively
than gold, which is a superb reector at those wavelengths.
3. It eects
f
50 50 nm domains of coplanar 111 gold facets.
Some obvious advantages of ILIT over the DLIT approach are:
No photolysis of the solution.
No photoelectrons are ejected into the solution.
No overheating, since DT
t
exhibits a rapid rise to the desired value
and then decays slowly, thus approximating a step change. The rise
ILIT Method 111
time and decay time can be tuned by changing the thickness of the
electrode (see Sec. IV).
Changes in the nature of the metal/solution interface as a function of
potential and time will have no eect on the eciency of light
absorption at the (quartz) substrate/metal interface.
III. RELEVANT ELECTRON TRANSFER THEORY:
MARCUSS DESCRIPTION OF HETEROGENEOUS
NONADIABATIC ELECTRON TRANSFER REACTIONS
Throughout this analysis we will assume that the entire potential across the
SAM is sensed by the redox moieties. Because of the relatively low
capacitance of most SAMs ( <10 AF/cm
2
) and the high concentration of
supporting electrolyte, the potential drop through the diuse layer will be
ignored. We will not consider cases where the redox moiety is buried (see
Ref. 77) in the diluent and subject to the unusual eects described by Smith
and White [78], where the attached redox moiety only senses a fraction of
the applied potential.
A. Chidseys Approach
Beginning in the 1950s, Marcus developed and rened a classical theoret-
ical description of redox reactions occurring in homogeneous solution [79
84]. He also extended the treatment to include electrode processes [8186].
Related treatments are the quantum models of Levich and Doganadze and
the works of Gerisher. More recently, Chidsey [29] presented a derivation
of Marcus nonadiabatic (MNA) behavior, characterized by the quadratic
free energy relationship for electrode processes:
e* k*FE*
4k*
6
where k*, E*, and e*, the dimensionless reorganization energy, dimension-
less electrode potential, and dimensionless electronic state energy (in the
metal), are dened by:
k*
k
k
B
T
7
E*
e
k
B
T
E E
oV

8
where E
oV
is the formal potential of the redox couple and
e*
e e
F
k
B
T
9
Feldberg et al. 112
where q
F
is the Fermi level of electrons in the metal. Two key assumptions
are that:
1. The density of states, q (units: eV
1
), in the metal electrode is
independent of the energy of the ith state, q
i
q
F
. The value of q
depends on assumptions about just how many metal electrode
atoms are coupled to a single redox moiety (see Refs. 29, 86, and
88 for a detailed discussion of q).
2. The density of states is sucient to be considered as a continuum.
We shall refer to these as the density-of-state assumptions. In the present
treatment we consider only the one-electron transfer reaction:
ox e V
k
MNA
ox
k
MNA
red
red 10
where k
MNA
ox
and k
MNA
red
are the rate constants for oxidation and reduction
and the superscript MNA denotes that they are evaluated using Mar-
cuss nonadiabatic formalism. Following the approach suggested by Chid-
sey [29], integrating the probabilities of electron transfer over all energy
states gives:
k
MNA
red

P
4pk*
p exp
E*
2

k*
4

l
l
exp
e* E*
2
4k*

2 cosh
1
=
2
e*
de*;
k
MNA
ox

P
4pk*
p exp
E*
2

k*
4

l
l
exp
e* E*
2
4k*

2 cosh
1
=
2
e*
de*
11
where P for nonadiabatic transfer is dened by
P
2pqH
j
2
DA
expbx x
0
t
12
and x is the distance between the electrode and the redox moiety, x
o
is a
convenient reference distance, b (units: cm
1
) is a constant characterizing
the exponential distance dependence of the electronic coupling [28,89,90],
H
B
DA
(units: eV) is the electronic coupling matrix element when x = x
o
and
t (units: J s) is h/2k where h is Plancks constant. Equations (11) are
mathematically identical to those derived by Chidsey [29], albeit in a
somewhat dierent form suggested by Liu [91], which makes some of
the properties of the equations more apparent. The expression for the
standard rate constant, kj, is obtained from Eqs. (11) when E* = 0:
ILIT Method 113
kj
P
4pk
p exp
k*
4

l
l
exp
e*
2
4k*

2 cosh
1
=
2
e*
de*
13
It is useful to express Eqs. (11) in terms of the dimensionless parameters
k
MNA
red
=kj and k
MNA
ox
=kj:
k
MNA
red
k
o

exp
E*
2

l
l
exp
e* E*
2
4k*

2 cosh
1
=
2
e*
de*
l
l
exp
e*
2
4k*

2 cosh
1
=
2
e*
de*
k
MNA
ox
k
o

exp
E*
2

l
l
exp
e* E*
2
4k*

2 cosh
1
=
2
e*
de*
l
l
exp
e*
2
4k*

2 cosh
1
=
2
e*
de*
14
A simple empirical approximation of the integral in Eq. (13) evolves from
the known Gaussian integral [the limiting case for Eq. (15) when k* Z 0]
and is accurate to
f
0.4% relative error for 0.01 Vk* V100, is (see Fig. 1)
l
l
exp
e*
2
4k*

2 cosh
1
=
2
e*
de*
p
1
p
k*
1 0:2256
p
k*
1
p
k*

2
15
Equation (15) can be simplied to:
l
l
exp
e*
2
4k*

2 cosh
1
=
2
e*
de*
p
1
p
k*
16
Feldberg et al. 114
Equation (16) is not as good an approximation as Eq. (15), but it is
adequate for most purposes (see Fig. 1). No comparably convenient
approximation of the full integral in [Eqs. (11) and (14)] has been produced
(see Ref. 92). Several general features of these equations are notable:
1. The value of the integral is independent of the sign of E*.
2. The ratio k
MNA
ox
=k
MNA
red
is dened by the Nernst relationship, i.e.:
k
MNA
ox
k
MNA
red
expE* 17
3. When E*/k* <
f
1.5 or E*/k* >
f
1.5, then k
MNA
red
/P = 1.00
or k
MNA
ox
/P = 1.00 [see Eq. (11)] and therefore become in-
dependent of the values of E* or k* and therefore independent of
temperature. When the limiting conditions are not met, the val-
ues of k
MNA
red
=Pand k
MNA
ox
=Pare temperature dependent (details of
that behavior will be explored in the next section). This limiting
condition is a direct consequence of the aforementioned density-
of-state assumptions, i.e., that the density of states is independent
FIG. 1. Integral for calculation of kj [Eq. (13)]: () exact numerical
integration lhs of Eq. (15) or (16); (ooo): best approximation, Eq. (15); (...):
poorer approximation, Eq. (16).
ILIT Method 115
of energy and that the density of states is large enough to justify
a continuum model [29].
Recall that the rate of electron transfer between a molecular acceptor
and donor exhibits a maximumwhen k =DGj (see e.g., Refs. 28,89). One
might argue that if the spacing of the energy states were larger, as might be
surmised for estimates of q
f
0.3/eV for gold [88], one would expect to see
a maximum in plots of ln[k
red
] and ln[k
ox
] vs. E E
o/
when ejE E
o/
j ck.
The relationship between experimentally measured values of ln[k
red
] and
ln[k
ox
] vs. E E
o/
for ferrocene (attached to gold by aliphatic ligands) was
shown by Chidsey [29] to be fully consistent with the continuum assump-
tion and Eqs. (11). The value of k (
f
0.85 eV) determined from chro-
noamperometric experiments at overpotentials as large as F1.0 V is
consistent with the value of k deduced from the temperature dependence
of kj, a correlation that requires the implicit assumption that the entropic
component of k is zero (see discussion in Sec. III.B). The plots of ln[k
red
]
and ln[k
ox
] vs. E E
o/
reported by Forster et al. [93] for an adsorbed
monolayer of [Os(OMe-bpy)2(p3p)Cl]
+
, where OMe-bpy is 4,4V-dime-
thoxy-2,2V-bipyidyl and p3p is 4,4V-trimethylenedipyridine, exhibit a very
marked attening ejE E
o/
j > 2k (easily attained since k = 0.28 eV), a
result that is also consistent with the density-of-state assumptions (Sec.
III.A). Forster also measured the temperature dependence of kj and
observed a signicant entropic contribution.
B. Temperature Dependence
As Chidsey neatly demonstrated, the dependence of k
MNA
red
and k
MNA
ox
on
EE
o/
permits direct evaluation of the reorganization energy, k [29]. Such
measurements require a large potential perturbation, i.e., e(EE
o
) = k,
which is not possible using ILIT. However, the enthalpy of activation,
DH
p
, can be evaluated from the slope of the Arrhenius plot of ln k
o
=
T
p
vs. 1/T. We shall show that DH
p
is directly and simply related to the
enthalpy of reorganization, DH
k
, which often, though not necessarily, can
be equated to k. The temperature dependence of kj for MNA electron
transfer is easily deduced by combining Eqs. (7), (13), and (16) (the simpler
approximation is adequate):
k
o
4pk*
p
p
1
p
k*
exp
k
4k
B
T

P
2
1
k*
p
exp
k*
4

18
Feldberg et al. 116
We divide through by
T
p
, which partially nullies the temperature
dependence of k* [see Eq (7)] to give:
k
o
T
p
P
2
T
p

1
k*
p
exp
k*
4

19
The reorganization energy, k, can be expressed as the sum of enthalpic and
entropic components:
k DH
k
TDS
k
20
Therefore (with Eq. 7),
k*
k
k
B
T

DH
k
k
B
T

DS
k
k
B
21
and
dk*
d
1
T
DH
k
k
B
22
Taking the natural logarithm of Eq. (19) and then the derivative with
respect to 1/T gives:
dln
k
o
T
p

d
1
T

DH
p
k
B
T
2

1
2p k*
dk*
d
1
T
1
4
dk*
d
1
T
T
2

DH
k
k
B
1
2p k*
1
4

23
and
DH
k

4DH
p
2k
B
T
1
2
p k*
24
where the value of T is midrange in the rather narrow range of temper-
ature that can be explored in most systems. In the experimental systems
ILIT Method 117
we have studied involving ferrocene and ruthenium redox moieties k c
DH
k
f
1 eV, and Eq. (24) may be written as
DH
k
c4DH
p
25
However, when k is small, the full expression, Eq. (24), should be used to
evaluate DH
k
(Forster et al. [93] have reported that k = 0.28 eV for
adsorbed [Os(OMe-bpy)
2
(p3p)Cl]
+
where OMe-bpy is 4,4V-dimethoxy-
2,2V-bipyidyl and p3p is 4,4V-trimethilenedipyridine).
Approximate the Marcus Formalism?
The voltage dependence of MNA electron transfer described by Eqs. (14)
is measurably dierent from that described by the Butler-Volmer expres-
sion [42], where:
k
BV
red
k
o
exp aE* ;
k
BV
ox
k
o
exp 1 aE* 26
where the transfer coecient, a, is generally
f
1/2, a value that comes
directly from Eqs. (14), where it is inherent in the term exp[FE
o/
*/2]. From
Eqs. (14) and (26), we obtain (assuming a = 1/2):
k
MNA
red
k
BV
red
k
MNA
ox
k
BV
ox
l
l
exp
e* E*
2
4k*

2 cosh
1
=
2
e*
de*
l
l
exp
e*
2
4k*

2 cosh
1
=
2
e*
de*
gk; E* 27
The ratios k
MNA
red
=k
BV
red
and k
MNA
ox
=k
BV
ox
are identical. For k*=10, 20, 40, and
80 over the range 1.5k*
0:4445
VE*
i
V1.5 k*
0:4445
, a plot of ln k
MNA
red
=k
BV
red

or
ln k
MNA
ox
=k
BV
ox

(= g(k*,E*) [computed from Eq. (27)] vs. ln[x] (= E*) is
shown as the solid lines in Fig. 2. The relationship is parabolic and is t
very well over this range by
ln gk*; E* ln
k
MNA
red
k
BV
red

ln
k
MNA
ox
k
BV
ox

0:1459
E
*
2
k
*
0:889
28
Feldberg et al. 118
as shown in Fig. 2. The parameters 0.1459 and 0.889 were obtained by
least-squares optimization. From Eqs. (27) and (28) we can construct a
modied version of the Butler-Volmer expression, which incorporates a
transfer coecient, a
MBV
, that is dependent upon E* and k*:
k
MNA
red
k
o
exp a
MBV
E*

;
k
MBV
ox
k
o
exp 1 a
MBV
E*

29
where, again assuming that when a
MBV
Z 1/2 as E* Z 0:
a
MBV
1
2
0:1459
E*
k
*
0:889
30
These relationships allow us to estimate the error in the rate constants
that obtain when using the simple Butler-Volmer expression (and assum-
ing that the MNA expression is correct): at any given potential the rate
constants deduced using the Butler-Volmer expression will be larger than
the MNA values by a factor of exp[0.1459 E*
2
/k*
0.889
]. The general ex-
pression for deducing the range of potential within which the Butler-
FIG. 2. (): Plot of g(x,E*) [Eq. (27)] computed for k*=10 (.), 20 (o), 40 (E),
and 80 (4) as a function of E*; points are computed from Eq. (28).
ILIT Method 119
Volmer rate constants will be within a given specied error, d (percent) for
a given value of k is
k
B
T
e
k
k
B
T

0:889
ln
d
100
1

0:1459
V E E
o=
V
k
B
T
e
k
k
B
T

0:889
ln
d
100
1

0:1459
31
or, assuming ln[1 + d/100] c d/100:
2:62
k
B
T
e
k
k
B
T

0:4445

d
100
V E E
o=
V 2:62
k
B
T
e
k
k
B
T

0:4445

d
100
32
As an example: for k = 1 eV, T = 298.2 K, and d = 20%, we deduce that
the rate constants using the simple Butler-Volmer expression will be within
20% of the MNA values as long as:
0:153 V V E E
o=
V 0:153 V 33
IV. ANALYSIS OF THE ILIT RESPONSE
In this section we will analyze the ILIT response, i.e., just how the open-
circuit potential depends upon temperature and upon the properties of the
redox couple that is attached to the electrode surface. In Sec. IV.A we will
focus on the ILIT response to a perfectly reversible electron-transfer. In
Sec. IV.B we will essentially state the kinetic response when the change in
the interfacial temperature is described by a step-function. In Secs. Cand D
we analyze the details of the actual change in the interfacial temperature,
the instrument response function, and nally the extraction of the physical
parameters from the experimental ILIT response.
Feldberg et al. 120
A. Response of the Open-Circuit Electrode Potential to a
Change in the Interfacial Temperature in the Presence
of a Perfectly Reversible Redox Couple Attached
to the Electrode Surface
We have spent some time discussing how to change the interfacial tem-
perature. What do we actually measure? We measure the open-circuit
potential, thereby passing virtually no external current and circumvent-
ing the limitations imposed by solution resistance and/or by the RC time
constant produced by the solution resistance in series with the double
layer capacitance of the electrode. In this section we will discuss just how
the open-circuit potential depends upon the interfacial temperature and
upon the interfacial electron-transfer kinetics.
The open-circuit ILIT response is analogous to the open-circuit re-
sponse for the coulostatic method [4345] where the initial perturbation is
eected by charge injection. We will conne our analysis to systems where
the redox species are attached to the electrode surface.
At a given temperature the open-circuit potential, V
oc
(unit: V), of a
gold (or any metal) electrode measured versus a given reference electrode
will be
V
oc

r
M
C
i
E
pzc
V
TJ
34
where C
i
(units: F/cm
2
) is the interfacial capacitance, r
M
(units: C/cm
2
)
is the charge density on the electrode side of the C
i
, E
pzc
(units: V) is the
potential of zero charge (pzc), i.e., the potential where j
M
=0 (E
pzc
is mea-
sured vs. the same given reference), and V
TJ
(units: V) is the sum of all
potentials caused by thermal junctions in the system comprising the
complete circuit including the Soret potential produced by the dierence
in temperature between the electrode surface and bulk solution (see Ref. 5
and Sec. V.E). The temperature of the reference electrode does not change
in the ILIT experiment. Prior to the ILIT perturbation, the entire system is
isothermal and V
i
TJ
, the initial value of V
TJ
, is zero. We will assume that the
potential associated with the diuse double layer, which may lie outside the
plane of the attached redox species [78], will be negligibly small for two
reasons: rst, the ILIT experiments are generally carried out with a high
concentration (0.11 M) of supporting electrolyte; and second, the inter-
facial capacitance, C
i
, associated with a self-assembled monolayer (our
ILIT Method 121
usual experimental system) will be small compared to the diuse layer
capacitance. The temperature dependence of V
oc
is
dV
oc
dT

r
M
C
2
i
dC
i
dT

1
C
i
dr
M
dT

dE
pzc
dT

dV
TJ
dT

r
M
C
i
dlnC
i
dT

1
C
i
dr
M
dT

dE
pzc
dT

dV
TJ
dT
35
The term dr
M
/dT reects the charge transfer across the interfacial capaci-
tance that is produced by the change in the interfacial temperature; this
can occur only if there is a redox species attached to the electrode;
*
we will
deduce its value shortly. In the absence of charge transfer, dr
M
/dT=0.
dV
oc
dT
E
i
pzc
V
i
oc

dlnC
i
dT

1
C
i
i
dr
M
dT

dE
pzc
dT

dV
TJ
dT
36
where the superscript i denotes that the parameter value is its initial
value just prior to the ILIT perturbation when the system is isothermal.
Because the system is initially isothermal, V
i
TJ
0.
Approximating parameter values with their initial values is valid as
long as the temperature changes are small; for an ILIT experiment, the
change in temperature is typically less than
f
5 K.
We assume that C
i
, E
pzc
, and V
TJ
respond instantaneously to the
change in interfacial temperature. It is only r
M
that might not respond
instantaneously. For a reversible (i.e., innitely fast) one-electron trans-
fer,
y
we can write
V
oc
E
redox
V
TJ
E
o=
V
TJ
k
B
T
e
ln
C
ox
C
red

37
where E
o/
is the formal potential of the redox couple and C
ox
and C
red
(molecules/cm
2
) are the surface concentrations of the oxidized and reduced
forms of the surface-attached redox moieties. We assume that the redox
moieties sense the full potential across the interfacial capacitance (the
*
We assume that there are no redox species in the bulk solutionthese could also
eect charge transfer across the interfacial capacitance.
y
Our ultimate interest is in the kinetics of electron transfer, where we will assume
a one-electron process.
Feldberg et al. 122
potential across the diuse layer on the solution side of the surface attached
redox species is assumed to be ignorably small [78]). Then
dV
oc
dT

dEj
dT

dV
TJ
dT

k
B
e
ln
C
ox
C
red

k
B
T
e
1
C
ox
dC
ox
dT

1
C
red
dC
red
dT

38
The total amount of redox species on the surface is xed, i.e.,
C
total
C
ox
C
red
39
and therefore
dC
ox
dT

dC
red
dT
40
For an open-circuit condition
dr
M
dT
e
dC
red
dT
41
Then Eqs. (37)(41) combine to give:
dV
oc
dT

dE
o
dT

dV
TJ
dT

V
i
oc
E
o;i
T
i

k
B
T
i
e
2
1
C
i
ox
1
C
i
red

dr
M
dT
42
To simplify subsequent algebra, we rewrite Eq. (42) as
dV
oc
dT
S Q
dr
M
dT
43
where
S
dEj
dT

dV
TJ
dT

V
i
oc
E
o;i
T
i
44
and
Q
k
B
T
i
e
2
1
C
i
ox
1
C
i
red

45
ILIT Method 123
We can express Eq. (36) as:
dV
oc
dT
A B 46
where
A E
i
pzc
V
i
oc

dlnC
i
dT

dE
pzc
dT

dV
TJ
dT
47
and
B
1
C
i
i
dr
M
dT
48
where A is the amplitude of the non-redox-related voltage response and B
is the amplitude of the redox-related voltage response. Combining Eqs.
(43) and (46) gives:
dV
oc
dT
S QC
i
i
dV
oc
dT
A

49
or
dV
oc
dT

S QC
i
i
A
1 QC
i
i
A B 50
Then
B
S QC
i
i
A
1 QC
i
i
A
S A
1 QC
i
i
dEj
dT

V
i
oc
E
o;i
T
i
E
i
pzc
V
i
oc

dlnC
i
dT

dE
pzc
dT
1
k
B
T
i
C
i
i
e
2
1
C
i
ox
1
C
i
red

51
From Eq. (37), remembering that the initial value of V
TJ
is zero, we can
write the Nernstian relationship:
x
i
C
i
ox
C
i
red
exp V
i
oc
E
o=i

52
Feldberg et al. 124
Combining Eqs. (39) and (52) gives:
C
i
ox

x
i
C
total
1 x
i
53
and
C
i
red

C
total
1 x
i
54
Combining Eqs. (53), (54), and (51) gives:
B
dE
o
dT

V
i
oc
E
o=i
T
i
E
i
pzc
V
i
oc

dlnC
i
dT

dE
pzc
dT
1
k
B
T
i
C
i
i
e
2
C
total
1 x
i
2
x
i
55
or
B
dE
o
dT

V
i
oc
E
o=i
T
i
E
i
pzc
V
i
oc

dlnC
i
dT

dE
pzc
dT
1 1 x
i
2
c
i
x
i
56
where c
i
, the dimensionless surface coverage, is dened by:
c
i
e
2
C
total
k
B
T
i
C
i
i
57
These somewhat laborious algebraic manipulations show that when a
perfectly reversible redox couple is attached to the electrode surface, the
temperature-induced change in the open-circuit potential, DV
oc
, is simply
described by
DV
oc
DTA B 58
This comes directly from Eq. (46) if we assume that DT is small enough so
that we can validly equate dV
oc
/dT and DV
oc
/DT.
ILIT Method 125
B. The Relaxation of the ILIT Response When the Rate
of Electron Transfer is Not Innitely Fast
Thus far we have assumed inntely fast electron transfer kinetics where
the redox species are always at equilibrium with the interfacial poten-
tial at the extant interfacial temperature. When the electron transfer ki-
netics are slow, the ILIT response will exhibit a relaxation whose ampli-
tude is B [Eq. (55)] and characterized by a rst-order rate constant, k
m
.
If the interfacial temperature change were a step function, we could sim-
ply write:
DV
oc
DT
step
A B1 expk
m
t 59
However, the interfacial temperature change is not a simple step func-
tion, and extracting the value of k
m
from the ILIT response requires a bit
more eort. In Sec. IV.D we will describe the shape of the ideal ILIT
perturbation where we assume that the laser pulse is a temporal delta
function and that the instrument responses are perfect. In Sec. IV.E we
will describe just how the instrument response time is introduced and
how we extract the value of k
m
from the experimental ILIT response.
Finally, in Sec. IV.F we will relate the value of k
m
to the standard rate
constant for the electron transfer, kj.
C. When Is the ILIT Response Purely Thermal (i.e., Devoid
of Kinetic Information)?
1. When the Rate of Electron Transfer Is Innitely Fast
We have already discussed the nature of the ILIT response when the
attached redox couple is perfectly reversible. Under those conditions the
response will be purely thermal, as dened by Eq. (58). Substituting into
Eq. (58) for A [from Eq. (47)] and B [from Eq. (56)] gives:
DV
oc
DT E
i
pzc
V
i
oc

dlnC
i
dT

dE
pzc
dT

dV
TJ
dT
dE
o
dT

V
i
oc
E
o=i
T
i
E
i
pzc
V
i
oc

dlnC
i
dT

dE
pzc
dT
1
1 x
i
2
c
i
x
i
60
Feldberg et al. 126
When (1 + x
i
)
2
/c
i
x
i
<< 1 (requiring c
i
>> x
i
), Eq. (60) simplies to:
DV
oc
DT
dV
TJ
dT

dE
o=
dT

V
i
oc
E
o=i
T
i

61
In other words, when c
i
is suciently large, the ratio C
ox
/C
red
is virtually
constant and the change in open-circuit potential, DV
oc
, is produced by
the thermally induced change in the junction potentials, DT(dV
TJ
/dT),
and by the thermally induced change in the equilibrium redox potential,
DT(dE
o/
/dT+(V
oc
i
E
o/i
)/T). When the condition (1 +x
i
)
2
/c
i
x
i
<<1 is
not met, the ILIT-induced charge-transfer will eect a change in the ratio
C
ox
/C
red
. We will examine this in greater detail in Sec. IV.F.
2. When the Rate of Electron Transfer Is Innitely Slow
When the rate of electron transfer is innitely slow on the time scale of
the ILIT perturbation, the result is eectively the same as when B = 0:
the ILIT response is described by [see Eq. (47)]:
DV
oc
DTA DT E
i
pzc
V
i
oc

dlnC
i
dT

dE
pzc
dT

dV
TJ
dT

62
3. When B = 0
There are also several experimental conditions that can eect a value of
B that is zero or virtually zero:
1. B = 0 when C
total
= 0
2. B= 0 when the numerator of Eq. (56) just happens (by bad luck)
to be zero, i.e.:
0
dE
o
dT

V
i
oc
E
o=i
T
i
E
i
pzc
V
i
oc

dlnC
i
dT

dE
pzc
dT
63
If this condition is met there will be no visible response to electron
transfer and therefore no measurable amplitude associated with
electron transfer kinetics. Quite simply it means that the exper-
imental conditions are such that the interfacial redox equilib-
rium is not disturbed by a change in the interfacial temperature;
the system was in equilibrium at T
i
and will be in equilibrium at
T
i
+ DT.
3. B Z 0 as (1 + x
i
)
2
/c
i
x
i
>> 1 [see Eqs. (56) and (57)]. Note,
however, that when (1 + x
i
)
2
/c
i
x
i
<< 1, the value of B is inde-
ILIT Method 127
pendent of the value of C
total
[23]. We will discuss this in more
detail in Sec. IV.E., which describes the kinetic dependence of
the ILIT response. When B is eectively zero, the ILIT response
depends only on the value of A [see Eq. (58)]. From Eq. (62) we
see that this predicts that a plot of DV
oc
/DT vs. V
i
oc
will be lin-
ear: the slope is dln[C
i
]/dT and the intercept is E
i
pzc
dln[C
i
]/
dT + dE
pzc
/dT + dV
TJ
/dT. In Fig. 3 we show such a plot of the
ILIT response, DV
oc
/DT, for two dierent SAMs, one formed
from HS(CH
2
)
10
CH
3
and the other from HS(CH
2
)
14
OH. When
DV
oc
/DT = 0 we specify the corresponding value of V
i
oc
as the
potential of zero response (pzr), or V
pzr
. The dierence in V
pzr
FIG. 3. ILIT responses for two dierent SAMs: one formed from HS(CH
2
)
10
CH
3
and the other formed from HS(CH
2
)
14
OH.
Feldberg et al. 128
for these two SAMs is nearly two volts. In Fig. 4 we show that
the value of V
pzr
is moderately insensitive to the length of SAMs
formed from methyl-terminated thiolswe assume that the pzr
values for a dierent lengths of hydroxy-terminated thiols are
similarly insensitive to the length of the alkyl chain. What are
the physical implications of this? When V
i
oc
= V
pzr
, Eq. (47)
becomes:
A
dV
oc
dT
0 E
i
pzc
V
pzr

dlnC
i
dT

dE
pzc
dT

dV
TJ
dT
64
If the same supporting electrolyte is used for all experiments
the value of dV
TJ
/dT cannot be aected by the change in the
FIG. 4. ILIT responses for a variety of SAMs comprising methyl terminated
aliphatic thiols of dierent specied lengths.
ILIT Method 129
thiol terminal group. Thus, the change in the V
pzr
, DV
pzr
, will
be expressed by:
DV
pzr
V
pzr
OH
V
pzr
CH
3
E
i
pzc
OH
E
i
pzc
CH
3
dE
pzc
dT

OH
dlnC
i
dT

OH
dE
pzc
dT

CH
3
d lnC
i
dT

CH
3
65
The values of E
i
pzc
OH
and E
i
pzc
CH
3
and their corresponding
derivatives will be functions of the dipole potentials associated
with the dierently terminated monolayers. Implicit in this ar-
gument is the assumption that the component of the total di-
pole associated with the gold-sulfur bond [34] is identical for the
-OH- and -CH
3
-terminated dipoles; however, it is also possible
that the surface concentrations of the two layers may be not be
the same and the contributions of the gold-sulfur dipoles may
not cancel.
Because our primary interest has been the evaluation of the elec-
tron transfer kinetics, we have not made a thorough study of the factors
that control V
pzr
. Nevertheless, even an approximate knowledge of the
pzr values of dierent diluents can indicate which diluent will maximize
the value of B [Eq. (51)]. Specic interactions between the diluent and the
redox species also play a role.
D. The Shape of the Ideal ILIT Perturbation
The ideal ILIT perturbation is assumed to be produced by a laser pulse
that can be described as a temporal delta function and whose energy is
deposited in an innitely thin layer at the quartz/gold interface. We assume
that the duration of the ILIT experiment, s
ILIT
, will be short enough to
ignore lateral thermal diusion and edge eects. For example, if s
ILIT
< 10
6
s, D
T
Au
f
1 cm
2
/s then
D
T
Au
s
ILIT
< 0.001 cm, negligibly small

compared to the
f
0.5 cm diameter of the electrode area that is heated.
The analytic solution describing the change in the interfacial temperature,
Feldberg et al. 130
DT
t,ideal
, of the metal/solution interface was derived for us by Johna Leddy
[94]
*
and experimentally tested [6]; the equation for DT
t,ideal
is
DT
t;ideal
DT
eq
d
2
1 r
23
1 r
12
pD
2
T
s
l
i 0
r
23
r
12
i
exp
2i 1
2
d
2
2
4D
2
T
s

66
and the normalized ideal change in temperature,DT
*
t;ideal
, is
DT*
t;ideal

DT
t;ideal
DT
eq
d
2
1 r
23
1 r
12
pD
2
T
s
l
i0
r
23
r
12
i
exp
2i 1
2
d
2
2
4D
2
T
s

67
where subscripts 1, 2, and 3 denote the substrate, metal, and solution
phases respectively, where d
2
is the thickness of phase 2 (the thicknesses of
phases 1 and 2 are assumed to be semi-innite); DT
eq
is the temperature the
electrode would achieve if all the deposited heat were uniformly distributed
within the electrode and if no heat were lost to the substrate or to the so-
lution, i.e.:
DT
eq

q
T
d
2
q
2
C
2
T
68
where q
T
(units: J/cm
2
) is the heat deposited per unit area of electrode.
Other terms in Eq. (67) are dened by the dimensionless reduced param-
eters:
r
jk

b
jk
K
jk
b
jk
K
jk
69
* Leddy has now developed an exact treatment of a four-phase model: quartz/
metal electrode/lm/solution, where it is assumed that the redox species are
located at the lm/solution interface and heat transport in all phases is diusion
controlled. Details will be published elsewhere [95]. This will be discussed further
in Sec. V.E.
ILIT Method 131
b
jk

D
j
T
D
k
T
70
and
K
jk

C
k
T
q
k
C
j
T
q
j
71
where subscripts j and k refer to the phases 1, 2, or 3, C
T
j
(units: J g
1
K
1
) is the heat capacity, q
j
(units: g cm
3
) is the density, and D
T
j
(units:
cm
2
s
1
) is the thermal diusion coecient of the j
th
medium. D
T
j
is directly
related to the thermal conductivity, j
T
j
, by
D
j
T
j
j
C
j
T
q
j
72
Corresponding variables with the subscript k are analogously dened.
Equation (67) gives an accurate picture of the ILIT perturbation if the
full-width-half-max of the laser pulse is very short, i.e., less than 1 ps
(generally that will not be the case, see sec. IV.E). Using the parameter
values for quartz, gold, and water as phases 1, 2, and 3 (see Table 1) we
will see that it is the thickness of the gold layer that is critical in deter-
mining the rise time of the change in the interfacial temperature. In Fig. 5
we show the shape of the perturbation plotted as DT
t;ideal
* vs. D
T
2
s=d
2
2
computed from Eq. (67) assuming parameter values given in Table 1. The
rise time, s
rise
(Fig. 5), is well approximated by
s
rise
c0:2
d
2
2
D
2
T
73
With D
T
2
= 1.26 cm
2
s
1
(for gold) and d
2
= 10
4
cm, s
rise
c2 10
9
s.
Reducing d
2
to 3.0 10
5
cm will decrease s
rise
to
f
2 10
10
s; this esti-
mate presumes that fwhm V 10
11
s for the initiating laser pulse, which is
easily achieved (see Sec. IV.G). Note that the rise time and decay time
cannot be independently controlled; however, the system can be tuned
by adjusting the thickness of the gold lm (most of our work used gold
lms approximately 1.0 10
4
cm thick). Perhaps as important as the
fast rise time is the step-like nature of the perturbation along with two
additional key features: (1) there is no overshoot of the interfacial temper-
ature, as there is with DLIT (see Sec. II.B), and (2) the decay of the tem-
perature is slow compared to the rise time.
Feldberg et al. 132
FIG. 5. Dependence of DT
t;ideal
* on D
T
2
t=d
2
computed from Eq. (67) using
parameter values given in Table 1.
ILIT Method 133
For simplicity we will limit our discussion of Eq. (67) to the con-
ditions that obtain for quartz, gold, and water as phases 1, 2, and 3, re-
spectively, a system we have actually used, with one small but important
modication: because gold does not form a robust layer when vapor-
deposited on quartz we rst vapor-deposit a few hundred angstroms of
titanium, which serendipitously accomplishes three goals described in Sec.
II.B. The thermal properties of the system are determined by the thermal
properties of the three major components (gold, quartz, water) summa-
rized in Table 1.
The ideal interfacial temperature change described by Eq. (67) can
be eectively achieved with modern laser technology. Laser pulses of su-
cient energy and with subpicosecond fwhm are available. For example,
we will describe some experimental results obtained using a titanium-
sapphire laser,*which operates at 10 Hz producing a 0.01 J, 50 fs fwhm
pulse at 798 nm. This laser pulse is eectively a temporal delta function,
as presumed for the derivation of Eq. (67). Most of our data, however, has
been obtained using a Continuum YG580A Nd-YAG laser [96], which
operates at 1 Hz and produces a 0.2 J, 10 ns fwhm pulse at 1064 nm laser.
To capture the open-circuit relaxation information with either system re-
quires ampliers and data collection systems which contribute signi-
cantly to the overall response function of the system. In the next section
we discuss the measurement of the system response function (which in-
cludes the shape of the laser pulse) and the convolution of the response
function with Eq. (67).
E. Nonidealities of the Shape of the ILIT Perturbation and
ResponseExtracting the Relaxation Rate Constant, k
m
Several instrumental factors will eect nonidealities in the ILIT pertur-
bation:
The laser perturbation may not be a temporal delta function (most
of work thus far used a Nd-YAG laser with
f
10 ns fwhm).
The response time of the ampliers for the ILIT response have a
combined bandwidth of
f
100 MHz.
* This laser is a component of Brookhaven National Laboratorys Laser Electron
Accelerator Facility (LEAF).
Feldberg et al. 134
The eective interfacial temperature response will be:
DT
t;eff
*

t
0
DT
s;ideal
*
R
ts
ds 74
where DT
*
t;ideal
was dened in Eq. (67) and R
t
is the response function
normalized such that
l
0
R
s
ds 1 75
R
t
is experimentally determined using a photodiode (ET-2030 Silicon
PIN Detector manufactured by Electro-Optics Technology, Inc., with a
rise time of
f
300 ps: this photodiode is probably a bit slow for proper
evaluation of the rise time of the fastest amplier). The photodiode is
subjected to the laser pulse whose intensity has been attenuated so as to
produce an approximately 5 mV response; that in turn is measured using
exactly the same ampliers and digital oscilloscope that are used to mea-
sure the ILIT response of an electrode. Most of our work was executed
using the Tektronix DSA602A (Tektronix, Inc., Corporate Headquarters,
14200 SWKarl Braun Drive, POBox 500, Beaverton, OR97077). In Fig. 6
the solid and open circles are the normalized experimental response func-
tions obtained for the old and new systems (the old system com-
prises the amplier, Tektronix 602A oscilloscope and the NdYAG laser
(k = 1060 nm; fwhm =
f
10 ns); the new system comprises the new
amplier, Tektronix 602A oscilloscope with the titanium-sapphire laser
(k = 798 nm; fwhm =
f
100 fs). All of our experimental measurements
on SAMs have used the old system. The new amplier response function
would be as much as a factor of 2 faster than shown here if a faster oscil-
loscope were used. As we noted earlier, proper evaluation of R
t
for the
fast amplier system, especially with a faster oscilloscope, will require a
faster photodiode.
It is convenient to nd a function that mimics the shape of the re-
sponse function. We have found the following expression to be eective:
R
t
pt
n
expbt 76
where [combined with Eq. (75)] we obtain n = 6.207, p = 3.54 10
58
s
7.207
, and b = 3.499 10
8
s
1
optimizes the t to the slower response
function and n =3.059, p =8.599 10
36
s
4.059
, and b =1.991 10
9
s
1
optimizes the t for the faster response function. The curves in Fig. 6
ILIT Method 135
were generated using Eq. (76) using the specied parameter values for p, n,
and b for the slower and faster response functions.
In the presence of a rst-order relaxation that is not inuenced by
mass transfer (diusion), an open-circuit potential change induced by an
ILIT perturbation is described by:
DV
oc
DT
eq
ADT
t;eff
* Bk
m
t
0
exp k
m
t s DT
s;eff
* ds 77
where DV
oc
is the experimentally observed change in the open-circuit
potential, DT
t;eff
* was dened in Eq. (74), the coecients A and B were
FIG. 6. Normalized response functions [see Eqs. (74) and (75) and associated
discussion] obtained using the fast amplier (o) with the titanium-sapphire laser
(k = 798 nm; fwhm =
f
100 fs) and using the old amplier (.) with the NdYAG
laser (k = 1060 nm; fwhm =
f
8 ns). Solid lines are optimized ts generated
using Eq. (76) with parameter values given in text.
Feldberg et al. 136
dened earlier [Eqs. (47) and (56)], and k
m
is the rate constant for the
relaxation. If DT
t,e
were a step function having changed from 0 to DT
eq
at t = 0, then it follows from Eq. (77) that:
DV
oc
DT
eq
A B 1 expk
m
t 78
We have essentially recovered Eq. (59). Rearranging Eq. (78), we obtain
DV
oc
ADT
eq
p 1
B
A
1 expk
m
t

79
The signs and magnitudes of A and B are independent of each other; thus,
there are several possible generic response shapes, as shown in Fig. 7, for
B/A = 1.5, 0.5, 0, 0.5, 1.0, and 1.5. Note that when B=0, the re-
sponse is a step function (see discussion, Sec. IV.B).
F. Correlating k
m
to Meaningful Physical Parameters
A correlation of the experimentally measured relaxation rate constant,
k
m
, to meaningful physical parameters was described by Smalley et al.
[23]. That approach was based on the Butler-Volmer expression. In this
FIG. 7. Ideal ILIT responses for dierent values of B/A. Computed from Eq.
(78).
ILIT Method 137
section we will develop a more general expression based on the Marcus
nonadiabatic formalism discussed in Sec. III.A.
The relaxation of the open circuit potential V
oc
following an ILIT
perturbation is a function of the thermal relaxation back to the initial
isothermal condition and the kinetics of an electron-transfer relaxation
that is characterized by the experimentally measured rate constant k
m
(see
Sec. IV.D)k
m
is a function of the electron-transfer resistance R
i
et
(units:
ohm cm
2
), the lm capacitance, C
i
i
(units: F/cm
2
), and the redox or
pseudocapacitance, C
i
redox
(units: F/cm
2
) (as before, the superscript i
denotes the equilibrium value of the variable prior to the perturbation).
The equivalent circuit for this relaxation, shown in Fig. 8, includes the area,
a (units: cm
2
), so that the circuit element R
i
et
=a has the units of ohms and a
C
i
i
and a C
i
redox
have the units of farads. The electron-transfer resistance
for MNA electron transfer is easily deduced. The basic equation for the
current is
i e C
i
red
k
MNA
ox
C
i
ox
k
MNA
red

80
With Eq. (14) we can write:
i ek
o
C
total
gk*; E*
C
i
red
C
total
exp
1
=
2
E*
C
i
ox
C
total
exp
1
=
2
E*

81
FIG. 8. Equivalent circuit for ILIT analysis. See text for details.
Feldberg et al. 138
where g(k*,E*) was dened in Eq. (27) along with an approximation in
Eq. (28) that is valid for the range 1.5k*
0.4445
V E*
i
V 1.5 k*
0.4445
. From
Eq. (81) we can write directly:
1
R
et
e
k
B
T
di
dE*

e
2
2k
B
T
k
o
C
total
gk*; E*
C
i
red
C
total
exp
1
=
2
E*
C
i
ox
C
total
exp
1
=
2
E*

e
2
k
B
T
k
o
C
total
gk*; E*
C
i
red
C
total
exp
1
=
2
E*
C
i
ox
C
total
exp
1
=
2
E*

dgk*; E*
dE*
82
With the Nernstian relationships, Eqs. (52)(54), Eq. (82) can be com-
pactly rewritten as:
1
R
et
e
2
2k
B
T
k
o
C
total
gk*; x
x
1=2
x
i
x
1=2
1 x
i
e
2
k
B
T
k
o
C
total
x
1=2
x
i
x
1=2
1 x
i
dgk*; x
dE*
83
where g(k*,x
i
) is identical to g(k*,E*
i
) and simply indicates that E*
i
on
the rhs of Eq. (82) has been replaced by the identity ln[x
i
]. When the
system is at its initial equilibrium prior to the ILIT perturbation (i.e.,
when x = x
i
) the second term on the rhs of Eq. (83) becomes zero and
the equation reduces to:
1
R
i
et
e
2
k
B
T
k
o
C
total
gk*; x
x
i 1=2
1 x
i
84
and the electron-transfer resistance at the initial (preperturbation) equi-
librium condition is
R
i
et

k
B
T1 x
i
e
2
k
o
C
total
gk*; x
i
x
i 1=2
85
R
i
et

1 x
i
k
o
C
i
i
c
i
x
i1=2
gk*; x
i
86
ILIT Method 139
The value of C
i
redox
(F/cm
2
) is deduced from the Nernst equation; with
Eqs. (52) and (57) we obtain:
C
i
redox

e
2
k
B
T
i
1
C
i
ox
1
C
i
red

c
i
x
i
1 x
i
2
C
i
i
87
The rate constant k
m
for the circuit shown in Figure 8 is simply the
inverse of the circuits RC time constant, which can be written by
inspection:
k
m

a
R
i
et
1
aC
i
i

1
aC
i
redox

1
R
i
et
1
C
i
i

1
C
i
redox

88
Combining Eqs. (86)(88) leads directly to
k
m
k
o
gk*; x
i
x
i 1=2
1 x
i
c
i
1 x
i
2
x
i

89
The maximum sensitivity of k
m
to c
i
occurs when c
i
c (1 + x
i
)
2
/x
i
:
when c
i
>> (1 + x
i
)
2
/x
i
, then k
m
/ k
o
c
i
and neither kj nor c
i
can be
independently determined; when c
i
<< 4 (the minimal value of (1 + x
i
)
2
/
x
i
which obtains when x
i
= 1), the dependence of k
m
upon c
i
disappears.
Typically, for systems we study involving SAMs comprising the redox
species and a diluent, 20 < c
i
< 100 [see Eq. (57)]. Thus, when c
i
= 100
the maximum sensitivity for the determination of c
i
will occur when (1 +
x
i
)
2
/x
i
= 100 or when x
i
= 100 or 0.01.
Since the value of k is generally not known initially, Eq. (89) is not
particularly convenient to use: attempting to evaluate the three unknowns
k
o
, c
i
, and k is impracticalthe dependence on k is weak. When g(k*,x
i
)
is set to unity, Eq. (89) reduces to the simpler Butler-Volmerbased ex-
pression:
k
m
k
o
x
i
1
=
2
1 x
i
c
i
1 x
i
2
x
i

90
Using Eq (90), only kj and c
i
need to be evaluated.
*
As long as the voltage
range examined is within the limits specied by Eq. (32) (which limits
* Because we have deduced Eq. (90) from the MNA expressions, the value of the
parameter corresponding to a in the Butler-Volmer expression is 1/2.
Feldberg et al. 140
depend upon a knowledge of the value of c), the values of parameters
extracted from a least-squares t to Eq. (90) give values of c
i
that are about
25% high and values of kj that are about 20% lowif k c 1 eV. The
relationship in Eq. (90) is plotted in Figure 9 as ln[k
m
/kj] vs. ln[x
i
)] for 0.5
Vc
i
V5000; c
i
= 500 roughly corresponds to a monolayer of redox species
assuming C
total
= 3 10
14
molecules/cm
2
and C
i
i
=4 10
6
F/cm
2
[Eq. (32)].The quality of the ts of the Butler-Volmerbased expression
[Eq. (90)] is uniformly good [23], and the values of c
i
thus obtained agree
well with the values of c
i
computed with Eq. (57) from the values of C
i
i
and
C
total
extracted from the corresponding cyclic voltammograms. We con-
sider this an essential check for self-consistency which conrms that the
faster ILIT response is sampling the same redox population sampled in the
cyclic voltammetric experiment. Because good ts can be obtained by
the manipulation of the values of kj and c
i
, it is obvious that attempts to
also extract values of E will be pointless. Extending the potential range
could, in principle, increase the sensitivity of the t to the value of k;
however, it is also true that the amplitude of the relaxation [see Eq. (55)]
FIG. 9. Plot of ln[k
m
/kj] vs ln[x
i
] based on the Butler-Volmer formalism [see
Eqs. (90) and (52)] for c
i
= 0.5 (lowest curve), 1, 2, 5, 10, 20, 50, 100, 200, 500,
1000, 2000, and 5000 (highest curve).
ILIT Method 141
will decrease with increasing value of (1 + x
i
)
2
/x
i
and evaluation of k
m
becomes increasingly dicult.
It is instructive exercise to generate a synthetic data set using the
MNA-based equation [Eq. (89)]. We do so for c
i
= 50 and kj = 1.0 10
6
s
1
with the assumption that k* = 36 (corresponding to k
f
0.925 eV at
298.2 K) and for the range -6 V ln[x
i
] V 6. Those data are shown as the
points in Figure 10. The line in Fig. 10 is a least-squares t of Eq. (90)
optimizing the values of kj (= 7.9 10
5
s
1
) and c
i
( = 64.5). For a given
range of ln[x
i
] the error will be larger for smaller values of k* and smaller
for larger values of k*.
Analyzing our experimental results in this same way we obtain values
of c
i
, which are in surprisingly good agreement with the values of c
i
extracted from cyclic voltammetric data, suggesting that the error is
tolerable. We have taken this as a good indication that the same redox
population is sampled by the ILIT and cyclic voltammetric experiments
(see Sec. VI.). In addition, we conclude that the excellent ts of the k
m
vs.
V
oc
i
data could not occur if there were one or more rate-controlling non-
FIG. 10. (O): Data synthesized from the MNA formalism [Eq. (89)] for kj =
1.0 10
6
s
1
, c
i
= 50, and k* = 36; (): computed from the Butler-Volmer
formalism [Eq. (90)] for kj = 7.9 10
5
s
1
and c
i
= 64.5.
Feldberg et al. 142
electrochemical processes (which are not potential dependent) coupled to
the electron transfer. Very fast coupled processes that remain reversible
on the time scale of the ILIT experiment would not be detected.
V. EXPERIMENTAL IMPLEMENTATION OF ILIT
A. The Cell
A schematic of the ILIT cell is shown in Fig. 11. The cell is designed so that
the irradiated area of the back side of the electrode is congruent with the
area of the electrode surface exposed to the electrolyte. Electrodes are
positioned so that there is a relatively clear path between the working
electrode and the acoustic pick up.
1. The Working, Reference, Counter, and Pseudo-Reference Electrodes
A schematic of the working electrode conguration is shown in Fig. 12.
The radius of the quartz disk, r
Q
, is 1/2 inch; the thickness, d
Q
, is 1/8 inch.
A thin layer of titanium (
f
5 10
6
cm) is vapor deposited on the quartz
disk followed by deposition of the desired thickness, of gold (generally
f
10
4
cm; where noted
f
0.3 10
4
cm). The radius of the deposited gold
layer, r
d
, is 1.1 cm; the radius of the exposed electrode, r
e
, is 0.475 cm.
Details of the preparation of the gold electrode are presented in Sec. V.B.
The entire assembly is held together with a simple clamp (not shown in
Fig. 12).
The reference electrode for most of our experiments was a saturated
sodiumcalomel electrode (SSCE) whose potential is 0.236 Vvs. the normal
hydrogen electrode (the potassium ion in a conventional calomel reference
electrode causes problems when working with perchlorate media). The
counter electrode was a platinum foil
f
2 cm
2
. The pseudo-reference
electrode was a platinum foil
f
1 cm
2
positioned as close as possible to
the working electrode. The area of the working electrode is pr
2
e
=0.71 cm
2
.
With the conservative assumption that the capacitance of each electrode is
f
10
5
farads/cm
2
, the series capacitance of the working and pseudo-
reference electrodes will be
f
10
5
farads. The input impedance of the open
circuit amplier is 10000 ohms (see Sec. V.D); thus, the RC time constant
will be
f
10
1
sorders of magnitude larger than the maximum ILIT
measurement time (
f
10
6
s). Thus, the capacitances of the working and
pseudo-reference electrodes can drive the open circuit amplier without
perturbing their potentials.
ILIT Method 143
FIG. 11. The ILIT cell.
Feldberg et al. 144
2. The Acoustic Sensor
The acoustic sensor was custom made for us by Panametrics (Panametrics,
Inc., Waltham, MA). The 10 MHz 0.25 inch diameter sensing element is
sealed in an all-plastic casing 0.44 inch in diameter; 1.25 inches long (the
usual sensor casing is metal that is grounded, as is the working electrode,
thereby producing an unacceptable ground loop when the metal-encased
FIG. 12. Conguration of the working electrode: radius of quartz disk, r
Q
, is
1/2 inch; thickness of quartz disk, d
Q
, is 1/8 inch; radius of deposited gold layer,
r
d
, is 1.1 cm; radius of active electrode, r
e
, is 0.475 cm.
ILIT Method 145
sensor is immersed in the electrolyte). A typical response is shown in
Fig. 13. The earliest response (DV
A
in Fig. 13) is the most reproducible
we believe that subsequent peaks are more likely to be subject to sonic
bounces o cell components that may not be precisely positioned from one
experiment to another.
It is easy enough to demonstrate that the acoustic response is linearly
related to the intensity of the laser beam by controlling the laser beam
intensity with neutral density ltersthese lters are easily calibrated
using a spectrophotometer to measure their transmittance at the wave-
length of interest (e.g., 1060 nm for the Nd-YAG laser). However,
calibrating the relationship between the selected acoustic response, DV
A
,
and the interfacial temperature change, DT, is a bit trickier. We can do this
in two dierent ways. The easy way is to measure the ILIT response of a
FIG. 13. Response of the acoustic sensor: DV
A
is the selected, reproducible
early response.
Feldberg et al. 146
bare gold electrode in an aqueous solution comprising 0.01 M K
3
Fe(CN)
6
,
0.01 M K
4
Fe(CN)
6
, and 1.0 M KF. For suciently high concentrations
the electron transfer will not be diusion limited and concentration polar-
ization can be ignored. Then [from Eqs. (37) and (52)] the response is
simply expressed:
DV
oc
DT

dV
oc
dT

dE
o=
dT

dV
TJ
dT

R
F
ln
c
ox
c
red

91
where c
ox
and c
red
are the bulk concentrations of the redox species. From
reported values of dE
o/
/dTand dV
TJ
/dT[97,98], we can then obtain a good
estimate of DT.
It is also possible to obtain an absolute calibration of DT using the
ILIT approach [24]. Combining the Nernst relationship,
RT
F
ln
c
ox
c
red

V
i
oc
E
o=
92
where V
i
oc
is the initial open-circuit potential, with Eq. (91) gives
DV
oc
DT
dE
o=
dT

dV
TJ
dT

DT
V
i
oc
E
o=
T
i

93
If a series of experiments are carried out at a given temperature, T
i
with
constant DT, but at several dierent values of V
i
oc
(V
i
oc
is changed by
changing the ratio c
ox
/c
red
), a plot of DV
i
oc
vs. V
i
oc
/T
i
will have a slope of
DT. Unfortunately, this is a tedious approach and not convenient for
routine calibration. However, it does serve to help us calibrate the ferri-
ferrocyanide system under the aforementioned conditions and determine
that dE
o/
/dT + dV
TJ
/dT =1.47 10
3
V/K in 1.0 M KF [24]. Then the
acoustic response can be directly calibrated from a single measurement of
the ILIT response for the ferri-ferrocyanide system.
Another straightforward (non-ILIT) way to calibrate the relevant
ferri-ferrocyanide thermal response is to construct a nonisothermal cell
in which the temperatures of the reference electrode and working
electrodes are independently controlled. Then dE
o/
/dT + dV
TJ
/dT is
directly and easily obtained. The Soret potential (see Sec. V.E), which
arises because of the temperature gradient in the solution, will exist
between the working and reference electrodes and will depend only upon
the dierence in temperature of the two electrodes and not upon the
shape of the gradient. Note that it is inappropriate to use the response
of an isothermal cell in which the temperature of both electrodes are the
samethe thermal junction potentials are eliminated but the response
ILIT Method 147
now depends upon the temperature dependences of the both the working
and reference electrodes.
Whatever method is deemed appropriate for establishing the rela-
tionship between DT and the responses of the acoustic sensor, it is im-
portant that the cell components are reproducibly positioned from one run
to another.
3. The Diuser
By using a diuser to expand the diameter of the laser beam, we can
improve the homogeneity of the beam intensity impinging on the back of
the electrode (see Fig. 11). We suggest positioning the diuser so that
the beam width is expanded to
f
2.5d
e
, where d
e
is the diameter of the
electrode.
Diusion Properties
The thin lm electrode is a key ingredient of the ILIT methodology.
Uniform deposition of the metal is vital, and accurate evaluation of the
thickness(es) of the layers is important for accurate analysis of the data (it
is possible to deduce the thickness from responses obtained in the absence
of electron transfer [6]). Virtually all the experimental work we describe
involved gold electrodes that are
f
1.0 10
4
cmthick. However, our most
recently developed ampliers improve the time resolution by about a factor
of 10 (see Fig. 6)with a fast laser system (producing a pulse with fwhmV
f
10 ps) and electrodes
f
3.0 10
5
5.0 10
5
cm thick, the rise time of
the interfacial temperature will be
f
1.54.0 10
10
s [see Eq. (73)]. We
have carried out a few preliminary experiments with these thinner electro-
des, and they appear to behave as expected.
1. Electrode Preparation
The electrodes for our work were prepared for us by Professor Christopher
Chidsey and his coworkers at the Department of Chemistry, Stanford
University. Quartz (fused silica) disks purchased from Heraeus-Amersil
(Buford, GA) were cleaned for several minutes in a freshly prepared
mixture of approximately 2 volumes H
2
SO
4
and 1 volume of 30% by
weight H
2
O
2
in water at
f
80jC.
*
The disks are then rinsed with copious
* This mixture is commonly referred to as pirhana solution and must be
handled with great caution: full face shield and body protection are required.
Feldberg et al. 148
amounts of Milli-Q(TM) water (4-bowl system with nal Organexk
cartridge; Millipore Corp, Bedford, MA).
The wet disks are immediately immersed into the vapor of reuxing
isopropanol. Once they reach the reux temperature, as noted by a
reduction in the rate of alcohol condensation on the disks, they are
removed into the room air where they rapidly become dry and slowly cool
to room temperature. At this point, the disks are individually weighed to
the nearest 0.1 mg. The disks are then loaded into a metal tray with large
circular regions on their bottom faces exposed and the tray placed in the
load lock of a vacuum chamber. The load lock is pumped to about 10
3
torr and then the tray is translated into the chamber and the load lock
sealed o from the chamber. The chamber is pumped to about 10
7
torr
with a cryopump. The major residual gas is water from the rotatable
rubber seal used between the evacuated space in the chamber and the water
ow path into and out of the rotatable copper crucible mentioned below.
Titanium vapor is sublimed from a Ti ingot by bombardment with about
8 keVelectrons while the periphery of the ingot is cooled by sparse physical
contacts with the water-cooled copper crucible in which it rests. The Ti
vapor condenses in the line of sight from the source onto the exposed
bottoms of the disks. The thickness of the deposited titanium, d
Ti
, is
monitored by a calibrated quartz crystal balance close to the quartz disks.
Typically d
Ti
f
5 10
6
cm.
Gold is similarly deposited by electron beam evaporation from a
small nugget of gold sitting in the bottom of a separate pocket in the water-
cooled copper crucible. To get the gold nugget under the e-beam and in the
line of sight of the disk bottoms, the crucible is rotated by a mechanism
attached to a mechanical feedthrough on the chamber wall. The source of
the gold nugget is either a coil of gold wire or set of gold pellets, which are
then melted into a nugget the rst time they are used. Because of the
temperature mismatch between the molten gold and the copper crucible,
the molten gold does not wet the (oxidized) copper surface and so stays as a
nearly spherical nugget that is easy to heat with the electron beam. After
the desired thicknesses of Ti and Au are deposited, the disks are removed to
the load lock, which is then vented to air. Typically d
Au
f
1.0 10
4
cm.
The disks are reweighed to determine the total mass of deposited
metal on each disk. The exact eective thicknesses of the metal layers on
each disk are derived assuming that the metals are pure and of bulk density
and that the quartz crystal monitor is only o by an unknown but constant
scale factor relative to the values it should have read if placed at the
location at which each disk was during the depositions.
ILIT Method 149
2. Thermal Diusion Properties
It is convenient to deduce the thickness of gold lm, d
//
Au
, that is equivalent
(in terms of thermal diusion) to a given thickness of titanium. The
relevant properties of gold and titanium are summarized in Table 1. By
equivalent, we mean that the rise times are identicalthe magnitude of the
temperature change will be a bit dierent. To equate the rise times we
equate:
D
T
Au
t
d
==
Au
D
T
Ti
t
d
Ti
94
Then, the thickness of the gold lm, d
/
Au
, that is equivalent to the thickness
of the combined Ti and Au layers is [with Eq. (94)]:
d
=
Au
d
Au
d
==
Au
d
Au
d
Ti
D
T
Au
D
T
Ti
95
Although we have computed an equivalent thickness of gold for the
combined Ti and Au electrode that will give us the correct time response,
C
Au
T
will not be the correct heat capacity. The eective average heat
capacity, C
Au
T/
, of the combined Ti and Au layers is
C
T=
Au

C
T
Ti
d
Ti
q
Ti
C
T
Au
d
Au
q
Au
d
Au
d
Ti
D
T
Au
D
T
Ti

q
Au
96
For all of the ILIT the work we have reported in the literature, d
Ti
/(d
Au
+
d
Ti
) V0.05 and the response time was of the order of 15 ns. Consequently,
for those systems we simply assumed d
Au
/
= (d
Au
+ d
Ti
)the error is
negligible [6].
C. Preparation of Self-Assembled Monolayers
Many techniques are available for the preparation of self-assembled
monolayersdetailed discussion of those methods is beyond the scope
of this chapter. Finklea has presented a thorough reviewof the preparation
and characterization of self-assembled monlayers [34]. Recently Brevnov,
Finklea, and Van Ryswyk described a unique method of chemically
attaching [4-aminomethylpyridine)Ru(NH
3
)
5
]
2+
to a previously self-as-
sembled mercaptocarboxylic acid monolayer (on gold) [37], and we discuss
some results obtained with such systems. Our preparations of systems
involving the ferrocene redox moiety [13] followed the conventional
Feldberg et al. 150
protocol of deposition of SAMs (on gold) from an organic (e.g., ethanol or
chloroform) solution containing of the SAM components.
D. The Electronics
1. The Special Potentiostat and Fast Amplier
Until 150 As before the laser pulse, the electrode potential vs. the reference
electrode is controlled by the special potentiostat. At 150 As before the
laser pulse, the special potentiostat changes its function and, for the next 5
ms, sustains a constant current at its 150 As value while the open-circuit
amplier follows the potential between the working and pseudo-reference
electrodes and amplies it: the old amplier used in all reported work
amplies by a factor of 50; the new faster amplier amplies by a factor
of 200. In the perfect system where the only redox species in the system are
those that comprise the redox couple that is attached to the electrode
surface, no current should be required to maintain a set potential; however,
there will always be adventitious background current and a small com-
pensating constant current is required to eliminate a small drift during the
measurementthe shorter the time of measurement, the less serious the
drift. We assume that this current is unaected by the laser pulse and that
the ILIT-induced open circuit response is due entirely to the electron
transfer process of interest. The input impedance of the amplier is 10000
ohmsthe capacitance of the electrode which is driving the amplier is
greater than 10
6
F; thus, the RC time constant for discharging the
capacitance is greater than 0.01 stoo long a time, by several orders of
magnitude, to be a problem. The latest version of the fast amplier is
attached to the electrode assembly to minimize the connection lengths and
inductive eects. The special potentiostat and fast amplier were designed
and constructed by the Brookhaven National Laboratory Instrumentation
Division; circuit diagrams can be accessed through the web.
*
2. Data Acquisition
Data were acquired using a Tektronix DSA602A (Tektronix, Inc. Beaver-
ton, OR) with a 1 Ghz bandwidth. Instrumentation with 10 Ghz band-
* Files containing circuit diagrams, part numbers, etc., for the potentiostat and
fast amplier are accessed using URLs http://www.inst.bnl.gov/cgi-bin/view_io.
pl?table=ionumbers&id=45 and http://www.inst.bnl.gov/cgi-bin/view_io.pl?
table=ionumbers&id=553 respectively. Click on the numeral in table entry Ar-
chive CD and follow instructions to download the desired les.
ILIT Method 151
width is now available. Single shot data is stored with 8 bits of resolution
full scale plus sign. We routinely average 10 shots and attain
f
10 bit
resolution full scale.
E. Potential Problems
1. The RF Noise Problem
Achallenging experimental problemis the suppression of the laser-induced
radiofrequency (RF) noise produced by the triggering of the ash and/or
the Pockels cell. A straightforward way to solve the problem is to put the
electrochemical cell in a Faraday cage; we have found that when we use the
NdYAGlaser, the cage must be fastidiously sealed to prevent signicant
RF interference. Remember that we are measuring signals whose magni-
tude is, at most, a few millivolts. On the other hand, we have carried out a
few experiments using the titanium-sapphire laser used in Brookhaven
National Laboratorys Laser Electron Accelerator Facility (LEAF). This
laser is physically well separated from the measurement station, and even
without shielding the RFnoise is virtually completely suppressed. Another
approach (which we have not used) is an optical delay line whereby the la-
ser beam is diverted over the distance required to obtain sucient time de-
lay (1 ns/ft) so that the RFhas died out before the ILIT process is initiated.
2. Temperature Eect
Our analysis of the ILITresponse assumes that the operative rate constant,
k
m
[see Eq. (77)], is eectively constant during the course of the ILIT
measurement. For a perfect step function the operative temperature would
be constant at T = T
i
+ DT
eq
[see Eq. (78)]. However, the ILIT
perturbation only approximates a step function. Even if the system is only
reasonably tuned
*
(see Fig. 5), the postperturbation change in the inter-
facial temperature will likely be much less than the typical maximum value
of DT
eq
V 5 K. The sensitivity of the rate constant to temperature is
deduced from:
Dlnkj
DH
p
RT
2
DT
k
97
* By properly tuned system we mean that the thickness of the gold electrode
has been chosen to optimize the shape of the temperature-time prole for the
kinetics of interest.
Feldberg et al. 152
where DT
k
is the change in interfacial temperature during the relaxa-
tion, typically less than 2 K. With T = 25jC and DH
p
= 2.2 10
4
J/mol
for the heterogeneous ferrocene electron transfer, we conclude [Eq. (97)]
Dln[kj]<0.12. Experimentally we routinely conrm that the value of kj is
independent of DT
eq
. Obviously it is advantageous to use the smallest
practical value of DT
eq
.
3. The Soret Eect
The Soret eect is eectively a liquid junction potential produced by a tem-
perature gradient in a homogeneous electrolyte. The activity coecients of
ions are generally not identical; consequently, a temperature gradient
produces a driving force, which will be opposed by potential. The impli-
cations of the Soret eect in an ILIT perturbation were discussed by
Smalley et al. [5], who showed that there is a linear relationship between the
Soret potential, DV
Soret
, and the temperature dierence between the elec-
trode surface and the bulk solution, DT, i.e.:
DV
Soret
b
Soret
DT 98
DV
Soret
is the solution component of V
TJ
[Eq. (34)]. Implicit in our analysis
is the assumption that the time constant for the establishment of the Soret
potential is innitely fast. We assume that the rate constant for this process,
k
Soret
, will be comparable to the rate constant for diuse double layer
relaxations [5,46,99,100]. For a symmetrical binary electrolyte:
k
Soret

e
2
k
B
Tee
o
D
1
D
2
cz
2
99
For typical parameter values for a 1 M binary electrolyte (e.g., KCl) in
aqueous media (T = 298.2, q = 78, q = 8.854 10
14
C V
1
cm
1
, D =
2 10
5
cm
2
/s, c = 6 10
20
molecules/cm
3
, and z
2
= 1), we calculate that
k
Soret
= 2 10
10
s
1
.
4. Pressure Eect
The local heating adjacent to the electrode (generally) causes the solvent to
expand, thereby eecting a local change in pressure. A simple analysis
allows us to estimate the largest possible local pressure change by assuming
that the solution does not move during the time of interest. Consider the
electrode/solution interface and a column of solution above it. The volume
of solution dened by the layer of thickness, d, and the cross section area,
ILIT Method 153
A, is assumed to be the volume subjected to an instantaneous change in
temperature, DT. For that volume we can write [101]:
dV 0
BV
BT

P
dT
BV
BP

T
dP 100
and therefore
BP
BT

V

BV
BT

P
BV
BP

T
a
P
b
P
101
where a
P
, the thermal expansivity, is dened by
a
P

1
V
o
BV
BT

P
102
where V
o
is a standard volume, usually the volume at 0j C, and where b is
the compressibility dened by
b
P

1
V
o
BV
BP
T
103
For water, at 30jCthe value of a
P
is 3.03 10
4
K
1
and b
P
is 4.46 10
11
cm
2
/dyne. From Eq. (101) we then deduce that (BP/BT)
V
= 6.8 10
6
dynes/cm
2
= 6.7 atm/K. Generally, DT associated with an ILIT experi-
ment is less than 4 K, so DPis less than
f
27 atm. How will this impact on a
rate constant? Weston and Schwarz [102] have noted that dln[k]/dP for
solution reactions is generally of the order of 0.001 atm
1
so it is unlikely
that a pressure change of 27 atm will have a measurable eect on an
electrochemical rate constant.
As discussed earlier (Sec. V.A) we use an acoustic sensor to measure
the intensity of the resultant pressure pulse and correlate that directly to the
change in the interfacial temperature.
5. Viscosity Eects
A number of workers have reported evidence that electron transfer rate
constants involving species in solution are a function of the viscosity of the
medium [32,54,103,104]. However, Smalley and Creager [105] and Creager
and Weber [106] saw no denitive viscosity eect on the electron transfer
rate constant for ferrocene attached to a gold electrode. Even if there were
a small eect, neither the temperature dependence nor the pressure depen-
Feldberg et al. 154
dence of viscosity [107] would be sucient to eect a measurable change
in the interfacial rate constants. Viscosity eects are discussed further in
Sec. VIII.
6. Delayed Heat Transport Through the Monolayer Film
Implicit in our ILIT analyses is the assumption that the redox moiety
instantaneously senses the interfacial temperature even though it is cova-
lently tethered at some distance, d
redox
, from the electrode. We adopt the
simplest argument, i.e., that thermal propagation through a single mole-
cule is very ecient, occurring through the vibrational modes of the system
at a speed roughly corresponding to the speed of sound, v
sound
. If we
assume that v
sound
= 1.5 10
5
cm/s, the speed of sound in water, we
conclude that the temperature front will take
f
7 10
13
s to go 10 A
and
f
2 10
12
s to go 30 A
. If we were to estimate these transit times based

on diusional heat transport, then the thermal transit time, H
T
, would be
f
d
2
redox
/D
T
; if D
T
= D
T
H
2
O
c 0.001 cm
2
/s, H
T
c10
11
s when d
redox
=
10
7
cm and c 10
10
s when d
redox
= 3 10
7
cm. However, our
experimental results suggest that the ferrocenium/ferrocene couple, one
of the most facile ET redox couples, tethered to gold by an oligophenyl-
enevinylene bridge will not be fast enough for these thermal transit times to
matter [3]: when d
redox
= 3.5 10
7
cm, kj = 5.5 10
5
s
1
the
corresponding relaxation time, 1/k
m
(at 25j C and E = E
o/
) is
f
6
10
8
s for c
i
= 60 and
f
2.2 10
7
s for c
i
= 12.2. The relationship
between k
m
and kj is complicated (see Sec. IV.F), but generally k
m
> kj;
nevertheless, the measured relaxation time will be orders of magnitude
longer than the thermal transit time. Leddy [95] has now developed an
exact treatment of a four-phase model, quartz/metal electrode/lm/solu-
tion, where it is assumed that the redox species are located at the lm/
solution interface, heat transport in all phases is diusion controlled, and it
is the temperature at the lm/solution interface that is of interest. Since the
thermal properties of the lm are not precisely known, this treatment will
serve best to indicate just when the three-phase model [Eq. (67)] is
inadequate (in the three-phase model we assume that thermal transport
through the lm is innitely fast).
7. A General Comment on the Error in Typical Measurements of kj
The error level for values of kj measured using ILIT is fairly represented
in Figs. 1416 where the results of a number of independent experiments
are displayed. The reproducibility in the value of kj (for a given temper-
ature and for a given redox system) is typically of the order of F5%. In
addition to conrming that the voltage dependence of the values of k
m
ob-
ILIT Method 155
tained for any given preparation is consistent with the theory (Sec. IV.F),
we also conrm that the values of parameters A [Eq. (47)] and B [see Eqs.
(48)(56)] are independent of the energy of the laser pulse (controlled
using neutral density lters) and that k
m
and kj are independent of the
energy of the laser pulse. We also conrm that the value of kj is inde-
pendent of the surface concentration of the attached redox species which
determines the value of c
i
[see Eq. (57)]. It is reasonable to expect that the
FIG. 14. Upper curve: ILIT response when V
oc
i
= E
oi
= 0.375 V vs. SSCE for
a SAM formed from a mixture of HS[ C u C-]
2
Fc and HS(CH
2
)
9
CH
3
[solid
line is least-squares curve for k
m
= 2.2 10
7
s
1
computed from Eq. (77)]; lower
curve: ILIT response for V
oc
i
= E
oi
= 0.300 V vs. SSCE for a SAM formed from
a mixture of HS[ C u C-]
2
Fc and HS(CH
2
)
9
OH [solid line is least-squares
curve for k
m
= 1.2 10
7
s
1
computed from Eq. (77)]. The dotted line for each
transient is the computed response for the same system assuming innitely slow
(i.e., zero) electron transfer.
Feldberg et al. 156
error level will increase as the time constant of the measurement 1/k
m
becomes comparable to the fwhm of the response function of the instru-
mentation (see Fig. 6). In Sec. V.G we discuss experimental protocols we
follow to obtain optimal results.
F. Energetic and Timing Considerations for Single
and Multiple Pulse Experiments
In these analyses we consider the width of the laser pulse to be a temporal
delta function.
FIG. 15. (O): Experimental k
m
values at T = 298.2 K plotted as a function of
V
oc
i
for a SAM formed from HS-[ CH=CH-]
2
Fc and HS(CH
2
)
9
CH
3
;
(.... : the best t obtained using the Butler-Volmerbased expression [Eq. (90)]
with kj = 3.8 10
6
s
1
and c
i
= 79.6; (): the best t obtained using the
Marcus-nonadiabaticbased expression [Eq. (89)] assuming k = 0.93 eV with kj
= 4.29 10
6
s
1
and c
i
= 71. The value of c
i
determined from the cyclic
voltammogram of the same SAM was 85.
ILIT Method 157
1. Laser-Pulse Energy Requirements for the Single Shot
Experiment
The energy, q
T
(units: J/cm
2
), that must be absorbed by the electrode to
produce a specied temperature change, DT
eq
, in a metal lm of thickness,
d
2
, has been dened in Eq. (68). To estimate the total laser energy (per
pulse) required, we must consider several factors:
Some fraction of the light, r
hv
, will be reected; r
hv
will depend upon
the metal (which in our case is the titanium under layer) and the
wavelength of the impinging light.
The diameter of the laser beam is expanded (using a Quartz diuser;
see Fig. 11 and discussion in Sec. V.A) to
f
5d
e
, where d
e
is the
FIG. 16. Arrhenius data for (o): SAM formed from A and HS(CH
2
)
9
CH
3
in
1.0 M HC1O
4
; (.): SAM formed from B and HS(CH
2
)
9
CH
3
in 1.0 M HC1O
4
;
(): SAM formed from A and HS(CH
2
)
9
CH
3
in 0.10 M HC1O
4
; (5): SAM
formed from A and HS(CH
2
)
11
CH
3
in 1.0 M HC1O
4
; (4): SAM formed from A
and HS(CH
2
)
9
CH
3
in 1.0 M H
2
SO
4
.
Feldberg et al. 158
diameter of the electrodethis ensures that the central portion of
the beam impinging on the electrode is very nearly uniform in
intensity.
*
Then, if we assume that the energy is uniformly distributed through out the
expanded area, the minimum energy, Q
L
(units: J), required for the laser
pulse is
Q
L
zf
p
41 r
hv
5d
e
2
q
T
f
6:25
pd
2
e
q
T
1 r
hv
104
or, with Eq. (68)
Q
L
z
f
6:25
pd
2
e
DT
eq
d
2
q
2
C
T
2
1 r
hv
105
where, as before, the subscript 2 denotes that the parameter is associated
with the metal electrode phase. For a gold lm, if d
e
= 1 cm, DT
eq
= 5 K
(the maximum that should be used in an ILIT experiment), with the
thermal properties for gold given in Table 1, and assuming that r
hv
(for
titanium)
f
0.65 (based on calculated reectivity for evaporated titanium
[108]), we estimate that:
Q
L
z
f
7:5 10
2
J=cmd
Au
106
where d
Au
is expressed in cm. When d
Au
= 3 10
5
cm [chosen to eect a
fast rise time, i.e.,
f
1.4 10
10
s; see Eq. (73)] q
T
c 4 10
4
J/cm
2
must be absorbed and Q
L
z
f
2.3 10
2
J. For the experiments reported in
the present work, d
Au
=1.0 10
4
cm(producing a rise time of 1.6 10
9
s)
q
T
c1.3 10
3
J/cm
2
must be absorbed and Q
L
z
f
7.5 10
2
J, a value
that was easily exceeded by the Nd-YAG laser used in those experiments.
2. Multiple Pulse Experiments
The deleterious eect of random noise in a single pulse experiment can be
diminished by averaging the ILIT responses for n
max
experimental repeti-
tions for which data are collectedthe signal-to (random)-noise ratio will
be increased by a factor of

n
max
p
.
y
Our usual practice was to signal average
the ILIT responses for 10 repetitions at 1 Hz, thereby increasing the signal-
to-noise ratio by approximately a factor of 3. We discuss here the
conditions required so that the ILIT responses are identical for all pulses
*
In our experimental work we only expanded to
f
2.5 d
e
.
y
The intensity of the laser pulse may not be absolutely constant. Consequently
the response of the acoustic sensor (Sec. V.A) was also averaged.
ILIT Method 159
and equivalent to the response for to a single pulse. There are, however,
two additional factors that must be considered: the spacing of the pulses
and the total energy dumped into the system, a function of total number
of pulses and of q
T
. In the following analyses we will ignore the role of
stirring which can enhance the rate of the systems return to its initial
thermal equilibrium.
a. Estimating the Minimum Time Between Pulses in a
Multipulse Experiment
The temporal spacing of the pulses, Dt
L
, in a multipulse experiment must
be such that each individual ILIT response is eectively independent of the
preceding response(s) and thus accurately described by the equations
developed in Sec. IV.C. We specify that that condition is adequately met
when DT
Dt
L
/DT
eq
V 0.01. From Eq. (77) we can demonstrate that the
interfacial potential will also have virtually returned to its initial value,
regardless of the value of k
m
. When DT
Dt
L
/DT
eq
<< 1, the temperature of
the metal lm is essentially uniform and the thermal diusion layers in the
substrate and the solution have become much larger than the thickness of
the lm, d
1
. The problem is now equivalent to the classic diusion problem
where the initial distribution of the diusing entity is a spatial delta
function and where the entity is diusing into two dierent phases, as
described in the Appendix. The solution to this problem is
DT
Dt
L

q
T
1 r
T
1=3
q
3
C
T
3
pD
T
3
t
107
With the denition of DT
eq
given in Eq. (68), we can now specify the
conditions required to eect DT
Dt
L
/DT
eq
V 0.01:
DT
Dt
L
DT
eq
d
2
q
2
C
T
2
1 r
T
1=3
q
3
C
T
3
pD
T
3
Dt
L
V 0:01 108
where [see Appendix, Eq. (120)]:
r
T
1=3

q
1
C
T
1
D
T
1
q
3
C
T
3
D
T
3
109
Using the values of the thermal parameters for quartz (phase 1), gold
(phase 2), and water (phase 3), we deduce that the condition DT
Dt
L
/DT
eq
V
0.01 will be met when
Feldberg et al. 160
Dt
L
z2:2 10
5
s=cm
2
d
2
Au
110
When d
Au
= 1.0 10
4
cmwe compute Dt
L
z2.2 10
3
s from Eq. (110);
in our experiments using the Nd-YAG laser Dt
L
is 1.0 s, more than
sucient to reestablish the initial conditions. Under these conditions,
5 10
3
s after the ILIT perturbation, the special potentiostat (see Sec.
V.D) imposed the initial potential and reestablished the compensating
current that would be sustained during the subsequent ILIT perturbation
(see Sec. V.D).
b. Estimating the Tolerable Maximum Number of Pulses
Application of multiple pulses will gradually heat up the substrate and the
solution in the vicinity of the electrode. If we assume that we will use n
max
multiple pulses spaced by Dt
L
, we now ask what value of n
max
will just
produce the maximum acceptable change, DT
max
i
, in the base level of the
initial interfacial temperature, T
i
. For the purposes of this estimate we
assume that the energy is absorbed at a constant rate, I
qT
= q
T
/Dt
L
for a
time n
max
Dt
L
. We further simplify the problem by assuming that virtually
all the energy will be in the substrate and solution phases adjacent to the
electrode and assume that the electrode is isothermal and innitely thin.
The solution to the problem is now a simple modication of the Sand
equation [42], which allows us to partition the heat ux into the substrate
and solution media (see Appendix):
DT
i
max

21 a
T
k
I
q
T
n
max
Dt
L
p
q
1
C
T
1
pD
T
1

2a
T
k
I
q
T
n
max
Dt
L
p
q
3
C
T
3
pD
T
3
111
where a
k
T
is the fraction of the heat diusing into the k
th
phase, 1 a
k
T
is the
fraction diusing into the j
th
phase, and is dened by Eq. (119). Then Eq.
(111) becomes:
DT
i
max

2q
T

n
max
p
1 r
T
1=3
q
3
C
T
3
pD
T
3
Dt
L
112
where r
1/3
T
is dened by Eq. (109). Combining with Eq. (68) gives:
DT
i
max

2DT
eq
d
2
q
2
C
T
2
n
max
p
1 r
T
1=3
q
3
C
T
3
pD
T
3
Dt
L
113
ILIT Method 161
With a bit of rearrangement we obtain the expression for n
max
:
n
max

p
4
DT
i
max
1 r
T
1=3
q
3
C
T
3
D
T
3
Dt
L
Dq
T
2
114
p
4
DT
i
max
1 r
T
1=3
q
3
C
T
3
D
T
3
Dt
L
DT
eq
d
2
q
2
C
T
2
2
Note that n
max
depends not only on Dt
L
but also depends strongly on Dq
T
or equivalently on DT
eq
[see Eq. (68)]. If we assume that DT
max
i
V 1 K,
DT
eq
=4 K, d
2
=0.0001, and Dt
L
=10
2
s, then, with the values in Table 1
for the relevant properties of quartz, gold, and water as phases 1, 2, and 3,
we conclude that n
max
V 718 or 7.18 s total time.
G. Some Suggested Experimental Protocols
The theory and experiments we have described suggest some experimental
guidelines and protocols that will enhance the quality and analysis of ILIT
responsessome will be obvious to the electrochemical practitioner; some
may not be. The following suggestions presume that the thickness of the
metal lm electrode is uniform and accurately known and that the SAM
itself has been appropriately characterized.
Execute cyclic voltammetry prior to the ILIT experiment to ensure
sensible behavior, to provide base information about the relevant
potential range of interest, and to estimate of C
total
and C
i
, and
therefore c
i
[see Eq. (57)].
Assemble the ILIT cell ensuring that the location of all cell com-
ponents is consistent with their location during calibration of the
acoustic response.
Conrm that the potentiostat current at any given potential should
less than
f
10
6
A.
Explore an adequate range of potential (E
o/
F 0.15 V).
Carry out experiments with dierent laser intensitiescontrolled
with calibrated neutral density lters (see Sec. V.A).
Repeat experiments at dierent T
i
exploring as large a temperature
range as possible.
Repeat one or two measurements at original T
i
to conrmstability (a
full set of experiments on a given electrode preparation can take
several hoursit is important to conrmthat the behavior has not
changed over this period of time).
Feldberg et al. 162
Repeat experiments with SAMs with dierent C
total
for a given redox
and diluent combination.
Data analysis: evaluate k
m
(Sec. IV.D) kj and c
i
(Sec. IV.F). Conrm
that kj is independent of laser intensity and C
total
. Conrm that c
i
estimated from cyclic voltammetry is (or is not) the same as c
i
evaluated from the ILIT experiments.
Do Arrhenius plots to determine DH
p
of k
o
. Compare to values of k
evaluated from overpotential measurements, if available.
VI. A FEW EXAMPLES OF MEASUREMENTS
OF INTERFACIAL KINETICS
The ILIT approach has allowed us to characterize some fast electron
transfers between a gold electrode and a surface-attached redox species [1
3]. Data prior to 2001 were analyzed using values of c
i
determined from
cyclic voltammetry (CV) and extracting only the value of kj fromthe k
m
vs.
V
oc
i
plots. However, those early data have been subjected to spot-check
reanalysis as described in Sec. IV.F, and we have conrmed that the values
of c
i
determined from CV and from the ILIT experiments agree within
experimental error. Thus, we have concluded from this that the CV and
ILIT techniques are sampling the same redox population in all cases we
have studied.
A. Some Typical Transients
In Fig. 7 we showed some possible theoretical ILIT responses. The nature
of the response will depend upon the relationship between the parameter A,
which denes the amplitude of nonkinetic component of the ILITresponse
[see Eq. (47)], and the parameter B, which denes the amplitude of the
kinetic component of the ILIT response (see Eq. (5) and see Sec. IV.A for
complete discussion). Parameter A will depend dramatically on the poten-
tial of zero response (pzr), which is primarily a property of the diluent
component of the SAM. A nice demonstration of this is shown in Fig. 14
for two ILIT transients (plotted as DV
oc
/DT
eq
vs. t) for two dierent
ferrocene-containing SAMs: one (upper curve) formed from a mixture
of HS-[ CuC-]
2
Fc and HS(CH
2
)
9
CH
3
and the other (lower curve)
from a mixture of HS-[ CuC]
2
Fc and HS(CH
2
)
9
OH. The qualitative
dierence in the two transients is caused by the dierence in the pzr of
the diluents (see Fig. 3, which shows that the pzr for the methyl- and hy-
droxy-terminated thiols diers by nearly 2 V). But that is not the only dif-
ference. There is also a dierence in the E
o/
(0.375 V vs. SSCE for the
ILIT Method 163
methyl-terminated diluent and 0.300 V vs. SSCE for the hydroxy-termi-
nated diluent). A subtler eect is that dEj/dT = 0.00027 V/K for the
methyl-terminated diluent and dEj/dT = 0.0 V/K for the hydroxy-ter-
minated diluent. Both responses in Fig. 14 are for the condition V
oc
i
= E
oi
,
giving k
m
= 2.2 10
7
s
1
for the upper curve and k
m
= 1.2 10
7
s
1
for
the lower curve [k
m
values are computed from Eq. (90)]* A full set of k
m
values obtained for dierent values of V
oc
i
- E
oi
allows us to deduce kj =
1.7 10
6
s
1
and c
i
= 21.9 for the upper curve and kj = 1.6 10
6
s
1
and
c
i
= 9.6 for the lower curve. Thus, the values of kj are immune to the
dierences in the composition of the SAMs and to the induced dierences
in the thermodynamic properties. The analysis of the electron transfer
kinetics of these oligophenyleneethynylene tethers has been reported ear-
lier [2].
B. Determining the Value of kjjj
In Sec. IV.E we mentioned two ways
y
to extract the value of the standard
rate constant, k
o
, from k
m
-vs.- V
oc
i
dataone approach is to use the
Butler-Volmer-based expression [Eq. (90)] and determine kj and, in most
cases, c
i
[the dimensionless surface coverage; see Eq. (57)]; the other way is
to use Marcus-nonadiabatic (MNA)based expression [Eq. (89)], which
requires knowing the value of the reorganization energy, k. In Fig. 15 we
show a set of k
m
-vs.- V
oc
i
data. Open circles were collected at T = 298.2 K
for a SAM formed from HS-[ CH=CH-]
2
Fc and HS(CH
2
)
9
CH
3
(see
Ref. 3 for a more complete discussion of the behavior of ferrocene tethered
to a gold electrode by the oligophenylenevinylene bridge). The dotted line
is the best t obtained using the Butler-Volmerbased expression [Eq. (90)]
with kj = 3.8 10
6
s
1
and c
i
= 79.6. The solid line is the best t obtained
using the Marcus-nonadiabatic (MNA)based expression [Eq. (89)] as-
suming k = 0.93 eV with kj = 4.29 10
6
s
1
and c
i
= 71. The value of c
i
determined from the cyclic voltammogram is 85. One might argue that the
dierence between the value of c
i
obtained directly from the CV (85) and
the value obtained from the ILIT data (71) using Eq. (82) is signicant.
That is possible. However, virtually all our analyses suggest that dier-
* At the time this work as carried out, we obtained the value of c
i
from the cyclic
voltammetric curves and obtained k
o
directly from Eq. (90).
y
In principle there is a third approach: estimate the value of c
i
from the cyclic
voltammetry and then use that value to evaluate kj from the ILIT response. The
danger with this approach is that the ILIT and CV experiments may not be sam-
pling the same redox population.
Feldberg et al. 164
ences in the values of c
i
determined in the dierent ways are not critical.
Using the MNA-based expression is problematic since it requires a
reasonable estimate of the reorganization energy, k, and that may not be
known. The least ambiguous evaluation of k is attained by a method
involving a large overpotential, e.g., cyclic voltammetry [30,109,110] or
chronoamperometry [29,50,93,111]. The slope of an Arrhenius plot reveals
the activation enthalpy, DH
p
, which is directly related to the reorganiza-
tion enthalpy, whose value may or may not be the same as k [see Eq. (24)
and sec. III.B]. We discuss an example in the next section.
p p
and DH
kkkkkk
In Fig. 16 we show the Arrhenius plots for SAMs formed from an aliphatic
diluent and one of two ferrocene moieties (A and B, see Ref. 3 for further
details):
The slopes and the rate constants are virtually identical, indicating that the
dierent bridge substituents do not modify either the rate constants or the
temperature dependencies
*
. The slope of this plot is 2.7 (F0.1) 10
3
K
corresponding to DH
p
=0.23 eV. If we assume the simple relationship
in Eq. (25), i.e., that DH
k
=4 DH
p
, we estimate that DH
k
= 0.92 eV.
Chidsey has reported values of k
f
0.85 eV for Fc attached to gold elec-
trodes by aliphatic tethers which are long enough to eect electron transfer
that is slow enough to permit facile chronoamperometric evaluation of k
*
Data for SAMs formed from Fc attached through oligophenyleneethynylene
tethers show some dramatic dierences that might be attributed to dierent sub-
stituents (e.g., compare the results of Sachs et al. [2] and Creager et al. [112]).
ILIT Method 165
[29]. Using this value of k and assuming a nominal temperature of
f
300 K
we can use Eq. (24) and estimate DH
k
=0.92 eV, essentially the same value
obtained using the simpler expression. Since DH
k c
k (within experimental
error) we conclude that DS
k c0. Weber and Creager [30], however, suggest
that there are measurable, albeit small, dierences between the values of
DH
k
and k; they conclude that DS
kc1.0 10
4
eV K
-1
. At T = 300 K
TDS
k
=0.03 eV, which is too small for us to detect.
VII. THE POTENTIAL OF THE ILIT APPROACH
We have already alluded to some of the advantages of the ILIT ap-
proachnotably that fast (nanosecond time domain) interfacial kinetics
can be measured at a macroelectrode. Even without a pump-probe
approach, we believe that our latest (but unfortunately unused) instru-
mentation can achieve 1 ns time resolution, and possibly a bit faster than
that. An eective optical probe sensitive enough to measure the small
changes in interfacial voltage and/or composition produced by the ILIT
perturbation could couple with ILIT to create a pump-probe technique
that would eect a dramatic improvement in time resolution. Ultimately
other limiting factors could come into play, e.g., the time constants for
double layer and Soret relaxations, for thermal transport through the
SAM, and for the thermal response of the interfacial capacitance. Theo-
retical calculations suggest that double layer and Soret relaxations will be
of the order of 10100 ps (in 1 M supporting electrolyte), but their
amplitudes may be small enough so that these relaxations can be ignored
(see Sec. V.E). However, the time constant for the thermal response of the
lmitself (the termA; see Sec. IV) is unknown. We have seen some evidence
of a slow response with a SAM formed from 11-mercaptoundecanoic acid
[113], but we have seen no evidence of such a relaxation with methyl- or
hydroxy-terminated thiols in the low ns time domain.
VIII. SOME THOUGHTS ABOUT FUTURE
EXPERIMENTS
The ILIT program at Brookhaven National Laboratory has been termi-
nated. We have demonstrated that our best instrumentation can now at-
tain nanosecond and possibly subnanosecond time resolution (Sec. IV.E;
Fig. 6). The holy grail for those who study ultrafast interfacial kinetics
remains the development of a pump-probe technique whose time resolu-
tion would be limited only by the operative physical chemical processes (see
Sec. V.E). We did not achieve that goal, but ILIT could be a component in
Feldberg et al. 166
combination with a yet-to-be-discovered optical probe which could sense
the small ILIT-induced changes in the interfacial potential. Nevertheless,
the present capability,
f
1 ns time resolution, could address some of the
following issues.
The response times of neat SAMs in the absence of a redox moiety: Our
theoretical analysis for the ILIT response has presumed that dC
i
/dt is
innitely fastif that is not the case, extracting meaningful values of k
m
will be dicult, if not impossible. We have already mentioned evidence of a
slow response of SAMs formed from 11-mercaptoundecanoic acid [113]
(Sec. VII). We expect that the results of these types of experiments will be
critically dependent upon the choice of the SAM constituent, the SAM
preparation, substrate metal, temperature, the solvent, and the electrolyte
ions (the less hydrophobic, the better). Athorough study of the potential of
zero response would be important and informative. Establishing which
lms exhibit the fast responses will be critical for any meaningful studies of
fast interfacial electron-transfer kinetics.
Electron-transfer kinetics using SAMs in nonaqueous solvent systems
(with obvious concern about the stability of SAMs in these solvents): The
decrease in the dielectric constant of the solvent coupled with a likely
decrease in the reorganization energy and a likely increase in the accessible
temperature range should yield some valuable information.
Electron-transfer kinetics using SAMs comprising a tethered redox
species and any of a variety of saturated and unsaturated diluents: The role
of the nature of the diluent deserves further study. Finklea and coworkers
have examined the eects of the lengths of the diluent molecules relative
to the length of the tethered redox moiety [77], but there has been no
study of the eect of using unsaturated diluents that might signicantly
improve electron coupling. We would expect the most dramatic eect
with an aliphatically tethered redox moiety and an unsaturated (e.g.,
oligophenyleneethynylene or oligophenylenevinylene) diluent [114] since
we have demonstrated that oligophenylenevinylene linkages are consid-
erably more eective as an electron conduit than aliphatic chains (see
Refs. [13]).
Eects of changes in the solution viscosity on the value of kj: In Sec.
V.E we alluded to several studies that suggest that an increase in viscosity
causes a decrease in the value of kj (units: cm/s) for redox species in bulk
solution [32,54,103,104]. No such eects have been noted in the analogous
preliminary studies of surface attached ferrocene species [105,106]. This
could be because the kj values for the systems studied were too small. A
search for a viscosity eect on larger kj values using ILIT (with the
improved response function) could be informative. The increase in solu-
ILIT Method 167
tion resistance that accompanies an increase in viscosity poses no serious
problem for ILIT; however, the time resolution of conventional electro-
chemical methods will be severely compromised by an increase in the
uncompensated resistance. Scanning electrochemical microscopy (SECM)
may be one of the best ways to minimize the eects of uncompensated
resistance [5254], and values of kj greater than 1 cm/s have been measured
for the ferrocenium/ferrocene couple in bulk solution [54]. Miao et al. [54]
suggest that their observations are consistent with a decrease in the
longitudinal relaxation time (caused by an increase in viscosity).* If this
argument is correct one might expect to see the eect with Fc moieties that
are attached to an electrode surface only when the kj is suciently large
[115]. Since Miao et al. were seeing eects with kj of the order of cm/s for
bulk redox species, one would expect to see the eect when kj (for an
attached species) is of the order of 10
8
s
1
for attached redox systems (note
dierent units for kj with the dierent types of systems). One can argue
that kj (s
1
for attached redox species)
c
b kj (cm/s for bulk redox species)
where b
c
10
8
cm
1
[88]. The faster version of ILIT(with time resolution of
f
1 ns) should be able to measure values of kj greater than
f
10
8
s
1
, larger
than any values reported thus far for any attached system. The largest
value we measured was
f
5 10
7
s
1
using the slower ILIT system for
ferrocene attached to a gold electrode with a phenylenevinylene tether [3].
Electron-transfer kinetics studied using SAMs self-assembled on a
variety of noble metals in addition to gold: Such studies will address issues
associated with the density-of-states assumptions described in Sec. III.A
[93] and will probe how the dierent orbitals of these metals contribute to
the overall electronic coupling [116] as the coupling gets stronger (and the
electron-transfer rate constant gets faster) with shorter or more eective
(e.g., unsaturated) tethers.
Activation energy as a function of distance: We have observed that the
activation energy, DH
p
, (obtained from Arrhenius plots) decreases with
decreasing distance between the electrode and the redox moieties [1,117].
The nature of this decrease does not depend upon the nature of the tether
and occurs for both ferrocene and rutheniummoieties (Fig. 17). The results
* The decrease in the longitudinal relaxation cannot explain the dramatic
viscosity eects on considerably smaller values of kj for ferrocenium/ferrocene or
ferri/ferrocyanide systems as reported by Zhang et al. [103]. Some other eect or
artifact must be operative.
Feldberg et al. 168
agree only very qualitatively with the predictions of Liu and Newton [118],
which considered a three-phase system (metal/dielectric lm/solvent) and
assumed that the reorganization energy was due entirely to solvent and lm
dipole reorientation and image charges in the metal electrode and in the
lm. We know that strong coupling can also eect smaller values of DG
p
and arguably of DH
p
[119,120]. Measurements of DH
p
values for shorter
tethers (requiring measurement of faster electron-transfer rate constants)
may clarify these observations.
FIG. 17. DH
p
/DH
x=l
p
for (.) ILIT measurement of directly linked ferrocene
formed from HS(CH
2
)
n
Fc [117]; (E) ester-linked ferrocene formed from
HS(CH
2
)
n
COOFc [1,29]; (z) formed by reaction of previously self-assembled
mercaptocarboxylic acid monolayer with [4-aminomethylpyridine Ru(NH
3
)
5
]
2+
)
[37]; corresponding open symbols are for DH
p
obtained from chronoampero-
metric or cyclic voltammetric data [121123]. Solid and dashed curves are
theoretical values of DG*/DG*
x=l
for Fc and Ru moieties (see Ref. 118).
ILIT Method 169
IX. GLOSSARY OF TERMS
Term Units Denition
a cm
2
Electrode area
A V/K Coecient for the amplitude of the open-circuit
thermal response in absence of electron transfer
[see Eqs. (47) and (58)]
b
jk
Reduced diusional parameter [see Eq. (70)]
B V/K Coecient for the amplitude of the open-circuit
thermal response eected by an electron-transfer
relaxation [see Eqs. (48)(56) and (58)]
c cm/s Speed of light (3 10
10
cm/s)
C
i
F/cm
2
Interfacial capacitance
C
i
i
F/cm
2
Initial interfacial capacitance prior to the ILIT
perturbation
C
i
redox
F/cm
2
Initial redox capacitance prior to the ILIT
perturbation [see Eq. (87)]
C
i
i
F/cm
2
Initial interfacial capacitance prior to the ILIT
perturbation
C
dl
F/cm
2
Diuse layer capacitance
C
j
T
Jg
1
K
1
Heat capacity of the material in the j
th
ILIT
material (Quartz, water, gold, titanium)
C
Au
T/
Jg
1
K
1
Eective heat capacity of the eective gold layer
of thickness d
/
Au
[see Eqs. (94)(96)]
Dd
delay
cm Distance dierence in light paths of the pump and
probe laser pulses
d
Q
cm Thickness of quartz disk supporting the working
electrode
d
e
cm Diameter of the working electrode
DLIT Direct laser-induced temperature-jump method
D
T
j
cm
2
/s Diusion coecient of heat in the j
th
medium
e C Elementary charge (1.60218 10
19
C)
E
o/
V Formal potential of a redox couple
E
oi
V Initial formal potential of a redox couple prior to
the ILIT perturbation at T
i
E
pzc
V Potential of zero charge of a metal electrode, i.e.,
the potential when j
M
= 0
E
pzc
i
V Initial value of E
pzc
at T
i
prior to the ILIT
perturbation
ET Abbreviation for electron transfer
fwhm s Full-width-half-max: characteristic temporal width
of a laser pulse
Feldberg et al. 170
g(k*,x
i
) Ratios k
red
BV
/k
red
NA
and k
ox
BV
/k
ox
NA
[see Eq. (27)]
H
o
DA
eV Coupling constant for electron transfer
ILIT Indirect laser-induced temperature-jump method
j,k Indices indicating ILIT medium 1 (quartz),
2 (metal electrode), or 3 (solution)
k
B
J mol
1
K
1
or eV/K
Boltzmanns constant
k
m
s
1
Rate constant describing the ILIT relaxation
(see Secs. IV.D and IV.E)
k
o
s
1
Standard electron-transfer rate constant for a
surface-attached species
K
jk
Reduced heat capacity-density parameter [see
Eq. (71)]; subscripts denote the phase: quartz (1),
metal (2), solution (3)
lhs Referring to the left hand side of an equation or
portion of an equation
MBV Connotes relevance to the modied Butler-Volmer
equation with a voltage-dependent a [see
Eqs. (29) and (30)]
MNA Connoting relevance to the Marcus nonadiabatic
formalism (see Sec. III)
pzc Potential of zero charge (see Sec. IV.A)
pzr Potential of zero response (see Sec. IV.A)
q
T
J/cm
2
Heat deposited per unit area of electrode in an
ILIT experiment
Q cm
2
/F Composite variable [see Eq. (45)]
rhs Referring to the right-hand side of an
equation or portion of an equation
r
d
cm Radius of vapor deposited gold layer on quartz disk
r
hv
Fraction of light reected from electrode
surface
r
Q
cm Radius of quartz disk supporting the
working electrode
R
et
ohm Electron transfer resistance [see Eq. (83)]
R
et
i
ohm Electron transfer resistance just prior to the
ILIT perturbation.
R
t
Normalized response function [see Eqs. (74)
and (75)]
R
u
ohm Uncompensated resistance
S V/K Composite variable [see Eq. (44)]
SECM Scanning electrochemical microscopy
SSCE Saturated sodium calomel electrode
(SSCE = 0.236 V vs. the normal
hydrogen electrode)
ILIT Method 171
Dt
delay
s Time delay in pump-probe method
Dt
L
s Repitition rate of laser in multipulse
experiment (see Sec. V.F.)
T K Temperature
T
i
K Temperature prior to the ILIT perturbation
DT K Generalized change in the interfacial
temperature
DT
eq
K Change in temperature of the gold lm for
given q
T
if there were no losses of heat to
the adjacent media [see Eq. (68)]
DT
t,ideal
K Ideal change in the interfacial temperature [see [Eqs.
(66) and (67)] and accompanying discussion]
DT
t,ideal
* DT
t,ideal
/DT
eq
DT
t
,
e
K Convolution of DT
t,ideal
with instrument
response function, R
t
[see Eq. (74) and
accompanying discussion]
DT
t,e
* DT
t,e
/DT
eq
ume Ultramicroelectrode
DV
A
V Portion of acoustic response selected for
calibration of DT (see Fig. 13)
V
oc
V Open-circuit potential of the (metal)
working electrode vs. SHE
V
oc
i
V The value of V
oc
at T
i
prior to the ILIT
perturbation
V
TJ
V Sum of potentials eected by temperature
dierences within the system including
the Soret potential created by the
temperature dierence between the
working and reference electrodes
V
TJ
i
V The initial value of V
TJ
at T
i
prior to the
ILIT perturbation; V
TJ
i
= 0 since the
system is initially isothermal
v
sound
cm/s Speed of sound in H
2
O: 1.5 10
5
cm/s
x cm Length of the thiolate tether attaching the redox
moiety to the electrode, dened as the shortest
distance between the carbon attached to the
surface and the atom of the redox moiety that is
linked to the tether
a Transfer coecient in Butler-Volmer equations
[see Eq. (26)]
a
j
T
Fraction of heat going into the j
th
phase
(see Appendix)
Feldberg et al. 172
X. APPENDIX: ONE-DIMENSIONAL THERMAL
DIFFUSION INTO TWO DIFFERENT PHASES
Assume that a quantity of heat, q
T
, is initially deposited as a spatial and
temporal delta function at x = 0 and t = 0. The subsequent one-
dimensional diusion of heat into a homogeneous medium and the re-
sultant temperature change as a function of x (distance from the position
a
MBV
Voltage dependent transfer coecient in modied
Butler-Volmer equations [see Eqs. (29) and (30)]
b cm
1
Parameter dening the exponential distance
dependence of the rate of electron transfer
[see Eq. (12)]
d cm Thickness of metal lm
c
i
Initial preperturbation value of the dimensionless
surface coverage [see Eq. (57)]
K
j
T
J s
1
cm
1
K
1
Thermal conductivity in the j
th
medium
q ohm cm Solution resistivity
q
j
g/cm
3
Density of j
th
medium
q eV
1
Density of electronic states in a metal
r
jk
Reduced thermal parameter [see Eq. (69)]
r
M
C/cm
2
Charge density on metal electrode
s
ILIT
s Duration of an ILIT experiment: typically less
than 10
6
s
s
T
s Transit time for heat to diuse from electrode
surface to redox species (see Sec. V.E)
s
rise
s Rise-time of interfacial temperature [see Eq. (73)]
C
ox
molecules/cm
2
Surface concentration of oxidized species
C
red
molecules/cm
2
Surface concentration of reduced species
C
ox
i
molecules/cm
2
Initial value of C
ox
at T
i
, i.e., prior to the ILIT
perturbation
C
red
i
molecules/cm
2
Initial value of C
red
at T
i
, i.e., prior to the ILIT
perturbation
C
total
molecules/cm
2
Total surface concentration of the redox species
k eV Reorganization energy
k
hv
nm Wavelength of laser pulse
v
el
s
1
Electron-transfer frequency
x exp[e/k
B
T
i
) (V
oc
E
o/
)]
x
i
Initial (preperturbation) value of
exp[(e/k
B
T
i
)(V
oc
i
- E
o/i
)]
ILIT Method 173
of the initially deposited energy) and t (time after energy deposition) is
described by the familiar equation [101]:
DT
t;x

q
T
2q
j
C
T
j
pD
T
j
t
exp
x
2
4D
T
j
t

115
where q
j
, C
j
T
, and D
j
T
are the density (g/cm
3
), heat capacity (J g
1
K
1
), and
thermal diusion cocient (cm
2
/s) of the j
th
phase, respectively. When
there are dierent phases on each side of the initial delta function we can
write instead:
DT
t;xV0

1 a
T
k
q
T
q
j
C
T
j
pD
T
j
t
exp
x
2
4D
T
j
t

and 116
DT
t;xz0

a
T
k
q
T
q
k
C
T
k
pD
T
k
t
exp
x
2
4D
T
k
t

where a
k
T
is the fraction of the heat diusing into the k
th
phase and 1 a
k
T
is the fraction of the heat diusing into the j
th
phase. Note that
q
j
C
T
j
0
l
DT
t;x V0
1 a
T
k
q
T
and 117
q
k
C
T
k
l
0
DT
t;x z0
a
T
k
q
T
indicating that the fraction of heat deposited into each phase is indepen-
dent of time. Since we are interested in the temperature at x = 0, we can
simplify Eqs. (116) to:
DT
t;x0

1 a
T
k
q
T
q
j
C
T
j
pD
T
j
t

a
T
k
q
T
q
k
C
T
k
pD
T
k
t
118
We can now solve for a
k
T
:
a
T
k

1
1 r
T
j=k
119
Feldberg et al. 174
where
r
T
j=k

q
j
C
T
j
D
T
j
q
k
C
T
k
D
T
k
120
Note that when r
j/k
T
= 1 then a
k
T
= 1/2 and Eq. (118) becomes identical to
Eq. (115) when x = 0.
ACKNOWLEDGMENTS
We are particularly indebted to Professor Christopher Chidsey at Stanford
University for his enthusiastic support and colleagueship throughout the
course of this project; several of his students deserve special mention for
their contributions over the course of this eort: Matthew Linford, Sandra
Sachs, Hadley Sikes, and Stephen Dudek. We also thank Harry Finklea,
Stephen Creager, Tal Nahir, and Edmond Bowden for graciously pro-
viding compounds and for their collaboration; Johna Leddy, Lin Geng,
David Reddy, Keli Chalfant, and Rory MacFarquhar are thanked for
their contributions in the early stages of this work. Andrew Cook is grate-
fully acknowledged for his assistance in experiments using the titanium-
sapphire laser in Brookhavens Laser Electron Accelerator Facility. At the
outset of this project discussions with Hans Coufal were invaluablehis
seminal work demonstrated that the ILIT concept would work. Numer-
ous conversations with Dwayne Miller were enlightening and very helpful.
The collaboration with Brookhaven National Laboratorys Instrumen-
tation Division has been a critical aspect of this eort; in particular, Lee
Rogers and Sergio Rescia are thanked for their contributions. The U.S.
Department of Energy, Contract no. DE-AC02-98CH10886, is thanked
for their support. Finally, JFS and SWF gratefully acknowledge the cour-
tesy extended to us by Creighton Wirick, Chair, Environmental Chemistry
Department.
REFERENCES
1. Smalley, J.F.; Feldberg, S.W.; Chidsey, C.E.D.; Linford, M.R.; Newton,
M.D.; Liu, Y.-P. J. Phys. Chem. 1995, 99, 13141.
2. Sachs, S.B.; Dudek, S.; Hsung, R.P.; Sita, L.R.; Smalley, J.F.; Newton,
M.D.; Feldberg, S.W.; Chidsey, C.E.D. J. Am. Chem. Soc. 1997, 119, 10563.
ILIT Method 175
3. Sikes, H.D.; Smalley, J.F.; Dudek, S.P.; Feldberg, S.W.; Newton, M.D.;
Chidsey, C.E.D. Science 2001, 291, 1519.
4. Smalley, J.F.; Kirshnan, C.V.; Goldman, M.; Feldberg, S.W.; Ruzic, I. J.
5. Smalley, J.F.; MacFarquhar, R.A.; Feldberg, S.W. J. Electroanal. Chem.
1988, 256, 21.
6. Smalley, J.F.; Geng, L.; Rogers, L.C.; Feldberg, S.W.; Chidsey, C.E.D.;
Leddy, J. J. Electroanal. Chem. 1993, 356, 181.
7. Gray, H.B.; Winkler, J.R. Ann. Rev. Biochem. 1996, 65, 537.
8. Newton, M.D.; Smalley, J.F.; Feldberg, S. W. Interfacial Electrochemistry;
Wieckowski, A., Ed.; Marcel Dekker: New York, 1999; 97 p.
9. Petty, M.C.; Bryce, M.R.; Bloor, D. An Introduction to Molecular Elec-
tronics, Oxford University Press: New York, 1995.
10. Jortner, J.; Ratner, M.A. Molecular Electronics; Blackwell Science Ltd:
Oxford, 1997.
11. Reed, M.A.; Tour, J.M. Sci. Am. 2000, 282, 86.
12. Nitzan, A. J. Phys. Chem. A 2001, 105, 2677.
13. Reed, M.A.; Zhou, C.; Muller, C.J.; Burgin, T.P.; Tour, J.M. Science 1999,
278, 252.
14. Chen, J.; Reed, M.A.; Rawlett, A.M.; Tour, J.M. Science 1999, 286, 1550.
15. Slowinski, K.; Chamberlain, R.V.; Bilewicz, R.; Majda, M. J. Am. Chem.
Soc. 1996, 118, 4709.
16. Slowinski, K.; Chamberlain, R.V.; Miller, C.J.; Majda, M. J. Am. Chem.
Soc. 1997, 119, 11910.
17. Slowinski, K.; Fong, H.K.Y.; Majda, M. J. Am. Chem. Soc. 1999, 121,
7527.
18. Slowinski, K.; Slowinska, K.; Majda, M. J. Phys. Chem. B 1999, 103, 8544.
19. Slowinski, K.; Majda, M. J. Electroanal. Chem. 2000, 491, 139.
20. Diau, E.W.-G.; Herek, J.L.; Kim, Z.H.; Zewail, A.H. Science 1998, 279,
847.
21. Yokoyama, K.; Silva, C.; Son, D.H.; Walhout, P.K.; Barbara, P.F. J. Phys.
Chem. A 1998, 102, 6957.
22. Lanzafame, J.M.; Palese, S.; Wang, D.; Miller, R.J.D. J. Phys. Chem. 1994,
98, 43.
23. Smalley, J.F.; Newton, M.D.; Feldberg, S.W. Electrochem. Commun. 2000,
2, 832.
24. Smalley, J.F.; Geng, L.; Chen, A.; Feldberg, S.W.; Lewis, N.S.; Cali, G. J.
25. Schreiber, F. Prog. Surf. Sci. 2000, 58, 151.
26. Li, T.T.-T.; Weaver, M.J. J. Am. Chem. Soc. 1984, 106, 6107.
27. Miller, J.R. Proceedings in Life Sciences: Protein Structure Molecular and
Electronic Reactivity; Austin, R., Buhks, E., Chance, B., de Vault, D.,
Feldberg et al. 176
Dutton, P.L., Frauenfelder, H., Goldanskii, V.I., Eds.;Springer-Verlag:
New York, 1987; 329 pp.
28. Closs, G.L.; Calcattera, L.T.; Green, N.J.; Peneld, K.W.; Miller, J.R. J.
Phys. Chem. 1986, 90, 3673.
29. Chidsey, C.E.D. Science 1991, 215, 919.
30. Weber, K.; Creager, S.E. J. Electroanal. Chem. 1997, 458, 17.
31. Weber, K.; Hockett, L.; Creager, S.E. J. Phys. Chem. B 1997, 101, 8286.
32. Koshtariya, D.E.; Dolidze, T.D.; Zusman, L.D.; Waldeck, D.H. J. Phys.
Chem. A 2001, 105, 1818.
33. Finklea, H.O.; Hanshew, D.D. J. Am. Chem. Soc. 1992, 114, 3173.
34. Finklea, H.O. Electoanalytical Chemistry; Bard, A.J., Rubinstein, I., Eds.;
Marcel Dekker: New York, 1996; Vol. 19, 110 p.
35. Brevnov, D.A.; Finklea, H.O. J. Electroanal. Chem. 2001, 488, 133.
36. Finklea, H.O. J. Electroanal. Chem. 2001, 495, 79.
37. Brevnov, D.A.; Finklea, H.O.; Van Ryswyk, H. J. Electroanal. Chem.
2001, 500, 100.
38. Czerlinski, G.; Eigen, M. Z. Elektrochem. 1959, 63, 652.
39. Eigen, M.; DeMaeyer. Techniques in Organic Chemistry; Weissberger, A.,
Ed.; Interscience: New York, 1963; Vol. III, Pt. II, 969 pp.
40. Eigen, M. Nobel Symposium #5; Claesson, S., Ed. Interscience: New York,
1967; 333 pp.
41. Forster, R.J. Encyclopedia of Analytical Chemistry; Myers, R.A., Ed.; J.
Wiley & Sons Ltd.: Chichester, 2000; 10142 pp.
42. Bard, A.J.; Faulkmer, L.R. Electrochemical Methods:Fundamentals and
Applications; 2nd Ed.; John Wiley and Sons: New York, 2001.
43. Barker, G.C. Trans. Symp. Elecrode Proc., Philadelphia, 1959; Yeager, E.,
Ed.; Wiley: New York, 1961.
44. Delahay, P. Anal. Chem. 1962, 34, 1267.
45. Reinmuth, W.H. Anal. Chem. 1962, 34, 1272.
46. Newman, J. J. Electrochem. Soc. 1966, 113, 501.
47. Brumlik, C.J.; Martin, C.R.; Tokuda, K. Anal. Chem. 1992, 64, 1201.
48. Wightman, R.M.; Wipf, D.O. Electrochemical Chemistry; Bard, A.J., Ed.;
Marcel Dekker: New York, 1989; 267 pp.
49. Morris, R.; Franta, D.J.; White, H.S. J. Phys. Chem. 1987, 91, 3559.
50. Robinson, D.B.; Chidsey, C.E.D. J. Phys. Chem. B 2002, 106, 10706.
51. Kirchner, V.; Xia, X.; Schuster, R. Acc. Chem. Res. 2001, 34, 371.
52. Mirkin, M.V.; Richards, T.C.; Bard, A.J. J. Phys. Chem. 1993, 97, 7672.
53. Bard, A.J.; Fan, F.R.; Mirkin, M.V. Electroanalytical Chemistry; Bard,
A.J., Ed.; New York: Marcel Dekker 1994; Vol.18, 243 pp.
54. Miao, W.; Ding, Z.; Bard, A.J. J. Phys. Chem. B 2002, 106, 1392.
55. Gerisher, H. Z. Elektrochem. 1960, 64, 86.
56. Strehlow, H.; K.S. Ber. Bunsenges 1966, 70, 139.
ILIT Method 177
57. Homann, H.; Yeager, E.; Stuehr, J. Rev. Sci. Instrum. 1968, 39, 649.
58. Eyring, E.M.; Bennion, B.C. Ann. Rev. Phys. Chem. 1968, 19, 129.
59. Beitz, J.V.; Flynn, G.W.; Turner, D.H.; Sutin, N. J. Am. Chem. Soc. 1970,
94, 4130.
60. Turner, D.H.; Flynn, G.W.; Sutin, N. J. Am. Chem. Soc. 1972, 94, 1554.
61. Dyer, R.B.; Gai, F.; Woodru, W.H.; Gilmanshin, R.; Callender, R.H.
Acc. Chem. Res. 1998, 31, 709.
62. Harima, Y.; Aoyagui, S. J. Electroanal. Chem. 1976, 69, 419.
63. Gabrielli, C.; Keddam, M.; Lizee, J.F. J. Electroanal. Chem. 1983, 148, 293.
64. Gabrielli, C.; Keddam, M.; Lizee, J.F. J. Electroanal. Chem. 1993, 359, 1.
65. Grundler, P.; Zerihun, T.; Kirbs, A.; Grabow, H. Anal. Chim. Acta 1995,
305, 232.
66. Babenko, S.D.; Benderskii, V.A.; Zolotovitskii, Y.M.; Krivenko, A.G. J.
67. Benderskii, V.A. J. Electroanal. Chem. 1977, 76, 223.
68. Benderskii, V.A.; Babenko, S.D.; Krivenko, A.G. J. Electroanal. Chem.
1978, 86, 223.
69. Benderskii, V.A.; Velichko, G.I. J. Electroanal. Chem. 1982, 140, 1.
70. Climent, V.; Coles, B.A.; Compton, R.G. J. Phys. Chem. B 2002, 106,
5258.
71. Climent, V.; Coles, B.A.; Compton, R.G. J. Phys. Chem. B 2002, 106, 5988.
72. Barker, G.C.; Gardner, A.W.; Bottura, G.J. Electroanal. Chem. 1973, 45,
21.
73. Cao, J.; Elsayed-Ali, H.E.; Miller, R.J.D.; Mantell, D.A.; Gao, Y. Phys.
Rev. B 1998, 58, 10948.
74. Miller, B. J. Electrochem. Soc. 1983, 130, 1639.
75. Coufal, H. Appl. Phys. Lett. 1983, 44, 59.
76. Coufal, H.; Heerle, P. Appl. Phys. Lett. A 1985, 38, 213.
77. Finklea, H.O.; Liu, L.; Ravenscroft, M.S.; Punturi, S. J. Phys. Chem. B
1996, 100, 18852.
78. Smith, C.P.; White, H.S. Anal. Chem. 1992, 64, 2398.
79. Marcus, R.A. J. Chem. Phys. 1956, 24, 979.
82. Marcus, R.A. Am. Rev. Phys. Chem. 1964, 15, 155.
84. Marcus, R.A. J. Phys. Chem. B 1998, 10071.
85. Marcus, R.A. Disc. Faraday Soc. 1960, 29, 21.
86. Hsu, C.-P.; Marcus, R.A. J. Chem. Phys. 1997, 106, 584.
87. Hsu, C.-P. J. Electroanal. Chem. 1997, 438, 27.
88. Royea, W.J.; Fajardo, A.M.; Lewis, N.S. J. Phys. Chem. B 1997, 101,
11152.
Feldberg et al. 178
89. Johnson, M.D.; Miller, J.R.; Green, N.S.; Closs, G.L. J. Phys. Chem. 1989,
93, 1173.
90. Chattoraj, M.; Paulson, B.; Shi, Y.; Closs, G.L.; Levy, D.H. J. Phys. Chem.
1993, 97, 13046.
91. Yi-Ping Liu, private communication.
92. Nahir, T.M. J. Electroanal. Chem. 2002, 518, 47.
93. Forster, R.J.; Loughman, P.; Keyes, T.E. J. Am. Chem. Soc. 2000, 122,
11948.
94. J. Leddy, private communication.
95. J. Leddy, private communication.
96. 3150 Central Expressway, Santa Clara, CA, 95051-0816.
97. Lin, J.; Brek, W.G. Can. J. Chem. 1965, 43, 229.
98. Koller, K.B.; Hawkridge, F.M. J. Am. Chem. Soc. 1985, 107, 7412.
99. Buck, R.P. J. Electroanal. Chem. 1969, 23, 219.
100. Feldberg, S.W. J. Phys. Chem. 1970, 74, 87.
101. Moore, W.J. Physical Chemistry. Prentice Hall, Inc: Englewood Clis, NJ,
1972.
102. Weston, R.E.; Schwarz, H.A. Chemical Kinetics; Prentice-Hall Interna-
tional Inc.: London, 1972.
103. Zhang, X.; Leddy, J.; Bard, A.J. J. Am. Chem. Soc. 1985, 107, 3719.
104. Williams, M.E.; Crooker, J.C.; Pyati, R.; Lyons, L.J.; Murray, R.W.
J. Am. Chem. Soc. 1997, 119, 10249.
105. J.F. Smalley and S.E. Creager, unpublished data.
106. Creager, S.E.; Weber, K.A. Proceedings of the Symposiumon NewDirections
in Electroanalytical Chemistry; Leddy, J., Wightman R.M., Eds.; The
Electrochemical Society: Pennington NJ., 1996.
107. Eisenberg, D.; Kauzmann, W. The Structure and Properties of Water.
Oxford University Press: New York, 1969.
108. Hass, G.; Hadley, L. American Institute of Physics Handbook; Gray, D.E.
Ed.; McGraw-Hill Inc.: New York, 1972; 6118 to 6161.
109. Tender, L.; Carter, M.T.; Murray, R.W. Anal. Chem. 1994, 66, 3173.
110. Carter, M.T.; Rowe, G.K.; Richardson, J.N.; Tender, L.; Terrill, R.H.;
Murray, R.W. J. Am. Chem. Soc. 1995, 117, 2896.
111. Forster, R.J.; Faulkner, L.R. J. Am. Chem. Soc. 1994, 116, 5444.
112. Creager, S.E.; Yu, C.J.; Bamdad, C.; OConnor, S.; MacLean, T.; Lam, E.;
Chong, Y.; Olsen, G.T.; Luo, J.; Gozin, M.; Kayhyem, J.F. J. Am. Chem.
Soc. 1999, 121, 1059.
113. Smalley, J.F.; Chalfant, K.; Feldberg, S.W.; Nahir, T.M.; Bowden, E.F.
J. Phys. Chem. B 1999, 103, 1676.
114. Jack Fajer, private communication.
115. Hynes, J.T. J. Phys. Chem. 1986, 90, 3701.
116. Gosavi, S.; Marcus, R.A. J. Phys. Chem. B 2000, 104, 2067.
ILIT Method 179
117. Smalley, J.F.; Finklea, H.O.; Chidsey, C.E.D.; Linford, M.R.; Creager, S.E.;
Ferraris, J.P.; Chalfant, K.; Zawodzinski, T.; Feldberg, S.W.; Newton, M.D.
J. Am. Chem. Soc. 2003, 128, 2004.
118. Liu, Y.P.; Newton, M.D. J. Phys. Chem. 1994, 98, 7162.
119. Sutin, N. Prog. Inorg. Chem. 1983, 30, 441.
120. Brunschwig, B.S.; Sutin, N. Coord. Chem. Rev. 1999, 187, 233.
121. Stephen Creager, private communication.
122. Christopher Chidsey, private communication.
123. Ravenscroft, M.S.; Finklea, H.O. J. Phys. Chem. 1994, 98, 3843.
Feldberg et al. 180
ELECTRICALLY CONDUCTING DIAMOND
THIN FILMS: ADVANCED ELECTRODE MATERIALS
FOR ELECTROCHEMICAL TECHNOLOGIES
Greg M. Swain
Michigan State University
East Lansing, Michigan, U.S.A.
I. INTRODUCTION 182
II. DIAMOND THIN FILM DEPOSITION, ELECTRODE
ARCHITECTURES, SUBSTRATE MATERIALS,
AND ELECTROCHEMICAL CELLS 185
III. ELECTRICAL CONDUCTIVITY OF DIAMOND
ELECTRODES 194
IV. CHARACTERIZATION OF MICROCRYSTALLINE
AND NANOCRYSTALLINE DIAMOND THIN
FILM ELECTRODES 195
V. BASIC ELECTROCHEMICAL PROPERTIES OF
MICROCRYSTALLINE AND NANOCRYSTALLINE
DIAMOND THIN FILM ELECTRODES 201
VI. FACTORS AFFECTING ELECTRON TRANSFER
AT DIAMOND ELECTRODES 212
VII. SURFACE MODIFICATION OF DIAMOND MATERIALS
AND ELECTRODES 216
VIII. ELECTROANALYTICAL APPLICATIONS 219
A. Azide Detection 219
B. Trace Metal Ion Analysis 221
C. Nitrite Detection 224
D. NADH Detection 225
E. Uric Acid Detection 225
F. Histamine and Serotonin Detection 226
G. Direct Electron Transfer to Heme Peptide and
Peroxidase 227
181
H. Cytochrome c Analysis 228
I. Carbamate Pesticide Detection 228
J. Ferrocene Analysis 229
K. Aliphatic Polyamine Detection 230
IX. ELECTROSYNTHESIS AND ELECTROLYTIC WATER
PURIFICATION 238
X. OPTICALLY TRANSPARENT ELECTRODES FOR
SPECTROELECTROCHEMISTRY 239
XI. ADVANCED ELECTROCATALYST SUPPORT
MATERIALS 251
A. Composite Electrode Fabrication and
Characterization 252
B. Oxygen Reduction Reaction 259
C. Methanol Oxidation Reaction 264
XII. CONCLUSIONS 267
REFERENCES 268
I. INTRODUCTION
Through the centuries, no material has been as cherished or treasured as
diamond. The outstanding combination of properties has long made it a
desirable material for mechanical, optical, and electronic applications.
Nature, of course, produces diamond under conditions of extreme pressure
and temperature. Driven by greed as well as the challenge, scientists for
more than a century have attempted diamond synthesis. The fascinating
history of this eort is eloquently chronicled by Robert M. Hazen in The
Diamond Makers. Attempts to produce diamond began back in the 1820s,
and up until the 1950s there were really no qualied successes. This was
due, in part, to the fact that researchers attempted to produce diamond
under conditions of only high pressure, failing to recognize the need for
high temperature.
Eorts to synthesize diamond took on added urgency in the United
States during the late 1940s and early 1950s as cold war tensions gripped
the nation. Diamond was needed for machining and polishing the carbide
tools required to fabricate components for war machinery and military
hardware. At that time the country was dependent primarily on imported
diamond from South Africa. The Norton Company, General Electric, and
Swain 182
Carborundum formed a consortium in the early 1940s with a goal of
producing diamond synthetically in order to establish a reliable state-side
supply of the technologically important material. From the mid-1940s to
the mid-1950s, the Norton Company led the synthetic eort combining
established high-pressure technology and know-how with high-tempera-
ture processing. For 10 years, scientists worked to grow diamond without
success. General Electric, during the 1950s, embarked on an ambitious
program to produce diamond, code-named Project Superpressure. The
research team met with success in 195455, an accomplishment that
required major advances in high-pressure and high-temperature materials
and devices, as well as new insights into the appropriate chemical envi-
ronment for growth. In 1957 the company began selling Man-Made
Diamonds, its trademarked synthetic diamond abrasive.
The equipment required for high-pressure and high-temperature
growth of diamond is expensive and massive in scale. Scientic inquiry
to produce diamond at low pressure using smaller and less costly equip-
ment began during the mid-1950s. William Eversole, a scientist at Union
Carbide, devised the rst reproducible, low-pressure synthesis of diamond
in 195758. Parallel eorts by the Derjaguin group in the former Soviet
Union and the Angus group at Case Western Reserve University in the
United States in the 1960s further demonstrated that low-pressure chem-
ical vapor deposition (CVD) growth of diamond was feasible. The work of
these researchers provided the foundation for modern-day diamond
synthesis at low pressures.
Vigorous research was conducted in the United States, the former
Soviet Union, Europe, and Japan, particularly during the 1970s, 1980s,
and into the 1990s. However, the Holy Grail application, high-power/high-
speed electronics, remained elusive to researchers, even to the present day.
The large breakdown voltage, high electron and hole mobilities, and high-
temperature stability are properties that make diamond attractive for this
use. Unfortunately, diculties with depositing low-defect diamond in a
cost-eective manner and controlled p- and n-type doping have prevented
this application from becoming a reality.
Diamond research in these areas continued to advance into the 1990s,
less so in the United States and more so in Europe and Japan. During this
period, the cost of producing diamond was signicantly reduced due to
improvements in the nucleation and growth rates. More was learned about
the mechanisms and kinetics of diamond synthesis. The use of diamond as
an electron emitter became an active area of research due to the negative
Electrically Conducting Diamond Thin Films 183
electron anity of properly treated diamond. Electrically conducting
diamond also began being tested in electrochemical technologies. The
dimensional stability and the chemical inertness are two properties that
stimulated the initial activity. A factor limiting the extent of research to
date has been the absence of a commercial supplier of diamond elec-
trodes. This situation has recently changed as there are now at least four
commercial entities marketing diamond electrodes worldwide.
No other material shows as much versatility as an electrode as does
electrically conducting, CVD diamond. The material can be used in
electroanalysis to provide low detection limits for analytes with superb
precision and stability; for high current density electrolysis (110 A/cm
2
) in
aggressive solution environments without any morphological or micro-
structural degradation; and as an optically transparent electrode (OTE) for
spectroelectrochemical measurements in the ultraviolet-visible (UV-Vis)
and infrared (IR) regions of the electromagnetic spectrum.
Electrically conducting diamond thin lm electrodes, fabricated by
CVD, provide scientists and engineers with a material that meets the
requirements for a wide range of applications. The eld has grown
signicantly over the past 10 years, as indicated by the increase in published
manuscripts, from 3 in 1993 to nearly 180 in 2002. The most signicant of
the published work has been in the elds of electroanalysis, electrolytic
water purication, and, more recently, as an OTE for spectroelectrochem-
istry. In the eld of electroanalysis, for instance, several papers have
demonstrated that diamond is useful for chemical measurements, oering
signicant advantages over other electrodes, especially carbon, in terms of
linear dynamic range, limit of detection, response time, response precision,
and response stability.
Diamond possesses a number of technologically important proper-
ties: chemical inertness, corrosion resistance, extreme hardness, wide
window of optical transparency, superb thermal conductivity, high sound
velocity transmission, etc. Diamond is also a wide bandgap material and is
normally a superb electrical insulator. However, when the material is
doped with impurities, such as boron, electrical conductivity is introduced
in a controlled manner. At doping levels below ca. 10
19
B/cm
3
, the material
behaves more like a semiconductor, and when doped at higher levels
behaves more like a semimetal. Electrical conductivities of 1001000 S/
cm can be introduced into the material, without compromising the other
unique properties, making it suciently conducting for many electrochem-
ical applications.
Swain 184
The purpose of this chapter is to acquaint the reader with the growth
and characterization of electrically conducting diamond, the basic prop-
erties and responsiveness of diamond electrodes, and some of the demon-
strated uses of this unique electrode material to date. The discussion is
conned to highly doped, semimetallic diamond electrodes and encom-
passes work from our group along with examples of published work by
other researchers. The chapter is organized according to the following
topics: (1) the basics of diamond lm deposition, as well as the electrode
architectures, substrate materials, and electrochemical cells commonly
used; (2) electrical conductivity of diamond electrodes; (3) characterization
of microcrystalline and nanocrystalline diamond thin lm electrodes; (4)
basic electrochemical properties of microcrystalline and nanocrystalline
diamond thin lm electrodes; (5) factors aecting electron transfer at
diamond electrodes; (6) surface modication of diamond materials and
electrodes; (7) electroanalytical applications; (8) electrosynthesis and
electrolytic water purication; (9) optically transparent diamond elec-
trodes for spectroelectrochemistry; and (10) advanced electrocatalyst
support materials.
II. DIAMOND THIN FILM DEPOSITION,
ELECTRODE ARCHITECTURES, SUBSTRATE
MATERIALS, AND ELECTROCHEMICAL CELLS
Diamond can be grown using one of several activation methods including
microwave plasma, hot-lament, or combustion ameassisted, CVD. The
most common method is microwave plasma, one reason being the com-
mercial availability of such reactors. While the mechanismof lmgrowth is
somewhat dierent from method to method, all serve to activate a car-
bonaceous source gas producing a carbon growth precursor in close
proximity to the substrate surface [1]. A CVD reactor consists of the
growth chamber and equipment associated with the particular activation
method (e.g., microwave power source), as well as various accessories, such
as mass ow controllers for regulating the source gases ows, a throttle
exhaust value and controller for regulating the system pressure, a pumping
system, temperature measurement capability, and the gas-handling system
for supplying the reaction gases. Figure 1 shows a block diagram of a
typical microwave plasma-assisted CVD reactor set-up. The microwaves
fromthe generator are directedtoandfocused ina quartz cavity, producing
a spherically shaped glow-discharge plasma directly above the substrate.
The substrate can either be positioned outside of (few mm) or immersed
within the intense discharge region. The plasma is where the reactive
growth precursor(s) is formed and transported to the substrate surface.
The key deposition parameters to control are the source gas composition,
microwave power, system pressure, and substrate temperature.
There are two types of synthetic diamond under investigation within
the scientic community: microcrystalline and nanocrystalline lms. Fig-
ure 2 shows SEM images of the two types of boron-doped lms deposited
on Si substrates. High-quality microcrystalline lms are deposited from
CH
4
/H
2
source gas ratios of 0.31.0%, microwave powers of 0.81 kW,
pressures of 3565 torr, and substrate temperatures of 700850jC. Under
these conditions, CH
3
.
is the primary growth precursor. The added
hydrogen serves several critical functions: (1) hydrogen abstraction from
CH
4
in the gas phase to produce CH
3
.
, (2) hydrogen abstraction from
chemisorbed methyl substituents on the surface to produce the active sites
for carbon addition, (3) passivation of dangling bonds on the surface to
minimize reconstruction to sp
2
bonding, and (4) rapid gasication of any
sp
2
-bonded, nondiamond carbon impurity that forms [1]. As can be seen
FIG. 1. Diagram of a typical microwave CVD reactor set-up.
Swain 186
FIG. 2. SEM images of boron-doped (A) microcrystalline and (B) nano-
crystalline diamond thin lms.
from the image, microcrystalline lms consist of well-faceted, polycrystal-
line diamond with crystallite sizes on the order of a few micrometers, or
greater, in lateral dimension. The crystallites are randomly oriented, and
there is signicant twinning. Of course, a grain boundary forms at the
junction between two or more crystallites.
Nanocrystalline lms are deposited from CH
4
/Ar source gas mix-
tures of 0.51.0% with little or no added hydrogen, microwave powers of
0.81 kW, pressures of 130160 torr, and substrate temperatures of 700
850jC. Gruen and coworkers discovered that phase-pure nanocrystalline
diamond can be grown from CH
4
/Ar gas mixtures with very little or no
added hydrogen [212]. The most remarkable dierence in lms grown
using hydrogen-poor Ar gas mixtures, compared with conventional hy-
drogen-rich mixtures, is the nanocrystallinity and smoothness (rms rough-
ness f1030 nmover large areas) of the former, as can be seen in the image.
Noble or inert gas addition to hydrogen-rich plasmas (e.g., 0.5% CH
4
/H
2
)
enhances the growth rate of lms by increasing the CH
3
.
and atomic
hydrogen in the plasma. The lowexcitation energy of inert gases, like Ar or
Xe, results in a plasma discharge with a higher electron density, and this
leads to the higher levels of CH
3
.
and H
.
. There is a fundamental change in
the plasma chemistry, growth process, growth rate, and lm properties as
one transitions from a hydrogen-rich to a hydrogen-poor plasma. In
particular, hydrogen-poor plasmas have high concentrations of C
2
, which
serves as both the growth and nucleation species. The nanocrystallinity is a
result of a growth and nucleation mechanism involving the insertion of
carbon dimer, C
2
, the primary growth precursor, into surface C-H bonds.
Apparently, there is sucient hydrogen from the CH
4
to minimize surface
reconstruction during growth to an sp
2
-bonded phase. The C
2
addition is
believed to occur by a two-step growth mechanism [2,3,6,7]. AC
2
molecule
approaches the unreconstructed monohydride surface and inserts into a C-
H bond. The C
2
molecule then rotates to insert its other carbon into a
neighboring C-H bond on the surface. A C
2
molecule then inserts into an
adjacent C-H bond, parallel to the newly inserted C
2
dimer. The original
state of the surface is recovered by the formation of a bond between carbon
atoms in the adjacent surface dimers. Very high rates of heterogeneous
renucleation are observed, on the order of 10
10
cm
-2
, and the resulting lms
consist of randomly oriented, phase-pure grains of diamond and well-
delineated grain boundaries [2,3,6,7].
Common substrate materials are Si, Mo, W, and Pt. Carbide-
forming materials tend to make good substrates for diamond growth.
Possessing a melting temperature well above the deposition temperature
Swain 188
and a coecient of thermal expansion similar to that of diamond are
important properties of the substrate. Generally, as long as the diamond
lm completely covers the substrate surface and the adhesion is good, the
electrochemical properties and responsiveness of diamond electrodes are,
for the most part, independent of the substrate material. The substrate
material could, however, inuence the electrode response of thin lms
(<1 Am) in at least three ways. First, pinholes and small crevices between
the grains could provide pathways for solution to permeate the lm and
reach the underlying substrate. In such cases, electroactive substrates (e.g.,
Pt, W, and Mo) would give rise to increased currents and potential depen-
dent features in the background voltammetric response. Undercutting and
delamination of the diamond lm could also result.
Second, impurities from either the substrate or the reactor walls
could be incorporated into the lm during deposition. Examples include Si
from the quartz reaction chamber during microwave plasma CVD [13] and
Wfromthe lament during hot lament growth [14]. Pt nanoparticles have
also been detected in diamond lms deposited on Pt substrates using
microwave-assisted CVD [15]. By a mechanism that is unclear, these Pt
metal impurities can be removed from the substrate at the deposition
temperature (ca. 850jC) and incorporated within the growing lm. The
inuence these metal impurities exert, directly or indirectly, on the lm
structure and electrochemical response could be important, depending
upon the metal and the redox reaction under study. Metal atoms exposed
at the surface, for instance, could catalyze electrochemical reactions.
Solvent and electrolyte could also react at these impurity sites causing an
increase in the background current. The resulting structural defects could
lead to localized breakdown of the lm during the imposition of harsh
electrochemical conditions of pH, temperature, and current density.
Third, diamond lms deposited on nondiamond substrates could
possess signicant internal stress, both intrinsic and thermal [16]. The role
that stress plays, macroscopically and microscopically, in the electrochem-
ical response is not understood at present. One possible manifestation of
localized stress could be less morphological and microstructural stability
during exposure to harsh electrochemical conditions.
The rate of diamond nucleation on untreated, smooth substrates is
routinely low. Therefore, substrate pretreatment is a prerequisite for
diamond deposition. A common method of pretreating substrates is to
seed the surface by either mechanically polishing with a small diameter
diamond powder (0.011 Amdiameter) or ultrasonicating the substrate in a
diamond powder/solvent slurry. For example, the diamond powder can be
suspended in ethanol and the sonication performed in this medium. Either
pretreatment results in the formation of scratches and other defects, as well
as the dispersion of diamond particles over the surface. Most pretreat-
ments also involve the subsequent washing of the substrate to remove
polishing debris and clusters of diamond particles. The washing typically
involves sequential sonications in isopropanol, acetone, and methanol.
Enhancement of the nucleation density results from both the physical
defects introduced and the highly dispersed diamond seed particles.
Figure 3 shows a series of AFM images (air) of (A) a smooth Si surface,
(B) the Si surface after mechanical polishing with diamond powder and
solvent cleaning, (C) the pretreated surface after a very short diamond
deposition period, and (D) a polycrystalline lm conformally coating the
pretreated substrate. Notice how clean the polishing scratches are after
washing (Fig. 3B) and how the initial nuclei form almost exclusively in the
clean scratches (Fig. 3C).
FIG. 3. AFM images of (A) an untreated Si (100) surface, (B) a mechanically
polished Si (100) surface with 0.1 Am diamond powder, (C) the mechanically
polished surface after a short period of diamond deposition, and (D) the
mechanically polished surface after complete diamond lm coverage.
Swain 190
Boron-doped diamond thin lms have also been deposited on other
types of substrates, such as sharpened Pt wires to form diamond micro-
electrodes, and Mo metal meshes to form high surface area electrode
architectures. Two examples are shown in scanning electron microscopy
(SEM) images presented in Fig. 4. The image in Fig. 4A reveals the well-
faceted, micrometer-sized grain morphology of a diamond lm deposited
on a sharpened Pt wire. The electrode radius at the end of the ber is
approximately 15 Am. Microelectrodes of 30 Am diameter or less oer
signicant advantages over their macroelectrode counterparts in terms of
enhanced diusional mass transport leading to larger analytical currents,
lower capacitive background currents, increased spatial and temporal
resolution, and the ability to make measurements in resistive media due
to lower iR
uncompensated
eects [17].
Metal meshes (e.g., Mo) have also been coated. Such electrodes could
eventually be useful for electrosynthesis, coulometric detectors, and as
supports for electrocatalysts. For many of the envisioned electrochemical
applications of diamond, the electrode needs to be in a nonplanar, high
surface area form. The low-magnication SEM image in Fig. 4B shows a
junction in the mesh between two wires coated with diamond. The lmwas
deposited using a CH
4
/H
2
source gas mixture (f0.5%). The mesh wires
are about 200 Am in diameter, and, although not easy to see, the diamond
lm thickness is several micrometers.
The diamond growth can also be patterned to produce microelec-
trode array structures [18,19]. Several possible microstructures are possi-
ble, such as microbands, microdiscs, and microcolumns. Micropyramids
are another microstructure that can be produced, and an image of such an
array is shown in Fig. 5. The SEM image reveals a monolithic diamond-tip
array. The tips are ca. 2 Am in base diameter and are equally positioned
over the surface with a spacing of ca. 5 Am.
In order to have sucient electrical conductivity for many electro-
chemical measurements (<0.1 ohm-cm or >10 S/cm), diamond lms must
be doped with boron at a concentration of 1 10
19
cm
-3
or greater, as this is
the most practical dopant to date. It should be pointed out, though, that
there is ongoing research with alternate dopants, like sulfur [20,21]. The
boron is added to the source gas mixture in the form of B
2
H
6
or B(CH
3
)
3
.
B
2
H
6
is the preferred gas as B(CH
3
)
3
not only adds boron to the source gas
mixture but also adds extra carbon. This extra carbon can decrease the lm
quality. The lms are rendered electrically conducting through incorpo-
ration of boron dopant atoms during deposition, although the electrical
conductivity depends in a complex manner on lattice hydrogen, defects,
FIG. 4. SEM images of a boron-doped diamond coated (A) Pt wire micro-
electrode and (B) Mo metal mesh.
Swain 192
dangling bonds, in addition to the doping level. For example, lms can be
doped as high as 10
21
B/cm
3
, with little alteration of the morphology or
microstructure, yielding resistivities of less than 0.05 V-cm.
A three-neck, single-compartment electrochemical cell constructed
of glass with about a 5 mL internal volume is routinely used in our lab-
oratory [22]. The diamond working electrode is pressed against a Vitonk
O-ring and clamped to the bottom of the glass cell. Ohmic contact is made
using either a bead of Ga/In alloy or Ag paste on the backside of the
scratched and cleaned substrate. The substrate is then placed in contact
with a clean current collector plate, either a Cu or Al. The counter and
reference electrodes are placed in two of the necks and are often a large
area carbon rod and a commercial Ag/AgCl electrode (saturated KCl)
(E
0
Ag/AgCl
= +0.197 V vs. NHE), respectively. The geometric area of the
working electrode is ca. 0.2 cm
2
in this particular design. Nitrogen-purge
gas is introduced into the solution through the third neck. For optimum
FIG. 5. SEM image of a micropyramid diamond electrode array. (Image cour-
tesy of Prof. J. L. Davidson and coworkers, Vanderbilt University.)
cell performance, the counter electrode should be placed normal to the
working electrode. This will keep the current density and the iR drop the
same at all points along the working electrode surface. The distance-
dependent potential drop between the reference and working electrodes,
iR
uncompensated
, causes an error in the accuracy of the applied potential.
This potential drop is minimized by positioning the tip of the reference
electrode as close to the working electrode as possible without interfering
with mass transport of electroactive species to the surface.
III. ELECTRICAL CONDUCTIVITY OF
DIAMOND ELECTRODES
As mentioned, the electrical conductivity of boron-doped diamond elec-
trodes depends in a complex manner on lattice hydrogen, defects, dangling
bonds, in addition to the doping level. The electronic properties (i.e.,
electrical conductivity) can be controlled, to a large extent, through
adjustments in the doping level from f10
17
to 10
21
B/cm
3
. Boron substi-
tutionally inserts into the growing diamond lattice and functions as an
acceptor, thereby increasing the carrier concentration. The activation
energy for carrier generation decreases from 0.37 eV at low doping levels
to below0.02 eVat high doping levels due to wave function overlap and the
formation of an acceptor state continuum [23,24]. The extent to which the
lattice hydrogen, defect density, and dangling bonds aect the conductivity
depends on the boron doping level, how the lms were grown, and the
procedure used to cool the lms after growth.
The carrier concentration at room temperature (i.e., density of
electronic states) is one of the factors that controls electrical conductivity
and is dependent on the activation energy for carrier generation [j(T) =
j
o
exp(E
a
/kT)]. The carrier concentration only includes electrons that
can easily be excited from occupied into unoccupied electronic states and
propagate through the lattice. In the absence of an external excitation (e.g.,
light), carrier generation occurs by thermal activation, which at room
temperature is ca. 0.03 eV. The conductivity is also related to the motion of
charge carriers in the lattice, either holes or electrons depending on the
dopant type, and is dened by the relationship
j nel
in which j is the conductivity (V
1
cm
1
or S/cm), n is the carrier con-
centration (cm
3
), e is the charge of the electron (1.6 10
19
coulombs),
and A is the carrier mobility (cm
2
/V-s). Commonly, boron-doped, semi-
Swain 194
metallic diamond lm electrodes (microcrystalline or nanocrystalline lms
used in our laboratory) have doping levels of ca. 10
20
10
21
B/cm
3
, as deter-
mined fromboron nuclear reaction analysis, and roomtemperature carrier
concentrations of ca. 10
18
cm
3
and hole mobilities of 1-50 cm
2
/V-s, both
as determined from Hall measurements. For a given doping level, current
data indicate that microcrystalline diamond has a slightly lower carrier
concentration and slightly higher mobility than nanocrystalline diamond.
The higher grain boundary density in nanocrystalline diamond reduces the
carrier mobility due to scattering. The discrepancy between the total boron
and the active carrier concentration is due, in part, to the fact that not all of
the incorporated boron substitutes for carbon in the lattice and, thus, does
not function electronically. Some of the inserted boron can aggregate in
grain boundaries or at other defect sites. Also, boron acceptors can be
compensated for by nitrogen donor impurities adventitiously incorporated
into the lattice or by complexation with H to form a B-H aggregates. The
adventitious nitrogen comes from low levels of atmospheric leakage into
reactors, as most CVDprocesses use only lowvacuumpumping (mtorr), or
from contamination of the source gases. The hydrogen, in the form of
atomic hydrogen, comes from the source gases used for the deposition.
Another important property to recognize about polycrystalline diamond is
the fact that homogeneous doping is not possible. The doping level
depends on the growth sector with the (111) crystallographic orientation
incorporating about 10 the amount of the (100) orientation [25].
Figure 6 shows Hall measurement data for a series of boron-doped
nanocrystalline diamond lms deposited with dierent levels of B
2
H
6
added to the source gas mixture. Measurements of the carrier concentra-
tion and mobility were made at dierent temperatures up to about 500jC.
At room temperature, the carrier concentration increases and the hole
mobility decreases as B
2
H
6
added to the source gas mixture with values in
the range of 10
18
10
19
cm
3
and 10100 cm
2
/V-s, respectively. The carrier
concentration and the doping level increase proportionally with B
2
H
6
added.
IV. CHARACTERIZATION OF MICROCRYSTALLINE
AND NANOCRYSTALLINE DIAMOND THIN
FILM ELECTRODES
Most of the published work with diamond electrodes has involved the use
of microcrystalline lms [19,22,26128]. Material characterization for
the purpose of verifying lm quality is an important component of any
FIG. 6. Carrier concentrations and mobilities (holes) in boron-doped nano-
crystalline diamond thin-lm electrodes as a function of the B
2
H
6
concentration
added to the source gas mixture. The measurements were made in a van der Pauw
geometry with Ti contacts by Dr. Toshihiro Ando at NIMS. The limit of the
temperature measurements was approximately 500 K.
diamond electrode research. Techniques such as scanning electron micros-
copy (morphology and texture), Raman spectroscopy (microstructure and
stress), x-ray diraction (crystallinity), x-ray photoelectron spectroscopy
(near-surface chemical composition), secondary ion mass spectrometry
(bulk chemical composition), boron nuclear reaction analysis (boron
concentration), Hall measurements (carrier type, concentration, and mo-
bility), and cathodoluminescence (impurity content) can be used for
characterization.
Figure 7 shows x-ray diraction patterns for moderately boron-
doped (f10
19
cm
3
) microcrystalline and nanocrystalline diamond thin
FIG. 7. X-ray diraction patterns for a boron-doped (a) microcrystalline and
(b) nanocrystalline diamond thin lm.
lms. Three broad reections are observed for both lms at 2-u values of
43.8, 75.5, and 91.5j. These reections are assigned to the (111), (220), and
(311) planes of cubic diamond, respectively [115,129]. The peaks for the
nanocrystalline lmare broader than they are for the microcrystalline lm,
both approximately the same thickness, due to the z2 order of magnitude
smaller diamond grain size. The peak widths increase with increasing
doping level. The diraction data, in both cases, reveal a bulk lmstructure
of sp
3
-bonded diamond.
Figure 8 shows visible-Raman spectra for moderately boron-doped
(f10
19
cm
3
) microcrystalline and nanocrystalline diamond thin lms.
The spectrum for the microcrystalline lm consists of the one-phonon di-
amond line centered at 1333 cm
1
. The line width (FWHM) is ca. 10 cm
1
and, to a rst approximation, is inversely related to the phonon lifetime
[123,130]. The line position is negligibly shifted from that for a reference
FIG. 8. Visible Raman spectra for a boron-doped (a) microcrystalline and (b)
nanocrystalline diamond thin lm.
Swain 198
piece of single crystal diamond, but the line width is larger (10 vs. 2 cm
1
)
due to the grain boundaries in the polycrystalline lm. The higher the
defect density, the larger the line width. There is negligible scattering inten-
sity between 1500 and 1600 cm
1
due to nondiamond sp
2
carbon impurity.
It is important to note that the qualitative assessment of lm quality using
Raman spectroscopy relies on the fact that the scattering cross section for
graphite, a model microstructure for the nondiamond carbon impurity,
which is actually amorphous, containing a mixture of sp
2
- and sp
3
-bonded
carbon, is a factor of 50 greater than that for diamond [130]. Hence, visible
Raman spectroscopy is highly sensitive to the impurity. This spectrum is
characteristic of high-quality diamond lms. A single, sharp peak at 1332
cm
1
in the Raman spectrum is frequently used as a signature of high-
quality, single crystal, or large-grained diamond [131,132].
The spectrum for the nanocrystalline lm is quite dierent. Broad
peaks are seen at 1150, 1333, 1470, and 1550 cm
1
. The peak at 1333 cm
1
,
atop a large background signal, is associated with the rst-order phonon
mode of sp
3
-bonded diamond [131,132]. Due to resonance eects, the
Raman cross-section scattering coecient (visible excitation) for sp
2
-
bonded carbon is larger (50) than that for the sp
3
-bonded carbon, and
the scattering intensity for the former can often greatly exceed that for the
latter [130]. The peak width (FWHM) for the diamond line is much
broader than that for microcrystalline diamond: 140 versus 10 cm
1
. This
is because of the small grain size in the nanocrystalline lm. There are two
possible causes for the line broadening. One possiblity is the well-estab-
lished connement model [132]. This model states that the smaller the
domain size, the larger the range of phonon modes (with dierent q vector
and energy) that are allowed to participate in the Raman process. Hence,
the line width results from the spread in phonon energy. Another, more
likely, explanation is phonon scattering by impurities and defects (i.e.,
grain boundaries) [132]. The scattering event shortens the lifetime of the
phonons and, thus, broadens the Raman line.
The peak at 1150 cm
1
is often used as a signature for high-quality
nanocrystalline diamond [132,133]. Prawer and coworkers, through the
study of clean nanocrystalline diamond particles (f5 nm diameter), have
attributed this peak to a surface phonon mode of diamond [133]. On the
other hand, Ferrari and Robertson have assigned this peak to sp
2
-bonded
carbon, specically transpolyactelylene segments at the grain boundaries
[134]. Their assignment of sp
2
-rather than sp
3
-bonded carbon, as has often
been proposed [131133], is based on the facts that the peak position
changes with excitation energy, the peak intensity decreases with increas-
ing excitation energy, and the peak is always accompanied by another peak
at f1450 cm
1
, which behaves similarly with excitation energy. Therefore,
the peaks at 1470 and 1550 cm
1
are tentatively assigned to disordered sp
2
-
bonded carbon in the grain boundaries. The 1550 cm
1
peak is down-
shifted from the expected 1580 cm
1
position for graphite. This shift
results because the sp
2
-bonded carbon is amorphous and is mixed with
sp
3
-bonded carbon. It is important to note that, very likely, the sp
2
-
bonded carbon is conned to the grain boundaries of the nanocrystalline
lm, producing a network of three- and four-fold coordinated carbon
atoms [135]. Additional research is needed to conrm the origins of the
1150 and 1470 cm
1
bands.
Raman imaging spectroscopy has been used to investigate the
structural and chemical inhomogeneities that exist in boron-doped dia-
mond lms [123]. A number of boron-related Raman lines centered at
approximately 610, 925, 1045, 1375, and 1470 cm
1
were evidenced and
found to vary from one region of the lm to another. Even if their origin is
still somewhat uncertain, their presence enabled the construction of spatial
maps of the boron signal intensity over the surface. It was conrmed that
boron is incorporated into CVD diamond with signicant inhomogene-
ities. In most cases, the boron signal intensity tracked the structure of the
grains. The high quality of the electrode studied allowed for detailed
analysis of both the grains and grain boundaries. A unique description
of the microstructure and chemical composition of the grains and grain
boundaries was not possible from the data, but there was evidence that the
nondiamond carbon impurity phases are incorporated into the grain
boundaries. The high-quality disk electrode was also nearly photolumi-
nescence-free. Photoluminescence, as excited by the 514.5 and 457.9 nm
lines of an argon ion laser, was only detected within some of the non-
diamond carbonrich grain boundaries. Localized stress variations were
also observed. The strained regions were not exclusively associated with
the incorporated nondiamond carbon. Grain boundaries and planar
defects were identied as the stress source.
Raman and XRD are two particularly useful lm characterization
techniques indeed. The diamond lms routinely used in our laboratory are
low in surface oxygen due to the hydrogen surface termination and,
consequently, are hydrophobic. X-ray photoelectron spectroscopy reveals
atomic oxygen/carbon ratios of 2% or less for both as deposited micro-
Swain 200
crystalline and nanocrystalline lms. The instantaneous water contact
angles range from 70 to 80 degrees on both surfaces.
V. BASIC ELECTROCHEMICAL PROPERTIES OF
MICROCRYSTALLINE AND NANOCRYSTALLINE
DIAMOND THIN FILM ELECTRODES
Boron-doped, hydrogen-terminated, diamond thin lms, both microcrys-
talline and nanocrystalline, possess a number of important and practical
electrochemical properties, clearly distinguishing them from other com-
monly used sp
2
carbon electrodes, like glassy carbon, pyrolytic graphite, or
carbon paste [22,2931,37,41,42,44]. These properties are (1) low and
stable voltammetric and amperometric background current, (2) wide
working potential window in aqueous media, (3) superb microstructural
and morphological stability even under very harsh electrochemical con-
ditions, (4) good responsiveness for several aqueous- and nonaqueous-
based redox analytes without any conventional pretreatment, (5) weak
adsorption of polar molecules and other potential surface contaminants,
(6) long-termresponse stability, and (7) optical transparency in the UV/Vis
and IR regions of the electromagnetic spectrum.
Diamond exhibits approximately an order of magnitude lower back-
ground current density (per geometric area) in voltammetric measure-
ments, as compared to glassy carbon [2931,48,50]. This leads to roughly
an order of magnitude improvement in the signal-to-background (S/B)
ratio and translates into lower detection limits. The lower background
current is attributable to two factors: a reduced pseudocapacitance from
the absence of redox-active or ionizable surface carbon-oxygen functional
groups, and a slightly lower internal-charge carrier concentration due to
the semimetal-like electronic properties. The data presented in Fig. 9
illustrates this comparison. The background voltammetric responses
shown are for freshly polished glassy carbon and a boron-doped diamond
thin lm and were obtained in 0.1 M HClO
4
at a potential sweep rate of
0.1 V/s. The background current for diamond is approximately an order of
magnitude less than that for glassy carbon.
The double-layer capacitance for glassy carbon in the potential re-
gion from500 to 1000 mV varies from 30 to 40 AF/cm
2
(geometric area),
while that for diamond ranges from 3 to 7 AF/cm
2
[22,28,33,38]. The
voltammogram for glassy carbon, shown in Fig. 9, shows some evidence
for characteristic oxidation and reduction peaks centered at 350 mV
associated with redox-active surface carbon-oxygen functional groups
[29,45,136] at the crystallite edge plane and defect sites. There is no
evidence for such redox-active functional groups in the voltammetric
response for diamond.
Diamond has a wide working potential window due to large over-
potentials (i.e., slow electrode reaction kinetics) for hydrogen and oxygen
evolution. Typical working windows of 3.25 V, or greater, are typical for
high-quality lms. The reason for the large overpotentials is not fully
understood but is likely related to the absence of low-energy adsorption
sites for the reaction intermediates on the hydrogen passivated surface:
OH
ads
for oxygen evolution and H
ads
for hydrogen evolution [53]. These
low-energy adsorption sites stabilize the reaction intermediates and lower
the activation energy for the electrochemical reaction. The lowering of the
activation barrier leads to an increase in the electrode-reaction rate. Large
overpotentials for electrochemical reactions at diamond is certainly not a
general rule. For example, diamond exhibits much lower overpotentials for
other surface-sensitive redox reactions, such as metal deposition.
FIG. 9. Background cyclic voltammetric i-E curves for (a) glassy carbon, (b) a
moderately boron-doped microcrystalline diamond lm electrode, and (c) a
heavily boron-doped microcrystalline diamond lm electrode in 0.1 M HClO
4
.
Electrode area = 0.2 cm
2
. Scan rate = 0.1 V/s.
Swain 202
The importance of the large working potential window for diamond
is that electrochemical reactions can be studied over a much wider
potential range than is possible with other carbon electrodes. Similarly,
wide working potential windows are also observed in nonaqueous media
(e.g., TBAClO
4
/CH
3
CN) [108].
Diamond electrodes can be exposed to extreme conditions, such as
high temperature and high current density, without degredation of the
morphology or microstructure. This is an amazing material property! For
example, excellent morphological and microstructural stability are ob-
served during chlorine evolution [47] and polarization in acidic uoride
solutions [45]. The superb dimensional stability has two important ram-
ications. First, positive (anodic) potentials are applied to an electrode for
the electrooxidative detection of many analytes, and the surface of sp
2
carbon electrodes is oxidized over a very wide potential range (>0.5 V vs.
Ag/AgCl, pH 1). In other words, the electrode surface microstructure is
unstable during detector operation. A variety of surface carbon-oxygen
functional groups are formed as a result of the microstructural alterations
(e.g., quinones, hydroxyls, carboxyls), and some of the functionalities are
redox active and ionizable, depending on the potential and solution pH.
Both of these factors lead to increased background currents and reduced
S/B ratios in electrochemical measurements. Moreover, the sp
2
carbon
electrode microstructure is dynamically altered by the oxidation process,
and this alteration leads to a progressive increase in the number of active
sites for oxidation. While diamond experiences changes in the surface
chemical composition during mild anodic-polarization conditions (e.g.,
conversion from H to O termination), the microstructure and morphology
are extremely stable [41,45,47]. There is no alteration of the morphology or
microstructure, and this leads to excellent long-term response stability.
Second, in electrocatalysis and electrosynthesis, sp
2
carbon electrodes are
often used as a catalyst support or as the direct electrode material.
Degradation of the microstructure causes detachment or aggregation of
the catalyst particles and reduced catalytic activity. There are limits on the
solution conditions, temperature, and current density to which sp
2
carbon
supports can be exposed, because of microstructural degradation and
corrosion (i.e., loss of carbon due to CO and CO
2
evolution). Diamond
electrodes do not have these limitations; therefore, because of the dimen-
sional stability and the material properties, the electrodes may nd future
use in electrocatalysis and electrosynthesis [4851,128]. The dimensional
stability and material properties are the impetus for eorts to incorporate
Pt and Pt/Ru nanoparticles into the diamond surface microstructure and
to use these materials in electrocatalysis, as is discussed further below.
Cyclic voltammetric studies of a number of aqueous [Fe(CN)
6
-3/-4
,
Ru(NH
3
)
6
+2/+3
, methyl viologen, Ru(bpy)
3
+3/+2
, chlorpromazine, ascor-
bic acid, 4-tert-butylcatechol, dopamine, 4-methylcatechol, Fe
+2/+3
, hy-
drazine, azide, etc.] and nonaqueous (ferrocene) redox systems have been
performed to probe the response of boron-doped diamond thin lm
electrodes as a function of the redox analyte type and concentration, scan
rate, and exposure time to the laboratory air [22,30,41,44,107,108,115].
The density of electronic states at the formal potential of the redox analyte
(i.e., how conducting the electrode is at dierent potentials) is important
for controlling the responsiveness of diamond electrodes [22,138]. It has
been shown using redox analytes with heterogeneous electron transfer rate
constants primarily sensitive to the density of electronic states that boron-
doped diamond electrodes possess excellent electrically conductivity at all
potentials within the working potential window [22,30,41,44,107,108,115].
For all analytes at the highest-quality lms, (1) the oxidation peak currents
vary linearly with the scan rate
1/2
(r
2
> 0.99), indicating current control
by semi-innite linear diusion of the analyte to the interfacial reaction
zone, (2) the oxidation peak currents vary linearly with the concentration
between 1 mM and 1 AM (r
2
> 0.99), (3) conventional pretreatment is
generally not required for an active response, and (4) the response remains
stable for weeks to months during exposure to the laboratory air. Using
these redox systems with vastly dierent standard reduction potentials, Ej,
has shown that the moderately to heavily boron-doped lms behave as a
semimetal with a high density of electronic states over a wide potential
range from +1.3 to 1.6 V vs. SCE and that the grain boundaries and
nondiamond sp
2
carbon impurity phases are not the sole sites for the
charge-transfer reactions [22,30,41,44,107,108,115].
There is a rationale for studying these redox systems, one that
originates largely from the work of McCreery and coworkers in recent
years with glassy carbon electrodes [139143]. Redox reactions are gener-
ally of two types. One type includes electrode reactions that proceed by
simple diusion of the analyte to the interfacial reaction zone with the
electrode serving solely as a source or sink for electrons. In this case, the
electrode reaction kinetics are relatively insensitive to factors such as
the surface chemistry and microstructure, but very sensitive to the density
of electronic states at the formal potential (so-called outer-sphere reac-
tion). The other pathway includes reactions that occur via some specic
Swain 204
interaction with the electrode surface, for example, reactions that proceed
through an adsorbed state. In this case, the electrode reaction kinetics
are sensitive to surface chemistry and microstructure as well as the density
of electronic states at the formal potential (so-called inner-sphere reac-
tion). The redox systems, Ru(NH
3
)
6
+2/+3
, methyl viologen, IrCl
6
2/3
,
Ru(bpy)
3
+3/+2
, and chlorpromazine, proceed by outer-sphere electron
transfer, and the electrode kinetics are relatively insensitive to the physi-
cochemical properties of diamond [22,30,41,44,107,115]. Apparent het-
erogeneous electron transfer rate constants, kj
app
, between 0.01 and 0.2 cm/
s are commonly observed for conducting polycrystalline lms (both
microcrystalline or nanocrystalline) without any kind of pretreatment
[22,30,41,44,107,115]. Normally, carbon and metal electrodes must be
pretreated (e.g., polishing) to achieve such rapid electrode kinetics.
Fe(CN)
6
3/4
proceeds through a more inner-sphere electron transfer
pathway, and the electrode kinetics are highly sensitive to the surface
termination of diamond [22,30,41,44,107,115]. Presumably, the reaction
proceeds through a specic surface site that is blocked by oxygen. kj
app
values ranging from 0.01 to 0.1 cm/s are commonly observed for clean,
hydrogen-terminated lms, but the rate constants decrease by over two
orders of magnitude for oxygen-terminated lms [41]. Fe
+2/+3
is also of
the inner-sphere type, with kj
app
being very sensitive to the surface carbon-
oxygen functionalities (e.g., carbonyls), at least for glassy carbon [120,139].
These functional groups are absent on hydrogen-terminated diamond, and
this is presumably the reason kj
app
is low, in the range of 10
4
10
6
cm/s.
The more complicated organic systems, dopamine, 4-methylcatechol, and
4-tert-butylcatechol, are inner-sphere systems with kj
app
values of 10
4
to
10
6
cm/s at hydrogen-terminated diamond. It is postulated that the slow
kinetics for these polar aromatic analytes result from weak adsorption on
the diamond surface [109,143]. It is supposed that k-k interactions through
the extended electronic conjugation are perhaps more inuential than
surface oxides in promoting molecular adsorption for some classes of
analytes on these surfaces (e.g., catechols). Further work is needed to fully
understand the structure-function relationships for the inner-sphere redox
systems.
Figure 10 shows a series of cyclic voltammetric i-E curves for
Fe(CN)
6
3/4
, Ru(NH
3
)
6
+2/+3
, IrCl
6
2/3
, and methyl viologen at a
boron-doped microcrystalline diamond thin lm. Nearly reversible vol-
tammetric responses are seen for all four redox systems at an untreated
electrode. Table 1 summarizes some of the cyclic voltammetric and
FIG. 10. Cyclic voltammetric i-E curves for 0.1 mM Fe(CN)
6
3/4
, 0.05 mM
Ru(NH
3
)
6
+3/+2
, 0.1 mM IrCl
6
2/3
, and 0.05 mM methyl viologen (MV
+2/+
) at
a boron-doped microcrystalline diamond thin-lm electrode. Scan rate = 0.1 V/s.
Electrolyte = 1 M KCl.
TABLE 1
Cyclic Voltammetric and Heterogeneous Electron Transfer Rate Constant Data
for Four Aqueous-Based Redox Systems at Boron-Doped Microcrystalline
Diamond Thin-Film Electrodes
Analyte DE
p
(mV) i
p
ox
(AA) k
o
app
(cm/s)
0.1 mM Fe(CN)
6
3/4
72 F 3 57 F 2 0.018 F 0.008
0.05 mM Ru(NH
3
)
6
+3/+2
76 F 8 59 F 2 0.025 F 0.008
0.1 mM IrCl
6
2/3
59 F 1 23 F 1 0.28 F 0.08
0.05 mM MV
+2/+
60 F 2 5.7 F 0.5 0.25 F 0.09
2
. Scan rates = 0.10.5 V/s. Peak potential separation and peak
current data are for 0.1 V/s.
Swain 206
heterogeneous electron transfer rate constant data for these redox systems
at boron-doped microcrystalline diamond electrodes.
The much smoother nanocrystalline diamond thin lms possess
interesting mechanical, tribological, and electrical properties owing to
the small grain size. For example, the diamond lms transition from an
electrically insulating to an electrically conducting material with a reduc-
tion in crystallite size from the micrometer to the nanometer level [2,3].
This is largely due to the presence of high-energy, high-angle grain
boundaries that contain k-bonded carbon atoms (i.e., a high density of
electronic states over a wide energy or potential range). The grain
boundaries are conducting because of k-states, and since their numbers
vastly increase with decreasing crystallite size, the entire lm becomes
electrically conducting and functions as an electrode material [2,3,137].
Theoretical calculations suggest that localized electronic states are intro-
duced into the band gap of these lms, due to the presence of sp
2
-bonded
dimers and sp
3
-hybridzed dangling bonds in the grain boundaries [2,3,7].
There is a lack of spatial connectivity among the sp
2
-bonded carbon sites,
therefore, the associated gap states do not form an extended k-system, but
rather are localized.
Nanocrystalline diamond thin lms can be grown with incorporated
nitrogen. These lms are electrically conducting and have been investi-
gated as electrodes [42,44]. As stated previously, the nanocrystalline
morphology results from a high renucleation rate. Such lms are typically
7501000 nm thick, continuous, and smooth at the nanometer scale. The
grain size is 310 nmat a thickness of 1 Amor less, and the grain boundaries
are 0.20.5 nm wide. Nitrogen does not function as a dopant per se, but
appears at the grain boundaries. The concentration (f10
20
atoms/cm
3
) of
incorporated nitrogen scales with the N
2
added to the source gas mixture
up to about the 5% level. The lms are void of pinholes and cracks, even
though very thin, and are electrically conducting, due in part, to the large
number of grain boundaries and the high concentration of nitrogen and/or
related defects (sp
2
carbon bonding) introduced by the incorporated
impurity. The electrical conductivity is retained regardless of the lm
thickness. The lms were found to possess semimetallic electronic proper-
ties over a wide potential range from at least -1.5 to 1.0 V vs. SCE and, like
boron-doped microcrystalline diamond, exhibit a wide working potential
window, a low background current, and good degree of electrochemical
responsiveness for redox systems, such as Fe(CN)
6
3/4
, Ru(NH
3
)
6
+3/+2
,
IrCl
6
2/3
, and methyl viologen (MV
+2/+
). More sluggish electrode
kinetics were observed for 4-methylcatechol (4-MC), presumably due to
weak adsorption on the surface.
The electrical conductivity of these nitrogen-containing lms scales
with the nitrogen content in the source gas mixture and the lms are
generally more conductive than the early forms of nanocrystalline dia-
mond without incorporated nitrogen [212,137]. Hall eect measurements
(mobility and carrier concentration) for some lms deposited with 10 and
20%N
2
indicated carrier concentrations of 2.0 10
19
and 1.5 10
20
cm
3
,
respectively [137]. The room temperature carrier mobilities were 5 and 10
cm
2
/V-s, respectively. A negative Hall coecient indicated that electrons
are the majority charge carrier. An explanation for the eect of nitrogen on
the electrical conductivity is that the impurity causes microstructural
changes within the grain boundaries (i.e., increased k-bonding), resulting
in an increase in the localized density of electronic states. Computations
indicate that the incorporation of nitrogen into the grain boundaries is
energetically favored by 35 eVover substitutional insertion into the grains
[2,3,137].
While the nitrogen-containing lms possess good electrochemical
behavior, much like those for high-quality microcrystalline diamond lms,
their electrical response is strongly linked to the physicochemical proper-
ties of the grain boundaries. Therefore, the electrochemical response can be
signicantly inuenced by changes in the k-bonded grain boundary carbon
atoms. Presumably, these electrodes would become deactivated by con-
taminant adsorption during extended exposure at anodic potentials. Such
alteration could occur during exposure to chemically harsh solutions that
lead to oxidative etching or disruption of the k-bonded carbon atoms. It
would be best if the nanocrystalline diamond lms exhibited a through-
grain conduction mechanism as a result of impurity incorporation, such as
boron doping. Such lms should exhibit electrical conductivity that scales
with the doping level and should have electrochemical properties that are
largely unaected by changes in the physicochemical properties of grain
boundaries (see Fig. 6).
Boron-doped nanocrystalline diamond lms have recently been
produced and evaluated as electrodes [115]. The lms consist of clusters
of diamond grains, 100 nm in diameter, and possess an rms surface
roughness of 34 nm over a 5 5 Am
2
area. The individual diamond grains
are approximately 1015 nm in diameter, as determined by TEM. Films
with a thickness ranging from 0.5 to 4 Am are deposited by microwave-
assisted CVD using a CH
4
/H
2
/Ar gas mixture (1/4/95%). B
2
H
6
, diluted
Swain 208
in hydrogen, is used as the source gas for doping. Unlike other nano-
crystalline diamond lms that have electrical properties dominated by the
high fraction of sp
2
-bonded grain boundaries, the present lms are doped
with boron and the electrical properties are dominated by the charge
carriers in the diamond. Good electrode responsiveness was observed for
Fe(CN)
6
3/4
, Ru(NH
3
)
6
+3/+2
, IrCl
6
2/3
, methyl viologen, Fe
+3/+2
, and
4-tert-butylcatechol. These new electrodes were also found to be useful for
the detection of aliphatic polyamines and trace metal ions.
Figure 11 shows cyclic voltammetric i-Ecurves for (A) Fe(CN)
6
3/4
,
(B) Ru(NH
3
)
6
+2/+3
, (C) IrCl
6
2/3
, (D) methyl viologen (MV
+2/+
) in 1 M
KCl, and (E) 4-tert-butylcatechol, and (F) Fe
+3/+2
in 0.1 M HClO
4
at a
boron-doped nanocrystalline diamond thin lm electrode. The potential
scan rate (r) was 0.1 V/s. The E
p
for these redox systems ranges from
FIG. 11. Cyclic voltametric i-E curves for (A) 1 mM Fe(CN)
6
3/4
, (B) 1 mM
Ru(NH
3
)
6
+3/+2
, (C) 0.25 mM IrCl
6
2/3
(D) 0.5 mM methyl viologen (MV
+2/+
),
(E) 1 mM dopamine, and (F) 1 mM Fe
+3/+2
at a boron-doped microcrystalline
diamondthin-lmelectrode. Scanrate =0.1 V/s. Electrolyte for ADwas 1 MKCl,
and for E and F was 0.1 M HClO
4
. (Reprinted with permission from Chem. Mater.
2003, 15, 879. Copyright (2003) American Chemical Society.) (From Ref. 115.)
approximately +800 to 1100 mV, so they are very useful for probing the
lms electronic properties over a wide potential range. A summary of
some of the cyclic voltammetric data is provided in Table 2.
There are several potential benets of using nanocrystalline rather
than microcrystalline diamond thin lms in electrochemical technologies:
(1) the ability to deposit continuous lms at nanometer rather than
micrometer thicknesses, leading to production time and cost savings; (2)
easier coating of irregular geometry substrates, like bers and high surface
area meshes; and (3) the dierent lm morphology and characteristic
electronic properties, compared to the microcrystalline lms, might lead to
unique electrochemical behavior.
Diamond exhibits low and weak adsorption of polar molecules in
aqueous media because of the hydrophobic and nonpolar nature of the
hydrogen-terminated surface [38,92]. Therefore, the surface resists con-
tamination and fouling by polar, aromatic adsorbates, resulting in excel-
lent long-term response stability. For example, relatively unchanging DE
p
values for Fe(CN)
6
3/4
and Ru(NH
3
)
6
+3/+2
have been observed at
diamond electrodes after exposure to toluene, pyridine, and hexane
vapors. The electrochemistry of one model compound, anthraquinone-
2,6-disulfonate (2,6-AQDS), has been studied in detail at glassy carbon,
hydrogenated glassy carbon (HGC), the basal plane of HOPG, and boron-
doped diamond using cyclic voltammetry and chronocoulometry [38].
The relationship between the 2,6-AQDS surface coverage, the double
layer capacitance, and heterogeneous electron transfer rate constant for
Fe(CN)
6
3/4
at these four carbon electrodes was presented [38]. The
TABLE 2
Cyclic Voltammetric Data for Aqueous-Based Redox Systems at a Boron-Doped
Nanocrystalline Diamond Thin-Film Electrode
Analyte DE
p
(mV) i
p
ox
(AA) i
p
ox
/i
p
red
1 mM Fe(CN)
6
3/4
63 69.4 1.0
1 mM Ru(NH
3
)
6
+3/+2
59 62.7 0.99
0.5 mM IrCl
6
2/3
61 22.7 0.98
0.5 mM MV
+2/+
60 34.5 1.1
Scan rate = 0.1 V/s; electrode area = 0.2 cm
2
; electrolyte = 1 M KCl.
Source: Ref. 115.
Swain 210
diamond and HGC surfaces are nonpolar and relatively oxygen-free with
the surface carbon atoms terminated by hydrogen. Also, HGC has an
extended k-electron system, whereas diamond does not. The polar 2,6-
AQDS does not adsorb on these surfaces to any measurable level, and the
electrolysis proceeds by a diusion-controlled reaction. Conversely, the
glassy carbon and HOPG surfaces are polar with the edge plane sites
terminated by carbon-oxygen functionalities. 2,6-AQDS strongly physi-
sorbs on both of these surfaces at near monolayer coverages or greater,
such that the electrolysis proceeds through a surface conned state. The
results implicate the important role of the surface carbon-oxygen function-
alities in promoting strong dipole-dipole and ion-dipole interactions with
polar and ionic molecules like 2,6-AQDS.
The adsorption tendencies of three structurally dierent catechols
have also been investigated using hydrogen-terminated, boron-doped
microcrystalline and nanocrystalline diamond thin lms. Comparison
measurements were made with polished glassy carbon. Table 3 summarizes
some coulometric data for the three redox systems. As the level of k-
bonding in the electrode microstructure increases (glassy carbon >>nano-
crystalline lm > microcrystalline lm), the surface coverage for the
catechols increases, and this correlates with the smallest cyclic voltam-
metric DE
p
[109]. For example, DE
p
for 4-methylcatechol is 719 F3 mVfor
microcrystalline diamond, 555 F 3 mV for nanocrystalline diamond, and
80 F 2 mV for glassy carbon (0.1 V/s). Surface oxides on glassy carbon
undoubtedly inuence the adsorption, but these preliminary data suggest
that the level of k-bonding in the surface microstructure of diamond is also
inuential at promoting adsorption.
TABLE 3
Molecular Adsorption Studies of Catechols at Diamond and Glassy Carbon
Electrodes (pmol/cm
2
)
Molecule Microcrystalline Nanocrystalline Glassy carbon
Catechol Not detected 3.2 F 0.1 29.4 F 1.7
4-Methylcatechol Not detected 3.9 F 0.3 164.3 F 3.0
4-tert-Butylcatechol Not detected 3.6 F 0.6 121.7 F 1.3
Potential step fromvs. Ag/AgCl. Electrolyte =0.1 MHClO
4
; analyte concentration =1 AM.
VI. FACTORS AFFECTING ELECTRON TRANSFER
AT DIAMOND ELECTRODES
The diamond electrodes responsiveness is governed by the (1) density
of electronic states at the formal potential, Ej, for the analyte, (2) sur-
face termination, (3) nondiamond sp
2
-bonded carbon impurity content,
and (4) defect density (morphology and microstructure). The extent to
which any one of these parameters aects the kinetics depends on the
electrochemical reaction mechanism for the particular analyte. In gen-
eral, the most active electrodes possessing the most reproducible prop-
erties are hydrogen-terminated. A characteristic of diamond is the fact
that pretreatment is not necessary to activate the electrodes. Due to the
absence of extended k-bonding within the lattice and the nonpolar
nature of the hydrogen-terminated surface, the electrodes resist fouling
by contaminants (i.e., deactivation). The normal protocol followed in
our laboratory is to place the hydrogen-terminated electrode in the elec-
trochemical cell and to expose the surface to a few milliliters of distilled
(clean) isopropanol [22,144]. In the event an electrode has been deac-
tivated, a useful method for reactivation involves a chemical cleaning
for 30 minutes each in warm 3:1 HNO
3
/HCl (v/v) and 30% H
2
O
2
/H
2
O
(v/v), followed by rehydrogenation in a hydrogen microwave plasma [22].
The density of electronic states is controlled, to a large extent, by the
boron doping level in the lattice. The electrodes appear to transition from
semiconducting to semimetallic electronic properties at doping levels
around 10
19
cm
-3
, although this has not been rigorously determined. The
eect of boron doping level on the electrode response for a fewanalytes has
been studied, but surprisingly, not much work has been done in this area
[145]. The cyclic voltammetric DE
p
for outer-sphere redox systems, like
Ru(NH
3
)
6
+3/+2
, decreases progressively as the doping level increases. The
carrier concentration (i.e., electrical conductivity) also tracks the doping
level, as expected.
The surface termination has a strong inuence on the electrode
response for some redox systems, like Fe(CN)
6
34
and ascorbic acid.
The two surface terminations that have been investigated are hydrogen and
oxygen. Cyclic voltammetric i-E curves are shown in Fig. 12 for 1 mM
ascorbic acid in 0.1 M HClO
4
at an electrode before and after anodic
polarization at dierent potentials. It can be seen that the E
p
ox
and i
p
ox
values at the as-deposited, hydrogen-terminated lm are 780 and 68 AA,
Swain 212
respectively. After the electrode has been anodically polarized at 1.5, 2.0,
and 2.5 V, respectively, for 5 minutes each in 1.0 M HNO
3
, E
p
ox
shifts
progressively positive and i
p
ox
progressively decreases. The anodic polar-
ization causes the addition of surface oxygen, making the surface more and
more hydrophilic. XPS data indicate a maximum atomic O/C ratio of ca.
0.15 after polarization at 2.5 V. Assuming that the total carbon volume
sampled is approximately 5 layers (ca. 5 10
15
atoms/layer), this ratio
corresponds to a surface coverage of approximately 0.75 of a monolayer.
Ascorbic acid is a surface-sensitive redox analyte in that the electrode
reaction rate and, presumably, the reaction mechanism are highly depen-
dent on the physicochemical nature of the diamond electrode surface.
Although further study is needed to understand this eect, it appears that
surface oxygen blocks a site involved in the redox reaction. Electrostatic
repulsion eects that are often invoked to explain the sluggish electron
transfer kinetics for highly anionic redox systems at severely anodized
FIG. 12. Cyclic voltammetric i-E curves for a boron-doped microcrystalline
diamond electrode in 1 mM ascorbic acid + 0.1 M HClO
4
before, after anodic
polarization at dierent potentials, and after rehydrogenation in a hydrogen
plasma.
glassy carbon do not seem to be a major cause for the sluggish kinetics at
oxidized diamond, at least in some cases. For example, unchanged cyclic
voltammetric DE
p
values have been observed for other anionic redox
systems, like IrCl
6
2/3
, at the same anodized diamond electrodes
[41,120]. Rehydrogenating the surface in a hydrogen plasma removes the
surface oxygen and reactivates the electrode as the E
p
ox
for ascorbic acid
shifts negative and i
p
ox
increases back to their original values. Similar
behavior is seen for Fe(CN)
6
3/4
. These eects are completely reversible
with multiple oxidation/hydrogenation cycles. There are no alterations in
the surface microstructure or morphology after polarization, simply
changes in the surface chemistry.
The nondiamond sp
2
carbon impurity level can inuence the elec-
trode response, depending on the redox system[30]. For example, as shown
below, the oxygen reduction reaction is very sensitive to the nondiamond
carbon impurity level while redox systems, like Fe(CN)
6
3/4
and
Ru(NH
3
)
6
+3/+2
are much less so. High-quality polycrystalline lms
normally contain minimal amounts of nondiamond sp
2
-bonded carbon.
However, sp
2
-bonded carbon can be introduced into the diamond, in a
controlled manner, by changing the CH
4
/H
2
sourcegas ratio used during
deposition, thus producing a wide range of sp
2
-sp
3
carbon microstructures.
Figure 13 shows a series of atomic force microscope (AFM) images of lms
deposited with dierent CH
4
/H
2
ratios. As the CH
4
/H
2
ratio increases, the
nominal crystallite size decreases and the extent of secondary nucleation
increases. The lms become thicker, more opaque, and have higher
incidence of nondiamond impurity and other defects with increased
CH
4
/H
2
ratio. The nondiamond carbon impurity tends to form in the
grain boundaries and defects. The higher ratio leads to higher rates of
renucleation and growth.
The level of sp
2
-bonded carbon can be detected either electrochem-
ically or by Raman spectroscopy. Figure 14 shows a series of Raman
spectra for the lms shown in Fig. 13. It can be seen that the one-phonon
diamond line at 1333 cm
1
decreases in amplitude and increases in width
from 12 to 43 cm
1
as the CH
4
/H
2
ratio increases. The scattering in-
tensity centered at ca. 1525 cm
1
also increases with increasing CH
4
/H
2
ratio, indicative of higher levels of sp
2
-bonded carbon. This nondiamond
carbon is actually a mixture of sp
2
- and sp
3
-bonded carbondiamond-
like carbon and is not graphitic in nature. The higher defect density, due
to the increased secondary nucleation, causes the increased line width
and the increased opacity from the nondiamond carbon causes the re-
Swain 214
duced line intensity. Above 3%, the Raman spectrum converts from one
representative of low-quality diamond to one more like glassy carbon.
The electrochemical behavior also tracks the microstructural conversion.
The growth rate increases with CH
4
/H
2
ratio and, thus, the reduced rst-
and second-order phonon modes for the Si substrate at 519 and ca. 900
cm
1
, respectively.
FIG. 13. Diamond lm morphologies as a function of the methane-to-hy-
drogen source gas ratio used during microwave plasma CVD. The area is 10 10
Am
2
. Z-axis range is 6 Am. Growth time in each case was 10 h.
VII. SURFACE MODIFICATION OF DIAMOND
MATERIALS AND ELECTRODES
Chemical modication of diamond surfaces is a ripe area for research.
Much can be learned about (1) how to control the electrode reaction
kinetics and mechanisms at diamond through alterations of the surface
chemistry and (2) using such modied surfaces as platforms for sensors and
other devices based on the material.
So far there have been only a few reports concerning the chemical
modication of diamond, either insulating or electrically conducting
diamond lms that would potentially be useful as electrodes. Smentkowski
and Yates reported on a facile approach for modifying the hydrogen-
FIG. 14. Raman spectra for diamond lms deposited with dierent CH
4
/H
2
source gas ratios. Spectra shown for the lms in Fig. 13. Excitation = 514.5 nm.
Integration time = 5 s.
Swain 216
terminated diamond surface using x-ray irradiation of adsorbed peruo-
roalkyl iodide layers [146]. Peruoroalkyl radicals are generated from the
irradiation, which then bond to the surface. These groups are then ther-
mally decomposed to produce strong surface C-F bonds. Angus and co-
workers also examined the eect of surface uorination on the electrode
properties of boron-doped diamond lms [53].
Ohtani and coworkers described an approach for covalently attach-
ing a quarternary pyridinium salt to the diamond surface [147]. The
reaction strategy rst involves introducing surface Cl functional groups
by irradiating the surface in the presence of Cl
2
. The chlorinated electrode
is then placed in hot pyridine to produce pyridine moieties on the surface.
Kim et al. used UV irradiation to convert adsorbed peruorobutyl iodide
on hydrogen-terminated diamond to form covalently attached peruo-
robutyl functional groups [148].
Kuo et al. modied the diamond electrode surface with substituted
aromatic groups (e.g., nitrophenyl, triuoromethylphenyl, and nitro-
azobenzene) [40]. The modication was accomplished using a well-
established chemistrythe electroreduction of substituted diazonium
salts [149153]. The reduction of such salts on diamond electrodes results
in the covalent attachment of phenyl moieties. Figure 15A shows the
proposed reaction mechanism. Figure 15B shows a series of cyclic vol-
tammetric i-E curves recorded during the derivatization step in 1 mM 4-
nitrophenyldiazonium salt plus 0.1 M TBABF
4
/CH
3
CN. A large current
is observed during the rst cathodic scan, which is associated with the
reduction of the diazonium to form the phenyl radical. The radical then
couples to the electrode surface, forming a stable adlayer. The reduction
current is signicantly attenuated during subsequent scans as the adlayer
passivates the surface toward this reaction. If there is good electronic
coupling between the substituted phenyl moiety and the electrode, then
the reversible reduction of the nitro group can be probed by cyclic vol-
tammetry, as is shown in Fig. 15C.
Hamers and coworkers have shown that the hydrogen-terminated,
electrically insulating diamond surface can be chemically modied with
functionalized alkenes (e.g., peruorodecene) [154]. They also showed that
it is possible to prepare diamond surfaces terminated with organic mole-
cules containing primary amine and carboxylic acid functional groups. In
their approach, chemical functionalization is accomplished with the aid of
UV irradiation. The diamond surface, in an enclosed vessel, is exposed to a
small volume of the particular organic compound. The solution environ-
Swain 218
ment is then purged with dry nitrogen prior to UV irradation (254 nm) for
times up to 2 hours. The authors speculate that the reaction mechanism
involves photoexcitation of electrons and holes in the space charge region
of the diamond followed by nucleophilic attack by the alkene at the surface.
The authors also demonstrated that nanocrystalline diamond thin lms
can be covalently modied with DNA olignucleotides [155]. This provides
an extremely stable, highly sensitive platform for subsequent surface
hybridization processes. A photochemical modication scheme is used in
which amine groups are formed on the diamond surface. The amine groups
serve as sites for DNA attachment. Once linked to the surface, hybridiza-
tion reactions between the DNA and uorescently tagged complementary
and noncomplementary oligonucleotides can be accomplished. The selec-
tivity between the matched and mismatched regions can be spectroscopi-
cally probed. The results are important because they demonstrate that
diamond is an ideal substrate for biological modication and sensing.
VIII. ELECTROANALYTICAL APPLICATIONS
One important electrochemical technology where diamond electrodes have
made a signicant impact is in the area of electroanalysis. CVD diamond
oers advantages over other electrodes, especially sp
2
carbon (e.g., glassy
carbon), in terms of linear dynamic range, limit of detection, response time,
response precision, and response stability. Some of the reported applica-
tions of diamond in electroanalysis are highlighted below. Unless stated
otherwise, all the diamond electrodes mentioned below are boron-doped,
microcrystalline thin lms deposited on a conducting substrate (e.g., Si).
A. Azide Detection
One of the rst demonstrations of diamonds usefulness in electroanalysis
was the oxidative detection of azide anion in aqueous media [29,30,37].
Sodium azide has been widely used commercially and, in the past, as an
inator for automotive airbags. Azide anion is highly toxic and presents a
FIG. 15. Diazonium salt reduction on a boron-doped nanocrystalline diamond
thin-lm electrode: (A) general reaction mechanism, (B) a series of cyclic vol-
tammetric i-E curves in 1 mM 4-nitrophenyldiazonium salt + 0.1 M TBABF
4
/
CH
3
CN during the surface derivatization step, and (C) a cyclic voltammetric i-E
curve for the derivatized surface in 0.1M TBABF
4
/CH
3
CN. Scan rate = 0.05 V/s.
2
.
health hazard, at relatively low levels, in the form of headaches, cyto-
chrome oxidase inhibition, and vasodilation. Therefore, industries pro-
ducing or using azide generally have tight controls on the levels of the anion
discharged in water.
Diamond provides a sensitive, reproducible, and stable response for
azide electrooxidation, leading to superior detection performance com-
pared with glassy carbon. The electrochemical reaction at diamond is
believed to be [29,30,37]
2N
3
!3N
2
2e
A peak-shaped oxidation response was observed in cyclic voltammetric

measurements (pH 7.2 phosphate buer) made with both electrodes, but
the background currents, upon which the faradaic response is measured,
were dramatically dierent. The oxidation peak potentials, E
p
ox
, were 1045
and 1100 mV, and the background-corrected peak currents, i
p
ox
, were
57 and 88 AA for glassy carbon and diamond, respectively. This 35%
dierence in current magnitude was an unusual case, as most diamond
electrodes yielded peak currents that were within 5% of those for glassy
carbon.
The background current for glassy carbon at 1200 mV, near the E
p
ox
for azide anion, was a factor of 200 higher than for diamond. The larger
background current results from a combination of oxygen evolution and
surface oxidation [29,30,37,38,45]. As a result, the azide oxidation signal is
superimposed on a large and rising background signal, while the response
for diamond is recorded on a low and unchanging background signal.
Azide anion was also detected by ow injection analysis (FIA) with
amperometric detection. Table 4 shows a comparison of data for diamond
and glassy carbon [37]. Clearly, diamond outperforms glassy carbon in
TABLE 4
Summary of FIA-EC Data for Azide at Diamond and Glassy Carbon Electrodes
Diamond Glassy carbon
Linear dynamic range (AM) 0.303300 1.03300
Sensitivity (nA/AM) 33 F 5 36 F 7
Detection limit (nM) S/N = 3 8 F 8 (0.3 ppb) 50 F 20 (2.1 ppb)
Reproducibility (% RSD) 0.55 620
Stability (response loss over 12 hr) 5% 50%
Source: Ref. 37.
Swain 220
terms of linear-dynamic range, limit of detection, response precision, and
response stability.
B. Trace Metal Ion Analysis
Diamonds electrochemical properties are ideally suited for the detection
of trace metal ions via anodic stripping voltammetry (ASV) [30,57,115]:
(1) a large overpotential for hydrogen evolution, (2) a large overpotential
for oxygen reduction, (3) no metal complexation with the diamond
surface, (4) excellent stability at extreme anodic and cathodic potentials,
and (5) chemical inertness and environmental friendliness. These prop-
erties make it superior, in some respects, to Hg and other alternative
electrode materials.
Metal deposition on diamond is an immensely complicated process
that is not fully understood. First, a particular metal must nucleate and
growon the surface. The sites at which this occurs, as well as the nucleation
and growth mechanismfor many metals, is unknown. In unpublished work
from our group, copper deposits have been observed to form equally well
on both the facets and grain boundaries of highly boron-doped micro-
crystalline diamond. The metal deposits by an instantaneous nucleation
and growth mechanism at low overpotentials and by progressive mecha-
nism at high overpotentials [125]. Silver deposits on microcrystalline
diamond have also been studied [97]. At low overpotentials, an instanta-
neous nucleation and growth mechanism was observed, while at high
overpotentials a progressive mechanism was operative. Zinc deposits on
microcrystalline diamond (unpublished work from our group) by a
progressive mechanism at both low and high overpotentials [125].
Second, when multiple metals are deposited simultaneously, as is the
case in a real stripping voltammetric measurement, not only is their
interaction with the diamond surface important, but equally critical is
their interaction with each other. There is a possibility of intermetallic
compounds or alloys forming, both of which will aect the oxidation or
stripping potential for each. When these heterogeneous deposits form, the
oxidation of a particular metal can occur from dierent sites on the
diamond surface or from another metal surface. Oxidation from these
multiple sites leads to peak broadening due to a spread in reaction kinetics.
Ideally, for this application, highly dispersed metal deposits of lowvolume,
without any intermetallic interactions, are desired. Even with these com-
plexities, it is supposed that diamond will become a useful electrode for the
determination of trace metal ions via anodic-stripping voltammetry.
The growing concerns about on-site monitoring in environmental
and clinical scenarios have prompted interest in Hg-free electrodes for
ASV. It has been shown that low ppb levels of Pb(II) (63 ppb), Cd(II) (34
ppb), and Zn(II) (98 ppb) can be detected with diamond by linear-sweep
voltammetry, using only a 5-minute deposition or preconcentration time
[30]. More recently, results for boron-doped nanocrystalline diamond have
revealed that these electrode materials can be used to sensitively and stably
to detect Ag(I), Cu(II), Pb(II), Cd(II), and Zn(II) in acetate buer (pH 4.5)
using dierential pulse voltammetry [15]. The preconcentration step
involved the application of 1200 mVfor 3 min (no stirring). The oxidation
peaks occurred at ca. 400, 110, 450, 710, and 1010 mV versus Ag/
AgCl, respectively, for 100 AM solutions of Ag(I), Cu(II), Pb(II), Cd(II),
and Zn(II), and shifted positive with increasing solution concentration
(i.e., increasing surface coverage). The limit of quantitation for the
individual metals was in the low ppb range. For instance, the limits of
quantitation (S/N > 3) for Ag(I), Cu(II), Pb(II), Cd(II), and Zn(II) were
0.11, 0.64, 2.1, 1.1, and 6.5 ppb, respectively. The limit of quantitation was
highest for Zn, and this is due, at least to some extent, to the hydrogen-
evolution reaction interfering with the coulometric eciency of the metal
deposition. Good response precision and stability (<4%) were seen for all
ve metals.
In another series of experiments, solutions containing mixtures of
Cd(II), Pb(II), and Ag(I) were analyzed using the same experimental
conditions. Figure 16 shows anodic stripping voltammograms for metal
ion concentrations ranging from 0.01 to 10 AM. The deposition was
performed at 1000 mV for 3 minutes (no stirring). The peaks were
identied based on the stripping peak potentials for the individual metals.
It can be seen that the oxidation peak potentials shift positive and increase
in width with increasing solution concentration (i.e., increasing surface
coverage). Also, the Pb(II) stripping peak splits into two components. The
more negative component is more apparent at higher solution concen-
trations, although the ratio of the two remains constant with increasing
coverage. One possible explanation for this split is metal stripping from
two kinetically dierent sites on the electrode. For example, the Pb may be
oxidizing from a homogeneous deposit on the diamond surface and
heterogeneous sites where multiple metals have codeposited. Another
possibility is the formation of an intermetallic compound. In a series of
separate experiments, it was observed that standard additions of Ag(I) to a
test solution of Pb(II) resulted in the formation of two Pb(II) stripping
Swain 222
peaks. Only one peak was observed when Pb(II) was tested alone.
Therefore, it is supposed that the peak splitting is caused by the formation
of a Ag-Pb intermetallic compound.
It is also observed that the Cd(II) stripping peak charge is suppressed
when co-deposited with Ag(I). Even though equimolar amounts of Pb(II)
and Cd(II) are present in solution, the stripping charge for Cd(II) is
signicantly less than that for Pb(II). It was observed in separate experi-
ments that the addition of Ag(I), Pb(II), or Cu(II) to a solution containing
Cd(II) caused some supression of the Cd(II) stripping peak, with the
greatest suppression seen after the addition of Ag(I). In fact, in the presence
of Cd(II), the stripping charge for Ag(I) is enhanced, suggesting that some
of the deposited Cd is stripping at the Ag oxidation potential. Therefore,
the suppression of the Cd(II) peak is tentatively attributed to formation of
an Ag-Cd intermetallic compound.
FIG. 16. Dierential pulse anodic-stripping voltammetric curves for Ag(I),
PB(II), and Cd(II) for a boron-doped nanocrystalline diamond thin-lm electrode
in 0.1 M acetate buer, pH 4.5. The metal ion concentrations are (a) 10, (b) 1, and
(c) 0.5 AM. Preconcentration at 1000 mV for 3 min (no stirring). (Reprinted
with permission from Chem. Mater. 2003, 15, 879. Copyright (2003) American
Chemical Society.) (From Ref 115.)
Even with the complications associated with the intermetallic com-
pound formation, the linear dynamic range extended from 0.01 to 100 AM
for Cd(II) and Pb(II) with linear regression correlation coecients of 0.992
each, and from0.001 to 50 AMfor Ag(I) with a linear regression correlation
coecient of 0.994. The limit of quantitation for Ag(I), Pb(II), and Cd(II)
in the mixture was 0.01 AM or 1, 2, and 1 ppb, respectively, at which the S/
N ratio of the Cd(II) peak was just over 3. Ag(I) and Pb(II) still showed a
good S/N (f30) at this concentration.
Manivannan et al. used diamond electrodes to detect Pb(II) in water
samples (pH = 1) by dierential pulse voltammetry, measuring concen-
trations down to 4 nM(0.8 ppb) [57]. The preconcentration step was rather
long at 1.0 Vvs. SCEfor 15 minutes. Prado et al. reported on the detection
of Pb(II) and Cu(II) at diamond electrodes using anodic stripping voltam-
metry [156]. The Pb(II) was detected by both anodic- (Pb metal) and
cathodic-stripping voltammetry. Pb(II) was detected in a river-water
sediment sample (pH = 1) and was quantied using a standard addition
method. The metal content in the sediment was found to be f200 mg/kg by
both anodic- and cathodic-stripping voltammetry. The electrode response
was improved after microwave activation (e.g., thermal treatment), with a
limit of detection of 21 ppb for a 20-second preconcentration time.
C. Nitrite Detection
Diamond electrodes are useful for the detection of other electroactive
anions, such as nitrite [30]. Nitrite is used in the manufacture of diazo dyes
and other organic compounds, dyeing and printing of textile fabrics, and in
meat curing and preserving processes. In recent years there has been
increasing public concern about levels of the compound in processed
foods. Exhaustive electrolysis revealed that the reaction proceeds as (30)
NO
2
H
2
O V NO
3
2H
2e
Ion chromatography (IC) (pH 7.2) with amperometric detection was

employed to separate and monitor nitrite. Some results are summarized
in Table 5.
The linear dynamic range is 5 orders of magnitude. The theoretical
detection limit (S/N = 3), based upon extrapolation of the linear-response
curve, is 7 nM (16 pg or 0.3 ppb). Experimentally, an ion chromatogram
for 50 ALof 0.01 AMnitrate had a peak easily detected with a S/Nof 5. The
response variability ranged from 2% at high concentrations to 8% at the
lowest injected concentration (10 nM).
Swain 224
D. NADH Detection
The electrochemical oxidation of nicotinamide adenine dinucleotide
(NADH) is of interest because it is a required cofactor in a large number
of dehydrogenase-based biosensors [60]. One of the major problems with
previously used electrodes is deactivation as a result of the irreversible
adsorption of oxidation products on the electrode surface. Another
problem inherent with the use of sp
2
carbon electrodes is the relatively
large background currents at the NADH oxidation potential. Rao et al.
recently showed that NADH can be quantitatively electrooxidized at
diamond at neutral pH [60]. They found that diamond electrodes oer
high resistance to deactivation and insensitivity to dissolved oxygen. Cyclic
voltammetry, amperometry, and rotating disk voltammetry were used to
characterize the oxidation reaction. Highly reproducible voltammetric
responses were observed. The diamond electrode response was stable over
several months of storage in ambient air, in contrast to glassy carbon
electrodes, which exhibited deactivation within 1 hour of air exposure. The
concentration limit of detection for NADH in the amperometric detection
mode was 10 nM (S/N = 7). Interestingly, NADH could be quantitatively
detected in the presence of an equimolar amount of ascorbic acid, a
common interferent. This speciation was made possible because E
p
ox
for
NADH is about 200 mV more positive than that for ascorbic acid in this
medium at hydrogen-terminated diamond.
E. Uric Acid Detection
Uric acid (2,6,8-trihydroxypurine, UA) and other oxypurines are the
principal products of purine metabolism. The normal concentration of
UA in the urine of a healthy human is in the low millimolar range, whereas
in the blood micromolar levels are normal. UA is electroactive and can be
oxidatively detected, but a major problem is interference from other
TABLE 5
Summary of IC-EC Data for Nitrate at Diamond
Linear dynamic range (AM) 0.011000
Sensitivity (nA/AM) 40
Detection limit (nM) S/N = 3 7 (0.3 ppb)
Response variability (%RSD) 28
Source: Ref. 30.
electroactive constituents, such as ascorbic acid. Ascorbic acid oxidizes at
potentials similar to those for UA at carbon electrodes. Popa et al.
discovered that, for reasons that remain unclear, electrochemically anod-
ized diamond electrodes are selective for the detection of UA, even in the
presence of high concentrations of ascorbic acid [61]. Dierential pulse
voltammetry and chronoamperometry were used to study the oxidation
reaction. Arather harsh anodic polarization of 75 minutes at 2.6 Vvs. SCE
in 0.1 MKOHwas used, which resulted in the passage of over 600 mC/cm
2
.
This treatment produced surface oxides that caused the oxidation peak for
ascorbic acid to shift more positively than the oxidation peak for UA,
resulting in some speciation. Using chronoamperometry, the authors
obtained linear-calibration curves for UA over a concentration range up
to 1 AM in 0.1 M HClO
4
, with the lowest concentration measured being 50
nM. Ascorbic acid in less than 20-fold excess did not interfere with the UA
response. An important aspect of this work was the use of the assay for
detecting UA in urine and serum samples from hospital patients without
any preparative treatment of the solution.
F. Histamine and Serotonin Detection
Histamine and serotonin, the latter also known as 5-hydroxytryptamine,
are important biogenic amines present in many food products and acting
as chemical messengers in biological systems [62]. The electrochemical
oxidation of these two compounds was investigated by Sarada et al. in
netural solution (pH 7.2) using diamond electrodes [62]. Highly repro-
ducible and well-dened cyclic voltammetric responses were obtained for
histamine, which has an E
p
ox
of 1.4 V vs. SCE. The voltammetric signal-to-
background (S/B) ratios at diamond were an order of magnitude higher
than those for glassy carbon at and above the 100 AM concentration level.
A linear dynamic range of 34 orders of magnitude was observed for
diamond in the voltammetric measurements with a limit of detection of 1
AM. Well-dened, peak-shaped voltammograms were also seen for sero-
tonin. No fouling or deactivation of the diamond electrode, i.e., by the
adsorption of reaction products, was observed during several hours of use.
FIA with amperometric detection was also used to assay these two
compounds at diamond electrodes. A detection limit of 0.5 AM (S/N =
13.8) was found for histamine, while a low, 10 nM detection limit was
observed for serotonin. Diamond exhibited a linear dynamic range from10
nM (S/N = 18.1) to 100 AM for serotonin and from 0.5 to 100 AM for
Swain 226
histamine. The response variabilities ranged from 3 to 7% for both
analytes (n = 15). The work demonstrated that diamond oers signicant
improvements over polished glassy carbon for the detection of these
bioanalytes.
G. Direct Electron Transfer to Heme Peptide and Peroxidase
Tatsuma et al. investigated the direct electron exchange between hydrogen-
terminated and oxygen-plasma treated diamond electrodes and adsorbed
heme undecapeptide and horseradish peroxidase in pH 7.4 phosphate
buer [157]. The primary objective of the work was to study the direct
electron transfer and to understand the inuence of the sp
2
and sp
3
carbon
electrode microstructures on the electrode kinetics. Comparison was made
between the responses for hydrogen- and oxygen-terminated diamond
electrodes, the edge plane of pyrolytic graphite, and glassy carbon. H
2
O
2
was added to the solution to chemically oxidize the adsorbed peptide or
peroxidase, and then the subsequent reduction of each biomolecule was
performed electrochemically at 300 mV vs. Ag/AgCl in pH 7.4 phosphate
buer. Both the modied sp
2
and sp
3
electrodes produced cathodic
currents that increased linearly with H
2
O
2
solution concentration between
0.1 and at least 10 AM. The dierence, however, between the electrodes was
the magnitude of the cathodic current and the background noise signal.
For both the peptide and peroxidase, the highest reduction currents and
lowest background signals were found for diamond with the best perform-
ance observed for the oxygen-plasmatreated surface. The oxygen-plasma
treatment introduces surface carbon-oxygen functionalities that, presum-
ably, facilitate strong adsorption of the biomolecules through dipole-
dipole or ion-dipole interactions. It was speculated that the best stability
was for biomolecules adsorbed through a cross-linking polymerization
process. Interestingly, the cathodic currents for the heme peptide-modied
oxygenated diamond were similar to those for edge plane pyrolytic
graphite, an electrode frequently used for the direct electron exchange
with redox proteins and enzymes due to its carbon-oxygen functional
groups and k-electron system. Also, the background currents for diamond
were signicantly lower. The responses for all electrodes increased linearly
with the H
2
O
2
solution concentration from 0.1 to 1000 AM.
Dierent results were obtained for the peroxidase-modied elec-
trodes. For this biomolecule, the edge plane of pyrolytic graphite and
glassy carbon produced larger cathodic currents than did either the
hydrogen- or oxygen-terminated diamond. The response for edge plane
pyrolytic graphite and glassy carbon increased linearly with the H
2
O
2
so-
lution concentration from 0.1 to 1000 AM, while the two diamond elec-
trodes exhibited a dynamic range only to 10 AM.
Clearly, the electrode reaction mechanism and kinetics for these two
biomolecules are dierent at diamond electrodes than they are at other SP
2
-
bonded carbon electrodes. Further research is needed to understand these
dierences. This work, however, reveals that direct electron exchange can
occur between biomolecules, such as peptides and proteins, and diamond
electrodes and that the response can be inuenced by the surface chemistry.
H. Cytochrome c Analysis
Direct electron transfer of horse heart cytochrome c was demonstrated
at a hydrophobic, hydrogen-terminated, boron-doped nanocrystalline
diamond thin lm electrode [119,124,158]. A quasi-reversible, diusion-
controlled cyclic voltammetric response was observed for untreated dia-
mond. The peak currents changed linearly with the concentration, and
importantly, there was no electrode fouling. The heterogeneous electron
transfer rate constant was estimated to be 1.1 10
-3
(F0.2) cm/s. This is an
important observation, considering that most reported measurements of
cytochrome c electrochemical kinetics have inferred the necessity of a
hydrophilic, negatively charged, and oxygen-rich electrode surface. It
means that hydrophobic interactions with the electrode surface are also
inuential in the electron transfer process. This preliminary work further
demonstrates that diamond electrodes may be useful for the study of
redox-active proteins.
Figure 17 shows cyclic voltammetric i-Ecurves for 1 and 25 AMhorse
heart cytochrome c in 50 mM NaCl + Tris HCl buer, pH 7.2, at a boron-
doped diamond thin lmdeposited on quartz. No faradaic response can be
seen above the background for the 1 AM concentration but a well-dened,
peaks-shaped response is seen for the 25 AM solution. At this can rate (0.1
V/s), the DE
p
is ca. 100 mVand i
p
red
is ca. 0.50 AA. The i
p
red
/i
p
ox
ratio is near 1
and was stable with multiple cycles. This response was observed without
any type of electrode pretreatment.
I. Carbamate Pesticide Detection
Rao et al. reported on the separation and amperometric detection of
several N-methyl carbamate pesticides using diamond electrodes [63]. The
Swain 228
pesticides (carbaryl, carboferran, methyl-2-benzimidazole carbamate,
bendiocarb) were separated by reversed-phase liquid chromatography in
an acetonitrile/phosphate buer moble phase (pH 2.3) and detected at a
positive potential of 1.45 V vs. Ag/AgCl. The limits of detection (S/N = 2)
ranged from 5 to 20 ppb, depending on the compound. The S/B ratios for
diamond were 510 times larger than for polished glassy carbon. The
response variability was estimated to be less than 2% (n = 15). Like the
azide detection discussed above, the sensitive and stable detection of these
pesticides is possible at diamond because of the wide working potential
window, the low background current, and the microstructural stability at
very positive potentials.
J. Ferrocene Analysis
Cyclic voltammetric measurements were made using well-characterized
microcrystalline, boron-doped diamond thin-lm electrodes to test the
materials responsiveness for ferrocene as a function of scan rate, solvent,
FIG. 17. Cyclic voltammetric i-E curves for 25 Am horse heart cytochrome c in
50 mM NaCl + Tris HCl buer, pH 7.2 at a boron-doped diamond thin-lm
electrode deposited on quartz. Scan rate = 0.1 V/s. Electrode area = 0.2 cm
2
.
and electrolyte composition [108]. This represents the rst signicant work
with a nonaqueous redox system at diamond. Apparent heterogeneous
electron transfer rate constants, k
app
j, of 0.042 F0.015, 0.048 F0.015, and
0.008 F 0.002 cm/s were observed in 0.1 M NaClO
4
/CH
3
CN, 0.1 M
TBAClO
4
/CH
3
CN, and 0.1M TBAClO
4
/CH
2
Cl
2
, respectively. These rate
constants, obtained for electrodes without any pretreatment, were similar
to those observed for freshly polished glassy carbon. The results demon-
strate that boron-doped diamond is a viable material for the electrochem-
ical analysis of nonaqueous analytes.
K. Aliphatic Polyamine Detection
One class of redox analytes for which diamond oers a somewhat unique
response, especially in the carbon electrode family, is the electrooxidation
of aliphatic polyamines, such as cadaverine (H
2
N-(CH
2
)
5
-NH
2
). Poly-
amines are ubiquitous components of all cells and are known to play
critical roles in the proliferation, dierentiation, maintenance, and neo-
plastic transformation of mammalian cells [159]. Polyamines are known to
be involved in angiogenesisa process essential for tumor growth and
metastasis (biological markers for cancer) [160,161].
Despite favorable thermodynamics, the oxidation of aliphatic poly-
amines in aqueous media is kinetically limited because the reactions
require transfer of oxygen from water, and most electrodes lack the
ability to support and sustain this complex mechanism [162164]. The
oxidation of amines in aqueous media has been typically limited to a few
select anode materials (Au, Ag, Ag-Au alloys, Ag-PbO
2
) in alkaline
media that are capable of transferring oxygen from water to the analyte
[165167]. Stable detection of the amines at these metals is often com-
plicated by electrode fouling with the reaction product(s), resulting in
poor response stability.
Importantly, these analytes can be quantitatively electrooxidized at
boron-doped diamond thin-lm electrodes without derivatization, the
use of pulsed voltammetric waveforms as is most often required for
detection at metal electrodes like gold, or fouling [36,39,43,110]. Boron-
doped diamond electrodes possess the requisite physicochemical proper-
ties needed to support the amine oxidation via an anodic oxygen-transfer
reaction (e.g., surface boron sites for amine coordination and localized
sp
2
carbon sites where the oxidant, OH
.
, is generated at lower overpo-
tential than the surrounding diamond matrix), and these properties can
Swain 230
be engineered into the material through variations in the deposition
conditions [36,39,43].
Figure 18 shows typical cyclic voltammetric i-E curves for 1.0 mM
(A) cadaverine (CAD), (B) putrescine (PUT), (C) spermidine (SPMD), and
(D) 0.8 mM spermine (SPM) in borax buer pH 11 (BBpH11) at a mi-
crocrystalline diamond thin lm deposited from a 0.5% CH
4
/H
2
ratio and
10 ppm B
2
H
6
. Similarly shaped curves were observed for several other
amines and polyamines (e.g., methylamine, ethylamine, propylamine,
ethylenediamine, 1,3-diaminopropane, 1,6-hexamethylenediamine, and
FIG. 18. Cyclic voltammetric i-E curves, background (dashed line) and total
current (solid line), for 1.0 mM (A) CAD, (B) PUT, (C) SPMD, and (D) SPM in
0.01 M borax buer/0.1 m NaCl, pH 11. The working electrode was a
microcrystalline diamond lm deposited from a 0.5% CH
4
/H
2
ratio and 10 ppm
B
2
H
6
. Scan rate = 0.1 V/s. Electrode geometric area = 0.2 cm
2
. (From Ref. 121.)
N-(3-aminopropyl)-1,3-propanediamine). The oxidation peak current for
each of the amines is clearly discernible from the background current. The
E
p
ox
values are 872, 868, 876, and 872 mV, and the E
1/2
values are 821, 817,
820, and 815 mV for CAD, PUT, SPMD, and SPM, respectively. The i
p
ox
values are 75, 93, 116, and 76 AA for 1 mM CAD, PUT, SPMD, and 0.8
mM SPM, respectively. A summary of the cyclic voltammetric data for
several amines and polyamines in this medium are presented in Table 6.
Whether a well dened peak shape is observed depends on the
physicochemical properties of the electrode. The E
p
ox
values are nearly
the same, independent of the amine molecular structure, indicating the
importance of OH
.
generation. The oxidation reactions all occur at a
potential near the onset of oxygen evolution. Very interestingly, there is an
oxidation peak present during the reverse sweep, which is clearly associ-
ated with the oxidation of amine. The current goes through a maximum at
about the same potential as on the forward sweep. At this potential,
consumption of OH
.
by the oxygen evolution reaction is minimal, and
these reactants are available for the oxygen transfer reaction. The current
during reverse sweep is lower than during the forward scan. This is partially
due to the depletion of the amine at the surface during forward scan, as at
TABLE 6
Cyclic Voltammetric Data for 0.1 mM Aliphatic Amine Oxidation at Diamond
Thin-Film Electrodes
Amine E
p
ox
(mV) E
1/2
(mV) Wave shape
Methylamine 865 801 Peak
Ethylamine 851 798 Peak
Propylamine 849 789 Peak
Ethylenediamine 848 780 Peak
1,3-Diaminopropane 860 797 Peak
Putrescine 851 770 Peak
Cadaverine 860 783 Peak
Spermidine 867 762 Peak/Wave
Spermine 869 769 Peak/Wave
Voltammetric data are for a 0.1 V/s scan rate. Reported potentials are versus a Ag/AgCl
reference electrode. Analytes were prepared in 0.01 M borax buer/0.1 M NaCl, pH11. E
1/2
values correspond to the potential at half the peak current.
Source: Ref. 121.
Swain 232
this scan rate there is not enough time for the new amine molecules to
diuse from the bulk solution and coordinate at the boron sites on the
electrode surface. Hence, a smaller number of molecules is available for
oxidation and a lower current results. The lower current during the reverse
sweep may also result from a limited number of surface boron sites
available for adsorption. Some time is necessary for desorption of the
oxidation products, freeing up boron sites for new amine coordination.
Cyclic voltammograms obtained at lower scan rates (510 mV/s) conrm
this supposition as it was observed that the current during the reverse
sweep tracked that for the forward sweep. The extent of hysteresis depends
on the scan rate, being larger at higher scan rates.
Repetitive potential cycling in the amine solution led to progressive
attenuation of the current. The voltammetric response could, however, be
regenerated by vigorous mixing and/or inserting a time delay interval from
3 to 5 minutes between the cycles. This behavior indicates there is no
permanent fouling of the electrode by the oxidation products, as in the case
for other electrodes. The current recovery is consistent with slow desorp-
tion of the product(s) from the surface.
The proposed mechanism for polyamine oxidation at diamond is
given below and is similar to the model for anodic oxygen-transfer
reactions put forth by Johnson and coworkers [162164].
S
ND
H
2
O !S
ND
OH H
OH radical generation at sp
2
sites
S
B
R X S
B
R
coordination of the amine at a boron surface site
S
ND
OH S
B
R !S
ND
S
B
ROH
oxygen-transfer reaction
The subscripts ND and B refer to nondiamond carbon and boron,
respectively. Boron-doped diamond thin-lm electrodes possess several
properties important for anodic oxygen-transfer reactions. First, high-
quality diamond lms are stable and resistant to corrosion in strongly
acidic and alkaline media. Therefore, at the anodic potentials used to
detect the polyamines, the electrode structure is stable. Second, lms may
contain sp
2
-bonded nondiamond carbon impurity distributed very locally
over the surface. These impurities can exist at the grain boundaries or as
extended defects within the diamond lm. These surface impurities, which
exhibit lower overpotential for oxygen evolution than does diamond, can
also be intentionally introduced into the lms by adjustment of the
diamond deposition conditions. This means that reactive OH
.
will be
generated locally at these sites at low overpotential and not to any
appreciable extent on the diamond lattice. Third, boron dopant atoms
located at the surface can serve as adsorption/coordination sites for the
lone pair of electrons on the N atoms of the polyamines. Boron atoms can
insert directly into the growing diamond lattice, but they can also clus-
ter and accumulate in the grain boundaries. The distribution of boron
atoms in the polycrystalline diamond lm is not homogeneous as dis-
cussed above. The polyamine adsorption/coordination at the boron sites
near the grain boundaries is believed to be important mechanistically, as
these are sites very near where the OH
.
is being generated at lower
overpotential.
The important roles of surface nondiamond carbon and boron on
the amine oxidation reaction mechanism have been determined [121].
Figure 19 shows cyclic voltammetric i-E curves for 1.0 mM CAD in
carbonate buer, pH 10 (CBpH10) at a 0.67% CH
4
/H
2
microcrystalline
diamond lm. Figure 19A shows the curve for a boron-doped, as-
deposited lm in the presence of 1 mM CAD along with the background
response (dashed line). The onset potential for oxygen evolution at the as-
deposited lm begins at about 1020 mV. The oxidation response for CAD
is well resolved from the background with an E
p
ox
of 990 mV, an E
1/2
of
920 mV, and an i
p
ox
of 32 AA for a 10 mV/s scan rate.
If the lm is acid-washed to remove the sp
2
-bonded nondiamond
carbon impurity from the surface and then hydrogen plasma treated to
rehydrogenate the surface, the voltammetric response in Fig. 19B is
observed. Clearly, the removal of the nondiamond carbon impurity causes
a total attenuation of the amine response [39]. The lack of an oxidation
response is due to the removal of the nondiamond carbon impurity at
which the OH
.
forms with lower overpotential than the surrounding
diamond. It is important to note that the acid-washing and rehydrogena-
tion leads to a higher overpotential for oxygen evolution with an onset
potential of about 1230 mV. The 200 mV dierence in potential between
the as-deposited and rehydrogenated electrodes is evidence for the suc-
cessful removal of sp
2
-bonded nondiamond carbon impurity. This result
provides compelling evidence for the importance of localized sp
2
-bonded
nondiamond carbon impurity sites in the amine oxidation reaction mech-
anism. It should be noted that this loss of nondiamond carbon impurity did
Swain 234
FIG. 19. Cyclic voltammetric i-E curves for 1 mM cadaverine in 0.01 M
carbonate buer/0.1 M NaClO
4
, pH 10.6, at a microcrystalline diamond lm
deposited from a 0.67% CH
4
/H
2
ratio. (A) As deposited and (B) acid-washed
and rehydrogenated diamond lm. Scan rate = 10 mV/s. (Reprinted with per-
mission from Anal. Chem., 71, 1188 (1999). Copyright (1999) American Chemical
Society.) (From Ref. 39.)
not have a deleterious eect on the electrode response for other redox
systems. For instance, DE
p
for Ru(NH
3
)
6
+3/+2
was largely unchanged.
Figure 20 shows cyclic voltammetric i-E curves for 1 mM CAD in
0.01 M borax buer, pH 10.6, along with the corresponding background
current, at (A) a nanocrystalline diamond thin lm deposited with inten-
tionally added boron and (B) a nanocrystalline lm deposited without any
intentionally added boron. Figure 20A shows a well-dened oxidation
wave, clearly distinguishable from the water electrolysis background
current, for the boron-doped lm. E
p
ox
is 927 mV and i
p
ox
is 136 AA. The
amine oxidation current scales with the doping level, at least to the extent
that this can be probed without altering the electrical conductivity. There
is a second oxidation peak on the reverse sweep at about the same
potential as on the forward sweep. This wave shape has been previously
discussed and is consistent with a redox reaction that involves a compet-
itive surface interaction [36,39,43]. Correspondingly, there is a poorly
dened oxidation wave for the unintentionally doped lm (Fig. 20B).
During the forward and reverse sweeps, the current is slightly increased
above the background, but there is no well-dened oxidation curve. The
absence of an oxidation response for CAD is not due to poor lm electrical
conductivity, as the DE
p
values for Fe(CN)
6
3/4
, Ru(NH
3
)
6
+2/+3
, and
IrCl
6
2/3
were 63, 68, 60 mV, and for methyl viologen, 4-tert-butyl
catechol, and Fe
+2/+3
were 60, 764, 396 mV. This result demonstrates
the importance of surface boron in the oxidation reaction mechanism of
aliphatic polyamines at diamond.
FIA and reversed-phase liquid chromatography (LC) with ampero-
metric detection were used to detect ethylene diamine, cadaverine (CAD),
putrescine (PUT), spermine (SPM), and spermidine (SPMD). The elec-
trode performance was evaluated in terms of the linear dynamic range,
limit of quantitation, response variability, and response stability [36,39,
43,110]. The polyamines were detected in alkaline solution (pH 10) with a
linear dynamic range from 1 to 1000 AM and a limit of quantitation of 1.0
AM, or 20 pmol injected (S/Nz3), for CAD, PUT, and SPMD. For SPM,
a linear-dynamic range from 0.32 to 1000 AM and a limit of quantitation
of 0.32 AM, or 6.4 pmol, were observed. A response variability as low as 2
4% was found when implementing a f3-minute delay period between
injections. The long-term response stability was good, with no evidence
for any progressive response attenuation or fouling by the reaction
product(s), although small deposits were observed to accumulate on the
electrode surface during extended use for several weeks [43]. Preliminary
Swain 236
FIG. 20. Cyclic voltammetric i-E curves for 1 mM CAD, along with the
corresponding background current (dashed line), in a 0.01 M borax buer, pH
11.2, at (A) a boron-doped nanocrystalline diamond thin-lm electrode and (B) a
nanocrystalline diamond thin-lm electrode deposited without intentionally
added boron. Scan rate = 0.1 V/s. Electrode geometric area = 0.2 cm
2
. (From
Ref. 121.)
chromatographic results demonstrated the possibility of separating and
detecting the polyamines by a simple reversed-phase methodology at
constant applied potential with detection limits in the high nM range
[110,121].
IX. ELECTROSYNTHESIS AND ELECTROLYTIC
WATER PURIFICATION
Another important electrochemical technology in which diamond elec-
trodes have made a signicant impact is in the area of electrolytic water
purication and decontamination. Electrochemical approaches are quite
attractive for chemical contaminant decontamination and remediation,
but often are limited by the stability of the electrode microstructure, mor-
phology, and responsiveness. Successful operation of any kind of electro-
chemical-based decontamination (oxidation) strategy would require a
dimensionally stable anode that can operate in chemically harsh environ-
ments with unchanging responsiveness, resist fouling by matrix compo-
nents (organic matter), possess a large overpotential for oxygen evolution,
and eciently support anodic oxygen transfer reactions. Diamond elec-
trodes meet all of these requirements. Decontamination and remediation
of organic and inorganic waste can be brought about by either direct or
indirect electrooxidation reactions. Chemical contaminants can be directly
oxidized at the anode or indirectly through the electrogeneration of oxi-
dants (e.g., OH
.
or Cl
2
) that subsequently chemically oxidize the contam-
inant. One mechanism for chemical contaminant oxidation is through
reaction with OH
.
generated during the initial stage of oxygen evolution.
Ideally, organic and inorganic chemicals are completely oxidized to CO
2
,
or at least partially oxidized to products that are less toxic than the original
form.
Agent chemical oxidant Z CO
2
A patent for the electrochemical treatment of various industrial
wastes, prior to discharge, was obtained by Carey et al. and represented
one of the rst demonstrations of diamond in electrochemistry [100].
Various chemical contaminants (e.g., phenols) from a Kodak industrial
lm-making process were oxidatively treated with diamond anodes. Up to
90%of the euent was oxidized in their electrolytic approach. Hagans and
coworkers used diamond anodes to oxidize phenol all the way to CO
2
in
acidic media [104]. The total organic content was eectively reduced from
Swain 238
f1% to less than 0.1% with no observable decrease in decomposition rate
over time. Gherardini et al. demonstrated the complete combustion of 4-
chlorophenol with high current eciency at boron-doped diamond anodes
[168]. Iniesta et al. found that it was possible to electrochemically oxidize 3-
methylpyridine partially to nicotinic acic or completely to CO
2
at diamond
anodes, depending on the current density (i.e., applied potential) [169].
Panizza and coworkers examined the oxidation of 2-naphthol in acidic
media at diamond anodes and observed that oxidation products adsorbed
and fouled the surface at low potentials, but complete incineration to CO
2
could be achieved at high potentials [170]. Foti et al. investigated the
electrochemical remediation of model organic systems at boron-doped
diamond thin lms [102]. Complete oxidation to CO
2
was observed with no
degradation of the diamond electrode.
Our group has shown that boron-doped diamond thin lmelectrodes
exhibit superb morphological and microstructural stability during anodic
polarization at elevated temperature in an acidic uoride media [45] and
that chlorine can be stably generated at current densities as high as 0.5 A/
cm
2
[47]. Katsuki et al. evaluated the eectiveness of boron-doped
diamond electrodes for water electrolysis in acidic solution [101]. They
discovered that ozone can be produced with a current eciency of a few
percent at ambient temperature. Most importantly, the ozone generation
produced no degradation of the electrode structure. Okino et al. studied
the electrouorination of 1,4-diuorobenzene at boron-doped diamond
thin lm electrodes in nonaqueous media [105]. Haenni et al. recently
reported on the use of diamond electrodes in a DiaCell
R
reactor to stably
and eciently generate chlorine in situ as a disinfectant for swimming pool
water [106].
X. OPTICALLY TRANSPARENT ELECTRODES FOR
SPECTROELECTROCHEMISTRY
The use of electrically conductive diamond as an optically transparent
electrode is a new eld of research [50,52,117,118]. Diamond possesses
attractive qualities as both an electrode and an optically transparent mate-
rial, making it an obvious choice for use as an OTE in spectroelectro-
chemical measurements. Diamond OTEs exhibit several technologically
useful properties: (1) the possibility of transmission measurements from
the near-UV to the far-IR (0.225100 Am); (2) low background current;
(3) wide working potential window; (4) good responsiveness for many
analytes without activation pretreatment; (5) electrical and optical stabil-
ity in most aqueous and nonaqueous solution environments during both
cathodic and anodic polarization (z10 mA/cm
2
); and (6) resistance to
fouling.
Diamond OTEs can be fabricated in several dierent forms, useful
for either transmission or reection measurements. Importantly, the opti-
cal and electrical properties can be engineered, to varying degrees, through
adjustments in the CVD conditions. One electrode form is a mechanically
polished (f30 nm roughness), free-standing disc of diamond. An optical
image of such an electrode is shown in Fig. 21. A disc can be prepared by
growing a thick lmof diamond on a refractory metal substrate, like Mo. If
the sample is cooled rapidly fromthe growth temperature, then the lmwill
separate from the substrate due to stress (compressive) resulting from the
dierence in thermal expansion coecients for diamond and the substrate.
The polycrystalline disc is then mechanically polished smooth to reduce the
surface roughness (i.e., light scattering). We previously reported on UV/
Vis transmission spectroelectrochemical measurements of the aqueous
FIG. 21. Optical images of several free-standing, boron-doped diamond disc
OTEs. Diamond discs fabricated and provided for the research by Dr. James E.
Butler at NRL.
Swain 240
redox analytes, ferri/ferrocyanide, and methyl viologen, using such an
electrode (8 mm diameter and 0.38 mm thick) [52].
Another electrode form is a thin lm deposited on an optically
transparent nondiamond substrate, such as undoped Si for IR or quartz
for UV/Vis spectroelectrochemical measurements. Diamond deposition
on Si for IR OTEs is rather straightforward and involves growth con-
ditions similar to those described above. The resulting lms are 24 Am
thick with micrometer-sized grains of diamond randomly oriented over the
surface. Deposition of thin lms of diamond on quartz is a little more
involved [118]. Figure 22 shows an optical image of a diamond/quartz
OTE. The lm has a blue hue to it due to the boron doping level. It is
FIG. 22. Optical image of a diamond/quartz OTE. Ruler is behind the dia-
mond/quartz OTE.
important to keep the diamond crystallite size smaller than the wave-
lengths of light to minimize scattering. This is achieved by mechanically
polishing the quartz surface with diamond powder (10 nm) in order to
produce a high density of scratches. This serves to increase the nucleation
rate such that small grains of diamond form in a continuous fashion over
the substrate surface. The lm thickness is 0.51 Am with aggregates of
diamond grains ca. 100200 nm in diameter. We recently described UV/
Vis-transmission spectroelectrochemical measurements of chlorpromazine
(CPZ), using a thin lm of diamond deposited on quartz [118]. Martin and
Morrison reported on a creative IR spectroelectrochemical approach for
characterizing carbon-oxygen functional groups formed on the surface of
diamond electrodes during anodic polarization, in which the OTE was a
thin lm of diamond deposited on undoped Si and employed in an internal
reection measurement [117].
The eld of spectroelectrochemistry dates back to 1964 when the rst
work was reported by Kuwana, et al. [171]. The authors described the use
of indium tin oxide (ITO), coated on glass, as an OTE for UV/Vis-
transmission spectroelectrochemical measurements of several inorganic
and organic redox analytes. The eld has evolved over the years to en-
compass research on a variety of topics: new electrode development, thin-
layer cell design, and construction. Additionally, many novel measurement
schemes have been developed for material characterization, structure-
function studies of biomolecules, and environmental contaminant moni-
toring [172175]. Several dierent OTEs have been used along the way,
with the most common type being ITO [173].
While highly functional for many applications, ITO has some
material property limitations. These include (1) highly variable optical
and electrochemical properties from source to source, (2) intolerance of
cathodic polarization, and (3) instability during exposure to strongly acidic
or alkaline media and to some chlorinated organic solvents. In contrast,
the optical and electrochemical properties of diamond OTEs are (1)
reproducible frombatch to batch and (2) very stable during either cathodic
or anodic polarization in all types of chemical environments. Clearly, the
major attributes of diamond OTEs are the wide optical window and the
reproducible and stable responsiveness, even in aggressive electrochemical
environments.
Several factors inuence the optical transparency of CVD diamond
in dierent regions of the electromagnetic spectrum [176,177]. These
include (1) reectance losses in the UV/Vis due to a high refractive index
Swain 242
(2.41 at 591 nm); (2) absorption attributable to the boron doping in the
near-IR extending into the IR and visible regions of the spectrum; (3)
scattering losses in the UV/Vis due to the rough, polycrystalline surface of
most lms or scattering centers in the lms; and (4) absorption by chemical
impurities and defect centers, such as sp
2
-bonded nondiamond carbon at
extended defects and adventitiously incorporated nitrogen and silicon. A
key to making a functional diamond OTE is obtaining a balance between
the boron-doping level required for the desired electrical conductivity and
the maintenance of sucient optical transparency [52,117,118].
Figure 23 shows a series of transmission spectra for diamond OTEs
in the UV/Vis and IR regions of the electromagnetic spectrum. Spectra for
several dierent electrode forms are shown for comparison. Figure 23A
shows transmission spectra for (1) ITO on quartz, (2) boron-doped nano-
crystalline diamond on quartz, (3) mechanically polished and boron-doped
diamond on an optically pure, white diamond substrate, and (4) a free-
standing, boron-doped, and mechanically polished diamond disc. The ITO
substrate has about 80% transparency at wavelengths in the visible region
of the electromagnetic spectrumwith an absorption edge starting at ca. 300
nm. The transmission of the boron-doped diamond thin lm on quartz has
an absorption edge at ca. 225 nm and a relatively constant transparency of
4050% between 300 and 800 nm. The reduced light throughput in this
region is attributable mainly to reectance losses. Given the large refractive
index, this is about the maximum transmission that can be expected for a
thin lm of diamond on quartz (for this boron-doping level, f10
19
B/cm
3
,
f0.1 ohm-cm). The polished boron-doped lm deposited on the optically
pure white diamond and the polished boron-doped, free-standing disc have
lower levels of light transmission than the diamond/quartz OTE, with a
continuous decrease in the transmitted light at longer wavelengths. This
results from IR absorption by the boron acceptor band which can extend
into the visible, and this is the reason such lms appear blue. The trans-
mission in the far-visible and near-IR decreases proportionally with the
boron-doping level.
Figure 23B presents IR transmission spectra for (5) an optically pure
and mechanically polished white diamond disc, (6) an undoped and
polished (both sides) Si substrate, and (7 and 8) moderately and heavily
boron-doped microcrystalline diamond thin lms deposited on the
undoped Si. The white diamond is relatively free of structural defects
and chemical impurities. There is reduced transparency between 2500 and
1500 cm
1
due to the two-phonon absorption. Diamond lms with more
FIG. 23. Transmission spectra for dierent materials in the (A) UV/Vis and (B)
IR regions of the electromagnetic spectrum. The electrodes in (A) are (1) a thin lm
of ITO on quartz, (2) a thin lm of boron-doped nanocrystalline diamond on
quartz, (3) a thin lm of mechanically polished and boron-doped diamond on an
optically pure, white diamond substrate, and (4) a free-standing, boron-doped, and
mechanically polished diamond disc. The electrodes in (B) are (5) an optically pure
and mechanically polished white diamond disc, (6) an undoped and polished (both
sides) Si substrate, and (7 and 8) moderately and heavily boron-doped microcrys-
talline diamond thin lms deposited on the undoped Si. (Reprinted with permission
from Interface 2003, 12, 33. Copyright (2003) The Electrochemical Society, Inc.)
(From Ref. 158.)
Swain 244
structural defects and chemical impurities lead to loss of crystal symmetry.
This causes a more intense one-phonon absorption between 1500 and 1000
cm
1
and increased absorption between 3000 and 2700 cm
1
, due to
vibrations of CUH bonds. With the exception of some weak absorption
below 1200 cm
1
, the Si substrate possesses about 75% transmittance
throughout the entire mid-IR. The reduced light throughput is due to
relatively constant reectance losses (f30%) over the entire wavelength
range.
When boron is introduced into the lattice, creating a conductive OTE
material, the transparency decreases, as shown in Fig. 23B (7 and 8),
particularly above 1500 cm
1
. Introduction of substitutional boron breaks
the lattice symmetry, resulting in increased absorption due to impurities
and one-phonon modes [177]. Specically, boron gives rise to absorption
bands at 2460 and 2790 cm
1
, due to the electronic transitions from the
ground to the rst and second excited states of the dopant atoms, res-
pectively [178]. However, for more heavily doped lms, a broad continuum
develops beyond 2000 cm
1
, as a result of the interaction of the boron
centers, resulting in boron acceptor level degeneracy. The mode at 1290
cm
1
is attributed to the one-phonon vibronic absorption induced by the
boron acceptor centers [178]. Despite the marked decrease in transmittance
observed in the region above 1500 cm
1
from boron incorporation, it is
important to note that the moderately doped lm retains a high level of
throughput in the region below 1500 cm
1
.
The electrical properties of the diamond lms or free-standing discs
are largely determined by the boron-doping level. Resistivities of useful
diamond OTEs are in the range of 0.50.05 V-cm. Boron-doping levels
associated with this resistivity are ca. 110 10
19
B/cm
3
, as determined by
boron nuclear reaction analysis measurements. Very preliminary Hall
eect measurements for the diamond/quartz (Fig. 23A, 2) and diamond/
Si (Fig. 23B, 7) OTEs have revealed carrier concentrations between 10
16
and 10
18
cm
3
and carrier mobilities (holes are the majority carrier in
boron-doped lms) of 1100 cm
2
/V-s.
One organic redox system that we have studied by transmission
spectroelectrochemical measurements is chlorpromazine (CPZ) [118]. The
electrochemistry and spectroscopy of CPZ have been investigated exten-
sively over the years [22,30,179]. Figure 24A shows a cyclic voltammetric
(background corrected) i-E curve for 100 AM CPZ + 10 mM HClO
4
. The
measurement was made in a specially designed, thin-layer electrochemical
cell with a path length of 150 Am and a cell volume of 5 AL [118]. The scan
FIG. 24. (A) Background-corrected, thin-layer voltammetric i-E curve for a
diamond/quartz OTE in 0.1 mM CPZ + 10 mM H
2
SO
4
. Scan rate = 2 mV/sec.
The lm was deposited from a source gas mixture of 0.5% C/H and 10 ppm B
2
H
6
.
(B) A series of UV-Vis absorbance dierence spectra, using the same lm, for 0.1
mM CPZ + 10 mM HClO
4
, as the potential was stepped from 0.32 to 0.47 V vs.
Ag-QRE. (Reprinted with permission from Anal. Chem. 2002, 74, 5924. Copy-
right (2002) American Chemical Society.) (From Ref. 118.)
Swain 246
was initiated at 0.20 V and scanned in the positive direction at 2 mV/s.
Well-dened oxidation and reduction peaks are observed, which corre-
spond to the following electrochemically reversible redox reaction:
CPZ X CPZ
.
e
A subsequent one-electron transfer occurs at more positive potentials,

f0.9 V, to form the dication, CPZ
+2
, which is quickly hydrolyzed [179].
The electrode reaction kinetics for this redox system (CPZ/CPZ
.+
) at
diamond are mainly inuenced by the density of electronic states at the
formal potential [22,30]. Rapid electrode-reaction kinetics have been
observed for boron-doped diamond electrodes, with no evidence of any
molecular adsorption [22,30]. The Q
p
ox
/Q
p
red
ratio for the CPZ/CPZ
.+
redox reaction is ca. 1. The i
p
ox
and i
p
red
values varied linearly with the scan
rate, while Q
p
ox
and Q
p
red
were independent of scan rate. These trends are
predicted for thin-layer voltammetric behavior.
Figure 24B shows a series of potential-dependent UV-Vis dierence
spectra for CPZ during the one-electron transfer reaction to form CPZ
.+
.
CPZ has an absorbance maximum at 253 nm, due to a k!k* transition (q
o
f10,000 AUcm
1
M
1
), while the radical cation (CPZ
.+
) has a maximum
at 275 nm, presumably also due to a k!k* transition (q
o
f20,000 AU
cm
1
M
1
) [180]. In neutral solution, the radical cation rapidly dispro-
portionates, limiting its spectroscopic observation. As a result, measure-
ments were carried out at pH 2, where the radical cation is more stable.
The spectra were obtained after 1-minute equilibration times at each
potential. It can be seen that as the potential is shifted stepwise from 320
to 470 mV, the peak at 253 nm progressively decreases and the peak at 275
nm increases. The spectra shown were background-corrected by subtract-
ing the spectrum for the fully reduced CPZ. When overlaid, the spectra
show well-dened isobestic points. An isobestic point near 260 nm
indicates that the species responsible for the absorbance peaks, on either
side of the point, are stoichiometrically related. In this case, these peaks are
attributed to CPZ and CPZ
.+
. The spectral trends were completely re-
versible with cycling.
Diamond OTEs are also useful for studying the electrochemical and
optical properties of important biomolecules, like cytochrome c. We
recently reported that boron-doped microcrystalline diamond thin lm
electrodes are quite responsive for horse heart cytochrome c, exhibiting a
very active and stable cyclic voltammetric response without any pretreat-
ment or surface modication [119,124]. Heterogeneous electron-transfer
rate constants in the mid to high 10
3
cm/s range are typical for both
boron-doped microcrystalline and nanocrystalline thin lms. The observa-
tion of a nearly reversible, diusion-controlled response at a hydrophobic
(hydrogen-terminated) and uncharged (no ionizable surface carbon-oxy-
gen functional groups) electrode contrasts with the conventional wisdom
regarding heterogeneous electron transfer of this protein, as the electrode
kinetics are known to be strongly dependent on a combination of interfa-
cial electrostatic and chemical interactions at other electrode surface [176].
Figure 25A shows a cyclic voltammetric i-E curve for 25 AM horse
heart cytochrome c in 1 mM Tris HCl buer (pH 7) containing 20 mM
NaCl [124]. The measurement was made in a specially constructed, thin-
layer electrochemical cell with a path length of approximately 100 Amand a
cell volume of approximately 6 AL. The diamond OTE was a free-standing
diamond disk (380 Am thick and 8 mm in diameter). The scan was initiated
at 0.15 V and recorded at 2 mV/s. The oxidation and reduction peaks
present at 0.075 and 0.010 V(DE
p
=85 mv) are not as well dened as they
are for other redox systems at this same electrode (e.g., ferrocene). The
cathodic current at 0.15 V is believed to be due to some residual oxygen
reduction, as the thin layer cell could not be eciently deoxygenated with
nitrogen.
Figure 25B shows a series of UV/Vis dierence spectra (reduced
minus oxidized) for 500 AM cytochrome c in 10 mM Tris HCl, pH 7, buer
containing 0.3 M NaCl [124]. The measurements were made in the same
thin-layer cell used for the voltammetry in Fig. 25A. The series of spectra
were collected by measuring the absorption spectrum for the progressively
reduced form of the protein in 50 mV increments from 0.30 to 0.20 V and
subtracting the spectrum for the fully oxidized form at 0.40 V. The spectra
were obtained after a 1-minute equilibration time at each potential. A
series of positive and negative peaks are observed at the dierent poten-
tials. The arrows depict the positive peaks that correspond to the reduced
form of the protein, and the peaks increase in amplitude as the potential is
made more negative.
The UV/Vis spectra of heme-containing proteins, such as cyto-
chrome c, contain two major bandsthe Soret and a-bands. The most
intense of these two is the Soret band, which occurs between 400 and 450
nm, depending on the nature of the heme iron and oxidation state of the
protein [181]. Both bands are attributed to the k ! k* transitions of the
heme and are, therefore, very sensitive indicators of the heme iron oxi-
dation state. The spectra presented, specically between 400 and 600 nm,
Swain 248
FIG. 25. (A) Cyclic voltammetric i-E curve for 25 AM horse heart cytochrome c
in 1 mM Tris HCl buer, pH 7, containing 20 mM NaCl at a boron-doped, free-
standing diamonddisc (see Fig. 23A). (B) UV/Vis reduced minus oxidized (at 0.4 V)
dierence absorbance spectra for horse heart cytochrome c collected at 0.050 V
intervals from 0.30 to 0.20 V OTE Fig. 23A, D 3.5 mM cytochrome c in 100 mM
phosphate buer, pH 7.4. Optical pathlength was 40 Am. (C) Reduced minus
oxidized (0.30 V) FTIR dierence absorbance spectra for cytochrome c acquired at
(A) a Au grid OTE and (B) a boron-doped diamond thin lm on undoped Si. The
noise spectrum is shown in C. 3.5 mM cytochrome c in 0.3 M NaCl + Tris HCl
buer, pH 7. (From Refs. 124 and 158.)
are the same as those reported previously [182] and indicate that the direct
electrochemical titration of this protein is possible. It is important to note
that the response is stable and reproducible with repeated cycling. The
stable isobestic points further indicate that no protein denaturation
occurred.
One additional spectral feature to be pointed out is the accessible
wavelength range possible with the diamond OTEs. The spectrum shows
that the region between 200 and 400 nm contains two absorption bands,
the Nand Lbands, which, like the Soret and a-bands, are attributed to k!
k* transitions of the heme Fe [183]. Many characteristic amino acid ab-
sorption transitions occur below 400 nm, such as tyrosine absorption at
280 nm. This range has been exploited in spectroelectrochemical measure-
ments of ferrocene at diamond OTEs [127].
The large Soret bands report on the oxidation state of the heme Fe,
but the absorbance of the a-band is routinely used to calculate the
cytochrome c concentration and to monitor heme reduction [184]. This
is due to the fact that the 550 nm absorption band is characteristic of the
reduced form of the protein. It is not inuenced by nearby overlapping
bands, as is the case for the Soret region where the oxidized and reduced
forms give rise to strong, broad absorption bands of nearly equal intensity.
The absorbance intensity at 550 nm, obtained from the direct redox
titration of cytochrome c as a function of the electrode potential, was
found to be a near perfect t to the Nernst equation. The formal reduction
potential determined from the plot was 0.072 V, which is in good
agreement (F0.01 V) with the accepted value of 0.06 V(vs. Ag/AgCl) [185].
Figure 25C(curve 2) shows a reduced (0.20 V) minus oxidized (0.30
V) IR dierence spectrum for 3 mM horse heart cytochrome c in 10 mM
Tris HCl buer, pH 7 at a diamond/Si OTE. All spectra presented
represent an average of 800 scans collected at 10jC with a path length of
10 Am. The electrochemical dierence IR spectrum has previously been
reported using a Au minigrid OTE [182]. This enabled us to evaluate the
performance of the new diamond OTE with respect to that of a conven-
tional OTE. The gure also contains our own spectroelectrochemical
measurement of the protein using a Au minigrid OTE (curve 1). Compar-
ison of the spectra for the diamond/Si and Au OTEs reveals excellent
agreement in that all features are retained [182]. There are many peaks in
the dierence spectrumrepresentative of amino acid side chain and peptide
backbone vibrations. A comprehensive interpretation of the spectrum is
beyond the scope of this manuscript. Finally, the gure shows the back-
Swain 250
ground-noise spectrum for the diamond OTE (curve 3). It is clear that the
background signal is at and relatively featureless, even in the areas of
strong water absorbance.
XI. ADVANCED ELECTROCATALYST
SUPPORT MATERIALS
Electrodes consisting of supported metal catalysts are used in electrosyn-
thesis and electrochemical energy conversion devices (e.g., fuel cells).
Nanometer-sized metal catalyst particles are typically impregnated into
the porous structure of an sp
2
-bonded carbon-support material. Typical
carbon supports include chemically or physically activated carbon, carbon
black, and graphitized carbons [186]. The primary role of the support is to
provide a high surface area over which small metallic particles can be
dispersed and stabilized. The porous support should also allow facile mass
transport of reactants and products to and from the active sites [187].
Several properties of the support are critical: porosity, pore size distribu-
tion, crush strength, surface chemistry, and microstructural and morpho-
logical stability [186].
A limitation of sp
2
-bonded carbon supports is their susceptibility to
corrosion and microstructural degradation during anodic polarization.
One possible carbon electrochemical corrosion reaction is [188]
C 2H
2
O !CO
2
4H
4e
The corrosion rate is strongly dependent on the electrode potential,

electrode microstructure, electrolyte composition, and pH. The rate is also
aected, at a given potential, by other factors, such as temperature and
vapor pressure [188]. In general, high temperature, high pressure, and high
operating potentials result in an increased rate of corrosion. Importantly,
electrochemical corrosion, even to a slight extent, can produce micro-
structural degradation and surface chemical changes, which generally lead
to lost catalytic activity, increased electrode resistance, or even catastro-
phic electrode failure.
Conductive sp
3
-bonded diamond is being developed as an advanced
catalyst support material. Boron-doped diamond thin-lm electrodes
possess excellent properties for this application, such as electrical conduc-
tivity, chemical inertness, extreme corrosion resistance, and dimensional
stability. Compared with more commonly used sp
2
-bonded carbon mate-
rials, diamond is highly resistant to electrochemical corrosion. For exam-
ple, polycrystalline diamond can withstand anodic current densities of 0.1
A/cm
2
, or greater, for 12 h in an acidic chloride media (E= 34 Vvs. SCE)
without any structural degradation [47]. The microstructure is also very
stable during polarization in 0.1 M HNO
3
+ 0.1 M NaF at 50jC, con-
ditions that lead to catastrophic failure of HOPG, Grafoilk, and glassy
carbon [45]. This makes diamond an ideal electrode for high-current
density electrolysis, particularly under demanding conditions (i.e., complex
matrix, high-current density and potential, high temperature, extremes in
pH, etc.).
Diamond is being used as an electrocatalyst support in the form of a
metal/diamond composite electrode [4851,125,126]. Nanometer-sized Pt
particles can be incorporated into the surface microstructure of micro-
crystalline diamond thin lms via a sequential diamond growth/Pt depo-
sition/diamond growth fabrication procedure. The dispersed Pt particles
are stabilized by the growth of a very thin diamond lm around their base
and are in good electrical communication with the current-collecting
substrate through the boron-doped diamond support. The research goal
is to develop state-of-the-art dimensionally stable, high surface area
catalytic composite electrodes that can withstand high temperatures
(200jC) and current densities (10 A/cm
2
) without structural degradation,
change in electrical resistance, or catalytic activity loss.
A. Composite Electrode Fabrication and Characterization
Figure 26 illustrates the three-step fabrication procedure for preparing
the composite electrodes. A boron-doped diamond thin lm is initially
deposited onto a substrate for 12 hours. The substrate could be planar in
the form of a plate of Si or a high surface area metal mesh (e.g., Mo).
The lm thickness at this stage is ca. 5 Am. The diamond growth is then
stopped and the substrate cooled to less than 300jC in the presence of
atomic hydrogen. After completely cooling to room temperature, the
coated substrate is removed from the reactor, and metal particles are gal-
vanostatically deposited (either continuous or pulsed) onto the surface
from a 1 mM K
2
PtCl
6
+ 0.1 M HClO
4
solution at a current density of
0.5 mA/cm
2
. Mass transport of reactant is by a combination of diusion
and convection, as nitrogen is slowly bubbled through the solution dur-
ing the deposition. Typical metal deposition times are 100400 s. Highly
dispersed, nanometer-sized metal particles are deposited over the lm sur-
face. The Pt-coated diamond lm is then placed back into the reactor
Swain 252
and boron-doped diamond deposited for an additional 3 hours. The sec-
ondary-growth lm thickness is on the order of 0.3 Am, eectively
entrapping the metal particles by anchoring and stabilizing them within
the microstructure. Diamond nucleates and grows, to some extent, on
both the Pt and the surrounding diamond matrix; however, the growth
rate on the diamond is much greater than on the Pt. Some of the smaller Pt
particles are completely covered during the process, as has been conrmed
in cross-sectional SEM measurements, but if the conditions are optimized
then the complete coverage of particles is minimized. Hydrocarbon
deposits also accumulate on the Pt surface, but these can be eectively
removed by potential cycling between the hydrogen and oxygen evolution
regimes.
FIG. 26. Fabrication process for the Pt/diamond composite electrodes. (From
Ref. 125.)
The concept behind the composite electrode is similar to that put
forward by Callstrom and coworkers in their metal-doped glassy carbon
studies [189,190]. The stable anchoring of the metal particles likely results
from a combination of physical entrapment by the secondary lm growth
and the formation of C-metal bonds. Metal deposits normally adhere
weakly to the hydrogen-terminated diamond surface. This is because of
C-H bond is strong and minimal C-metal bonding occurs. However dur-
ing the secondary diamond growth, the temperature is high enough (ca.
800jC) that removal of the surface hydrogen occurs due to thermal de-
sorption and hydrogen abstraction occurs by atomic hydrogen produced
in the plasma. Both processes activate the surface carbon atoms and make
them available to form C-metal bonds.
The poor metal particle adhesion typical of hydrogen-terminated
diamond surfaces was previously demonstrated [49,125,126]. Cyclic vol-
tammetric i-E curves for a Pt-coated diamond electrode were recorded
during long-term potential cycling in 0.1 M HClO
4
. One thousand cycles
were performed between 400 and 1500 mV vs. Ag/AgCl with a maximum
anodic and cathodic current density of ca. 1 mA/cm
2
. There was progres-
sive loss of the Pt voltammetric features with cycle number. The hydrogen
desorption charge prior to cycling was 800 AC/cm
2
(1617 ng/cm
2
active Pt)
and decreased signicantly to 25 AC/cm
2
(51 ng/cm
2
active Pt) after
cycling, representing a decrease of over 96%. The loss of electrodeposited
Pt results from the turbulence of gas evolution, which physically dislodges
the weakly adhering deposits. The lack of C-metal bonding due to
hydrogen passivation is also likely a reason for the weak adhesion.
Proof that the metal particles are strongly anchored to the surface by
the secondary diamond growth has been demonstrated, and an example is
given in Fig. 27 [48,49,51,125,126]. The dimensional stability of the Pt/
diamond composite electrodes was examined during a 2-hour exposure to
85% H
3
PO
4
acid at 170jC and 0.1A/cm
2
. Electrochemical and AFM
measurements revealed no loss in Pt activity and no degradation of the
diamond microstructure after 2 hours of electrolysis. Figure 27 shows
cyclic voltammetric i-E curves for a Pt/diamond composite electrode in 0.1
M HClO
4
before and after two 1-hour periods of anodic polarization. The
curve for the electrode prior to the polarization (dashed line) reveals the
presence of Pt with the expected voltammetric features. After the two
polarizations, the voltammetric features are unchanged and reveal that
there is no loss of catalyst activity due to degradation of the diamond
microstructure and morphology. Importantly, there is no loss in the charge
Swain 254
associated with hydrogen ion adsorption and desorption. Such loss would
be expected if the Pt catalyst particles were detached from the surface due
to an oxidizing and corroding diamond support. In fact, the charge
associated with the hydrogen ion adsorption actually increases after the
electrolysis. The cathodic charge is 355 AC/cm
2
before and increases to 420
and 455 AC/cm
2
after the two 1-hour polarizations, respectively. The
increased charge is attributed to minor surface cleaning and crystallo-
graphic changes in the deposits that occur during the vigorous gas evo-
lution. One type of minor cleaning that is possible is the oxidative removal
of residual carbon deposits formed during the secondary diamond depo-
sition. These deposits do not aect the stability of the metal particles but,
rather, inuence their surface activity toward faradaic electron transfer
processes. There was no signicant change in the particle size and coverage
after polarization, at least as revealed by AFM. The most signicant
change in the voltammograms is the reduced overpotential for oxygen
evolution after the polarizations. The current associated with the Pt-oxide
FIG. 27. Cyclic voltammetric i-E curves for a Pt/diamond composite electrode
in 0.1 M HCIO
4
before (dashed line) and after two 1-h polarizations (solid lines)
in 85 wt% H
3
PO
4
at 170jC and an anodic current density of 0.1 A/cm
2
. (Re-
printed with permission from Electrochem. Solid State Lett. 5, E4 2002. Copy-
right (2002) The Electrochemical Society, Inc.) (From Ref. 51.)
reduction increases after the electrolysis, and the current maximum shifts
to slightly more negative potentials. There is also a minor decrease in the
overpotential for hydrogen evolution after the polarization.
Figure 28 shows ex situ AFM images of the Pt/diamond composite
electrode before and after the two 1-hour polarizations. A well-faceted,
polycrystalline morphology is observed before and after electrolysis. The
crystallites are randomly oriented with hemispherical Pt dispersions (10
300 nm) decorating both the facets and grain boundaries. Clearly, there
is no evidence of any morphological or microstructural damage such as
lm delamination, grain roughening, or pitting. The similarity of the
image features before and after polarization is consistent with the
voltammetric data.
Figure 29 shows top-view, 10 10 Am
2
AFM images (force mode,
air) of diamond surfaces after galvanostatic deposition of Pt from a 1 mM
FIG. 28. AFM images (air) of a Pt/diamond composite electrode (A) before
and (B) after anodic polarization in 85 wt% H
3
PO
4
at 170jC and an anodic cur-
rent density of 0.1 A/cm
2
. (Reprinted with permission from Electrochem. Solid
State Lett. 5, E4, 2002. Copyright (2002) The Electrochemical Society, Inc.) (From
Ref. 51.)
Swain 256
K
2
PtCl
6
+0.1 MHClO
4
(constant current density =0.5 mA/cm
2
) for 100,
200, 300, and 400 s periods of time, respectively. The nominal particle size,
variances, and distribution density for each deposition time are summa-
rized in Table 7. The metal particles decorate the entire surface with good
distribution. After 100 s of deposition (Fig. 29A), distinguishable Pt par-
ticles are uniformly dispersed over the diamond surface decorating both
the microcrystallite facets and the grain boundaries. The particle size
ranges from 20 to 200 nm, with an average particle distribution of 0.5
10
9
cm
2
. Slightly higher distribution is observed after 200 s of deposition
FIG. 29. AFM images (air) of boron-doped diamond thin lms after
galvanostatic deposition of Pt from a solution of 1 mM K
2
PtCl
6
+ 0.1 M HClO
4
.
The deposition times are (A) 100, (B) 200, (C) 300, and (D) 400 s, respectively.
Current density = 0.5 mA/cm
2
. Electrode area = 0.2 cm
2
. (Reprinted with per-
mission from J. Electrochem. Soc. 150, E24, 2003. Copyright (2003) The Electro-
chemical Society, Inc.) (From Ref. 126.)
T
A
B
L
E
7
P
t
P
a
r
t
i
c
l
e
S
i
z
e
,
D
i
s
t
r
i
b
u
t
i
o
n
,
H
y
d
r
o
g
e
n
D
e
s
o
r
p
t
i
o
n
C
h
a
r
g
e
,
a
n
d
R
o
u
g
h
n
e
s
s
F
a
c
t
o
r
f
o
r
P
t
/
D
i
a
m
o
n
d
C
o
m
p
o
s
i
t
e
E
l
e
c
t
r
o
d
e
s
B
e
f
o
r
e
a
n
d
A
f
t
e
r
S
e
c
o
n
d
a
r
y
D
i
a
m
o
n
d
G
r
o
w
t
h
B
E
F
O
R
E
A
F
T
E
R
D
e
p
o
s
i
t
i
o
n
t
i
m
e
(
s
)
P
a
r
t
i
c
l
e
s
i
z
e
(
n
m
)
D
i
s
t
r
i
b
u
t
i
o
n
d
e
n
s
i
t
y
(
1
0
9
c
m
2
)
H
y
d
r
o
g
e
n
d
e
s
o
r
p
t
i
o
n
c
h
a
r
g
e
(
m
C
/
c
m
2
)
R
o
u
g
h
n
e
s
s
f
a
c
t
o
r
P
a
r
t
i
c
l
e
s
i
z
e
(
n
m
)
D
i
s
t
r
i
b
u
t
i
o
n
d
e
n
s
i
t
y
(
1
0
8
c
m
2
)
H
y
d
r
o
g
e
n
d
e
s
o
r
p
t
i
o
n
c
h
a
r
g
e
(
m
C
/
c
m
2
)
R
o
u
g
h
n
e
s
s
f
a
c
t
o
r
S
u
r
f
a
c
e
L
o
s
s
(
%
)
1
0
0
1
0
2
0
0
0
.
5
F
0
.
0
7
0
.
5
1
2
.
4
3
0
2
5
0
0
.
4
F
0
.
0
5
0
.
3
2
1
.
5
3
7
.
3
2
0
0
2
0
2
5
0
0
.
9
F
0
.
0
4
0
.
8
0
3
.
8
5
0
3
0
0
1
F
0
.
0
6
0
.
5
6
2
.
7
3
0
.
0
3
0
0
5
0
3
0
0
2
F
0
.
0
5
1
.
1
9
5
.
7
5
0
3
5
0
2
F
0
.
0
3
0
.
8
1
3
.
9
3
1
.
9
4
0
0
6
0
4
0
0
4
F
0
.
0
4
1
.
4
3
6
.
8
6
0
5
0
0
4
F
0
.
0
2
0
.
9
2
4
.
4
3
5
.
7
S
o
u
r
c
e
:
R
e
f
.
1
2
6
.
Swain 258
(Fig. 29B). The particle size ranges from 10 to 250 nm, and the distribution
increases to 0.8 10
9
cm
2
. A further increase in the distribution is seen
after 300 s of deposition (Fig. 29C). The particle size ranges from 50 to 300
nm, and the distribution is 1.0 10
9
cm
2
. Finally, somewhat larger
particles and a higher distribution are observed after 400 s of deposition
(Fig. 29D). The particle size ranges from60 to 500 nm, and the distribution
is 1.2 10
9
cm
2
. The image features are consistent with a progressive
nucleation and three-dimensional growth mechanism in which new Pt
nuclei form progressively as a function of time while the existing nuclei
increase in size [191]. Nuclei form on both the grain boundaries and the
facet surfaces. This indicates that both regions are electrochemically active
and support the ow of current.
Figure 30 shows 5 5 Am
2
AFM images (force mode, air) of Pt-
coated lms after 3 additional hours of diamond growth. The images are of
the same Pt-coated lms shown in Fig. 29. Numerous particles are
randomly distributed over the surface on both the microcrystallite facets
and in the grain boundaries, particularly for 200, 300, and 400 s electro-
deposition times. The nominal particle size is slightly increased by the
secondary diamond growth, ranging from 30 to 500 nm, but the distribu-
tion is lower. The process of entrapping the particles with diamond reduces
the total active Pt exposed. This is seen by comparing the hydrogen
desorption charge in Table 7. There is a 3037% loss in Pt activity, as
measured in the hydrogen desorption voltammetric charge between 0 and
300 mV. This is consistent with the secondary diamond growth around the
base (lower third) of the metal deposits.
While the composite electrode possesses superb dimensional stability
and catalytic activity, the nominal particle size is too large, at present, for a
real world catalyst. We are presently pursuing pulsed deposition
approaches for reducing the particle size and increasing the dispersion.
For example, recent results have indicated that a series of 1-second pulses
at 0.5 mA/cm
2
results in the formation of Pt particles with a nominal
diameter of ca. 50 F 12 nm and an estimated loading of <0.1 mg/cm
2
. It
should be possible with further work to reduce the nominal particle size to
the 110 nm range.
B. Oxygen Reduction Reaction
Platinum is the best known electrocatalyst for the oxygen-reduction re-
action (ORR), and it is widely used in electrochemical-energy conversion
devices. Pt particles are generally dispersed onto a high surface area con-
ductive support, such as carbon powder or a porous membrane, in order
to obtain high electrocatalytic activity and optimized Pt utilization. The
electrode kinetics for the ORR have been extensively studied at Pt-
supported catalytic electrodes in both acidic and alkaline electrolytes
[192206]. The electrocatalytic activity of the Pt/diamond composite elec-
trodes for oxygen reduction was investigated in 0.1 M H
3
PO
4
, H
2
SO
4
, and
HClO
4
at room temperature. Tafel slopes and exchange current densities
were determined as a function of the electrolyte composition and Pt
FIG. 30. AFMimages (air) of Pt-coated diamond thin lms (as indicated in Fig.
29) after a 3-h secondary diamond growth. The corresponding Pt deposition times
are (A) 100, (B) 200, (C) 300, and (D) 400 s, respectively. (Reprinted with per-
mission from J. Electrochem. Soc. 150, E24, 2003. Copyright (2003) The Electro-
chemical Society, Inc.) (From Ref. 126.)
Swain 260
loading. Figure 31 shows a typical cyclic voltammetric i-Ecurve for oxygen
reduction on a Pt/diamond composite electrode. The experiments were
carried out in O
2
-saturated 0.1 M HClO
4
at room temperature. The
potential sweep is initiated at 1.2 V, whereupon the Pt surface is covered
with a thin oxide lm, scanned to 0.25 V, and then scanned back to 1.2 V.
The response for the Pt/diamond composite electrode resembles that for
the clean Pt foil (data not shown). During the cathodic scan direction,
discernible reduction current commences at ca. 0.72 V, close to the open
circuit potential in this electrolyte. A dramatic increase in the cathodic
current is observed at potentials below 0.55 V, accompanied by the
reductive removal of the Pt oxide, in agreement with the fast ORR kinetics
on oxide-free Pt surface. The current reaches a maximum of 0.33 mA at ca.
0.45 V and decays rapidly on the low potential side of the peak due to the
depletion of oxygen at the electrode surface (diusion controlled). Oxygen
reduction continuously contributes to the total current even during the
reverse scan, as there is net cathodic current, until the Pt oxides begin to
form on the electrode surface.
FIG. 31. Linear sweep voltammetric i-E curves for a Pt/diamond composite
electrode in 0.1 M HClO
4
at dierent scan rates. Inset shows a plot of peak
potential vs. logarithm of scan rate. (From Ref. 125.)
The rate of oxygen reduction on the oxide-free Pt surface is limited by
the mass transport of dissolved oxygen to the electrode surface. The
dependence of the oxygen reduction on the scan rate, r, was examined.
Figure 31 shows the linear potential sweep voltammetric i-E curves for a
Pt/diamond composite electrode in O
2
-saturated 0.1 M HClO
4
[125]. The
peak current, i
p
, increases linearly with r
1/2
when the scan rate is varied
from 50 to 400 r/s, indicative of a semi-innite linear diusion-controlled
process. However, i
p
rather approaches a proportionality with r at scan
rates higher than 400 r/s. This is expected as the reaction shifts from being
mass transport limited to control by the surface adsorption process.
Figure 32 shows slowscan linear sweep voltametric i-Ecurves (1 mV/
s) for a Pt/diamond composite electrode in oxygen-saturated 0.1 MHClO
4
,
FIG. 32. Slowscan linear sweep voltammetric i-Ecurves for a Pt/diamond com-
posite electrode in oxygen-saturated 0.1 M (A) HClO
4
, (B) H
2
SO
4
, and (C) H
3
PO
4
.
The Pt deposition time is 300 s. Apparent loading = 75.8 Ag/cm
2
. Scan rate = 1
mV/s. The insert shows corresponding Tafel slopes constructed from the rising
portionof the i-Ecurves. (ReprintedwithpermissionfromJ. Electrochem. Soc. 150,
E24, 2003. Copyright (2003) The Electrochemical Society, Inc.) (From Ref. 126.)
Swain 262
H
2
SO
4
, and H
3
PO
4
, respectively [126]. The polarization curves have been
corrected for the background current, which was measured in nitrogen-
purged solution. Peak-shaped curves are seen with E
p
red
varying with acid
type. Values of 460, 390, and 335 mV are seen for HClO
4
, H
2
SO
4
, and
H
3
PO
4
, respectively. The peak current, i
p
red
, is 31.5, 29.5, and 28.0 AA
(f0.15 mA/cm
2
or f2.0 mA/mg Pt), respectively. The decrease in the
oxygen reduction current in the kinetically controlled region and the
negative shift of E
p
red
can be explained by an anion adsorption eect.
H
2
PO
4
has been reported to strongly adsorb on the Pt surface in the

potential region of the oxygen-reduction current onset [207,208]. H
2
PO
4
adsorption can slow down the electrode reaction kinetics by blocking sites
for the initial oxygen chemsorption [208]. Changes in the double layer
structure as a result of the adsorption could also be inuential [209]. The
i
p
red
in all the acids increases linearly with the square root of scan rate (50
200 mV/s), indicating the reaction rate is limited by semi-innite linear
diusion of oxygen to the electrode surface.
The rising portionof the i-Ecurves in the kinetically controlledregion
was analyzed, and Tafel plots were constructed, based on the equation
logi logi
o
a
c
nF=2:303 RTg
Plots of g vs. log i for the composite electrode are shown as an inset in Fig.
32. No correction for any mass-transfer eects was made in the analysis of
these data. The Tafel plots for all three electrolytes are linear at low
overpotentials, most likely being inuenced at higher overpotentials by
mass transfer eects. The slopes are 63, 69, and 80 mV/dec, respec-
tively, for HClO
4
, H
2
SO
4
, and H
3
PO
4
. These results are similar to those
reported by Kita et al. [207] and Damjanovic et al. [210,211]. Mass-transfer
corrected Tafel plots for oxygen reduction at Pt in acid normally yield two
slopes, ca. 60 mV/dec at low overpotentials and 120 mV/dec at high
overpotentials [211]. The low Tafel slope region corresponds to a potential
regime where the oxygen reduction occurs on an oxide-covered surface.
Adsorbed O and OH species aect the adsorption of O
2
and other
reaction intermediates (Temkin conditions). The high Tafel slope region
(120 mV/dec) corresponds to a potential region where the reaction
proceeds on an oxide-free surface (Langmuir conditions). Due to the
inuence of mass-transfer eects, this higher-slope region is not observed
in the present data. The slightly higher Tafel slopes observed in H
2
SO
4
and
H
3
PO
4
are likely attributable to anion adsorption inuencing the O
2
chemsorption on the surface [207]. The exchange currents can be deter-
mined by extrapolating the linear Tafel region back to g = 0. Normalizing
these currents to the electrochemically active surface area of the Pt (0.78
cm
2
, as estimated from the roughness factor), yields exchange current
densities, i
0
, of 1.9 10
10
Acm
2
in HClO
4
, 4.6 10
11
Acm
2
in
H
2
SO
4
, and7.1 10
12
Acm
2
inH
3
PO
4
. Thus, the reactionrate increases
in order of H
3
PO
4
< H
2
SO
4
< HClO
4
. This is a well-known phenomenon
for smooth Pt surfaces and can be explained by the dierent degrees of
adsorption of phosphate/biphosphate, sulfate/bisulfate, and perchlorate
anions [207]. These preliminary results indicate that the kinetic parameters
for the oxygen-reduction reaction at Pt/diamond composite electrodes are
similar to the parameters for clean polycrystalline Pt [211]. In addition, the
diamond matrix exerts little inuence on the oxygen-reduction response.
C. Methanol Oxidation Reaction
Preliminary studies of the methanol oxidation reaction (MOR) have also
been performed. The MOR kinetics were initially studied by cyclic
voltammetry. Figure 33 (solid line) presents a typical cyclic voltammetric
i-E curve for methanol oxidation at a Pt/diamond composite electrode.
The experiments were carried out at room temperature in nitrogen-purged
0.1 M HClO
4
containing 0.2 M CH
3
OH. The background voltammetric
response is also shown (dashed line) for comparison. The potential sweep
was initiated at 0 mV and scanned in the positive direction at a scan rate of
50 mV/s. The background i-E curves for both electrodes are shown as
dashed lines. The response for the Pt/diamond composite electrode
resembles that for the clean Pt foil. The voltammogram displays peaks
that are also characteristic of MOR on supported Pt catalysts in acidic
solution [212]. During the forward or anodic sweep, discernible faradaic
current begins at ca. 0.25 V, corresponding to the dehydrogenation
reactions of chemisorbed methanol. The current is low due to surface
poisoning by the rapid formation of chemisorbed organic residues, mainly
CO. As the applied potential becomes more positive, dissociative chemi-
sorption of water takes place. Accordingly, there is a sharp rise in the
current at E > 0.4 V, which is attributed to the oxidative removal of the
adsorbed organic residues. An oxidation peak is observed at 0.65 V with a
peak current, i
p1
ox
, of 420 AA. The oxidation charge is greater than that
predicted for the oxidation of a monolayer of adsorbed CO, so in addition
to the oxidation of adsorbed CO, oxidation of additional CH
3
OH mole-
Swain 264
cules is occurring. The decline in current above 0.65 V reects the
inhibition of the MOR by surface oxide formation at these potentials.
However, some catalytically active surface oxides are formed again at
potentials greater than 0.85 V, accounting for the second anodic peak near
1.2 V (Reaction 6.10). Oxygen evolution takes place at potentials greater
than 1.3 V. The surface is inactive toward CH
3
OHoxidation on the reverse
sweep until the reduction of the surface oxides at 0.55 V (i.e., formation of
bare Pt). An oxidation peak is then observed at ca. 0.42 V with a peak
current, i
p2
ox
, of 225 AA. At potentials below 0.2 V, the active surface sites
become blocked by CO formation from the dehydrogenation of chem-
isorbed CH
3
OH. Consequently, hydrogen adsorption/desorption is sig-
nicantly suppressed.
The OCPs for the Pt/diamond composite electrodes with varying
platinum loadings were measured in nitrogen-saturated methanol + 0.1
FIG. 33. Cyclic voltammetric i-E curves for Pt/diamond composite electrode
and polycrystalline Pt foil electrode in 0.1 M HClO
4
+ 0.2 M CH
3
OH (solid line)
and 0.1 M HClO
4
(dashed line). Scan rate = 50 mV/s. Active surface area of the
composite electrode is 0.44 cm
2
(Pt deposition was at 0.5 mA/cm
2
for 200 s).
(From Ref. 125.)
M HClO
4
. Prior to the measurement, the electrodes were subjected to a
potential cycling activation procedure between the hydrogen and oxygen
evolution regimes. The potential cycling was then stopped at 0.2 V vs.
Ag/AgCl, a potential at which the electrode surface is oxygen-free.
Activated electrodes were then exposed to nitrogen-saturated CH
3
OH/
HClO
4
solutions and allowed to stabilize for 5 min. Steady-state OCP
values are presented in Table 8 [125]. The OCP is, on average, about 0.20 V
vs. Ag/AgCl (0.393 Vvs. NHE), indicating a ca. 0.55 Vpotential dierence,
with respect to the oxygen electrode, in the same electrolyte. The OCP
decreases with increasing methanol concentration, a trend predicted by the
Nernst equation. The results in Table 8 also demonstrate that the OCP
slightly decreases with increasing the Pt loading, most notably at concen-
trations less than 1 M. However, this decrease is less evident at high
concentrations.
One of the gures of merit for catalytic activity is the turnover
number (tn) dened as the number of the molecules converted per surface
site per second. For oxidative heterogeneous catalysis, the turnover
number should be in the range from 10 to 100 to make a given material
an eective catalyst. However, to date, the turnover number for methanol
oxidation at the noble metal catalyst in the operational fuel cell is ca. 0.06,
still far below the needed performance level. The following formula can be
used to calculate the tn from the current density [213]:
tn
molecules
s site

imA cm
2
6:02 10
20
nF 1:3 10
15
TABLE 8
Open Circuit Potentials for Pt/Diamond Composite Electrodes with Varying Pt
Loading in 0.1 M HClO
4
as a Function of Methanol Concentration
Methanol concentration (M)
Electrode 0.2 0.6 1.0 1.5 2.0
Diamond 1 (25.2 Ag/cm
2
) 0.240 0.221 0.210 0.182 0.169
2
) 0.232 0.212 0.200 0.179 0.168
2
) 0.230 0.200 0.193 0.178 0.170
Source: Ref. 125.
Swain 266
There are 1.3 10
15
platinum sites per 1 cm
2
of the real platinum surface
area, and n = 6 for the methanol oxidation reaction yielding CO
2
as the
product. Collection of the constants yields the formula
tn 0:8 imA cm
2
Calculated turnover numbers are listed in Table 9. The turnover number

increases with methanol concentration as expected.
Data are normalized to the real surface area, 0.51 cm
2
, estimated
from the hydrogen desorption charge. The charge density was measured
during a 20-minute potential step to 0.60 V vs. Ag/AgCl. The current
density was measured at the end of the 20-minute period, and the turnover
number was calculated from this current.
XII. CONCLUSIONS
The future for diamond electrodes in a number of electrochemical tech-
nologies is bright indeed. Diamond electrodes will continue to impact
various electrochemical technologies as more is learned about their
properties and commercial suppliers emerge to provide these electrodes
to the general community. Research with the material is still in its infancy,
so much remains to be understood about the structure-function relation-
ship. Diamond oers signicant advantages over other electrodes, in
particular sp
2
carbon electrodes, in terms of linear dynamic range, limit
of detection, response time, response precision, and response stability. The
electrode can be used to provide a sensitive, reproducible, and stable
TABLE 9
Quantitative Data on Reactivity of Methanol Oxidation at Pt/Diamond
Composite Electrode in 0.1 M HClO
4
Solution as a Function of Methanol
Concentration
Methanol concentration (M)
Electrode 2 0.2 0.6 1.0 1.5 2.0
Charge density (C/cm
2
) 0.20 0.30 0.45 0.50 0.55
Steady-state current density (mA/cm
2
) 0.08 0.15 0.22 0.26 0.32
Turnover number (molecules/site-s) 0.06 0.12 0.18 0.21 0.26
Source: Ref. 125.
response in electroanalytical measurements, for spectroelectrochemical
measurements in both the UV/Vis and IR as an optically transparent
electrode, and as a morphologically and microstructurally stable electrode
for high current density electrolysis. Some important areas of future work
include growth and characterization of diamond lms deposited on ber-
ous substrates of 10 Amdiameter or less, chemical modication of diamond
surfaces to control adsorption and electron-transfer kinetics, and pattern-
ing of electrically conductive diamond electrodes into microelectrode array
geometries.
ACKNOWLEDGMENTS
This chapter is dedicated to Prof. Ted Kuwana for his many years of
mentoring and friendship. The results reported herein were obtained
through the hard work and dedication of a number of graduate students,
postdoctoral researchers and visiting faculty (Michael Granger, Jishou Xu,
Qingyun Chen, Jian Wang, Jason Bennett, Jason Stotter, Maggie Witek,
Matt Hupert, Shannon Haymond, Prerna Sonthalia, Grace Muna, Gloria
Pimienta, Jinwoo Park, Josef Cvacka, Zuzana Cvackova, Yoshiyuki
Show, Jerzy Strojek, and Jerzy Zak). The diamond electrode research has
been generously supported in recent years by the NSF, DOE, USDA, NIH,
NASA, and the ACS-PRF, support that is gratefully acknowledged.
REFERENCES
1. Butler, J.E.; Woodin, R.L. Phil. Trans. R. Soc. Lond. A 1993, 342, 209.
2. Gruen, D.M. MRS Bull. 1998, 23, 32.
3. Gruen, D.M. Annu. Rev. Mater. Sci. 1999, 29, 211.
4. Gruen, D.M.; Liu, S.; Krauss, A.R.; Luo, J.; Pan, X. Appl. Phys. Lett.
1964, 64, 1502.
5. Zuiker, C.D.; Krauss, A.R.; Gruen, D.M.; Carlisle, J.A.; Terminello, L.J.;
Asher, S.A.; Bormett, R.W. In Applications of Synchrotron Radiation
Techniques to Material Sciences III, Proceedings of the Materials Research
Society, April 812; Terminello, L.J. Mini, S.M., Ade, H., Perry, D.L.;
Eds.; San Francisco, CA, 1996; Vol. 437, 211.
6. McCauley, T.G.; Gruen, D.M.; Krauss, A.R. Appl. Phys. Lett. 1998, 73,
1646.
7. Redfern, P.C.; Horner, D.A.; Curtiss, L.A.; Gruen, D.M. J. Phys. Chem.
1996, 100, 11654.
Swain 268
8. Gruen, D.M. R&D Mag 1997, 4, 57.
9. Zhou, D.; Gruen, D.M.; Qin, L.C.; McCauley, T.G.; Krauss, A.R. J. Appl.
Phys. 1998, 84, 1981.
10. Qin, L.C.; Zhou, D.; Krauss, A.R.; Gruen, D.M. Appl. Phys. Lett 1998, 72,
3437.
11. Zhou, D.; McCauley, T.G.; Qin, Ley C.; Krauss, A.R.; Gruen, D.M. J.
Appl. Phys. 1998, 83, 540.
12. Erdemir, A.; Halter, M.; Fenske, G.R.; Krauss, A.R.; Gruen, D.M.;
Pimenov, S.M.; Konov, V.I. Surf. Coat. Technol. 1997, 94, 537.
13. Sakaguchi, I.; Nishitani-Gamo, M.; Loh, K.P.; Haneda, H.; Hishita, S.;
Ando, T. Appl. Phys. Lett. 1997, 71, 629.
14. Wolden, C.; Mitra, S.; Gleason, K.K. J. Appl. Phys. 1992, 72, 3750.
15. Tachibana, T.; Yokota, Y.; Miyata, K.; Onishi, T.; Kobashi, K.; Tarutani,
M.; Takai, Y.; Shimizu, R.; Shintani, Y. Phys. Rev. B 1997, 56, 15967.
16. Michler, J.; Mermoux, M.; von Kaenel, Y.; Haouni, A.; Lucazeau, G.;
Blank, E. Thin Solid Films 1999, 357, 189.
17. Wightman, R.M. Science 1998, 240, 415.
18. Wisitsora-at, A.; Kang, W.P.; Davidson, J.L.; Kerns, D.V.; Kerns, S.E. J.
Vac. Sci. Tech. B: Microelectron Nanometer Struct. 2001, 19, 971.
19. Tsunozaki, K.; Einaga, Y.; Rao, T.N.; Fujishima, A. Chem. Lett. 2002,
502.
20. Eaton, S.C.; Anderson, A.B.; Angus, J.C.; Evstefeeva, Y.E.; Pleskov, Y.V.
Electrochem. Solid State Lett. 2002, 5, G65.
21. Saada, D.; Adler, J.; Kalish, R. Appl. Phys.Lett. 2000, 77, 878.
22. Granger, M.C.; Witek, M.; Xu, J.; Wang, J.; Hupert, M.; Hanks, A.;
Swain, G.M. Anal. Chem. 2000, 72, 3793.
23. Okano, K.; Naruki, H.; Akiba, Y.; Kurosu, T.; Iida, M.; Hirose, Y.;
Nakamura, T. Jpn. J. Appl. Phys. 1989, 28, 1066.
24. Mort, J.; Okumura, K.; Machonkin, M. Philosoph Mag. B 1991, 63, 1031.
25. Ushizawa, K.; Watanabe, K.; Ando, T.; Sakaguchi, I.; Nishitani-Gamo,
M.; Sato, Y.; Kanada, H. Diam. Rel. Mater. 1998, 7, 1719.
26. Iwaki, M.; Sato, S.; Takahashi, K.; Sakairi, H. Nucl. Instrum. Meth. 1983,
209, 1129.
27. Patel, K.; Hashimoto, K.; Fujishima, A. Denki Kagaku 1992, 60, 659.
28. Swain, G.M.; Ramesham, R. Anal. Chem. 1993, 65, 345.
29. Xu, J.; Granger, M.C.; Chen, Q.; Lister, T.E.; Strojek, J.W.; Swain, G.M.
Anal. Chem. 1997, 69, 591A.
30. Granger, M.C.; Xu, J.; Strojek, J.W.; Swain, G.M. Anal. Chim. Acta,
1999, 397, 145.
31. Swain, G.M.; Anderson, A.B.; Angus, J.C.J.C. MRS Bull. 1998, 23, 56.
32. Swain, G.M. Adv. Mater. 1995, 6, 388. Awada, M.; Strojek, J.W.; Swain,
G.M. J. Electrochem. Soc. 1994, 142, L42.
33. Alehashem, S.; Chambers, F.; Strojek, J.W.; Swain, G.M.; Ramesham, R.
Anal. Chem. 1995, 67, 2812.
34. Strojek, J.W.; Granger, M.C.; Swain, G.M.; Dallas, T.; Holtz, M.W. Anal.
Chem. 1996, 68, 2031.
35. DeClements, R.; Hirsche, B.L.; Granger, M.C.; Xu, J.; Swain, G.M. J.
Electrochem. Soc. 1996, 143, L150.
36. Jolley, S.; Koppang, M.; Jackson, T.; Swain, G.M. Anal. Chem. 1997, 69,
4041.
37. Xu, J.; Swain, G.M. Anal. Chem. 1998, 70, 1502.
38. Xu, J.; Chen, Q.; Swain, G.M. Anal. Chem. 1998, 70, 3146.
39. Koppang, M.D.; Witek, M.; Blau, J.G.; Swain, G.M. Anal. Chem. 1999,
71, 1188.
40. Kuo, T.C.; McCreery, R.L.; Swain, G.M. Electrochem. Solid State Lett.
1999, 2, 288.
41. Granger, M.C.; Swain, G.M. J. Electrochem. Soc. 1999, 146, 4551.
42. Fausett, B.M.; Granger, M.C.; Hupert, M.L.; Wang, J.; Swain, G.M.
Electroanalysis 2000, 12, 7.
43. Witek, M.A.; Swain, G.M. Anal. Chim. Acta 2001, 440, 119.
44. Chen, Q.; Gruen, D.M.; Krauss, A.R.; Corrigan, T.D.; Witek, M.; Swain,
G.M. J. Electrochem. Soc. 2001, 148, E44.
45. Swain, G.M. J. Electrochem. Soc. 1994, 141, 3382.
46. DeClements, R.; Swain, G.M. J. Electrochem. Soc. 1997, 144, 856.
47. Chen, Q.; Granger, M.C.; Lister, T.E.; Swain, G.M. J. Electrochem. Soc.
1997, 144, 3086.
48. Wang, J.; Swain, G.M.; Tachibana, T.; Kobashi, K. J. N. Electrode Mat.
Electrochem. Sys. 2000, 3, 75.
49. Wang, J.; Swain, G.M.; Tachibana, T.; Kobashi, K. Electrochem. Solid-
State Lett. 2000, 3, 286.
50. Witek, M.; Wang, J.; Stotter, J.; Hupert, M.; Haymond, S.; Sonthalia, P.;
Zak, J.K.; Chen, Q.; Gruen, D.M.; Butler, J.E.; Kobashi, K.; Tachibana,
T.; Swain, G.M. J. Wide Bandgap Semiconductors 2002, 8, 1.
51. Wang, J.; Swain, G.M. Electrochem. Solid State Lett. 2002, 5, E4.
52. Zak, J.K.; Butler, J.E.; Swain, G.M. Anal. Chem. 2001, 73, 908.
53. Martin, H.B.; Argoitia, A.; Landau, U.; Anderson, A.B.; Angus, J.C. J.
Electrochem. Soc. 1996, 143, L133.
54. Cooper, J.B.; Pang, S.; Albin, S.; Zheng, J.; Johnson, R.M. Anal. Chem.
1998, 70, 464.
55. Sarada, B.V.; Rao, T.N.; Tryk, D.A.; Fujishima, A. J. Electrochem. Soc.
1999, 146, 1469.
56. Cvacka, J.; Park, J-W.; Swain, G.M. in preparation.
57. Manivannan, A.D.; Tryk, A.; Fujishima, A. Electrochem. Solid State Lett.
1999, 2, 455.
Swain 270
58. Popa, E.; Notsu, H.; Miwa, T.D.; Tryk, A.; Fujishima, A. Electrochem.
Solid State Lett. 1999, 2, 49.
59. Sopchak, D.; Miller, B.; Kalish, R.; Avyigal, Y.; Shi, X. Electroanalysis
2002, 14, 473.
60. Rao, T.N.; Yagi, I.T.; Miwa, D.; Tryk, A.; Fujishima, A. Anal. Chem. 1999,
71, 2506.
61. Popa, E.; Kubota, Y.; Tryk, D.A.; Fujishima, A. Anal. Chem. 2000, 72, 1724.
62. Sarada, S.; Rao, T.N.; Tryk, D.A.; Fujishima, A. Anal. Chem. 2000, 72,
1632.
63. Rao, T.N.; Loo, B.H.; Sarada, B.V.; Terashima, C.A.; Fujishima, A. Anal.
Chem. 2002, 74, 1578.
64. Natishan, P.M.; Morrish, A. Mater. Lett. 1989, 8, 269.
65. Sakharova, A.; Sevastyanov, A.E.; Pleskov, Y.; Templitskaya, G.L.; Suri-
kov, V.V.; Voloshin, A.A. Electrokhimiya 1991, 27, 239.
66. Sakharova, A.; Nyikos, L.; Pleskov, Y. Electrochimica Acta 1992, 37, 973.
67. Zhang, X.; Wang, R.; Yao, Y.; Chen, C.; Zhu, J.; Liu, X.; Wu, J.; Zhang, G.
In 2nd International Conference on the Applications of Diamond Films and
Related Materials; Yoshikawa, M. Murakawa, M., Tzeng, Y., Yarbrough,
W.A., Eds.; MYU, Tokyo, 1993; 65.
68. Tenne, R.; Patel, K.; Hashimoto, K.; Fujishima, A. J. Electroanal. Chem.
1993, 347, 409.
69. Miller, B.; Kalish, R.; Feldman, L.C.; Katz, A.; Moriya, N.; Short, K.;
White, A.E. J. Electrochem. Soc. 1994, 141, L41.
70. Peilio, Z.; Jianzhong, Z.; Shonzhong, Z.; Xikang, Z.; Guoxiong,; Fre-
senius, Z. J. Anal. Chem. 1995, 353, 171.
71. Reuben, C.; Galun, E.; Cohen, H.; Tenne, R.; Kalish, R.; Muraki, Y.;
Hashimoto, K.; Fujishima, A.; Butler, J.M.; Levy-Clement, C.J. J. Elec-
troanal. Chem. 1995, 396, 233.
72. Sakharova, A. Ya.; Pleskov, Yu.V.; Di Quarto, F.; Piazza, S.; Sunseri, C.;
Teremetskaya, I.G.; Varin, V.P. J. Electrochem. Soc. 1995, 142, 2704.
73. Pleskov, Yu.V.; Mishuk, V. Ya.; Abaturov, M.A.; Elkin, V.V.; Krotova,
M.D.; Varin, V.P.; Teremetskaya, I.G. J. Electroanal. Chem. 1995, 396,
227.
74. Vinokur, N.; Miller, B.; Avyigal, Y.; Kalish, R. J. Electrochem. Soc. 1996,
143, L238.
75. Pleskov, Yu.V.; Elkin, V.V.; Abaturov, M.A.; Krotova, M.D.; Mishuk, V.
Ya.; Varun, V.P.; Teremetskaya, I.G. J. Electroanal. Chem. 1996, 413,
105.
76. DeClements, R.; Swain, G.M.; Dallas, T.; Holtz, M.W.; Herrick, R., III;
Stickney, J.L. Langmuir 1996, 12, 6578.
77. Boonma, L.; Yano, T.; Tryk, D.A.; Hashimoto, K.; Fujishima, A. J. Elec-
trochem. Soc. 1997, 144, L142.
78. Modestov, A.D.; Pleskov, Yu.V.; Varnin, V.P.; Teremetskaya, I.G. Elec-
trokhimiya 1997, 33, 60.
79. Li, L-F.; Totir, D.; Miller, B.; Chottiner, G.; Argoitia, A.; Angus, J.C.;
Scherson, D.A. J. Am. Chem. Soc. 1997, 119, 7875.
80. Ramesham, R.; Rose, M.F. Diam. Rel. Mater. 1997, 6, 17.
81. Ramesham, R.; Rose, M.F. Thin Solid Films 1997, 300, 144.
82. Ramesham, R.; Rose, M.F. J. Mater. Sci. Lett. 1997, 16, 799.
83. Pleskov, Yu.V.; Evstefeeva, Yu.E.; Krotova, M.D.; Elkin, V.V.; Mazin,
V.M.; Mishuk, V. Ya.; Varnin, V.P.; Teremetskaya, I.G. J. Electroanal.
Chem. 1998, 455, 139.
84. Evstefeeva, Yu. E.; Krotova, M.D.; Pleskov, Yu.V.; Elkin, V.V.; Varnin,
V.P.; Teremetskaya, I.G. Elektrokhimiya 1998, 34, 1171.
85. Evstefeeva, Yu. E.; Krotova, M.D.; Pleskov, Yu. V.; Mazin, V.V.; Elkin,
V.V.; Mishuk, V. Ya.; Varnin, V.P.; Teremetskaya, I.G. Elektrokhimiya
1998, 34, 1493.
86. Yano, T.; Tryk, D.A.; Hashimoto, K.; Fujishima, A. J. Electrochem. Soc.
1998, 145, 1870.
87. Compton, R.G.; Marken, F.; Goeting, C.H.; McKeown, R.A.J.; Foord,
J.S.; Searsbrook, G.; Sussmann, R.S.; Whitehead, A.J. Chem. Commun.
1998, 1961.
88. Goeting, C.H.; Jones, F.; Foord, J.S.; Eklund, J.C.; Marken, F.; Compton,
R.G.; Chalker, P.R.; Johnston, C. J. Electroanal. Chem. 1998, 442, 207.
89. Bouamrane, R.; Tadjeddine, A.; Tenne, R.; Butler, J.E.; Kalish, R.; Levy-
Clement, C. J. Phys. Chem. B 1998, 102, 134.
90. Ohtani, B.; Kim, Y.; Yano, T.; Hashimoto, K.; Fujishima, A.; Uosaki, K.
Chem. Lett. 1998, 9, 953.
91. Tenne, R.; Levy-Clement, C. Israel J. Chem. 1998, 38, 57.
92. Chen, Q.; Swain, G.M. Langmuir 1998, 14, 7017.
93. Evstefeeva, Yu. E.; Krotova, M.D.; Pleskov, Yu.V.; Laptev, V.A. Elek-
trokhimya 1999, 35, 138.
94. Pleskov, Yu. V.; Evstefeeva, Yu. E.; Krotova, M.D.; Elkin, V.V.; Baranov,
A.M.; Dementev, A.P. Diam. Rel. Mater. 1999, 8, 64.
95. Yoshihara, S.; Shinozaki, K.; Shirakashi, T.; Hashimoto, K.; Tryk, D.A.;
Fujishima, A. Electrochim. Acta 1999, 44, 2711.
96. Yano, T.; Popa, E.; Tryk, D.A.; Hashimoto, K.; Fujishima, A. J. Elec-
trochem. Soc. 1999, 146, 1081.
97. Vinokur, N.; Miller, B.; Avyigal, Y.; Kalish, R. J. Electrochem. Soc. 1999,
146, 125.
98. Tsunozaki, K.; Einaga, Y.; Rao, T.N.; Fujishima, A. Chem. Lett. 2002,
502.
99. Ohnishi, K.; Einaga, Y.; Notsu, H.; Terashima, C.; Rao, T.N.; Park, S-G.;
Fujishima, A. Electrochem. Solid State Lett. 2002, 5, D1.
Swain 272
100. Carey, J.J.; Christ, C.S., Jr.; Lowery, S.N. US Patent 5399247, 1995.
101. Katsuki, N., et al. J. Electrochem. Soc. 1998, 145, 2358.
102. Foti, G.; Gandini, D.; Comninellis, C.; Perret, A.; Haenni, W. Electro-
chem. Solid State Lett. 1999, 2, 228.
103. Rodrigo, M.A.; Michaud, P.A.; Duo, I.; Panizza; Cerisola, G.; Comni-
nellis, C. J. Electrochem. Soc. 2001, 148, D60.
104. Hagans, P.L.; Natishan, P.M.; Stoner, B.R.; OGrady, W.E. J. Electro-
chem. Soc. 2001, 148, E298.
105. Okino, F.; Shibata, H.; Kawasaki, S.; Touhara, H.; Momota, K.; Nishitani-
Gamo, M.; Sakaguchi, I.; Ando, T. Electrochem. Solid State Lett. 1999, 2,
382.
106. Haenni, W.; et al. Fall Meeting of The Electrochemical Society, Inc., San
Francisco, CA, Sept. 27, 2001. Paper #1323.
107. Wang, J.; Swain, G.M.; Mermoux, M.; Lucazeau, G.; Zak, J.; Strojek, J.W.
New Diamond Frontier Carbon Tech. 1999, 9, 317.
108. Haymond, S.; Babcock, G.T.; Swain, G.M. Electroanal. 2003, 15, 249.
109. Cvacka, J.; Swain, G.M. unpublished.
110. Witek, M.A.; Swain, G.M. Anal. Chem. 2003. submitted.
111. Muck, A.; Swain, G.M. unpublished.
112. Prado, C.; Flechsig, G.; Gruendler, P.; Foord, J.S.; Marken, F.; Compton,
R.G. Analyst 2002, 127, 329.
113. Mutaaga, G.; Swain, G.M. submitted.
114. Terashima, C.; Rao, T.N.; Sarada, B.V.; Tryk, D.A.; Fujishima, A. Anal.
Chem. 2002, 74, 895.
115. Show, Y.; Witek, M.A.; Sonthalia, P.; Swain, G.M. Chem. Mater. 2003,
15, 879.
116. Martin, H.B.; Wightman, R.M.; Angus, J.C. Pittcon, presentation #1046,
2002.
117. Martin, H.B.; Morrison, P.W., Jr. Electrochem. SolidState Lett. 2001, 4, E17.
118. Stotter, J.; Zak, J.; Behler, Z.; Show, Y.; Swain, G.M. Anal. Chem. 2002,
74, 5924.
119. Haymond, S.; Babcock, G.T.; Swain, G.M. J. Am. Chem. Soc. 124, 10634.
120. Yagi, I.; Notsu, H.; Kondo, T.D.; Tryk, D.A.; Fujishima, A. J. Elec-
troanal. Chem. 1999, 473, 173.
121. Witek, M. Dissertation. Michigan State University, 2002.
122. Cvacka, J.; Quaiserova, V.; Park, J.; Show, Y.; Muck, A.; Swain, G.M.
Anal. Chem., 2003, 75, 2678.
123. Mermoux, M.; Marcus, B.; Swain, G.M.; Butler, J.E. J. Phys. Chem. B
2002, 106, 10816.
124. Haymond, S. Dissertation. Michigan State University, 2002.
125. Wang, J. Dissertation. Michigan State University, 2002.
126. Wang, J.; Swain, G.M. J. Electroanal. Chem. 2003, 150, E24.
127. Haymond, S.; Zak, J.K.; Butler, J.E.; Babcock, G.T.; Swain, G.M. Ana-
lytica Chimica Acta 2003. in press.
128. Wang, J.; Swain, G.M. J. Electrochem. Soc. 2003, 149, E24.
129. ASTM 6-0675.
130. Knight, D.S.; White, W.B. J. Mater. Res. 1989, 4, 385.
131. Nemanich, R.J.; Glass, J.T.; Lucovsky, G.; Shroder, R.E. J. Vac. Sci.
Technol. 1988, A6, 1783.
132. Bergman, L.; Nemanich, R.J. J. Appl. Phys. 1995, 78, 6709.
133. Prawer, S.; Nugent, K.W.; Jamieson, D.N.; Orwa, J.O.; Bursill, L.A.; Peng,
J.L. Chem. Phys. Lett. 2000, 332, 93.
134. Ferrari, A.C.; Robertson, J. Phys. Rev. B 2001, 63, 121405.
135. Gruen, D.M.; Krauss, A.R.; Zuiker, C.D.; Csencsits, R.; Terminello, L.T.;
Carlisle, J.A.; Jimenez, I.; Sutherland, D.G.J.; Shu, D.K.; Tong, W.;
Himpsel, F.J. Appl. Phys. Lett. 1996, 68, 1640.
136. McCreery, R.L. In Electroanalytical Chemistry; Bard, A.J., Ed.; New
York: Marcel Dekker, 1991; 17, 221374.
137. Bhattacharyya, S.; Auciello, O.; Birrell, J.; Carlisle, J.A.; Curtiss, L.A.;
Goyette, A.N.; Gruen, D.M.; Krauss, A.R.; Schlueter, J.; Sumant, A.;
Zapol, P. Appl. Phys. Lett. 2001, 79, 1441.
138. Gerischer, H. Electrochim. Acta 1990, 35, 1677.
139. McDermott, C.A.; Kneten, K.R.; McCreery, R.L. J. Electrochem. Soc.
1993, 140, 2593.
140. Kneten, K.R.; McCreery, R.L. Anal. Chem. 1992, 64, 2518.
141. Chen, P.; Fryling, M.A.; McCreery, R.L. Anal. Chem. 1995, 67, 3115.
142. Chen, P.; McCreery, R.L. Anal. Chem. 1996, 68, 3958.
143. DuVall, S.H.; McCreery, R.L. J. Am. Chem. Soc. 2000, 122, 6759.
144. Ranganathan, S.; Kuo, T-C.; McCreery, R.L. Anal. Chem. 1999, 71, 3574.
145. Ferreira, N.G.; Silva, L.L.G.; Corat, E.J.; Trava-Airoldi, V.J. Diam. Rel.
Mat. 2002, 11, 1523.
146. Smentkowski, V.S.; Yates, J.T., Jr. Science 1996, 271, 193.
147. Ohtani, B.; Y-ho. Kim, T.; Hashimoto, Yano.K.; Fujishima, A.; Uosaki,
K. Chem. Lett. 1998, 953.
148. Kim, C.S.; Mowrey, R.C.; Butler, J.E.; Russell, J.N., Jr. J. Phys. Chem. B
1998, 102, 9290.
149. Allongue, P.; Delamar, M.; Desbat, B.; Fagebaume, O.; Hitmi, R.; Pinson,
J.; Saveant, J.M. J. Am. Chem. Soc. 1997, 119, 201.
150. Saby, C.; Ortiz, B.; Champagne, G.Y.; Belanger, D. Langmuir 1997, 13,
6805.
151. Yang, H.H.; McCreery, R.L. Anal. Chem. 1999, 71, 4081.
152. Kariuki, J.K.; McDermott, M.T. Langmuir 1999, 15, 6534.
153. Harnisch, J.A.; Gazda, D.B.; Anderegg, J.W.; Porter, M.D. Anal. Chem.
2001, 73, 3954.
Swain 274
154. Strother, T.; Knickerbocker, T.; Russell, J.N., Jr.; Butler, J.E.Smith, L.M.;
Hamers, R.J.Langmuir 2002, 18, 968.
155. Yang, W.; Auciello, O.; Butler, J.E.; Cai, W.; Carlisle, J.A.; Gerbi, J.E.;
Gruen, D.M.; Knickerbocker, T.; Lasseter, T.L.; Russell, J.N., Jr.; Smith,
L.M.; Hamers, R.J. Nature Mater. 2002, 1, 253.
156. Prado, C.; Wilkins, S.J.; Marken, F.; Compton, R.G. Electroanalysis 2002,
14, 262.
157. Tatsuma, T.; Mori, H.; Fujishima, A. Anal. Chem. 2000, 72, 2919.
158. Stotter, J.; Haymond, S.; Zak, J.K.; Show, Y.; Cvackova, Z.; Swain, G.M.
Interface 2003, 12, 33.
159. Tabor, C.W.; Tabor, N. Annu. Rev. Biochem. 1984, 53, 749.
160. Pegg, A.E. Cancer Res. 1988, 49, 759.
161. Marton, L.J.; Feuerstein, B.G. Pharm. Res. 1986, 3, 311.
162. Vitt, J.E.; Johnson, D.C. J. Electrochem. Soc. 1992, 139, 774.
163. He, L.; Anderson, J.R.; Franzen, H.F.; Johnson, D.C. Chem. Mater. 1997,
9, 715.
164. Treimer, S.E.; Feng, J.; Scholten, M.D.; Johnson, D.C.; Davenport, A.J. J.
Electrochem. Soc. 2001, 148, E459.
165. Ge, J.; Johnson, D.C. J. Electrochem. Soc. 1995, 142, 1525.
166. Dobberpuhl, D.A.; Johnson, D.C. J. Chrom. A 1995, 694, 391.
167. Hoekstra, J.C.; Johnson, D.C. Anal. Chim. Acta 1999, 390, 45.
168. Gherardini, L.; Michaud, P.A.; Panizza, M.; Comninellis, Ch.; Vatistas, N.
J. Electrochem. Soc. 2001, 148, D78.
169. Iniesta, J.; Marchaud, P.A.; Panizza, M.; Comninellis, Ch. Elecrochem.
Comm. 2001, 3, 346.
170. Panizza, M.; Michaud, P.A.; Cerisola, G.; Comninellis, Ch. J. Electroanal.
Chem. 2001, 507, 206.
171. Kuwana, T.; Darlington, R.K.; Leedy, D.W. Anal. Chem. 1964, 36, 2023.
172. Kuwana, T.; Winograd, N. In Electroanalytical Chemistry; Bard, A.J., Ed.;
New York: Marcel Dekker, 1974; 7, 1.
173. Donley, C.; Dunphy, D.; Paine, D.; Carter, C.; Nebesny, K.; Lee, P.; Al-
loway, D.; Armstrong, N.R. Langmuir 2002, 18, 450.
174. Landrum, H.L.; Salmon, R.T.; Hawkridge, F.M. J. Am. Chem. Soc. 1977,
99, 3154.
175. Stegemiller, M.L.; Heineman, W.R.; Ridgway, T.H.; Seliskar, C.J.; Bryan,
S.A.; Hubler, T.; Sell, R.L. Environ. Sci. Technol. 2002, 37, 123.
176. Zaitsev, A.M. Optical Properties of Diamond. Berlin: Springer-Verlag.
177. Pankove, J.I.; Qui, C.-H. In Synthetic Diamond: Emerging CVD Science
and Technology; Spear, K.E., Dismukes, J.P., Eds.; John Wiley & Sons,
Inc.: New York, 1994; 401.
178. McNamara, K.M.; Williams, B.E.; Gleason, K.K.; Scruggs, B.E. J. Appl.
Phys. 1994, 76, 2466.
179. Cheng, H.Y.; Sackett, P.H.; McCreery, R.L. J. Am. Chem. Soc. 1978, 100,
962.
180. Ates, S.; Somer, G. J. Chem. Soc., Faraday Trans. 1 1981, 77, 859.
181. Horie, S.; Morrison, M. J. Biol. Chem. 1963, 238, 2859.
182. Moss, D.; Nabedryk, E.; Breton, J.; Ma ntele, W. Eur. J. Biochem. 1990,
187, 565.
183. Eaton, W.A.; Hofrichter, J. Meth. Enzymol. 1981, 76, 175.
184. Dawson, R.M.C.; Elliot, D.C.; Elliot, W.H.; Jones, K.M. Data for
Biochemical Research; 3rd Ed. New York: Oxford University Press, Inc,
1986.
185. Hawkridge, F.; Kuwana, T. Anal. Chem. 1973, 45, 1021.
186. Auer, E.; Freund, A.; Pietsch, J.; Tacke, T. Appl. Catalysis A: General 1998,
173, 259.
187. Sepulveda-Escribano, A.; Coloma, F.; Rodriguez-Reinoso, F. Appl. Catal-
ysis A: General 1998, 173, 247.
188. Appleby, A.J. Corrosion 1987, 43, 398.
189. Callstrom, M.R.; Neenan, T.X.; McCreery, R.L.; Alsmeyer, D.C. J. Am.
Chem. Soc. 1990, 112, 4955.
190. Pocard, N.L.; Alsmeyer, D.C.; McCreery, R.L.; Neenan, T.X.; Callstrom,
M.R. J. Am. Chem. Soc. 1992, 114, 769.
191. Scharifker, B.; Hills, G. Electrochim. Acta 1983, 28, 879.
192. Zeliger, H. J. Electrochem. Soc. 1967, 114, 144.
193. Blurton, K.F.; Greenberg, P.; Osmin, H.G.; Butt, D.R. J. Electrochem.
Soc. 1972, 119, 559.
194. Bett, J.; Lundquist, J.; Washington, E.; Stonehart, P. Electrochim. Acta
1973, 18, 343.
195. Kunz, H.R.; Gruver, G.A. J. Electrochem. Soc. 1975, 122, 1279.
196. Vogal, W.M.; Baris, J.M. Electrochim. Acta 1977, 22, 1259.
197. Sattler, M.L.; Ross, P.N. Ultramicroscopy 1986, 20, 21.
198. Peuchert, M.; Yoneda, T.; Betta, R.A.D.; Boudart, M. J. Electrochem. Soc.
1986, 133, 944.
199. Ticianelli, E.A.; Deroulin, C.R.; Srinivasan, S. J. Electrochem. Soc. 1988,
251, 275.
200. Kinoshita, K. J. Electrochem. Soc. 1990, 137, 845.
201. Perez, J.; Tabaka, A.A.; Gonzales, E.R.; Ticianelli, E.A. J. Electrochem.
Soc. 1994, 141, 432.
202. Poirier, J.A.; Stoner, G.E. J. Electrochem. Soc. 1994, 141, 425.
203. Takasu, Y.; Ohashi, N.; Zhang, X.-G.; Murakami, H.M.Y.; Sato, S.; Yahi-
kozawa, K. Electrochim. Acta 1996, 41, 2595.
204. Gojkovic, S.L.; Zecevic, S.K.; Savinell, R.F. J. Electrochem. Soc. 1998,
145, 3713.
205. Bregoli, L. Electrochim. Acta 1978, 23, 489.
Swain 276
206. Kinoshita, K. Electrochemical Oxygen Technology. NewYork: JohnWiley &
Sons, Inc., 1992.
207. Kita, H.; Lei, H.; Gao, Y. J. Electroanal. Chem. 1994, 379, 407.
208. Ross, P.N.; Andricacos, P.C. J. Electroanal. Chem. 1983, 154, 205.
209. Hsueh, K.L.; Gonzalez, E.R.; Srinivasan, S.; Chin, D.T. J. Electrochem.
Soc. 1984, 131, 823.
210. Damjanovic, A.; Genshaw, M.A.; Bockris, J.O.M. J. Electrochem. Soc.
1967, 114, 466.
211. Damjanovic, A.; Brusic, V. Electrochim. Acta 1967, 12, 615.
212. Gojkovic, S.L.; Vidakovic, T.R. Electrochim. Acta 2001, 47, 633.
213. Chrzanowaki, W.; Wieckowski, A. In Interfacial Electrochemistry: Theory,
Experiment and Applications; Wieckowski, A., Ed.; Marcel Dekker: New
York, 1999; 937.
214. Chrzanowski, W.; Kim, H.; Tremiliosi-Filho, G.; Wieckowski, A.; Grzy-
bowska, B.; Kulesza, P. J. New Mater. Electrochem. Sys. 1998, 1, 31.
215. Chrzanowski, W.; Wieckowski, A. Langmuir 1998, 14, 1967.
Abaturov MA, 195, 271
Adler J, 191, 269
Akiba Y, 194, 269
Albin S, 195, 270
Alehashem S, 195, 201, 270
Allongue P, 217, 274
Alloway D, 242, 275
Alsmeyer DC, 254, 276
Anderegg JW, 217, 274
Anderson AB, 191, 195, 201, 202,
203, 217, 252, 269, 270
Anderson JR, 230, 275
Ando T, 189, 195, 239, 269, 273
Andricacos PC, 263, 277
Angus JC, 191, 195, 202, 203, 217,
252, 269, 270, 272, 273
Angus JCJC, 195, 201, 269
Aoyagui SJ, 108, 178
Appleby AJ, 251, 276
Aramata A, 65, 99
Argoitia A, 195, 202, 203, 217,
252, 270, 272
Armstrong NR, 242, 275
Asher SA, 188, 208, 268
Ataka K, 34, 97
Ataka TJ, 11, 95
Ates S, 247, 276
Atkinson JD, 33, 97
Auciello O, 204, 207, 208, 219,
274, 275
Auer E, 251, 276
Auge J, 11, 95
Avaca LA, 33, 54, 65, 97
Avyigal Y, 195, 221, 271, 272
Awada M, 195, 269
Babcock GT, 195, 228, 247, 250,
273, 274
Babenko SD, 109, 178
Bacskai J, 53, 98
Bamdad C, 164, 179
Banavar JR, 17, 96
Bandey HL, 6, 10, 94, 95
Baranov AM, 195, 272
Barbara PF, 104, 176
Bard AJ, 106, 107, 154, 159, 167,
177, 179
Baris JM, 260, 276
Barker GC, 106, 109, 121, 177,
178
Barnes C, 11, 95
Barrat JL, 21, 22, 96
Beck R, 6, 11, 26, 71, 94, 96, 99
Behler Z, 195, 239, 241, 242, 243,
245, 273
Behling C, 10, 95
Beitz JV, 108, 178
Belanger D, 217, 274
Benderskii VA, 109, 132, 178
Bennion BC, 108, 178
Ber KS, 108, 177
Bergman L, 199, 200, 274
Bett J, 260, 276
Betta RAD, 260, 276
279
AUTHOR INDEX
Italicized pages locate the full title of the work cited.
Bhattacharyya S, 204, 207, 208,
274
Bilewicz R, 104, 176
Birrell J, 204, 207, 208, 274
Birss VI, 54, 55, 58, 98
Bitziou E, 65, 99
Blank E, 189, 269
Blau JG, 195, 230, 231, 236, 270
Bloor D, 103, 176
Blurton KF, 260, 276
Bockris JO, 43, 97
Bockris JOM, 263, 277
Bocquet L, 21, 22, 96
Boonma L, 195, 271
Borges G, 26, 50, 81, 96
Bormett RW, 188, 208, 268
Bottom VG, 11, 95
Bottura GJ, 109, 178
Bouamrane R, 195, 272
Boudart M, 260, 276
Bowden EF, 166, 167, 179
Bregoli L, 260, 276
Brek WG, 134, 179
Breton J, 250, 276
Brett DJ, 76, 81, 82, 99
Brevnov DA, 105, 150, 177
Brinkman HC, 30, 97
Brown MJ, 10, 95
Bruckenstein S, 5, 17, 26, 33, 34,
36, 55, 58, 60, 65, 81, 86, 88,
94, 96, 97, 98, 99
Bruckenstein SJ, 6, 11, 14, 94, 95
Brugger K, 37, 97
Brumlik CJ, 107, 177
Brunschwig BS, 168, 169, 180
Bruschi L, 16, 96
Brusic V, 263, 264, 277
Bryan SA, 242, 275
Bryce MR, 103, 176
Buck RP, 153, 179
Bund A, 6, 71, 76, 77, 95, 99
Burgin TP, 104, 176
Bursill LA, 199, 200, 274
Butler JE, 185, 186, 195, 198, 200,
201, 203, 217, 219, 239,
241, 243, 250, 252, 268,
270, 272, 273, 274, 275
Butler JM, 195, 271
Butt DR, 260, 276
Buttry DA, 2, 8, 11, 53, 86, 94, 98
Cady WG, 11, 95
Cai W, 219, 275
Calcattera LT, 105, 113, 116, 177
Cali GJ, 105, 176
Callender RH, 108, 178
Callstrom MR, 254, 276
Cao J, 109, 110, 111, 178
Carey JJ, 195, 238, 273
Carlin A, 16, 96
Carlisle JA, 188, 199, 200, 204,
207, 208, 219, 268, 274, 275
Carter C, 242, 275
Carter MT, 163, 179
Cavicvlasak BA, 22, 96
Cerisola G, 195, 239, 273, 275
Cernosek RW, 10, 95
Chalfant K, 166, 167, 168, 179,
180
Chalker PR, 195, 272
Chamberlain RV, 104, 176
Chambers F, 195, 201, 270
Champagne GY, 217, 274
Chang M, 54, 55, 58, 98
Chattoraj M, 113, 179
Chen A, 105, 176
Chen C, 195, 271
Chen J, 104, 176
Author Index 280
Chen P, 204, 274
Chen Q, 195, 201, 202, 203, 204,
205, 207, 210, 219, 220,
239, 252, 269, 270, 272
Cheng HY, 245, 247, 276
Chiarello RP, 15, 16, 96
Chidsey C, 169, 180
Chidsey CED, 103, 105, 107, 112,
113, 116, 131, 148, 150,
161, 162, 163, 165, 167,
168, 175, 176, 177, 180
Chidsey CEDJ, 103, 105, 150, 161,
162, 164, 167, 175
Chin DT, 263, 277
Chong Y, 164, 179
Chottiner G, 195, 272
Christ CS Jr, 195, 238, 273
Chrzanowaki W, 267, 277
Chung FL, 11, 14, 95
Churaev NN, 20, 96
Climent V, 109, 178
Closs GL, 105, 113, 116, 177,
179
Cochran HD, 20, 96
Cohen H, 195, 271
Coles BA, 109, 178
Coloma F, 251, 276
Comninellis C, 195, 239, 273
Comninellis CH, 239, 275
Compton RG, 109, 178, 195, 224,
272, 273, 275
Conway BE, 54, 98
Cooper JB, 195, 270
Corat EJ, 212, 274
Corrigan TD, 195, 201, 204, 205,
207, 270
Coufal H, 111, 178
Cox AP, 54, 55, 57, 58, 98
Creager S, 169, 180
Creager SE, 105, 154, 163, 164,
167, 168, 177, 179, 180
Crooker JC, 154, 167, 179
Csencsits R, 199, 200, 274
Cummings PT, 20, 96
Curtiss LA, 188, 204, 207, 208,
268, 274
Cvacka J, 195, 205, 270, 273
Cvackova Z, 228, 275
Czerlinski G, 106, 108, 177
Daikhin L, 5, 6, 7, 19, 23, 25, 27,
28, 29, 30, 31, 33, 34, 35, 36,
37, 40, 41, 42, 43, 44, 46, 47,
48, 58, 59, 63, 64, 66, 72, 73,
74, 75, 94, 95, 96, 97, 99
Daikhin LJ, 6, 76, 78, 79, 80, 81,
82, 95
Dallas T, 195, 270, 271
Daly C, 16, 96
Damaskin BB, 57, 98
Damjanovic A, 263, 264, 277
Darlington RK, 242, 275
Daujotis V, 6, 65, 76, 81, 82, 95,
98, 99
Davenport AJ, 230, 275
Davidson JL, 191, 269
Dawson RMC, 250, 276
De Cli SV, 76, 81, 82, 99
de Gennes JP, 21, 96
Deakin MR, 53, 65, 98, 99
DeClements R, 195, 270, 271
Delahay P, 39, 97, 106, 121, 177
Delamar M, 217, 274
Dementev AP, 195, 272
Deroulin CR, 260, 276
Desbat B, 217, 274
Devanathan MAV, 43, 97
Di Quarto F, 195, 271
Author Index 281
Diau EWG, 104, 176
Diesing D, 49, 98
Digel JJ, 15, 16, 21, 96
Ding Z, 107, 154, 167, 177
Dismukes JP, 242, 245, 275
Dobberpruhl DA, 230, 275
Doblhofer K, 6, 94
Dolidze TD, 105, 154, 167, 177
Donley C, 242, 275
Dudek S, 103, 105, 150, 161, 162,
164, 167, 175
Dudek SP, 103, 105, 150, 161, 162,
167, 176
Duncan-Hewitt WC, 5, 11, 22, 71,
94, 96
Dunphy D, 242, 275
Duo I, 195, 273
Durning CJ, 87, 99
Dutton PL, 105, 113, 116, 177
DuVall SH, 204, 205, 274
Dyer RB, 108, 178
Eaton SC, 191, 269
Eaton WA, 250, 276
Ebersole RC, 11, 96
EerNisse EP, 37, 97
Emov IO, 50, 98
Eichelbaum F, 11, 95
Eigen M, 106, 108, 177
Einaga Y, 191, 195, 269, 272
Einzel D, 32, 33, 97
Eisenberg D, 154, 179
Eklund JC, 195, 272
Elkin VV, 195, 271, 272
Elliot DC, 250, 276
Elliot WH, 250, 276
Elsayed-Ali HE, 109, 110, 111,
178
Erdemir A, 188, 208, 269
Evstefeeva YE, 191, 195, 269, 272
Eyring EM, 108, 178
Fagebaume O, 217, 274
Fajardo AM, 113, 178
Fan FR, 107, 177
Farias GA, 86, 99
Faulkmer LR, 106, 107, 159, 177
Faulkner LR, 163, 179
Fausett BM, 195, 201, 207, 270
Feinman R, 17, 96
Feldberg SW, 103, 105, 128, 131,
138, 141, 148, 150, 153,
161, 162, 164, 166, 167,
168, 175, 176, 179, 180
Feldman LC, 195, 271
Feng J, 230, 275
Fenske GR, 188, 208, 269
Fensore A, 26, 96
Ferrari AC, 199, 200, 274
Ferraris JP, 168, 180
Ferreira NG, 212, 274
Feuerstein BG, 230, 275
Finklea HO, 105, 112, 150, 167,
168, 169, 177, 178, 180
Flechsig G, 195, 273
Flynn GW, 108, 178
Fong HKY, 104, 176
Foord JS, 195, 272, 273
Forster RJ, 106, 116, 163, 177, 179
Foss RP, 11, 96
Foti G, 195, 239, 273
Franta DJ, 107, 177
Franzen HF, 230, 275
Frauenfelder H, 105, 113, 116, 177
Fresenius Z, 195, 271
Freund A, 251, 276
Frey G, 5, 11, 23, 26, 71, 94
Frey GC, 11, 95
Author Index 282
Frubose C, 6, 94
Frumkin AN, 57, 98
Fryling MA, 204, 274
Fujishima A, 191, 195, 205, 214,
217, 221, 224, 225, 226,
227, 228, 269, 270, 271,
272, 273, 274, 275
Fujishima AJ, 195, 270
Furtak TE, 65, 70, 98
Futamata M, 49, 98
Gabrielli C, 108, 178
Gai F, 108, 178
Gaidamauskas E, 65, 98
Galun E, 195, 271
Gandini D, 195, 239, 273
Gao Y, 109, 110, 111, 178, 263,
277
Gardner AW, 109, 178
Gazda DB, 217, 274
Ge J, 230, 275
Geng L, 103, 105, 131, 148, 150,
176
Genshaw MA, 263, 277
Gerbi JE, 219, 275
Gerischer H, 204, 274
Gerisher H, 108, 177
Gewirth AA, 65, 66, 99
Gherardini L, 239, 275
Gileadi E, 4, 5, 6, 7, 23, 25, 27, 34,
35, 40, 41, 42, 43, 44, 45, 46,
47, 48, 58, 59, 61, 62, 63, 64,
66, 70, 72, 73, 74, 75, 76, 78,
79, 80, 81, 82, 87, 88, 94, 95,
96, 97, 98, 99
Gileadi EJ, 19, 34, 36, 37, 40, 88,
96
Gilmanshin R, 108, 178
Glass JT, 199, 200, 274
Gleason KK, 189, 245, 269, 275
Glidle A, 6, 11, 14, 94, 95
Gloaguen F, 58, 59, 98
Goeting CH, 195, 272
Gojkovic SL, 260, 264, 276, 277
Goldanskii VI, 105, 113, 116,
177
Goldman M, 103, 105, 176
Gonsalves M, 6, 94, 95
Gonzales ER, 260, 276
Gonzalez ER, 263, 277
Gootzen JFE, 54, 55, 57, 58, 98
Gordon JG, 5, 17, 26, 29, 50, 94,
96
Gordon JS, 36, 97
Gosavi S, 168, 169, 179
Goyette AN, 204, 207, 208, 274
Gozin M, 164, 179
Grabow H, 108, 178
Granger MC, 193, 195, 201, 202,
203, 204, 205, 207, 212,
214, 219, 220, 221, 222,
224, 225, 239, 245, 247,
252, 269, 270
Granick S, 19, 96
Gray HB, 103, 176
Green NJ, 105, 113, 116, 177
Green NS, 113, 116, 179
Greenberg P, 260, 276
Gruen DM, 188, 195, 199, 200,
201, 203, 204, 205, 207,
208, 219, 239, 252, 268,
269, 270, 274, 275
Gruendler P, 195, 273
Grundler P, 108, 178
Gruver GA, 260, 276
Grzybowska B, 267, 277
Guoxiong, 195, 271
Gupta SA, 20, 96
Author Index 283
Hadley L, 163, 179
Haenni W, 195, 239, 273
Hagans PL, 195, 238, 273
Halter M, 188, 208, 269
Hamers RJ, 217, 219, 275
Haneda H, 189, 269
Hanks A, 193, 195, 201, 204, 205,
212, 245, 247, 269
Hanshew DD, 105, 177
Haouni A, 189, 269
Harima Y, 108, 178
Harnisch JA, 217, 274
Hashimoto K, 195, 269, 271, 272
Hashimoto YK, 217, 274
Hass G, 163, 179
Hauptmann P, 11, 95
Hawkridge F, 250, 276
Hawkridge FM, 242, 275
Haymond S, 195, 201, 203, 204,
228, 230, 239, 247, 248,
249, 250, 252, 270, 273,
274, 275
He L, 230, 275
Heerle P, 111, 178
Heineman WR, 242, 275
Henderson MJ, 65, 99
Hepel M, 2, 65, 94, 98, 99
Herek JL, 104, 176
Herrick R III, 195, 271
Hervet H, 20, 22, 96
Heusler KE, 50, 98
Hillman AR, 6, 10, 11, 14, 60, 63,
65, 89, 94, 95, 98, 99
Hillman ARJ, 26, 96
Hills G, 259, 276
Himpsel FJ, 199, 200, 274
Hirose Y, 194, 269
Hirsche BL, 195, 270
Hishita S, 189, 269
Hitmi R, 217, 274
Hockett L, 105, 177
Hoekstra JC, 230, 275
Homan H, 108, 178
Hofrichter J, 250, 276
Holtz MW, 195, 270, 271
Holzle MH, 49, 98
Hoogvliet JC, 87, 99
Hook F, 87, 99
Horanyi GJ, 53, 65, 88, 98, 99
Horie S, 248, 276
Horner DA, 188, 207, 208, 268
Hsu CP, 112, 113, 178
Hsueh KL, 263, 277
Hsung RP, 103, 105, 150, 161, 162,
164, 167, 175
Hubler T, 242, 275
Hui KC, 76, 81, 82, 99
Hupert M, 193, 195, 201, 203, 204,
205, 212, 239, 245, 247,
252, 269, 270
Hupert ML, 195, 201, 207, 270
Hwang E, 53, 58, 81, 98
Hwang TS, 87, 99
Hynes JT, 168, 179
Iida M, 194, 269
Iijima M, 5, 94
Ikeda N, 33, 54, 55, 58, 60, 65, 81,
97, 98, 99
Ikeda NJ, 54, 58, 98
Iniesta J, 239, 275
Inzelt G, 53, 65, 88, 98, 99
Irish DE, 37, 97
Israel J, 195, 272
Iwaki M, 195, 269
Jabbarzadeh A, 33, 97
Jack Fajer, 167, 179
Author Index 284
Jackson T, 195, 230, 231, 236, 270
Jamieson DN, 199, 200, 274
Jasaitis D, 82, 99
Jeerey M, 54, 98
Jerey M, 33, 97
Jerkiewicz G, 54, 86, 98, 99
Jianzhong Z, 195, 271
Jimenez I, 199, 200, 274
Joanny JF, 30, 97
Johna Leddy, 131, 179
Johnson DC, 36, 97, 230, 275
Johnson DCJ, 230, 275
Johnson MD, 113, 116, 179
Johnson RM, 195, 270
Johnston C, 195, 272
Jolley S, 195, 230, 231, 236, 270
Jones F, 195, 272
Jones JL, 30, 97
Jones KM, 250, 276
Jortner J, 103, 176
Joseph IS, 11, 95
Josse F, 11, 95
Jusys Z, 33, 34, 36, 97
Kalish R, 191, 195, 221, 269, 271,
272
Kalish RJ, 195, 271
Kanada H, 195, 269
Kanazawa KK, 5, 10, 17, 26, 29,
50, 81, 94, 95, 96
Kang WP, 191, 269
Kanige K, 65, 98
Kariuki JK, 217, 274
Karube I, 11, 95
Kasemo B, 21, 87, 96, 99
Katsuki N, 195, 239, 273
Katz A, 195, 271
Katz G, 6, 7, 27, 63, 64, 72, 73, 74,
75, 76, 78, 79, 80, 81, 82, 95
Kautek W, 33, 36, 97
Kauzmann W, 154, 179
Kawaguchi T, 65, 99
Kawasaki S, 195, 239, 273
Kayhyem JF, 164, 179
Keddam M, 108, 178
Kelly JJ, 87, 99
Kern P, 43, 97
Kerns DV, 191, 269
Kerns SE, 191, 269
Keyes TE, 116, 163, 179
Kim CS, 217, 274
Kim H, 267, 277
Kim KB, 87, 99
Kim MS, 87, 99
Kim T, 217, 274
Kim Y, 195, 272
Kim ZH, 104, 176
Kimura K, 11, 14, 95
Kinoshita K, 260, 276, 277
Kipling AL, 22, 96
Kirbs A, 108, 178
Kirchner V, 107, 177
Kirshnan CV, 103, 105, 176
Kita H, 263, 277
Kita HJ, 55, 98
Kitamura F, 34, 97
Klein J, 19, 96
Kneten KR, 204, 205, 274
Knickerbocker T, 217, 219, 275
Knight DS, 198, 199, 274
Kobashi K, 189, 195, 201, 203,
222, 239, 252, 253, 269, 270
Kobashi KJ, 195, 201, 203, 252,
253, 270
Kolb DM, 49, 98
Koller KB, 134, 179
Kondo TD, 195, 205, 214, 273
Konov VI, 188, 208, 269
Author Index 285
Koplik J, 17, 96
Koppang M, 195, 230, 231, 236,
270
Koppang MD, 195, 230, 231, 236,
270
Koshtariya DE, 105, 154, 167, 177
Krauss AR, 188, 195, 199, 200,
201, 204, 205, 207, 208,
268, 269, 270, 274
Krauss ARAR, 188, 208, 269
Krim J, 15, 16, 20, 21, 96
Krivenko AG, 109, 132, 178
Krotova MD, 195, 271, 272
Kubota Y, 195, 226, 271
Kuhn AT, 68, 99
Kulesza PJ, 267, 277
Kunz HR, 260, 276
Kuo TC, 195, 212, 270, 274
Kurosu T, 194, 269
Kuwana T, 242, 250, 275, 276
Lachenwitzer A, 66, 68, 99
Lam E, 164, 179
Lamy C, 58, 59, 98
Landau U, 195, 202, 203, 217, 252,
270
Landolt D, 43, 97
Landrum HL, 242, 275
Lanzafame JM, 104, 176
Laptev VA, 195, 272
Lasseter TL, 219, 275
Laurat P, 19, 96
Leddy J, 103, 105, 131, 148, 150,
154, 155, 167, 176, 179
Lee P, 242, 275
Lee WW, 19, 96
Leedy DW, 242, 275
Leger JM, 58, 59, 98
Leger L, 20, 22, 96
Lei H, 263, 277
Lei HW, 53, 98
Lessard J, 86, 99
Levy-Clement C, 195, 272
Levy-Clement CJ, 195, 271
Levy DH, 113, 179
Lewis NS, 105, 113, 176, 178
Li FB, 60, 63, 89, 98
Li LF, 195, 272
Li TTT, 105, 176
Li Z, 26, 96
Liess HD, 11, 96
Lin J, 134, 179
Lin Z, 89, 99
Lin ZX, 11, 95
Linford MR, 103, 105, 150, 161,
167, 168, 175, 180
Lipkowski J, 97
Lister TE, 195, 201, 202, 203, 219,
220, 239, 252, 269, 270
Liu L, 112, 167, 178
Liu M, 32, 33, 97
Liu S, 188, 208, 268
Liu X, 195, 271
Liu Y, 103, 105, 150, 161, 167, 168,
175
Liu YP, 168, 180
Lizee JF, 108, 178
Loh KP, 189, 269
Loo BH, 195, 228, 271
Loughman P, 116, 163, 179
Louwerix E, 2, 94
Lowery SN, 195, 238, 273
Lubetkin SD, 60, 63, 89, 98
Lucazeau G, 189, 195, 204, 205,
269, 273
Lucklum R, 10, 95
Lucovsky G, 199, 200, 274
Lundquist J, 260, 276
Author Index 286
Luo J, 164, 179, 188, 208, 268
Lyons LJ, 154, 167, 179
MacFarquhar RA, 103, 105, 153,
176
Machado SAS, 33, 54, 58, 65, 97,
98
Machonkin M, 194, 269
MacLean T, 164, 179
Magnussen CJ, 66, 68, 99
Majda MJ, 104, 176
Malyszko J, 41, 90, 99
Manivannan AD, 195, 221, 224,
270
Mantele W, 250, 276
Mantell DA, 109, 110, 111, 178
Mao Y, 53, 58, 81, 98
Maradurin AA, 86, 99
Marchaud PA, 239, 275
Marcus B, 195, 198, 200, 273
Marcus RA, 112, 113, 168, 169,
178, 179
Marken F, 195, 224, 272, 273, 275
Marques CM, 30, 97
Martin CR, 107, 177
Martin HB, 195, 202, 203, 217,
239, 242, 243, 252, 270, 273
Martin SJ, 5, 10, 11, 23, 26, 71, 94,
95
Marton LJ, 230, 275
Mazin VM, 195, 272
Mazin VV, 195, 272
McCauley TG, 188, 208, 268, 269
McCreery RL, 195, 202, 204, 205,
212, 217, 245, 247, 254,
270, 274, 276
McDermott CA, 204, 205, 274
McDermott MT, 217, 274
McKeown RAJ, 195, 272
McNamara KM, 245, 275
Melory OJ, 26, 50, 53, 65, 96, 98,
99
Mermoux M, 189, 195, 198, 200,
204, 205, 269, 273
Meyer H, 6, 95
Miao W, 107, 154, 167, 177
Michaud PA, 195, 239, 273, 275
Michler J, 189, 269
Miller B, 195, 221, 271, 272
Miller BJ, 111, 178
Miller CJ, 104, 176
Miller JR, 105, 113, 116, 176, 177,
179
Miller RJD, 104, 109, 110, 111,
176, 178
Mirkin MV, 107, 177
Mishuk V, 195, 271
Mishuk V Ya, 195, 272
Mistura G, 16, 96
Mitra S, 189, 269
Miura M, 65, 99
Miwa D, 195, 225, 271
Miwa DW, 33, 54, 58, 65, 97, 98
Miwa TD, 195, 271
Miyata K, 189, 222, 269
Modestov AD, 195, 272
Mohilner DM, 48, 97, 98
Momota K, 195, 239, 273
Moore WJ, 154, 174, 179
Moret H, 2, 94
Mori H, 227, 275
Morin MJ, 51, 98
Moriya N, 195, 271
Morris R, 107, 177
Morrish A, 195, 271
Morrison M, 248, 276
Morrison PW Jr, 195, 239, 242,
243, 273
Author Index 287
Mort J, 194, 269
Mortimer CJ, 68, 99
Moss D, 250, 276
Mowrey RC, 217, 274
Muck A, 195, 273
Muller CJ, 104, 176
Muller K, 43, 97
Murakami HMY, 260, 276
Muraki Y, 195, 271
Muramatsu H, 11, 14, 95
Murray RW, 154, 163, 167, 179
Mutaaga G, 195, 273
Nabedryk E, 250, 276
Nahir TM, 115, 166, 167, 179
Nakamura T, 194, 269
Nanbu N, 34, 97
Nandi NJ, 48, 98
Naruki H, 194, 269
Natishan PM, 195, 238, 271, 273
Nebesny K, 242, 275
Neenan TX, 254, 276
Nemanich RJ, 199, 200, 274
Newman JJ, 107, 153, 177
Newton MD, 103, 105, 128,
138, 141, 150, 161, 162,
164, 167, 168, 175, 176,
180
Niece BK, 65, 66, 99
Nikolsky P, 91, 99
Nishitani-Gamo M, 189, 195, 239,
269, 273
Nitzan AJ, 104, 176
Noel MAM, 6, 11, 95, 96
Nomura T, 5, 94
Notsu H, 195, 205, 214, 271, 272,
273
Nugent KW, 199, 200, 274
Nyikos L, 195, 271
OConnor S, 164, 179
OGrady WE, 195, 238, 273
Ohashi N, 260, 276
Ohnishi K, 195, 272
Ohsaka T, 11, 34, 95, 97
Ohtani B, 195, 217, 272, 274
Okajima T, 11, 95
Okano K, 194, 269
Okino F, 195, 239, 273
Okumura K, 194, 269
Olsen GT, 164, 179
Onishi T, 189, 222, 269
Ortiz B, 217, 274
Orwa JO, 199, 200, 274
Osawa M, 34, 97
Osmin HG, 260, 276
Oyama N, 6, 11, 64, 95
Paine D, 242, 275
Palese S, 104, 176
Pan X, 188, 208, 268
Pang S, 195, 270
Panizza M, 195, 239, 273, 275
Pankove JI, 242, 245, 275
Panzer P, 32, 33, 97
Park J, 195, 273
Park JW, 195, 270
Park SG, 195, 272
Patel K, 195, 269, 271
Paulson B, 113, 179
Pegg AE, 230, 275
Peilio Z, 195, 271
Peneld KW, 105, 113, 116, 177
Peng JL, 199, 200, 274
Perez J, 260, 276
Perret A, 195, 239, 273
Persson BNJ, 15, 16, 21, 32, 96
Petrii OA, 57, 98
Petty MC, 103, 176
Author Index 288
Peuchert M, 260, 276
Piazza S, 195, 271
Pietsch J, 251, 276
Pimenov SM, 188, 208, 269
Pit R, 20, 22, 96
Pitterman U, 6, 11, 26, 71, 94, 96,
99
Plausinaitis D, 6, 76, 81, 95, 99
Pleskov YV, 191, 195, 269, 271,
272
Pocard NL, 254, 276
Poirier JA, 260, 276
Polish J, 6, 76, 81, 95
Popa E, 195, 226, 271, 272
Porter MD, 217, 274
Prado C, 195, 224, 273, 275
Prawer S, 199, 200, 274
Punturi S, 112, 167, 178
Pyati R, 154, 167, 179
Qin LC, 188, 208, 269
Qu D, 51, 98
Quaiserova V, 195, 273
Qui CH, 242, 245, 275
Rahman KMA, 87, 99
Rajakovic LV, 22, 96
Ramesham R, 195, 201, 269, 270,
272
Ranganathan S, 212, 274
Rao TN, 191, 195, 225, 226, 228,
269, 270, 271, 272, 273
Raudonis R, 6, 65, 76, 81, 82, 95,
99
Ravenscroft MS, 112, 167, 169,
178, 180
Raviv U, 19, 96
Rawlett AM, 104, 176
Redfern PC, 188, 207, 208, 268
Reed MA, 103, 104, 176
Reinmuth WH, 106, 121, 177
Reuben C, 195, 271
Ricco A, 5, 11, 23, 26, 71, 94
Richards TC, 107, 177
Richardson JN, 163, 179
Richer J, 37, 97
Ridgway TH, 242, 275
Roberts DJ, 60, 63, 89, 98
Robertson J, 199, 200, 274
Robinson DB, 107, 163, 177
Rodahl M, 21, 87, 96, 99
Rodrigo MA, 195, 273
Rodriguez-Reinoso F, 251, 276
Rogers LC, 103, 105, 131, 148,
150, 176
Rose MF, 195, 272
Rosler S, 11, 95
Ross PN, 260, 263, 276, 277
Rowe GK, 163, 179
Royea WJ, 113, 178
Rubinstein I, 11, 95
Russell JN Jr, 217, 219, 274, 275
Saada D, 191, 269
Saby C, 217, 274
Sachs SB, 103, 105, 150, 161, 162,
164, 167, 175
Sackett PH, 245, 247, 276
Sahimi M, 30, 97
Sahre M, 33, 36, 97
Sakaguchi I, 189, 195, 239, 269,
273
Sakairi H, 195, 269
Sakharova A, 195, 271
Sakurai H, 11, 95
Salmon RT, 242, 275
Santos MC, 33, 54, 58, 65, 97, 98
Sarada BV, 195, 228, 270, 271, 273
Author Index 289
Sarada S, 195, 226, 271
Sato S, 195, 260, 269, 276
Sato Y, 195, 269
Sattler ML, 260, 276
Sauerbrey GZ, 2, 94
Savinell RF, 260, 276
Scendo M, 41, 90, 99
Scharifker B, 259, 276
Scherson DA, 195, 272
Schlueter J, 204, 207, 208, 274
Schneider M, 71, 76, 99
Scholten MD, 230, 275
Schreiber F, 105, 176
Schumacher R, 2, 6, 8, 26, 50, 53,
58, 60, 81, 86, 94, 95, 96, 98
Schuster R, 107, 177
Schwarz HA, 154, 179
Schwitzgebel G, 6, 76, 77, 95, 99
Scruggs BE, 245, 275
Searsbrook G, 195, 272
Segal I, 54, 55, 58, 98
Seliskar CJ, 242, 275
Sell RL, 242, 275
Senturia S, 5, 11, 23, 26, 71, 94
Sepulveda-Escribano A, 251, 276
Sevastyanov AE, 195, 271
Shana ZA, 11, 95
Shay M, 5, 17, 33, 54, 55, 60, 65,
81, 86, 94, 97
Shi X, 195, 271
Shi Y, 113, 179
Shibata H, 195, 239, 273
Shimazu K, 55, 65, 98, 99
Shimizu R, 189, 222, 269
Shinozaki T, 195, 272
Shintani Y, 189, 222, 269
Shirakashi T, 195, 272
Shonzhong Z, 195, 271
Short K, 195, 271
Show Y, 195, 198, 204, 205, 207,
210, 221, 228, 239, 241,
242, 243, 245, 273, 275
Shroder RE, 199, 200, 274
Shu DK, 199, 200, 274
Sikes HD, 103, 105, 150, 161, 162,
167, 176
Silva C, 104, 176
Silva LLG, 212, 274
Sita LR, 103, 105, 150, 161, 162,
164, 167, 175
Slowinska K, 104, 176
Slowinski K, 104, 176
Smalley JF, 103, 105, 128, 131,
138, 141, 148, 150, 153,
154, 161, 162, 164, 166,
167, 168, 175, 176, 179, 180
Smentkowski VS, 217, 274
Smith CP, 112, 123, 178
Smith LM, 217, 219, 275
Soares DM, 6, 11, 33, 36, 94, 95,
97
Sobolev VD, 20, 96
Solina DH, 15, 16, 96
Somer G, 247, 276
Somov ANJ, 20, 96
Son DH, 104, 176
Sonthalia P, 195, 198, 201, 203,
204, 205, 207, 210, 221,
239, 252, 270, 273
Sopchak D, 195, 271
Spear KE, 242, 245, 275
Srinivasan S, 260, 263, 276, 277
Stegemiller ML, 242, 275
Stockel W, 60, 81, 98
Stolberg L, 37, 97
Stonehart P, 260, 276
Stoner BR, 195, 238, 273
Stoner GE, 260, 276
Author Index 290
Stotter J, 195, 201, 203, 228, 239,
241, 242, 243, 245, 252,
270, 273, 275
Strehlow H, 108, 177
Strojek JW, 195, 201, 202, 204,
205, 214, 219, 220, 221,
222, 224, 225, 245, 247,
269, 270, 273
Strother T, 217, 275
Stuehr J, 108, 178
Sumant A, 204, 207, 208, 274
Sunseri C, 195, 271
Sussmann RS, 195, 272
Sutherland DGJ, 199, 200, 274
Sutin N, 168, 169, 180
Sutin NJ, 108, 178
Swain GM, 193, 195, 198, 200,
201, 202, 203, 204, 205,
207, 210, 212, 214, 219,
220, 221, 222, 224, 225,
228, 230, 231, 236, 238,
239, 241, 242, 243, 245,
247, 250, 252, 253, 254,
258, 263, 269, 270, 271,
272, 273, 274, 275
Swathirajan S, 65, 88, 99
Tabaka AA, 260, 276
Tabor CW, 230, 275
Tabor N, 230, 275
Tachibana T, 189, 195, 201, 203,
222, 239, 252, 253, 269,
270
Tacke T, 251, 276
Tadjeddine A, 195, 272
Taguchi S, 65, 99
Takahashi K, 195, 269
Takai Y, 189, 222, 269
Takasu Y, 260, 276
Tamiya E, 11, 95
Tang Y, 65, 70, 98
Tanner RI, 33, 97
Tarutani M, 189, 222, 269
Tatsuma T, 6, 64, 95, 227, 275
Templitskaya GL, 195, 271
Tenan MA, 6, 95
Tender L, 163, 179
Tenne R, 195, 271, 272
Terashima C, 195, 272, 273
Terashima CA, 195, 226, 228,
271
Teremetskaya IG, 195, 271, 272
Terminello LJ, 188, 208, 268
Terminello LT, 199, 200, 274
Terrill RH, 163, 179
Terui S, 65, 99
Theodore RB, 48, 97
Thompson M, 5, 11, 14, 22, 26, 71,
94, 95, 96
Thompson PA, 20, 21, 96
Thorston RN, 37, 97
Ticianelli EA, 260, 276
Tokuda K, 11, 34, 95, 97, 107,
177
Tomi TL, 53, 98
Tong W, 199, 200, 274
Topart PA, 6, 11, 95, 96
Totir D, 195, 272
Touhara H, 195, 239, 273
Tour JM, 103, 104, 176
Trava-Airoldi VJ, 212, 274
Treimer SE, 230, 275
Tremiliosi-Filho G, 267, 277
Troian SM, 20, 21, 96
Tryk A, 195, 221, 224, 225, 270,
271
Tryk DA, 195, 205, 214, 226, 270,
271, 272, 273
Author Index 291
Tsionsky V, 4, 5, 6, 7, 19, 23, 25,
27, 34, 35, 36, 37, 40, 41, 42,
43, 44, 45, 46, 47, 48, 58, 59,
63, 64, 66, 72, 73, 74, 75, 76,
78, 79, 80, 81, 82, 87, 88, 94,
95, 96, 97
Tsionsky VJ, 61, 62, 63, 70, 87,
98
Tsunozaki K, 191, 195, 269, 272
Turner DH, 108, 178
Uchida H, 33, 53, 54, 55, 58, 60,
65, 81, 97, 98, 99
Uosaki K, 195, 217, 272, 274
Urbakh M, 5, 6, 7, 23, 25, 27, 28,
29, 30, 31, 33, 43, 44, 48, 63,
64, 72, 73, 74, 75, 94, 95, 96,
97, 99
Ushizawa K, 195, 269
Van Bennekom WP, 87, 99
Van Ryswyk H, 105, 150, 177
Van Veen JAR, 54, 55, 57, 58,
98
Varin VP, 195, 271
Varnin VP, 195, 272
Varun VP, 195, 271
Vaskevich A, 70, 99
Vatankhah G, 86, 99
Vatistas NJ, 239, 275
Velichko GI, 109, 178
Vidakovic TR, 264, 277
Vieil EJ, 65, 99
Vigier F, 58, 59, 98
Vinokur N, 195, 221, 271, 272
Visscher W, 54, 55, 57, 58, 98
Vitt JE, 230, 275
Vogal WM, 260, 276
von Kaenel Y, 189, 269
Waldeck DH, 105, 154, 167,
177
Walhout PK, 104, 176
Wandlowski TH, 49, 98
Wang D, 104, 176
Wang J, 11, 96, 193, 195, 201, 203,
204, 205, 212, 221, 239,
245, 247, 252, 253, 254,
258, 261, 262, 263, 265,
266, 267, 269, 270, 273,
274
Wang JG, 195, 201, 207, 270
Wang R, 195, 271
Ward MD, 2, 11, 19, 89, 94, 95, 96,
99
Wasberg M, 53, 98
Washington E, 260, 276
Wasle S, 6, 95
Watanabe K, 195, 269
Watanabe M, 33, 54, 55, 58, 60,
65, 81, 97
Watanabe MJ, 53, 54, 58, 65, 98,
99
Watts ET, 15, 16, 21, 96
Weaver MJ, 57, 98, 105, 176
Weber K, 105, 163, 177
Weber KA, 154, 167, 179
Weil KG, 6, 11, 26, 71, 94, 96,
99
West AC, 87, 99
Weston RE, 154, 179
White AE, 195, 271
White HS, 19, 96, 107, 112, 123,
177, 178
White WB, 198, 199, 274
Whitehead AJ, 195, 272
Widom A, 15, 16, 96
Wieckowski A, 2, 94, 103, 176,
267, 277
Author Index 292
Wightman RM, 107, 177, 191,
195, 269, 273
Wilde CP, 65, 76, 81, 82, 99
Wilkins SJ, 224, 275
Williams BE, 245, 275
Williams ME, 154, 167, 179
Winkle JR, 103, 176
Winograd N, 242, 275
Wipf DO, 107, 177
Wisitsora-at A, 191, 269
Witek M, 193, 195, 201, 203, 204,
205, 207, 212, 230, 231,
232, 234, 236, 237, 238,
239, 245, 247, 252, 269,
270, 273
Witek MA, 195, 198, 204, 205,
207, 210, 221, 230, 231,
236, 238, 270, 273
Wolden C, 189, 269
Woodin RL, 185, 186, 268
Woodru WH, 108, 178
Woods R, 33, 97
Woods RJ, 54, 98
Wu J, 195, 271
Wunsche M, 6, 95
Xia X, 107, 177
Xikang Z, 195, 271
Xu J, 193, 195, 201, 202, 204, 205,
212, 214, 219, 220, 221,
222, 224, 225, 245, 247,
269, 270
Yagi I, 195, 205, 214, 273
Yagi IT, 195, 225, 271
Yahikozawa K, 260, 276
Yamamoto N, 6, 64, 95
Yamane T, 6, 64, 95
Yang HH, 217, 274
Yang M, 5, 11, 22, 26, 71, 94, 95,
96
Yang MS, 11, 14, 95
Yang W, 219, 275
Yano K, 217, 274
Yano T, 195, 271, 272
Yao Y, 195, 271
Yates J, 217, 274
Ye X, 11, 95
Yeager E, 108, 178
Yi-Ping Liu, 113, 179
Yip CM, 11, 95
Yokota Y, 189, 222, 269
Yokoyama K, 104, 176
Yoneda T, 260, 276
Yoshihara S, 195, 272
Yotsuyanagi T, 34, 97
Yu CJ, 164, 179
Zagidulin D, 6, 7, 27, 63, 64, 72,
73, 74, 75, 95
Zaitsev AM, 242, 248, 275
Zak J, 195, 204, 205, 239, 241, 242,
243, 245, 273
Zak JK, 195, 201, 203, 228, 239,
241, 243, 250, 252, 270,
274, 275
Zapol P, 204, 207, 208, 274
Zawodzinski T, 168, 180
Zecevic SK, 260, 276
Zeliger H, 260, 276
Zerihun T, 108, 178
Zewail AH, 104, 176
Zhang G, 195, 271
Zhang MJ, 65, 76, 99
Zhang X, 154, 167, 179, 195,
271
Zhang XG, 260, 276
Zheng J, 195, 270
Author Index 293
Zhou C, 104, 176
Zhou D, 188, 208, 269
Zhu J, 195, 271
Zhu YX, 19, 96
Zilberman G, 23, 25, 34, 35, 41, 42,
43, 44, 45, 46, 47, 48, 58, 59,
66, 87, 88, 96, 97
Zolfaghari A, 54, 86, 98, 99
Zolotovitskii YM, 109, 132,
178
Zuiker CD, 188, 199, 200, 208,
268, 274
Zusman LD, 105, 154, 167,
177
Author Index 294
Acoustic sensor
indirect laser-induced tempera-
ture-jump (ILIT), 145, 146
Activation energy
distance, 168169
Active electrode
radius, 145
Admittance
vs. frequency
gold, 7
viscosity, 6
Adsorbate/electrolyte interface
slippage, 2326
velocity elds, 23
Adsorption studies, 4360
organic substances, 4353
Aliphatic polyamine detection
electrically conducting diamond
thin lms, 230238
Alkenes
hydrogen-terminated
electrically insulating
diamond surface, 217
Ampliers
ture-jump (ILIT), 134
Angus group, 183
Anodic stripping voltammetry
(ASV), 221
Arrhenius plots, 163165
Ascorbic acid, 213
ASV, 221
Atomic hydrogen
adsorption, 5860
Azide
diamond electrodes, 220
electrically conducting
diamond thin lms,
219221
glassy carbon electrodes, 220
Boron-doped diamond discs
optical images, 240
Boron-doped diamond electrodes
electrical conductivity, 194
Boron-doped diamond thin lm
AFM image, 257
deposition, 191
electrodes
redox systems, 204205
Boron-doped diamond thin
molybdenum metal mesh
SEM images, 192
Boron-doped diamond thin
platinum wire
microelectrode
SEM images, 192
Boron-doped microcrystalline
diamond electrodes
cyclic voltammetric, 213
Boron-doped microcrystalline
diamond thin lms
cyclic voltammetric constant
data, 206
295
SUBJECT INDEX
[Boron-doped microcrystalline
diamond thin lms]
cyclic voltammetric i-E curves,
202, 206, 209
SEM image, 187
visible Raman spectra, 198200
x-ray diraction patterns, 197
Boron-doped nanocrystalline
diamond thin electrodes,
208209
cyclic voltammetric data
aqueous-based redox
systems, 210
237
pulse anodic-stripping
voltammetric curves, 223
Boron-doped nanocrystalline
diamond thin lms
carrier concentrations and
mobilities, 196
SEM image, 187
x-ray diraction patterns, 197
Brinkmans equation, 30, 31
Brookhaven National
Laboratory
Laser Electron Accelerator
Facility (LEAF), 109, 152
Butanethiol
adsorption, 51
calculated equivalent weights,
51
Butanol
in water
frequency shift vs. surface
charge density, 45
Butler-Volmer expression, 141,
163
vs. MNA electron transfer,
118120
Butterworth-van Dyke equivalent
circuit, 12, 14
Cadaverine
detection, 236
Carbamate pesticide detection
thin lms, 228229
Carbide-forming materials
diamond growth, 188189
Carborundum, 183
Cell
indirect laser-induced
temperature-jump (ILIT),
143148
schematic, 144
Charged species
electrostatic adsorption
thin-lm model, 3738
Charge-injection method, 107
Chemical vapor deposition
(CVD), 183
microwave plasma
diamond lm morphology,
215
reactor, 185
diagram, 186
Chidseys approach
electron transfer theory,
112113
Chlorpromazine, 245, 247
Chromium underlayer
removal, 8788
Coecient of sliding friction,
15
Composite electrode fabrication
and characterization
electrically conducting dia-
mond thin lms, 252259
Contact angle, 22
Subject Index 296
Copper
copper electrodeposition,
6163
Correlation function, 2829
Coulostatic technique, 106107
Counter electrodes
143
Cr underlayer
removal, 8788
Current distribution, 8687
CVD. See Chemical vapor
deposition (CVD)
Cyclic voltammetry
electrochemical crystal micro-
balance (EQCM), 8990
gold, 55
gold electrodes, 79
Cytochrome c, 228
249
Delayed heat transport
monolayer lm, 155
Density-of-state assumptions,
113
Deposited gold layer
radius, 145
Derjaguin group, 183
Diamond
catechols
molecular adsorption studies,
211
electrical conductivities, 184
electron emitter, 183184
metal deposition, 221
polar molecule adsorption
aqueous media, 210
polyamine oxidation, 233
[Diamond]
substrate pretreatment,
189190
technologically important
properties, 184
working window, 202203
Diamond discs
boron-doped
optical images, 240
Diamond electrodes
azide, 220
boron-doped
electrical conductivity, 194
carrier concentration, 194
dimensional stability, 203
electrical conductivity,
194195
electron transfer, 212216
extreme conditions, 203
surface modication, 216219
Diamond lm
electrically insulating to
material, 207
electrical properties, 245
Raman spectra, 216
Diamond Makers, 182
Diamond materials
surface modication, 216219
Diamond/platinum composite
electrode
265
linear sweep voltammetric i-E
curves, 261
methanol oxidation, 267
open circuit potentials, 266
slow scan linear sweep
voltammetric i-E curves,
262
Subject Index 297
Diamond powder
AFM image, 190
Diamond/quartz optically
transparent electrode
optical image, 241
voltammetric i-E curve, 246
Diamond synthesis, 182183
Diamond thin lm
boron-doped. See Boron-doped
diamond thin lm
electrically conducting. See
Electrically conducting
diamond thin lms
XRD, 200201
Diamond thin lm electrodes
aliphatic amine oxidation
cyclic voltammetric data, 232
Diamond thin molybdenum metal
mesh
boron-doped
SEM images, 192
Diamond thin platinum wire
microelectrode
boron-doped
SEM images, 192
Diuse double layer
properties, 44
viscosity
thin-lm model, 3839
Diuser
Direct laser-induced
temperature-jump method
(DLIT), 109111
vs. indirect laser-induced
111112
problems, 109
Distance
activation energy, 168169
DLIT, 109111
vs. indirect laser-induced
111112
problems, 109
Double layer region
electrochemical crystal
microbalance (EQCM),
3435
frequency shift
gold, 35
roughness, 82
Elastic displacement
wave equation, 26
Electrical double layer/
electrostatic adsorption,
3343
frequency potential
dependence, 3643
results, 3436
Electrical equivalent circuit, 11
Electrically conducting diamond
thin lms, 181268
advanced electrocatalyst
support materials,
251267
aliphatic polyamine detection,
230238
azide detection, 219221
carbamate pesticide detection,
228229
composite electrode fabrication
and characterization,
252259
corrosion rate, 251252
deposition, 185194
Subject Index 298
[Electrically conducting diamond
thin lms]
electroanalytical applications,
219238
electrochemical cells, 185194
electrode architectures, 185194
electrosynthesis and electrolytic
water purication,
238239
ferrocene analysis, 229230
histamine detection, 226227
methanol oxidation reaction
(MOR), 264267
nicotinamide adenine
dinucleotide (NADH),
225
nitrite detection, 224
optically transparent electrodes
spectroelectrochemistry,
239251
oxygen reduction reaction,
259264
serotonin detection, 226227
substrate materials, 185194
trace metal ion, 221224
uric acid detection, 225226
Electrochemical case, 7682
Electrochemical cells
diamond thin lms,
185194
Electrochemical crystal
microbalance (EQCM)
adsorption studies, 4360
equation, 48
metal, 44
cyclic voltammetry, 8990
double layer region, 3435
electrochemical case, 7682
[Electrochemical crystal
microbalance (EQCM)]
experimental application,
8687
frequency shift
gold on gold electro-
deposition, 62
gold on gold substrate, 64
hydrogen adsorption
platinum, 5860
impedance spectrum, 58
inorganic species
adsorption, 5360
metal deposition, 6070
solution height and bubbles,
8889
surface roughness, 84
sweep rate, 70
thiols
adsorption, 4953
uracil
adsorption, 4953
Electrochemical deposition
Sauerbrey equation, 6061
Electrochemical impedance
spectroscopy, 11
Electrode(s)
active
radius, 145
compartment
frequency, 88
roughness, 81
working ILIT, 143
conguration, 145
preparation, 146147
thermal diusion properties,
146150
Electrode/redox center
linkage, 105
Subject Index 299
Electrolytic water purication
thin lms, 238239
Electron emitter
diamond, 183184
Electronics
151152
Electron transfer
diamond electrodes, 212216
Electron transfer theory, 112120
Chidseys approach, 112113
temperature dependence,
116118
Electroplating
mass-transport limitation
surface roughing, 63
Electrostatic adsorption, 3343
charged species
thin-lm model, 3738
frequency potential
dependence, 3643
results, 3436
Electrosynthesis and electrolytic
water purication
diamond thin lms,
238239
EQCM. See Electrochemical
crystal microbalance
(EQCM)
Ethylene diamine
detection, 236
Eversole, William, 183
Experimental protocols
160
Faradaic eciency, 60
Fast amplier
151
Fast interfacial electron transfer
101175
measurement, 103104
Ferri-ferrocyanide thermal
response
calibration, 147
Ferrocene
diamond thin lms,
229230
169
FIA, 220, 236
Film
viscosity
comparison, 41
potential function, 3940
Flow injection analysis (FIA),
220, 236
Frequency
vs. admittance
gold, 7
pyridine/water/gold system, 47
Frequency potential dependence
electrical double layer/electro-
static adsorption, 3643
Frequency shift, 24
gold, 55
gold surfaces, 80
pyridine/water/silver system, 47
Subject Index 300
[Frequency shift]
vs. surface charge density
gold electrodes, 42
silver electrodes, 42
Gas-phase adsorption
quartz crystal microbalance
(QCM), 45
Gaussian random roughness, 29
General Electric, 182, 183
Glassy carbon
catechols
molecular adsorption studies,
211
202
electrodes
azide, 220
Gold
admittance vs. frequency, 7
cycling voltammogram, 55
frequency shift, 55
deposition, 69
gold electrodeposition, 6163
hydrogen evolution
overvoltage, 68
parameter values, 110
Gold electrodes
cycling voltammetry, 79
frequency shift vs. potential, 59
charge density, 42, 50
oxide formation, 5455
Gold layer
deposited
radius, 145
Gold/pyridine/water system
frequency, 47
Gold surfaces
frequency shift, 80
resonance width, 80
Gouy-Chapman theory, 37
Hazen, Robert M., 182
Heat transport
delayed
monolayer lm, 155
Height-height pair correlation
function, 27
Heme peptide
direct electron transfer, 227228
Heterogeneous nonadiabatic
electron transfer reactions
Marcuss description, 112120
Histamine detection
thin lms, 226227
Homogeneous kinetics
temperature-jump approach,
108
Hydrogen adsorption
platinum, 5860
surface oxide formation, 5458
Hydrogen-poor AR gas mixture
nanocrystalline lms, 188
Hydrogen-terminated electrically
insulating diamond surface
alkenes, 217
ILIT. See Indirect laser-induced
temperature-jump (ILIT)
Impedance
microbalance (EQCM),
58
(QCM), 812
Subject Index 301
Indirect laser-induced
temperature-jump (ILIT)
acoustic sensor, 145, 146
analysis, 120143
cell, 143148
schematic, 144
counter electrodes, 143
data acquisition, 151152
diuser, 148
vs. direct laser-induced
temperature-jump
method (DLIT), 111112
eective interfacial
temperature, 135
electronics, 151152
energetic and timing issues,
156160
equivalent circuit, 138
evolution, 108112
experimental implementation,
143160
experimental protocols, 160
fast amplier, 151
fast electron transfer, 126127
fast interfacial electron transfer,
101175
ferrocene, 169
future experiments, 166170
ideal interfacial temperature
change, 134
ideal perturbation shape,
130134
one-dimensional thermal
diusion, 173175
open-circuit potential change,
136
perturbation shape and
response
nonidealities, 134137
[Indirect laser-induced
temperature-jump (ILIT)]
potential problems, 152156,
166
pressure eect, 153154
pseudo-reference electrodes,
143
reference electrodes, 143
relaxation, 126
RF noise, 152
slow electron transfer, 127
Soret eect, 153
special potentiostat, 151
temperature eect, 152153
terminology, 170173
thermal, 126130
typical transients, 161162
viscosity eect, 154155
working electrodes, 143
conguration, 145
146150
Indium tin oxide (ITO), 242
transmission spectra, 244
Inorganic species
adsorption, 5360
Interfacial friction coecient,
15
Interfacial kinetics
measurement examples,
161165
temperature-jump approach,
108112
Interfacial layer
liquid ow, 31
Ions
adsorption, 5354
electrostatic adsorption
viscosity, 84
Subject Index 302
ITO, 242
transmission spectra, 244
Kozeny-Carman equation, 3031
Laser Electron Accelerator
Facility (LEAF)
Brookhaven National
Laboratory, 109, 152
Laser perturbation
134
Laser-pulse energy requirements
single shot experiments,
156157
LEAF
Brookhaven National
Laboratory, 109, 152
Liquid(s)
quartz crystal
rough surfaces, 2632
(QCM), 5
roughness, 7082
Liquid/adsorbate interface
velocity distribution, 18
Liquid-induced frequency shift,
31
Liquid/solid interface
slippage, 2023
Man-Made Diamonds, 183
Marcus nonadiabatic (MNA)
behavior, 112113, 163
electron transfer
vs. Butler-Volmer expression,
118120
formalism, 142
Marcuss description
heterogeneous nonadiabatic
electron transfer reactions,
112120
Mass loading, 78
Mechanical impedance, 71
Mechanical resonance, 15
Metal deposition
balance (EQCM), 6070
foreign substrate
early stages, 6470
Metal lm
(QCM), 4
Metal substrate
same
deposition, 6064
Methanol oxidation
platinum/diamond composite
electrode, 267
Methanol oxidation reaction
(MOR)
thin lms, 264267
Methyl terminated aliphatic thiols
self-assembled monolayer
(SAM)
129
Microbalance
(QCM), 3
Microcrystalline diamond
electrodes
boron-doped, 213
Subject Index 303
Microcrystalline diamond thin
lm
microcrystalline diamond
thin lms
vs. nanocrystalline diamond
thin lm, 210
silver deposits, 221
Microcrystalline diamond thin
lm electrodes
characterization, 195201
231
electrochemical properties,
201211
Microcrystalline lms, 186
Microelectrode array structures,
191
Micropyramids, 191
diamond electrode array
SEM image, 193
Microwave plasma, 185, 186
chemical vapor deposition
(CVD)
diamond lm morphology,
215
MNA behavior, 112113,
163
MNA electron transfer
vs. Butler-Volmer expression,
118120
MNA formalism, 142
Monolayer lm
delayed heat transport, 155
MOR
diamond thin lms,
264267
Motional branch, 11
Multiple pulse experiments
157160
temporal spacing, 158159
tolerable maximum number,
159160
Multiscale roughness
schematic representation, 75
NADH
diamond thin lms, 225
Nanocrystalline diamond thin
electrodes
nanocrystalline diamond
thin electrodes
lm
nanocrystalline diamond
thin lms
cyclic voltammetric, 236
vs. microcrystalline diamond
thin lms, 210
nitrogen, 207208
lm electrodes
characterization, 195201
electrochemical properties,
201211
Nanocrystalline lms, 186
hydrogen-poor AR gas
mixture, 188
noble gas, 188
Nano-tribology, 85
Navier-Stokes equation, 17, 23,
26, 31
Subject Index 304
N-butanol
in water
charge density, 45
Nernstian relationship, 124, 139
Nicotinamide adenine
dinucleotide (NADH)
Nitrite detection
Nitrogen
nanocrystalline diamond thin
lm, 207208
Noble gas
nanocrystalline lms, 188
Nonelectrochemical case, 7176
Nonslip boundary condition
quartz crystal
liquid contact, 1720
Nonuniform lm
rigidly attached to surface,
1314
surface, 86
Norton Company, 182, 183
Octanethiol
adsorption, 51
calculated equivalent weights,
51
One-dimensional thermal
diusion
opd, 66
Open-circuit indirect laser-
induced temperature-jump
(ILIT) response, 121125
Open-circuit potential change
136
Optically transparent electrode
(OTE), 184
diamond/quartz
optical image, 241
spectroelectrochemistry
diamond thin lms,
239251
Organic substances
adsorption studies, 4353
OTE. See Optically transparent
electrode (OTE)
Over potential deposition (opd),
66
Oxide formation
gold electrodes, 5455
platinum, 5455
Oxygen reduction reaction
diamond thin lms,
259264
Peruoroalkyl radicals, 217
Permeability length scale, 32
Peroxidase
direct electron transfer, 227228
Perturbation theory, 30
Platinum
oxidation-reduction cycles
frequency response, 77
oxide formation, 5455
Platinum-coated diamond thin
lms
AFM image, 260
Subject Index 305
Platinum composite electrodes
AFM image, 256
cyclic voltammetric i-E curve,
255
fabrication, 253
Platinum diamond composite
secondary diamond growth,
258
Platinum/diamond composite
electrode
265
linear sweep voltammetric i-E
curves, 261
methanol oxidation, 267
open circuit potentials, 266
slow scan linear sweep
voltammetric i-E curves,
262
Platinum electrodes
frequency shift vs. potential, 59
hydrogen adsorption, 5860
Pressure eect
153154
Prole function, 27
Project Superpressure, 183
Pseudo-reference electrodes
143
Pump-probe method, 104
Putrescine
detection, 236
Pyridine
adsorption
frequency shift, 46
in butanol
[Pyridine]
charge density, 45
in water
charge density, 45
Pyridine/water/gold system
frequency, 47
Pyridine/water/silver system
frequency shift, 47
QCM. See Quartz crystal
microbalance (QCM)
QCS, 3
Quartz
Quartz crystal
liquid contact
nonslip boundary condition,
1626
slip boundary conditions,
2026
thin liquid lm, 1920
rough surfaces, 2633
liquids, 2632
shear-mode oscillations, 86
Quartz crystal microbalance
(QCM), 2
application, 45
gas-phase adsorption, 45
impedance, 812
liquids, 5
microbalance, 3
roughness, 3233
induced response, 26
liquids, 7082
slippage, 3233
theoretical interpretation, 833
thin surface lms, 1216
Subject Index 306
Quartz crystal resonator
liquid contact
schematic sketch, 9
Quartz crystal sensor (QCS), 3
Quartz disk
radius, 145
Quartz optically transparent
electrode
optical image, 241
Raman spectra
diamond lm, 216
visible
boron-doped micro-
crystalline diamond thin
lms, 198200
Redox center
linkage, 105
Redox systems
boron-doped diamond
thin lm electrodes,
204205
Reference electrodes
143
Resonance frequency, 14, 18
shift, 19, 28
viscosity, 6
Resonance width, 18, 81
gold surfaces, 80
shift, 28, 31
slip boundary conditions, 21
Resonant frequency shift
slip boundary conditions, 21
Reversed-phase liquid
chromatography
amperometric detection, 236
RF noise
Roughness
liquids, 7082
(QCM), 3233
Rough surfaces
frequency shift, 78
quartz crystal, 2633
SAM. See Self-assembled
monolayer (SAM)
Saturated sodium calomel
electrode (SSCE), 143
Sauerbrey equation for
electrochemical
deposition, 6061
Scanning electrochemical
microscopy (SECM), 167
Self-assembled monolayer
(SAM), 105
electron-transfer kinetics
noble metals, 168
nonaqueous solvent systems,
167
tethered redox species, 167
128129
preparation, 150151
response times
redox moiety absence,
166167
SEM images
boron-doped diamond thin
platinum wire
microelectrode, 192
Subject Index 307
Serotonin detection
diamond thin lms,
226227
Shear mechanical impedance, 26
Shear-mode oscillations
quartz crystal, 86
Shear mode resonator, 8
Silver
gold
deposition, 67
silver electrodeposition, 6163
Silver deposits
microcrystalline diamond thin
lm, 221
Silver electrodes
charge density, 42
Silver/pyridine/water system
frequency shift, 47
Single-crystal electrodes, 85
Single shot experiments
laser-pulse energy require-
ments, 156157
Sliding friction coecient, 2122
Slight roughness, 27
Slip boundary conditions
quartz crystal
resonance width, 21
resonant frequency shift, 21
Slip length, 22
Slippage
(QCM), 3233
Solid/liquid interface
slippage, 2023
Soret bands, 250
Soret eect
153
Soret potential, 147
Special potentiostat
151
Spectroelectrochemistry, 242
optically transparent electrodes
diamond thin lms,
239251
Spermidine
detection, 236
Spermine
detection, 236
SSCE, 143
Strong roughness, 27, 3031
Substrate pretreatment
diamond, 189190
Supporting gas method, 4
Surface charge density
vs. frequency shift
gold electrodes, 42
silver electrodes, 42
Surface oxide formation
hydrogen adsorption, 5458
Surface roughness, 71
balance (EQCM), 84
Surface tension
thin-lm model, 3637
Sweep rate, 89
microbalance (EQCM), 70
Symbols, 9194
Subject Index 308
Tektronix DSA602A, 135, 151
152
Temperature dependence
electron transfer theory,
116118
Temperature eect
152153
Terminology
170173
Tert-butanol
in water
charge density, 45
Thallium
deposition, 66
Thin-lm model, 3643
charged species
3738
diuse double layer
viscosity, 3839
surface tension, 3637
Thin lm/solid interface
slippage, 1516
Thin-layer technology
(QCM), 4
Thin liquid lm
quartz crystal
Thin surface lms
(QCM), 1216
Thiols
adsorption, 4953, 52
Titanium
Titanium-sapphire laser
fast amplier, 136
Titanium underlayer
removal, 8788
Trace metal ion
diamond thin lms,
221224
Typical transients
161162
Under potential deposition (upd),
6466, 68
Uniform lm
rigidly attached to surface,
1213
upd, 6466, 68
Uracil
adsorption, 4953
Uric acid detection
diamond thin lms,
225226
Velocity decay length, 73, 74
Velocity distribution
liquid/adsorbate interface,
18
Velocity elds
adsorbate/electrolyte interface,
23
Viscosity
admittance spectra, 6
diuse double layer
thin-lm model, 3839
Subject Index 309
[Viscosity]
ions
84
resonance frequency, 6
Viscosity eect
154155
Visible Raman spectra
boron-doped microcrystalline
diamond thin lms,
198200
Water
Water purication
electrolytic
[Water purication]
diamond thin lms,
238239
Water/pyridine/gold system
frequency, 47
Water/pyridine/silver system
frequency shift, 47
Wave equation
elastic displacement, 26
Working electrodes
143
conguration, 145
preparation, 146147
146150
Subject Index 310

Electroanalytical Chemistry A Series of Advances Volume 22 Electroanalytical Chemistry

Cargado por

Información del documento

Descripción original:

Título original

Derechos de autor

Formatos disponibles

Compartir este documento

Compartir o incrustar documentos

Opciones para compartir

¿Le pareció útil este documento?

¿Este contenido es inapropiado?

Copyright:

Formatos disponibles

Electroanalytical Chemistry A Series of Advances Volume 22 Electroanalytical Chemistry

Cargado por

Copyright:

Formatos disponibles

ELECTROANALYTICAL

can be a complex number,

contains the same information

d RexpiKR < nRVnRVR > 40

ion replaces one water molecule, we can estimate the contribu-

, reaching zero in the case of F

, in the supporting electro-

varies between 1.5 and 60), and

: Thus, for the homogeneous

. Analysis of the EQCMresponse on

k wave vector of shear waves in quartz, x

l correlation length of surface roughness and of non-

positive and negative excesses of charge

electrochemical roughness factor

Dc change of surface tension

< 0.001 cm, negligibly small

. If we were to estimate these transit times based

A peak-shaped oxidation response was observed in cyclic voltammetric

Ion chromatography (IC) (pH 7.2) with amperometric detection was

A subsequent one-electron transfer occurs at more positive potentials,

The corrosion rate is strongly dependent on the electrode potential,

has been reported to strongly adsorb on the Pt surface in the

Calculated turnover numbers are listed in Table 9. The turnover number

También podría gustarte