MPB0000026 PDF

Chemical Sensors
Korotcenkov
Edited by Ghenadii Korotcenkov, Ph.D., Dr. Sci.
Momentum Press is proud to bring you Chemical Sensors Volume 1, the newest addition to The Sensors
Technology Series, edited by Joe Watson. In this long-awaited first volume, Fundamentals of Sensing
Material: General Approaches, you will find a comprehensive introduction to all key materials used for
chemical sensors, beginning with the essential foundations like desired properties, as well as methods of
synthesis, methods of deposition, and methods for modification of sensing materials properties. Inside,
you will find background and guidance on:
CHEMICAL SENSORS
• Basic principles of operation for all major classes of sensors, including electrochemical sensors,
Fundamentals of Sensing Material: General Approaches

capacitance sensors, field-effect transistor sensors, Schottky Diode-based sensors, and Langmuir-
Blodgett Film sensors
• How to attain desired properties for sensors like surface properties, stability properties, electrophysical
properties, and structural properties
• Synthesis methods for sensing materials, including vacuum evaporation and vacuum deposition,
sputtering, chemical vapor deposition , the Sol-Gel process, powder technologies and polymer
technologies
• How to improve the operating characteristics of sensors through metal oxides engineering
Chemical sensors are integral to the automation of a myriad industrial processes, as well as everyday
monitoring of such activities as public safety, testing and monitoring, medical therapeutics, and many
more. This massive reference work will cover all major categories of both the materials used for chemi-
cal sensors and their applications. This will become THE reference work on sensors used for chemical
detection and analysis.
The Chemical Sensors references books will span six volumes and cover in-depth details on both mate-
rials used for chemical sensors and their applications, with the first three volumes exploring the materials
used for chemical sensors — their properties, their behavior, their composition, and even their manufac-
turing and fabrication. Volumes 4 through 6 will explore the great variety of applications for chemical
sensors — from manufacturing and industry to biomedical uses.
About the Editor
Ghenadii Korotcenkov received his Ph.D. in Physics and Technology of Semiconductor Materials and
Devices in 1976 and his Habilitate Degree (Dr. Sci.) in Physics and Mathematics of Semiconductors and
Dielectrics in 1990. He was for many years the leader in the Gas Sensor Group at the Technical Univer-
sity of Moldova. He is currently a research professor at Gwangju Institute of Science and Technology, in
Gwangju, Republic of Korea. Dr. Korotcenkov is the author of five previous books and has authored over
180 peer-reviewed papers. His research has received numerous awards and honors, including the Award
volume i
of the Supreme Council of Science and Advanced Technology of the Republic of Moldova.
A volume in the Sensor Technology Series

Edited by Joe Watson
Published by Momentum Press®
ISBN: 978-1-60650-103-0
90000
www.momentumpress.net 9 781606 501030

CHEMICAL SENSORS
FUNDAMENTALS OF
SENSING MATERIALS
VOLUME 1:
GENERAL APPROACHES
CHEMICAL SENSORS
FUNDAMENTALS OF
SENSING MATERIALS
VOLUME 1:
GENERAL APPROACHES
EDITED BY
GHENADII KOROTCENKOV
GWANGJU INSTITUTE OF SCIENCE AND TECHNOLOGY

GWANGJU, REPUBLIC OF KOREA
MOMENTUM PRESS, LLC, NEW YORK

Chemical Sensors: Fundamentals of Sensing Materials. Volume 1: General Approaches
Copyright © Momentum Press®, LLC, 2010
All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or
transmitted in any form or by any means—electronic, mechanical, photocopy, recording or any other—
except for brief quotations, not to exceed 400 words, without the prior permission of the publisher.
First published in 2010 by

Momentum Press®, LLC
222 East 46th Street, New York, NY 10017
www.momentumpress.net
ISBN-13: 978-1-60650-103-0 (hard back, case bound)

ISBN-10: 1-60650-103-8 (hard back, case bound)
ISBN-13: 978-1-60650-105-4 (e-book)
ISBN-10: 1-60650-105-4 (e-book)
DOI forthcoming
Cover design by Jonathan Pennell

Interior design by Derryfield Publishing, LLC
First Edition: September 2010
10 9 8 7 6 5 4 3 2 1
Printed in Tawian, ROC

CONTENTS
PREFACE TO CHEMICAL SENSORS: FUNDAMENTALS OF SENSING MATERIALS xi

PREFACE TO VOLUME 1: GENERAL APPROACHES xiii
ABOUT THE EDITOR xv
CONTRIBUTORS xvii
1 BASIC PRINCIPLES OF CHEMICAL SENSOR OPERATION 1
M. Z. Atashbar
S. Krishnamurthy
G. Korotcenkov
1 Introduction 1
2 Electrochemical Sensors 1
2.1 Amperometric Sensors 3
2.2 Conductometric Sensors 4
2.3 Potentiometric Sensors 5
3 Capacitance Sensors 8
4 Work-Function Sensors 9
5 Field-Effect Transistor Sensors 12
6 chemFET-Based Sensors 15
7 Schottky Diode–Based Sensors 16
8 Catalytic Sensors 19
9 Conductometric Sensors 20
10 Acoustic Wave Sensors 24
10.1 Thickness Shear Mode Sensors 25
10.2 Surface Acoustic Wave Sensors 27
v
vi • CONTENTS
11 Mass-Sensitive Sensors 31
12 Optical Sensors 33
12.1 Fiber Optic Chemical Sensors 36
12.2 Fluorescence Fiber Optic Chemical Sensors 38
12.3 Absorption Fiber Optic Chemical Sensors 39
12.4 Refractometric Fiber Optic Chemical Sensors 39
12.5 Absorption-Based Sensors 39
12.6 Surface Plasmon Resonance Sensors 43
13 Photoacoustic Sensors 45
14 Thermoelectric Sensors 45
15 Thermal Conductivity Sensors 47
16 Flame Ionization Sensors 49
17 Langmuir-Blodgett Film Sensors 50
References 52
2 DESIRED PROPERTIES FOR SENSING MATERIALS 63

G. Korotcenkov
1 Introduction 63
2 Common Characteristics of Metal Oxides 65
2.1 Crystal Structure of Metal Oxides 65
2.2 Electronic Structure of Metal Oxides 68
2.3 Role of the Electronic Structure of Metal Oxides in Surface Processes 70
3 Surface Properties of Sensing Materials 77
3.1 Electronic Properties of Metal Oxide Surfaces 77
3.2 Role of Adsorption/Desorption Parameters in Gas-Sensing
Effects 79
3.3 Catalytic Activity of Sensing Materials 84
4 Stability of Parameters in Sensing Materials 88
4.1 Thermodynamic Stability 88
4.2 Chemical Stability 92
4.3 Long-Term Stability 94
5 Electrophysical Properties of Sensing Materials 97
5.1 Oxygen Diffusion in Metal Oxides 97
5.2 Conductivity Type 101
5.3 Band Gap 106
CONTENTS • vii
5.4 Electroconductivity 107

5.5 Other Important Parameters for Sensing Materials 110
6 Structural Properties of Sensing Materials 112
6.1 Grain Size 113
6.2 Crystal Shape 120
6.3 Surface Geometry 123
6.4 Film Texture 127
6.5 Surface Stoichiometry (Disordering) 129
6.6 Porosity and Active Surface Area 131
6.7 Agglomeration 140
7 Outlook 144
8 Acknowledgments 144
References 145
3 COMBINATORIAL CONCEPTS FOR DEVELOPMENT OF SENSING MATERIALS 159

R. A. Potyrailo
1 Introduction 159
2 General Principles of Combinatorial Materials Screening 160
3 Opportunities for Sensing Materials 162
4 Designs of Combinatorial Libraries of Sensing Materials 163
5 Discovery and Optimization of Sensing Materials Using Discrete Arrays 165
5.1 Radiant Energy Transduction Sensors 165
5.2 Mechanical Energy Transduction Sensors 177
5.3 Electrical Energy Transduction Sensors 183
6 Optimization of Sensing Materials Using Gradient Arrays 192
6.1 Variable Concentration of Reagents 192
6.2 Variable Thickness of Sensing Films 193
6.3 Variable 2-D Composition 194
6.4 Variable Operation Temperature and Diffusion-Layer Thickness 196
7 Emerging Wireless Technologies for Combinatorial Screening of
Sensing Materials 196
8 Summary and Outlook 202
References 203
viii • CONTENTS
4 SYNTHESIS AND DEPOSITION OF SENSOR MATERIALS 215

G. Korotcenkov
B. K. Cho
1 Deposition Technology: Introduction and Overview 215

2 Vacuum Evaporation and Vacuum Deposition 216
2.1 Principles of Film Deposition by the Vacuum Evaporation Method 216
2.2 Disadvantages of the Vacuum Evaporation Method 216
2.3 Film Deposition by Thermal Evaporation 218
3 Sputtering Technology 219
3.1 Principles of Deposition by Sputtering 219
3.2 Sputtering Techniques 221
3.3 Advantages and Disadvantages of Sputtering Technology 224
3.4 Properties of Films Deposited by Sputtering 225
4 The RGTO Technique 227
4.1 Particulars of the RGTO Method 227
4.2 Advantages and Disadvantages of RGTO 229
5 Laser Ablation or Pulsed Laser Deposition 229
5.1 Principles of Pulsed Laser Deposition 229
5.2 Advantages and Disadvantages of PLD 231
5.3 Technical Approaches to Improving PLD Results 231
5.4 Some Particulars of Film Deposition by the PLD Method 232
6 Ion-Beam–Assisted Deposition (IBAD) 233
6.1 Introduction 233
6.2 Principles of the IBAD Method 233
7 Chemical Vapor Deposition 235
7.2 Principles of the CVD Process 235
7.3 Chemical Precursors and Reaction Chemistry 237
7.4 Particulars of CVD Technology 237
7.5 Advantages and Disadvantages of CVD 239
7.6 Variants of CVD Methods 241
8 Deposition from Aerosol Phase 245
8.2 Principles and Mechanism of the Deposition Process 246
8.3 Atomization Techniques 248
8.4 Advantages and Disadvantages of Deposition from an Aerosol Phase 252
CONTENTS • ix
8.5 Technology of the Pyrolysis Process 253

8.6 Regularities of Metal Oxide Growth During Spray Pyrolysis
Deposition 255
9 Deposition from Aqueous Solutions 259

9.2 Chemical Bath Deposition (CBD) 259
9.3 Selective Ion-Layer Adsorption and Reaction (SILAR) or Successive
Ionic-Layer Deposition (SILD) 262
9.4 Liquid-Phase Deposition (LPD) 264
9.5 Electroless Deposition (ED) 265
9.6 Electrochemical Deposition (ECD) 267
9.7 Ferrite Plating 269
9.8 Liquid Flow Deposition (LFD) 269
9.9 Electrophoretic Deposition (EPD) 270
9.10 Photochemical Deposition (PCD). Applying External Forces
or Fields 271
9.11 Summary 272
10 The Sol-Gel Process 272

10.2 Principles of the Sol-Gel Process 273
10.3 Sol-Gel Chemistry and Technology 273
10.4 Advantages and Disadvantages of Sol-Gel Techniques 275
10.5 Calcination of Sol-Gel–Obtained Oxides 277
10.6 Organic–Inorganic Hybrid Materials (OIHM) 279
10.7 Summary 280
11 Powder Technology 280

11.2 Gas Processing Condensation (GPC) 280
11.3 Chemical Vapor Condensation (CVC) 282
11.4 Microwave Plasma Processing (MPP) 283
11.5 Combustion Flame Synthesis (CFS) 283
11.6 Nanopowder Collection 285
11.7 Mechanical Milling of Powders 285
11.8 Summary 287
12 Polymer Technology 288

12.2 Methods of Polymer Synthesis 288
12.3 Fabrication of Polymer Films 292
x • CONTENTS
13 Deposition on Fibers 294

13.1 Specifics of Film Deposition on Fibers 294
13.2 Coating Design and Tooling 295
14 Outlook 298
References 301
5 MODIFICATION OF SENSING MATERIALS: METAL OXIDE MATERIALS ENGINEERING 313
G. Korotcenkov
1 Introduction 313
2 Control of Sensor Response Through Structural Engineering of Metal Oxides 314
2.1 Structural Engineering—What Does It Mean? 314
2.2 Structural Engineering of Metal Oxides—Technical Approaches 321
3 Sensor Response Control Through Modification of Metal Oxide Composition 331
3.1 Phase Modification of Metal Oxides 331
3.2 Methods of Phase Modification 332
3.3 Influence of Additives on Structural Properties of Multicomponent
Metal Oxides 333
3.4 Gas-Sensing Properties of Multicomponent Metal Oxides 337
4 Sensor Response Control Through Surface Modification of Metal Oxides 344
4.1 Methods of Surface Modification 345
4.2 Influence of Surface Modification on Gas-Sensing Properties of
Metal Oxides 348
4.3 Surface Additives as Active and Passive Filters 355
5 Improved Operating Characteristics of Gas Sensors Through Materials
Engineering of Metal Oxides: What Determines the Choice? 357
5.1 Device Application 358
5.2 The Nature of the Gas to Be Detected 359
5.3 Detection Mechanism 361
5.4 Environmental Conditions During Use 361
5.5 Required Rate of Sensor Response 365
5.6 Required Sensitivity 366
5.7 Sensor Response Selectivity 367
5.8 Compatibility with Peripheral Measuring Devices 368
6 Summary 368
References 370
INDEX 379
PREFACE TO
CHEMICAL SENSORS:
FUNDAMENTALS OF SENSING MATERIALS
Sensing materials play a key role in the successful implementation of chemical and biological sen-
sors. The multidimensional nature of the interactions between function and composition, preparation
method, and end-use conditions of sensing materials often makes their rational design for real-world
applications very challenging.
The world of sensing materials is very broad. Practically all well-known materials could be used for
the elaboration of chemical sensors. Therefore, in this series we have tried to include the widest pos-
sible number of materials for these purposes and to evaluate their real advantages and shortcomings.
Our main idea was to create a really useful “encyclopedia” or handbook of chemical sensing materials,
which could combine in compact editions the basic principles of chemical sensing, the main properties
of sensing materials, the particulars of their synthesis and deposition, and their present or potential ap-
plications in chemical sensors. Thus, most of the materials used in chemical sensors are considered in
the various chapters of these volumes.
It is necessary to note that, notwithstanding the wide interest and use of chemical sensors, at the
time the idea to develop these volumes was conceived, there was no recent comprehensive review or
any general summing up of the fundamentals of sensing materials The majority of books published
in the field of chemical sensors were dedicated mainly to analysis of particular types of devices. This
three-volume review series is therefore timely.
This series, Chemical Sensors: Fundamentals of Sensing Materials, offers the most recent advances in
all key aspects of development and applications of various materials for design of chemical sensors.
Regarding the division of this series into three parts, our choice was to devote the first volume to the
fundamentals of chemical sensing materials and processes and to devote the second and third volumes
to properties and applications of individual types of sensing materials. This explains why, in Volume 1:
General Approaches, we provide a brief description of chemical sensors, and then detailed discussion of
desired properties for sensing materials, followed by chapters devoted to methods of synthesis, deposi-
tion, and modification of sensing materials. The first volume also provides general background infor-
mation about processes that participate in chemical sensing. Thus the aim of this volume, although not
xi
xii • PREFACE TO CHEMICAL SENSORS: FUNDAMENTALS OF SENSING MATERIALS
exhaustive, is to provide basic knowledge about sensing materials, technologies used for their prepara-
tion, and then a general overview of their application in the development chemical sensors.
Considering the importance of nanostructured materials for further development of chemical
sensors, we have selected and collected information about those materials in Volume 2: Nanostructured
Materials. In this volume, materials such as one-dimension metal oxide nanostructures, carbon nano-
tubes, fullerenes, metal nanoparticles, and nanoclusters are considered. Nanocomposites, porous semi-
conductors, ordered mesoporous materials, and zeolites also are among materials of this type.
Volume 3: Polymers and Other Materials, is a compilation of review chapters detailing applications
of chemical sensor materials such as polymers, calixarenes, biological and biomimetric systems, novel
semiconductor materials, and ionic conductors. Chemical sensors based on these materials comprise a
large part of the chemical sensors market.
Of course, not all materials are covered equally. In many cases, the level of detailed elaboration was
determined by their significance and interest shown in that class of materials for chemical sensor design.
While the title of this series suggests that the work is aimed mainly at materials scientists, this is not
so. Many of those who should find this book useful will be “chemists,” “physicists,” or “engineers” who
are dealing with chemical sensors, analytical chemistry, metal oxides, polymers, and other materials and
devices. In fact, some readers may have only a superficial background in chemistry and physics. These
volumes are addressed to the rapidly growing number of active practitioners and those who are inter-
ested in starting research in the field of materials for chemical sensors and biosensors, directors of in-
dustrial and government research centers, laboratory supervisors and managers, students and lecturers.
We believe that this series will be of interest to readers because of its several innovative aspects.
First, it provides a detailed description and analysis of strategies for setting up successful processes for
screening sensing materials for chemical sensors. Second, it summarizes the advances and the remain-
ing challenges, and then goes on to suggest opportunities for research on chemical sensors based on
polymeric, inorganic, and biological sensing materials. Third, it provides insight into how to improve the
efficiency of chemical sensing through optimization of sensing material parameters, including composi-
tion, structure, electrophysical, chemical, electronic, and catalytic properties.
We express our gratitude to the contributing authors for their efforts in preparing their chapters.
We also express our gratitude to Momentum Press for giving us the opportunity to publish this series.
We especially thank Joel Stein at Momentum Press for his patience during the development of this
project and for encouraging us during the various stages of preparation.
Ghenadii Korotcenkov
PREFACE TO
VOLUME 1: GENERAL APPROACHES
This volume provides an introduction to the fundamentals of sensing materials. We have tried to to
provide here the basic knowledge necessary for understanding chemical sensing through a brief de-
scription of the principles of chemical sensor operation and consideration of the processes that take
place in chemical sensors and that are responsible for observed operating characteristics. In spite of the
seeming extreme simplicity of chemical sensor operation and application, understanding the mecha-
nisms involved in the process of chemical sensing is usually not so simple. Chemical sensing as a rule
is a multistage and multichannel process, which requires a multidisciplinary approach. Therefore, in this
volume we provide a description of the important electronic, electrophysical, and chemical properties,
as well as diffusion, adsorption/desorption, and catalytic processes.
To our knowledge, this volume is the first attempt to analyze in detail the interrelationships be-
tween properties of sensing materials and operating parameters of chemical sensors. This volume
describes the properties of sensing materials by emphasizing the specificities of these materials. We
consider analyses that have been performed as bridging the gap between scientists studying properties
of materials and researchers using these materials for actual chemical sensor design. We hope that the
information included in this volume will help readers to approach soundly the selection of either sens-
ing materials or technology for sensing material synthesis or deposition.
Detailed consideration of various materials properties with respect to their application in chemical
sensors provides a clear idea of the complexity and ambiguity involved in selecting an optimal sensor
material. Research has demonstrated that there is no universal sensing material, and selection of an
optimal material is determined by the type of chemical sensor being designed and the requirements that
device will have to meet. This volume also illustrates the complementary nature of functionality in sens-
ing materials; for example, high sensitivity usually conflicts with stability. This richness and complexity
in behavior cannot be ignored.
This volume is intended to provide readers with a good understanding of the techniques used
for synthesis and deposition of sensing materials. Readers will find descriptions of different tech-
niques such as various methods of film deposition, sol-gel technology, deposition from solutions, col-
loidal processing, the peculiarities of polymers synthesis, techniques used for depositing coatings on
fibers, and so on. Description of various methods of synthesis and deposition, accompanied by detailed
xiii
xiv • PREFACE TO VOLUME 1: GENERAL APPROACHES
analysis of the advantages and shortcomings of those methods, provides the understanding necessary
for considered selection of a technology for forming a sensitive layer.
Analysis of metal oxide modification methods highlights the opportunities for control of the
properties of sensing materials, and demonstrates that a choice of methods should be based on consid-
eration of all possible consequences of the technical decision that is made.
Combinatorial and high-throughput materials screening approaches analyzed in this volume will be
also of interest to researchers working on materials design for chemical sensors.
We are confident that the present volume will be of interest of anyone who works or plans to start
activity in the field of chemical sensor design, manufacturing, or application.
Ghenadii Korotcenkov
ABOUT THE EDITOR
Devices in 1976, and his Habilitate Degree (Dr.Sci.) in Physics and Mathematics of Semiconductors
and Dielectrics in 1990. For a long time he was a leader of the scientific Gas Sensor Group and manager
of various national and international scientific and engineering projects carried out in the Laboratory
of Micro- and Optoelectronics, Technical University of Moldova. Currently, he is a research professor
at Gwangju Institute of Science and Technology, Gwangju, Republic of Korea.
Specialists from the former Soviet Union know G. Korotcenkov’s research results in the study
of Schottky barriers, MOS structures, native oxides, and photoreceivers based on Group III–V com-
pounds very well. His current research interests include materials science and surface science, focused
on metal oxides and solid-state gas sensor design. He is the author of five books and special publica-
tions, nine invited review papers, several book chapters, and more than 180 peer-reviewed articles. He
holds 16 patents. He has presented more than 200 reports at national and international conferences. His
articles are cited more than 150 times per year. His research activities have been honored by the Award
of the Supreme Council of Science and Advanced Technology of the Republic of Moldova (2004), The
Prize of the Presidents of Academies of Sciences of Ukraine, Belarus and Moldova (2003), the Senior
Research Excellence Award of Technical University of Moldova (2001, 2003, 2005), a Fellowship from
the International Research Exchange Board (1998), and the National Youth Prize of the Republic of
Moldova (1980), among others.
xv
CONTRIBUTORS
Beongki Cho (Chapter 4)

Department of Material Science and Engineering
Gwangju Institute of Science and Technology
Gwangju, 500-712, Republic of Korea
Ghenadii Korotcenkov (Chapters 1, 2, 4, and 5)

Department of Material Science and Engineering
Gwangju Institute of Science and Technology
Gwangju, 500-712, Republic of Korea
and
Technical University of Moldova
Chisinau, 2004, Republic of Moldova
Massood Zandi Atashbar (Chapter 1)

Department of Electrical and Computer Engineering
Western Michigan University
Kalamazoo, Michigan 49008-5066, USA
Radislav A. Potyrailo (Chapter 3)

Chemical and Biological Sensing Laboratory
Chemistry Technologies and Material Characterization
General Electric Global Research
Niskayuna, New York 12309, USA
Sridevi Krishnamurthy (Chapter 1)

Department of Electrical and Computer Engineering
Western Michigan University
Kalamazoo, Michigan 49008-5066, USA
xvii
CHAPTER 1
BASIC PRINCIPLES OF CHEMICAL

SENSOR OPERATION
M. Z. Atashbar
S. Krishnamurthy
G. Korotcenkov
1. INTRODUCTION
In recent years increased knowledge has led to significant development of chemical sensors for detec-
tion and quantification of chemical species. Chemical sensors have found a wide range of applications
in clinical, industrial, environmental, agricultural, and military technologies.
Chemical sensors are characterized by parameters such as sensitivity, selectivity, response and re-
covery time, and saturation. Chemical sensors can be classified in a number of ways, depending on their
principle of operation. In this chapter we describe the fundamental concepts of the following classes
of chemical sensors: electrochemical, conductometric, capacitive, work function, chemFET, catalytic,
Schottky diode, acoustic wave, mass-sensitive, optical, chemoluminesese, photoacoustic, thermal, sur-
face plasmon resonance, thermoelectric, thermal conductivity–based, flame ionization, and Langmuir-
Blodgett sensors.
2. ELECTROCHEMICAL SENSORS
Electrochemical sensors constitute the largest and most developed group of chemical sensors and have
taken a leading position with respect to commercialization in the fields of clinical, industrial, environ-
1
2 • CHEMICAL SENSORS: FUNDAMENTALS. VOLUME 1: GENERAL APPROACHES
mental, and agricultural analysis (Korotcenkov et al. 2009). Electrochemical sensors are based on the
detection of electroactive species involved in chemical recognition processes and make use of charge
transfer from a solid or liquid sample to an electrode or vice versa. An electrochemical sensing method
essentially requires a closed electrical circuit, which enables the flow of direct or alternating current to
make measurements. Figure 1.1 illustrates a basic structure of an electrochemical sensor. For gas sen-
sors, the top of the casing has a membrane which can be permeated by the gas sample.
For current to flow, at least two electrodes are needed. One electrode is an active electrode or
working electrode (WE), where the chemical reaction takes place; the other is a return electrode,
which is called a counter or auxiliary electrode (AE). A current is created as positive ions flow to the
cathode and negative ions flow to the anode. To measure the electrochemical potentials produced by
the electrodes and the electrolyte, a third electrode, called a reference electrode (RE), can be used. The
reference electrode corrects the error introduced as a result of polarization of the working electrode.
The working electrode is often made up of a layer of noble metal or catalytic metal such as platinum-,
palladium-, or carbon-coated materials. This working electrode is covered with a hydrophobic mem-
brane that acts as a transport barrier for the target chemical molecules (analyte), allowing the chemi-
cal species to diffuse to the working electrode and also preventing leakage of the electrolyte (Zao et
al. 1992). To produce a measurable signal, the electrodes must have a large surface area in order to
maximize the contact area with the analytes. Electrodes generally have a special coating to improve
their rate of reaction and to extend their working life. An electrolytic medium is required to carry the
ionic charges. Selective reactions can be accomplished by careful choice of the electrolyte and hence
it is the first stage in enhancing selectivity to a particular analyte. The sensor formed by the combi-
nation of an electrolyte, electrodes, and an external circuit is called an electrochemical cell. The cell
can be configured to extract electrical signals such as current, potential, conductance, or capacitance
(Fradeen 2003).
A common application for potentiometric and amperometric sensors is for water analysis. The
most common is the pH sensor system. A wide range of gaseous analytes such as oxygen, carbon
oxides, nitrogen oxides, sulfur oxides, and combustible gases also can be detected and quantified using
electrochemical sensors. Gases such as oxygen, nitrogen oxides, and chlorine, which are electrochemi-
gas Gas inlet gas

Electrolyte Porous working electrode
Gas membrane
Electrolyte
Working electrode Counter electrode Counter electrode Reference electrode

(a) (b)
Figure 1.1. Schematic diagram of electrochemical gas sensors with (a) two- and (b) three-electrode
configurations. (Reprinted with permission from Korotcenkov et al. 2009. Copyright 2009 American
Chemical Society.)
BASIC PRINCIPLES OF OPERATION • 3
cally reducible, are sensed at the cathode, while electrochemically oxidizable gases such as carbon mon-
oxide, nitrogen dioxide, and hydrogen sulfide are sensed at the anode.
Remarkable detectability, simple construction, and low cost have been the important advantages of
electrochemical sensors in comparison to optical, mass, and thermal sensing methods. Electrochemical
sensors can be classified into three categories based on their operating principles: amperometric, con-
ductometric, and potentiometric sensors (Janata 1989; Madou and Morrison 1989; Göpel et al. 1991;
Gardner 1994).
2.1. AMPEROMETRIC SENSORS

Electrochemical amperometric sensing involves measuring the flow of current at a constant applied
potential between working and auxiliary electrodes. The applied potential serves as the driving force
for the electron-transfer reaction of the electroactive species. The resulting current is a direct measure
of the rate of the electron-transfer reaction. It thus reflects the rate of the recognition event and is
proportional to the concentration of the target analyte.
The earliest design example of this sensor type was the two-electrode Clark oxygen sensor (Clark
et al. 1953). The counter electrode current arises by two oxygen–reduction steps:
O2 + 2H2O + 2e − → H2O2 + 2OH −

H2O2 + 2e − → 2OH − (1.1)
The basic structure of an amperometric sensor is similar to the one shown in Figure 1.1. A fixed
potential is applied to the reference and working electrodes by means of a potentiostatic circuit. Current
flow is recorded between the working and counter electrodes as a function of analyte concentration.
Upon exposures to an electroactive analyte, the analyte diffuses into the electrochemical cell through
the membrane and then to the working electrode. Depending on the type of analyte, the electrochemi-
cal reaction involves either oxidation or reduction, thus producing a current at the working electrode.
For instance, carbon monoxide oxidizes to carbon dioxide, resulting in a flow of electrons from the
working electrode to the counter electrode through the external circuit. On the other hand, gases such
as oxygen reduce to water, resulting in a flow of electrons from the counter electrode to the working
electrode. This current is related directly to the rate of reaction at the working electrode surface. If there
are no other species other than the analyte, the current varies in proportion to the amount of analyte
at the working electrode. Gold-based sensors respond to ozone and to nitrogen and sulfur compounds
(e.g., NO2, NO, HNO3, O3, H2S, SO2, HCN); platinum-based sensors respond to all of those gases plus
carbon–oxygen compounds (CO, formaldehyde, alcohols, etc.). Some examples with the reactions at the
counter and working electrodes are shown in Table 1.1.
Although amperometric sensors have long been in wide use, microfabrication of such sensors
has recently started. To name a few, Somov and co-workers (2000) developed multielectrode zirconia
electrolyte amperometric sensors, Benammar (2004) developed zirconia-based oxygen sensors, and Lee
and co-workers (2004) developed solid-state amperometric CO2 sensors. For example, Lee and co-
workers (2004) used Na2CO3 as a porous coating, platinum paste as counter and working electrodes, and
Table 1.1. Reactions at working and counter electrode in amperometric sensors,

with their common electrocatalyst for the reaction mechanism
WORKING ELECTRODE COUNTER ELECTRODE ELECTRODE

(ANODE) (CATHODE) ELECTROCATALYST
1
CO H2 O CO 2 2H 2e O2 2H 2e 2H2 O Platinum
2
1
SO2 H2 O H2 SO4 2H 2e O2 2H 2e 2H2 O Gold
2
NO 2H2 O HNO3 3H 3e O2 4H 4e 2H2 O Gold
1
NO2 2H 2e NO H2 O HO2 O2 2H 2e Gold
2
1
Cl 2 2H 2e 2HCl HO2 O2 2H 2e Platinum
2
NASICON (Na3Zr2Si2PO12, a sodium super-ionic conductor) as an electrolyte material with a constant

potential of 0.1 V. Figure 1.2 shows the transient response of the sensor to various CO2 concentrations.
As can be seen from Figure 1.2, the current increases at higher concentrations of CO2.
2.2. CONDUCTOMETRIC SENSORS

In an electrochemical conductometric sensor, the change in conductivity of the electrolyte in an electro-
chemical cell is measured at a series of frequencies. This type of sensor may have a capacitive com-
ponent due to the polarization of the electrodes and the charge-transfer process. In a homogenous
electrolytic solution, the conductance is (Fradeen 2003)
Figure 1.2. CO2 amperometric transient response.

(Reprinted with permission from Lee et al. 2004.
Copyright 2004 Elsevier.)
σA
G= (1.2)
L
where σ is the specific conductivity of the electrolyte, which is related to the concentration of the charg-
es and ionic species, L is the segment of the solution along the electric potential; and A is the cross-
sectional area perpendicular to the electric field. For measurements, a Wheatstone bridge configuration
is usually used, in which the electrochemical cell forms one of the arms of the bridge. Measurement
of the conductivity of the liquid is complicated due to the charge-transfer process at the electrode
surface and polarization of the electrodes at the operating voltage. Therefore the measurement needs
to be performed at low voltage so that charge transfer does not take place. When an electrode is in an
electrolyte solution, a potential is generated due to the unequal distribution of charges across the inter-
face. Two oppositely charged layers, one on the electrode surface and one inside the electrolyte, form
a “double layer,” which behaves as a parallel-plate capacitor. This results in Warburg impedance, which
depends on the frequency of the potential perturbation. At high frequencies the Warburg impedance is
small, since diffusing reactants do not have to move very far. At low frequencies the reactants have to
diffuse farther, thereby increasing the Warburg impedance. Employing a high-frequency, low-amplitude
alternating signal can minimize both the charge-transfer and double-layer effects.
2.3. POTENTIOMETRIC SENSORS

In potentiometric sensors, the analytical information is obtained by converting the recognition process
into a potential signal, which is proportional (in a logarithmic fashion) to the concentration (activity) of
species generated or consumed in the recognition event. Such devices rely on the use of ion-selective
electrodes to obtain the potential signal. A permselective ion-conductive membrane (placed at the tip of
the electrode) is designed to yield a potential signal that is due primarily to the target ion. The response
is measured under conditions of essentially zero current.
Upon equilibrium at the analye–electrolyte interface in the electrochemical cell, an electrical poten-
tial develops due to the “redox” reaction. The reaction occurs at one of the electrodes and is called a
half-cell reaction. The reactions for H2 sensors are shown in Figure 1.3.
Under quasi-thermodynamic equilibrium conditions, the Nernst equation is applicable (Janata 1989):
RT ⎛ C 0* ⎞
E = E0 + ln ⎜ ⎟ (1.3)
nF ⎝ C R* ⎠
where R, T, n, F, C 0*, and C R* are the gas constant (8.314 J mol−1 K−1), temperature (in K), number of
electrons transferred, Faraday constant (96,485 C), concentrations of oxidant and reduced product,
respectively, and E0 is the electrode potential at the standard state.
In a potentiometric sensor, two half-cell reactions take place at each electrode; however, only the
reaction at the working electrode needs to be involved in the measurements. The measurement of the
cell potential should be made under zero-current or quasi-equilibrium condition. Therefore, a very-
high-input impedance amplifier should be used.
P(H2)(I) > P(H2)(II)

½ H 2 H+ + e- H+ + e- ½ H2
H2 H2
H+
P(H2)(I) P(H2)(II)
EMF
- +
E0cell=RT/2F ln[PH2(I)/PH2(II)]
Figure 1.3. Working principle of hydrogen concentration cell using a proton-conducting membrane.
(Adapted with permission from Sundmacher et al. 2005. Copyright 2005 Elsevier.)
Potentiometric sensors are very attractive for field operations because of their high selectivity,
simplicity, and low cost. They are, however, less sensitive and often slower than their amperometric
counterparts. In the past, potentiometric devices were more widely used, but the increasing amount of
research on amperometric probes would gradually shift this balance. Detailed theoretical discussions of
amperometric and potentiometric measurements are available in many textbooks and reference works
(Korotcenkov et al. 2009).
2.3.1. Ion-Selective Electrodes

Ion-selective electrodes (ISEs) are potentiometric sensors that measure the electric potential of a spe-
cific ionic species in the solution. This potential is measured against a stable reference electrode with
constant potential. These sensors are also known as specific-ion electrodes (SIEs). A schematic diagram
illustrating the principle of operation of ISEs is shown in Figure 1.4.
The pH sensor based on a glass membrane is the oldest type of ISE. Credit for the first glass sensing
pH electrode is given to Cremer, who first described it in a 1906 paper (Meyerhoff and Opdeycke 1986).
In 1949, George Perley published an article on the relationship of glass composition to pH function
(Frant 1994). Since then, many other types of ion-selective membranes have been developed and char-
acterized, including polymeric membranes having various ionophores, which have extended the sens-
ing capabilities of ISEs to a wide range of ions. Other ions that can be measured include ions such as
metal fluorides (Apostolakis et al. 1991) and bromides (Shamsipur et al. 2000), cadmium (Hirata and
Higashiyama 1971), copper (Gupta and D’Arc 2000), and gases in solution such as ammonia (Radomska
et al. 2004) and carbon dioxide (Johan et al. 2003). Ion-selective electrodes offer direct and selective
detection of ionic activities in water samples. Such potentiometric devices are simple, rapid, inexpen-
sive, and compatible with online analysis. The inherent selectivity of these devices is attributed to highly
selective interactions between the membrane material and the target ion. Depending on the nature of
the membrane material used to impart the desired selectivity, ion-selective electrodes can be divided into
three groups: glass, solid, and liquid electrodes.
The ISE ideally follows a Nernst equation (Taylor and Schultz 1996):
⎛ RT ⎞
E = K + ⎜ 2.303 × ⎟ log a i (1.4)
⎝ ZF ⎠
where E is the potential, K is a constant representing the effect of various sources such as liquid junc-
tion potentials, R is the universal gas constant (8.314 J mol−1 K−1), T is the absolute temperature, Z is
ionic charge, and ai is the activity of ion i. The activity of an ion i in solution is related to its concentra-
tion ci :
ai = γi ci (1.5)
where γi is the activity coefficient, which depends on the types of ions present and the total ionic
strength of the solution.
At present, many ion-selective electrodes are commercially available and are used routinely in vari-
ous fields. Electrochemical analysis systems based on ISEs may be applied to measure ionic concentra-
tions in biomedical, food processing, water quality, and pollution monitoring applications.
Figure 1.4. Schematic of a typical potentiometric cell assembly incorporating an ISE and reference
electrode as the galvanic half-cells and a high-impedance voltmeter for measurement of the cell EMF.
3. CAPACITANCE SENSORS
A capacitor is an electrical circuit component which can store electrical energy in response to an ap-
plied electric field. It consists of a dielectric material which is sandwiched between two parallel metal
electrodes. When a voltage is applied to these plates, equal and opposite charges accumulate on the
electrodes. Capacitance can be expressed as
ε 0 εr A
C= (1.6)
d
where ε0 represents the permittivity in vacuum, εr is the relative permittivity of the dielectric, A is the
area of the electrode, and d is the distance between the two electrodes.
The main operating principle of capacitance sensors is that the analyte of interest can change εr ,
A, or d, and therefore, by measuring the change in the capacitance, the presence of the measurand can
be detected (Ishihara et al. 1998). With reference to polymer coating, the mechanisms which can be ac-
companied by the change of these parameters are shown in Figure 1.5. Upon analyte absorption, the
physical properties of the polymer layer change as a result of the incorporation of the analyte molecules
into the polymer matrix. Changes in two physical properties influence the sensor capacitance: (1) vol-
ume (swelling) and (2) dielectric constant. The resulting capacitance change is detected by the read-out
electronics (Kummer et al. 2004).
A change in area is not common in capacitor sensors, but changes in relative permittivity and
dielectric thickness are commonly used (Seiyama et al. 1983; Matsuguch et al. 1998; Yamazoe et al.
1989; Harrey et al. 2002). Interdigitated electrode structures as shown in Figure 1.5 are normally used
for capacitor sensors, where the capacitance of the sensor is measured rather than the conductance.
In this case, the sensitive layer can be a polymer film or ceramic (Sheppard et al. 1982; Ishihara et al.
1998). The interdigitated capacitors used for this purpose have capacitance of the order of 1 pF and
therefore the capacitance changes due to the sensing mechanism are in the range of attofarads (10−18 F).
Figure 1.5. Schematic of sensing principle, showing analyte absorption and the two relevant effects
that affect the sensor capacitance: a change in the dielectric constant and swelling. The interdigi-
tated electrodes (+, −) on the substrate (black) are coated with a polymer layer (gray). Big and small
globes represent analyte and air molecules, respectively. Analyte molecules are polarized (v) in the
electric field (solid lines). Analyte-induced polymer swelling is indicated by dashed lines (right side).
(Reprinted with permission from Kummer et al. 2004. Copyright 2004 American Chemical Society.)
Since it is difficult to transfer these small signals through wires and cables, on-chip circuitry is required
(Hierlemann et al. 2003). For this purpose, two capacitors, one a sensing capacitor and the other a refer-
ence capacitor, are used in a switched capacitor configuration along with a sigma-delta modulator for
analog-to-digital conversion (Hagleitner et al. 1999). The differential capacitance changes are converted
to frequency changes by the on-chip circuitry.
Capacitive microsensors are promising devices for chemical sensing, because no micromechanical
parts, such as membranes or cantilevers, are needed to render the chip more rugged, and there are fewer
postprocessing steps, and hence, lower costs. Moreover, the low power consumption favors their use
in hand-held devices.
One of the most well known capacitive gas sensors based on changes in relative permittivity is the
humidity sensor. Polymer films such as polyimides or cellulose acetates are used as humidity-sensitive
dielectrics because the difference in their relative permittivities at room temperature is very large. Pure
water has a dielectric constant of 78.5, whereas those of the polyimides are in the range of 3–6.
Therefore, the capacitance in these sensors is highly sensitive to the adsorption of water (Ishihara et al.
1998; Hierlemann et al. 2003).
Capacitive sensors based on changes in thickness rely on the formation of depletion layers in the
dielectric of the capacitor. Depletion layers formed between p- and n-type semiconductors, metal ox-
ide semiconductors, or metal semiconductors have been investigated for this purpose. The electronic
interaction of the analyte and the interface changes the carrier concentration in the semiconductor, re-
sulting in changes in the depletion-layer thickness (Schoeneberg et al. 1990; Ishihara et al. 1998; Nagai
et al. 1998). Chemical capacitance sensors have also been investigated for detection of other gases such
as CO (Balkus et al. 1997), ammonia and methane (Domansky et al. 2001), and NOx (Ishihara et al.
1995, 1996).
4. WORK-FUNCTION SENSORS
In the solid phase, the work function (Ф) of a material is defined as the minimum energy required by an
electron to escape from the Fermi level of the bulk material to the vacuum level (Streetman 1990).
Ф = −qθ − EF (1.7)
EF = −Ф − qθ (1.8)
The Fermi energy (EF) level corresponds to the electrochemical potential of an electron. At a given
temperature, the Fermi level is the highest occupied energy level in the band gap. The energy difference
between EF and the vacuum (Volta) energy level (−qθ) is the work function, as illustrated in Figure 1.6
(Janata and Josowicz 2002; Anh et al. 2004).
When an electroactive material is brought into electrical contact with a metallic electrode, charge
transfer occurs until the Fermi energies of both materials reach thermal equilibrium and the Fermi lev-
els of the two materials become equal. However, their Volta energy levels are different. The difference
in their work functions is equal to their difference in Volta energy. This can be represented as (Anh et
al. 2004)

Tɽ
9DFXXPOHYHOT
&RQGXFWLRQ :RUNIXQFWLRQˇ
%DQG
)HUPLOHYHO()
9DOHQFH%DQG
Figure 1.6. Illustration of energy levels in an energy-band diagram.
−φMet − q θMet = −φMat − q θMat (1.9)
M = φMet − φMat = −q (θMet − θMat )

φMet (1.10)
where φMet and φMat are the work functions of electrons in the metal and the electroactive material,
respectively; –qθMet and −qθMat are the Volta energies of the metal and the electroactive material, re-
spectively; and φMet
Mat is the difference between the work function of the electrons in the metal and that
of the material (Anh et al. 2004). The redox properties of materials can be studied by measuring their
work function using a Kelvin probe, which consists of a vibrating capacitor (Janata and Josowicz 1997).
Using this method, the change in the work function due to the redox reaction of the material can be
measured. When the surface of the metal and the electroactive material are held parallel to each other
and separated by a dielectric material, there is a potential difference across the surface due to transfer of
charges (Bergstrom et al. 1997). This potential difference is given by the difference between the work
functions of the two materials and is given by a bulk term and a surface dipole term. Gases interact
with the surface dipole of one or both materials that constitute the Kelvin probe. The extent to which
the gases can interact with the surface dipole determines the amount of work function shift (Bergstrom
et al. 1997).
In measuring the contact potential differences between two surfaces using the Kelvin probe meth-
od (see Figure 1.7), one of the two surfaces oscillates periodically with respect to the other. This oscil-
lation induces a charge across the surface. This structure can be modeled as a capacitor with an induced
charge Q(t ) that depends on the time-varying displacement of the two plates. This induced charge can
be expressed in terms of the capacitance C(t) and the work function difference V12 between the sensing
plate and the reference plate as (Bergstrom et al. 1997)
ε0 A
Q( t ) = C ( t ) (V12 + Vb ) = (V12 + Vb ) (1.11)
d (t )
where Vb is the backing potential which is used to nullify the charge when V12 = −Vb, A is the area of
the plates, and d(t) is the time-varying distance between the plates and is given by
d (t ) = d0 + d1 sin ωt (1.12)
Figure 1.7. Setup of a Kelvin probe. A metallic grid oscil-

lates over a sample. The counter voltage VG is adjusted so
that the current i vanishes. (Reprinted with permission from
Barsan and Weimar 2003. Copyright 2003 IOP.)
where d0 is the plate spacing, d1 is the displacement, and ω is the frequency of oscillation. The oscillating
plate capacitively generates a sinusoidal current in the sensing film, and this current is proportional to
the work function difference V12. Therefore, the resolution of the current measurement determines the
overall resolution of the system (Bergstrom et al. 1997). For a given work function difference, resolution
improves with the magnitude of the current. An example of NO2 sensing using the work function princi-
ple is presented by Karthigeyan and co-workers (2001), who used nanoparticulate SnO2 films. The SnO2
films used in the sensor contained particles with an average size of 15 nm. The work function measure-
ments were performed with a Kelvin probe. Figure 1.8 shows the response of the SnO2 work function–
based sensor to 100 ppm NO2 at a temperature of 130°C. It can be seen from this figure that sensors
have almost a negligible baseline drift after repeated exposure to NO2 (Karthigeyan et al. 2001).
Work function–type gas sensors for detection of CO2 (Ostrick et al. 1999), ammonia (Ostrick et
al. 2000), oxygen (Bergstrom et al. 1997), and ozone (Doll et al. 1996; Zimmer et al. 2001) have been
explored. Work-function sensors based on field-effect transistors (FETs) have also been studied in
Figure 1.8. Response of SnO2 sensor to 100-ppm NO2 at a working temperature of 130º. (Reprinted
with permission from Karthigeyan et al. 2001. Copyright 2001 Elsevier.)
detail for gas sensing. These are called adsorption-based FETs (ADFETs) (Middlehoek and Audet 1989;
Janata 1989; Madou and Morrison 1989), ion-sensitive FETs (ISFETs) (Harame et al. 1987; Bergveld
1991), or chemically modified FETs (chemFETs) (Eddowes 1987). ChemFETs are discussed in detail
in the following section.
5. FIELD-EFFECT TRANSISTOR SENSORS

The basic metal-oxide semiconductor field-effect transistor (MOSFET) consists of four terminals, a
source, drain, gate, and substrate, as shown in Figure 1.9. The source and drain are highly conducting
semiconductor regions, whereas the gate is a metal region separated from the source and drain by a gate
oxide. The most important parameter of a MOSFET is the threshold voltage (VT), which is defined as
the gate voltage at which there is an onset of inversion layer at the surface of the semiconductor. The
carrier concentration of the inversion layer has opposite polarity to that of the carrier concentration in
the bulk of the semiconductor. The formation of the inversion layer is controlled by the applied gate
voltage. The net result is that the current between the drain and source is controlled by the voltage that
is applied to the gate and the drain-to-source voltage.
The drain current of a MOSFET is given by the following equation (Razavi 2000):
⎛ 1 ⎞
I D = β ⎜VGS − VT − VDS ⎟VDS (1.13)
⎝ 2 ⎠
where VGS and VDS are the gate-to-source and drain-to-source voltages, respectively; VT is the threshold
voltage of the MOSFET; and β is the sensitivity parameter determined by the dimensions of the gate
and given by
VGS
VDS
Gate
Gate oxide
Source Drain
L
p-type substrate
Figure 1.9. Schematic of a four-terminal MOSFET.

W
β = μC ox (1.14)
L
where μ is the mobility of electrons in the channel, Cox is the gate oxide capacitance per unit area, W is
the channel’s width, and L is the length of the channel. The threshold voltage (VT) of the MOSFET is
given by (Razavi 2000)
Qdep
VT = φMS + 2φF + (1.15)
C ox
where φMS is a function of the difference between the work function of the gate and the silicon sub-
strate, Qdep is the charge in the depletion region, and
kT ⎛ N sub ⎞
φF = ln ⎜ ⎟ (1.16)
q ⎝ ni ⎠
where Nsub is the doping concentration of the substrate and ni is the intrinsic carrier concentration.
A theory for the operation of these FET-based devices and metal insulator semiconductor (MIS)
capacitors as gas sensors was developed by Lundström and co-workers. (Lundström 1991; Ekedahl et al.
1998). The basic theory regarding a gas FET or MIS capacitor with a Pd electrode (or similar catalytic
metal, such as Pt) sensitivity to hydrogen is produced by a modulation in the flat band potential. If we
consider a gas FET with a thick Pd-gate film (≈ 100 nm), hydrogen gas molecules are absorbed at the
surface of the palladium by disassociating into atomic hydrogen. These atoms diffuse through the metal
and absorb on the inner surface of the gate insulator junction, where they become polarized as shown
in Figure 1.10. The resulting dipole layer is in equilibrium with the outer layer of chemisorbed hydrogen
Figure 1.10. Mechanism of MOSFET operation in a hydrogen atmosphere. (Reprinted with permis-
sion from Lundstrom et al. 2007. Copyright 2007 Elsevier.)
and hence in phase with the gas. The dipole layer produces an abrupt rise in the surface potential at the
metal oxide interface.
When the hydrogen is removed from the ambient atmosphere, the atomic hydrogen within the
MOSFET recombines into hydrogen molecules, which desorb from the surface. If the atmosphere
contains oxygen, then this recombination is dominated by the production of water. Once the hydrogen
has been removed from the metal–insulator interface, the potential between the Pd and the insulator
returns to its initial level. It has been shown that the reaction is totally reversible, though the recovery
rate can be very slow and may take a number of hours. The reaction described above occurs at room
temperature, though to increase the speed of the response and to remove water molecules rapidly from
the metal surface, the sensors are normally operated at 150–200°C.
The gas FET response is measured as a shift in either gate-source voltage (VGS ) or drain-source
current (IDS ) (see Figure 1.11), and it has been shown that this response is related to a shift in the
threshold voltage. This shift in threshold voltage is usually defined as
VT = VT0 – ΔV (1.17)
where VT0 is the initial threshold voltage and ΔV is the potential produced by the formation of the
dipole layer.
Experiments have shown that this sensitivity to hydrogen can be measured over a dynamic pressure
range of 14 orders of magnitude without the sensor saturating.
Earlier gas FET sensors were fabricated mainly based on silicon. However, in recent years, great
success has been achieved in development of such devices on the basis of silicon carbide. The use
of SiC as the semiconductor substrate in field-effect devices led to the development of gas sensors
Figure 1.11. “Classical” schematic drawing illustrat-

ing the hydrogen-sensitive field-effect devices, in
which hydrogen atoms adsorbed at the metal–oxide
interface cause a shift of the electrical characteris-
tics along the voltage axis in devices having catalytic
metal (Pd) gates. (Reprinted with permission from
(Lundstrom et al. 2007. Copyright 2007 Elsevier.)
operating at high temperatures. It was demonstrated that SiC-based sensors respond to a change in
ambient between oxygen and propane even at 1000°C (Lundstrom et al. 2007).
6. chemFET-BASED SENSORS
The schematic of a chemFET is similar to that of a MOSFET except that the gate structure is slightly
modified. Based on the gate structure, the family of chemFETs can be divided into three main types,
ion-sensitive field-effect transistors (ISFETs), enzymatically selective field-effect transistors (ENFETs),
and basic or work function chemFETs (Wilson et al. 2001).
The ion-sensitive field-effect transistor is a chemFET structure without a conductive gate. The ion-
selective layer is placed on top of the insulator layer of the FET structure. The enzymatically selective
field-effect transistor is also a chemFET structure without a conductive gate. The chemically sensitive
enzyme layer forms a part of or the entire insulator layer of the FET. The basic or work function chem-
FET is a chemFET structure with a conductive gate. The conductive gate is the chemically selective
layer (Wilson et al. 2001). All of these chemFET structures work on the basic principle that the surface
charge changes at the interface between the insulator and the overlying layer (the overlying layer is the
ion-selective layer in an ISFET, the enzyme-selective layer in the ENFET, and the gate layer in the basic
or work function chemFET) (Yotter and Wilson 2004). This results in a change in the work function,
and therefore, the threshold voltage changes in accordance with Eq. (1.15).
Extensive research has been conducted on chemFETs for monitoring ammonium (Brzozka et al.
1997; Senillou et al. 1998), cadmium and lead (Reinhoudt 1995; Ali et al. 2000), Cu2+ (Taillades et al.
1999), hydrogen (Domansky et al. 1998), and pH (Bausells et al. 1999; Cho and Chiang 2000).
Ion-sensitive field-effect transistors were first developed in the early 1970s (Bergveld 1970). Figure
1.12 is a schematic diagram of an ISFET. As can be seen, its structure is the same as that of a MOSFET.
The main component of an ISFET is an ordinary metal-oxide silicon field-effect transistor (MOSFET)
with the gate electrode replaced by a chemically sensitive membrane, solution, and a reference electrode
Figure 1.12. Schematic diagram of an ISFET.

(RE). As in a MOSFET, the channel resistance in an ISFET depends on the electric field perpendicular
to the direction of the current. Charges from solution accumulate on top of this insulating membrane
and do not pass through the ion-sensitive membrane. Anisotropic ions accumulate at the contact inter-
face between an electrochemically active surface and a liquid electrolyte. Because of their different sizes
and charges, the ions form a well-confined electric double layer close to the surface, and a diffuse layer
of outer charges exists between the Helmholtz planes and the neutral bulk of the solution (Yuqing et
al. 2003). Of course, to build a selective sensor, it is necessary to select a membrane that allows only
one chemical reaction to occur.
Because of the similarity between the ISFET and the MOSFET, most papers dealing with the
ISFET operational mechanism start with the theoretical description of a MOSFET. Therefore, a gen-
eral expression for the drain current of the ISFET and MOSFET in nonsaturated mode is as described
before [see Eq. (1.13)].
The fabrication processes for MOSFETs and ISFETs are the same, which results in identical
threshold voltages as shown in Eq. (1.18). However, in the case of ISFETs, two additional parameters
are involved: the constant potential of the reference electrode, Eref, and the interfacial potential at the
solution/oxide interface, which is the chemical input parameter, shown to be a function of the solution
pH, and is the surface dipole potential of the solvent and thus has a constant value. Hence the expres-
sion for the ISFET threshold voltage becomes (Bergveld 2003)
Qdep
VT = Eref − Ψ + χsol + 2φF + (1.18)
C ox
The fabrication of chemical microsensors based on ISFETs and selective ionic membranes has
been extensively studied (Bergveld 2003). Among the various aspects to consider for a good imple-
mentation of selective membranes on such devices, the most critical are the adhesion of the film to
the sensor surface and the reproducibility of the deposition process, which in most cases is carried out
manually. The use of Langmuir-Blodgett (LB) nanostructures is a good technique, since the deposition
method provides films with well-ordered molecular orientation and good thickness control. Thin selec-
tive monolayers can be obtained using a phospholipid matrix containing ionophores as specific recogni-
tion molecules (Jimenez et al. 2004).
Because of their robustness and the need for minimal maintenance, ISFET sensors can be used
for online process measurements, eliminating the need for sampling and later analysis in the laboratory.
In addition, ISFETs have very fast response, high sensitivity, batch processing capability, microsize, and
the potential for on-chip circuit integration. What is most attractive is that ISFETs are a preferred trans-
ducing element for biosensors because the SiO2 surface contains reactive SiOH groups, which can be
used for covalent attachment of organic molecules and polymers. Also, ISFETs can be used to develop
immunobiosensors (Yuqing et al. 2003).
7. SCHOTTKY DIODE–BASED SENSORS

The Schottky diode is a semiconductor-based diode and consists of a metal–semiconductor junction
instead of a semiconductor–semiconductor junction as in conventional diodes. Schottky diodes use the
metal–semiconductor junction as the Schottky barrier. Since the barrier height is lower in metal–semi-
conductor junctions than in conventional p–n junctions, Schottky diodes have lower forward voltage
drop. An n-type semiconductor is normally used in Schottky diodes. Due to the absence of the p-type
semiconductor region in these diodes, faster switching times are possible since the mobility of electrons
is three times that of holes. These diodes can be used as sensors for various chemicals if the adsorbed
analyte changes the electrical characteristics of the Schottky diode.
Based on the thermionic field emission conduction mechanism of the Schottky diode, the I–V char-
acteristics of the diode for forward bias voltage greater than 3 kT is given by (Chen and Chou 2003)
⎛ qV ⎞
I = I SAT exp ⎜ ⎟ (1.19)
⎝ nKT ⎠
where K is the Boltzmann constant, T is the temperature in kelvin, n is the ideality factor, and ISAT is the
saturation current defined as (Chen and Chou 2003)
⎛ −q φb ⎞
I SAT = SA ∗∗T 2 exp ⎜ ⎟ (1.20)
⎝ KT ⎠
where S is the area of the junction, A** is the effective Richardson constant, and φb is the barrier height.
The ideality factor can be extrapolated from Eq. (1.19), and the barrier height can be calculated from
Eq. (1.20).
A cross section of a typical chemical sensor based on a Schottky barrier is shown in Figure 1.13.
Each Schottky diode comprises two contacts, or junction areas, by definition: (1) between metal I
and the semiconductor, forming the Schottky barrier junction, which is the origin of the sensor signal;
and (2) between metal II and the semiconductor, forming the ohmic contact, which has to be inactive
and hence should not contribute to the sensor response. The type of contact is defined by the work
function of the materials used for fabrication.
The changes in the I–V characteristics of the Schottky diode in the presence of the analyte as a
rule are conditioned by the change of the Schottky barrier height. The response can be due either to
adsorption of the species of interest at the metal surface affecting interfacial polarization by formation of
a dipole layer or to absorption of gases or vapors of interest by the semiconductor and their interaction
Figure 1.13. Cross-sectional view of a typical geometric arrangement of a Schottky barrier di-
ode used for chemical sensing applications. (Reprinted with permission from Potje-Kamloth 2008.
Copyright 2008 American Chemical Society.)
with the semiconductor, which changes its work function and hence the contact potential or built-in
voltage of the diode (Trinchi et al. 2004; Potje-Kamloth, 2008)
It is necessary to note that Schottky diodes with catalytic contact metals have been widely studied
mainly for their application in the detection of hydrogen and hydrogen-producing compounds (see
Figure 1.14) (Trinchi et al. 2003; Tsai et al. 2003; Lin et al. 2004). Schottky diodes are advantageous for
gas-sensing applications due to the simple electrical circuitry required to operate them. In hydrogen
sensors based on Schottky diodes, the key role is played by the Schottky metallization, mostly belonging
to the platinum group metals (Pd, Pt), which ensures the catalytic dissociation of molecular hydrogen
into hydrogen atoms. It should be noted that the semiconducting substrate is only used to provide
sufficiently high Schottky barrier heights, necessary for good sensing performance, but is not the gas-
sensitive part of the sensor.
According to the dipole model, originally developed by Lundstrom, the hydrogen sensitivity of
Schottky diode gas sensors is based on the dissociation of hydrogen molecules on the catalytic metal
surface and the diffusion through the metal film to form a polarized layer at the metal/insulator inter-
face. When atomic hydrogen has been formed on the outer metal surface, which is exposed to the ambi-
ent (see Figure 1.14), an equilibrium between the hydrogen concentration at this metal surface and that
at the metal/semiconductor interface is reached. Based on this model, the change in the I–V characteris-
tics and the decrease in the Schottky barrier height are strongly related to the hydrogen concentration.
Schottky-based sensors also respond to hydrogen-containing molecules such as hydrocarbons,
provided that the molecules are also dissociated on the catalytic metal surface (Trinchi et al. 2003).
Oxygen atoms and NOx (Tuyen et al. 2002) are also dissociated on the catalytic metal surface. Water
formation with oxygen atoms from oxygen-containing molecules consumes hydrogen and therefore
decreases the sensor response. In other words, catalytic metal gates have a direct response to hydrogen
Figure 1.14. Hydrogen response of Schottky diode with a forward bias current of 1 mA at 310°C.
(Reprinted with permission from Trinchi et al. 2004. Copyright 2004 Elsevier.)
and hydrocarbons as well as an indirect response to oxygen molecules, the effect of which is to decrease
the direct response.
Besides Si, compound semiconductors such as GaAs, InP, GaN, Ga2O3, and SiC have been alterna-
tively employed as substrate materials for Schottky diode–type hydrogen sensors (Chen and Chou 2003;
Trinchi et al., 2004). GaN, Ga2O3, and SiC have a wide bandgap, and hence operating temperatures up
to 900°C are achievable, as compared to silicon substrates, for which operating temperatures are, limited
up to 250°C (Trinchi et al. 2003).
8. CATALYTIC SENSORS
The catalytic sensors widely known as “catalytic bead” or “pellistors” were among the first chemical
sensors. Such detectors have been in widespread use for more the 50 years in portable, transportable,
and fixed multipoint gas alarms. State-of-the-art catalytic sensors are stable, reliable, accurate, rug-
ged, and have a long operating life. The output is linear because the platinum wire has a good linear
coefficient of thermal resistance. Although many design improvements have been made to detectors
of this type over the years, in essence, the basic concept has not changed (Symons 1992; Miller 2001;
Korotcenkov 2007).
In operation, the pellet and consequently the catalyst layer is heated by passing a current through
the underlying coil. In the presence of a flammable gas or vapor, the hot catalyst allows oxidation to
occur in a chemical reaction similar to combustion. Just as in combustion, the catalytic reaction releases
heat, which causes the temperature of the catalyst together with that of its underlying pellet and coil
to rise. This rise in temperature results in a change in the electrical resistance of the coil, and it is this
change in electrical resistance which constitutes the signal from the sensor.
The sensor temperature rise can be detected via an increase in the Pt coil resistance, typically by
incorporating the sensing element in a Wheatstone bridge circuit. A measurement voltage is applied
across both arms of the bridge, and the resistance in each arm is matched so that the potential differ-
ence measured across the center of the bridge is zero. Any change in the resistance of the platinum wire
will now result in a change in this measured voltage.
The detection elements may take various forms according to the origin of their design. Possible
versions of detection elements in simplified form are presented in Figure 1.15. The simplest form is
Figure 1.15. Schematic views of various pellistors. (Reprinted with permission from Korotcenkov
2007b. Copyright 2007 Elsevier.)
the platinum wire filament (Figure 1.15a), still used in some gas detection instruments. However, the
“bead” construction is the most effective and promising. The “bead” (see Figure 1.15c) takes the form
of a ceramic pellet supported on a platinum wire coil. Usually, pellistors consist of small-diameter Pt
spirals (approximately 20 turns) surrounded by a porous refractory bead, typically alumina impregnated
with a precious metal catalyst such as palladium, thorium, or another catalyst. A noble metal acts as a
catalyst to promote exothermic oxidation of flammable gases. The presence of catalyst coating reduces
the temperature needed to achieve a stable signal for hydrocarbons. In this case the sensitive element is
normally electrically self-heated to between 400 and 800°C for methane and other gases. This helps in
extending the life of the filament.
We need to note that pellistors are sensitive only to flammable gases and vapors. Hot-wire sensors
(pellistors) are traditionally used to detect the percentage of the lower explosive limit (% LEL) levels
of flammable gases and vapors. The LEL is usually expressed as percent by volume of combustible gas
in air. Of course, pellistors have been used for detection of parts-per-million concentrations of some
gases and vapors, but in this case they require active sampling and other instrument design consider-
ations. Moreover, the pellistors are nonselective sensors (Korotcenkov 2007).
9. CONDUCTOMETRIC SENSORS
Conductometric sensors have a simple structure and their operating principle is based on the fact that
their electrical conductivity (or electrical conductance) can be modulated by the presence or absence of
some chemical species that comes in contact with the device (Wohltjen et al. 1985). Figure 1.16 shows
a typical structure of a conductometric sensor. It consists of two elements, a sensitive conducting layer
and contact electrodes. These electrodes are often interdigitated and embedded in the sensitive layer. To
make the measurement, a DC voltage is applied to the device and the current flowing through the elec-
trodes is monitored as the response. The main advantages of these sensors are easy fabrication, simple
operation, and low production cost, which means that well-engineered metal-oxide conductometric
sensors can be mass produced at reasonable cost. Moreover, these sensors are compact and durable. As
a result, they are amenable to being placed in situ in monitoring wells.
The basis of the operation of conductometric sensors is the change in resistance under the ef-
fect of reactions (adsorption, chemical reactions, diffusion, catalysis) taking place on the surface of

Sensitive layer IDTs electrodes
Insulating substrate
Insulating substrate
Figure 1.16. Structure of a typical conductometric sensor.

the sensing layer. The chemical species interact with the sensitive layer and thus modulate its electrical
conductivity. This can be measured as a change in the current, which is correlated to the concentration
of the chemical species.
The two main categories of conductometric gas sensors are thin- and thick-film sensors. Their
preparation involves different techniques. The different structural properties due to the fabrication and
film thickness lead to different sensing properties.
In thick-film sensors the film thickness is typically in the range of 2–300 μm. Thick-film technol-
ogy, based mainly on screen printing, is the most common fabrication technique for such sensors. The
layer itself is a porous body, so the inner surface also becomes a working surface. Therefore, the gases
diffuse into it, leading to good sensitivity. The microstructure of the sensing layer is a function of tem-
perature parameters of grains sintering. The conductance usually is controlled by the contact resistance
between the grains.
In thin-film sensors the film thickness is typically 5–500 nm, in special cases even 1 μm. The most
common preparation process is thin-film technology based on sputtering, evaporation, chemical vapor
deposition, spray pyrolysis, chemical deposition, or laser ablation methods of forming the sensing layer.
Sensing layers in thin-film sensors have more dense structure in comparison with thick-film sensors.
At present, for conductometric sensors, mainly oxide semiconductors—electron or mixed con-
ductors—are used as base materials (Korotcenkov 2007a). Polymers, thin-film metals, and the carbon
nanotubes, which exhibit changes in their conductivity when exposed to an analyte, also can be used for
fabricating such sensors. Extensive literature is available for a variety of polymers (Harris et al. 1997;
Wang et al. 1997; Su and Tsai 2004), metal oxides (Patel et al. 1997; Atashbar et al. 1998; Tan et al. 2003;
Yu et al. 2005), and mixed oxides (Moon et al. 2002). Metals are normally substituted into polymers, and
the choice of metal governs the sensitivity of the conductometric sensors. Metal-substituted phthalo-
cyanines are commonly used due to their stability toward oxidation at high temperatures (Wohltjen et al.
1985). Middlehoek and Audet (1989) used Cu-substituted phthalocyanine conductive polymer as a CCl4
sensor. Previous studies used sublimation techniques to deposit thin films for chemical vapor sensing.
These devices, however, had problems due to poor control of film thickness, size, and morphology
(Van Oirschot et al. 1972; Bott and Jones 1984). To overcome these problems, Wohltjen and co-workers
(1985) deposited monolayers of Cu-substituted phthalocyanine on interdigitated Au electrodes using
Langmuir-Blodgett methods to sense ammonia and NO2. Ultrathin metallic films have been used suc-
cessfully for detecting variuous species. Platinum thin films were employed by Majoo and Patel for
hydrogen gas sensing (Majoo et al. 1996; Patel et al. 1999). A variety of metallic nanowires has also been
reported for gas sensing, such as palladium nanowires for hydrogen sensing (Yang et al. 2009; Atashbar
and Singamaneni 2005) and silver nanowires for amine vapor (Murry et al. 2005).
The most accepted mechanism, explaining sensitivity of n-type metal oxide–based sensors, in-
cludes consideration of the role of the chemisorbed oxygen (Barsan et al., 1999; Gurlo, 2006). Oxygen
chemisorption means the formation of O2−, O−, O2− species on the surface. Among these, O− proved
to be more reactive than O2−, while O2− is not stable. So the dominant species is the O− species. The
oxygen chemisorption results in a modification of the space charge region toward depletion. The re-
sistance corresponding to this state is considered the base resistance. The appearance of a reducing
gas leads to partial consumption of the adsorbed oxygen, resulting in a decrease in resistance, while
the appearance of oxygen increases the surface oxygen coverage, and hence the resistance. The above
mechanisms suggest the existence of a grain boundary (connected to the modification of the space
charge region (see Figure 1.17). The relationship between the amount of change in resistance to the
concentration of a combustible gas can be expressed by a power-law equation (Barsan et al. 1999):
1
G= = kC n (1.21)
R
where C is the concentration of analyte and k and n are individual constants, which depend on the
mechanism of sensitivity and must be determined empirically by calibration.
To activate reactions of oxygen chemisorptions and surface catalysis, high temperature (>200°C) is
required. For these purposes, metal-oxide gas sensors have incorporated heaters, which are electrically
isolated from the sensing layer. One possible variant of conductometric sensor configuration is shown
in Figure 1.18.
Figure 1.17. Schematic representation of the basic steps in the detection of gas molecules us-
ing conductometric metal-oxide gas sensors. Surface and bulk reactions lead to changes of the
overall DC or AC conductance, which may include frequency-dependent contributions from (1) an
undoped or doped surface, (2) bulk, (3) three-phase boundary or contacts, and (4) grain boundaries.
Equivalent circuits with different resistance/capacitance (RC) units describe the frequency behavior
formally. Each unit corresponds to a characteristic charge carrier transport. (Reprinted with permis-
sion from Gopel and Schierbaum 1995. Copyright 1995 Elsevier.)
Figure 1.18. Layout of planar alumina substrate with Pt electrodes and Pt heater. The SnO2 layer is
printed on top of the interdigitated electrodes. The heater on the back keeps the sensor at the opera-
tional temperature. (Reprinted with permission from Barsan and Weimar 2003. Copyright 2003 IOP.)
The sensitivity of conductometric metal-oxide sensors can be improved using different approaches
based on the tools of materials engineering: bulk doping, surface modification, structure optimization,
etc. (Korotcenkov 2005). The selectivity can be improved by selecting or preparing proper materials
(Korotcenkov 2007a). The incorporation of a catalyst results in improvement of surface reaction se-
lectivity. An applied filter selectively removes the undesired component(s). However, the filter saturates
and requires regeneration. To improve the stability, a materials science approach, i.e., the use of chemi-
cally stable materials, seems to be the most efficient.
The main considerations in the fabrication of a stable conductometric sensor are careful selection
of the sensitive layer, film-coating technique, bias potential, and measurement scheme. The sensitive
layer should have affinity to the target measurand, and the film coating technique needs to be compat-
ible with the device fabrication process. In addition, its chemical composition, thickness, and morphol-
ogy should be reproducible. A thorough review paper has been published by Korotcenkov (2008) on
the influence of morphology and crystallographic structure of metal oxides and characteristics such as
thickness, grain size, agglomeration, porosity, faceting, grain network, surface geometry, and film tex-
ture on conductometric gas sensors. With regard to the measurement scheme, Harris and co-workers
(1997) presented signal-to-noise ratio (SNR) analysis for conductivity measurements of thin films de-
posited between microelectrodes. They compared the AC and DC excitation of the sensor, and their
noise measurements revealed that passage of DC current generates a high level of 1/ƒ noise charac-
teristic. Remarkable improvement in SNR can be obtained by applying low-frequency (a few kilohertz)
excitation. They also found that the conduction mechanism is nonohmic at frequencies below 1 MHz,
but considerably more ohmic at frequencies of tens of MHz. In the case of an AC excitation, however,
the distinction between the conductivity and capacitive measurements becomes challenging, since the
sensitive layers have both resistive and capacitive elements (Kovacs 1998). Wohltjen and co-workers
(1985) also noted that a differential DC conductivity measurement can be made to reduce temperature-
dependent errors by keeping the reference sensor isolated from the analyte.
The major advantages of metal oxides are reversibility, rapid response, longevity, low cost, and
robustness. The materials and manufacturing techniques easily lend themselves to batch fabrication.
Unfortunately, several disadvantages limit more widespread sensor utilization and continue to challenge
sensor development. The metal-oxide sensor generally requires a high temperature for measurable gas
response, often leading to high power consumption. Metal-oxide sensors are not highly selective, and
much effort has been involved in devising materials and methods of operation to improve specificity.
These sensors are generally not appropriate for applications in vacuum or inert atmospheres. Baseline
sensor drift may require periodic calibration for certain applications.
Unlike the metal-oxide gas sensors, devices based on polymers can operate at room temperature. A
more versatile sensing system that may work at room temperature is based on a carbon black–polymer
blend, in which the carbon particles provide electrical conductivity and the polymer (any polymer) pro-
vides the sensor function. The sensor response is a result of the swelling of the polymer, which causes
the conductivity of the sensor to change. The amount of swelling corresponds to the concentration of
the chemical vapor in contact with the absorbent. Weak van der Waals forces between the polymer and
the target gas molecules are responsible for the swelling of the sensing layer; therefore, the change is
purely physical and is reversible, which makes the sensor reusable. Some hysteresis can occur when the
sensor is exposed to high concentrations.
Polymer-based conductometric gas sensors have better selectivity. So far, however, problems such
as low stability, high sensitivity to humidity, and degradation under UV radiation and during interaction
with ozone have not been resolved for polymers.
10. ACOUSTIC WAVE SENSORS
Acoustic wave–based sensors offer a simple, direct, and sensitive method for probing the chemical and
physical properties of materials. Acoustic wave sensors use piezoelectric material as substrates to gener-
ate acoustic waves. The type of acoustic wave generated in a piezoelectric material depends mainly on
the substrate material properties, the crystal cut, and the structure of the electrodes utilized to trans-
form the electrical energy into mechanical energy.
Acoustic wave sensors have been broadly used for many applications in detecting chemical and
biological components in either gas or liquid media. To monitor a specific gas or vapor, a sensitive layer
is generally employed. The utilization of acoustic wave devices for chemical sensing applications relies
on their sensitivity toward small changes (perturbations) occurring at the “active” surface. The interface
selectively adsorbs target molecules in the analyte to the surface of the sensing area. This adsorption
results in changes in the electrical and mechanical properties of the device (acoustic wave–based gas
and vapor sensors measure the change in the mass at the surface or in the sheet conductivity of the
sensitive layer), which in turn causes changes in the acoustic wave characteristics, namely, amplitude,
velocity, and the resonance frequency of the device. In the presence of an analyte species, the waves’
properties become perturbed in a measurable way that can be correlated to the analyte concentration
(Ippolito et al. 2009).
It has been established that acoustic-wave-based sensors have high sensitivity toward surface
perturbations and exhibit good linearity with low hysteresis. They are also typically small, relatively
inexpensive, and inherently capable of measuring a wide variety of input quantities. They offer a simple
means of analyte monitoring, with several advantages over other solid-state sensors (Ippolito et al.
2009). These include well-established fabrication processes, chemical inertness of the substrate materi-
als, and high structural rigidity. Additionally, the microelectronics industry has developed numerous
piezoelectric transducer platforms, which in turn has widely encouraged the development and improve-
ment of various acoustic wave–based chemical sensors.
However, it is necessary to note that, similar to other solid-state chemical sensors, with acoustic
wave–based sensors it is inherently difficult to satisfy the requirements for all sensing applications with a
single sensor design. Therefore, in most cases, the selection of materials and structural parameters results
in design trade-offs among device sensitivity, long-term stability, selectivity, and operating temperature.
Acoustic wave devices come in a number of configurations, each with its own distinct acoustic and
electrical characteristics. Major classes of acoustic devices for sensor applications are thickness shear
mode, Rayleigh and shear horizontal surface acoustic wave, acoustic plate mode, Love wave mode, and
flexural plate wave devices. Several different groups of acoustic wave devices that are commonly em-
ployed for gas sensing will be discussed in the following.
10.1. THICKNESS SHEAR MODE SENSORS
The thickness shear mode (TSM) resonators were first studied and used by Sauerbrey (1959) to deter-
mine the thickness of thin layers adhering to the surface of the device. Over time, the use of TSM reso-
nators has been extended for detection of various chemical and biological species by coating the surface
of the device with a suitable sensitive layer. These devices are also referred to as quartz crystal microbal-
ances (QCMs) and consist of an AT-cut quartz crystal with thin-film metal electrodes patterned on both
sides forming a single electrical port (typical electrode materials are gold and aluminum). These devices
fall into the category of bulk wave acoustic sensors, as application of a voltage produces an electric
field throughout the bulk of the device. The piezoelectric nature of quartz in turn causes this electric
field to produce a shearing motion transverse to the normal of the crystal. The resonant frequency
of the device is inversely proportional to the thickness of the crystal (which is half a wavelength) and
is typically in the range of 5–15 MHz. QCMs have been used extensively for thickness monitoring in
metal evaporation (McCallum 1989); protein sensing (Barnes et al. 1992; Imai et al. 1994; Atashbar et al.
2005), and gravimetric immunoassays (Muramatsu et al. 1987; Caruso et al. 1996; Anzai et al. 1996; Sakti
et al. 2001). One of the advantages the QCM is the small temperature dependence of the operating
frequency. The electronic circuitry and the measurements associated with QCMs are relatively simple
and straightforward. Further, these devices have a very high quality factor. The most serious limitation
of these devices is their low sensitivity, which is limited by the thickness of the crystal. Although the
device’s sensitivity can be increased by decreasing the crystal thickness, it makes the crystal much more
fragile. Figure 1.19 shows top and side views of a typical QCM crystal. Useful reviews of TSM devices
have been provided by Lu and co-workers (1984) and by Benes and co-workers (1995).
If we assume that the adsorbed mass is of infinitely small thickness with negligible phase change
across the film thickness and there is no acoustic energy dissipation, then it is easy to derive the Sauerbrey
equation (Cheeke and Wang 1999):

Gold (Upper electrode)
Quartz
Gold (Lower electrode)
(a) (b)
Figure 1.19. (a) Top view of a QCM crystal. (b) Side view of a QCM crystal.
Δf − 2 f 0 Δm S − 2 Δm S
= = (1.22)
f0 μ q ρq ρq λ 0
where f0 is the resonance frequency, μq is the substrate shear stiffness, ρq is the substrate mass density,
λ0 is the acoustic wavelength of the resonator, and mS is the area mass loading on the surface of the
resonator. Hence, a measurement of the resonant frequency will give a measure of mS. In practice, the
device must be calibrated.
More sensors designed to respond to different gas components are made by coating the metallizing
layers with an additional acceptor layer, which can absorb the sample component. This acceptor layer
should interact with the desired component as selectively as possible, without being changed by other
parameters. If an uncoated QCM with a fundamental frequency of fa is coated with a acceptor layer, the
fundamental frequency of the QCM will drop to fb . When this coated QCM is then exposed to a vapor,
some of the vapor will be absorbed into the acceptor layer, changing the mass of the crystal and hence
its fundamental frequency.
Ideal coatings for QCM applications should be nonvolatile so that the coating stays on the crys-
tal surface and allows rapid and easy diffusion of vapors into and out of the material. The material
should be stable over prolonged use and not undergo any hysteresis effects. Selectivities of QCM sen-
sors differ according to their acceptor layer. For instance, palladium is highly selective for hydrogen,
organic amines are only weakly selective for sulfur dioxide, and carbowaxes (commonly used as the
stationary phase in chromatography) interact with numerous nonpolar gases, hence they are nonselec-
tive sensors for hydrocarbons. We need to note that polymers are the most widely usable coatings for
QCM-based sensors.
QCM sensors are fast, cheap, reliable sensors that can be used for a variety of applications ranging
from solvent detection to food quality analysis. They are inexpensive because they are mass produced
for oscillator circuits, and their lifetime is approximately 1 year. They have good linearity and can be
incorporated into an array and used to build up a library of characteristic responses for individual
analytes, allowing for rapid and reliable identification of unknown compounds. They can be used for
early-warning detection of possible contaminants in many industrial processes, such as beer and to-
bacco production.
Application of piezoelectric sensors in liquids is much more difficult than working with gases, since
the surrounding liquid phase acts strongly, damping the oscillating crystal. Although the problems have
been solved, real analytical applications are rarely encountered even today. The best-investigated applica-
tions are antibody layers immobilized on quartz crystals, which are used in immunoassays. Reaction with
the corresponding antigen is highly selective, and the resulting mass change is measurable. Piezoelectric
sensors in liquid phase do not obey the Sauerbrey equation.
10.2. SURFACE ACOUSTIC WAVE SENSORS

Surface acoustic wave (SAW) or Rayleigh sensors consist of a piezoelectric substrate with electrodes
patterned on the surface in the form of interdigital transducers (IDTs), as shown in Figure 1.20.
This is the most commonly utilized structure for gas-sensing applications, with the sensitive layer
normally deposited in between the two IDT ports. However, if the layer is of low conductivity, it may
be deposited over the entire device. This configuration generates the Rayleigh-mode SAW, which pre-
dominantly has two particle displacement components in the sagittal plane. Surface particles move in
elliptical paths with a surface-normal and a surface-parallel component. The surface-parallel component
is parallel to the direction of propagation. The electromagnetic field associated with the acoustic wave
travels in the same direction. The wave velocity is determined by the substrate material and the cut of
the crystal. Most of the energy of the SAW is confined to a few wavelengths depth of the substrate
(Wohltjen 1984). The acoustic wavelength is determined by the IDT periodicity. These sensors are
normally operated in a “delay line” configuration, with one set of IDTs acting as a transmitter and the
other set acting as a receiver. These devices work at higher frequencies (a typical frequency of operation
being 30–500 MHz) and in comparison to TSM resonators have higher sensitivity, because the wave
propagates not through the bulk of the device but near the surface. The use of Rayleigh SAW sensors
is applicable only to gas media, as the Rayleigh wave is severely attenuated in liquid media (Calabrese
et al. 1987).
The basic principle in a SAW-based gas sensor is the detection of small deviations in acoustic
wave propagation characteristics, which are caused by perturbations on the active surface of the device.
The most common technique is to measure the deviation in propagation velocity of the surface wave
as a change in resonant frequency of the device. However, perturbations that affect acoustic phase
Wave propagation
Sensitive film
IDT IDT
Piezoelectric substrate
Figure 1.20. Schematic view of a SAW sensor.

velocity can be caused by many factors, each of which represents a potential sensor response (Ricco and
Martin 1991): T (temperature), ε (permittivity), E (electric field), σ (electrical conductivity), c (stiffness),
η (viscosity), ρ (density), p (pressure), μ (shear elastic modulus), m (mass). These interactions change
the boundary conditions, producing a measurable shift in the propagating SAW mode’s phase velocity.
Equation (1.23) illustrates the change in acoustic phase velocity as a result of possible external perturba-
tions, under the assumption that the perturbations are small and linearly combined:
Δv 1 ⎛ δv δv δv δv δv δv ⎞
≅ ⎜ ΔT + Δε + ΔE + Δσ + Δm + Δp + …⎟ (1.23)
v 0 VP ⎝ δT δε δE δσ δm δp ⎠
Since acoustic wave devices use piezoelectric materials for the excitation and detection of acoustic
waves, the nature of almost all the parameters involved in sensor applications concerns either mechani-
cal or electrical perturbations. Furthermore, the temperature dependence of each parameter and the
overall temperature coeffcient of the structure must also be considered. Therefore, the sensor response
may be due to a combination of the above parameters. However, it is necessary to note that during
chemical sensing the change of mass and conductivity of functionalizing (sensing) layer has the stron-
gest influence.
In the simplest case, mass loading (ML) of adsorbed species reduces the phase velocity in the sub-
strate, leading to a reduction in frequency so that
Δv
= −c m f 0 ·Δm S (1.24)
v0
where cm is a mass sensitivity factor.

Some excellent articles describing the application of SAW sensors for the measurement of param-
eters such as temperature, pressure, electrical conductivity, mass, and viscoelastic changes of thin films
have been reported in a number of reviews (D’Amico and Verona 1989; McCallum 1989; Ballantine
and Wohltjen 1989; Khlebarov et al. 1992; Grate et al. 1993; Caliendo et al. 1997; Ballantine et al. 1999).
SAW sensors have found extensive use in gas sensing (Atashbar and Wlodarski 1997; Varghese et al.
2003; Fontecha et al. 2004; Shen et al. 2004).
SAW devices are well established in the electronics industry and have been employed for gas-
sensing applications for approximately 30 years. They can be designed to offer high sensitivity with a
reasonably large dynamic range, good linearity, and low hysteresis. This makes them extremely suitable
for sensing applications at parts-per-million (ppm) and parts-per-billion (ppb) concentrations (Ippolito
et al. 2009). When a good coating is available, it is usually possible to detect species at concentration
levels of 10–100 ppb. The mass detection limit is in the range of 0.05 pg/mm2.
Many SAW sensor applications use a dual-delay-line configuration; one delay line can be coated
with the sorptive or reactive film, while the other remains inert or protected from environmental effects.
Typically, the frequency difference is measured, which is of the order of kilohertz and can be easily
sampled. The uncoated or protected resonator acts as a reference to compensate undesired effects, for
instance, frequency fluctuations caused by temperature or pressure changes.
10.2.1. Shear-Horizontal Surface Acoustic Wave Sensors

Shear-horizontal surface acoustic wave (SH-SAW) sensors belong to the category of SAW sensors in
which the application of a potential to the IDTs produces an acoustic wave having particle displacement
perpendicular to the direction of wave motion and in the plane of the crystal (shear displacement). As
in SAW devices, the acoustic wavelength is determined by the transducer periodicity. An example of
a piezoelectric substrate that supports SH-SAW is lithium tantalate (LiTaO3), in which the dominant
acoustic wave propagating on 36º rotated Ycut, X propagating LiTaO3 is a shear horizontal mode. In
the SH-SAW, the acoustic wave propagation is not severely attenuated when the surface is loaded with
a liquid, as is the case with the Rayleigh-mode SAW-based devices. SH-SAW sensors have been applied
to a number of enzyme-detecting (Kondoh et al. 1994) and liquid-sensing applications (Nomura et al.
2003; Martin et al. 2004).
10.2.2. Shear-Horizontal Acoustic Plate Mode Sensors

Shear-horizontal acoustic plate mode (SH-APM) sensors are essentially Rayleigh sensors with the dif-
ference that their substrate thickness is of the order of a few acoustic wavelengths (Déjous et al. 1995;
Esteban et al. 2000). Under such conditions, IDTs in addition to the Rayleigh waves also generate
the shear horizontal waves. These waves are not confined to the surface only but travel through the
bulk, being reflected between the top and the bottom surfaces of the substrate which now acts as an
acoustic waveguide (see Figure 1.21). The particle displacement for SH-APM devices is parallel to the
surface (in plane) and transverse to the direction of wave propagation. The Rayleigh waves typically
propagate at much lower velocities than those of the SH-APM modes. The frequency of operation
is determined by the material properties, the thickness-to-wavelength ratio of the substrate, and the
IDT finger spacing.
SH-APM devices are used mainly for liquid sensing and offer the advantage of using the back
surface of the plate as the active sensing area. Consequently, the IDTs can be isolated from the liquid
media, thereby preventing direct chemical attack on the IDTs. The interfaces where the wave interacts
with liquid and air should be polished and smooth, since any irregularities will produce additional noise
signals (Gardner et al. 2001). Considerable research has been devoted to developing various sensing
applications. Applications such as vapor and viscosity measurements (Ricco et al. 1988; Esteban et al.
Input IDT Output IDT

Particle displacement
Wave propagation
Piezoelectric plate and reflections
Figure 1.21. Schematic view of a SH-APM device structure.

2000), identification via the electrical surface perturbation (Kelkar et al. 1991; Dahint et al. 1993), and
biosensing (Andle et al. 1995; Déjous et al. 1995) have been reported.
10.2.3. Love-Mode Acoustic Wave Sensors

Love wave devices are characterized by acoustic waves that propagate in a layered structure consisting
of a piezoelectric substrate and a guiding layer (Bender et al. 2000; Kalantar-zadeh et al. 2001) (Figure
1.22). They have a pure shear polarization, with the particle displacements perpendicular to the nor-
mal of the surface plane. An essential condition for existence of Love waves is that the shear acoustic
velocity in the guiding layer is less than the shear acoustic velocity in the substrate (Haueis et al. 1994;
White et al. 1997). Under such conditions, the elastic waves generated in the substrate are coupled to
this surface guiding layer, which traps the acoustic energy near the surface of the device and thereby
increases the sensitivity of the device to mass loadings.
The mass sensitivity of a Love-mode sensor is determined mainly by the thickness of the guiding
layer and its operating frequency. Hence the guiding layer should be one with low density and low shear
velocity. The guiding layer also serves to passivate the IDTs from the contacting liquid (Josse et al. 2001).
The most commonly used substrates for SH-SAW sensors and Love-mode devices are quartz, lithium
niobate (LiNbO3), and lithium tantalate (LiTaO3 ). The guiding layer may be SiO2 (Harding 2001), certain
polymers such as poly(methyl methacrylate), also known as PMMA (Gizelli 2000), or ZnO (Kalantar-
zadeh et al. 2001). Top and side views of a Love-wave sensor are shown in Figure 1.22.
Love mode sensors are advantageous for gravimetric sensing applications such as organic vapors
absorbed by a thin polymer film (Jakoby et al. 1999).
10.2.4. Lamb Wave Sensors

Lamb wave sensors are essentially Rayleigh wave sensors in which the waves propagate in a thin plate
(membrane), which has a thickness less than an acoustic wavelength. Particle displacement is transverse
(a)
Interdigital Transducers Guiding Layer Substrate
(b)
Figure 1.22. (a) Top view of Love wave sensor. (b) Side view of a Love wave sensor.
to the direction of wave propagation and is parallel and normal to the plane of the surface. Application
of a potential across the electrodes gives rise to both symmetric and antisymmetric plate modes. The
lowest-order antisymmetric mode (A0 ) of the Lamb wave is known as the flexural plate wave (FPW),
and it is this mode that is used in liquid sensing applications (Wang et al. 1998). The IDTs form a delay
line for launching the flexural plate waves that set the whole membrane into motion. The wave veloci-
ties and hence the frequency of operation of these devices therefore depends on the plate material and
its thickness. Figure 1.23 shows top and side views of a flexural plate wave sensor. These FPW devices
have been investigated for applications such as chemical gas sensors (Wenzel and White 1989; Grate et
al. 1991), pressure sensors (Tirole et al. 1993, Choujaa et al. 1995), and for the determination of liquid
properties (Vellekoop et al. 1994). Development of a FPW-based biosensor has also been investigated
(Andle and Vetelino 1995; Wang et al. 1997).
11. MASS-SENSITIVE SENSORS

Mass-sensitive chemical sensors are the simplest form of gravimetric sensors that respond to the mass
of species attached to the sensing layer (Ballantine et al. 1999). In principle, any species that can be
immobilized on the sensor can be sensed. The mass changes can be monitored by either deflecting a
micromechanical structure due to stress changes or mass loading, or by assessing the frequency char-
acteristics of a resonating structure or a traveling acoustic wave upon mass loading. As we know, the
change of the resonant frequency f is given by (Cimalla et al. 2007)
(a)
Particle
displacement
(b)
Input IDT Output IDT
Thin piezoelectric
plate
Wave
propagation
Figure 1.23. (a) Top view of a FPW sensor. (b) Side view of a FPW sensor.
1 Δm
Δf = − f (1.25)
2 m
Among the different geometric shapes of mass-sensitive components, clamped-free beams, also
called cantilevers, represent the simplest microelectromechanical system (MEMS). Although cantile-
vers play a role as a basic building block for complex MEMS devices, they are also widely used as
force-sensor probes in atomic force microscopy (AFM). Recent developments in AFM have enabled
mass fabrication of micrometer-sized cantilevers with various geometric shapes and made of various
materials (silicon, silicon nitride, etc.). The free-standing cantilever beam end is coated with a sensitive
layer (see Figure 1.24). The detection principle of such devices is based on measurement of the change
in cantilever deflection or the change in resonance frequency of the cantilever (Lavrik et al. 2004).
These changes are induced by the adsorption of chemical species on the functionalized surface of a
microcantilever (Chen 1995). Micromachined cantilever arrays have already proven their suitability as
chemical and biological sensors for gas- (Battiston et al. 2001; Kim et al. 2001; Lange et al. 2001; Raiteri
et al. 2001; Moulin et al. 2003; Then et al. 2006) and liquid-phase interactions (Raiteri et al. 2002).
The sensitivity of such sensors to additional mass loadings can be increased by reducing the inher-
ent active mass of the resonator, which is the reason for the ongoing research toward nano-electro-
mechanical systems (NEMS). The detection of monolayers of molecules, single viruses, and carbon
particles are outstanding results that have been demonstrated recently (Cimalla et al. 2007).
At present, the measurement of microcantilever deflection with subangstrom resolution can be
organized into two approaches: optical and electrical (Raiteri et al. 2001). The most widely used method
in AFM is the so-called beam bounce or optical lever technique (see Figure 1.25). The visible light from
a low-power laser diode is focused on the free apex of the cantilever, which acts as a mirror. The reflec-
Sensitive
chemical Layer
(a)
Reacting chemical species
(b)
Figure 1.24. Schematic view of the working principle of a microcantilever-based chemical sensor:
The mass deposition that follows the chemical reaction causes a stress that leads to a detectable
bending of the cantilever.
Figure 1.25. Optical lever deflection detection method. (Reprinted with permission from Raiteri et al.
2001. Copyright 2001 Elsevier.)
tion beam hits a position-sensitive photodetector or a split photodetector, as illustrated schematically in

Figure 1.25. When the cantilever bends, the reflected laser light moves along the photodetector surface.
The distance traveled by the laser beam is proportional to the cantilever deflection and linearity magni-
fied by the cantilever–photodetector distance as the arm of a lever.
The interference of a reference laser beam with the one reflected by the cantilever is another
optical deflection method that has been implemented. Intrerferometry is highly sensitive and provides
a direct and absolute measurement of the displacement. The capacitance method (electrical method)
measures displacement as a change in the capacitance of a plane capacitor. This technique is highly
sensitive and can provide absolute displacement.
12. OPTICAL SENSORS

Optical sensors can be used for detection and determination of chemical or physical parameters by
measuring changes in optical properties. Techniques for using optical chemical sensors have been a
growing research area over the last three decades. Two basic operation principles are utilized for opti-
cally sensing chemical species (Baldini et al. 2006):
• An intrinsic optical property of the analyte is utilized for its detection.

• Indicator (or label) sensing is used when the analyte has no intrinsic optical property. For exam-
ple, pH is measured optically by immobilizing a pH indicator on a solid support and observing
changes in the absorption or fluorescence of the indicator as the pH of the sample varies with
time.
The first principle is used mainly in infrared and ultraviolet spectroscopes. The most widely used
techniques employed in optical chemical sensors are optical absorption, fluorescence, and chemilumi-
nescence. In this case, optical chemical sensors detect the intensity of photon radiation that arrives at a
sensor. Sensors based on other optical parameter spectroscopies, such as refractive index and reflectiv-
ity, have also been well developed.
Absorption is based on the transmitted light remaining after incident light is passed through a
medium. The medium can be a gas, a liquid, or a solid. There are no perfectly transparent materials that
would allow light or electromagnetic energy to pass through them without any change. Some energy
must be absorbed, depending on the physical properties of the transmitting medium. Variations in
light absorption can be due to the vibrational and rotational movements for different chemical bonds
in the molecules in the light path which absorb energy at different wavelengths. For example, water is
fairly transparent to visible and ultraviolet light, but it strongly absorbs infrared radiation and begins to
absorb in the far-ultraviolet range. The optical properties of a liquid medium can also be affected by the
concentration (C, moles per liter) of other chemical solutes that are dissolved in it. The Lambert-Beer
law is often used to characterize the intensity (I ) of light transmitted through a uniform, homogenous
medium as a function of the incident light (I0). The intensity is given by
I = I0·exp(−kĆ·Δx) (1.26)
where k´ is an extinction coefficient and Δx is the thickness of the medium. Light scattering in the
sample is not considered.
Fluorescent chemical sensors occupy a prominent place among the optical devices because of their
superb sensitivity, combined with the required selectivity that photo- or chemiluminescence imparts to
the electronic excitation. This is due to the fact that the excitation and emission wavelengths can be se-
lected from those of the absorption and luminescence bands of the luminophore molecule (the emitted
photon has a fluorescence peak at longer wavelength than the absorption peak); in addition, the emis-
sion kinetics and anisotropy features of the latter add specificity to luminescent measurements (Baldini
et al. 2006). One can simply measure the intrinsic fluorescence of the target analyte or design sensors
based on the variation of the fluorescence of an indicator dye with the determinand concentration. In
the latter case, the probe molecule can be immobilized onto a (thin) polymer support (sometimes even
the waveguide surface itself) and placed at the distal end or in the evanescent field of an optical fiber or
integrated optics sensor.
Provided it is optically diluted, the relationship between a luminophore (“luminescence bearer”)
concentration and the intensity of its emission is a linear one:
IL = ФLI0κελ·lc (1.27)
where ФL is the luminescence quantum yield (i.e., the ratio of emitted to absorbed photons per unit of
time, an intrinsic feature of the luminophore), I0 is the intensity of the excitation light, κ is an instru-
mental parameter (related to the geometry of the luminescent sample, the source, and the detector, i.e.,
the emission collection efficiency), ελ is the luminophore absorption coefficient at the excitation wave-
length (in dm3·mol–1·cm–1), l is the optical path length (in cm), and c is the luminophore concentration
(in mol·dm–3). Such relationship helps us to understand potential problems that may arise when design-
ing and operating optical sensors based on polymer-supported luminescent probes (Lakowicz 1994).
Both absorption and fluorescence signals need an excitation light source, so they are not as com-
patible with other miniaturized solid-state chemical sensors. Chemiluminescence does not need an
excitation light source. Chemiluminescence occurs in the course of some chemical reactions when an
electronically excited state is generated. Chemiluminescence measurements consist of monitoring the
rate of production of photons, and thus, the light intensity depends on the rate of the luminescent
reaction. Consequently, light intensity is directly proportional to the concentration of a limiting reactant
involved in a luminescence reaction. With modern instrumentation, light can be measured at a very
low level, and this has allowed the development of very sensitive analytical methods based on these
light-emitting reactions (Baldini et al. 2006). The first chemiluminescence sensor for hydrogen peroxide
analysis was reported by Freeman and Seitz in 1978.
In many cases, optical sensors are based on optical fiber technology and planar waveguide con-
figurations (Figure 1.26). Generally, an optical fiber consists of a fiber as the core surrounded by a
cladding layer and a light-impermeable jacket. The refractive index of the core (n1 ) is always higher
than that of the cladding (n2 ) layer. As a result, light beams are reflected toward the inside of the guide
by internal total reflection. A planar waveguide sensor can consist of thin planar layers instead of opti-
cal fibers. The principle of measurement is shown in Figure 1.26b. The evanescent field is the basis
of such sensors. There are different methods for inputting light into the active layer. Commonly, light
is coupled to the device by means of prisms (prism coupler sensors). Alternatively, the light can be
coupled by miniature diffraction gratings at the sensor ends (grating coupler sensors, GCS). With grat-
ing couplers, preferably, laser beams are used which are coupled to the planar medium by diffraction
at the grating with a precisely defined angle of incidence. Light is totally reflected numerous times at
the interfaces as a result of the small thickness of the active layer. This means an extraordinarily in-
tensive interaction with the interface, i.e., with the sample medium. The reason is the energy quantum
which is transferred at each reflection from the extension of the evanescent field across the interface
(Gründler 2006).
Increasing flexibility in measurement modes, such as evanescent wave, surface reflectance, and
surface plasmon reflectance, has resulted in their rapid development. Recent developments in the field
have been driven by factors such as the availability of low-cost miniature optoelectronic components
(light sources, waveguides, and detectors). Additional driving factors are the need for multianalyte
array-based sensors, particularly in the area of biosensing, as well as advances in microfluidics and
imaging technology. Portable UV/visible, Raman, and fluorescence instruments are now available for
field measurements.

(a) (b)
Figure 1.26. Basic setup of (a) cylindrical light guides, e.g., fiber optic guides. (b) Planar waveguide
configuration.
12.1. FIBER OPTIC CHEMICAL SENSORS

Fiber optic sensors are a class of sensors that use optical fibers to detect chemical contaminants. Light
is generated by a light source and is sent through an optical fiber. The light then returns through the
optical fiber and is captured by a photodetector. According to Rogers and Poziomek (1996), fiber optic
sensor configurations fall into three main groups: end-of-fiber, side-of-fiber, and porous or interrupted
fiber configurations (see Figure 1.27). For end-of-fiber sensors, the optical fiber acts as a conduit to
carry light to and from the sample. This method involves sending a light source directly through the
optical fiber and analyzing the light that is reflected or emitted by the contaminant. The refractive index
of the material at the tip of the optical fiber can be used to determine what phases (vapor, water, or any
solution) are present. The modulation of intensity for a given range of wavelengths is dependent on
the absorbance or fluorescence of the analyte, indicator, or analyte–indicator complex. The indicator
compound can be trapped behind a membrane, in a polymer, or covalently immobilized to the end of
the fiber. The side-of-fiber configuration typically relies on the use of the evanescent wave. This effect
occurs when light is propagated down an optical waveguide. An electromagnetic wave is generated at
the fiber surface and decays exponentially into the medium surrounding the fiber. This evanescent zone,
Figure 1.27. Fiber optic sensor configurations. (Adapted with permission from Rogers and Poziomek
which is usually limited to less than 100 nm, can be used to detect the presence of optical indicators or
changes in refractive index at the surface of the unclad fiber. In the case of fluorescent indicators, the
method is very sensitive.
Cladding-based chemical fiber-optic sensors are examples of the side-of-fiber configuration.
These sensors can be made using microporous or other types of sensitive claddings (Lieberman 1993).
Evanescent wave refractometric cladding–based sensors detect the absorption of the species in the
polymeric cladding, which leads to a variation of its refractive index and thus to a variation of the over-
all transmission efficiency of the fiber. Porous polymer cladding sensors can be used, for instance, for
humidity measurements (Ogawa et al. 1988). The quantity of moisture absorbed by the microporous
cladding varies with humidity. The optical power level of the transmitted light in the core varies accord-
ing to the moisture absorption because of the change in the refractive index of the cladding. Another
example is the application of organopolysiloxanes as cladding materials, which can change refractive
index when they adsorb and/or absorb molecules from the surrounding medium (Harsanyi 2000).
For porous or interrupted fiber configurations, indicator chemistry is typically incorporated directly
into the structure. For several reasons, this configuration has the potential to be extremely versatile. For
example, because of the large surface area provided by the porous fiber core, this method is particularly
well suited for absorbance measurements. Further, because the porous regions are intrinsically coupled
with the fiber (i.e., they are part of the fiber), measurements can be made at multiple locations along
a single fiber. Porous fibers may exhibit very high gas permeability and liquid impermeability, so they
can be used for the detection of gases in liquids. Vapors permeating into the porous zone can produce
a spectral change in transmission. For better sensitivity and selectivity, colorimetric reagents can be
trapped in the pores.
However, we need to note that, in general, optical fiber sensors are usually classified as intrinsic
and extrinsic types. In the extrinsic type, the optical fiber is used only as a means of light transport to an
external sensing system, i.e., the fiber structure is not modified in any way for the sensing function.
Regarding intrinsic fiber optic chemical sensors, one can note that there are four general sensor
designs: fiber refractometers, evanescent spectroscopy, active coating, and active core. Peculiarities of
the design of these sensors are discussed in detail by Baldini et al. (2006). Coating-based sensors are the
largest class of intrinsic fiber optic chemical sensors; in this design, a small section of the optical fiber
passive cladding is replaced by an active coating. The analyte reacts with the coating to change the opti-
cal properties of the coating—refractive index, absorbance, fluorescence, etc.—and is then coupled to
the core to change the transmission through the optical fiber.
The basic fiber optic chemical sensor (optode) design consists of a source fiber and a receiver fiber
connected to a third optical fiber by a special connector as shown in Figure 1.28. The tip of the third
fiber is coated with a sensitive material, usually by a dip coating procedure. The chemical to be sensed
may interact with the sensitive tip by changing the absorption, reflection, scattering properties, lumines-
cence intensity, refractive index, or polarization behavior, hence changing the reflected light properties.
The fiber in this case acts as a light pipe transporting light to and from the sensing region.
Commonly, the following advantages of optodes are cited (Wolfbeis 1990):
• Optodes do not require a “reference electrode.” There is no electric circuit to close, hence it is
not necessary to work with two probes.
Figure 1.28. Fiber optic optode for chemical sensing.
• Optodes can be miniaturized easily.

• Optical fibers with high transparency allow signal transmission over long distances (up to 1000
m) without loss of quality. This characteristic is valuable in particular for long-distance measure-
ments when the instrument cannot be placed near the sensor.
• The sensor signal is primarily optical in nature and consequently insufficient to block electrical
interference from the environment. Such interference may provoke serious problems for electro-
chemical sensors.
• Materials for optical sensors, such as fused silica, are highly inert and may remain in contact with
aggressive media for a long time.
The main problem associated with this design is that the area of interaction between the chemical
and the material is very small (Cimalla et al. 2007), i.e., 8–10 μm in diameter in the case of a single-mode
fiber and 50–200 μm in diameter in the case of a multimode fiber, which directly affects the sensitivity
achieved using this type of sensor. In addition, ambient light may distort the measurement, and some
of the reagents immobilized at the sensor surface, mainly dyes, are unstable and can be bleached by UV
radiation or washed out by solvents.
Fiber optic chemical sensors can also be categorized based on the signal detection mechanism
employed. These include fluorescence, absorption (colorimetric and spectroscopic), and refractometry.
12.2. FLUORESCENCE FIBER OPTIC CHEMICAL SENSORS

In order to transform a standard optical fiber into an intrinsically fluorescence-based optical chemical
sensor, it is necessary to impart analyte-sensitive fluorescence to the fiber in some manner. This may
be achieved by replacing the cladding of the fiber over a portion of its length with a solid matrix that
contains a fluorescent compound. This process involves removing a portion of the original cladding
of the fiber and coating the decladded region with a sensing material, which is subsequently cured to
form a solid fluorescent cladding. A variation of the fluorescent cladding configuration was employed
by Ahmad et al. (2005), who described a system based on two “positive” fibers attached to either end
of a “negative” fiber. Park et al. (2005) described development of fiber optic sensors for detection of
inter- and intracellular dissolved oxygen. Preejith et al. (2006) realized a tapered fluorescent fiber optic
evanescent wave—based sensor for serum protein.
12.3. ABSORPTION FIBER OPTIC CHEMICAL SENSORS
Absorption-based optical sensors can be colorimetric or spectroscopic in nature. Colorimetric sensors,

as the name suggests, are based on detection of an analyte-induced color change in the sensor material,
while spectroscopic absorption–based sensors rely on detection of the analyte by probing its intrinsic
molecular absorption. The sensors that utilize these techniques have been modified to facilitate the
interaction of the light guided within the fiber with either the color-changing indicator or the analyte
itself, depending on the nature of the sensor. In fact, the modified cladding configuration described
above has been used extensively in a number of modes for development of absorption-based sensors.
Examples include sensors for ammonia (Moreno et al. 2005), pH (Gupta and Sharma 2002; James and
Tatam, 2006), and ethanol (King et al. 2004).
12.4. REFRACTOMETRIC FIBER OPTIC CHEMICAL SENSORS
A range of refractometric optical sensors has been developed in the past decade, which involves ad-
dition of refractive index–sensitive optical structures to the optical fiber. Such structures include fiber
Bragg gratings. Fiber optic chemical sensors based on fiber Bragg gratings detect changes in effective
refractive index and require decladding of the fiber section bearing the fiber Bragg gratings in order to
impart sensitivity toward the refractive index of the environment in which the fiber is placed. Changes
in the intensity of reflected light may accurately represent physical as well as chemical events that occur.
It is possible to have an evanescent wave that penetrates only a very short distance (<100 nm) into the
medium, allowing study of localized surface phenomena. In other applications, it may be desirable to
have the light penetrate as deeply as possible into the medium.
Examples of refractometric-based sensors include sugar and propylene detection in glycol solu-
tions (Sang et al. 2006), detection of analytes such as sodium chloride (Falciai et al. 2001), ethylene
glycol, antibodies (DeLisa et al. 2000), dissolved ammonia (Pisco et al. 2006), organic solvents (Liu et al.
2006), and a humidity sensor for breath monitoring (Kang et al. 2006).
12.5. ABSORPTION-BASED SENSORS
12.5.1. Infrared and Near-Infrared Absorption

For many applications, infrared (IR) and near-infrared (NIR) spectroscopic detection has been a reliable
method of detecting chemical species. The growing interest in IR and NIR spectroscopy is probably a
direct result of its major advantages over other analytical techniques, namely, easy sample preparation
without any pretreatments, the possibility of separating the sample measurement position and the spec-
trometer by use of fiber optic probes, and the prediction of chemical and physical sample parameters
from a single spectrum (Reich 2005), which means that a single spectrum allows several analytes to be
determined simultaneously. Moreover, gas absorption spectroscopy enables identification and quantifi-
cation of chemicals in liquid and gas mixtures with little interference from other gases.
IR and NIR measure the vibrational transitions of molecular bonds. In the mid-IR region of the
spectrum, the incident radiation excites the fundamental transitions between the ground state of a
vibrational mode and its first excited state. A broad range of molecules, including most organic com-
pounds, absorb in the IR and NIR range, and therefore IR and NIR spectroscopy is a valuable analytical
tool in the quantitative and qualitative analysis of organic matter. Typical infrared spectra are shown in
Figure 1.29.
NIR absorption bands are typically broad, overlapping, and 10–100 times weaker than their cor-
responding fundamental mid-IR absorption bands. These characteristics severely restrict sensitivity in
the classical spectroscopic sense and call for chemometric data processing to relate spectral information
to sample properties. The low absorption coefficient, however, permits high penetration depth and,
thus, an adjustment of sample thickness. This aspect is actually an analytical advantage, since it allows
direct analysis of strongly absorbing and even highly scattering samples, such as turbid liquids or solids
in either transmittance or reflectance mode without further pretreatments.
The most prominent absorption bands occurring in the NIR region are related to overtones and
combinations of fundamental vibrations of –CH, –NH, –OH (and –SH) functional groups, character-
istic of organic matter (Reich 2005). Therefore, IR and NIR spectroscopy have been widely used for
detection of gases, for example, CO2 (Zhang et al. 2000), CO (Zhang and Wu 2004), NO2 (Ashizawa et
al. 2003), and CH4 (Chévrier et al. 1995).
In its simplest form, the technique involves confining a sample of target material (gas) in an opti-
cal absorption cell and measuring its absorption at specific IR wavelengths, which are characteristic of
the vibrational modes of the molecule. The system components usually include an IR source, a mono-
chromator (or optical filters to select specific absorption wavelengths), a sample holder or a sample
Figure 1.29. Infrared spectra of stearic acid in a basalt matrix: A, basalt powder with 50 ppm stearic
acid; B, basalt powder with 50 ppm stearic acid and reference basalt spectrum subtracted; C, refer-
ence spectrum of stearic acid; D, samples acquired at four-wavenumber resolution with a Biorad FTS
6000 FTS and a Pike diffuse reflectance attachment, with KBr as background and offset and scaled
for clarity. (Reprinted with permission from Anderson et al. 2005. Copyright 2005 American Institute
of Physics.)
Figure 1.30. Basic near-infrared spectrometer configuration. (Reprinted with permission from Reich
presentation interface, and a detector, allowing for transmittance or reflectance measurements at the
wavelength of interest (see Figure 1.30). The light source is usually a tungsten halogen lamp, since it is
small and rugged (Reich 2005).
The appropriate NIR measuring mode will be dictated by the optical properties of the sample.
Transparent materials are usually measured in transmittance. When, for example, gas in the path ab-
sorbs energy from the source, the detector receives less radiation than without the gas present, and the
detector can quantify the difference. Turbid liquids or semisolids and solids may be measured in diffuse
transmittance, diffuse reflectance, or transflectance, depending on their absorption and scattering char-
acteristics (Reich 2005). Qualitative analysis involves comparison of the sample spectrum with spectral
libraries set up previously for known samples of known products. Sophisticated software assists in spec-
tral matching, utilizing preset threshold values which provide stringent identification and discrimination
criteria (Rinnan and Rinnan 2007).
As we indicated earlier, NIR spectroscopy requires no sample preparation and no reagents. This
has helped the development of quartz fiber optic–based devices that allow the spectra of samples with
different characteristics to be recorded simply by selecting the most suitable mode for each (usually,
reflectance for solids, transmittance for liquids, and transflectance for emulsions and turbid liquids).
The ability to perform field measurements instead of having to collect samples for subsequent
analysis in the laboratory is another advantage of NIR spectroscopy. Some NIR spectrophotometers
can make measurements on site. The miniaturization of optical components has boosted development
of portable NIR spectrophotometers. Currently available models (www.asdi.com) using such optical de-
vices include hand-held instruments and equipment that can be carried in a backpack or mounted on a
vehicle. Moreover, because water shows relatively weak absorption in the NIR region and NIR radiation
is relatively low-energy, NIR spectroscopy enables noninvasive analysis. Therefore, NIR spectroscopy
has become a widely used analytical method in the agricultural, pharmaceutical, chemical, medical, and
petrochemical industries (Siesler et al. 2002; Ciurczak and Drennen, 2002).
However, it is necessary to know that IR sensors can only monitor specific gases that have nonlin-
ear molecules. In addition, they can be affected by humidity and water; they can be expensive; and dust
and dirt can coat the optics and impair response, which is a concern in in-situ environments.
12.5.2. UV Absorption
Many molecules absorb ultraviolet or visible light of different wavelengths. Therefore, these molecules
can be detected using UV spectroscopy. When a molecule absorbs a photon from the UV/visible re-
gion, the corresponding energy is captured by one (or several) of its outermost electrons. Ultraviolet-
visible spectroscopy (UV = 200–400 nm, visible = 400–800 nm) corresponds to electronic excitations
between the energy levels that correspond to the molecular orbitals of the systems. In particular, transi-
tions involving π orbitals and ion pairs (n = nonbonding) are important, and so UV/visible (UV/Vis)
spectroscopy is used mostly for identifying conjugated systems which tend to have stronger absorptions
(Baldini et al. 2006). Absorbance is directly proportional to the path length and the concentration of the
absorbing species. This means that an absorption spectrum can show a number of absorption bands
corresponding to structural groups within the molecule (Rouessac and Rouessac 2007).
Modern absorption instruments can usually display the data as transmittance, percent transmit-
tance, or absorbance. The transmittance (T ) is a measure of the attenuation of a beam of monochro-
matic light based on comparison between the intensities of the transmitted light (I ) and the incident
light (I0) according to whether the sample is placed, or not, in the optical pathway between the source
and the detector. T is expressed as a fraction or a percentage:
I
T= T (1.28)
I0
Absorbance (formerly called optical density) is defined by
A = −log T (1.29)
For compounds of simple atomic composition, provided the spectrometer possesses high enough
resolution, the fundamental transitions appear as if isolated. In extreme situations the positions of the
absorptions are recorded in cm−1 (wavenumber), a unit better adapted for precise pointing than nm.
An unknown concentration of an analyte can be determined by measuring the amount of light that a
sample absorbs and applying Beer’s law (Baldini et al. 2006), which is presented here in its current form:
A = ελlC (1.30)
where A is the absorbance, an optical parameter without dimension that is accessible with a spectro-
photometer, l is the thickness (in cm) of the solution through which the incident light is passed, C is the
molar concentration, and ελ is the molar absorption coefficient (L mol−1 cm−1) at wavelength λ, at which
the measurement is made. If the absorptivity coefficient is not known, the unknown concentration can
be determined using a working curve of absorbance versus concentration derived from standards. The
light source is usually a hydrogen or deuterium lamp for UV measurements and a tungsten lamp for
visible measurements. The wavelengths of these continuous light sources are selected with a wavelength
separator such as a prism or grating monochromator.
At present, direct UV absorption sensing has been used mainly in environmental applications to
monitor pollutants such as heavy metals, hydrocarbons, and volatile organic compounds (VOCs) in air
and water. Examples include a sensor to detect Cr in water (Tau and Sarma 2005), ozone and NO2 in
the atmosphere (Wu et al. 2006), and VOCs and aromatic hydrocarbons in ambient air in the vicinity of
a refinery (Lin et al. 2004).
12.6. SURFACE PLASMON RESONANCE SENSORS

Over the past two decades, the surface plasmon resonance (SPR) technique has been developed for
detection of chemical and biological species. In principle, SPR sensors are thin-film refractometers that
measure changes in the refractive index occurring at the surface of a metal film supporting a surface
plasmon. Surface plasmon resonance is a charge-density oscillation that may exist at the interface of two
media with dielectric constants of opposite signs, for instance, a metal and a dielectric. The resonance
condition depends on the wavelength, the angle at which the light strikes the substrate–metal interface,
and the dielectric constants of all the materials involved (Karlsen et al. 1995; Bardin et al. 2002). It was
established that at optical wavelengths, this condition is fulfilled by several metals, of which gold and
silver are the most commonly used. The charge-density wave is associated with an electromagnetic
wave, the field vectors of which reach their maxima at the interface and decay evanescently into both
media. This surface plasma wave (SPW) is a TM-polarized wave (the magnetic vector is perpendicular
to the direction of propagation of the SPW and parallel to the plane of the interface). The propagation
constant of the surface plasma wave propagating at the interface between a semi-infinite dielectric and
a metal is given by the following expression (Homola et al. 1999):
ε m n S2
β=k (1.31)
ε m + n s2
where k denotes the free-space wave number, εm is the dielectric constant of the metal, and ns is the
refractive index of the dielectric. A surface plasmon excited by a light wave propagates along the metal
film, and its evanescent field probes the medium (sample) in contact with the metal film. A change in
the refractive index of the dielectric gives rise to a change in the propagation constant of the surface
plasmon, which, through the coupling conditions, alters the characteristics of the light wave coupled to
the surface plasmon (e.g., coupling angle, coupling wavelength, intensity, phase).
So, the operational principle of SPR is based on an interface between two transparent media with
different refractive indices. The light coming from the side of higher refractive index is partly reflected
and partly refracted. Above a certain critical angle of incidence, no light is refracted across the interface,
and total internal reflection is observed. While incident light is totally reflected, the electromagnetic
field component penetrates a short distance (tens of nanometers) into the medium of a lower refrac-
tive index, creating an exponentially evanescent wave (Homola et al. 1999). If the interface between the
media is coated with a thin layer of metal (e.g., gold), and the light is monochromatic and p-polarized
(the electric field vector component is parallel to the plane of incidence), then the energy carried by
photons of the light can be coupled to the electrons of the metal film. Energy transfer takes place
when there is a match (resonance) between the incident light energy photons and the electrons at the
metal film surface. The coupling occurs at a particular wavelength which is dependent on the metal
and the environment of the illuminated metal surface. The coupling phenomenon results in creation of
plasmons, which are a group of highly excited electrons that behave like a single electrical entity. All the
light except that at the resonant wavelength is absorbed. This beam loses most of its energy to the surface
plasmon wave, and this is called surface plasmon resonance. This wavelength modulation depends on the
magnitude of chemical changes. The material adsorbed onto the thin metal film influences the resonance
conditions (Pockrand et al. 1979; Jorgenson and Yee 1993). Generally, a linear relationship is found be-
tween resonance energy and mass concentration of (bio)chemically relevant molecules such as proteins,
sugars, and DNA. The SPR signal, which is expressed in resonance units, is therefore a measure of mass
concentration at the sensor chip surface. This means that the analyte and ligand association and disso-
ciation can be observed, and ultimately, rate constants as well as equilibrium constants can be extracted.
These SPR biosensors have a wide variety of applications; the most popular commercial instruments for
SPR biosensing are Biocore instruments.
Generally, an SPR optical sensor comprises an optical system, a transducing medium which inter-
relates the optical and (bio)chemical domains, and an electronic system that supports the optoelectronic
components of the sensor and allows data processing. The transducing medium transforms changes in the
quantity of interest into changes in the refractive index, which may be determined by optically interrogat-
ing the SPR. The optical part of the SPR sensor contains a source of optical radiation and an optical struc-
ture in which SPW is excited and interrogated. In the process of interrogating the SPR, an electronic signal
is generated and processed by the electronic system. Major properties of an SPR sensor are determined by
properties of the sensor’s subsystems. The sensor sensitivity, stability, and resolution depend on properties
of both the optical system and the transducing medium. The selectivity and the response time of the sen-
sor are determined primarily by the properties of the transducing medium (Homola et al. 1999).
This enhancement and subsequent coupling between light and a surface plasmon are performed in
a coupling device (coupler). The most common couplers used in SPR sensors include a prism coupler,
a waveguide coupler, and a grating coupler (Figure 1.31). Prism couplers represent the most frequently
used means of optical excitation of surface plasmons.

(a) (b)
Figure 1.31. Most widely used configurations for SPR sensors: (a) prism coupler–based SPR sys-
tem (ATR method); (b) grating coupler–based SPR system. (Reprinted with permission from Homola
2008. Copyright 2008 American Chemical Society.)
Considerable literature is available on the principle of SPR sensing and various configurations
(Homola et al. 1999). Applications such as hydrogen sensing (Benson et al. 1999), methylene sensing
(Ideta and Arakawa 1993), sensing of hydrocarbons such as ethane and methane (Urashi and Arakawa
2001), and nitrogen dioxide sensing (Wright et al. 1995), to name a few, have been reported.
13. PHOTOACOUSTIC SENSORS

The generation of acoustic waves by modulated optical radiation is called the photoacoustic effect. This phe-
nomenon was discovered in 1880 by A. G. Bell. Bell found that sound could be generated by exposing
optically absorbing materials to chopped sunlight. Today’s photoacoustic gas sensors utilize absorption
of light by a certain species. The absorbed light leads to an increase in the temperature of the sample
and the surrounding gas and results in thermal expansion of the latter. If the light source is modulated
with a certain frequency, then the thermal expansion will also be modulated and form a sound wave
proportional to the absorbed energy. To improve the detection limit, in most cases the acoustical signal
is further amplified by an acoustic resonator (Keller et al. 2005).
There are excellent review articles describing the principle and application of photoacoustic sen-
sors (Miklós et al. 2001; Elia et al. 2006) for the measurement of gases. Photoacoustic sensors have
been used for sensing a wide range of species, including H2O, NH3 (Paldus 1999), NH3, CO2, CH3OH
(Hofstetter et al. 2001), O3 (Da Silva et al. 2004), NO (Elia et al. 2005), HMDS (Elia et al. 2006), and
smoke (Keller et al. 2005).
14. THERMOELECTRIC SENSORS

The thermoelectric heat power transducer for calorimetric chemical sensors is gaining more attention
as an alternative to conductometric gas sensors. The thermoelectric effect, also known as the Seebeck effect,
is the conversion of temperature differences directly into electricity. For small temperature differences,
the relation can be approximated as follows (Göpel et al. 1994):
US = (αA − αB)(T1 − T0) = αA/B(T1 − T0) (1.32)
where αA and αB denote the Seebeck coefficients (thermopower or thermoelectric power) of materials
A and B. They express specific transport properties determined by the band structure and the carrier
transport mechanisms of the materials. Us is always created in an electrically conducting material when
a temperature gradient is maintained along the sample, but it cannot be observed with two legs of the
same material, for reasons of symmetry.
The use of the Seebeck effect to detect thermal responses from calorimetric gas sensors was first
reported by McAleer and co-workers (1985). In their work, the thermoelectric sensors consisted of
monolithic structures in which a catalyst layer was coated on one side of a tin dioxide pellet. Seebeck
voltage was measured when hydrogen gas reacted with oxygen and generated heat. To achieve se-
lectivity and reversibility, these devices required temperatures in excess of 240°C. More recently, a
number of researchers have reported planar devices, in both thick tin oxide (Širok 1993; Ionescu
1996) and thin indium oxide film forms (Papadopoulos et al. 1996; Vlachos et al. 1997). Although
these devices are undoubtedly better than monolithic devices in terems of sensor size, sensitivity, and
in some cases selectivity, they all require temperatures in excess of 240°C for the reversible detection
of trace combustibles.
Thermoelectric metal oxide sensors fabricated using thick-film technology, as well as Si or SiGe
thin-film thermopile sensors prepared by conventional CMOS processes, provide, for example, for
hydrogen, detection limits in the low parts-per-million range. Because of the small thermal inertia of
silicon thermopile sensors and the inherent reference temperature compensation effects, they can be
operated in a high-speed temperature scanning regime up to 600°C and with cycle times of a few milli-
seconds (Schreiter et al. 2006). With respect to differences in the activation energies of the involved gas
components, the shape of the signals is a characteristic of the investigated mixture; i.e., the application
of pattern-recognition methods should lead to enhanced selectivity of the sensor.
A thermoelectric sensor that operated at room temperature was also reported by Markinkowska et
al. (1991). Their devices were coated with hydrophobic oxidation catalyst granules, with the other mate-
rial surrounded by a bed of noncatalytic hydrophobic catalyst support material.
Pyroelectricity is related to piezoelectricity. Sometimes, both effects can be encountered in the same ma-
terial. One could imagine that a volume increase of a piezo crystal at a higher temperature should result
in effects similar to those caused by mechanical compression or expansion. Spontaneous polarization
will be a strong function of temperature, since the atomic dipole moments vary as the crystal expands
or contracts. Heating the crystal tends to desorb the surface neutralizing ions as well as change the po-
larization, so that a surface charge may then be detected. Thus, the crystal appears to have been charged
by heating. This is called the pyroelectric effect.
The electric field developed across a pyroelectric crystal can be remarkably large when it is sub-
jected to a small change in temperature. We define a pyroelectric coefficient, p, as the change in flux
density (D) in the crystal due to a change in temperature (T ) (Anderson et al. 1990):
δD
p= (1.33)
δT
Using a capacitor, the pyroelectric voltage signal ΔU is
ΔT
ΔU = pd (1.34)
εr ε 0
where εr ε0 is the permittivity and d is the thickness of the pyroelectric film.

The performance characteristics of the sensors depend on the sensing material, the preparation
and geometry of the electrodes, the use of absorbing coatings, the thermal design of the structure,
and the nature of the electronic interface as well. Since the output is dependent on the rate of change,
the radiant flux incident on the sensor must be chopped, pulsed, or otherwise modulated. Important
parameters of the materials are the heat capacity per volume, the heat conductivity, and the penetra-
tion depth of the temperature wave in the pyroelectric material at a given modulation frequency
(Harsanyi, 2000).
Figure 1.32. Schematic of pyroelectric sensor layout. (Reprinted with permission from Schreiter
et al. 2006. Copyright 2006 Elsevier.)
One example of a pyroelectric sensor is shown in Figure 1.32. The array is formed by a thin-film
capacitor deposited on a supporting SiO2/Si3N4 membrane. The dielectric is made of a thin film of
ferroelectric lead zirconate titanate (PZT). To provide sufficiently high thermal isolation, the underlying
silicon is removed by bulk micromachining. To enable elevated working temperatures, a local low-power
thin-film heater arranged on bulk-micromachined membranes and integrable with the pyroelectric sen-
sor structure was developed and tested. The functionalization of active elements of the pyroelectric de-
tector array was accomplished by coating them with a chemically sensitive layer of polydimethylsiloxan
(PDMS) for the detection of VOCs or with Pt-cluster arrays grown on bacterial surface layers (S-layers)
serving as catalyst for hydrogen oxidation. Another commonly used pyroelectric material is lithium tan-
talate (LiTaO3 ). For use in sensors, the material is doped with lanthanum traces. Polyvinylidene fluoride
(PVDF) is a synthetic organic pyroelectric polymer.
15. THERMAL CONDUCTIVITY SENSORS

Thermal conduction is one form of thermal energy transfer from one object to another. Physical con-
tact between two bodies is required for heat conduction. Kinetic energy is transferred to a cooler body
from a warmer body by thermally agitating its particles. As a result, the cooler body gains heat while
the warmer body loses heat. For instance, heat passage through a rod is governed by a law similar to
Ohm’s law; the heat flow rate is proportional to the thermal gradient across the material (dT/dx) and
its cross-sectional area (A), or
dQ dT
H= = −kA (1.35)
dT dx
where k is called the thermal conductivity. The minus sign indicates that heat flows in the direction of
temperature decrease. A good thermal conductor has a high value of thermal conductivity, whereas
thermal insulators have low values of k.
The general principle of thermal conductivity sensors is as follows. A known temperature differ-
ence is maintained between a “cold” element and a “hot” element (Devoret et al. 1980). Heat is trans-
ferred from the “hot” element to the “cold” element via thermal conduction through the investigated
gas. The power needed to heat the “hot” element is therefore a direct measure of the thermal conduc-
tivity. Heat loss due to radiation, convection, and heat conduction through the suspensions of the “hot”
element need to be minimized by sensor design. The precise mechanism is quite complex, because the
thermal conductivity of gases varies with temperature and convection as well as conductivity. Most
gases produce a linear output signal, but not all.
In the simplest case, power loss of a single filament thermistor by heat conduction via the ambient
gas can be expressed as
P = kTC λ·ΔT (1.36)
where P is the power dissipation of the heater by thermal conduction of the gas, λ is the thermal con-
ductivity of the given gas–gas mixture, kTC is a constant that is characteristic for a given geometry, and
ΔT is the temperature difference between the heater and the ambient gas.
In the sensor, a heated thermistor or platinum filament is mounted so that it is exposed to the
sample. Another element, which acts as a reference, is enclosed in a sealed compartment. If the sample
gas or vapor has a thermal conductivity higher than the reference, heat is lost from the exposed element
and its temperature decreases, whereas if its thermal conductivity is lower than that of the reference, the
temperature of the exposed element increases. These temperature changes alter the electrical resistance,
which is measured using a bridge circuit.
Figure 1.33 shows a schematic diagram of a thermistor-based sensor which can be used to mea-
sure humidity, using the thermal conductivity of gas (Miura 1985). Because of the amount of power
required to heat these sensors, they have to be mounted in a flameproof enclosure, in the same way as
pellistors are. Two tiny thermistors (T1 and T2 ) are supported by thin wires to minimize thermal conduc-
tivity loss to the housing. Small venting holes are used to expose thermistor T1 to the outside gas, and
thermistor T2 is sealed in dry air. Both thermistors are connected into a bridge circuit (R1 and R2 ), which
is powered by voltage +E. Due to the passage of electric current, the thermistors develop self-heating.
Their temperatures rise up to 170ºC over the ambient temperature. To establish a zero reference point
initially, the bridge is balanced in dry air. As absolute humidity rises from zero, the output of the sensor
gradually increases.
It is necessary to know that the presence of gases that have thermal conductivities relative to
air of >1 leads to cooling of the exposed thermistor or filament. These gases are often measured by
the TC technique—the higher their thermal conductivity, the lower the concentration which can be
measured. Gases with thermal conductivities of <1 are more difficult to measure, partly because water
vapor may cause interference. Gases with thermal conductivities close to 1 cannot be measured by this
technique. These include ammonia, carbon monoxide, nitric oxide, oxygen, and nitrogen (Barsony et
al. 2009).
Thermometric and calorimetric sensors are discussed here only for their use for gases. In the liquid
phase, the conditions are much less advantageous, since the thermal conductivities of liquids are orders
of magnitude higher than those of gases. Calorimetric measurements must be performed in closed
+E
R0
Dry Air
R1 Venting Holes
R2
T1 T2 VOUT
Figure 1.33. Schematic diagram of a thermistor-based humidity sensor with self-heating

thermistors.
vessels with thermal isolation, or at least in a flowing stream. In some applications, thermometric sen-
sors are used as detectors for flow injection analysis.
16. FLAME IONIZATION SENSORS

The presence of ions in flames has long been known, and a large volume of literature exists on the sub-
ject of flame ionization. The literature includes descriptions of the mechanisms of ion formation and
the electrical properties of the flame. There are thorough review articles (Calcote 1957; Fialkov 1997)
on this subject, and it is widely accepted that the key mechanism enabling the flow of electrical current
through hydrocarbon flames results from the chemi-ionization of the formyl radical, CHO*
CH + O → CHO* → CHO+ + e– (1.37)
Flame ionization–based sensors are used in gas chromatography (Morgan 1961). The first flame
ionization detector was developed more than five decades ago by Commonwealth Scientific and
Industrial Research Organization (CSIRO) scientists in Melbourne, Australia. A typical flame ionization
sensing technique consists of at least two electrodes arranged so that a voltage potential can be applied
across the flame, or a part of the flame. The electrical properties of the flame facilitate a measurable
current that is related to a parameter of interest. For example, a typical flame ionization detector (FID)
used in a hydrocarbon analyzer has electrodes arranged so that a voltage potential is applied across the
entire flame, and the amount of electrical current flow through the flame is linearly proportional to the
hydrocarbon concentration (Cheng et al. 1998).
Most commonly, flame ionization–based sensors are attached to a gas chromatography system.
Figure 1.34 shows a general schematic diagram of a of a flame ionization sensor. The gas enters the
detector’s oven through the gas chromotography column, where it is mixed with the hydrogen fuel and
then with the oxidant as it moves up. The effluent/fuel/oxidant mixture continues to travel up to the
nozzle head, where a positive bias voltage exists. This positive bias helps to repel the reduced carbon
ions created by the flame, thus pyrolyzing the eluent. The ions are then repelled up toward the collec-
tor plates, which are connected to a very sensitive ammeter, which detects the ions hitting the plates
and then feeds that signal to an amplifier, integrator, and display system. The products of the flame are
finally vented out of the detector through the exhaust port.
17. LANGMUIR-BLODGETT FILM SENSORS

Langmuir-Blodgett sensors are named for Irving Langmuir and Katherine Blodgett, who discovered
unique properties of thin films in the early 1900s. Langmuir’s original work involved the transfer of
monolayers from liquid to solid substrates. Blodgett expanded on Langmuir’s research to include the
deposition of multilayer films on solid substrates (Blodgett 1935). The structure of the film can be
controlled at the molecular level by transferring monolayers of organic material from a liquid to a solid
substrate. These Langmuir-Blodgett films exhibit various electrochemical and photochemical properties.
Figure 1.34. Schematic diagram of a flame ionization sensor.

Langmuir-Blodgett (LB) films basically consist of molecules which are amphiphilic in nature, i.e.,
they have a hydrophilic (polar) head group that can be bonded to various solid surfaces and a long
hydrophobic (nonpolar) tail that extends outward (Grandke and Ko 1996). Amphiphilic substances are
generally saturated fatty acids such as palmitic acid, magaric acid, stearic acid, arachidic acid, etc. The
head group is a carboxylic acid and the tail is a straight-chain hydrocarbon. When these amphiphilic
molecules come in contact with water, they accumulate at the air/water interface, causing a decrease in
the surface tension of the water and exhibiting a tendency to orient with the head group into the water
and the tail out of the water (see Figure 1.35b). The hydrophilic forces of the head group and the hydro-
phobic forces of the tail group allow the material to form oriented monolayers. The material dissolves
in water if the head group is too strongly attracted to the water, and there is no layer formation if the tail
group is too hydrophobic. LB films can be deposited on a substrate by dipping it up and down through
the monolayer while simultaneously keeping the surface pressure constant. Multilayers can be formed
by successive deposition of these monolayers on the same substrate. Figure 1.35c shows a schematic
representation of the deposition of the floating monolayer on a solid substrate.
Comprehensive information about the properties and methods of preparation of LB films can
be found in the literature; for example, see Roberts (1990). LB films have been widely studied, because
the end of the tail group can be functionalized to form precisely arranged molecular arrays for various
applications ranging from simple ultrathin insulators and lubricants to complex chemical and biologi-
cal sensors. LB films have been deposited on conductometric, capacitive, optoelectric, and FET-based
Figure 1.35. Schematic representation of (a) an amphiphilic component, (b) the orientation of am-
phiphilic molecules, with their head groups on the water surface and their tails sticking up, and (c)
deposition of the floating monolayer on a solid substrate.
devices for sensing applications. Review papers describe the fundamental scheme of LB films for chem-
ical sensors, fields of major interest in chemical sensors, and future trends in LB sensor studies (Roberts
1983; Bubeck 1988; Moriizumi 1988; Barraud 1990; Izumi et al. 1998).
REFERENCES
Ahn S., Kulis D.M., Erdner D.L., Anderson D.M., and Walt D.R. (2006). Fiber-optic microarray for simultaneous
detection of multiple harmful algal bloom species. Appl. Environ. Microbiol. 72, 5742–5749.
Ahmad M., Chang K.P., King T.A., and Hench L.L. (2005) A compact fibre-based fluorescence sensor. Sens.
Actuators A 119, 84–89.
Ali M.B., Kalfat R., Sfihi H., Chovelon J.M., Ouada H.B., and Jaffrezic-Renault N. (2000) Sensitive cyclodextrin-
polysiloxane gel membrane on EIS structure and ISFET for heavy metal ion detection. Sens. Actuators B 62,
233–237.
Anderson J.C., Leaver K.D., Rawlings R.D., and Alexander J.M. (1990) Materials Science. Chapman & Hall, London.
Anderson M.S., Andringa J.M., Carlson R.W., Conrad P., Hartford W., Shafer M., Soto A., Tsapin A.I., Dybwad J.P.,
Wadsworth W., and Hand K. (2005) Fourier transform infrared spectroscopy for Mars science. Rev. Sci. Instrum.
76, 034101.
Andle J.C. and Vetelino J.F. (1995) Acoustic wave biosensors. Sens. Actuators A 44, 167–176.
Andle J.C., Weaver J.T., Vetelino J.F., and McAllister D.J. (1995) Selective acoustic plate mode DNA sensor. Sens.
Actuators B 24–25, 129–133.
Anh D.T.V., Olthuis W., and Bergveld P. (2004) Work function characterization of electroactive materials using an
EMOSFET. IEEE Sensors J. 4, 284–287.
Anzai J., Guo B., and Osa T. (1996) Quartz-crystal microbalance and cyclic voltammetric studies of the adsorption
behaviour of serum albumin on self-assembled thiol monolayers possessing different hydrophobicity and
polarity. Bioelectrochem. Bioenerg. 40, 35–40.
Ashizawa H., Yamaguchi S., Endo M., Ohara S., Takahashi M, Nanri K., and Fujioka T. (2003) development of a
nitrogen dioxide gas sensor based on mid-infrared absorption spectroscopy. Rev. Laser Eng. 31, 151–155.
Atashbar M.Z., Bejcek B., Vijh A., and Singamaneni S. (2005) QCM biosensor with ultra thin polymer film. Sens.
Actuators B 107, 945–951.
Atashbar M.Z., Sun H.T., Gong B., Wlodarski W., and Lamb R. (1998) XPS study of Nb-doped oxygen sensing
TiO2 thin films prepared by sol-gel method. Thin Solid Films 326, 238–244.
Atashbar M.Z. and Wlodarski W. (1997) Design, simulation and fabrication of doped TiO2-coated surface acoustic
wave oxygen sensor. J. Intell. Mater. Syst. Struct. 8, 953–959.
Atashbar M.Z., Sadek A.Z., Wlodarski W., Sriram S., Bhaskaran M., Cheng C.J., Kaner R.B., and Kalantar-zadeh K.
(2009) Layered SAW gas sensor based on CSA synthesized polyaniline nanofiber on AlN on 64 degrees YX
LiNbO3 for H2 sensing. Sens. Actuators B 138, 85–89.
Apostolakis J.C., Georgiou C.A., and Koupparis M.A. (1991) Use of ion-selective electrodes in kinetic flow in-
jection: determination of phenolic and hydrazino drugs with 1-Fluoro-2,4-dinitrobenzene using a fluoride-
selective electrode. Analyst 116, 233–237.
Baldini F., Chester A.N., Homola J., and Martellucci S. (eds.) (2006) Optical Chemical Sensors. Springer-Verlag,
Dordrecht, The Netherlands.
Balkus K.J., Ball L.J., Gnade B.E., and Anthony J.M. (1997) A capacitive type chemical sensor based on AlPO4-5
molecular sieves. Chem. Mater. 9, 380–386.
Ballantine D.S., White R.M., Martin S.J., Ricco A.J., Frye G.C., Zellers E.T., and Wohltjen H. (1997) Acoustic Wave
Sensors: Theory, Design, and Physico-Chemical Applications. Academic Press, San Diego, CA.
Ballantine D.S. Jr. and Wohltjen H. (1989) Surface acoustic wave devices for chemical analysis. Anal. Chem. 61,
704–712.
Bardin F., Kašik I., Trouillet A., Matějec A., Gagnaire H., and Chomát M. (2002) Surface plasmon resonance sensor
using an optical fiber with an inverted graded-index profile. Appl. Opt. 41, 2514–2520.
Barnes C., D’Silva C., Jones J.P., and Lewis T.J. (1992) Lectin coated piezoelectric crystal biosensors. Sens. Actuators
B 7, 347–350.
Barraud A. (1990) Chemical sensors based on LB films. Vacuum 41, 1624–1628.
Barsan N., Schweizer-Berberich M., and Göpel W. (1999) Fundamental and practical aspects in the design of na-
noscaled SnO2 gas sensors: a status report. Fresenius J. Anal. Chem. 365, 287–304.
Barsan N. and Weimar U. (2003) Understanding the fundamental principles of metal oxide based gas sensors; the ex-
ample of CO sensing with SnO2 sensors in the presence of humidity. J. Phys.: Condens. Matter 15, R813–R829.
Bausells J., Carrabina J., Errachid A., and Merlos A. (1999) Ion-sensitive field-effect transistors fabricated in a com-
mercial CMOS technology. Sens. Actuators B 57, 56–62.
Benammar M. (2004) Design and assembly of miniature zirconia oxygen sensors. IEEE Sensors J. 4(1), 3–8.
Battiston F.M., Ramseyer J.-P., Lang H.P., Baller M.K., Gerber C., Gimzewski J.K., Meyer E., and Güntherodt H.-J.
(2001) A chemical sensor based on a microfabricated cantilever array with simultaneous resonance-frequency
and bending readout. Sens. Actuators B 77, 122–131.
Barsony I., Ducso C., and Furjes P. (2009) Thermometric gas sensing. In: Comini E., Faglia G., and Sberveglieri G.
(eds.), Solid State Sensing. Science-Business Media, chap. 7.
Bell A.G. (1880) On the production and reproduction of sound by light: the photophone. Am. J. Sci. 3(20),
305–324.
Bender F., Cernosek R.W., and Josse F. (2000) Love wave biosensors using cross-linked polymer waveguide on
LiTaO3 substrates. Electron. Lett. 36(19), 1–2.
Benes E., Gröschl M., Burger W., and Schmid M. (1995) Sensors based on piezoelectric resonators. Sens. Actuators A
48, 1–21.
Benson D.K., Tracy C.E., Hishmeh G.A., Ciszek P.A., Lee S.-H., Pitts R., and Haberman D.P. (1999) Low cost
fiber-optic chemochromic hydrogen gas detector. In: Proceedings of the 1999 US DOE Hydrogen Program Review,
NREL/CP-570-26938.
Bergstrom P.L., Patel S.V., Schwank J.W., and Wise K.D. (1997) A micromachined surface work-function gas sensor
for low-pressure oxygen detection. Sens. Actuators B 42, 195–204.
Bergveld P. (1970) Development of an ion-sensitive solid-state device for neurophysiological measurements. IEEE
Trans. Biomed. Eng. 17, 70–71.
Bergveld P. (1991) Future applications of ISFETs. Sens. Actuators B 4, 125–133.
Bergveld P. (2003) Thirty years of ISFETOLOGY What happened in the past 30 years and what may happen in
the next 30 years. Sens. Actuators B 88, 1–20.
Blodgett K.B. (1935) Films built by depositing successive monomolecular layers on a solid surface. J. Am. Chem.
Soc., 57, 1007–1022.
Bott B. and Jones T.A. (1984) A highly sensitive NO2 sensor based on electrical conductivity changes in phthalo-
cyanine films. Sens. Actuators 5, 43–52.
Bowden M., Song L.N., and Walt D.R. (2005) Development of a microfluidic platform with an optical imaging
microarray capable of attomolar target DNA detection. Anal. Chem. 77, 5583–5588.
Brzozka Z., Dawgul M., Pijanowska D., and Trobicz W. (1997) Durable NH4+ sensitive chemFET. Sens. Actuators
B 44, 527–531.
Bubeck C. (1988) Reactions in monolayers and Langmuir-Blodgett films. Thin Solid Films 160(1–2), 1–14.
Calabrese G., Wohltjen H., and Roy M.K. (1987) Surface acoustic wave devices as chemical sensors in liquids.
Evidence disputing the importance of Rayleigh wave attenuation. Anal. Chem. 59, 833–837.
Calcote, H.F. (1957) Mechanism for the formation of ions in flames. Combust. Flame 1, 385–403.
Caliendo C., Verardi P., Verona E., D’Amico A., Di Natale C., Saggio G., Serafini M., Paolesse R., and Huq S.E.
(1997) Advances in SAW-based gas sensors. Smart Mater. Struct. 6, 689–699.
Caruso F., Rodda E., and Furlong D.N., (1996) Orientational aspects of antibody immobilization and immunologi-
cal activity on quartz crystal microbalance electrodes. J. Colloid Interface Sci. 178, 104–115.
Cimalla V., Niebelschutz F., Tonisch K., Foerster Ch., Brueckner K., Cimalla I., Friedrich T., Pezoldt J., Stephan
R., Hein M, and Ambacher O. (2007) Nanoelectromechanical devices for sensing applications. Sens. Actuators
B 126, 24–34.
Cheeke J.D.N. and Wang Z. (1999) Acoustic wave gas sensors. Sens. Actuators B 59, 146–153.
Chen H.I. and Chou Y.I. (2003) A comparative study of hydrogen sensing performances between electroless plated
and thermal evaporated Pd/InP Schottky diodes. Semicond. Sci. Technol. 18, 104–110.
Chen G.Y., Thundat T., Wachter E.A., and Warmack R.J. (1995). Adsorption-induced surface stress and its effects
on resonance frequency of microcantilevers. J. Appl. Phys. 77(8), 3618–3622.
Cheng W.K., Summers T., and Collings N. (1998) The fast-response flame ionization detector. Prog. Energy Combust.
Sci. 24, 89–124.
Chévrier J.-B., Baert K., Slater T., and Verbist A. (1995) Micromachined infrared pneumatic detector for gas sensor.
Microsyst. Technol. 1, 71–74.
Cho J.-C. and Chiang J.-L. (2000) Ion sensitive field effect transistor with amorphous tungsten trioxide gate for pH
sensing. Sens. Actuators B 2, 81–87.
Choujaa A., Tirole N., Bonjour C., Martin G., Hauden D., Blind P., Cachard A., and Pommier C. (1995) AlN/Silicon
Lamb-wave microsensors for pressure and gravimetric measurements. Sens. Actuators A 46, 179–182.
Ciurczak E.W. and Drennen J.K. (2002) Pharmaceutical and Medical Applications of Near-Infrared Applications (Practical
Spectroscopy). Marcel Dekker, New York.
Clark L.C., Wolf R., Granger D., and Taylar Z. (1953) Continuous recording of blood oxygen tensions by polaro-
graphy. J. Appl. Physiol. 6, 189–193.
D’Amico A. and Natale C.D. (2001) A contribution on some basic definitions of sensors properties. IEEE Sensors J.
1, 183–190.
D’Amico A. and Verona E. (1989) SAW sensors. Sens. Actuators 17, 55–66.
Dahint R., Shana Z.A., Josse F., Riedel S.A., and Grunze M. (1993) Identification of metal ion solutions using acou-
stic plate mode devices and pattern recognition. IEEE Trans. Ultrason. Ferroel. Freq. Contr. 40(2), 114–120.
Davide F., Di Natale C., and D’Amico A. (1993) Sensor arrays figure of merits: definitions and properties. Sens.
DeLisa M.P., Zhang Z., Shiloach M., Pilevar S., Davis C.C., Sirkis J.S., and Bentley W.E. (2000) Evanescent wave
long period fiber Bragg grating as an immobilized antibody biosensor. Anal. Chem. 72, 2895–2900.
Déjous C., Savart M., Rebière D., and Pistré J. (1995) A shear-horizontal acoustic plate mode (SH-APM) sensor for
biological media. Sens. Actuators B 26–27, 452–456.
Devoret M., Sullivan N.S., Esteve D., and Deschamps P. (1980) Simple thermal conductivity cell using a miniature
thin film printed circuit for analysis of binary gas mixtures. Rev. Sci. Instrum. 51(9), 1220–1224.
Doll T., Lechner J., Eisele I., Schierbaum K.-D., and Göpel W. (1996) Ozone detection in the ppb range with work
function sensors operating at room temperature. Sens. Actuators B 34, 506–510.
Domansky K., Baldwin D.L., Grate J.W., Hall T.B., Li J., Josowicz M., and Janata J. (1998) Development and cali-
bration of field-effect transistor-based sensor array for measurement of hydrogen and ammonia gas mixtures
in humid air. Anal. Chem. 70, 473–481.
Domansky K., Liu J., Wang L.-Q., Engelhard M.H., and Baskaran S. (2001) Chemical sensors based on dielectric
response of functionalized mesoporous silica films. J. Mater. Res. 16, 2810–2816.
Eddowes M.J. (1987) Response of an enzyme-modified pH sensitive ion-selective device; analytical solution for the
response in the presence of pH buffer. Sens. Actuators 11, 265–274.
Elia A., Lugarà P.M., and Giancaspro C. (2005) Photoacoustic detection of nitric oxide by use of a quantum cas-
cade laser. Opt. Lett. 30(9), 988–990.
Elia A., Rizzi F., Di Franco C., Lugarà P.M., and Scamarcio G. (2006) Quantum cascade laser-based photoacoustic
spectroscopy of volatile chemicals: application to examethyldisilazane. Spectrochim. Acta A 64, 426–429.
Ekedahl L.-G., Eriksson M., and Lundström I. (1998) Hydrogen sensing mechanisms of metalinsulator interfaces.
Accounts Chem. Res. 31, 249–256.
Esteban I., Déjous C., Rebière D., Pistré J., Planade R., and Lipskier J.F. (2000) Mass sensitivity of SH-APM sensors:
potentialities for organophosphorous vapors detection. Sens. Actuators B 68, 244–248.
Falciai R., Mignani A.G., and Vannini A. (2001) Long period gratings as solution concentration sensors. Sens.
Fialkov A.B. (1997) Investigations on ions in flames. Prog. Energy Combust. Sci. 23, 399–528.
Fontecha J., Fernández M.J., Sayago I., Santos J.P., Gutiérrez J., Horrillo M.C., Gràcia I., Cané C., and Figueras E.
(2004) Fine-tuning of the resonant frequency using a hybrid coupler and fixed components in SAW oscillators
for gas detection. Sens. Actuators B 103, 139–144.
Freeman T.M. and Seitz W.R. (1978) Chemiluminescence fiber optical probe for hydrogen peroxide based on the
luminol reaction. Anal. Chem. 50, 1242–1246.
Fradeen J. (2003) Handbook of Modern Sensors, 3rd ed. Springer-Verlag, New York.
Frant M.S. (1994) History of the early commercialization of ion-selective electrodes. Analyst 199, 2293–2301.
Sauerbrey G. (1959) Verwendung von Schwingquarzen zur Wägung dünner Schichten und zur Mikrowägung. Z.
Physik 155, 206–222.
Gardner J.W. (1994) Microsensors. Wiley, Chichester, UK.
Gardner J.W., Varadan V.K., and Awadelkarim O.O. (2001) Micro Sensors, MEMS and Smart Devices. Wiley, Chichester, UK.
Gizelli E. (2000) Study of the sensitivity of the acoustic waveguide sensor. Anal. Chem. 72, 5967–5972.
Göpel W., Hesse J., and Zemel J.N. (eds.) (1991) Sensors: A Comprehensive Survey. Chemical and Biological Sensors. VCH,
Weinhiem.
Gopel W. and Schierbaum K.D. (1995) SnO2 sensors: current status and future prospects. Sens. Actuators B 26–27, 1–12.
Göpel W., Hesse J., and Zemel J.N. (eds.) (1994) Sensors: A Comprehensive Survey, Vol. 1. VCH, Weinheim.
Grate J.W., Martin S.J., and White R.M. (1993) Acoustic wave microsensors—Part II. Anal. Chem. 65(22), 987–996.
Grate J.W., Wenzel S.W., and White R.M. (1991) Flexural plate wave devices for chemical analysis. Anal. Chem. 63,
1552–1561.
Gründler P. (2006) Chemical Sensors: An Introduction for Scientists and Engineers. Springer-Verlag, Berlin.
Gupta K.C. and D’Arc M.J. (2000) Performance evaluation of copper ion selective electrode based on cyanocopo-
lymers. Sens. Actuators B 62, 171–176.
Gupta B.D. and Sharma N.K. (2002) Fabrication and characterization of U-shaped fiber-optic pH probes. Sens.
Gurlo A. (2006) Interplay between O2 and SnO2: Oxygen ionosorption and spectroscopic evidence for adsorbed
oxygen. ChemPhysChem. 7, 2041–2052.
Hagleitner C., Koll A., Vogt R., Brand O., and Baltes H. (1999) CMOS capacitive chemical microsystem with active
temperature control for discrimination of organic vapors. Tech. Dig. Transducers 2, 1012–1015.
Harame D.L., Bousse L.J., Shott J.D., and Meindl J.D. (1987) Ion-sensing devices with silicon nitride and borosilicate
glass insulators. IEEE Trans. Elect. Dev. 34, 1700–1707.
Harding G.L. (2001) Mass sensitivity of Love-mode acoustic sensors incorporating silicon dioxide and silicon-oxy-
fluoride guiding layers. Sens. Actuators B 88, 20–28.
Harrey P.M., Ramsey B.J., Evans P.S.A., and Harrison D.J. (2002) Capacitive-type humidity sensors fabricated using
the offset lithographic printing process. Sens. Actuators B 87, 226–232.
Harris P.D., Arnold W.M., Andrews M.K., and Partridge A.C. (1997) Resistance characteristics of conducting poly-
mer films used in gas sensors. Sens. Actuators B 42, 177–184.
Harsanyi G. (2000) Sensors in Biomedical Applications: Fundamentals, Technology and Applications. CRC Press, Boca Raton, FL.
Haueis R., Vellekoop M.J., Kovacs G., Lubking G.W., and Venema A. (1994) A Love-wave based oscillator for
sensing in liquids. In: Proceedings of the Fifth International Meeting of Chemical Sensors, 1, Lake Tahoe, CA, pp.
126–129.
Hierlemann A., Brand O., Hagleitner C., and Baltes H. (2003) Microfabrication techniques for chemical/biosensors.
Proc IEEE 91, 839–863.
Hirata H. and Higashiyama K. (1971) Analytical study of a cadmium ion-selective ceramic membrane electrode.
J. Anal. Chem. 257, 104–107.
Hofstetter D., Beck M., Faist J., Nagele M., and Sigrist M.W. (2001) Photoacoustic spectroscopy with quantum
cascade distributed-feedback lasers. Opt. Lett., 26, 887–889.
Homola J., Yee S.S., and Gaugiltz G. (1999) Surface plasmon sensors: review. Sens. Actuators B 54, 3–15.
Homola J. (2008) Surface plasmon resonance sensors for detection of chemical and biological species. Chem. Rev.
108(2), 462–493.
Ideta K. and Arakawa T. (1993) Surface plasmon resonance study for the detection of some chemical species. Sens.
Actuators B 13–14, 384–386.
Imai S.H., Mizuno H., Suzuki M., Takeuchi T., Tamiya E., Mashige F., Ohkubo A., and Karube I. (1994) Total
urinary protein sensor based on a piezoelectric quartz crystal. Anal. Chim. Acta 292, 65–70.
Ionescu R. (1998) Combined Seebeck and resistive SnO2 gas sensors, a new selective device, Sens. Actuators B 48,
392–394.
Ippolito S.J., Trinchi A., Powell D.A., and Wlodarski W. (2009) Acoustic wave gas and vapor sensors. In:
Comini E., Faglia G., and Sberveglieri G. (eds.), Solid State Gas Sensing. Science-Business Media, New York,
pp. 261–304.
Ishihara T. and Matsubara S. (1998) Capacitive type gas sensors. J. Electrocer. 2, 215–228.
Ishihara T., Sato S., Fukushima T., and Takita Y. (1996) Capacitive gas sensor of mixed oxide CoO-In2O3 to selec-
tively detect nitrogen monoxide. J. Electrochem. Soc. 143, 1908–1914.
Ishihara T., Sato S., and Takita Y. (1996) Capacitive-type sensors for the selective detection of nitrogen oxides. Sens.
Actuators B 24–25, 392–395.
Izumi M., Ohnuki H., Kato R., Imakubo T., Nagata M., Noda T., and Kojima K. (1998) Recent progress in metallic
Langmuir-Blodgett films based on TTF derivatives. Thin Solid Films 327–329, 14–18.
Jakoby B., Ismail G.M., Byfield M.P., and Vellekoop M.J. (1999) A novel molecularly imprinted thin film applied to
a love wave gas sensor. Sens. Actuators A 76, 93–97.
James S.W. and Tatam R.P.J. (2006) Fibre optic sensors with nano-structured coatings. Opt. A: Pure Appl. Opt. 8,
S430–S445.
Janata J. (1989) Principles of Chemical Sensors. Plenum Press, New York.
Janata J. and Josowicz M. (1997) A fresh look at some old principles: the Kelvin probe and the Nernst equation.
Anal. Chem. 69, 293A–296A.
Janata J. and Josowicz M. (2002) Conducting polymers in electronic chemical sensors. Nature Mater. 2, 19–24.
Jimenez C., Rochefeuille S., Berjoan R., Seta P., Desfours J.P., and Dominguez C. (2004) Nanostructures for chemi-
cal recognition using ISFET sensors. Microelectron. J. 35, 69–71.
Johan B., Magdalena M.Z., and Andrzej L. (2003) Carbonate ion-selective electrode with reduced interference from
salicylat. Biosens. Bioelectron. 18, 245–253.
Jorgenson R.C. and Yee S.S. (1993) A fiber optic chemical sensor based on surface plasmon resonance. Sens.
Actuators B 12, 213.
Josse F., Bender F., and Cernosek R.W. (2001) Guided shear horizontal surface acoustic wave sensors for chemical
and biochemical detection in liquids. Anal. Chem. 73, 5937–5944.
Kalantar-zadeh K., Trinchi A., Wlodarski W., Holland A., and Atashbar M.Z. (2001) A novel SAW Love mode
device with nanocrystalline ZnO film for gas sensing applications. In: Proceedings of the 1st IEEE Conference on
Nanotechnology, IEEE-NANO, Hawaii, pp. 556–561.
Kang Y., Ruan H., Wang Y., Arregui F.J., Matias I.R., and Claus R.O. (2006) Nanostructured optical fibre sensors
for breathing airflow monitoring. Meas. Sci. Technol. 17, 1207–1210.
Karlsen S.R., Johnston K.S., Jorgenson R.C., and Yee S.S. (1995) Simultaneous determination of refractive indi-
ces and absorbance spectra of chemical samples using surface plasmon resonance. Sens. Actuators B 24–25,
747–749.
Karthigeyan A., Gupta R.P., Scharnagl K., Burgmair M., Zimmer M., Sharma S.K., and Eisele I. (2001) Low tempe-
rature NO2 sensitivity of nano-particulate SnO2 film for work function sensors. Sens. Actuators B 78, 69–72.
Keller A, Ruegg M, Forster Loepfeb M., Pleischb M., Nebikerb P., and Burtschera H. (2005) Open photoacoustic
sensor as smoke detector. Sens. Actuators B 104, 1–7.
Kelkar U.R., Josse F., Haworth D.T., and Shana Z.A. (1991) Acoustic plate waves for measurements of electrical
properties of liquids. Microchem. J. 43, 155–164.
Khlebarov Z.P., Stoyanova A.I., and Topalova D.I. (1992) Surface acoustic wave sensors. Sens. Actuators B 8,
33–40.
Kim B.H., Prins F.E., Kern D.P., Raible S., and Weimar U. (2001) Multicomponent analysis and prediction with a
cantilever array based gas sensor. Sens. Actuators B 78, 12–18.
King D., Lyons W.B., Flanagan C., and Lewis E. (2004) Interpreting complex data from a three-sensor multipoint
optical fibre ethanol concentration sensor system using artificial neural network pattern recognition. Meas. Sci.
Technol. 15, 1560–1567.
Kondoh J., Matsui Y., Shiokawa S., and Wlodarski W.B. (1994) Enzyme-immobilized SH-SAW biosensor. Sens.
Kovacs G.T.A. (1998) Micromachined Transducers Source Book. McGraw-Hill, New York.
Kang Y., Ruan H., Wang Y., Arregui F.J., Matias I.R., and Claus R.O. (2006) Nanostructured optical fibre sensors
for breathing airflow monitoring. Meas. Sci. Technol. 17, 1207–1210.
Korotcenkov G., Han S.-D., and Stetter J.R. (2009) Review of electrochemical hydrogen sensors. Chem. Rev. 109(3),
1402–1433.
Korotcenkov G. (2008) The role of morphology and crystallographic structure of metal oxides in response of
conductometric-type gas sensors. Mater. Sci. Eng. R 61, 1–39.
Korotcenkov G. (2007a) Metal oxides for solid state gas sensors. What determines our choice? Mater. Sci. Eng. B
139, 1–23.
Korotcenkov G. (2007b) Practical aspects in design of one-electrode semiconductor gas sensors: status report. Sens.
Korotcenkov G. (2005) Gas response control through structural and chemical modification of metal oxides: state
of the art and approaches. Sens. Actuators B 107, 209–232.
Kummer A.M., A. Hierlemann A., and Baltes H. (2004) Tuning sensitivity and selectivity of complementary metal
oxide semiconductor-based capacitive chemical microsensors. Anal. Chem.76, 2470–2477.
Lakowicz J.R. (ed.) (1994) Topics in Fluorescence Spectroscopy, Vol. 4: Probe Design and Chemical Sensing. Plenum Press,
New York.
Lange D., Hagleitner C., Brand O., and Baltes H. (2001) CMOS resonant beam gas sensing system with on-chip
self excitation. In: Proceedings of IEEE Micro Electro Mechanical Systems 14th International Conference, Interlaken,
Switzerland, pp. 547–552.
Lavrik N., Sepaniak M., and Datskos P. (2004) Cantilever transducers as a platform for chemical and biological
sensors. Rev. Sci. Instrum. 75, 2229–2253.
Lee J.S., Lee J.H., and Hong S.H. (2004) Solid state amperometric CO2 sensor using a sodium ion conductor. J. Eur.
Cer. Soc. 24, 1431–1434.
Lieberman R.A. (1993) Recent progress in intrinsic fiber-optic chemical sensing II. Sens. Actuators B 11, 43–55.
Lin K.-W, Chen H.-I., Chuang H.-M., Chen C.-Y., Lu C.-T., Cheng C.-C., and Liu W.-C. (2004) Characteristics of
Pd/InGaP Schottky diodes hydrogen sensors. IEEE Sensors J. 4, 72–79.
Lin T., Sree R., Tseng S., Chiu K., Wu C., and Lo J. (2004) Volatile organic compound concentrations in ambient air
of Kaohsiung petroleum refinery in Taiwan. Atmos. Environ. 38, 4111.
Liu N., Hui J., Sun C.Q., Dong J.H., Zhang L.Z., and Xiao H. (2006) Nanoporous zeolite thin film-basedfiber intrin-
sic Fabry-Perot interferometric sensor for detection of dissolved organics in water. Sensors 6, 835–847.
Lu C., Czanderna W., and Townshend A. (1984) Applications of piezoelectric quartz crystal microbalances: book
review. Anal. Chim. Acta 199, 279.
Madou M.J. and Morrison S.R. (1989) Chemical Sensing with Solid Devices. Academic Press, Boston.
Lundström I. (1991) Field effect chemical sensors. In: Gopel W., Hesse J., and Zemel J.N. (eds.), Sensors: A
Comprehensive Survey. VCH, Weinheim, Vol. 1, pp. 467–529.
Lundstrom I., Sundgren H., Winquist F., Eriksson M., Krantz-Rulcker C., and Lloyd-Spetz A. (2007) Twenty-five
years of field effect gas sensor research in Linkoping. Sens. Actuators B 121, 247–262.
Majoo S, Gland J.L., Wise K.D., and Schwank J.W. (1996) A silicon micromachined conductometric gas sensor with
a maskless Pt sensing film deposited by selected-area CVD. Sens. Actuators B 36, 312–319.
Markinkowska K., McGauley M.P., and Symons E.A. (1991) A new carbon monoxide sensor based on a hydropho-
bic oxidation catalyst. Sens. Actuators B 5, 91–96.
Martin F., Newton M. I., McHale G., Melzak K.A., and Gizeli E. (2004) Pulse mode shear horizontal-surface acou-
stic wave (SH-SAW) system for liquid based sensing applications. Biosens. Bioelectron. 19, 627–632.
Matsuguch M., Kuroiwa T., Miyagishi T., Suzuki S., Ogura T., and Sakai Y. (1998) Stability and reliability of capaci-
tive-type relative humidity sensors using crosslinked polyimide films. Sens. Actuators B 52, 53–57.
McCallum J.J. (1989) Piezoelectric devices for mass and chemical measurements: an update. Analyst 114,
1173–1189.
Meyerhoff M.E. and Opdycke W.N. (1986) Ion selective electrodes. Adv. Clin. Chem. 25, 1–47.
Middlehoek S. and Audet S.A. (1989) Silicon Sensors for Chemical Signals. Academic Press, Boston.
Miller J.B. (2001) Catalytic sensors for monitoring explosive atmospheres. IEEE Sensors J. 1(1), 88–93.
Miura T. (1985) Thermistor humidity sensor for absolute humidity measurements and their applications. In: Chaddock,
J.B. (ed.), Proceedings of International Symposium on Moisture and Humidity, ISA, Washington, DC., pp. 555–573.
Moon W.J., Yu J.H., and Choi G.M. (2002) The CO and H2 gas selectivity of CuO-doped SnO2–ZnO composite
gas sensor. Sens. Actuators B 87, 464–470.
Moreno J., Arregui F.J., and Matias I.R. (2005) Fiber optic ammonia sensing employing novel thermoplastic poly-
urethane membranes. Sens. Actuators B 105, 419–424.
Morgan D.J. (1961) Construction and operation of a simple flame-ionization detector for gas chromatography.
J. Sci. Instrum. 38, 501–503.
Moriizumi T. (1988) Langmuir-Blodgett films as chemical sensors. Thin Solid Films 160, 413–429.
Moulin A.M., O’Shea S.J., and Welland M.E. (2000). Microcantilever-based biosensors. Ultramicroscopy 82, 23–31.
Muramatsu H.M., Tamiya E., and Karube I. (1987) Determination of microbes and immunoglobulins using a
piezoelectric biosensor. J. Membr. Sci. 41, 281–290.
Nagai M. and Nishio T. (1988) A new type of CO2 gas sensor comprising porous hydroxyapatite ceramics. Sens.
Actuators 15, 145–151.
Nomura T., Saitoh A., and Miyazaki T. (2003) Liquid sensor probe using reflecting SH-SAW delay line. Sens.
Actuators B 91(1–3), 298–302.
Ogawa K., Tsuchiya S., Kawakami H., and Tsutsui T. (1988) Humidity-sensing effects of optical fibers with micro-
porous SiO2 cladding. Electron. Lett. 24(1), 42–43.
Ostrick B., Muhlsteff J., Fleischer M., Meixner H., Doll T., and Kohl C.-D. (1999) Adsorbed water as key to room
temperature gas-sensitive reactions in work function type sensors: the carbonate-carbon dioxide system. Sens.
Ostrick B., Pohle R., Fleischer M., and Meixner H. (2000) TiN in work function type sensors: a stable ammonia
sensitive material for room temperature operation with low humidity cross sensitivity. Sens. Actuators B 68,
234–239.
Paldus B.A., Spence T.G., Zare R.N., Oomens J., Harren F.J.M., Parker D.H., Gmachl C., Capasso F., Sivco D.L.,
Baillargeon J.N., Hutchinson, A.L., and Cho A.Y. (1999) Photoacoustic spectroscopy using quantum cascade
lasers. Opt. Lett. 24, 178–180.
Papadopoulos C.A, Vlachos D.S., and Avaritsiotis J.N. (1996) A new planar device based on Seebeck effect for gas
sensing applications. Sens. Actuators B 34, 524–527.
Patel S.V., Wise K.D., Gland J.L., Zanini-Fisher M., and Schwank J.W. (1997) Characteristics of silicon-micromachi-
ned gas sensors based on Pt/TiOx thin films. Sens. Actuators B 42, 205–215.
Park E.J., Reid K.R., Tang W., Kennedy R.T., and Kopelman R. (2005) Ratiometric fiber optic sensors for the de-
tection of inter- and intra-cellular dissolved oxygen. J. Mater. Chem. 15, 2913–2919.
Patel S.V., Gland J.L., and Schwank J.W. (1999) Film structure and conductometric hydrogen-gas-sensing characte-
ristics of ultrathin platinum films. Sens. Actuators B 15, 3307–3311.
Pisco, M., Consales, M., Campopiano, S., Viter, R., Smyntyna, V., Giordano, M., and Cusano, A. J. (2006) A novel
optochemical sensor based on SnO2 sensitive thin film for ppm ammonia detection in liquid environment.
Lightwave Technol. 24, 5000–5007.
Pockrand I., Swalen J.D., Gordon JG. II, and Philpott M.R. (1979) Exciton-surface plasmon interactions. J. Chem.
Phys. 70, 3401.
Potje-Kamloth K. (2008) Semiconductor junction gas sensors. Chem. Rev. 108, 367–399.
Preejith P.V., Lim C.S., and Chia T.F. (2006) Serum protein measurement using a tapered fluorescent fibre-optic
evanescent wave-based biosensor. Meas. Sci. Technol. 17, 3255–3260.
Radomska A., Bodenszac E., Glab S., and Koncki R. (2004) Creatinine biosensor based on ammonium ion selective
electrode and its application in flow-injection analysis. Talanta 64(3), 603–608.
Raiteri R., Grattarola M., Butt H.-J., and Skládal P. (2001). Micromechanical cantilever-based biosensors. Sens.
Raiteri R., Grattarola M., and Berger R. (2002) Micromechanics senses biomolecules. Mater. Today 5(1), 22–29.
Razavi B. (2000) Design of Analog CMOS Integrated Circuits. McGraw-Hill, New York.
Reich G. (2005) Near-infrared spectroscopy and imaging: basic principles and pharmaceutical applications. Adv.
Drug Deliv. Rev. 57, 1109–1143.
Reinhoudt D.N. (1995) Durable chemical sensors based on field-effect transistors. Sens. Actuators B 24–25, 197–200.
Ricco A.J., Martin S.J., Frye G.C., and Niemczyk T.M. (1988) Acoustic plate mode devices as liquid phase sensors.
In: Proceedings of the IEEE Ultrasonics Symposium, Hilton Head Island, SC, USA, pp. 23–26.
Ricco A.J. and Martin S.J. (1991) Thin metal film characterization and chemical sensors: monitoring electronic
conductivity, mass loading and mechanical properties with surface acoustic wave devices. Thin Solid Films 206,
94–101.
Rinnan R. and Rinnan A. (2007) Application of near infrared reflectance (NIR) and fluorescence spectroscopy to
analysis of microbiological and chemical properties of arctic soil. Soil Biol. Biochem. 39, 1664–1673.
Roberts G. (ed.) (1990) Langmuir-Blodgett Films. Plenum Press, New York.
Roberts G.G. (1983) Transducer and other applications of Langmuir-Blodgett films. Sens. Actuators 4, 131–145.
Rogers K.R. and Poziomek E. (1996) Fiber optic sensors for environmental monitoring. Chemosphere 33,
1151–1174.
Rouessac F. and Rouessac A. (2007) Chemical Analysis: Modern Instrumentation Methods and Techniques. Wiley,
Chichester UK.
Sakti S.P., Hauptmann P., Zimmermann B., Buhling F., and Ansorge S. (2001) Disposable HSA QCM-immunosensor
for practical measurement in liquid. Sens. Actuators B 78, 257–262.
Sang X.Z., Yu C.X., Yan B.B., Ma J.X., Meng Z.F., Mayteevarunyoo T., and Lu N.G. (2006) Temperature-insensitive
chemical sensor with twin bragg gratings in an optical fiber. Chin. Phys. Lett. 23, 3202–3204.
Schierbaum K.D., Weimar U., and Göpel W. (1990) Multicomponent gas analysis: an analytical chemistry approach
applied to modified SnO2 sensors. Sens. Actuators B 2, 71–78.
Schoeneberg U., Hosticka B.J., Zimmer G., and Maclay G.J. (1990) A novel readout technique for capacitive gas
sensors Sens. Actuators B 1, 58–61.
Schreiter M., Gabl R., Lerchner J., Hohlfeld C., Delan A., Wolf G., Bluher A., Katzschner B., Mertig M., and Pompe
W. (2006) Functionalized pyroelectric sensors for gas detection. Sens. Actuators B 119, 255–261.
Seiyama T., Yamazoe N., and Arai H. (1983) Ceramic humidity sensors. Sens. Actuators 4, 85–96.
Senillou Jaffrezic-Renault N., Martelet C., and Griffe F. (1998) A miniaturized ammonium sensor based on the
integration of both ammonium and reference FETs in a single chip. Mater. Sci. Eng. C 6, 59–63.
Shamsipur M., Rouhani S., Mohajeri A., and Ganjali M.R. (2000) A bromide ion-selective polymeric membrane
electrode based on a benzo-derivative xanthenium bromide salt. Anal. Chim. Acta 418(2), 197–203.
Shen C.Y., Huang C.P., and Huang W.T. (2004) Gas-detecting properties of surface acoustic wave ammonia sensors.
Sens. Actuators B 101, 1–7.
Sheppard N.F., Day D.R., Lee H.L., and Senturia S.D. (1982) Microdielectrometry. Sens. Actuators 2, 263–274.
Siesler H.W., Ozaki Y., Kawata S., and Heise H.M. (2002) Near-Infrared Spectroscopy: Principles, Instruments, Applications.
Wiley-VCH, Weinheim.
Da Silva M.G., Vargas H., Miklós A., and Hess P. (2004) Photoacoustic detection of ozone using a quantum cascade
laser. Appl. Phys. B 78, 677–680.
Široký K. (1993) Use of the Seebeck effect for sensing flammable gas and vapours. Sens. Actuators B 17, 13–17.
Somov S.I., Reinhardt G., Guth U., and Göpel W. (2000) Tubular amperometric high-temperature sensors: simulta-
neous determination of oxygen, nitrogen oxides and combustible components. Sens. Actuators B 65, 68–69.
Streetman B.G. (1990) Solid State Electronic Devices, 3rd ed. Prentice Hall, Englewood Cliffs, NJ.
Su P.-G. and Tsai W.-Y. (2004) Humidity sensing and electrical properties of a composite material of nano-sized
SiO2 and poly(2-acrylamido-2-methylpropane sulfonate). Sens. Actuators B 100, 425–430.
Sundmacher K., Rihko-Struckmann L.K., and Galvita V. (2005) Solid electrolyte membrane reactors: status and
trends. Catal. Today 104, 185–199.
Symons E.A. (1992) Catalytic gas sensors. In: Sberveglieri G. (ed.), Gas Sensors. Kluwer Academic, Dordrecht, The
Netherlands, pp. 169–185.
Taillades G., Valls O., Bratov A., Dominguez C., Pradel A., and Ribes M. (1999) ISE and ISFET microsensors based
on a sensitive chalcogenide glass for copper ion detection in solution. Sens. Actuators B 59, 123–127.
Taylor R. and Schultz J. (1996) Handbook of Chemical and Biological Sensors. IOP Publishing, Philadelphia.
Tan O.K., Cao W., Zhu W., Chai J.W., and Pan J.S. (2003) Ethanol sensors based on nano-sized α-Fe2O3 with SnO2,
ZrO2, TiO2 solid solutions. Sens. Actuators B 93, 396–401.
Tau S. and Sarma T.V.S. (2006) Evanescent-wave optical Cr VI sensor with a flexible fused-silica capillary as a
transducer. Opt. Lett. 31, 1423–1425.
Then D., Vidic A., and Ziegler Ch. (2006) A highly sensitive self-oscillating cantilever array for the quantitative and
qualitative analysis of organic vapor mixtures. Sens. Actuators B 117, 1–9.
Tirole N., Choujaa A., Hauden D., Martin G., Blind P., Froelicher M., Pommier J.C., and Cachard A. (1993) Lamb
waves pressure sensor using an AlN/Si structure. In: Proceedings of the IEEE Ultrasonics Symposium, Hawaii,
pp. 371–374.
Trinchi A., Galatsis K., Wlodarski W., and Li Y.X. (2003) A Pt/Ga2O3-ZnO/SiC Schottky diode-based hydrocar-
bon gas sensor. IEEE Sensors J. 3, 548–553.
Trinchi A., Woldarski W., and Li Y.X. (2004) Hydrogen sensitive Ga2O3 Schottky diode sensor based on SiC. Sens.
Tsai Y.-Y., Lin K.-W., Chen H.-I., Lu C.-T., Chuang H.-M., Chen C.-Y., and Liu W.-C. (2003) Comparative hydrogen
sensing performances of Pd- and Pt-InGaP metal-oxide-semiconductor Schottky diodes. J. Vac. Sci. Technol. B
21, 2471–2475.
Tuyen L.T.T., Vinh D.X., Khoi P.H., and Gerlach G. (2002) Highly sensitive NOx gas sensor based on a Au/n-Si
Schottky diode. Sens. Actuators B 84, 226–230.
Urashi T. and Arakawa T. (2001) Detection of lower hydrocarbons by means of surface plasmon resonance. Sens.
Van Oirschot T.G.J., VanLeeuwen D., and Medema J. (1972) The effect of gases on the conductive properties of
organic semiconductors. J. Electroanal. Chem. 37, 373–385.
Varghese O.K., Gong D., Dreschel W.R., Ong K.G., and Grimes C.A. (2003) Ammonia detection using nanoporous
alumina resistive and surface acoustic wave sensors. Sens. Actuators B 94, 27–35.
Vlachos D.S., Papadopoulos C.A., and Avaritsiotis J.N. (1997) A technique for suppressing ethanol interference
employing Seebeck effect devices with carrier concentration modulation. Sens. Actuators B 44, 239–242.
Vellekoop M.J., Lubking G.W., Sarro P.M., and Venema A. (1994) Integrated-circuit compatible design and techno-
logy of acoustic-wave based microsensor. Sens. Actuators A 44, 249–263.
Wang A.W., Kiwan R., White R.M., and Ceriani R.L. (1997) A silicon-based ultrasonic immunoassay for detection
of breast cancer antigens. In: Proceedings of the Ninth International Conference on Solid-State Sensors and Actuators,
Transducers’97, Chicago, pp. 191–194.
Wang A.W., Kiwan R., White R.M., and Ceriani R.L. (1998) A silicon-based ultrasonic immunoassay for detection
of breast cancer antigens. Sens. Actuators B 49, 13–21.
Wang H., Feng C.-D., Sun S.-L., Segre C.U., and Stetter J.R. (1997) Comparison of conductometric humidity-
sensing polymers. Sens. Actuators B 40, 211–216.
Wenzel S.W. and White R.M. (1989) Flexural plate-wave sensor: chemical vapor sensing and electrostrictive excita-
tion. In: Proceedings of the IEEE Ultrasonics Symposium, New York, pp. 595–598.
White R.M., Martin S.J., Ricco A.J., Zellers E.T., Frye G.C., and Wohltjen H. (1997) Acoustic Wave Sensors: Theory,
Design, and Physico-Chemical Application. Academic Press, San Diego, CA.
Wilson D.M., Hoyt S., Janata J., Booksh K., and Obando L. (2001) Chemical sensors for portable, handheld field
instruments. IEEE Sensors J. 1, 256–274.
Wohltjen H. (1984) Mechanisms of operation and design considerations for surface acoustic wave device vapor
sensors. Sens. Actuators 5, 307–325.
Wohltjen H., Barger W.R., Snow A.W., and Jarvis N.L. (1985) A vapor-sensitive chemiresistor fabricated with pla-
nar microelectrodes and a Langmuir-Blodgett organic semiconductor film. IEEE Trans. Electron. Dev. ED-32,
1170–1174.
Wolfbeis O.S. (1990) Chemical sensors—survey and trends. Fresenius J. Anal. Chem. 337, 522–527.
Wright J.D., Cado A., Peacock S.J., Rivalle V., and Smith A.M. (1995) Effects of nitrogen dioxide on the surface
plasmon resonance of substituted pthalocyanine films. Sens. Actuators B 29, 108–114.
Wu B., Chang C., Sree U., and Chiu K. (2006) Measurement of non-methane hydrocarbons in Taipei city and their
impact on ozone formation in relation to air quality. J. Anal. Chim. Acta 576, 91–99.
Yamazoe N. and Shimuzu Y. (1989) Humidity sensors: principle and applications. Sens. Actuators 10, 379–398.
Yotter R.A. and Wilson D.M. (2004) Sensor technologies for monitoring metabolic activity in single cells—Part II:
nonoptical methods and applications. IEEE Sensors J. 4, 412–429.
Yuqing M., Jianguo G., and Jianrong C. (2003) Ion sensitive field effect transducer-based biosensors. Biotechnology
Adv. 21, 527–534.
Zao Z., Buttner W.J., and Stetter J.R. (1992) The properties and applications of amperometric gas sensors.
Electroanalysis 4, 253–266.
Zhang G., Lui J., and Yuan M. (2000) Novel carbon dioxide gas sensor based on infrared absorption. Opt. Eng.
39, 2235.
Zhang G. and Wu X. (2004) A novel CO2 gas analyzer based on IR absorption Opt. Lasers Eng. 42, 219–231.
Zimmer M., Burgmair M., Scharnagl K., Karthigeyan A., Doll T., and Eisele I. (2001) Gold and platinum as ozone
sensitive layer in work-function gas sensors. Sens. Actuators B 80, 174–178.
CHAPTER 2
DESIRED PROPERTIES FOR

SENSING MATERIALS
G. Korotcenkov
1. INTRODUCTION
The problem of optimization is a key factor in both design and manufacturing of electronic devices.
In the case of either chemical sensors or solid-state gas sensors (GS), this problem has some specific
peculiarities due to the absence of strict quantitative theory which would describe their operation. Some
semiquantitative approaches can be found in the literature (Clifford 1983; Morrison 1987; Zemel 1988;
Geistlinger 1993, 1994; Moseley et al. 1991; Brinzari et al. 2000; Barsan and Weimar 2001). However,
since the number of physical and chemical parameters that characterize sensor properties is large, and
some of these parameters are difficult to control, the problem of optimization is largely empirical and
remains a kind of art.
At present, therefore, the field of chemical sensors is characterized by a search for optimal sensing
materials and design of adequate theoretical models that may promote optimization of chemical sen-
sors. The term optimization here means achieving either necessary or the best available values of sensitiv-
ity, selectivity, and response time of chemical sensors at given operating conditions.
For example, according to at least one earlier view of the problem of gas sensor design (Moseley et
al. 1991), almost any metal oxide can serve as the basis for a solid-state gas sensor. For this purpose one
needs only to prepare this metal oxide as a sufficiently fine dispersed porous substance with properties
controlled by the surface state.
However, as requirements for gas sensors became more complex and precise, and understanding
of the nature of the gas-sensing effects was became more fundamental, our conception of material
63
compatibility for gas sensor elaboration began to change. We began to understand that for implementa-
tion of such requirements as
• High response to the target agent

• Low cross-sensitivity
• Fast and reversible interaction with analytes
• Low sensitivity of the signal to a change in air humidity
• Absence of long-term drift
• Short time to operational status
• Effective low-cost technology
• High reproducibility
• Uniform and strong binding to the surface of the substrate
• Easy connection to control units, etc.
materials for chemical sensors have to possess specific combination of physical and chemical proper-
ties, and not every material can fulfill these requirements.
A systematic consideration of the required properties of materials for chemical sensor applications
indicates that the key properties which determine a specific choice include:
• Adsorption ability
• Electronic, electrophysical, and chemical properties
• Catalytic activity
• Permittivity
• Thermodynamic stability
• Crystallographic structure
• Interface quality
• Compatibility with current or expected materials to be used in processing
• Reliability, etc.
Many different materials appear to be favorable in terms of some of these properties, but very few
materials are promising with respect to all of these requirements.
As confirmation of this statement, let us examine parameters which can determine a material’s
compatibility for use as a chemical sensor. Our attention will focus mainly on solid-state gas sensors,
which are the most complicated in terms of their principles of operation (Kupriyanov 1996; Barsan et
al. 1999; Kohl 2001; Barsan and Weimar 2003); however, the information in this chapter is general in
nature and can be used as a guide for both the analysis of behavior and the choosing of materials for
all types of chemical sensors, including optical sensors, in which the basic sensitive parameter is the
coefficient of optical refraction. Even in such devices, the change of the sensing element’s parameters
takes place through adsorption processes, which will be thoroughly reviewed in this chapter. The same
requirements also apply to metal oxides for electrochemical sensors. According to Trasatti (1980), these
materials must possess parameters such as large surface area, high electrocatalytic activity, excellent
long-term stability, high electrical conductivity, etc.
DESIRED PROPERTIES FOR SENSING MATERIALS • 65
This brief review cannot cover all promising materials that have been developed for chemical sen-
sors. In addition to binary oxides and various polymers, there are numerous ternary, quaternary, and
complex metal oxides, which are of interest for particular applications. For simplicity of presentation,
priority has been given to examining binary oxides. Properties of more complicated oxide phases will
be discussed in Chapter 5.
Readers who are interested in further information about these metal oxides and polymers may con-
sult relevant reviews and monographs (Sandier and Karo 1974; Samsonov 1973; Jolivet 2000). The bulk
properties of simple binary oxides are fairly well understood, and some excellent reviews and books
treat the thermodynamics (Johnson 1982), the structures, and nonstoichiometric aspects (Sorensen
1981; Catlow 1997) which are particularly important for oxides. Spectroscopy aspects are described by
Hamnett and Goodenough (1984), and mechanical properties by Sorensen (1981). Present knowledge
about the surface science of metal oxides has been comprehensively reviewed by Dufour and Nowotny
(1988), Freund and Umbach (1993), Henrich and Cox (1994), Szuber and Gopel (2000), and Calatayud
et al. (2003). Detailed information about transition metal oxides is given by Krilov and Kisilev (1981),
Hamnett and Goodenough (1984), and Cox (1992). Electrochemistry of metal oxides and interface
phenomena are reviewed by Gellings and Bouwmeester (1997). The interaction of metal oxide surfaces
with agents dissolved in water is discussed in detail by Blesa et al. (2000).
2. COMMON CHARACTERISTICS OF METAL OXIDES
2.1. CRYSTAL STRUCTURE OF METAL OXIDES
In most cases we do not know exactly where atoms are on the surface of metal oxides, but their bulk
crystal structures are known very accurately, thanks to about a century of x-ray crystallography study
of various crystalline materials. The most prevalent crystallographic structures of metal oxides used for
chemical sensors, with some examples, are listed in Table 2.1 (Samsonov 1973, 1982; Cox 1992; Henrich
and Cox 1994). It is necessary to note that the crystal structure of metal oxides is complicated; most
metal oxides have large unit cells which include many atoms.
The most important determinants of crystal structure are the stoichiometry, or relative numbers
of the different types of atoms present, and the coordination of ions, i.e., the number of ions of one
type surrounding another and their geometric arrangement. Commonly, metal oxides consist of several
structural units, which may be triangular, tetrahedral, octahedral, or cubic (see Figure 2.1). A consider-
ation of oxide structures suggests that metal-ion coordination numbers and geometries show reason-
ably systematic features that carry over among different structure types. For example, it has been estab-
lished that the ratio between the ionic radii of cations and anions (the “ratio rule”) serves as a rough
prediction of preferred coordination in the unit cells of metal oxides (see Table 2.2) (Pauling 1929).
According to Pauling, the size of the O2− anion is 0.14 nm, and the size of cations in metal oxides is in
the range from 0.05 to 0.15 nm. As follows from Table 2.1, sixfold octahedral coordination is the most
common metal-ion geometry in the metal oxides used for chemical sensors.
Another very important crystallographic parameter of sensing materials is the concentration of
point defects, i.e., the degree of deviation of the chemical composition from the ideal chemical formula.
Table 2.1. Oxide structures of some stable metal oxides
FORMULA EXAMPLES NAME SYMMETRY COORDINATION

M2O Cu2O, Ag2O Cuprite Cubic 2: Linear
Na2O Antifluorite Cubic 4: Tetrahedral
MO MgO, NiO, CoO, CaO Rocksalt Cubic 6: Octahedral
(no oxides) Zincblende Cubic 4: Tetrahedral
ZnO Wurtzite Hexagonal
PdO Pd O Tetragonal 4: Planar
CuO Tenorite Monoclinic
PbO PbO Tetragonal 4: Pyramidal
M3O4 Fe3O4 Magnetite Cubic Two M6 (oct.);
(inverse spinel) One M4 (tet.)
M2O3 α-Al2O3, Rh2O3 Corundum Hexagonal 6: Octahedral
α-Ga2O3 α-Fe2O3
Cr2O3, V2O3
In2O3, Sc2O3, Mn2O3 Mn2O3 Cubic 6: Octahedral
MO2 ZrO2, UO2, CeO2 Fluorite Cubic 8: Cubic
TiO2, SnO2, RuO2 Rutile Tetragonal 6: Octahedral
GeO2
M2O5 α-V2O5 V2O5 Orthorhombic 5+1: Distorted
MO3 ReO3, WO3 ReO3 Cubic 6: Octahedral
(distorted variants) (pseudo-cubic)
MoO3 MoO3 Monoclinic 6: Distorted
α-MoO3 α-MoO3 Orthorhombic
AMO3 SrTiO3, BaTiO3 Perovskite Cubic M6 (oct.), A12 (cub. oct )
CaTiO3, NaNbO3 (or distorted)
LiNbO3, FeTiO3 Ilmenite Trigonal M and A6 (oct.)
A2MO4 La2CuO4, Nb2CuO4 Layer perovskite Tetragonal M6 (oct.), A12 (cub. Oct.)
(K2NiF4) (or distorted)
The structure of point defects in metal oxides, especially in transition-metal oxides, has been extensively
studied for many years (Mrowec 1978; Murch and Nowick 1984; Nowick 1991; Cox 1992). It has been
established that many physical and chemical properties of nonstoichiometric metal oxides depend on
the nature, concentration, and mobility of defects in the crystalline lattice of these materials. In particu-
lar, the reactivity of solids, as well as their sintering, semiconducting, catalytic, and many other proper-
ties, depend to a greater degree on the nature of the defects than on the kind of substance itself.
It is necessary to note that point defects in oxides behave quite differently from point defects in
metals and standard semiconductors. According to Nowick (Murch and Nowick 1984; Nowick 1991),
the reasons are as follows:
Figure 2.1. Basic building blocks of metal oxide crystal structures: (a) triangle; (b) tetrahedron; (c)
octahedron; (c) cube.
1. In a binary oxide there are two different sublattices. The diffusion of cations and of anions pro-
ceeds each on its own sublattice, involving different point defects. Accordingly, cation and anion
diffusion rates can be very different and, in fact, usually are.
2. The exact composition of a large class of oxides is a sensitive function of the ambient partial
pressure of oxygen, P(O2 ). Oxides can exchange oxygen with the ambient atmosphere. Thus, the
stoichiometry, and therefore the concentration, of point defects can be controlled by the anneal-
ing of a metal oxide specimen at a specified P(O2 ) at a given temperature.
3. The melting temperatures of many oxides (SiO2, Al2O3, MgO, CaO, La2O3, etc.) are very high,
so, below approximately 1000°C the intrinsic point defect concentrations in these oxides are
very small. This implies that the presence of accidentally or intentionally introduced impurities
strongly affects the defect concentrations in such metal oxides. Therefore, unintentional and
uncontrolled defects may dominate the transport behavior.
4. Most important, oxides are highly ionic, so that, in general, point defects carry an effective
charge. This implies that the behavior of point defects is controlled by charge compensation
that results in the simultaneous introduction of point defects of opposite charge to maintain the
condition of charge neutrality. The existence of charge-compensation defects on opposite sub-
lattices and the corresponding equilibrium conditions through mass action equations means that
the defect concentrations in the two sublattices are coupled—i.e., the introduction of defects in
one sublattice can profoundly affect the defect structure in the other. Furthermore, the presence
Table 2.2. Relationship between cation coordination number, shape of the

coordination polyhedron, and cation-to-anion size ratio
COORDINATION RATIO OF
CATION POLYHEDRON IONIC RADII
COORDINATION GEOMETRY (Rc /Ra )
3 Triangle 0.155–0.225
4 Tetrahedron 0.225–0.414
6 Octahedron 0.414–0.732
8 Cube 0.732–1.00
Source: Data from Pauling 1929.
of oppositely charged point defects results in the formation of associated pairs or larger clusters,
especially when the temperature is not very high.
Because of the potential problem of impurities controlling the defect structure, it is desirable to
deliberately control the introduction of defects. Probably the two best ways to accomplish this are (1) to
introduce aliovalent impurities and (2) to take the material off stoichiometry through appropriate P(O2 )
annealing. However, it is necessary to remember that at high concentration of point defects the mutual
interaction among point defects can be accompanied by formation of complexes and defect clusters
called extended defects (Mrowec 1978). These extended defects may become further ordered, leading to
superstructure ordering and to formation of intermediate phases. In some cases, point defects may be
eliminated by the process of crystallographic sheer, which is connected to formation of a whole series
of intermediate phases.
It is important to note that, in principle, most of the point defects mentioned above can also be
present at free surfaces of a crystal, but their energies are very different from those of the defects in
the bulk sample (Gellings and Bouwmeester 2000). This means that the defect concentrations at the
surface also differ from those in the bulk. Depending on the sign of the energy difference, this can
lead either to increased concentration, also called positive defect segregation to the surface, or the
reverse, called negative (or reverse) segregation. Due to the changes in band energies near the surface
in semiconducting oxides, the so-called band bending or band curvature, the electron or electron-hole
concentrations in the conduction or valence bands, respectively, are different close to the surface from
those in the bulk.
2.2. ELECTRONIC STRUCTURE OF METAL OXIDES
Metal oxides exhibit a very wide range of electrophysical properties. Their electrical behavior ranges
from the best insulators (e.g., Al2O3 and MgO) through wide-band-gap and narrow-band-gap semicon-
ductors (TiO2, SnO2, and Ti2O3, respectively) to metals (V2O3, NaxWO3, and ReO3 ) and superconduc-
tors (including reduced SrTiO3 ). The range of electronic structures of oxides is so wide that metal oxide
compounds have been divided into two categories: (1) transition-metal oxides (Fe2O3, NiO, Cr2O3, etc.)
and (2) non–transition-metal oxides, which include (a) pre–transition-metal oxides (Al2O3, etc.) and (b)
post–transition-metal oxides (ZnO, SnO2, etc.).
The common feature of the non–transition-metal oxides is that the valence orbitals of the metal
atoms have s and p symmetry. In transitional-metal oxides, however, the d atomic orbitals assume cru-
cial importance. Many of the complications in using transition-metal oxides stem from this difference,
because of the different bonding properties associated with d orbitals. These complexities include the
existence of variable oxidation states, the frequent failure of the band model, and the crystal-field
splitting of the d orbitals (Henrich and Cox 1994). Table 2.3 shows the electron configuration of 3d
transition-metal oxides.
Henrich and Cox (1994) gave the following explanation of the difference in behavior between
non–transition and transition-metal oxides. The non–transition-metal oxides contain elements that with
some exceptions have only one preferred oxidation state. Other states are inaccessible because too
Table 2.3. Electron configurations for 3d transition-metal oxides
ELECTRON
CONFIGURATION METAL OXIDES
3d 0 Sc2O3, TiO2,V2O5, CrO3
3d 1 VO2, Ti2O3
3d 2 Cr O2, V2O3, TiOx
3d 3 β-MnO2, Cr2O3, VOx
3d 4 Mn2O3, Mn3O4
3d 5 α-Fe2O3, MnO, Fe3O4
3d 6 FeO, Co3O4
3d 7 CoO
3d 8 NiO
3d 9 CuO
3d 10 Cu2O, ZnO
Source: Data from Krilov and Kisilev 1981.
much energy is needed to add or remove an electron from the cations when they are coordinated with
O2− ligands. Transition-metal oxides behave differently in that the energy difference between a cation
d n configuration and either a d n+1 or d n−1 configuration is often rather small. The most obvious conse-
quence is that many transition elements have several stable oxides with different compositions. It is also
much easier than with non–transient-metal oxides to create defects that have different electron configu-
rations. As a result of high defect concentrations, the bulk and surface chemistries of transition-metal
oxides are very complicated.
Trends in the stability of different oxidation states are very important in surface chemistry, as they
control both the types of defect that may be formed easily and the type of chemisorption that may take
place. The d 0 configuration represents the highest oxidation state that can ever be attained; thus, pure TiO2,
V2O5, etc., cannot gain any more oxygen, although they can lose oxygen to form defects or other bulk
phases. On the other hand, d n oxides with n ≥ 1 are potentially susceptible to oxidation as well as reduction.
The stability of high oxidation states declines with increasing atomic number across a given series.
This big difference between the behavior of non–transition-metal and transition-metal oxides
means that transition-metal oxides are more sensitive to a change of external conditions. It seems,
therefore, that these oxides might be preferable for use in chemical sensors, and gas sensors particularly.
However, in practice, simple transition-metal oxides are not being used for chemical sensor design.
Structure instability, and nonoptimality of other parameters that are important for chemical sensors,
such as Eg and electroconductivity, considerably limit their use.
Only transition-metal oxides with d 0 and d 10 electronic configurations find real application. As we
know, the post–transition-metal oxides, such as ZnO and SnO2, have cations with the filled d 10 con-
figuration. The d 0 configuration is found in binary transition-metal oxides such as TiO2, V2O5, WO3,
and perovskites such as ScTiO3, LiNbO3, etc., as well. These compounds share many features with the
non–transition-metal oxides. They have a filled valence band of predominantly O2p character and a gap
between this and an empty conduction band. Typical band gaps are 3–4 eV. Stoichiometric d 0 oxides
are therefore good isolators, diamagnetic, and have no electronic excitations at energies less than the
band gap.
So, the main differences between transition-metal oxides and non–transition-metal oxides are that
(1) the lower part of the conduction band is based on metal d, rather than s, orbitals, and (2) many
transition-metal oxides are relatively easily reduced to form semiconducting or metallic phases.
The band model predicts that most oxides that have a partially filled d band—i.e., for d n with 0 <
n < 10—should be metallic. These expectations are frequently not fulfilled because of the intervention
of various types of electron–electron and electron–lattice interactions. Nevertheless, “simple” metallic
behavior is found with a number of oxides of elements in the 4d and 5d series (ReO3, RuO2). Some
oxides of the 3d series (Ti2O3, VO2, Fe2O3) also have high conductivity in the metallic range.
However, limited use of pure transition-metal oxides for conductometric gas sensor fabrication
does not mean that transition-metal oxides are not of interest to designers of chemical sensors. On the
contrary, unique surface properties, plus high catalytic activity, make them very attractive for various sen-
sor applications, such as modification the properties of more stable and wide-band-gap oxides, and for-
mation of more complicated nanocomposite materials (Kanazawa et al. 2001). For example, for optical
wave-guide sensors, where the change of optical refraction coefficient is more important than the change
of electroconductivity, the most attractive metal oxides are transition-metal oxides such as WO3 (for H2
and alcohol detection), Mn2O3, Co3O4, and NiO (for CO detection) (Dakin and Culshaw 1988).
2.3. ROLE OF THE ELECTRONIC STRUCTURE OF METAL OXIDES IN

SURFACE PROCESSES
In spite of many experimental and theoretical studies (Henrich and Cox 1994; Noguera 1995; Lantto
et al. 2000; Maki-Jaskar and Rantala 2002), many surface properties of metal oxides are still not well
understood. It is established, however, that adsorption on metal oxides involves very rich chemistry.
To analyze the adsorption and sensing properties of metal oxides, one can sometimes use the
concepts of the acid/base properties of the metal oxide surface. In this case, the behavior of differ-
ent oxides can be rationalized to some extent in terms of electrostatic and chemical bonding modes
(Noguera 1995). Henrich and Cox (1994), however, are skeptical of the direct application of Lewis
acid/base properties to explain the adsorption properties of real surfaces. Nevertheless, they think that
it is possible to apply these ideas to predict some adsorption properties of metal oxides (Horiuchi et al.
1998) as well as surface activity in solutions (Trasatti 1987).
Henrich and Cox (1994) gave the following explanation of Lewis acid/base properties of metal
oxides. Lewis acidity should depend on the existence of exposed metal cations at the surface, having
empty orbitals and positive charges that can interact with the filled orbitals and/or negative charges or
dipoles of donor molecules. Lewis acid strength is expected to depend on factors such as the ion charge,
the degree of coordinative unsaturation, and the availability of empty orbitals (i.e., the band gap). For
example, the stable (100) surfaces of alkaline-earth oxides should show very weak acidity because of
the relatively low cation charge and the large band gap. Stronger acidity is expected with more highly
charged ions such Ti4+, and also with post–transition-metal ions as in ZnO, where the band gap is
smaller and bonding with empty Zn 4s and 4p orbitals is possible.
Lewis basicity on oxide surfaces is related to the availability of a pair of 2p electrons associated with
oxygen ions. One again, coordination is expected to be important. However, the charge of the metal
cation should now play a role inverse to that expected in acid behavior. Low cation charge and large
cation radius lead to weaker bonding and hence more basic O2− ions. Thus, the most basic of the oxides
considered in this chapter is BaO, which is easily corroded by acidic adsorbates such as CO2 (see Figure
2.2). These arguments show that not only the composition but also the particular crystal surface under
consideration are important in determining the acid/base properties. However, the reported behavior
does not always follow simple expectations. One reason is that defects may often be important, even
on surfaces supposedly prepared to be defect-free. Surface defects create sites where surrounding ions
have lower coordination and thus may be more active than acidic or basic centers.
In the frame of the approach discussed, Dimitrov and Komatsu (2002) tried to classify simple
metal oxides. For this purpose they used the values of refractive index, the energy of the band gap, ion
polarizability, cation polarizability, bulk basicity, O1s binding energy determined by x-ray photoelectron
spectroscopy (XPS) measurements, metal (or nonmetal) binding energy, and the Yamashita-Kurosawa
interaction parameter of the oxides. Some of these parameters are presented in Table 2.4.
Analysis of these electronic properties of metal oxides allowed Dimitrov and Komatsu (2002) to
propose a classification of the oxides, as shown in Table 2.5. According to this classification, the simple
oxides can be separated into three groups. The first group of semicovalent, predominantly acidic oxides
includes BeO, B2O3, P2O5, SiO2, Al2O3, GeO2, and Ga2O3, which have low oxide ion polarizability,
high O1s binding energy, low cation polarizability, high metal (or nonmetal) outermost binding energy,
comparatively low bulk basicity, and strong interionic interaction, leading to the formation of strong
covalent bonds.
Cs2O BaO
Rb2O SrO
K2O Na2O CaO La2O3 Nb2O3
MgO Pr2O3
CeO2
Al2O3
Figure 2.2. Heat of CO2 adsorption versus electronegativity of the metal oxide cation. (Reprinted
with permission from Horiuchi et al. 1998. Copyright 1998 Elsevier.)
Table 2.4. Electronic properties of simple metal oxides

BINDING ENERGY (eV)
ELEMENT OXIDE CATION ORBITAL ELEMENT (O1s) OPTICAL BASICITY
(a)
Be BeO Be2+ 1s2 113 — 0.375
B B2O3 B3+ 1s2 191 533.2 0.425
P P2O5 P5+ 2p6 133 533.5 0.33 (0.40)
Si SiO2 Si4+ 2p6 102 532.8 0.50
Al Al2O3 Al3+ 2p6 74 531.2 0.60
Mg MgO Mg2+ 2p6 51 530.9 0.68
Ge GeO2 Ge4+ 3d 10 31 531.3 0.70
Ga Ga2O3 Ga3+ 3d 10 20 530.6 0.755
(b)
Li Li2O Li+ 1s2 56 — 0.87
Ca CaO Ca2+ 3p6 25 529.8 1.00
Sc Sc2O3 Sc3+ 3p6 31 — 0.87
Ti TiO2 Ti4+ 3p6 37 529.7 0.97
V V2O5 V5+ 3p6 40 530.0 1.04
Mn MnO Mn2+ 3d 5 — 529.8 0.95
Fe Fe2O3 Fe3+ 3d 5 — 530.0 1.02
Co CoO Co2+ 3d 7 — 529.9 0.98
Ni NiO Ni2+ 3d 8 — 530.0 0.915
Cu CuO Cu2+ 3d 9 — 530.3 1.10
Zn ZnO Zn2+ 3d 10 10 530.3 1.08
Y Y2O3 Y3+ 4p6 25 529.3 0.99
Zr ZrO2 Zr4+ 4p6 29 529.9 0.825
Nb Nb2O5 Nb5+ 4p6 35 — 1.05
Mo MoO3 Mo6+ 4p6 38 530.4 1.07
In In2O3 In3+ 4d 10 19 530.1 1.07
Sn SnO2 Sn4+ 4d 10 25 530.1 0.85
Te TeO2 Te4+ 5s2 14 530.5 0.93
Ce CeO2 Ce4+ 5p6 18 529.1 1.01
Ta Ta2O5 Ta5+ 4f 14 25 — 0.94
W WO3 W6+ 4f 14 34 530.2 1.045
(c)
Na Na2O Na+ 2p6 31 529.7 —
Sr SrO Sr2+ 4p6 20 529.0 1.14
Cd CdO Cd2+ 4d 10 11 528.6 1.115
Sb Sb2O3 Sb3+ 5s2 7 — 1.18
Cs Cs2O Cs+ 5p6 12 529.4 —
Ba BaO Ba2+ 5p6 16 528.2 1.22
Pb PbO Pb2+ 6s2 3 529.7 1.18
Bi Bi2O3 Bi3+ 6s2 8 — 1.19
Source: Reprinted with permission from Dimitrov and Komatsu 2002. Copyright 2002 Elsevier.
Table 2.5. Classification of simple oxides (group of oxides) according to their electronic properties
OXIDE ION O1S BINDING CATION METAL-BINDING BULK INTERACTION

OXIDES POLARIZABILITY ENERGY POLARIZABILITY ENERGY BASICITY PARAMETER BONDING
Semicovalent Low High Low High Acidic Strong Large overlap
(predominantly interionic between O2p and
acidic oxides) interaction valence metal orbitals
BeO, B2O3, P2O5,
SiO2, Al2O3, MgO, Strong covalent
GeO2, Ga2O3 bonds
Ionic (basic) High Medium High Low Basic Mainly weak Smaller overlap
Li2O, CaO, Sc2O3, range interionic between O2p and
TiO2, V2O5, MnO, interaction metal valence orbitals
Fe2O3, CoO, NiO,
CuO, ZnO, Y2O3, Bonds with increased
ZrO2, Nb2O5, MoO3, ionicity
In2O3, SnO2, TeO2
Very ionic Very high Low Very high Very low Very Very weak Small overlap
(very basic) basic interionic between O2p and
Na2O, SrO, CdO, interaction metal valence orbitals
Sb2O3, Cs2O, BaO
PbO, Bi2O3 Very ionic bonds
Source: Reprinted with permission from Dimitrov and Komatsu 2002. Copyright 2002 Elsevier.
The second main group includes so-called ionic or basic oxides such as CaO, In2O3, SnO2, and
TiO2, and most of these transition-metal oxides show relatively high oxide ion polarizability, O1s bind-
ing energy in a very narrow medium range, high cation polarizability, and low metal (or nonmetal) bind-
ing energy. Their bulk basicities vary in a narrow range and are close to that of CaO.
The third group of very ionic or very basic oxides includes CdO, SrO, and BaO, as well as PbO,
Sb2O3, and Bi2O3, which possess very high oxide ion polarizability, low O1s binding energy, very high
cation polarizability, and very low metal (or nonmetal) binding energy. Their bulk basicities are higher
than that of CaO, and the interionic interaction is very weak, giving rise to the formation of very ionic
chemical bonds.
It is necessary to note that this metal oxide classification is very productive from our point of view,
and can be applied for understanding the behavior of metal oxides in interacting with analytes that have
different properties. For example, the results presented in Figure 2.2 may be understood in the frame
of this approach.
“Redox” properties (ionosorption processes) is another characteristic which can be used to describe
the surface behavior of metal oxides during their interaction with gas and liquid surroundings. “Redox”
reactions involve an electron transfer, either directly or through the removal or addition of an oxygen
atom in the metal oxide lattice. Pre–ransition-metal oxides (e.g., MgO) are expected to be quite inert,
since they can neither be reduced nor oxidized easily. In terms of electronic structure, this property is
related to the large band gap, which means that neither electrons nor holes can be formed easily.
On the other hand, the post–transition-metal oxides ZnO, In2O3, and SnO2, as well as most tran-
sition-metal oxides, are active in redox reactions, since the electron configuration of the solid may be
altered. Stoichiometric ZnO, SnO2, and d 0 transition-metal oxides may be reduced but not oxidized.
Reaction with oxidizing species such as O2 is therefore expected only with samples that have been bulk
reduced or in which the surfaces have been made oxygen-deficient. The detailed behavior with reducing
molecules may be different for different metal oxides. On ZnO, the reduction leads to the formation of
free carriers, which greatly increase the surface conductivity, a fact that is crucial for sensor applications.
Reduction of transition-metal oxides, by contrast, is more likely to lead to local changes in electron
configuration, with electrons trapped at specific ions in the surface region. For SnO2 there is the pos-
sibility of both localized formation and free carriers. In this case it is necessary to remember that the d n
transition-metal oxides may, in principle, be both reduced and oxidized, so that their redox chemistry is
likely to be particularly complex.
For solid-state conductometric gas sensors, it is important that the change in electroconductivity
takes place exactly through ionosorption processes, because ionosorption means charge carrier ex-
change with the bands of the bulk solids. Local bonding in principle does not affect the conductivity
of the semiconductor, because such injection or capture of electrons is not implied. Local bonding of
an adsorbate to a surface stats may affect the conductivity only indirectly, because the surface state may
exchange electronic carriers with the bands of the semiconductor (Henrich and Cox 1994).
For stoichiometric metal oxides in which the metals are in their highest oxidation state, the metal at-
oms have lost all their valence electrons. Then they can be understood as insulators. The valence band has
a predominant bonding character, and the crystal orbitals are mainly localized on O2−. The conduction
band, on the contrary, has a predominantly antibonding character, and the crystal orbitals are localized
mainly on Men+. The most stable surfaces do not show intrinsic surface states in the band gap (Henrich
and Cox 1994; Calatayud et al. 2003). However, when the oxide deviates from the stoichiometric state
because of the presence of defects such as vacancies or adatoms (atoms that lie on the surface and there-
fore can be thought of as the opposite of surface vacancies), the oxidation state of the surface atoms
varies and the electron count does not correspond to that of an insulator. All the atoms are not in their
highest oxidation state. If oxygen vacancies are present, the metal oxide is reduced, some electrons filling
the bottom of the conduction band or levels in the gap. If oxygen adatoms are present, the valence band
is not completely filled (and some O atoms have an oxidation state of −1 instead of −2).
According to Post et al. (1999), if initially there are perfect surfaces, adsorption through an acid–
base mechanism is the best way to maintain adsorbed species. Electron-rich molecules (Lewis bases)
will interact at the cationic site (Me+), and electron-poor ones (Lewis acids) will interact at the anionic
site (O2−). MgO and TiO2 surfaces clearly appear to be predominantly acidic, and molecules that do not
dissociate generally bind to the metal cation. Sites of low coordination are in general more reactive than
sites of high coordination. The former is more common, and metal oxide surfaces are predominantly
acidic surfaces when organic molecules are adsorbed without dissociation (Minot 2001). The adsorp-
tion on the surface of oxygen atoms involves only the cations that result from heterolytic cleavage, the
metal atoms and a few specific adsorbates. Acid–base reactions do not modify the oxidation states of
the atoms. In a molecular orbital description, the occupied orbitals of the base are stabilized and the
vacant orbitals of the acid are destabilized. It follows that stoichiometric oxides remain insulating. The
electron count for the whole system is determining.
The reactivity at surfaces that are not stoichiometric differs from that on ideal surfaces. The differ-
ence originates not only from the modification of the coordination number but also from the electron
count. In this case (a defective surface), the most favorable adsorption scheme is the redox mechanism.
It is thus generally believed that the presence of defects, adatoms or oxygen vacancies, creates active
sites and enhances the surface activity. Oxygen vacancies modify the adsorption scheme by decreasing
the coordination of the surface atoms and by modifying the electron count. The first factor enhances
the activity of the acidic sites close to the vacancy. The second factor, the electronic factor, can operate
in both ways, either improving or diminishing the reactivity. When the oxide is irreducible, the elec-
tron pair is trapped in an F center that is extremely basic. The geometry of the oxide does not change
extensively, the electron pair replacing the anion and preserving the structure of alternating charges.
However, Calatayud et al. (2003) believe that the acid–base mechanism of adsorption at defective sur-
face is nevertheless still possible. The deviation from the stoichiometry in that case modifies the acidic
and basic properties of the surface atoms.
According to a results presented by many researchers, molecules that adsorb without dissociation
always bind to one or several metal cations (Minot et al. 1995; Minot 2001). The oxygen ions, anionic
sites with a formal charge of −2, are therefore less reactive. They should adsorb Lewis acids, but we
have seen that CO2 is not a strong enough Lewis acid for that. Adsorption at the O2−.site is the excep-
tion. On terraces, it takes place as a secondary interaction complementing the main interaction with the
metal cation, or it happens at very uncoordinated oxygen sites. Basic properties are revealed when the
molecules dissociate and when the cationic fragment adsorbs at the oxygen surface sites. However, the
adsorption on O2− of bases or radicals requires electron transfer, either a pair or a single electron, to
the cations of the metal oxide—i.e., the adsorption takes place through a redox mechanism. Calatayud
et al. (2003) have shown that this is easier for an electron-acceptor adsorbate and more difficult for an
electron-donor one. The adsorption of a radical species is incompatible with the electron count being
maintained in the stoichiometric metal oxide. A first possibility is to couple the electrons and form
two opposite ions adsorbed on the two surface sites, as for H2/MgO, using an acid–base mechanism.
Another possibility for adsorbing radicals is via an electron transfer to/or from the metal oxide (a re-
dox mechanism). Adsorptions with donor-like or acceptor-like adsorbates implies a redox mechanism
and an electron transfer from (to) the adsorbate to (from) the metal oxide. After accepting electrons,
acceptor-like moieties behave like Lewis bases and bind to the metal cations. When the adsorbate is an
electropositive group (a donorlike adsorbate such as NO), the unpaired electron can be transferred to a
reducible metal cation of the metal oxide. When the radical is an electronegative atom (an acceptor-like
adsorbate such as Cl), it can capture an electron from the metal oxide, provided that it is in a reduced
form. Then, the first adsorption corresponds to a Lewis acid (such as NO+) on the oxygen atom, and
the second one to that of a Lewis base (such as Cl−) on the metal atom.
SO2 chemisorption on the surface of metal oxides may also be understood in the frame of Lewis
acid/base properties (Ziolek et al. 1996). The amount of SO2 chemisorbed on metal oxides such as
MgO, Al2O3, ZrO2, TiO2, and CeO2 corresponds to the Lewis basicity strength. On all metal oxides
that have been studied by IR spectroscopy, SO2 chemisorption leads to the creation of Brønsted acidic
sites. Results of catalysis have shown that the strength of these acid sites depends on the metal oxide: It
is quite weak on MgO, more pronounced on γ-Al2O3, ZrO2, and TiO2 anatase, and high on rutile. The
creation of such sites on ceria is accompanied by a drastic modification of its redox properties.
Adsorption of metal atoms on the oxide surface has been extensively studied during the last de-
cade. It has been established that a metal cation has acidic properties and binds to the oxygen anions
of the surface. The adsorption of neutral metal atoms, on the other hand, is weak. The charge transfer
from the metal to the cations that would equilibrate the charges is not very important. This charge trans-
fer also depends on the difference in electronegativity and is thus weak in first approximation. Electron
transfer is significant for K and Ca only when the metal oxide is oxidized. Considering the first row of
the transition metals, the curve of the heats of adsorption as a function of the atomic number (Figure
2.3) resembles that of the cohesive energies. They are small for the alkali and the noble metals. They
are larger for the transition metals, with a depression for Cr due to the high stability of the atom in the
high-spin state. Heats of adsorption are from Krilov and Kisilev (1981).
It was shown by Calatayud et al. (2003) that even though interaction of the metal atom with O2− is
weak, it splits the metal atomic levels. The σ levels, 4s, and 3dz2 (or 4pz ) are mainly nonbonding and
antibonding M–O orbitals; it is important that the latter remains unoccupied. When it is occupied by
two electrons (as in Zn), the heat of adsorption is negligible. When it is occupied by one electron (as
in Cr, Mn, and Cu), the heat of adsorption is weak. The heat of interaction is large for Ti, since there
are no antibonding occupied orbitals, while the π levels are occupied for V–Co (high-spin state). The
trend from K to Ti is due to the variation of electronegativity. The lower the metal atomic levels are, the
stronger is the interaction with the oxygen orbitals. This also explains the high heat of adsorption for Ni
and Cu. Since we qualified the metal adsorption as a physisorption, the effect is of small amplitude.
Finally, it is necessary to note that the reaction medium plays an important role in adsorption
processes. Clean surfaces exist in dry conditions when the surface is exposed at low gas pressure. In
hydrated conditions, when the metal oxide surface is covered by water, the surface sites are not available
for other molecules. As a consequence, adsorption at a hydroxylated surface is possible only in two situ-
Figure 2.3. Binding energies of the transition metals on the MgO(100) surface at low coverage as a
function of their atomic numbers.
ations: Either the adsorption is strong enough to allow the desorption of the water molecules that are
bound directly to the clean surface (or a substitution of hydroxyl groups), or it occurs directly on these
groups through hydrogen bonds. For example, experimental research reported by Leblanc et al. (2000)
has shown that the presence of hydroxyl groups at the surface of SnO2 strongly modifies the nature and
the content of species formed by NO2 adsorption. In the presence of hydroxyl groups, an additional
interaction is observed, with the formation of hydrogenonitrate species. These groups appear to be
more stable that the unidentate nitrato complexes that form in the absence of OH groups.
3. SURFACE PROPERTIES OF SENSING MATERIALS

3.1. ELECTRONIC PROPERTIES OF METAL OXIDE SURFACES
The work function, W, is one of the most important properties of a surface. The work function is
defined as the energy necessary to remove an electron from the Fermi level in a material and put it at
rest in vacuum an infinite distance away from the material. Averaged work functions measured for clean
surfaces of some metal oxides are presented in Table 2.6.
The work function is extremely sensitive to the state of the surface (Henrich and Cox 1994). In
fact, for metal oxides, W is so sensitive that its absolute value has little significance. However, this is not
to say that work functions of metal oxides are entirely useless. Relative changes in W reflect variation of
surface parameters and contain important information about the surface state. Therefore, measurement
of these changes is often used both in studies of adsorption effects (De Fresart et al. 1982; Barsan et al.
1999; Barsan and Weimar 2003) and for design of room-temperature gas sensors (Doll et al. 1998). It
should be noted that, in selecting a material for gas sensors, the absolute value of the surface potential
is not as important as its change upon chemisorption.
Table 2.6. Average work functions of some metal oxides
WORK FUNCTION
OXIDE (eV)
BaO 1.6–2.7
CaO 2.5–2.8
ZnO 4.2–4.9
SnO2 4.3–4.7
TiO2 5.1–5.9
ZrO2 5.0–6.6
SrTiO3 4.2–5.2
V2O3 4.9
Fe2O3 5.4
NiO 4.3–4.65
Concentration of native surface states is another metal-oxide surface parameter that is important
for chemical sensors, especially solid-state gas sensors. To achieve effective operation of solid-state gas
sensors, the concentration of these states must be minimized, because in this case the surface Fermi level
will not be pinned. A low concentration of surface states creates a possibility for modulating the surface
potential of a semiconductor with a change of surrounding atmosphere, because the charge of the na-
tive surface state becomes commensurable with, or less than, the charge of the chemosorbed particles.
A similar correlation with the concentration of the native surface state was observed for Schottky
barrier heights at the metal–semiconductor interface (Shaw 1985). The height of the Schottky barrier at
the semiconductor surface can be represented as
U S = K (WMe − WS ) (2.1)
where WMe and WS are the work functions for emission of an electron from a metal and a semiconduc-
tor, respectively, and K is a chemical parameter that depends on the nature of the semiconductor. The
parameter K is defined as a coefficient of a linear dependence relating US and (WMe − WS), and can be
interpreted as a demonstration of the sensitivity of the electronic properties of a semiconductor to the
state of its surface. Figure 2.4 shows the parameter K in relation to the difference in electronegativy ΔΧ
(by Pauling) between the anion and cation that form the semiconductor. The abrupt change in the K
value at ΔΧ = 0.8 corresponds to a transition from materials with covalent bonding (Si, Ge, GaAs) to
those in which ionic bonds predominate (ZnO, SiO2, SnO2 ).
If we employ this correlation for gas sensing, the greatest sensitivity to changes in the concentra-
tion of molecules adsorbed onto the surface and, consequently, to changes in the gas-phase composi-
tion, will be exhibited by materials with predominantly ionic bonding—e.g., materials such as CdS,
ZnS, and SiO2. However, standard ionic semiconductors such as ZnS and CdS have low chemical and
thermal stability, which limits their possible application in the design of chemical sensors. Metal oxides,
which also have low concentrations of native surface state, have much higher thermal and temporal
stability, which allows their successful use in chemical sensors. Furthermore, this statement is true only
for conducting oxides. Oxides that are characterized by exceedingly high resistivity, such as SiO2 and
Al2O3, are not used as sensing materials in resistive gas sensors because of the difficulties encountered
in measuring electrical conductivity. Some fluorides, such as LaF3, are more thermodynamically stable
than oxides in air at high temperature, but they often are more volatile.
Brinzari et al. (1999a, 2000, 2001) illustrated the influence of electronic parameters of metal oxide
surfaces on gas-sensing effects, for example, CO detection by SnO2 gas sensors. Unfortunately, reliable
data about SnO2 surface properties, which would be necessary for theoretical simulation, are practically
nonexistent in the literature, so this example can serve only as preliminary guidance for experimental
confirmation and further development.
Brinzari et al.’s (2001) results for Us(PCO) dependencies are presented in Figure 2.5. Here we can see
that the density of native surface states, Nss, really is a factor in determining the behavior of Us: The slope
of the Us(PCO) curve decreases with an increase of Us. Moreover, it was shown that when Nss → N*, where
N* is the concentration of adsorption sites, the surface charge associated with SnO2 native defects and
adsorbed charged species ( Q = QSS + Q* ) is not affected by the surrounding ambient gas. In other words,
the surface potential is pinned and is not sensitive to a change in gas atmosphere. Further, the increase of
density of the native surface states also leads to an increase in the sensor’s threshold of sensitivity.
3.2. ROLE OF ADSORPTION/DESORPTION PARAMETERS IN

GAS-SENSING EFFECTS
Much research has shown that for effective chemisorptions, sensor materials should have particular
combinations of adsorption/desorption parameters for oxygen and detecting gases (Brinzari et al.

ZnO SiO2 KTaO3
ZnS
Ga2O3
GaS SnO2 Al2O3
CdS
CdTe ZnSe
SiC
GaP CdSe
Ge GaAs
InSb
, eV
Figure 2.4. Influence of electronegativity on the value of K in US = K ΔX. (Adapted with permission
from Kurtin et al. 1969. Copyright 1969 American Physical Society.)

SnO2
10-4 10-3 10-2 10-1

r
Partial pressure, PCO
Figure 2.5. Influence of the concentration of surface states on the surface potential dependencies
of undoped SnO2 films on relative CO pressure (PCO /PO2): PO2 = 2 × 104 Pa; βCO/βRO = 10−3; αCO/βRO =
1–10−5; Nd = 1019 cm−3; N* = 1015 cm−3; Toper = 300°C: (1) Nss = 0; (2) Nss = 6 × 1012 cm−2; and (3) Nss
= 1013 cm−2 (Adapted with permission from Brinzari et al. 2001. Copyright 2001 Elsevier.)
1999a, 2000, 2001). It is known that the smaller the activation energy of chemisorption and the higher
the activation energy of desorption, the greater is the gas-sensing effect of adsorption-type sensors
(Lundstrem 1996; Brinzari et al. 2000). At the same time, we have to take into account that at excessively
large activation energy of adsorbed species, the desorption might lead to a considerable increase of
recovery time after changing the surrounding atmosphere, which is not acceptable for practical applica-
tions. That is why for chemisorptions-type chemical sensors a material with optimal activation energy
of desorption for the given work temperature is needed. Otherwise, to reduce recovery time, it would
be necessary to increase the working temperature, which could lead to a sharp drop of the sensor’s reli-
ability and durability. For instance, simulations carried out by Brinzari et al. (2000) showed that at an
operating temperature of 300°С, an activation energy for oxygen desorption of ~1.0 eV is optimal.
Brinzari et al. (2000) also considered the influence of some adsorption/desorption parameters on
the surface potential (ΔUs ) and SnO2 conductivity (ΔG ) during CO detection. The pattern of the influ-
ence of the main physical-chemical parameters of a reducing gas (R) on ΔUs and ΔG is presented in
(2.2), where ↑ indicates an increase and ↓ indicates a decrease:
αR/αO↑, β4↑, βRO↑, N*↓, Nss↓, β3↓, βR↓ ⇒ ΔUs↑, ΔG↑ (2.2)
where αR and αO are the coefficients of R and O2 adsorption; βR and βRO are the coefficients of R and
RO desorption; β3 and β4 are the coefficients of charging and neutralization of RO; and N* and Nss
are the total number of adsorption sites and sites originating from native (biographic) surface charge.
The scheme shows the directions of adsorption/desorption parameters changes, which are necessary
to improve sensor response.
According to the model of solid-state gas sensors presented by Brinzari et al. (2000, 2001), the
temperature dependence of gas sensitivity for undoped metal oxide films (SnO2 ) is determined by the
following processes:
• Dissociative adsorption of oxygen (molecular oxygen does not interact with reducing gas). This
process begins at T > 170°C.
• Adsorption of R molecules (by impact or another mechanism) and conversion to RO (catalytic
oxidation). This process depends on the type of reducing gas—for example, in the case of CO
molecules, it begins at T >250°C; the difference between CO and H2 adsorption results from an
associative mechanism for H2 adsorption.
• Decrease of conversion rate of R (reverse desorption of R).
• Desorption of products of catalytic reactions.
• Desorption of chemisorbed oxygen.
The first two processes determine the drop in sensitivity at low operating temperatures, the last three
determine it at high temperatures (see Figure 2.6). Competition among these processes determines both
the temperature of maximum sensitivity and the half-width of the S(Toper ) sensitivity curves.
Brinzari et al. (2000) simulated the influence of some adsorption/desorption (A/D) parameters on
S(T ) dependence. The influence of some A/D parameters on S(T ) dependence is shown in Figure 2.7.
Here ECO is the activation energy of CO adsorption, and qCO is the activation energy of CO desorp-
tion. One can see that the decrease of activation energy of CO and O2 adsorption processes may really
change the sensor sensitivity and shift greatly the temperature of the maximum sensor response. In
addition, this process may be accompanied by a considerable decrease of response and recovery times.
Brinzari et al. (2000) found that the position of the Us(PCO) curve along the PCO axis depends on the
αCO/αO ratio and N*. The US(PCO ) curve shifts in the direction of lower PCO with an αCO/αO increase
CO
3 ~q
CO CO q CO 2
~ E CO E O 4
O
or O
Figure 2.6. Adsorption/desorption parameters controlling temperature dependence of metal-oxide

sensor response to CO.

103
SnO2
102
101
100
Figure 2.7. Simulation of the influence of adsorption/desorption parameters on temperature depen-

dence of SnO2 conductivity response to CO. ECO: 1, 0.9 eV; 2, 0.85 eV; 3, 0.8 eV; 4, 0.7 eV; 5, 0.6
eV. qCO: 1, 2, 3, 4, 1.6 eV; 5, 1.4 eV. (Reprinted with permission from Korotcenkov 2007. Copyright
2007 Elsevier.)
and an N* decrease (curve 3 in Figure 2.5). This means that indicated changes of adsorption/desorption
parameters promote an increase in sensor response. However, in this case we need to find the optimum
value for N*, because with a decreased number of adsorption sites, a decrease of sensor response is
observed.
The main methods of influencing the electronic parameters of adsorbed species are changing the
composition of the metal oxide film, i.e., change from simple binary to multioxide films, and/or dop-
ing the metal oxide surface by adding catalyst particles (Yamazoe 1983; Korotcenkov et al. 2003). For
example, doping with metal catalyst additives (Pd, Pt) seems to result in a decrease of αCO/αO, at least
on metal catalyst particles, because modifying the surface with noble metals increases adsorption of
dissociative oxygen. The impact adsorption of CO (by Redeal-EIley mechanism [Masel 1996]) is less
affected by doping. However, due to the competition of molecular and atomic forms of oxygen on the
SnO2 surface, there is an “apparent” decrease of αO and increase of αCO/αO with surface doping by
these noble metals. The result is shifts of both the S(T ) and S(PCO) curves at lower values of Toper and
PCO. Simulations of S(PCO) dependencies for doped SnO2:Pd films are shown in Figure 2.8.
Analysis of gas detection reactions indicates that materials for gas sensors should also be stable to
surface poisoning, i.e., they should have acceptable desorption energy of catalytic reaction products.
Otherwise these products may accumulate on the surface of sensitive elements (see Figure 2.9), and
parameters of interest may worsen. “Sulfur poisoning” (Somorjai 1981; Satterfield 1991) is one such type
of poisoning. One can also consider species such as lead tetraethylene, metal hydride vapors, chlorides, and
metal organics (Madou and Morrison 1987) as reagents which could poison active sites of the sensors.
In addition, it is desirable that surface coking does not take place. This process is one of the
most important reasons for a decrease of catalytic activity during use (Somorjai 1981; Satterfield 1991).
Another source of poisoning is other compounds that may be reducible to metals and elements under

SnO2
102
101

100
10-1
10-6 10-4 10-2 100

r
Partial pressure, PCO
Figure 2.8. Simulation of the influence of metal catalysts on sensor response dependencies on CO
r
pressure PCO = PCO /PO2; PO2 = 2–10−4 Pa; Nss = 6 × 1012 cm−2; Nd = 1019 cm−3; Toper = 300°C; N* = 2 ×
10 cm : 1, undoped film; 2, modified film [αO(doped)/αO(undoped) = 105; αCO(doped)/αCO(undoped)
13 −2
= 102 ]. (Adapted with permission from Brinzari et al. 2001. Copyright 2001 Elsevier.)
reaction conditions. Elements such as As, Fe, P, etc., may alloy with the catalytically active metals and
metal oxides, reducing their effectiveness. Arsenic is present in trace amounts in various feedstocks.
Iron is a ubiquitous material of construction and can be a serious poison to platinum-group catalysts.
Phosphorus compounds are typically used in lubricating oils for pumps, blowers, fans, and other ma-
chinery. In this context, one should note that metal oxides are more resistant to certain types of poison-
ing (especially halogens, As, Pb, and P) than noble metals. These effects of poisoning have been dis-
cussed in terms of occupancy (site blocking) and of electronic effects. It is now quite clear that strong
electronic effects play a fundamental role in these changes (Ustaze et al. 1998).
Analogous requirements exist for materials for adsorption sensors, for example, such as SAW and
cantilever sensors, where the change in weight of the sensing element is a determining factor (Houser
Figure 2.9. Schematic illustration of surface reactions during H2 detection.

2001). It is known that the role of the sensing material in such devices is selectively and reversibly to
sorb an analyte of interest from sampled air or liquid, and to concentrate it to achieve lower concentra-
tion detection capabilities. Therefore, maximum and reversible sorption of specific analytes or classes
of analytes, with rapid sorption kinetics and minimal sorption of interferents, are key aims in the de-
velopment of a successful chemoselective coating for SAW and work-function sensors (Thomson and
Stone 1997).
As an example, the capability of various metal oxides for adsorption of isopropanol and methanol
is shown in Table 2.7. These measurements were conducted by Wachs and co-workers (Badlani and
Wachs 2001; Kulkari and Wachs 2002). The active surface site densities were calculated from the gain in
weight of the samples after adsorption of isopropanol and methanol for 1 h at 110°C and were found
to be 0.2–4 μmol m−2 on the majority of metal oxide surfaces Some of the active metal oxides (MgO,
La2O3, Cr2O3, Sb2O3) possessed higher active site density, which gives those oxides an advantage for
the design of adsorption sensors sensitive to isopropanol and methanol. Cr2O3, WO3, and BaO show
somewhat lower active surface site density. SiO2 is extremely unreactive and has a low Ns in spite of
having a high surface area.
However, we have to note that this conclusion cannot be considered to hold in all cases. Every
technical task requires individual solution, considering both the nature of the gas tested and the operat-
ing conditions. For example, CO2 adsorption (see Figure 2.2) is entirely different.
3.3. CATALYTIC ACTIVITY OF SENSING MATERIALS
In many chemical sensors, the sensitivity is determined by the effectiveness of catalytic reactions with
the gas being detected at the surface of the sensing material. Results of experiments directed toward
simultaneous control of sensor response and efficiency of conversion of the gas detected confirm this
conclusion. Therefore, high catalytic reactivity of the surface, and especially selectivity of this reaction
to detected gas, are important advantages for a sensor material. As a result, control of the catalytic activ-
ity of a material is often the main method used in preliminary evaluation of the material’s suitability as a
gas sensor, and in determining operating temperatures. As a rule, maximum sensor response is observed
at a temperature corresponding to 50% conversion of detected gas (Yamazoe et al. 1983; Li et al. 1999;
Cabot et al. 2002) (see Figure 2.10).
It is necessary to note that the maximum catalytic activity to different gases may be observed at
different temperatures. This is in fact a favorable property for gas-sensing materials, because by chang-
ing the operating temperature we may be able to influence the selectivity of gas sensors. For example,
the peak in sensitivity (oxidation) for methane is often at higher temperatures than for CO and other
hydrocarbons, suggesting that a higher temperature may be desirable for methane-selective sensors,
while a lower temperature will be desirable for CO-selective ones.
However, in spite of the obvious correlations between chemical sensing and heterogeneous ca-
talysis, the choice of a material for gas sensor applications is not determined just by catalytic activity:
Catalytic activity is an important parameter, but not a determining one; in confirmation, see Figures 2.11
and 2.12. Figure 2.12 (Krilov and Kisilev 1981) presents results on the relative catalytic activity of metal
oxides in logarithmic units for different reactions.
Table 2.7. Number of active surface sites on metal oxides during isopropanol and
methanol adsorption
NUMBER OF ACTIVE SITES

(μmol m –2)
ISOPROPANOL METHANOL
OXIDE ADSORPTION ADSORPTION
MgO 8.9 22.5

CaO 2.7 5.4
SrO — 4.3
BaO 0.7 3.8
Y2O3 3.1 4.9
La2O3 5.3 34.1
TiO2 3.2 3.7
ZrO2 2.6 1.1
HfO2 3.1 2.6
CeO2 2.5 4.2
V2O5 1.6 0.7
Nb2O5 2.0 2.6
Ta2O5 3.6 4.6
Cr2O3 0.7 12.4
MoO3 1.2 0.8
WO3 0.7 2.3
Mn2O3 2.6 1.6
Fe2O3 7.9 3.7
Co3O4 3.9 2.8
Rh2O3 2.7 8.1
NiO 2.8 6.5
PdO 8.3 9.9
PtO 2.1 7.2
CuO 4.3 8.4
Ag2O 5.8 12.0
Au2O3 36.5 —
ZnO 1.7 0.3
Al2O3 2.8 5.6
Ga2O3 3.4 4.1
In2O3 2.0 2.7
SiO2 0.5 0.2
SnO2 2.1 1.6
Bi2O3 2.0 2.1
P2O5 — 3.6
Sb2O3 — 11.3
TeO2 — 4.1
Source: Data from Badlani and Wachs 2001;
Kulkame and Wachs 2002.
SnO2: Me
Catalyst temperature, T50 , oC

Figure 2.10. Correlation between catalysis temperature of 50% conversion (T50 ) and temperature of
SnO2 sensor response maximum (Tm ). (Reprinted with permission from Korotcenkov 2007. Copyright
2007 Elsevier.)
SnO2
Co3O4
CaO
La2O3
NiO
CuO WO3
In2O3
Fe2O3
ZnO
SnO2 La2O3
In2O3
CuO NiO Cr2O3 TiO2
Fe2O3 Al2O3
Co3O4
Cr2O3
TiO2 MgO

Figure 2.11. (1) Reaction rate of nitrous oxide decomposition at 773 K in oxygen-containing
atmosphere (8% O2, 1% N2O) as a function of the heat of oxide formation. (2) Sensitivity to 300 ppm
N2O for the various single metal oxides. (Experimental data from Satsuma et al. 2001 and Kanazawa
et al. 2001.)
Numerous experiments conducted by various authors have indicated that, as a rule, oxides with
electron configurations of d 3 (Cr2O3, MnO2) and d 6–8 (Co3O4, NiO) are the most catalytically active.
Minimum activity is observed for oxides with electron configurations of d 5 (Fe2O3, MnO), d 0 (CaO,
Sc2O3, TiO2), and d 10 (ZnO, Cu2O), although the activity of oxides with an electron configuration of
d 5 is much higher than the activity of oxides with d 0 and d 10 configurations. In practice, however, as we
have mentioned before, metal oxides with electron configurations such as d 0 (TiO2), d 10 (ZnO, SnO2,
Cu2O, Ga2O3), and more rarely d 5 (Fe2O3), which are least active in terms of catalysis, are being con-
sidered today as the most promising gas-sensing materials. Therefore, catalytic activity, in spite of the
temperature of maximum sensitivity being equal to the temperature of 50% conversion of detecting
gas, cannot explain the choice of d 10 and d 0 oxides as base materials for conductometric gas sensors.
Selection is in fact determined by the totality of the material properties, not just one. For example, based
on the data presented by Krilov and Kisilev (1981) (see Figure 2.12), one can conclude that oxygen
bond energy at the surface of transition-metal oxides of the fourth period is a parameter that is more
useful than catalytic activity in terms of determining the adaptability of particular metal oxides for
solid-state gas sensor design.
TiO2 Cr2O3 Fe2O3 NiO CuO Cu2O

CaO Mn2O3 Co3O4 ZnO
V2O5 Ga2O3
d0 d3-4 d5 d6-8 d10
Figure 2.12. Relative catalytic activity of metal oxides in different reactions (1–4) and bonding en-
ergy of oxygen on the surface of transition metal oxides (5): 1, N2O decomposition; 2, isopropanol
dehydration; 3, CO oxidation; 4, reaction of H2–D2 exchange. (Experimental data from Krilov and
Kisilev 1981.)
At the same time, it is true that the choice of a metal oxide as an additive to modify the properties
of another metal oxide is often connected to the catalytic properties of the oxides (Yamazoe et al. 1983;
Oelerich et al. 2001). For example, the catalytic activity toward a particular gas is the most important
parameter for applications in membranes and is used to improve the selectivity of sensor response.
Room-temperature (RT) chemical sensors is another field of application of catalytically active oxides.
Doll et al. (1998) found that RT work-function sensors based on catalytically active oxides such as CeO,
Fe2O3, and NiO show good operating parameters.
4. STABILITY OF PARAMETERS IN SENSING MATERIALS
4.1. THERMODYNAMIC STABILITY
Materials for chemical sensors that must work at high temperature have to possess high thermodynamic
stability. The better a material’s thermodynamic stability, the higher will be the temperatures at which
chemical sensors incorporating this material will work, especially in the presence of reducing gases.
Sensor materials with high thermodynamic stability should also have better temporal stability. This
condition can be attained by suppressing grain-size increases during use. In this case, the opportunity to
use materials with small crystallites exists, which is necessary to achieve both high sensitivity and a good
rate of response. Both high heat of formation and high melting temperature characterize such materi-
als. Table 2.8 (data from Samsonov 1973, 1982; Jonhnson 1982; Lamereaux et al. 1987; Weast et al. 1988;
Badlani and Wachs 2001; etc.) lists relevant parameters of the metal oxides that are most widely used
in design of chemical sensors. Here the heats of formation are normalized to the number of oxygen
atoms required for reaction. Oxides in air are in the lowest free-energy state of almost all metals in the
Periodic Table, which leads to their high thermodynamic stability.
The data in Table 2.8 can be considered as parameters that characterize the reactivity of elements
with respect to oxygen. Table 2.9 shows some other relevant data (Dean 1972; Samsonov 1973; Henrich
and Cox 1994). The elements are ordered according to the heat of formation of the most stable oxide,
avaluated as ΔHf in kJ mol–1 of oxygen atoms.
The more reactive metals are those with a more negative heat of oxide formation, lower down in
Table 2.8. They should be able to reduce the oxides of metals above them. There are, however, various
reasons why the predictions of bulk thermodynamics may not be followed. There is the possibility that
surface phases formed may have thermodynamic stability different from those of bulk oxides, but it is
also important to remember that surface reactions of this kind require extensive migration of atoms,
a process that may have a high activation energy (Henrich and Cox 1994). Such reactions are therefore
more likely at higher temperatures.
The thermal program reduction (TPR) technique may be used to probe the stability of different
metal oxides. In this method, diluted hydrogen is used to reduce metal oxides. Similar to methanol
chemisorption, hydrogen reduction of a metal oxide proceeds through dissociative adsorption of H2,
which reacts with lattice oxygen to form surface hydroxyl species. Subsequently, H2O leaves the surface
by eliminating the surface hydroxyl species. The TPR threshold temperatures of the metal oxides, which
reflect the reducibility of the metal oxides, are shown in Table 2.8. The initial reduction temperatures
Table 2.8. Thermodynamic stability of metal oxides
∆Hf FOR METAL OXIDE TEMPERATURE-

MELTING FORMATION PER OXYGEN PROGRAMMED
TEMPERATURE ATOM, −∆Hf @ 298 K REDUCTION, TPR THERMAL STABILITY IN
MATERIAL (°C) (kJ mol−1) (°C) OXYGEN ATMOSPHERE
SiC 2700
MgO 2800–2820 601.7 N.R. Thermally stable (T.S.)
СaO 2587–2620 635.1 300 T.S.
SrO 2430–2650 590.7 326 T.S.
BaO 1923–2015 553 330 T > 500°C → BaO2
Y2O3 586.2 325 T.S.
La2O3 2300 699.7 468 T.S.
TiO2 1855 470.8 N.R. T.S.
ZrO2 2690 547.4 N.R. T.S.
HfO2 2790 556.8 N.R. T.S.
CeO2 2727 544.6 594 T.S.
V2O5 690 311.9 550 T > 700°C, evaporates
with partial dissociation
Nb2O5 1512 381.1 N.R. T.S.
Ta2O5 1879 409.9 340 T.S.
Cr2O3 2300–2435 380.0 219 T.S.
MoO3 795 251.7 575 T > 650°C, sublimates
WO3 1470 280.3 544 T > 1000°C, sublimates
Mn2O3 1347 323.9 184 T > 750°C, decomposes
Fe2O3 1347 247.7 200 T > 1400°C, dissociates
Co3O4 1562 202.3 288 T > 900°C → CoO
Rh2O3 1115 95.3 100 T.S.
NiO 1957 245.2 278 T.S.
PdO 877 85.0 90
PtO 507 71.2 345 Decomposes
CuO 1336 157.0 268 T > 800°C, decomposes
Ag2O 187 30.6 200 T > 200°C, decomposes
Au2O3 26.9 T > 160°C, decomposes
ZnO 1800–1975 348 N.R. T.S.
Al2O3 2050 558.4 N.R T.S.
Ga2O3 1740–1805 360 320 T.S.
In2O3 1910–2000 308.6 350 T.S.
SiO2 1720 429.1 N.R T.S.
SnO2 1900–1930 290.5 500 T.S.
Bi2O3 817 192.6 400
P2O5 563 306.3 N.R. T > 359°C, sublimates
Sb2O3 655 233.2 563 Easily sublimates
TeO2 2127 162.6 355 T > 450°C, sublimates
N.R., no reduction detected between 150 and 700°C.
Source: Reprinted with permission from Korotcenkov 2007. Copyright 2007 Elsevier.
Table 2.9. Reactivity of elements toward oxygen

HEAT OF OXIDE FORMATION,
∆Hf (kJ PER MOLE O) ELEMENTS
0–50 Au, Ag, Pt
50–100 Pd, Rh
100–150 —
150–200 Ru, Cu
200–250 Re, Co, Ni
250–300 Na, Fe, Mo, Sn, Ge, W
300–350 Rb, Cs, Zn, V
350–400 K, Cr, Nb, Mn
400–450 Si, Ta
450–500 Ti
500–550 U, Ba, Zr
550–600 Al, Sr, La, Ce
600–650 Mg, Th, Ca, Sc
Source: Data from Dean 1972; Henrich and Cox 1994.
vary in the range from 100°C (Rh2O3) to more than 700°C (CeO2 ). Most of the bulk metal oxides ex-
hibit multiple peaks in their TPR profiles due to their multiple oxidation states when they are extensively
reduced. Note, however, that only the onset reduction temperatures are reported in Table 2.8.
For some metal oxides, such as HfO2, MgO, ZnO, TiO2, Al2O3, SiO2, ZrO2, P2O5, and Nb2O5,
no detectable H2 consumption was observed in the temperature range of 150–700°C. Some of these
samples (TiO2, Al2O3, ZrO2, and Nb2O5) probably experienced slight surface reduction because their
color changed after a TPR run and the color quickly disappeared when the sample was exposed to
ambient conditions (Badlani and Wachs 2001; Kulkari and Wachs 2002). It can be seen that the initial
reduction temperature is in full accordance with the heat of oxide formation, which characterizes the
oxide’s thermodynamic stability. The less stable the metal oxide is, the more easily the surface is reduced
to form oxygen adsorption sites. The higher the energy of a stable oxide’s formation is, the higher is
the initial reduction temperature. Therefore, sensors based on such oxides will have better parameter
stability in reducing atmosphere.
As has been mentioned, the initial reduction temperature is a very important parameter for chemi-
cal sensors, because if this temperature is exceeded, the metal oxide may be reduced to the metal during
interaction with a reducing gas. So, the applicability of such oxides as Сr2O3, Mn2O3, Fe2O3, and NiO
for high-temperature solid-state gas sensor design is very limited, because their working temperature in
a reducing atmosphere cannot exceed 200°С.
For more complex oxides there are also some correlations which may predict their thermodynamic
stability (Kreuer 1997). For example, for simple perovskites of the ABO3 type with alkaline-earth met-
als on the A site, it was established that the thermodynamic stability of oxides is determined mainly by
the choice of the B cation. In accordance with an increasing perovskite tolerance factor RA/RB (where
RA and RB are the ionic radiuses of the A and B cations), one observes increasing stability in the order
cerates → zirconates → titanates. Even higher stability of perovskites, apparently, can be expected upon
introduction of Nb+5 an the B position, due to a more advantageous perovskite tolerance factor.
There are also stability criteria for polymers. According to Sandier and Karo (1974), a polyamide
with high resistance to degradation should have the following properties: (1) high melting/softening
point, (2) low weight loss as determined by thermogravimetric analysis, and (3) a structure that is not
susceptible to degradative chain scission or intra- or intermolecular bond formation. Intensive research
is being carried out in this area. For example, when a segment of the aliphatic polymer’s main chain
is replaced by a ring segment, the melting temperature and hence the thermal stability increases due
to the decrease in the flexibility of the polymer chain (Bhuiyan 1984). However, accomplishing this is
quite complicated, because it is necessary to attain high stability while maintaining the high activity of
the polymer.
In addition, there are some fundamental restrictions to achieving the required thermal stability
with polymers. Figure 2.13 shows that the melting point of the polymer decreases as the chain length
increases. This means that a complicated polymer will inevitably have a lower melting temperature and
therefore lower stability. Figure 2.13 shows relevant data for nylon-type polymers (Bhuiyan 1984).
Results given by Ryabtsev et al. (1999) show the importance of high thermal stability. Fe2O3:Pt-
based sensors operating at Toper ≈ 200–250°С had maximum sensitivity to acetone compared to SnO2-,
CdO-, and Nb2O5-based sensors. However, Fe2O3:Pt-based sensors were not used in the instrument
prototype for acetone vapor analysis, due to the strong dependence of the long-term stability on oper-
ating temperature. Increasing Toper to higher than 250°C resulted in a sharp worsening of gas-sensing
characteristics. It is important to note that this behavior of Fe2O3-based sensors corresponds to data
presented in Table 2.8.
Figure 2.13. Influence of the number of carbon atoms in the monomer on the approximate melting
temperature of nylon-type polymers. (Experimental data from Bhuiyan 1984.)
4.2. CHEMICAL STABILITY

Sensing materials should be characterized by high chemical stability. This property means the absence
of corrosion during interaction with gases and solutions, i.e., an opportunity to work in corrosive me-
dia. The specificity of interaction of some materials with chemical reagents is given in Table 2.10 (data
from Samsonov 1973, 1982; Jolivet 2000; etc.).
The chemical activity of sensor materials is also important in medical applications. Sensing ele-
ments, as well as sensor construction elements, often contact patients’ blood, so prevention of infection
is critical in the application of chemical sensors in medicine.
When polymers were first considered for chemical sensors, it was feared that they might not pro-
vide the necessary durability in organic solvents. However, results presented by Matsuguchi et al. (1998)
showed that this need not be a problem. Polymers such as cross-linked polyamide (C-PI) and cross-
linked fluorinated polyamide (C-FPI) can have high stability against many strong solvents. After inter-
action with solvents such as saturated acetone vapor, formadehyde, ethylene oxide gas, chlorine-type
sterilizer, and ampholytic surfactant, the change in resistance of C-PI and C-FPI polymers did not
exceed 2–4%.
At the same time, however, excessive chemical inactivity of a material may actually create difficul-
ties during sensor design, when one wants to localize the sensing material on the surface of the chemical
sensor platform, i.e., when creating the sensor’s configuration. Such widely used oxides as tin oxide and
aluminum oxide have this increased chemical resistance. To create the necessary surface configuration,
one must use passive masks, dry etching, or elaborate special technological methods for local deposition
of these materials in required spots (Majoo et al, 1996; Semancik et al. 1996).
On the other hand, the chemical activity of some materials with respect to certain reagents can be
used to advantage. As an example, two-phase systems such as SnO2-CuO and SnO2-AgO may be used
in gas sensors that will be sensitive to H2S (Tamaki et al. 1998; Jianping et al. 2000). The high sensitivity
of sensors based on these materials is a consequence of the following reactions:
CuO + H2S ⇒ CuS + H2O↑ CuS + O2 ⇒ CuO + SO2↑ (2.3)
Or
AgO + H2S ⇒ AgS + H2O↑ AgS + O2 ⇒ AgO + SO2↑ (2.4)
These reactions lead to changes in chemical composition and physical properties of material, forming
intercrystallite interlayers in the gas-sensing matrix.
The same principle is used in designing solid-state chemical sensors for СО2 detection, for exam-
ple, using the La2O3/Li2CO3 (1/10) system (Bogue 2002). In the presence of CO2, the La2O3 is con-
verted to lanthanum carbonate, which alters the sensor’s conductivity.
La2O3 + 3CO2 ⇒ La2(CO3)3 (2.5)
It is important to note here that the use of such materials for sensor design is possible only when
the reactions are completely reversible and take place at an acceptable rate.
Table 2.10. Chemical activity of metal oxides
METAL OXIDE REAGENT REACTION

MgO Dilute acids Dissolves
H2O Interacts with hydroxide formation
Al2O3 Acids and alkalis Does not interact
SiO2 HF Interacts slightly
Molten alkalis Dissolves
CaO H2O Interacts with CaCO3 formation
CO2 Dissolves
SrO2 Acids Does not interact
TiO2 H2SO4 (conc.) Interacts slowly during heating
V2O5 H2SO4 (conc.) Interacts during heating
Alkalis Dissolves
Cr2O3 Acids and alkalis Dissolves slightly
Mn2O3 H2SO4 Interacts
HCl Interacts
Fe2O3 H2SO4 (conc.) Interacts
HCl Interacts
Co3O4 Acids Dissolves slowly
NiO H2SO4 (conc.) Dissolves
Cu2O Ammonia Dissolves
HCl Dissolves
Dilute H2SO4, HNO3 Dissolves
ZnO Dilute acetic acid Dissolves
Dilute mineral acids (HCl, H3PO4 ) Dissolves
Ammonia, ammonium carbonate Interacts
Fixed alkalis Dissolves
Ga2O3 Alkalis Dissolves
Acids Dissolves
ZrO2 H2SO4 (conc.) Dissolves during heating
Nb2O5 HCl Dissolves
H2SO4 Dissolves
In2O3 Acids Dissolves
SnO2 Acids and alkalis Does not interact
Sb2O3 Alkalis Dissolves
HF Dissolves
CeO2 Acids Dissolves
HfO2 HF Dissolves
WO3 Acids and alkalis Dissolves with difficulty
ReO2 HNO3 Interacts
HCl Interacts
La2O3 Mineral acids Dissolves
Bi2O3 Acids Dissolves
4.3. LONG-TERM STABILITY

Well-designed chemical sensors should provide long-term use regardless of the operating conditions.
In general, it is required that, for example, any gas-sensing device should exhibit stable and reproduc-
ible signal for a period of at least 2–3 years (17,000–26,000 h). Therefore, high temporal stability of
bulk and surface properties of sensing materials, even in corrosive media, is very important for sensor
applications. This is not possible for standard semiconductors, such as Si, InP, GaAs, and GaP, because
surface oxidation takes place in oxygen atmosphere, which inevitably leads to changes in the electronic,
adsorption, and catalytic properties of the surface. Gas sensors based on ionic compounds, such as
CuBr, have unstable parameters as well (Bendahan et al. 2002).
Organic materials, polymers, and especially biological systems also do not have the necessary tem-
poral parameter stability. Receptors based on biological recognition elements have excellent selectivity
for some reagents, but the temporal and thermal stability of these materials is very poor. For example,
according to some estimates, electrochemical sensors, using the organic polymer Nafion may retain the
capability for work for up to 1 year. However, to achieve this result, the Nafion must be wetted using a
wicking system to a reservoir (Pasierb et al. 2004). Polymer sensors used for environmental control have
another big problem: their sensitivity to ultraviolet (UV) radiation and the presence of oxidizing gases.
It has been reported that ozone and other oxidizing components (NOx ) of the polluted atmosphere
of industrial centers may be initiators or accelerators of the photochemical destruction of polymers
(Razumovskii and Zaikov 1982; Heeg et al. 2001). Because of either polymerization or destruction,
their properties change irreversibly within a fairly short period of time. As a result, gas sensors based
on such materials have short life spans, especially in normal atmosphere containing water and active
gases. Among other polymers, undoped PPY is fairly stable toward UV irradiation (Fang et al. 2002).
However, the stability of PPY against UV irradiation depends on the type of dopant present in the
polymer and the power density of the UV irradiation (Rabek 1995). Moreover, UV irradiation may
change the thickness and surface roughness of even PPY films (Fang et al. 2002). As a result, in spite
of the wide range of gas sensor prototypes designed on polymer films, very few have found their way
to the market. Even those that show excellent analytical qualities are often not suitable for industrial
fabrication, because of low technological effectiveness of the fabrication process as well as insufficient
reliability and stability.
All recognize that to realize the advantages of polymers that have a rare combination of electri-
cal, electrochemical, and physical properties, it is very important to increase their processability and
their environmental and thermal stability (Sandler and Karo 1974; Kumar and Sharma 1998). Intensive
research is being carried out in this area. For example, when a segment of an aliphatic polymer’s main
chain is replaced by a ring segment, the melting temperature, and hence the thermal stability, increases
due to the decrease in flexibility of the polymer chain (Razumovskii and Zaikov 1982; Bhuiyan 1984).
However, this task with reference to gas sensor design is rather complicated, because it is necessary to
attain high stability while maintaining the polymer’s activity. This again confirms that stability and reli-
ability are determinants for practical use of any sensing material.
The same situation is observed for silicon-based chemical sensors, such as field-effect transis-
tors (FETs) or humidity sensors. No Si-based humidity sensors are currently produced commercially
because of the technological and fundamental problems of reproducibility and stability. Crystalline
porous Si creates a problem because PSi is very reactive and is easily oxidized. Chemical sensors based
on Si generally need an aging treatment for them to have reliable and repeatable sensitivity. Even then,
lifetimes of chemical sensors based on standard semiconductors (InP, GaAs, GaP), and especially on
porous Si, can be short (Han et al. 2001).
Only metal oxides and wide-band semiconductors such as SiC and GaN with a dielectric cover-
ing have the necessary stability of surface and bulk properties in both oxygen atmosphere and water
environment to make them of wide practical use in real devices for long-term use. Thus the long-term
stability of signal of chemical sensors is one of the most important factors determining the practical
use of such devices.
Zirconia-based solid electrolytes have high stability as well: Zirconia-based solid electrolytes retain
their electroconductivity even at Т > 1000°С. However, as established by Badwal et al. (Badwal 1992;
Badwal et al. 2000), to achieve such results it is necessary to use special additives, such as Sc2O3 or Y2O3
(6–10 mol%) to stabilize the parameters of zirconia-based ceramics. Note that, with respect to other
solid electrolytes, stabilized zirconia ceramics have a minimal electronic contribution to total conductiv-
ity in the oxygen partial pressure p(O2) range from approximately 100–200 atm down to 10−25–10−20
atm. This is important for practical applications.
Potentiometric sensors based on carbonates also have yet not achieved the required durability.
However, recent research has indicated progress in this direction. Stable behavior of potentiometric
sensors was observed for a eutectic phase of Li:K:Na carbonates used as an auxiliary electrode material
(Seo et al. 2000). Another example of progress in improving sensor stability and performance was exact
definition of reference electrode influence on sensor parameters. The use of two-phase systems of
Na2Ti6O13–TiO2, Na2Ti3O7–Na2Ti6O13, and Na2SnO3–SnO2 (Maier et al. 1994; Holzinger et al. 1996) or
LiCoO2–Co3O4 (Zhang et al. 1997) allowed experimenters to obtain sensors with stable signal, in which
no shift of signal value was observed over several weeks. This research shows that all elements in chemi-
cal sensors play their parts in achieving acceptable operating parameters for practical applications.
Pasierb et al. (2004) have shown that solid-state potentiometric CO2 sensors of the type M2CO3–
BaCO3 |Nasicon|-X (where M = Li or Na, and X is a Na2O activity buffer) with a Na2Ti6O13/Na2Ti3O7
reference electrode had stable signal over 1000 h at Toper = 748 K. This is a very good result for poten-
tiometric sensors working at high operating temperatures.
These results are in accordance with the conclusions of Meixner and Lampe (1996) that the main
reasons for long-term instability of solid-state gas sensors are the following:
• A change in metal oxide parameters caused by (1) a change in the crystallite size, a consequence
of insufficient preaging by tempering; (2) irreversible reactions with the gas phase—i.e., reduc-
tion during interaction, or reactions with active gases such as SO2, Cl2, etc., that create new
phases; and (3) reactions with the substrate
• A change of metalization (sensor heating) elements, contacts for metal oxides
• Instability of the wire contacts
• Interaction with an insutable sensor casing
An additional source of temporal drift might be ionic drift, which can modify electrophysical and sur-
face properties of metal oxides (Madou and Morrison 1987).
This strong influence of both contacts and substrate shows that the problem of stability of
chemical sensors is not an easy one to resolve, and all the affecting factors need to be considered.
Attempts to improve stability should be targeted to removing all possible causes of instability. High-
temperature devices such as solid-state gas sensors do not have secondary elements (components).
These devices operate in extreme conditions, and degradation of any component may be responsible
for long-term instability.
To see how important this is, consider the results of Esch et al. (2000). In analyzing the behavior
of Ti–Pt layers destined for the heater elements of chemical sensors, Esch et al. (2000) established that
during use the Pt surface lost its metallic luster, and at a temperature of 650°C, hillocks appeared on the
heater surface. It was found that this thermal treatment induces diffusion and oxidation of Ti, especially
for annealing times up to 2 h at 450°C. Longer heat treatments did not further affect the chemical com-
position. Annealing at 650°C gave rise to hillock formation and strong adhesion problems.
Chemical sensors should work in both water solutions and in atmospheres containing water vapors.
As we know, the relative humidity of the surrounding atmosphere can reach 100%. One can judge
the importance of water vapor influence on sensor parameters by analyzing the results of research by
Skouras et al. (1999). They established that adsorption of water is a dominant factor in the formation
of surface characteristics, both with respect to adsorption of other species and to surface catalysis. For
example, hydroxylation of a SnO2 surface was found to inhibit sorption for all gas mixtures (CH4, CO,
CO2, O2 ) examined.
A hydroxylated surface is formed on an oxide by the chemisorption of a monolayer of water. Water
may also catalyze reactions taking place on the surface of a chemical sensor. The adsorption of water
also has an effect on the electronic properties of semiconducting metal oxides, usually acting as a do-
nor. Morrison (1982) has shown that hydroxylation is an intermediate stage in the interaction of water
with the metal oxide. It is intermediate between hydration of the surface and physical adsorption of
water. Long exposure can lead to hydration of the surface layer, and correspondingly to drift of chemi-
cal sensor characteristics. Therefore, a low tendency to hydration is an important requirement for a
material intended for practical use. Only this property can provide stable operation in wet atmospheres.
For example, early research in the field of humidity sensors showed that ceramic humidity sensors had
progressive drift in resistance, which was caused by the gradual formation of stable chemisorbed OH
groups on the oxide surface after prolonged exposure to humid environments (Kohl 1990). Given the
ionic-type humidity-sensing mechanism, proton hopping was adversely affected by the surface presence
of hydroxyl ions instead of water molecules, thereby resulting in a decrease in surface conductivity
(Traversa et al. 1996). As a result, most commercial humidity sensors based on ceramic sensing elements
were equipped with a heater for regeneration before each operation (Nitta 1980, 1981). Unfortunately,
this makes it necessary to expend energy for the recovery of sensitivity of the porous ceramics, and
during the cleaning operation the sensor is unable to give information about humidity.
Eventually, however, this problem was solved by using materials with different humidity-sensing
mechanisms (Yanagida 1990). For example, Arshak et al. (2002), because of a happy choice of com-
ponents (SiO2/In2O3 = 75%/25%) and, probably, of parameters of thermal stabilization, succeeded
in obtaining very good use parameters and high temporal stability with metal oxide conductometric
humidity sensors operating at room temperature. A humidity sensitivity of 0.25%/RH% was achieved.
The samples exhibited low drift over a 1-year time span (0.0013 RH%/year), low hysteresis (0.34 RH%),
good linearity (±2 RH%), and a reasonably fast time response (18 s). This stability was achieved without
using any additional thermal treatments.
Experiments carried out by Matsuguchi et al. (1998) showed that proper choice of polymer al-
lows considerable improvement of temporal stability of capacitance-type humidity sensors operating at
room temperatures as well. Long-term stability tests of cross-linked polyamide (C-PI) and cross-linked
fluorinated polyamide (C-FPI) sensor exposed to hot and wet atmospheres showed that C-FPI sensors
are quite stable after exposure to hot and wet atmospheres for long times. During these tests, sensor
elements were placed in a vessel in which the atmosphere was maintained at 40°C and 90% RH. The
humidity dependence of capacitance was measured over a humidity cycle of 10–90% RH at 25°C. It
was established that the drift of capacitance is very small.
The reason for the drift of sensor output in hot and wet atmospheres is not clear at the present
time. Hydrolysis of the film is one possible explanation. Another possibility is a change in the amount
and the state of sorbed water with time, caused by morphology (distribution of microvoids) changes
in the film. It is well known that a fluorinated polymer is water-repellent (Matsuguchi et al. 1998). The
introduced fluorine atom also reduces the cohesive forces between the polymer segments. These char-
acteristics of the fluorinated polymer reflect the stability of the sensor in hot and wet atmospheres.
Because the melting points of cross-linked polymers are high and they are heat-resistant, cross-linked
polymer sensors can have good stability even under high-temperature conditions.
All the work discussed in this section testifies that the problem of temporal stability of chemical
sensors parameters, in spite of its complexity, is a temporary problem, which will be resolved in time.
5. ELECTROPHYSICAL PROPERTIES OF SENSING MATERIALS
5.1. OXYGEN DIFFUSION IN METAL OXIDES
In the signal-determining elementary interaction processes in oxide-based chemical sensors, one may
distinguish between thermodynamically controlled chemisorptions and kinetically controlled catalytic
reactions of the molecules to be detected, as well as between thermodynamically controlled bulk point
defect equilibria (Gopel et al. 1989; Mosely 1997). Semiconductor oxides are in general nonstoichio-
metric, and the oxygen vacancies are the main bulk point defects. This means that changes in oxygen
partial pressure at operating temperatures may cause changes in bulk conductance of metal oxides. For
example, the oxygen vacancies can diffuse from the interior of the grains to the surface and vice versa,
and the bulk of the oxide has to reach an equilibrium state with ambient oxygen. So, the coefficient of
oxygen diffusion, which controls the equilibration time between the concentration of bulk point defects
in metal oxides and the surrounding gas, is an important parameter, just like the other physical-chemical
properties we have discussed.
Considering this, one can conclude that, depending on the type of solid-state gas sensor, materials
may be needed that have extreme properties, e.g., very high coefficients of bulk diffusion of oxygen
and point defects, or very low ones. The first type of material is necessary for chemical and gas sensors
whose function depends on a change of the bulk properties of materials. In such sensors the change
in bulk conductivity is a reflection of the equilibration between the oxygen activity in the oxide and the
oxygen content (oxygen partial pressure, PO2 ) of the surrounding atmosphere. Usually, in explaining
their behavior, the following equation is used (Sberveglieri 1995; Wang et al. 1998):
⎛ E ⎞ ±1/n
G = GO exp ⎜ − a ⎟ PO2 (2.6)
⎝ kT ⎠
where Go is a constant and Ea is the activation energy for conductivity. The value and sign of 1/n are
determined by the type of dominant bulk point defect involved in the equilibration process. Positive
and negative signs of 1/n correspond to p-type and n-type conduction, respectively. The sensitivity of a
semiconducting gas sensor is determined by the value of 1/n—the higher the value of 1/n, the greater
the sensitivity of the sensor. A high diffusion coefficient in such devices decreases both operating tem-
perature and response time.
The main application of this kind of solid-state gas sensor is in the measurement of oxygen partial
pressure as required in combustion-control systems, particularly the feedback control of the air/fuel
ratio of automobile engine exhaust gases near the λ point, in order to improve fuel economy and reduce
the harmful emission of gases such as CO, NOx , and hydrocarbons (Gopel et al. 1989; Moseley 1997).
Normally, electrochemical cells based on solid-state electrolytes such as ZrO2 are used as λ sensors.
At lower temperatures, the change in ambient gas concentration does not necessarily lead to equili-
bration of bulk properties of metal oxides and the surrounding gas. The surrounding gas affects elec-
trical properties through surface reactions. Therefore, for chemosorption-type sensors, the diffusion
of oxygen in the bulk of crystallites is a source of temporal drift of parameters (Jamnik et al. 2002).
This means that such sensors should utilize materials in which the coefficient of oxygen diffusion is
minimized.
However, according to Fleischer and Meixner (1997), for any metal oxide sensor, there are three
temperature regions for operation under specific conditions. At high temperatures, the kinetics of the
rate processes are fast enough that equilibrium is rapidly established between the partial pressure of
oxygen in the surrounding gas and the composition of the oxide. At intermediate temperatures, we
can observe reaction of the gas with the metal oxide lattice, but because of the small coefficient of
bulk oxygen diffusion, the chemical composition of the material does not reach equilibrium during the
time of gas detection. This is the so-called “redox” (reduction/reoxidation) mechanism. At still lower
temperatures, chemisorption (adsorption/desorption) processes can dominate in surface reactions. For
these modes of operation, there are no fixed temperature borders (Figure 2.14)—they are fairly diffuse
and can be shifted by exchanging one metal oxide for another. One can only say that the first border is
in the temperature range 200–500°С, while the second border occurse at temperatures of 400–700°С.
It is impossible to say which mode of operation is preferable for practical use. Every mode has
advantages and disadvantages. Some of them, relating to solid-state gas sensors, are given in Table 2.11.
SnO2- and In2O3-based sensors, operating in the temperature range 200–450°С, are an example of low-
temperature sensors, while TiO2-, WO3-, SrTiO3-, and Ga2O3-based sensors, functioning at tempera-
tures higher than 450–500°С, are high-temperature sensors.
The sensing mechanisms of low-temperature gas sensors based on the widely used SnO2 oxide are
in general well investigated (Fleischer and Meixner 1997, 1998; Barsan et al. 1999; Brinzari et al. 2000).
This sensor material is characterized by a grain boundary controlled sensing mechanism and a tendency
200-500 oC 400-700 oC
Operating temperature,
Figure 2.14. Processes controlling conductivity response of metal-oxide gas sensors. (Reprinted
with permission from Korotcenkov 2007. Copyright 2007 Elsevier.)
to suffer from surface contamination due to a low operating temperature range, 100–400°C. This tem-
perature is not high enough to completely burn out organic deposits or desorb certain adsorbates. This
problem limits long-term stability of the electrical output signal. However, low-temperature chemo-
sorption sensors could be easily adapted into modern microelectronics, having sufficient limitations in
temperature of use. Using these, it is easier to maintain good selectivity and to create sensor arrays for
design of an “electronic nose” (Gardner and Bartlett 1992, 1999).
On the other hand, high-temperature sensors are better for conducting in-situ control of many high-
temperature technological processes, including control of combustion engines (Fleisher and Meixner
1997, 1998). The sensing behavior of such devices is mostly explainable and predictable and is based
Table 2.11. Advantages and disadvantages of sensors operating in different modes
OPERATING TEMPERATURE RANGE ADVANTAGES DISADVANTAGES

Low operating temperatures Low dissipated power of sensor Strong dependence on relative air
(Toper < 400°C) Low threshold of sensitivity humidity
Long lifetime Long response and recovery times
Wide choice of sensitive materials Necessity of prolonged aging
Good compatibility with before start of exploitation
micromachining technology
High operating temperatures Weak dependence on air humidity High dissipated power
(Toper > 500°C) signal Lower reliability
Good signal reproducibility Lower sensitivity
Short response time Strict requirements for sensing
Fast recovery of initial state material and sensor construction
Poor compatibility with standard
silicon technology
Source: Adapted with permission from Korotcenkov 2007. Copyright 2007 Elsevier.
on well-established thermodynamic principles (Brynn and Tseung 1979). In addition, during a study of
semiconducting metal oxides operated at high temperatures (400–900°C), it was established that for
some materials, there is a grain boundary–independent conduction mechanism and self-cleaning ef-
fects on the sensor surfaces, thus indicating progress toward stability and reproducibility (Fleisher and
Meixner 1997).
As we noted earlier, virtually all oxides can function as high-temperature gas sensors (Wang et al.
1998). In practical applications, however, the usefulness of a metal oxide for this temperature range is
determined by parameters such as material stability, response time, Eg , type of conductivity, etc. Many
semiconducting oxides have been investigated; BaTiO3, SrTiO3, Ga2O3, WO3, Nb2O3, CoO, MoO3,
CeO2, and BaSnO3 are some examples. All these oxides are stable enough (see Table 2.8) and can pro-
vide sensors that operate with more or less effectiveness right up to 900°С.
For example, for Ga2O3 sensors, the optimal operating temperature is Toper ≈ 600–800°C, whereas
for oxygen sensors based on SrTiO3, operating temperature is 1000°C. It is necessary to note that
considerable attention to complex metal oxides such as SrTi0.65Fe0.35O3−δ is determined by unusual
temperature-independent conductivity of these materials above 700°C and PO2 > 1 Pa (Litzelman et al.
2005; Rothschild et al. 2005). It was established that at an intermediate composition of X = 0.35, the
band-gap energy is such that the Fermi energy lies just far enough above the valence band to compen-
sate for the temperature dependence of mobility, yielding a zero temperature coefficient of resistance
(TCR) from the product of the free carrier (hole) concentration and mobility terms (Rothschild et al.
2005). Strong sensitivity to oxygen partial pressure variation and negligible cross-sensitivity to tempera-
ture fluctuations make these metal oxides promising candidates for oxygen sensors in lean-burn engines
(Litzelman et al. 2005; Rothschild et al. 2005).
If one considers that the presence of structural vacancies in the lattice promotes an increase in
the coefficient of oxygen bulk diffusion, it is possible to assume that for design of sensors in which
the appearance of diffusion processes worsens exploitation parameters, materials which do not contain
structural vacancies are preferable. At the same time, for design of sensors in which bulk diffusion con-
trols the sensor’s parameters, materials with native structural vacancies are preferable.
The perovskite materials, investigated intensively in recent years for high-temperature sensors, have
these very structural properties. The structures of the perovskite phases can be viewed as distortions
arising from ordering of oxygen vacancies in the cubic lattice of SrFeO3. Post et al. (1999) established
that as the content of oxygen drops, the phase changes from cubic perovskite (x ≈ 0.5) to tetragonal
(x ≈ 0.35), to orthorhombic (x ≈ 0.25), to the brownmillerite structure (x ≈ 0). These phases can be
represented as a repeating sequence of octahedral and tetrahedral layers. The SrFeO2.5+x compounds
are normally prepared and operated as sensors at high temperatures, where the oxygen-ion mobility is
reasonably high. The ability to prepare these phases at room temperature was interpreted in terms of
high mobility of oxygen anions along extended defects in the material as it formed (Post et al. 1999).
The mobility along defects is several orders of magnitude greater than in an ordered crystal. It was
postulated that the defects responsible for the increased mobility are the stacking faults in the repeating
sequence of octahedral and tetrahedral layers. Rapid mobility of the oxygen anions or vacancies along
the stacking faults would be followed by diffusion over very small distances within the ordered domains,
on the order of some nanometers. Thus, the overall process is accelerated as a result of high diffusion
coefficients over long distances and small diffusion coefficients over short distances.
Another interesting conclusion regarding requirements for materials based on dual-phase compos-
ites was made by Chena et al. (1996). On the basis of a mixed oxides study, it was established that for
large oxygen permeability, the following requirements must be met: (1) percolation of the electronic
conducting metal phase in the oxide matrix, allowing electrons to pass through; (2) large ionic conduc-
tivity of the oxide phase, allowing oxygen ions to migrate through; and (3) high catalytic activity of
the metal phase toward surface oxygen exchange. Bismuth oxide–silver composite fulfils these require-
ments, showing the best oxygen permeability. For example, for a 1.60-mm-thick bismuth oxide–silver
composite, an oxygen flux of 1.19 10−7 mol cm−2s−1 was observed at 800°C under conditions of PO2(h)
= 0.21 atm and PO2(l ) = 0.026 atm. For comparison, zirconia-based composites had oxygen perme-
ability at the same conditions almost two orders less. Such dual-phase composites are promising for
applications in oxygen separation, oxygen-enriched combustion, and catalytic membrane reactors.
It is necessary to note that considerable uncertainty exists in the choice of material for low-tem-
perature sensors, for a temperature range lower than 450°С. The gas response of such sensors can be
controlled by either chemisorption processes or “redox” processes (Korotcenkov et al. 2004). Sensors
based on tin dioxide, the most studied example, are sensors of the first type, whereas In2O3-based
sensors, studied intensively earlier, are sensors of the second type. However, the fact that SnO2-based
sensors have been better studied and are more widely used is not the basis for concluding that the
chemisorption mechanism of sensitivity has advantages over the “redox” mechanism for sensor design.
SnO2-based sensors have better sensitivity to reducing gases, and better stability during operation in re-
ducing atmospheres. However, In2O3-based sensors have better sensitivity to oxidizing gases, and show
less dependence of parameters on changes in air humidity (Korotcenkov et al. 2004a, 2004b, 2004c).
Which is better, high response to reducing gases, or to oxidizing ones? The answer to this question
depends on the user’s requirements.
One can make a similar comparison for other pairs of metal oxides, for example, SnO2–CTO, and
SnO2–WO3 (see Table 2.12).
All of the above indicates that the choice of materials and operation modes is determined by such
factors as the type of gas sensor (see Tables 2.9–2.11), the objective (apparatus, device) for which the
sensor is being designed, and the construction (structure) chosen for the sensor’s fabrication. However,
any competition between sensing materials can be ignored if the devices are intended for incorporation
into an “electronic nose.” Different behavior during interaction with the same gas is one of the most
important requirements for sensors designed for this application (Gardner and Bartlett 1992, 1999).
5.2. CONDUCTIVITY TYPE
Metal oxides can have either n- or p-type conductivity, a difference which is very important in terms of
their application. For materials with p-type conductivity, the conductivity increases with an increase in
oxygen pressure; whereas for n-type oxides, the conductivity decreases with an increase in oxygen pres-
sure (see Table 2.13).
If one considers oxides in their main valence states, there is a regular change of conductivity type
depending on the metal’s position in the Periodic Table. The stable metal oxides are shown in Figure
2.15. There is some information in the literature about their type of conductivity. Oxides of transition
Table 2.12. Advantages and disadvantages of gas-sensing materials
MATERIAL ADVANTAGES DISADVANTAGES

SnO2 High sensitivity Low selectivity
Good stability in reducing atmosphere Dependent on air humidity
WO3 Good sensitivity to oxidizing gases Low sensitivity to reducing gases
Good thermal stability Dependent on air humidity
Slow recovery process
Ga2O3 High stability Low selectivity
Possibility to operate at high temperatures Average sensitivity
In2O3 High sensitivity to oxidizing gases
Fast response and recovery
Low sensitivity to air humidity Low stability at low oxygen partial
pressure
CTO (CrTiO) High stability Average sensitivity
Low sensitivity to air humidity
Source: Reprinted with permission from Korotcenkov 2007. Copyright 2007 Elsevier.
elements at the beginning of big periods have n-type conductivity, while the ones at the end usually have
p-type conductivity.
According to Krilov and Kiselev (1981), the metal oxides of highest valence states as a rule are dis-
posed to partial reduction, which leads to an appearance of stoichiometric excess of metal in the lattice
of an n-type semiconductor. Oxides of lowest degrees of oxidation are disposed to partial oxidation,
and both stoichiometric excess of oxygen and p-type conductivity is a consequence of this fact. For
example, both CrO2 and CrO3 are oxides with n-type conductivity, while CrO is a p-type semiconductor.
Cr2O3 may have either n-type or p-type conductivity, depending on oxygen pressure.
If one considers the main gas-sensing materials according to their conductivity type, it turns out
that all the most effective chemisorptions-type sensors are based on materials with n-type conductiv-
ity, such as SnO2, TiO2, WO3, ZnO, and In2O3, providing the opportunity for oxygen chemisorption.
Previous research has shown that, in general, all n-type metal oxides are thermally stable and have low
oxygen availability in comparison with p-type metal oxides (Gordon et al. 1996). It is known that inter-
Table 2.13. Expected resistance responses for reducing and oxidizing gases on
n- and p-type semiconducting oxides
DECREASED REDUCING OXIDIZING

MATERIAL PO2 GASES GASES
n-Type − − +
p-Type + + −
+, Resistance rises; −, resistance falls.

Figure 2.15. Dependence of the metal oxide’s type of conductivity on the position of metals in the
Periodic Table. Hatched, metal oxides with a predominance of p-type conductivity. Unhatched, metal
oxides with a predominance n-type conductivity.
action with reducing gas decreases the resistance of n-type metal oxides. This is the preferred direction
for detection of reducing gases, leading to simpler compatibility with peripheral measuring devices,
and probably better reproducibility. In addition, Madou and Morrison (1987) showed that many p-type
oxides are relatively unstable because of their tendency to exchange lattice oxygen easily with air.
However, this does not mean that p-type materials are not applicable for sensor design. For exam-
ple, the metal oxide Cr2-xTixO3 (x < 0.4) (CTO), a prospective for gas sensors, is a p-type material
(Williams and Pratt 1997, 1998). According to Williams and Pratt (1997, 1998), CTO is in many ways
an excellent material. It shows minimal effects from humidity, unlike SnO2, for which humidity can
dominate overall response. This may be attributed to metal oxides having separate surface sites for
catalytic combustion, electrical and water vapor response, as discussed by Williams and Pratt (1998) for
SnO2. In addition, labile bridging oxygens, characteristic of the rutile structure of SnO2, are absent in
the corundum crystal structures of Cr2O3 and CTO, allowing a longer time constant for the surface
modification of CTO. This, in turn, may explain the superior resilience of CTO to surface poisoning
(Williams and Pratt 1997).
Stable perovskites, for example, ReCoO3 (Re = La, Nd, etc.), promising candidates for chemical
sensor application, are also p-type semiconductors. It was demonstrated on the example of LaCoO3
that these materials have high activity in the oxidation of CO, reduction of NO (Voorhoeve 1972), and
reduction of SO2 in the presence of CO (Brynn and Tseung 1979). Concerning the catalytic mecha-
nism, it has been clarified that a fundamental role is played by the oxygen present in the structure (bulk)
and by the one adsorbed on the surface. The rate of the catalytic reactions may depend on the surface
reaction and/or the oxygen bulk diffusion. In particular, at high temperatures, when the oxygen is fairly
mobile, the contribution of bulk oxygen becomes more relevant. Moreover, the cations exposed to the
surface have some dangling bonds and these, together with their incomplete d shells, favors chemisorp-
tion and electron transfer between the solid phase and the interacting molecules (Kosima et al. 1983;
Chan et al. 1994).

Figure 2.16. Typical view of the influence of oxygen partial pressure on the conductivity of per-
ovskite-type metal oxides with p-type conductivity. (Adapted with permission from Moseley 1997.
Copyright 1997 IOP.)
The fact that perovskites can have p-type conductivity gives them additional advantages for applica-
tions in high-temperature oxygen sensors. It was found that the temperature dependence of conduction
at high temperatures is considerably less in the p-type range than in the n-type one; i.e., in the p-type
range, the isothermal data lines are very much closer together than in the n-type range (see Figure 2.16).
Parameters of the dependence of conductivity on both oxygen partial pressure and temperature for
some metal oxides are shown in Table 2.14 (data from Lee and Park 1990; Sberveglieri 1995; Moseley
1997; Menesklou et al. 2000).
Moseley (1997) gives the following explanation: In general, as the temperature is raised at a fixed
oxygen partial pressure, the oxides of transition metals lose oxygen; if they are in the p-type range of
oxygen partial pressure, then conductivity will tend to decrease. This effect is opposite to the thermal
Table 2.14. Dependence of high-temperature conductance on oxygen partial

pressure and temperature for some metal oxides
OXYGEN SENSITIVITY ACTIVATION ENERGY

MATERIAL (n) (eV) CRYSTAL STRUCTURE
TiO2 −4, −6 1.5 Rutile
Ga2O3 −4 1.2–1.9 Corundum
SnO2 −4, −6 1.3–2.0 Casseterite
CeO2 −4, −6 Fluorite
SrTiO3 −4, −6 1.24 Perovskite
+4 ~0
BaFe0.8Ta0.2O3 +5 ~0 Perovskite
promotion of charge carriers, so the possibility arises of doping the oxide so that the two effects cancel
and the effect of a temperature change is to leave the conductivity little altered. Two points need to
be considered here. First, if such a balance of thermal effects is achieved, it need not alter the effect
of changes in oxygen partial pressure at a fixed temperature. Second, such a balance is only possible
for oxides in the p-type regime, where the thermal activation of charge carriers and the stoichiometry
effect are opposed. The data in Figure 2.16 show that for the materials tested, the thermal activation
and stoichiometric effects are not quite in balance, so the activation energy is small but not zero. A
family of doped perovskite structure ferrates has been found in which balance is achieved, and the
activation energy is almost zero over a considerable temperature range (Moseley 1997). Materials such
as BaFe0.8Ta0.2O3 (see Table 2.14) retain substantial oxygen sensitivity and thus should be useful in the
construction of oxygen sensors without the need for additional temperature control or compensation
elements. Sr(Ti0.65Fe0.35)O3 has the same properties (Menesklou et al. 2000).
CuO and ferrum oxides also have p-type conductivity. These oxides are effective additives to both tin
and indium oxides for the formation of nano-composite-based sensors with extremely high conductivity
response to H2S and series of other specific gases (Ivanovskaya et al. 2003; Kotsikau et al. 2004). In addi-
tion, materials with p-type conductivity are being used successfully in adsorption electrochemical sensors.
Metal oxides of p-type conductivity also show specific catalytic properties. Metal oxides of p-type
are generally active oxidation catalysts. For example, p-type semiconducting bulk oxides (Co3O4, MnO2,
MoO3, and V2O5 ) have been found to promote catalytic oxidation of carbonaceous residues through
the thermal-redox cycle by releasing lattice oxygen to initiate combustion under pyrolytic conditions in
the 450–600°C temperature range (Gordon et al. 1996). It turns out that ozone dissociation takes place
with greater effectiveness on the surface of p-type oxides (Dhandapani and Oyama 1997). Metal oxides
of n-type are generally not active oxidation catalysts. For example, Dhandapani and Oyama (1997)
found that catalytic activity of unsupported oxides based on equal volume gave the following order of
reactivity in terms of ozone decomposition:
Ag2O > NiO > Fe2O3 > Co3O4 > CeO2 > Mn2O3 > CuO
> Pb2O3 > Bi2O3 > SnO2 > MoO3 > V2O5 > SiO2 (2.7)
Unfortunately in this progression, surface areas have not been considered. Later research yielded
a different progression:
MnO2 > Co3O4 > NiO > Fe2O3 > Ag2O > Cr2O3 > CeO2 > MgO > V2O5 > CuO > MoO3 (2.8)
According to Gordon et al. (1996), n-type bulk oxides do not exhibit redox activity to any great
extent under these conditions of temperature and reaction environment, because surface oxygen mobil-
ity is too low.
It is possible that distinction of catalytic properties of p-type conductivity oxides is a consequence
of a particular surface’s materials properties. Kohl (1990) notes that oxygen cannot be chemisorbed
on undoped stoichiometric n-type oxides. Oxygen ions can only be adsorbed if their negative charge is
compensated by ionized bulk donors in a space-charge layer. In addition, for thermodynamic reasons,
only a small fraction of the oxygen monolayer can be chemisorbed on the surface of n-type oxides
(the Veisz limitation). In contrast, on p-type semiconductors, a full monolayer of oxygen ions typically
occurs, because the metal ions of the lattice can be oxidized into a higher oxidation state. However, it is
necessary to note that, as a rule, with a rise in temperature, in such oxides the probability of transition
of adsorbеd oxygen into lattice oxygen due to its incorporation into the metal oxide lattice increases
sharply. As a result, one can observe the following distinction in the behavior of n- and p-type metal
oxides (Sprivey 1987): If n-type metal oxides have lost oxygen upon heating in air, p-type metal oxides
gain oxygen during such thermal treatments. Numerous researchers have identified that the multiple
stable oxidation states and high concentration of positive “holes” in p-oxides stimulate surface oxygen
mobility to a greater extent than n-type oxides. Each “hole” provides a vacancy where the free electrons
of mobile surface oxygen species can be stabilized (Gordon et al. 1996).
5.3. BAND GAP
The electronic structure of metal oxides is much more complex than that of most metals and stan-
dard semiconductors. For example, the bulk electronic structure of the 3d transition-metal oxides lies
somewhere between itinerant and localized, and neither theoretical approach–the former of which
has been developed for metals and semiconductors and the later that is used to describe molecules–is
entirely appropriate. The great number of atoms per unit cell in many structures, coupled with the re-
duced symmetry at the surface and the necessity to calculate many unit cells to separate surface and bulk
effects, means that first-principles electronic structure calculations still require a great deal of computer
time (Maki-Jaskar and Rantala 2001, 2002). Therefore, in this section we will limit our consideration to
the band gap, which may be determined experimentally.
Fairly large band gap (Eg) and small activation energy of the centers, responsible for metal oxide
conductivity, is an optimal combination of parameters for materials designed for semiconductor solid-
state chemical sensors. This combination of activation energies is necessary in order to avoid sensor
operation in the region of self-conductance. In this case the influence of the surrounding temperature
on the sensor’s parameters is reduced. At that, as a rule, the higher the operating temperature, the larger
should Еg be (see Figure 2.17). The curve in Figure 2.17 was built on the basis of results presented
by Bogue (2002a, 2000b). As follows from experimental results, for solid-state gas sensors operating
at temperatures exceeding Т > 300°С, the optimal band gap must be larger than 2.5 eV. For sensors
working at room temperature, Еg can be considerably smaller. Moreover, for sensors that function at
room temperature, a small Eg may be an advantage. Doll and Eisele (1996, 1998) have shown that the
average work function shifts in an atmosphere of dry oxidizing gases (Cl2, NO2, SO2 ), increasing when
the energy band gap of metal oxides decreased.
Note that the ability to operate at higher temperatures is an important advantage of solid-state
chemical sensors, because this allows one to reduce considerably the influence of air humidity on gas-
sensing characteristics. It has been established that, as a rule, the lower the operating temperature, the
greater is the sensitivity of the sensor’s parameters to relative air humidity (Korotcenkov et al. 2003).
Analyzing the data in Table 2.15 (data from Samsonov 1973; Kindery et al. 1976; Walton 1990; Cox
1992; Gordon et al. 1996; Kumar and Sharma 1998; Sol and Tilley 2001; Colladet et al. 2004; Wallace
2004; etc.), one can see that most known metal oxides satisfy this requirement. Note that low dielectric
constant is required for design of humidity sensors.

Figure 2.17. Influence of band gap on the maximum operating temperature of semiconductor de-
vices. (Data from Bogue, 2002a, 200b.)
Many polymers also have fairly large band gaps. Table 2.15 lists only a range of possible Еg values,
because there are so many known polymers. For example, Harsanyi (1994, 2000) lists more than 10
types of polymers that are used for gas sensors, including PMODS [poly(methyl-octadecyl siloxane)],
polychloroprene, polyvinylpyrrolidone, polyepichlorohydrin, polycaprolactone, DEGA (diethylene gly-
col adipate), PEVA (18% VA) [polyethylene-co-vinyl acetate (18% vinyl acetate)], polystyrene-butadi-
ene, PEVA (45% VA), and polyethylene oxide. Taking into account the polymers’ limited temperature
range, however, they need not necessarily have wide band gaps. It is also necessary to know that the
Еg for polymers is only a conditional value, which may change within a certain range depending on the
mode of polymerization. For example, the Еg for thienylene-PPV polymers may vary within the range
of 1.4 to 2.4 еV (Colladet et al. 2004).
A large band gap is also an advantage for ion semiconductors, because in this case the contribution of
electron conductivity to that of the entire material, especially at high operating temperatures, is reduced.
5.4. ELECTROCONDUCTIVITY
Certainly a large variety of metal oxides can be used in many types of chemical sensors. However, this does
not mean that metal oxides have no limitationss in application. For example, for a chemisorptional con-
ductometric gas sensor, the sensing material should be conducting,, i.e. the concentration of point defects
in the metal oxide should be fairly high. Experimental results have indicated that the optimum lies in the
range 1017–1019 сm−3, which corresponds to an electroconductivity of the metal oxide of 10−2–101 Sm/
cm. Too high a concentration of point defects, i.e., high electroconductivity, reduces the influence of the
surface on the bulk concentration of charge carriers in the grains. In this case the effects on the surface are
screened and cannot be observed. As a result, metals are usually not used for conductometric sensors.
However, results have shown that metals can be used for these purposes if they are in the form of
superthin films with thickness less than 20–40 nm (Najafi et al. 1994; Johnson et al. 1994; Galdikas et
Table 2.15. Electrophysical parameters of sensing materials
BAND GAP TYPE OF RESISTANCE DIELECTRIC COEFFICIENT

MATERIAL (eV) CONDUCTIVITY (ohm m−1) CONSTANT OF REFRACTION
Semiconductors
Si 1.1 n, p 103–10−5 11.8 3.4–3.55
InP, GaAs 1.35–1.41 n, p 108–10−5 12
SiC 3.27 n 9.7 3.4
GaN 3.6 n
Diamond 5.4 n 1013 5.5
Polymers
Polymers 0.3–3.5 n, p 10−18–104
trans-Polyacetylene 1.4 n, p 10−12–104
Polyphenylene 3.4 n, p 10−16–103
Polypyrrole 2.7–3.0 p 10−14–103 8
Polythiophene 2.0 p 10−14–102
Metal oxides
MgO 7.7–7.8 n 1013–1014 3–3.5 1.74
CaO 6.9 n >1010 3 1.84
SrO 5.3 102 1.87
BaO 4.4 >106–108 4 1.98
Y2O3 5.5–5.6 p >109 14–15
La2O3 4.3 p >1010 20–30
TiO2 3.0–3.4 n 1011 50–86 2.00–2.70
ZrO2 4.5–5.8 n 104–08 17–25 2.10
HfO2 5.7 n, p 109–1011 25 1.98–2.02
CeO2 3.4–5.4 p >106–1010 21
V2O5 2.2–2.45 n 101–105 13–15
Nb2O5 3.3–3.9 n 11–40 2.0
Ta2O5 3.5–4.0 n >103 21–27 2.5
Cr2O3 1.8–3.3 n, p >102–106 9 2.1–2.5
MoO3 n 105–108
WO3 2.4–2.85 n >101–104 20
Mn2O3 p 101–106 2.35
Fe2O3 1.9–2.2 n, p 102–106
Co3O4 1.0–1.1 p 102–104
Rh2O3
ReO2 10−6–10−3
NiO 2.0–4.3 p 109–1011 10 2.23–2.37
PdO 1.0 p
PtO
CuO 1.2–1.4 n, p 10−1–101 10 2.63–2.84
(Continued on following page)
Table 2.15. (Continued )
BAND GAP TYPE OF RESISTANCE DIELECTRIC COEFFICIENT

MATERIAL (eV) CONDUCTIVITY (ohm m−1 ) CONSTANT OF REFRACTION
Metal oxides (Continued)

Ag2O p 10−1
Au2O3
ZnO 3.2–3.4 n 10−4–102 8–10 2.00–2.02
Al2O3 6.2–8.7 n >1012–1013 8–9 1.6–1.70
Ga2O3 4.2–4.4 n >106–108
In2O3 3.2–3.7 n 10−2–102
SiO2 8.9 >1012–1014 3.5–4.0 1.461–1.54
SnO2 3.6 n 1–104 14 1.99–2.09
Bi2O3 p 106–1010 2.42
P205
Sb2O3 1.6 p 2.087–2.18
TeO2 ~6–7 p
Si3N4 5.1 7
al. 1996, 1998). Prototypes of such gas sensors were produced using thin films of various metals such
as Pt, Au, and Ni. Reversible resistance response was obtained for the new sensors in air contaminated
with various gases. The sensitivity and selectivity of the new sensors are comparable to those of the
metal-oxide gas sensors that have already been commercialized. However, the stability of the param-
eters has to be studied thoroughly before these new sensors can be used for practical applications. The
results of first experiments carried out by Galdikas et al. (1998) indicated that the electric parameters are
stable for the metal sandwiches within fixed intervals of working temperatures. The high-temperature
limit depends on the metal used for the bottom layer of the sandwich. The upper limit of temperature
also depends on the amount of contaminating gases in some sandwiches. Overheating the sensors leads
to limited drift of the parameters, probably caused by oxidation of the metal films. Tin oxide coating
protects metal film sensors against oxidation. The coating also stabilizes the electrical parameters and
extends the limits for stable operation of the sandwiches.
For conductometric sensors, excessively low concentration of free charge carriers (n < 1016 cm−3,
i.e., σ < 10−4–10−5 Sm/cm) is also not acceptable. In nanosize structures it reduces the modulation limits
of the position of the Fermi level and leads to a sharp increase in the resistance of the sensing material.
For other chemical sensors, however,such as sorptional sensors, fiber-optic sensors, sensors utiliz-
ing the fluorescence effect, etc., where conductivity is not a controlled parameter, there is no need to
impose restrictions on the electroconductivity of the materials used. The materials may be either insula-
tors or have metal-type conductivity. For example, metals may be used successfully in devices such as
metal-insulator-semiconductor sensors and work-function sensors, which exploit the catalytic proper-
ties of metals (Eisele et al. 2001). Similarly, such insulators as Аl2О3 are good materials for humidity
sensors. Materials for high-temperature sensors (Т > 800°С) may be insulators at room temperature.
Conductivity in such materials may become apparent only at sufficiently high temperatures.
Table 2.15 provides data on the electroconductivity of various oxides used in chemical sensors
(data from Samsonov 1973; Kindery et al. 1976; Harsanyi 1994; Gellinger and Bouwmeester 1997; etc.).
Applicability of metals, semiconductors, and dielectrics for the most common types of sensors is given
in Table 2.16.
Note that for electrochemical sensors, in contrast to metal-oxide conductometric gas sensors, high
electroconductivity is an advantage. Therefore, metal oxides with metallic-type conductivity, such as
RuO2, Co3O4, PbOx, and MnO2, may be used successfully in electrochemical technology as so-called
activated electrodes (Trasatti 1987).
Figure 2.18 shows the possible range of electroconductivity change for some polymers (data from
Kumar and Sharma 1998; Skotheim et al. 1998; etc.). One can see that polymers cover a much wider
range of conductivity than other materials, demonstrating that polymers can satisfy any requirement for
electroconductivity. Polymers can exhibit conductivity across a range of some 15 orders of magnitude;
however, the conductivity may involve different mechanisms within different ranges. Conductivity of
polymers is influenced by a variety of parameters, including polaron length (the number of rings or
double bands spanned), the conjugation length (the extent of delocalization along the chain before a
defect is encountered), the overall chain length (which also affects other physical properties of the poly-
mer), and, importantly, charge transfer to adjacent molecules. This latter involves interchain hopping
within a fibril or polymer particle and finally interparticle hopping (Walton 1990).
Also, it should be noted that since conducting polymers are generally insoluble and able to inter-
act, they are not amenable to conventional methods of purification and characterization, and some
discrepancy and apparent irreproducibility may originate from changes in preparation procedures that
are sufficient to alter the exact nature of the polymer under study. In addition, the mechanism of
sensitivity—for example, the sensitivity of polymer materials to gases—differs from the mechanism
of the sensitivity of metal oxides to gases. Therefore, the limitations described above for metal oxide
sensors do not apply to polymer sensors.
5.5. OTHER IMPORTANT PARAMETERS FOR SENSING MATERIALS

Taking into account that at present a large variety of optical methods may be used in chemical sensors, in-
cluding ellipsometry, luminescence, fluorescence, phosphorescence, Raman spectroscopy, interferometry,
surface Plasmon, etc., one can conclude that for chemical sensor design based on these methods, param-
Table 2.16. Applicability of various materials for chemical sensor design
TYPE OF CHEMICAL SENSOR

FIBER RT SENSOR
CONDUCTOMETRIC HT OPTIC OPTIC (SAW, WORK ELECTROCHEMICAL
MATERIAL GAS SENSOR SENSOR SENSOR SENSOR MEMBRANE FUNCTION) CELL
Metal + − − + + + +
Semiconductor + + + + + + +
Insulator − – + + + + –

Figure 2.18. Comparison of conductivity ranges for several conducting polymer systems compared
with conventional materials.
eters such as refractive index, absorbance and fluorescence properties of analyte molecules or chemo-
optical transducing elements will have definitive importance. For example, as noted earlier, in order to
shift the operating point of surface plasmon resonance gas sensors toward an aqueous environment, a
thin, high-refractive-index dielectric overlayer can be employed (Koudelka-Hep 2000, 2001). Use of an
overlayer with high refractive index allows for a thinner overlayer and potentially better sensor sensitiv-
ity. Analysis of data presented in Table 2.17 indicates that tantalum pentoxide, which has high refractive
index and good environmental stability, may be used for this purpose.
However, metal oxides such as Co3O4, NiO, Mn3O4, CuO, and WO3 are more suitable for use in
optochemical sensors based on optical absorption change during interaction with analyte (Ando et al.
1997, 2001). The reversible absorbance change in the visible–near IR range and relatively fast response
make these oxides potential candidates for optical detection of CO, H2, and air humidity. Co3O4-based
optochemical sensors can operate at room temperature (Ando et al. 1997). Of course, these sensors
Table 2.17. Refractive indices of some sensing materials
MATERIAL REFRACTIVE INDEX

Semiconductors
Si, InP, GaAs 3.4–3.55
Metal oxides
Al2O3, SiO2 1.4–1.7
MgO, CaO, SrO 1.7–2.0
BaO, ZrO2, HfO2, Nb2O5 ~2–2.1
ZnO, SnO2, Sb2O3 2.1–2.5
Cr2O3, Fe2O3, NiO, Bi2O3, >2.5
TiO2, Ta2O5, CuO
Source: Reprinted with permission from Korotcenkov
2007. Copyright 2007 Elsevier.
do not possess as high sensitivity as standard semiconductor gas sensors. However, this type of sensor
has some unusual features which can be utilized in real applications. The advantages of optochemical
sensors over conventional electricity-based gas sensors are higher resistivity to electromagnetic noise,
compatibility with optical fibers, and the potential for multiple-gas detection using differences in the
intensity, wavelength, phase, and polarization of the output light signals (Ando et al. 2001).
Dielectric constant is another important parameter of materials intended for chemical sensor ap-
plications (see Table 2.18) and should be considered when selecting materials for capacitance-type gas
sensors (Ishihara and Matsubara 1998). Capacitive-type sensors have good prospects given that the
capacitor structure is so simple, enabling miniaturization and achieving high reliability at low cost. In
addition, application of capacitance is easily performed by oscillator circuits and thus, capacitive-type
sensors enable sensitive detection. In addition, oscillator circuits consist of only a standard resistor and
a sensor capacitor. Therefore, the signal treatment circuit is also very simple and low in cost. The hu-
midity sensor is the most well known capacitive-type sensor. Since water has an abnormally high dielec-
tric constant, the adsorption of water into porous metal oxide changes the relative permittivity of the
gas-sensing matrix. In relation to other types of humidity sensors, the capacitive type has the advantage
of high sensitivity over a wide humidity range. Porous Al2O3 is the best-known metal oxide for using in
such sensors (Ishihara and Matsubara 1998).
6. STRUCTURAL PROPERTIES OF SENSING MATERIALS
It has been established that materials in different structural states can be used for chemical sensors.
These states include the amorphous-like, glass-like, nanocrystalline, polycrystalline, and single-crystalline
Table 2.18. Dielectric constants of some sensing materials
MATERIAL DIELECTRIC CONSTANT

Metal oxides
MgO, CaO, BaO, SiO2 3–5
Cr2O3, NiO, CuO, ZnO, Al2O3 5–10
Y2O3, ZrO2, V2O5, WO3, SnO2 10–20
La2O3, HfO2, CeO2, Nb2O5, Ta2O5 20–50
TiO2 >50
Semiconductors
Si, InP, GaAs 11.8–12
SiC 9.7
Diamond 5.5
Polymers
Polypyrrole 8
Source: Reprinted with permission from Korotcenkov 2007.
Copyright 2007 Elsevier.
states. Each state has its own unique properties and characteristics that can affect sensor performance.
However, in practice, nanocrystalline and polycrystalline materials have found the greatest application
in solid-state chemical sensors (Gopel et al. 1989–1995; Sberveglieri 1992a; Ihokura and Watson 1994;
Schierbaum et al. 1992; Henrich and Cox 1994; Barsan et al. 1999; Kohl 2001; Korotcenkov 2007,
2008b). Nanocrystalline and polycrystalline materials have the optimal combination of critical proper-
ties for sensor applications, including high surface area due to small crystallite size, cheap design tech-
nology, and stability of both structural and electrophysical properties
Typically, amorphous-like and glassy materials are not stable enough for gas-sensing applications,
especially at high temperature (Korotcenkov 2007a, 2008b; Tsiulyanu et al. 2004). Single-crystalline and
epitaxial materials have maximum stability, and therefore the use of materials in these states for gas sen-
sors may improve the temporal stability of the sensor. Unlike polycrystalline materials, devices based
on epitaxial and single-crystalline materials will not be plagued with the problem of instability of grain
size. However, the high cost and technological challenges associated with their deposition limit their
general use in gas sensors.
One-dimensional structures, which are single-crystalline materials, can be synthesized using simple,
inexpensive technology (Barsan et al. 1999; Dai et al. 2002; Lu et al. 2006). Wide use of one-dimensional
structures is, however, impeded by the great difficulties encountered in their separation and manipula-
tion (Deb et al. 2007; Huang and Choi 2007). During the synthesis of one-dimensional structures, one
may observe a considerable diversity of geometric parameters. In polycrystalline and even nanocrystal-
line material we work with averaged grain size, whereas in using one-dimension structures, each sensor
is characterized by the specific geometry of the one-dimensional crystal. Therefore, reproducibility
of performance parameters for sensors based on one-dimensional structures depends on the unifor-
mity of those structures. Unfortunately, the problem of separation, sizing, and manipulation of one-
dimensional structures is not yet resolved. To achieve uniform sizing and orientation, advanced new
technologies will need to be implemented, and these will probably be expensive and not accessible for
wide use. There are a few interesting proposals for controlling one-dimensional structures (Heo et al.
2004), but they require further improvement for practical implementation.
Thus, chemical sensors based on individual one-dimensional structures are not yet readily avail-
able commercially. Further, the manufacturing cost of sensors based on one-dimensional structures
would far exceed that of polycrystalline devices. Therefore, in the near future, polycrystalline materi-
als will probably remain the dominant platform for solid-state gas sensors. However, nano- and poly-
crystalline materials are very complicated objects to study, because a lot of structural parameters in-
fluence on their properties and therefore control operating characteristics of chemical sensors based
on these materials.
6.1. GRAIN SIZE
At present, either the “grains” model or the “necks” models (see Figure 2.19) are applied to rationalize
the electrophysical properties of polycrystalline materials, which depend strongly on their microstruc-
ture (Xu et al. 1991; Barsan et al. 1999; Brynzari et al. 1999, 2000; Barsan and Weimar 2001; Rothschild
and. Komem 2004). The grain size and the width of the necks are also the main parameters that control
LS
S EC
EF
S
EC
EF
Rnarrow > Rbroad
Rlong > Rshort
Rclosed >> Ropened

Figure 2.19. Diagram illustrating the role of necks in the conductivity of a polycrystalline metal oxide
matrix and the potential distribution across the neck. (Reprinted with permission from Korotcenkov
gas-sensing properties in metal oxide films. Moreover, in the frame of modern gas-sensor models, the
influence of grain size and neck size on sensor response may be attributed to the fundamentals of gas-
sensor operation (Yamazoe 1991; Xu et al. 1991; Barsan et al. 1999; Brynzari et al. 1999, 2000a; Gas’kov
and Rumyantseva 2001; Rothschild and. Komem 2004; Schierbaum et al. 1991). Usually the relationship
is expressed through the so-called dimension effect, i.e., a comparison of the grain size (d ) or neck
width (X ) with the Debye length (LD ):
εkT
LD = (2.9)
e2N
where k is the Boltzmann constant, T is the absolute temperature, ε is the dielectric constant of the
material, and N is the concentration of charge carriers.
In brief, as an explanation of the dimension effect on the gas-sensing effect, we offer the following
argument (see Figure 2.19) (Schierbaum et al. 1991; Barsan and Weimar 2001). For large crystallites with
grain size diameter d >> 2LS, where LS is the width of the surface space charge L S = L D eVS / kt ,
and for a small neck width (d < LS), the conductance of both the film and ceramics usually is limited by
Schottky barriers (VS ) at the grain boundary. In this case the gas sensitivity is practically independent of d.
In the case of d ≈ 2Ls , every conductive channel in necks between grains is overlapped (see Figure
2.19). If the number of long necks is much larger than the intergrain contacts, they control the conduc-
tivity of the gas-sensing material and define the size dependence of the gas sensitivity.
If d < 2Ls , every grain is fully involved in the space-charge layer and the electron transport is af-
fected by the charge at the adsorbed species. Ogawa et al. (1982) demonstrated that when the grain size
becomes comparable to twice the Debye length, a space-charge region can develop in the whole crys-
tallite. The latter case is the most desirable, since it allows achievement of maximum sensor response.
More detailed descriptions of models used for explanation of both grain size and neck influence on
gas-sensing effects may be found in the literature (Xu et al. 1991; Yamazoe 1991; Yamazoe and Miura
1992; Wang et al. 1995; Barsan and Weimar 2001).
Note that the applicability of the “grains” or “necks” model depends strongly on the techno-
logical routes used for metal-oxide synthesis or deposition and the sintering conditions (Yamazoe
1991). Usually, the appearance of necks is a result of high-temperature annealing (Tan > 700–800°C).
Considering processes which take place at intergrain interfaces during high-temperature annealing, one
can assume that the forming of long necks in intergrain spaces is a consequence of mass transport from
one grain to another. According to Xu et al. (1991), for metal oxide samples after high-temperature
annealing, the neck size (X ) is proportional to the grain size (d ) with a proportionality constant (X/d )
of 0.8 ± 0.1. However, this constant depends on the sintering parameters and may vary. For thin metal
oxide films and ceramics which are not subject to high-temperature treatments, the gas-sensing matrix
is formed from separately grown grains. Therefore, in such metal oxides, the necks between grains are
very short or are absent. This means that we can use the “grains” model to describe of their gas-sensing
properties. According to this model, there are Schottky-type contacts between grains with the height of
the potential barrier depending on the surrounding atmosphere. In this approach, the grain boundary
space charge or band bending on intergrain interfaces are the main parameters controlling the conduc-
tivity of nanocrystalline metal oxides. The adequacy of this model was estimated from results obtained
during the impedance spectroscopy of metal oxides. (Bose et al. 2006; Kaur et al. 2005; Labeau et al.
1995; Labidi et al. 2005; Vasiliev et al. 2006).
Regarding experimental confirmation of the influence of grain size on sensor response, a dramatic
increase in sensitivity for metal oxides with grain size smaller than a Debye length has been demonstrat-
ed many times for various materials, such as SnO2 (Suzuki and Yamazaki 1990; Xu et al. 1991; Yamazoe
and Miura 1992; Brinzari et al. 2001; Korotcenkov, 2005, 2008), WO3 (Kanda and Maekawa 2005; Gillet
et al. 2005), and In2O3 (Gurlo et al. 1997; Korotcenkov et al. 2004, 2007).
This effect is illustrated in Figures 2.20 and 2.21 for In2O3 and SnO2-based sensors. Shimizu and
Egashira (1999) established that the sensitivity of sensors based on tin oxide nanoparticles increased
dramatically when the particle size was reduced to 6 nm. Below this critical grain size, sensor sensitivity
decreased rapidly. As the calculated Debye length of SnO2 is LD = 3 nm at 250°C (Ogawa et al. 1982),
the greatest sensitivity was actually reached when the particle diameter was 2 nm.
S~t -3
S~t-1.3
In2O3

Figure 2.20. Influence grain size on sensor response to (1) ozone (~1 ppm) and (2) NO2 (~1 ppm)
of undoped In2O3-based devices. (1) Thin-film technology. Films deposited by spray pyrolysis at Tpyr
= 475°C (Toper = 270°C; RH = 1–2%). (2) Thick-film technology. In2O3 powders were synthesized by
the sol-gel method (d ≈ 200 nm; Toper = 150°C; RH = 50%). (Experimental data from Korotcenkov et
al. 2004a and Gurlo et al. 1997.)
However, it is necessary to note that the threshold value of grain sizes determined by Shimizu
and Egashira (1999) is not an invariable constant of SnO2. This parameter must be dependent on the
material’s properties and the doping [see Eq. (2.9)]. For example, the Debye length in SnO2 (~3 nm),
estimated by Ogawa et al. (1982), corresponds to a donor concentration equal to nd = 3.6 × 1024 m−3.
The decrease in concentration of free charge carriers—e.g., by improving the metal oxide stoichiometry
through annealing or doping, can considerably increase this threshold value (Xu et al. 1991). Thus, Al-
doped SnO2 (see Figure 2.21, curve 8) shows high sensitivity with increasing grain size even at d > 20
nm, while Sb-doped SnO2 (see Figure 2.21, curve 6) is insensitive in the whole d region (see Figure 2.19).
If d >> LD, the increase of sensor response with grain size decrease is not so significant.
It is important to note that the relationship to grain size is dependent on the type of metal oxide,
the detection mechanism, and the gas being analyzed. For example, in In2O3-based gas sensors, the size
of the crystallites is appreciably less important for the detection of reducing gases in comparison with
oxidizing gases (Korotcenkov et al. 2004). Korotcenkov et al. (2004b) found that for certain deposition
conditions, the sensor signal to the reducing gases can either increase or decrease with increasing grain
size. In other words, for reducing gases, the In2O3 grain size is less important than for SnO2-based gas
sensors (Korotcenkov et al. 2004b).
For In2O3, it was also found that sensors fabricated using thin films with minimal crystallite size
had minimal response time (see Figure 2.22) in addition to maximum sensor response (see Figure 2.20).
For example, a decrease in grain size from 60–80 nm to 10–15 nm decreased τres during ozone detec-
Figure 2.21. (1–5) theoretical and (6–8) experimental dependencies of SnO2 sensor response to
reducing gas on grain size for (7) undoped SnO2 ceramics and ceramics doped by (6) Sb and (8) Al:
1, Nd = 1017 cm−3; 2, 3 × 1017 cm−3; 3, 1018 cm−3; 4, 3 × 1018 cm−3; 5, 1019 cm−3. Porous sensor ele-
ments were fabricated on an alumina tube with Pt wire electrodes. Elements were sintered at 700°C
for 4 h. SnO2 powders were synthesized by conventional hydrolysis of SnCl4. Foreign oxides (5 at.%)
were added by an impregnation method. Sensitivity to H2 (800 ppm) was estimated at 300°C. (Data
from Xu et al. 1991 and Korotcenkov et al. 2001. Reprinted with permission from Korotcenkov 2008.
In2O3
~t2
Figure 2.22. Influence of grain size on (1, 2) the response and (3) recovery times of the undoped
In2O3-based sensors during ozone detection: Tpyr = 475°C (0.2 M), Toper = 270°C, 1 and 2, RH = 1–2%;
3, RH = 45–50%. (Adapted with permission from Korotcenkov 2004a. Copyright 2004 Elsevier.)
tion in dry air by a factor of 50–100 times. In a humidified atmosphere, however, the correlation was
significantly weakened (Korotcenkov et al. 2004a, 2004c).
On the other hand, the decrease in grain size cannot be unlimited. At some critical dimension, the
number of free electrons in the grain can become zero even at VS = 0. This leads to a grain resistance
that does not depend on changes in the surrounding atmosphere. For charge-carrier concentrations in
metal oxides of 1021 cm−3, the critical crystallite size is 1 nm.
The use of finely dispersed small crystallites can also have a deleterious effect on the temporal
stability of the sensor (Korotcenkov 2005; Korotcenkov et al. 2005). An excessive decrease of grain
size leads to a loss of structural stability (Korotcenkov 2005; Korotcenkov et al. 2005b) and, as a con-
sequence, to changes in both the surface properties and the catalytic properties of the material (Rao et
al. 2002). For example, it was shown that for SnO2 with a grain size of about 1–4 nm, the grain growth
process already begins at temperatures of ~200–400°С (see Figure 2.23) (Shek et al. 1999; Korotcenkov
et al. 2005b). In contrast, SnO2 crystallites with average sizes ranging from 1.7 to 4.0 μm were stable
up to 1050°С.
It thus becomes clear that the presence of a finely dispersed fraction with a grain size smaller than
2–5 nm will lead to some structural instability of the metal oxide matrix at moderate operating tempera-
tures (Т < 600°С). This is true even for films with average grain sizes greater than 100 nm. Therefore,
future design methods for nano-scale devices, which assure grain size stabilization during long-term
operation at high temperature, will gain priority over the design of methods that produce nano-scaled
materials with minimal grain size. The production of both metal oxide powders and deposition of
SnO2

Figure 2.23. Influence of annealing temperature on grain size in SnO2 thin films, thick films, and
ceramics fabricated using different manufacturing methods. (Reprinted with permission from
Korotcenkov 2005b. Copyright 2005 Elsevier.)
Figure 2.24. Influence of pyrolysis temperature on sensitivity to air humidity during (1) H2 (5000 ppm)
detection by undoped In2O3 sensors (Toper = 370°C) and (2) O3 (1 ppm) detection by undoped SnO2
sensors (Toper = 270°C), and (3) on the grain size of In2O3- and SnO2-based sensors. Dry air ~1% RH,
wet air ~35–45% RH. (Adapted with permission from Korotcenkov 2008. Copyright 2008 Elsevier.)
films with a small dispersion of grain sizes is an effective method for improving the temporal stability
of solid-state gas sensors as well. Utilizing thin films in gas sensors also leads to improved stability of
the gas-sensing matrix. Research has shown that the grain size during the annealing process changes
considerably less in thin films than in thick films. Therefore, unless there is a pressing need to improve
detection limits (sensitivity), it is not advisable to design sensors with excessively reduced grain sizes.
This problem is especially serious for operation in atmospheres of reducing gases.
As the grain size decreases, there is one more interesting effect. With decreased grain size, a signifi-
cantly increased sensitivity to humidity was observed (see Figure 2.24). The same effect was observed
experimentally by Williams and Coles (1998b). Williams and Coles found that alumina-based humidity
sensors also increased their sensitivity by an order of magnitude when they were prepared from 13-nm
particles as compared to 300-nm particles. This means that the humidity influence of sensor parameters
is structure-dependent and may be controlled through structural engineering of the metal oxides used.
It has been confirmed that water adsorption on the surface of metal oxides depends on the crystallo-
graphic structure (Emiroglu et al. 2001; Golovanov et al. 2005; Batzill et al. 2005, 2006a). The difference
in water attachment on different crystallographic SnO2 planes appeared not only in concentration, but
also in the type of bonds between hydroxyl groups and the SnO2 surface. The OH groups participate
in gas-detection reactions, and water adsorption/desorption processes can control the kinetics of gas
response (Korotcenkov et al. 1998, 2004d, 2004e; Michel et al. 1998; Barsan et al. 1999). Thus, if mini-
mum sensitivity to humidity is required, the size of the crystallites in the sensor material should be the
maximum permissible.
Mechanical properties of materials are connected directly to their crystal structure as well. Zhang et
al. (2003) showed that with a decrease in grain size, the multiplication and mobility of the dislocations
are hindered, and the hardness of materials increases according to the Hall-Petch relationship, H(d ) =
HО + Kt-1/2. This effect is especially prominent for grain size (t ) down to tens of nanometers (see Figure
2.25). However, dislocation movement, which determines the hardness and strength in bulk materials,
has little effect when the grain size is less than approximately 10 nm. At this grain size, further reduction
in grain size brings about a decrease in strength because of grain boundary sliding. Softening caused
by grain boundary sliding is attributed mainly to a large amount of defects in grain boundaries, which
allows fast diffusion of atoms and vacancies under stress.
6.2 CRYSTAL SHAPE

At present, the effects of grain shape and faceting of crystallites on gas sensing has not been studied.
However, there are reasons to believe that the role of these parameters is undeservedly understated
because, depending on the external form of the nanocrystallites, a nanostructured gas-sensing matrix
will have a unique combination of structural, electronic, and adsorption/desorption process parameters
(Henrich and Cox 1994; Lucas et al. 2001), these parameters control the operating characteristics of all
types of chemical sensors (Brinzari et al. 2001, 2002; Korotcenkov 2005; Golovanov et al. 2005; Batzill
et al. 2005; Korotcenkov et al. 2005c; Batzill, 2006d). So, the determination of crystallographic planes
with optimal combinations of adsorption/desorption and catalytic parameters, and the development
of methods for metal-oxide deposition or synthesis with desired faces of grains can be considered an
important contemporary goal for thin-film technology as applied to metal-oxide gas sensors. Current
research results (Lucas et al. 2001; Kawabe et al. 2001; Brinzari et al. 2002; Batzill et al. 2005) can be
considered as the first step in this direction. For example, research carried out by Batzill et al. (2005)

Figure 2.25. Schematic diagram of the influence of grain size on the hardness of the covering.
(Adapted with permission from Zhang et al. 2003. Copyright 2003 Elsevier.)
has shown that the (110), (110), and (101) SnO2 planes have different surface energies, different phase-
transition conditions, and different energy spectra of surface states generated during their reduction.
Different crystallographic planes also have different peculiarities of interaction with water (Golovanov
et al. 2005; Batzill, 2006a). Kawabe et al. (2001) established that SnO2 films with different texturings
have different catalytic activity for selective oxidation of CH4. In these investigations, one type of SnO2
film had predominant orientation in the (110) crystallographic direction and another in the (211) and
(301) directions. Lucas et al. (2001) observed that monodentate adsorption on MgO and CaO is pre-
ferred on the edge/corner sites. This effect is unique to small nanocrystallites (spherulites). In contrast,
bidentate adsorption is favored by flat planes, which are more prevalent on the larger nanocrystals. This
feature of surface species adsorption indicates that the control of surface roughness and grain shape
can be exploited for improvement of gas-sensing characteristics such as absolute magnitude and selec-
tivity. Such control may be as effective as the use of catalytic additives. This last factor is important for
improving the stability of solid-state gas sensors (Korotcenkov 2005a).
Detailed studies (Brinzari et al. 2002; Korotcenkov 2005a; Korotcenkov et al. 2005a, 2005cc) con-
ducted on SnO2 films deposited by a spray pyrolysis method confirmed the appropriateness of this ap-
proach for modifying gas-sensing properties. For example, it was established that the growth of grains,
especially in the range from nanometers to micrometers, during which transition from spherulites to
nanocrystallites and from nanocrystallites to nanocrystals and crystals takes place, is accompanied by
changes in both the size and the external shape of the crystallites (Brinzari et al. 2002; Korotcenkov et
al. 2005). Possible morphologies of SnO2 films deposited by spray pyrolysis are shown in Figure 2.26.
A proper choice for crystallite deposition technology or synthesis with a necessary grain facet can
be also one method to decrease humidity effects in gas sensors. For example, Golovanov et al. (2005),
using a Mulliken population analysis, have shown that the chemisorption of OH groups on the (110)
face is accompanied by the localization of negative charge to a greater extent than the chemisorption of
OH groups at the (011) surface of SnO2. This means that adsorption/desorption processes and surface
reactions with water vary with different SnO2 crystallographic planes.
Understandably, the preparation of polycrystalline metal oxides with necessary grain faceting is
difficult to control, but it is achievable for one-dimensional sensors and should be a high-priority area
of research. One-dimensional structures are crystallographically perfect and have clear faceting with a
fixed set of planes, which may be modified via control of synthesis parameters (Liang et al. 2001; Dai
et al. 2003; Kong et al. 2003; Li et al. 2003).
Semiconducting one-dimensional metal oxide structures with well-defined geometry and perfect
crystallinity might represent a perfect model material family for systematic experimental study and
theoretical simulation of gas-sensing effects in metal oxides. Many parameters used for characterizing
polycrystalline materials lose meaning for one-dimensional structures, because they are single-crystalline
materials. These parameters include film thickness, porosity, grain size, grain network, grain boundary,
agglomeration, and texturing, i.e., all the parameters that we have considered that influence the metal
oxide sensor response. The main structural and morphologic parameters that characterize one-dimen-
sional structures are geometric size, characterizing the profile of one-dimensional structures, and crys-
tallographic planes, framing these one-dimensional structures (see Figure 2.27).
The minimal distance between faceting planes in one-dimensional structures plays the same role
in gas-sensing effects as grain size plays in polycrystalline material. Undoubtedly, the decreased number

Figure 2.26. SEM images of SnO2 films with different thicknesses (Tpyr = 450–475°C): (a) d ≈ 40–50
nm; (b) d ≈ 80–90 nm; (c) d ≈ 310–380 nm. (Reprinted with permission from Korotcenkov 2005a.
Figure 2.27. Schematic diagram of the geometric configuration of SnO2 nanobelts. (Reprinted with
permission from Korotcenkov 2008. Copyright 2008 Elsevier.)
of parameters which control the sensor response of one-dimensional structures should facilitate better
understanding of the nature of the observed effects. Theoretical simulation of gas-sensing effects can
become significantly simpler. As noted above, only two planes usually participate in gas-sensing effects
in one-dimensional structures. This means that in simulations, one-dimensional structures should be
considered as single crystals with limited sizes. On the other hand, in one-dimensional gas sensors the
role of contacts increases because of their small area and therefore greater specific resistance.
It should be noted that the indicated simplified approach to consideration of gas-sensing effects in
one-dimensional structures is possible only when sensors are formed based on single one-dimensional
structures with invariability along the length parameters. In the presence of segmented morphology,
small-diameter segments may act as necks between particles in traditional thin-film gas sensors, but
with the significant advantage of greater morphological integrity and stability. Studies conducted by
Dmitriev et al. (2007) have shown that the responsiveness of these structurally modulated nanowires to
gases is improved over that of straight nanowires of the same average diameter. The better tolerance
of such nanostructures toward contact effects is another advantage of such chemiresistors based on
quasi-one-dimensional single-crystal nanostructures.
Sensors based on nanowire arrays behave similarly to sensors based on the usual nanoparticle films.
Impedance spectroscopy studies have showed that the gas-sensing mechanism for sensors based on
networked nanowire thin films involves changes in both the nanowire and the internanowire bound-
ary resistances (Deb et al. 2007). Thin films containing nanowires in a highly networked fashion show
promise for gas sensor design. A study carried out by Deb et al. (2007) showed that the sensors based
on these films had behavior similar to that of single nanowire devices without much postprocessing.
According to Deb et al. (2007), this is possible only when the individual nanowires within the gas-
sensing matrix are bonded to each other and form two- and three-dimensional networks of nanowires
6.3. SURFACE GEOMETRY
The role of surface geometry in gas-sensing effects, which includes the role of surface steps, terraces,
crystallographic defects, and corners, has not been studied in detail, in spite of the fact that the influence
of surface faceting in gas adsorption phenomena and catalysis is a well-known phenomenon (Madey et al.
1999; Pederson et al. 2005). The concept of an “active site” is one of the fundamentals in heterogeneous
catalysis. This notion reflects experimental evidence that, in general, catalytic surfaces are not uniformly
active. Rather, only sites distinguished by a special arrangement of surface atoms (including defects), or by
a particular chemical composition, are actually reactive. For example, numerous experiments have shown
that in many cases monoatomic steps are highly preferential sites for many surface reactions. To address
this issue, Zambelli et al. (1996) studied the dissociative chemisorption of NO on the (0001) surface of
ruthenium, which is known to be the most selective catalyst. The surface was exposed to a small dose
of NO (0.3 Langmuir, 1 L = 10−6 torr s) at room temperature, and scanning tunnel microscopy (STM)
images were recorded. From the distribution of product atoms observed in the STM micrographs, it was
inferred that adsorbed NO molecules diffuse rapidly across surface terraces until they meet a step, where
they are observed to dissociate with high probability. Another good example of the structural sensitivity
of a gas interaction with a solid-state surface is the behavior of hydrogen at the surface of SnO2. Kohl
(1990, 2001) showed that H2 molecules are not activated on smooth SnO2 surfaces of single crystals.
The activation means excitation of a bond and, as a possible consequence, ionization, dissociation, or
formation of radicals. The same effect was observed for WO3 and V2O5. However, on the rough surface
of sintered SnO2, activation and reaction to H2O occurs at 470 K. In contrast to H2, as shown by Kohl
(1990, 2001), the methane can be activated on a smooth SnO2 surface, forming a CH3 radical.
The important role of surface steps in diffusion processes was established for perovskites as well.
Chen et al. (2002) found that the in-diffusion of surface oxygen vacancies takes place mostly at the step
edges. Wang et al. (2007a), using a Monte Carlo simulation of the deposition process and a mean-field
theory model of the SrTiO3 surface structure, established that, during deposition by laser molecular
beam epitaxy, the concentration of oxygen vacancies close to step edges is greater than that on flat
terraces and remains stable. It was suggested that oxygen vacancies diffusing at the surface tend to ac-
cumulate near step edges due to their slow in-diffusion rate there, and that this in-diffusion dominates
the oxidation of the as-deposited film. The step edges act as a route for oxygen vacancies at the surface
to move into the film. Wang et al. (2007a) concluded that the in-diffusion rate of oxygen vacancies is
limited by step edges. No matter how large the surface diffusion rate is, the oxidation process is domi-
nated by the in-diffusion of oxygen vacancies near step edges.
These observations illustrate that the adsorption properties of many gases are very sensitive to the
microscopic surface geometry of metal oxides. Therefore, by expanding our understanding of these
processes along with the development of necessary technologies, this phenomenon could be used to
improve the selectivity of solid-state gas sensors. For example, it is possible that the increased sensitiv-
ity to nitrogen oxide observed after high-temperature annealing and subsequent mechanical milling
of SnO2 (Dieguez et al. 2000; Tan et al. 2000; Abe et al. 2005) is a consequence of the appearance of
microsteps formed as a result of the mechanical crushing.
In the context of the role of surface morphology in sensor performance, the results of Batzill et
al. (2006) deserve special attention. Batzill et al. (2006) analyzed the interaction of water with a SnO2
(101) surface and found that the water interacts weakly with the reduced surface; the small amount of
water at the SnO2 surface for T > 130 K may be connected with water adsorption at surface defect sites
such as step edges (Batzill et al. 2005). This implies that surface geometry may control water influence
on gas response as well.
The effect of noble-metal surface clustering, which has not been studied in terms of gas sensing,
may be directly connected with surface geometry as well. There are statements (Yamazoe et al. 1991;
Kabbabi et al. 1994; Meier et al. 2004; Castañeda 2007) indicating that the dispersion state of surface
catalyst particles, which can be characterized by particle size and population, can favorably affect gas-
sensing characteristics. It has been shown that improved sensitivity can be achieved if the aggregate
distribution of noble metals such as Pd and Pt in the film is characterized by very small particle size
coupled with a high number density (Matko et al. 2002). However, it was impossible to find in the lit-
erature any correlation between the size of surface clusters and gas sensitivity.
Recent research, however, has shown that the process of surface clustering is structurally sensitive
(Chen et al. 2002), and therefore the consequences of surface modification may depend on the surface
structure of the metal oxides. The most important consequences pertain to such structural parameters
as surface geometry, i.e., the presence of steps, terraces, and facets, the size of planes, the degree of
surface reduction, and other parameters (El-Azab et al. 2002; Min et al. 2006).
Recent work (Valden et al. 1998; El-Azab et al. 2002; Lauritsen et al. 2006) has provided direct
experimental confirmation that, following thermal treatments, noble-metal clusters accumulate at the
step edges of metal oxides (see Figure 2.28). Due to this behavior of noble metals, an increase of ter-
race area should be accompanied by an increase in the size of clusters, which can be accumulated at the
step edges of this terrace. This means that for the same degree of surface coverage by noble metals, the
number of clusters will be smaller, while the distance between the clusters will be bigger on the surface
with the bigger terrace area (see Figure 2.29).
From this it follows that the appearance of extended, atomically flat surface planes facilitates the
process of cluster growth. At the same time, the presence of atomically stepped-like surfaces will pro-
vide with high probability the conditions for atomic dispersion of noble metals at the surface of metal
oxides. Min et al. (2006) established that Au clusters had higher densities on highly reduced TiO2. It was
suggested that reduced Ti sites act as active sites for the nucleation and growth of Au clusters.
Figure 2.28. STM image of Au/TiO2(110) sur-

face after annealing at 850 K for 2 min. The Au
coverage was 0.25 monolayer (ML). (Reprinted
with permission from Lai et al. 1998. Copyright
1998 Elsevier.)
(b)
(a)
(c)

Figure 2.29. Schematic illustration of the influence of terrace size on the size of noble metal clusters.
(Reprinted with permission from Korotcenkov 2005. Copyright 2005 Elsevier.)
The size of clusters also depends on the nature of the metal deposited. For example, Alfredsson
et al. (2004), analyzing the behavior of Pt and Pd atoms on a ZrO2 surface, found that the Pd atoms
have higher surface mobility and more easily form larger metal clusters than Pt. The higher mobility of
the Pd agrees with other experimental data (Alfredsson et al. 2004). It was established that after anneal-
ing at 700 K, the cluster size distribution of Pd on the TiO2 (111) surface moves readily toward larger
cluster sizes. Alternatively, the size distribution of Pt is unchanged after annealing, with smaller clusters
being stable at higher temperatures. Furthermore, it was found that Pd shows a higher probability of
adsorbing during deposition on both terrace and step sites, while Pt is characteristically observed on the
steps (see Figure 2.30).
The thickness of deposited noble-metal layers appreciably influences the cluster size as well.
As a rule, for maximum sensor response, the concentration of noble-metal additives in the bulk of
metal oxides should not exceed 1–2 wt% (Matsushima et al. 1988; Arbiol et al. 2002; Matko et al.
2002). According to Matko et al. (2002), Pt clusters in SnO2 have a mean size smaller than 1.5–2.0 nm.
Experimental results indicate that the thickness of the deposited noble-metal layer for films prepared
by rheotaxial growth and thermal oxidation (RGTO) (Sberveglieri et al. 1991) should not exceed 2.5
nm to attain the same effect during surface doping. For thin metal oxides this thickness may be even
less (Tsud et al. 2001; Veltruska et al. 2001; Korotcenkov et al. 2003a). Research has shown that for
Figure 2.30. Possible positions of (a) Pd and (b) Pt clusters on the surface of ZrO2. (Reprinted with
permission from Alfredsson et al. 2004. Copyright 2004 Elsevier.)
the formation of an atomically dispersed Pd layer at the SnO2 surface, the Pd thickness should not
exceed 1 monolayer. The deposition of several milliliter equivalents of Pd leads to the formation of
three-dimensional Pd particles with an average size in the range 3–5 nm (Tsud et al. 2001; Veltruska et
al. 2001). Approximately the same size Pd clusters on SnO2 grain surfaces were observed upon bulk
doping when the concentration of doping additives was 5 wt% (Arbiol et al. 2002). Similar results were
obtained by Alfredsson et al. (2004). Auger spectroscopic studies on Pd and Pt supported on ZrO2
show that both Pd and Pt have layer-by-layer growth at low metal coverage and temperatures. At el-
evated temperatures, or following annealing, the Pd-clusters are more raftlike (two-dimensional), while
the Pt-clusters show three-dimensional islands (Alfredsson et al. 2004).
Based on the catalytic activity of gold (Au) clusters (Haruta 1997; Valden et al. 1998; Wahlstrom
et al. 2003; Campbell 2004), one can judge how important it is to form and stabilize noble-metal clus-
ters with a specified size. Gold has long been known to be essentially inert chemically (and therefore
catalytically inactive) in its bulk form (Hammer and Norskov 1995). Quite recently, however, it was
found that when gold is dispersed in the form of nanoscale clusters and supported on transition-metal
oxide surfaces such as TiO2 (Valden et al. 1998; Wahlstrom et al. 2003), Au exhibits very high catalytic
activity at low temperature for the partial oxidation of CO, hydrocarbons, hydrogenation of unsatu-
rated hydrocarbons, and reduction of nitrogen oxides. It was shown that the catalytic properties of Au
nanoparticles depend specifically on the support, the preparation method, and, critically, on the size
of the Au clusters, which are most active when their diameter is smaller than 3.5 nm (Bamwenda et
al. 1997; Valden et al. 1998) (see Figure 2.31). Cluster diameters smaller than 3 nm lead to a decrease
in reactivity.
It is known that the activity of palladium increases substantially with a decrease in cluster size. For
example, Maier et al. (2004) reported that catalytic activity toward electrochemical proton reduction is
enhanced by more than two orders of magnitude as the diameter of the palladium particles, measured
parallel to the support surface, decreases from 200 to 6 nm. Thus, during fabrication of surface-mod-
ified metal-oxide gas sensors, one should learn not to create clusters of specified size and to stabilize
their size in the desired range.
It is important to note that each surface catalyst has its own optimal size for providing maximum
response for each analyte. Research results (Kohl 1990; Yamazoe et al. 1991; Williams 1999; Cabot et
al. 2001; Van de Krol and Tuller 2002; Korotcenkov et al. 2003a) provide good examples of this ef-
fect. It is also important to note that the size of noble metal clusters that are optimal for gas sensitivity
does not coincide with the optimal size for heterogeneous catalysis (Kabbabi et al. 1994). It therefore
follows that the surface noble-metal clusters in gas sensors and heterogeneous catalysis perform dif-
ferent functions.
6.4. FILM TEXTURE
Film texture has not been considered so far because in ceramics all grains are oriented arbitrarily. Only
since the transition to film technology has film texture started attracting attention. For example, Ning
et al. (2007) established that the highly preferred orientation of homogeneous ZnO films decreased the
electrical resistance of ZnO ceramics at room temperature. Ning et al. (2007) also found that textured
Figure 2.31. STM image (bottom left) and schematic (upper left) of Au clusters on TiO2(110) whose
population is dominant (lower right) for a Au/TiO2 catalyst most active for CO oxidation (upper right).
The activity (at 350 K) is expressed as turnover frequency (TOF) in units of CO2 molecules produced
per Au site per second. (Reprinted with permission from Choudhary and Goodman 2005. Copyright
2005 Elsevier.)
ZnO films have more low-angle grain boundaries in comparison with films without texture. Another in-
teresting effect was observed by Korotcenkov et al. (2004a, 2004c): They found that for ozone detection
using In2O3-based sensors, the recovery time increases as the film texture increases (see Figure 2.32). For
reducing gases this dependence is observed for both response (τres ) and recovery times (τrec ). Moreover,
the change in τres and τrec in the sensor response to reducing gases was subject to the same regularity of
Tpyr and film thickness influence as was the change in τrec for ozone detection. In other words, the faster
response times were obtained from structures formed from arbitrarily oriented crystals. It has been
assumed that the influence of the film’s texture on gas-sensing characteristics takes place through the
change of porosity of the gas-sensing matrix. Korotcenkov et al. (2002, 2004a, 2004b) studied In2O3
films textured in the (001) direction perpendicular to the substrate. The degree of texture increases
when both the sintering temperature and the film thickness increase. One can assume that the indicated
parameters of In2O3 film deposition promote more compact crystallite packing. The decrease of film
porosity and gas permeability may be the result of such changes in the In2O3 film structure.
An attempt to estimate the influence of film texture on the In2O3 conductivity response to NO2 was
made by Steffes et al. (1999). It was found that In2O3 films with preferential orientation in the (211) direc-
tion had higher sensor response to NO2 than the layers with preferential (222) texture. Films studied by
Korotcenkov et al. (2002, 2004a, 2004b) had a different texture than the films used by Steffes et al. (1999).
Therefore, it was not possibly to compare the results discussed by Korotcenkov et al. (2002, 2004a,
2004b) and Steffes et al. (1999). As a result, we can neither confirm nor refute the above conclusions.
However, the existence of such an effect, in addition to the findings presented by Korotcenkov et al.
(2005), allows us to conclude that modification of the film texture of polycrystalline films as well as de-
creasing grain size and film thickness may open a new way to optimize thin-film gas-sensor parameters.
6.5. SURFACE STOICHIOMETRY (DISORDERING)

Surface stoichiometry (disordering) of the metal oxide grains may be as important a parameter in gas
sensing as film thickness and grain size. This parameter determines the adsorption ability and the sur-
face charge through the number of oxygen vacancies, and as a result, controls the initial surface band
bending (eVs ) of the metal oxide and the change of eVs upon replacement of the surrounding gas.
Considering that the sensing properties of thin-film sensors, such as sensitivity, selectivity, and
stability, are strongly related to their microstructures and to the exact stoichiometry of their surfaces, ac-
curate control of these parameters is extremely important for the production of sensors with reproduc-
ible behavior (Korotcenkov et al. 2005b; Mašek et al. 2006). For example, a change in the surface oxygen
composition of SnO2 is accompanied by a change in the electronic structure of the surface Sn from a
valence of IV for the stoichiometric surface to a valence of II for the reduced surface. This change in
composition also has a strong influence on the chemical and gas-sensing properties. Batzill et al. (2005)
showed for the adsorption of water that water dissociates on the stoichiometric surface but adsorbs
weakly on the reduced surface. While this easy reduction of the SnO2 surface can play an important
Figure 2.32. Influence of In2O3 film texture on recovery time during ozone detection. Films with
thicknesses of 40–300 nm were deposited by spray pyrolysis from 1.0 M InCl3–water solution (Toper =
270°C). (Reprinted with permission from Korotcenkov 2004a. Copyright 2004 Elsevier.)
role in the gas-sensing behavior of this material, it is unlikely to be of general importance for other gas-
sensing metal oxides. The particular behavior of the SnO2 surfaces arises from the dual valency of Sn.
For other gas-sensing materials, e.g., ZnO, no variation in the lattice-oxygen composition is expected.
Excessive lattice disordering of the surface layer may be accompanied by significant growth of
native surface states. Theoretical simulations presented by Brynzari et al. (1999, 2000a, 2001) indicate
that an increase of these surface defects may lead to pinning of the surface Fermi-level position, and
correspondingly to a drop in response of gas sensors. This means that if we want to maintain effective
operation of solid-state gas sensors, the concentration of those states as well as the lattice disordering
of the surface layer of metal oxide grains should be minimized. Only in this case will the surface Fermi
level not be pinned, because the charge of native surface states becomes comparable to or less than the
charge of chemisorbed particles. The indicated surface property creates a condition for modulation of
surface band bending of semiconductors with the change of the surrounding atmosphere. The change
of surface stoichiometry of metal oxide films may result from changes in deposition and annealing
temperature (Brinzari et al. 2000b).
High-resolution transmission electron microscopy (HRTEM) images of SnO2 grains presented by
Nayral et al. (2000) (see Figure 2.33) indicate that this effect may occur in metal oxides. In many cases,
the particles of SnO2 consist of a well-crystallized core covered with an amorphous layer of tin oxide.
Further, such amorphous layers may appear in undoped tin dioxide fabricated by low-temperature tech-
nological routes, or in doped material at superfluous concentrations of doping additives.
The same conclusion was reached by Dieguez et al. (2001) on the basis of an analysis of the com-
plete Raman spectrum of nanometric SnO2 particles. Dieguez et al. (2001) found that for grains smaller
than 7 nm, two bands appeared in the high-frequency region of spectrum. They proposed that these
Amorphous tin oxide layer

Tin oxide core

Figure 2.33. HREM images of SnO2 nanoparticles showing the crystallized tin oxide core and the
amorphous tin oxide shell: (a) SnO2:Pd; (b) undoped SnO2. (Adapted with permission from Nayral et
al. 2000. Copyright 2000 Elsevier.)
bands were due to a surface layer of nonstoichiometric SnO2 with a different symmetry than SnO2. The
thickness of this layer was calculated to be ~1.1 nm, which is about two to three unit cells.
One can conclude that the presence of such a layer on the surface of small grains is a main reason
for lower thermal stability of the metal oxide structure and the enhanced sensitivity to air humidity of
sensors fabricated of fine-dispersed material with grain size smaller than 5–7 nm.
The presence of a nonstoichiometric surface layer may also be responsible for a strong interac-
tion between noble metals and metal oxides. The activity of noble-metal catalysts incorporated on
solid-state sensors is determined by ether their chemical state, aggregation form, or interaction with
metal oxide (Castañeda 2007). The formation of alloys is the main reason for the indicated dependence
(Hanys et al. 2006).
Wahlstrom et al. (2004) assumed that the surface stoichiometry of metal oxides, through a charge
transfer–induced diffusion mechanism for O2 molecules adsorbed on a metal oxide surface, can also
control the surface diffusion of oxygen molecules. Time-resolved scanning tunneling microscopy of
the TiO2 (110) surface has shown that the O2 hopping rate depends on the number of surface donors
(oxygen vacancies), which determines the density of conduction band electrons. Wahlstrom et al. (2004)
assumed that the metal oxides act as reservoirs for oxygen; and the O2 diffusion may be a rate-limiting
step in oxidation processes on these metal oxides. Diffusion of oxygen molecules on a metal oxide
surface plays a vital role in gas-sensing effects, and therefore these results may have implications for
understanding their nature. This mechanism is expected to be an important one for such reducible
oxides as TiO2, Fe2O3, SnO2, and ZnO, where shallow donor states provide a rise to a high density of
electrons in the conduction band
6.6. POROSITY AND ACTIVE SURFACE AREA
From an analysis of the numerous parameters which can affect chemoresistance sensors, one can con-
clude that in order to achieve maximum gas sensitivity it is necessary either to increase the role of surface
conductivity or to increase the contribution of intercrystalline barriers by decreasing either the contact
area or the width of the neck (Brynzari et al. 1999, 2000; Barsan and Weimar 2001) (see Figure 2.19).
At present there is no doubt that the simplest method to achieve maximum sensitivity indepen-
dently on gas-sensing material is to increase the film’s porosity (Park et al. 1996; Basu et al. 2001; Ahmad
et al. 2003; Hyodo et al. 2003; Lee et al. 2005; Jin et al. 2005; Tesfamichael et al. 2007). In the case of
compact layers, the interaction with gases takes place only at the geometric surface (see Figure 2.34).
In porous material the volume of the layer is accessible to the gases, and therefore the active
surface is much greater than the geometric area and the sensor response is higher (Barsan et al. 1999).
Higher porosity results in a smaller number of contacts with the necks that are not overlapped under
interaction with the surrounding gas. Experimental results obtained from a study of gas-sensing charac-
teristics of SnO2 films synthesized by the RGTO technique (Sberveglieri et al. 1991; Sherveglieri 1992b)
are in good agreement with this conclusion. The films had high sensitivity in spite of the high level of
agglomeration. Further, it was observed that films with minimal contact area between agglomerates
showed maximum sensitivity. This was clearly illustrated by the SEM images presented in (Sberveglieri
et al. 1991; Sherveglieri 1992b). Increased porosity also decreased the probability of forming so-called

Gas Gas
products products
Current flow
Figure 2.34. Schematic representation of the influence of material density on gas penetrability of
gas-sensing matrix. (Reprinted with permission from Korotcenkov 2008. Copyright 2008 Elsevier.)
capsulated zones in the volume of the gas-sensing layer. Capsulated zones are isolated from contact
with the atmosphere, and therefore their resistance does not depend on the surrounding gas (McAleer
et al. 1987). Density, porosity, pore size, and gas permeability can be determined by appropriate mea-
surement techniques (Ishizaki et al. 1998; Nguyen and Do 1999; Wang et al. 2007b). As has been estab-
lished, all these parameters have a direct correlation with grain size (see Figure 2.35).
The specific surface area parameter, which is rarely used in characterizing gas-sensing materials,
deserves more attention. Rumyantseva et al. (2005) have expressed the same opinion. Experiments
(Yamazoe et al. 1979; Yamazoe 1991; Li et al. 1999; Ahmad et al. 2003; Rumyantseva et al. 2005) have
indicated that the use of material with high specific surface area may lead to the development of highly
sensitive metal-oxide gas sensors (see Figure 2.36).
Specific surface area has a more universal meaning than generally used parameters, because it com-
bines such parameters as porosity and grain size. From data on specific surface areas, average grain size
may be calculated using the equation (Zhang and Gao 2004; Rumyantseva et al. 2005)
6
d SA = (2.10)
ρA
where dSA is the average grain size of spherical particles, A is the surface area of the powder, and ρ is the
theoretical density of SnO2. Moreover, in many cases the specific surface area shows better correlation
with gas response than does crystallite size (Gas’kov and Rumyantseva 2001; Ahmad et al. 2003).
As a rule, an increase in specific surface area of a gas-sensing material is accompanied by a shift in
the sensitivity maximum into the range of lower operating temperatures (see Figure 2.37). One can as-
sume that this effect is caused by the thermally activated nature of gas diffusion inside the gas-sensing
matrix (Ruiz et al. 2005).
Specific surface area data, estimated by classical multipoint Brunauer, Emmett, and Teller (BET)
adsorption techniques, and grain size, calculated using x-ray diffraction (XRD) data, provide a more
Figure 2.35. Schematic illustration of the influence of grain size on parameters of sensing materials.
complete and reliable description of the gas-sensing material. The difference between grain sizes de-
termined by those methods (see Table 2.19) may be used for porosity characterization (Rumyantseva et
al. 2005). However, the specific surface area should be determined only after conducting all operations
used during the sensor’s fabrication.

SnO2
H2
CO
Figure 2.36. Relationship between surface areas of SnO2 sensors and their sensitivity to 500 ppm
of H2 and CO at 300°C. Sensors were fabricated by pressing and annealing at 450°C calcined SnO2
powders into 14-mm-diameter and 1-mm-thick pellets with two Pt electrodes. SnO2 for these experi-
ments were prepared by a surfactant-templating method. The as-synthesized SnO2 samples were
calcined at 450°C. (Data from Li et al. 1999. Reprinted with permission from Korotcenkov 2008.

SnO2
500 ppm H2
Figure 2.37. Influence of surface area of SnO2-based thick-film sensors on response to 500 ppm of
H2: 1, 3 m2/g; 2, 54 m2/g; 3, 69 m2/g; 4, 77 m2/g; 5, 92 m2/g; 6, 99 m2/g. Tested sensors were fabri-
cated using technology described in Figure 2.38. (Data from Li et al. 1999. Reprinted with permission
from Korotcenkov 2008. Copyright 2008 Elsevier.)
Thus, as the porosity and active surface area of the gas-sensing material increases, the sensor re-
sponse increases as well (see Figures 2.36 and 2.37). This is consistent with the conclusion of Matko
et al. (2002) that the porosity and specific surface area are basic factors that influence the solid/gas
interactions and ultimately the material performance in gas detection. This is an important conclusion.
However, it is necessary to know that in some cases, dense, nonporous metal oxides may compensate
for their lower sensitivity and other shortcomings by having higher temporal and thermal stability (Min
and Choi 2004; Kocemba et al. 2001).
Sever studies (Park et al. 1996; De Souza Brito et al. 1998; Korotcenkov et al. 2004a) have shown
that gas sensors with higher porosity have faster response. Park et al. (1996) established that the refrac-
tive index correlates with film porosity. If the porosity is higher, the refractive index is smaller. It is
known that only nonporous metal oxides have refractive indices equal to the tabular data estimated for
these materials. Pores in metal oxides can be considered as a component with a refractive index equal
to 1.0. Therefore, films with smaller refractive index along with invariability of the main parameters
must have a higher degree of porosity and better gas permeability. Figure 2.38 shows that In2O3 films
with low refractive index really have minimal τres and τreс. Korotcenkov et al. (2004b) concluded that for
the detection of reducing gases in In2O3 films, the porosity or the gas permeability is a more important
factor than the grain size.
Within the framework of the above discussion, results reported by De Souza Brito et al. (1995)
are of interest. The structural properties of SnO2 films synthesized using sol-gel technology were
analyzed, and it was found that the pore size increased during sintering within the temperature range
from 200 to 700°C as indicated by growth in the grain size. Moreover, it was established that in the
low-temperature region (TS < 400°C), the average micropore diameter increased slightly. In the high-
temperature region (TS > 400°C), this process was intensified, and the growth of average pore diameter
and crystallite size were followed by a drastic reduction of the surface area (see Figure 2.39). The struc-
tural evaluation during sintering of compacted SnO2 sol-gel powders was performed using nitrogen
adsorption isotherm analysis.
This means that ceramic-based gas sensors with larger grain size due to bigger pore size could
have a faster response than sensors fabricated of fine-dispersion metal oxide. As a result, in some
cases, annealing at temperatures between 400 and 700°C might result in decreased response times. The
linear dependence between average pore size and crystallite size (see Figure 2.39) reveals that the mean
coordinate of the pore is constant as the sintering temperature increases from 400 to 700°C (De Souza
Brito et al. 1995). Established regularities provide an opportunity for understanding many effects which
are discussed in various papers. For example, the growth of porosity of the SnO2 gas-sensing layer with
increased grain size is a good explanation for both the increase of NO2 response in films with larger-
sized crystallites (Rumyantseva et al. 2005) and the increase of response to CO for SnO2 sensors after
high-temperature calcination (Jin et al. 1998).
Note, however, that with an increase in porosity, the influence of film thickness on both the mag-
nitude of sensor response and the response time is appreciably weakened (see Figures 2.40 and 2.41). In
Table 2.19. Characteristics of SnO2 samples prepared by different routes
ROUTE Tan, °C t (XRD), nm Ssurf, m2 g−1 t (BET), nm tBET /t EXD S (NO2)

g 300 4 122 7 1.8 15
g 500 9 35 25 2.8 60
g 700 22 9 96 4.4 110
g 1000 35 — — — 130
k 300 4 175 5 1.3 168
k 500 9 65 13 1.4 190
k 700 17 28 31 1.8 77
k 1000 26 — — — —
h 300 4 135 6 1.5 14
h 500 11 26 33 3.0 25
h 700 34 12 72 2.1 48
h 1000 35 — — — —
s 300 3 180 5 1.7 53
s 500 11 69 13 1.2 41
s 700 22 14 62 2.8 130
s 1000 36 — — — —
SnO2 powders elaborated by wet chemical synthesis using (g) conventional hydrolysis of
SnCl4, (k) eryosol technique, and (h, s) specific hydrolysis in various solutions. Films calcined
at 500°C for 6 h. Sensor response estimated in dry atmosphere as S ( NO2 ) = (R – R O )/R O.
Source: Data from Rumyantseva et al. 2005. Reprinted with permission from Korotcenkov
Figure 2.38. Time constants of sensor response to CO versus In2O3 film refraction index: (1, 2) Toper
= 370°C; (3) Toper = 270°C; (1) d ≈ 40–60 nm; (2, 3) d ≈ 400 nm; (A) In2O3 films were deposited from
0.2 M InCl3–water solution; (B) from 1.0 M InCl3–water solution. (Reprinted with permission from
Korotcenkov 2004a. Copyright 2004 Elsevier.)
SnO2
Figure 2.39. Average pore size and surface area as a function of mean SnO2 crystallite size. The
structural evolution during sintering of compacted SnO2 sol-gel powders was investigated using nitro-
gen adsorption isotherm analysis. Pressed cylindrical pellets had diameter of 8 mm and height ~2.5
mm. (Data from De Souza Brito et al. 1995. Reprinted with permission from Korotcenkov 2008.
Figure 2.40. Dependences of response time during H2 detection (Toper = 370°C) on the thickness of
In2O3 films deposited from (2) high- and (1) low-concentration InCl3-sprayed solutions (Tpyr = 475°C):
(1), 0.2 M; (2) 1.0 M InCl3 solution. (Reprinted with permission from Korotcenkov 2007a. Copyright
2007 Elsevier.)
particular, Korotcenkov and co-workers (Korotcenkov 2005; Korotcenkov et al. 2001, 2007), studying
the influence of the modes of deposition on the structural and gas-sensing properties of SnO2 and
In2O3 films deposited by spray pyrolysis, found that for dense films an increase of film thickness in
the range 30–200 nm was accompanied by a decrease in sensitivity to ozone by almost two orders of
magnitude; for porous films, the same change of thickness did not affect gas sensitivity.
Figure 2.41. Influence of film porosity on sensor response dependence from SnO2 film thickness: (1)
porous films (Tpyr = 460°C); (2) dense film (Tpyr = 520°C). (Reprinted with permission from Korotcenkov
2007c. Copyright 2007 Elsevier.)
In addition to the temperature of deposition or synthesis, results presented by Matko et al. (2002)
showed that metal oxide doping may also affect the porosity of gas-sensing materials. For example,
SnO2:Pt films in the range of 0–0.6% Pt concentration had a maximum porous fraction. However, with
a doping concentration higher than 2%, no significant number of pores could be detected.
As shown by Shimuzu et al. (1998), control of porosity provides a means to influence selectiv-
ity. Shimuzu et al. (1998) proposed that the limited diffusion of O2 into the interior region of a thick
film should lead to a lowered oxygen partial pressure inside the gas-sensing matrix. This then leads to
decreased coverage of chemisorbed oxygen on the metal oxide surface. This should result in increased
relative sensitivity to H2. The mechanism of this effect is shown in Figure 2.42.
Experiments described by Shimuzu et al. (1998) with thick SnO2 films (d > 100 μm) covered by
a dense SiO2 layer confirmed this assumption. Thus, porosity control can be one of the methods of
influencing gas-sensor selectivity to H2 and other gases for which H2 is one of the products of dissocia-
tion. Gas molecules with larger diameters than H2 and O2 will not be able to penetrate into the interior
region of a metal oxide film, so the gas sensors will not be sensitive to these gases. Results reported by
Hyoda et al. (2002) (see Figure 2.43) confirmed that this effect requires that the pores not exceed a size
of 5–10 nm. Only in this case can good selectivity to H2 be obtained.
However, this method can be used only with thick and ceramic gas sensors, where the thickness
of the gas-sensing matrix exceeds 10–100 μm. Further, the decrease of pore size will often be accom-

O2
H2, CH4
S(H2)~S(CH4)

O2
O2
H2, CH4 S(H2)>>S(CH4)
H2
CH4 Figure 2.42. Schematic illustration of gas distribu-
tion in dense and porous gas-sensing matrix for
gases with different size (D) of molecules. (Reprinted
with permission from Korotcenkov 2008. Copyright
D(CH4)>D(O2)>D(H2) 2008 Elsevier.)

d ~ 85-125 m
H2 Toper=350 oC
CO
Figure 2.43. Influence of pore size on response of ceramic-type SnO2-based gas sensors to H2 and
CO. Tested sensors were fabricated using thick-film technology. Paste of the SnO2 powders was
calcined at 600°C for 5 h and printed on alumina substrate, on which interdigitated Pt electrodes
had been formed. (Data from Hyoda et al. 2002. Reprinted with permission from Korotcenkov 2008.
panied by an increase in response and recovery times, i.e., by a degradation of other important sensor
performance parameters. For example, the results presented by Hyoda et al. (2002) indicate that for
dense SnO2 ceramics with pore diameters ~3–5 nm, even at Toper ≈ 350–450°C, the response and re-
covery times exceed 10 min. Response kinetics were especially effective for CO detection as well as the
recovery process for H2. This should be expected, because the diffusion contribution to the total mo-
lecular transport through the micropore material depends on the pore size and the length of the path
along which the molecule travels. The same correlation has been established by Ahmad et al. (2003).
Sensors with minimal active surface area had maximum recovery and response times (see Table 2.20).
The crystallite size in all sensors was the same.
Korotcenkov et al. (2004a, 2004b) has shown that porosity effects can be used only for thick films.
For example, for In2O3 films with thickness less than 70–80 nm, no differences in response or recovery
times were observed between the “porous” and “dense” films (see Figure 2.40). This means that metal
oxide films with thickness d < 60–80 nm possess high gas permeability, and such structural parameters
as porosity no longer affect the kinetics of gas sensing for thin In2O3-based gas sensors.
The relationship between material porosity and sensor sensitivity to humidity has not attracted
designers’ attention. However, it is known that the smaller the pore size, the more probable that water
condensation is in it. Numerous capacitive-type humidity sensors have been designed based on this
effect. Therefore, one can suppose that during the sensor’s storage or break-in period, and when the
temperature of the active zone does not exceed room temperature, water condensation in small pores
is possible. Long-time contact of water with an oxide’s surface could substantially modify the oxide’s
surface properties, and thus change a sensor’s performance parameters.
Table 2.20. Influence of engineering procedure on structural parameters of SnO2-

based sensors and their time constants of response to reducing gas
D (XRD), Ssurf, D (BET), RESPONSE TIME RECOVERY TIME

SENSOR nm m2 g−1 nm (350°C), s (350°C), s
Sn-CM 30.15 122.44 7.0 10–20 50–60
Sn-DC 10.06 37.3 22.97 20–30 70–80
Sn-HC 12.33 27.7 30.94 50–55 150–160
Sn-HM 10.77 33.4 26.22 70–75 190–200
Source: Data from Ahmad et al. 2003. Reprinted with permission from Korotcenkov
Material with extremely small pores shows a decreased resistance to poisoning. This is another im-
portant problem of gas-sensing materials that is related directly to the need to provide maximum tem-
poral stability of gas-sensor parameters. This problem was discussed earlier for pellistors. Simulations
carried out by Gentry and Walsh (1982) for flammable gas-sensing elements showed that the poison
resistance of the porous sensors decreased with an increase of the parameter ho ~ (1/rDp )1/2, where r is
the pore radius and Dp is the pore diffusion coefficient.
Accumulation of solid reaction products at or on the sensor surface may also block the active sur-
face sites. With small pores, this process will be accompanied by a sharp decrease of active surface area
and a decrease of sensor response. Sulfides and graphite products (Brinzari et al. 2000b) are particularly
troublesome for metal oxides
6.7. AGGLOMERATION
The agglomeration and aggregation of matter are ubiquitous phenomena that may be expected to occur
almost everywhere in nature (Kaye et al. 1989). Their appearance ranges from metallic polycrystals to
colloidal aggregates and clusters, as well as the lipid–protein viscoelastic matrices that are often called
biomembranes. In most cases the metal oxide films are agglomerated as well (Barsan et al. 1999). One
example of a SnO2 agglomerated film is shown in Figure 2.44. However, the role of agglomeration in
gas-sensing effects has not been studied in detail yet.
One should note that an agglomeration problem exists for thin-film as well as ceramic and thick-
film sensors. It has been established that in many cases synthesized powders agglomerate into big par-
ticles. SEM images obtained by Chabanis et al. (2003) indicate that, as a rule, the agglomerates formed
during ceramic synthesis are more porous and are larger in size than those formed by the methods of
thin-film technology. The agglomerate size for SnO2 prepared using the sol-gel method exceeded 5 μm
(Chabanis et al. 2003). For films deposited by spray pyrolysis, the size of agglomerate did not exceed
200–500 nm (Brinzari et al. 2002; Korotcenkov et al. 2002, 2004a, 2005a).
Such a difference in the properties of ceramics and films is natural and follows from the various
conditions of synthesis. Agglomerates formed in ceramic-type and thin-film-type metal oxides corre-
spond to two different types of aggregation. In ceramic materials, aggregation is diffusion-controlled,

in which the attachment of particles and clusters occurs instantaneously at first contact. In thin films,
aggregation is controlled by chemical reaction (Derjaguin 1989; Gadomski et al. 2005). In the latter
case, the “first touch” sticking principle does not hold, and the microstructural evolution has to fol-
low certain attachment–detachment sequences of events. Finally, the incorporation takes place in the
form of a particle or some cluster of particles with different types of microstructures (Anderson and
Lekkerkerker 2002). Chemical reaction–controlled aggregations are compact compared to their diffu-
sive counterparts, which look more irregular, viz., fractal, meaning that their arms or branches expand
more visibly. Therefore, in general, in diffusion-controlled reactions the aggregations are not dense
(Jullien and Botet 1987; Schmelzer et al. 1999).
The gas-sensing matrix of agglomerated polycrystalline films can shown schematically as an equiv-
alent circuit (see Figure 2.44). We consider the gas-sensing matrix as a three-dimensional network of
crystallites (grains) forming a gas-sensing film. Here R(a–a) represents the resistance of interagglomerate
contacts, Rc is the resistance of intercrystallite (intergrain) contacts, Rb is the bulk resistance of the crys-
tallites (grains), and Ragl is the resistance of the agglomerate.
From this scheme one can see that in a gas-sensing matrix there are four gas-sensitive elements
which can affect the sensor response:
1. Intergrain (intercrystallite) contacts

2. Interagglomerate contacts
3. Agglomerates
4. Grains

Figure 2.44. SEM image of SnO2 agglomerated film deposited by spray pyrolysis and diagrams
illustrating peculiarities of electroconductivity and gas-sensing reactions in agglomerated films.
Here the agglomerate resistance is an integral resistance, representing a three-dimensional grain net-
work. Thus, if there are alterations in the porosity of either agglomerates or the gas-sensing matrix,
the role of the various gas-sensing elements in sensor response changes. For example, a decrease of
grain size increases the role of grains in gas-sensing effects. A low gas permeability of agglomerates
promotes an increase in the influence of interagglomerate contact on sensor response. These obser-
vations are in agreement with simulations (Williams and Pratt 2000) affirming that agglomeration of
small crystallites into large aggregations is a key phenomenon and can result in great variations of the
apparent response characteristics.
Research has shown that both the sensor response and the kinetics depend on the properties of
agglomerates. Larger and denser agglomerates exhibit slower response and recovery times. It is impor-
tant to note that this effect depends on the nature of the detected gas, particularly on the reactivity
and diffusion coefficient of this gas in the oxide matrix. A high level of agglomeration also promotes
the formation of capsulated zones, i.e., zones with closed porosity. Properties of metal oxides in these
zones do not depend on a change in the surrounding gas (McAleer et al. 1987).
Based on an analysis of SnO2 sensors fabricated using the successive ionic layer deposition (SILD)
method, Korotcenkov et al. (2003b) concluded that the small size of crystallites (grains) was an es-
sential, but not sufficient, condition to achieve both maximum gas sensitivity and fast response. First,
the gas-sensing matrix should be highly porous and the size of the agglomerates must be minimal.
Moreover, a paradoxical situation appears for strongly agglomerated structures. In such structures the
small size of grains (crystallites) is not an advantage. Agglomerates from smaller grains are more densely
packed—i.e., they have smaller gas penetrability. Probably this was the case in the work of Hyoda et al.
(2002), where, in spite of minimum grain size (~2 nm), the gas response of SnO2 sensors to H2 was
unexpectedly low. Therefore, in accordance with the above concept, such structures can have poorer
gas-sensing characteristics than films with larger crystallites. A similar conclusion was drawn by De
Souza Brito et al. (1995). They established that films consisting of larger crystallites were more porous
than films consisting of small crystallites. Therefore, the former had higher sensor response. In this
context, it becomes clear that to perform a comparative analysis of gas-sensing characteristics of metal
oxide–based sensors, it is necessary first to consider the size and the porosity of agglomeration, and
only after that the crystallite size.
Usually, dense and strongly agglomerated ceramics have a small active surface area (Tan et al. 2000).
As was previously shown, this has an important negative impact on the approach for optimization of gas-
sensing characteristics. Numerous studies have established that mechanical milling is one effective method
for resolving this problem (Koch 1997; Dieguez et al. 2000; Tan et al. 2000, 2004; Abe et al. 2005). Abe et
al. (2005) showed that mechanical milling increased the sensitivity to CO more than four times (see Figure
2.45). Dieguez et al. (2000) established that milling allows the possibility to fabricate SnO2 sensors with
very good sensitivity to NO2 and a negligible cross-sensitivity to CO, along with high long-term stability.
It is impossible to fully explain this effect by a decrease of grain size, because the grain size was
decreased only by 10% as a result of this procedure. At the same time, the influence of mechanical
milling on agglomerate size is considerable. The consequences are clearly shown in Figure 2.46 (Tan et
al. 2000). Taking into account estimates made from transmission electron microscopy (TEM) micro-
graphs, mechanical milling decreased the agglomerate size more than 20-fold. After mechanical milling,
the sensitivity to ethanol increased more than 20-fold as well.
Figure 2.45. Influence of milling time on (1) SnO2 grain size and (2) xSnO2-(1-x)αFe2O3 sensor re-
sponse to ethanol (1000 ppm). Sensors were fabricated onto the Al2O3 substrate with interdigital Au
electrodes using thick-film technology. (Data from Tan et al. 2000. Reprinted with permission from
Korotcenkov 2008. Copyright 2008 Elsevier.)
However, it is necessary to note that the development of nanocrystalline microstructures during

mechanical milling takes place in three stages (Hadjipanayis and Siegel 1994):
• Stage 1. Deformation is localized in shear planes containing a high dislocation density.

• Stage 2. Dislocation annihilation/recombination/rearrangement forms a cell/subgrain structure
with nanoscale dimensions; further milling extends this structure throughout the sample.
• Stage 3. The orientation of the grains becomes random.

Figure 2.46. TEM micrographs for SnO2 powders after (a) 2 h and (b) 129 h milling using high-
energy milling in a Fritsh Pulverisette 5 planetary ball milling system. (After Tan et al. 2000. Reprinted
with permission from Tan et al. 2000. Copyright 2000 Elsevier.)
Thus, as a result of mechanical milling, additional mechanical tensions appear in the material. The ap-
pearance of mechanical tensions and the change in the microstructure may be an important additional
factor that influences gas-sensing characteristics. Mechanical tensions should be considered when de-
veloping procedures for sensor fabrication. Only thermal treatment at Т ≈ 600–800°C seems to relieve
these tensions (Cirera et al. 2000). Thus, an additional thermal treatment after mechanical milling may
be required.
Based on these facts, one can conclude that thin-film sensors with minimum agglomeration will
always have a maximum rate of response. As Тpyr increases, the agglomeration of the film decreases.
However, at the same time, considerable growth of both grain size and film density takes place, which
is undesirable for maximum sensor sensitivity and minimum response time. Therefore, we have to seek
a compromise between grain size and the possibility of agglomerate formation when preparing films.
Many studies have shown that the easiest approach to this problem is to use ultrathin films with thick-
nesses of less than 40–60 nm. In such films, neither the agglomerate size nor the film thickness reaches
the critical value at which gas penetrability starts to influence the sensor response kinetics.
However, as agglomerates become denser and bigger, the size of the interagglomerate pores in-
creases. Under certain conditions, qualitative changes may take place in the gas-sensing matrix which
may affect the metal oxide structure and design parameters for achieving maximum sensitivity. In some
cases, mainly for oxidizing gases, the sensor response may increase with growth of film thickness and
agglomerate size. The clearest example of such a situation was presented by Rumyantseva et al. (2005).
This seems to be a contradiction to what was said before, but the idea becomes clear if one realizes that
gas sensitivity may be controlled, not by intercrystallite contacts, but by interagglomerate contacts. The
growth of agglomerates increases the size of interagglomerate pores,which facilitates access of active
gas into the inner volume of the metal oxide matrix and thus makes it more sensitive. This mechanism
was suggested by Korotcenkov et al. (2003) to explain the high sensitivity to ozone of SnO2 films that
were synthesized by the SILD method and are characterized by a high level of agglomeration.
7. OUTLOOK
In this chapter we have shown that no material satisfies all the desired properties for a sensing material.
Further, every new application creates new requirements in terms of properties. advances its own whole
requirements to sensors. The literature is therefore replete with examples of materials that have been
tested for their applicability to the design of chemical sensors. This process is still ongoing, involving re-
search on entire new families of compounds. Fullerenes, carbon and metal oxide nanotubes, nanowires,
nanocomposites, etc., are presently being explored (Baraton et al. 1997; Chao and Shih 1998; Varghese
et al. 2001, 2003; Gas’kov et al. 2001; Penza et al. 2001; Baena et al. 2002; Cantalini et al. 2003; Valentini
et al. 2003, 2004). We don’t know yet where eventually a qualitative leap to a world of nano-sized struc-
tures may lead. However, early results are promising.
8. ACKNOWLEDGMENTS
The author thanks the Korean BK21 Program for support of his research.
REFERENCES
Abe S., Choi U.S., Shimanoe K., and Yamazoe N. (2005) Influences of ball-milling time on gas-sensing properties
of Co3O4–SnO2 composites. Sens. Actuators B 107, 516–522.
Adhikari B. and Majumdar S. (2004) Polymers in sensor applications. Prog. Polym. Sci. 29, 699–766.
Ahmad A., Walsh J., and Wheat T.A. (2003) Effect of processing on the properties of tin oxide-based thick-film
gas sensors. Sens. Actuators B 93, 538–545.
Alfredsson M., Richard C., and Catlow A. (2004) Predicting the metal growth mode and wetting of noble metals
support on c-ZrO2. Surf. Sci. 561, 43–56.
Anderson V.J. and Lekkerkerker H.N.W. (2002) Insights into phase transition kinetics from colloid science. Nature
416, 811–816.
Ando M., Chabicovsky R., and Haruta M. (2001) Optical hydrogen sensitivity of noble metal-tungsten oxide com-
posite films prepared by sputtering deposition. Sens. Actuators B 76, 13–17.
Ando M., Kobayashi T., Iijima S., and Haruta M. (1997) High impact applications, properties and synthesis of
exciting new materials. J. Mater. Chem. 7, 1779–1783.
Arbiol J., Peiro F., Cornet A., Morante J.R., Perez-Omil J.A., and Calvino J.J. (2002) Computer image HRTEM
simulation of catalytic nanoclusters on semiconductor gas sensor materials supports. Mater. Sci. Eng. B 91–92,
534–536.
Arshak K., Twomey K., and Heffernan D. (2002) Development of a microcontroller-based humidity sensing sys-
tem. Sensor Rev. 22(2), 150–156.
Badlani M. and Wachs I.E. (2001) Methanol: a “smart” chemical probe molecule. Catal. Lett. 75(3–4), 137.
Badwal S.P.S. (1992) Zirconia-based solid electrolytes: microstructure, stability and ionic conductivity. Solid State
Ionics 52, 23–32.
Badwal S.P.S., Ciacchi F.T., and Milosevic D. (2000) Scandia-zirconia electrolytes for intermediate temperature solid
oxide fuel cell operation. Solid State Ionics 136–137, 91–99.
Baena J.R., Gallego M., and Valcarcel M. (2002) Fullerenes in the analytical sciences. Trends Anal. Chem. 21(3),
187–198.
Bamwenda G.R., Tsubota S., Nakamura T., and Haruta M. (1997) The influence of the preparation methods on the
catalytic activity of platinum and gold supported on TiO2 for CO oxidation. Catal. Lett. 44, 83.
Baraton M.I., Merhari L., Wang J., and Gonsalves K.E. (1997) Dispersion of metal oxide nanoparticles in con-
jugated polymers: investigation of the TiO2/PPV nanocomposite. In: MRS Symposium Proceedings, Vol. 501,
Pittsburgh, PA, pp. 59–64.
Barsan N. and Weimar U. (2001) Conduction model of metal oxide gas sensors. J. Electroceram. 7, 143–167.
Barsan N. and Weimar U. (2003) Understanding the fundamental principles of metal oxide based gas sensors: the
example of CO sensing with SnO2 sensors in the presence of humidity. J. Phys. Condensed Matter 15, R1–R27.
Barsan N., Heilig A., Kappler J., Weimar U., and Gopel W. (1999b) CO-water interaction with Pd-doped SnO2 gas
sensors: simultaneous monitoring of resistance and work function. In: Proceedings of the 13th European Conference on
Solid State Transducers, EUROSENSORS XIII, September 12–15, 1999, the Hague, The Netherlands,
pp. 367–369.
Barsan N., Schweizer-Berberich M., and ·Göpel W. (1999a) Fundamental and practical aspects in the design of
nanoscaled SnO2 gas sensors: a status report. Fresenius J. Anal. Chem. 365, 287–304.
Basu S., Saha M., Chatterjee S., Mistry K.Kr., Bandyopadhay S., and Sengupta K. (2001) Porous ceramic sensor for
measurement of gas moisture in the ppm range. Mater. Lett. 49, 29–33.
Batzill M. (2006b) Surface science studies of gas sensing materials: SnO2. Sensors 6, 1345–1366.
Batzill M., Bergermayer W., Tanaka I., and Diebold U. (2006a) Tuning the chemical functionality of a gas sensitive
material: water adsorption on SnO2(101). Surf. Sci. 600, L29–L32.
Batzill M., Katsiev K., Burst L.M., Diebold U., Chaka A.M., and Delley B. (2005) Gas-phase-dependent properties
of SnO2 (110), (100), and (101) single-crystal surfaces: structure, composition, and electronic properties. Phys.
Rev. B 72, 165414.
Bendahan M., Lauque P., Lambert-Mauriat C., Carchano H., and Seguin J.L. (2002) Sputtered thin films of CuBr for
ammonia microsensors: morphology, composition and ageing. Sens. Actuators B 84, 6–11.
Bhuiyan A.L. (1984) Some aspects of the thermal stability action of the structure in aliphatic polyamides and poly-
acrylamides. Polymer 25, 1699–1710.
Blesa M.A., Weisz A.D., Morando P.J., Salfity J.A., Magaz G.E., and Regazzoni A.E. (2000) The interaction of metal
oxide surfaces with complexing agents dissolved in water. Coord. Chem. Rev. 196, 31–63.
Bogue R. (2002b) Advanced automotive sensors. Sens. Rev. 22(2), 113–118.
Bogue R.W. (2002a) The role of materials in advanced sensor technology. Sens. Rev. 22(4), 289–299.
Bose A.C., Thangadurai P., and Ramasamy S. (2006) Grain size dependent electrical studies on nanocrystalline
SnO2. Mater. Chem. Phys. 95, 72–78.
Brinzari V., Korotcenkov G., and Golovanov V. (2001) Factors influencing the gas sensing characteristics of tin
dioxide films deposited by spray pyrolysis: understanding and possibilities for control. Thin Solid Films 391(1/2),
167–175.
Brinzari V., Korotcenkov G., and Matolin V. (2005) Synchrotron radiation photoemission study of indium oxide
surface prepared by pyrolysis method. Appl. Surf. Sci., 243(1–4), 335–344.
Brinzari V., Korotcenkov G., and Schwank J. (1999a) Optimization of thin film gas sensors for environmental mo-
nitoring through theoretical modeling. In: S. Buettgenbach (ed.), Chemical Microsensors and Applications II, Proc.
SPIE 3857, 186–197.
Brinzari V., Korotcenkov G., Schwank J., Lantto V., Saukko S., and Golovanov V. (2002) Morphological rank of
nano-scale tin dioxide films deposited by spray pyrolysis from SnCl4.5H2O water solution. Thin Solid Films 408,
51–58.
Brinzari V., Korotcenkov G., Veltruska K., Matolin V., Tsud N., and Schwank J. (2000b) XPS study of gas sensitive
SnO2 thin films. In: Proceeding of Semiconductor International Conference CAS’2000, 10–14 October 2000, Sinaia,
Romania, Vol. 1, pp. 127–130.
Brinzari V., Korotchenkov G., and Dmitriev S. (1999b) Simulation of thin film gas sensor kinetics. Sens. Actuators
B 61(1–3), 143–153.
Brinzari V., Korotchenkov G., and Dmitriev S. (2000a) Theoretical study of semiconductor thin film gas sensitivity:
attempt to consistent approach. J. Electron. Technol. 33, 225–235.
Brynn D.H. and Tseung C.C. (1979) The reduction of sulphur dioxide by carbon monoxide on La0.5Sr0.5CoO3
catalyst. J. Chem. Technol. Biotechnol. 29(12), 713.
Cabot A., Dieguez A., Romano-Rodriguez A., Morante J.R., and Barsan N. (2001) Influence of the catalytic intro-
duction procedure on the nano-SnO2 gas sensor performances. Where and how stay the catalytic atoms? Sens.
Cabot A., Vila A., and Morante J.R. (2002) Analysis of the catalytic activity and electrical characteristics of different
modified SnO2 layers for gas sensors. Sens. Actuators B 84, 2–20.
Calatayud M., Markovits A., Menetrey M., Mguig B., and Minot C. (2003) Adsorption on perfect and reduced sur-
faces of metal oxides. Catal. Today 85, 125–143.
Campbell C.T. (2004) The active site in nanoparticle gold catalysis. Science 306, 234–235.
Cantalini C., Valentini L., Lozzi L., Armentano I., Kenny J.M., and Santucci S. (2003) NO2 gas sensitivity of carbon
nanotubes obtained by plasma enhanced chemical vapor deposition. Sens. Actuators B 93, 333–337.
Castañeda L. (2007) Effects of palladium coatings on oxygen sensors of titanium dioxide thin films. Mater. Sci. Eng.
B 139, 149–154.
Catlow C.R.A. (ed.) (1997) Computer Modeling in Inorganic Crystallography. Academic Press, New York.
Chabanis G., Parkin I., and Williams D.E. (2003) A simple equivalent circuit model to represent microstructure
effects on the response of semiconducting oxide-based gas sensors. Meas. Sci. Technol. 14, 76–86.
Chan K.S., Ma J., Jaenicke S., and Chuan G.K. (1994) Catalytic carbon-monoxide oxidation over strontium, cerium
and copper-substituted lanthanum manganates and cobaltates. Appl. Catal. A 107, 201.
Chao Y.C. and Shih J.S. (1998) Adsorption study of organic molecules on fullerene with piezoelectric crystal detec-
tion system. Anal. Chim. Acta 374, 39–46.
Chen F., Zhao T., Fei Y.Y., Lu H.B., Cheng Z.H., Yang G.Z., and Zhu X.D. (2002) Surface segregation of bulk
oxygen on oxidation of epitaxially grown Nb-doped SrTiO3 on SrTiO3(001). Appl. Phys. Lett. 80, 2889.
Chena C.S., Kruidhof H., Bouwmeestera H.J.M., Verweij H., and Burggraaf A.J. (1996) Oxygen permeation
through oxygen ion oxide-noble metal dual phase composites. Solid State Ionics 86–88, 569–572.
Choudhary T.V., and Goodman D.W. (2005) Catalytically active gold: the role of cluster morphology. Appl. Catal.
A 291, 32–36.
Cirera A., Cornet A., Morante J.R., Olaizola S.M., Castano E., and Gracia J.R. (2000) Comparative structural study
between sputtered and liquid pyrolysis nanocrystalline SnO2. Mater. Sci. Eng. B 69–70, 406–410.
Clifford P.K. (1983) Microcomputational selectivity enhancement of semiconductor gas sensors. In: Proceeding of
International Meeting on Chemical Sensors, Fukuoka, Japan, 19–22 September 1983, pp. 153–158.
Colladet K., Nicolas M., Goris L., Lutsen L.L., and Vanderzande D. (2004) Low-band gap polymers for photovol-
taic applications. Thin Solid Films 451–452, 7–11.
Cox P.A. (1992) Transition Metal Oxides: An Introduction to Their Electronic Structure and Properties. Clarendon Press,
Oxford.
Dai Z.R., Gole J.L., Stout J.D., and Wang Z.L. (2002) Tin oxide nanowires, nanoribbons, and nanotubes. Phys. Chem.
B 106, 1274–1279.
Dai Z.R., Pan Z.W., and Wang Z.L. (2003) Novel nanostructures of functional oxides synthesized by thermal eva-
poration. Adv. Funct. Mater. 13, 9–24.
Dakin J. and Culshaw B. (eds.) (1988) Optical Fiber Sensors: Principles and Components, Vol.1. Artech House, Boston.
De Fresart E., Darville J., and Gilles J.M. (1982) Influence of the surface reconstruction of the work function and
surface conductance of (110) SnO2. Appl Surf. Sci. 11/12, 637–651.
De Souza Brito G.E., Santilli C.V., and Pulcenelli S.H. (1995) Evolution of the fractal structural during sintering of
SnO2 compacted sol-gel powders. Colloids Surf. A 97, 217–225.
Dean J.A. (ed.) (1972) Lange’s Handbook of Chemistry, 13th ed. McGraw-Hill, New York.
Deb B., Desai S., Sumanasekera G.U., and Sunkara M.K. (2007) Gas sensing behaviour of mat-like networked
tungsten oxide nanowire thin films. Nanotechnology 18, 285501.
Derjaguin B.V. (1989) Theory of Stability of Colloids and Thin Films. Consultants Bureau, New York, London.
Dhandapani B. and Oyama S.T. (1997) Gas phase ozone decomposition catalysis. Appl. Catal. B 11, 129–166.
Dieguez A., Romano-Rodriguez A., Vila A., and Morante J.R. (2001) The complete Raman spectrum of nanometric
SnO2 particles. Appl. Phys. 90, 1550–1557.
Dieguez A., Romano-Rodrýguez A., Alay J.L., Morante J.R., Barsan N., Kappler J., Weimar U., and Gopel W. (2000)
Parameter optimisation in SnO2 gas sensors for NO2 detection with low cross-sensitivity to CO: sol–gel prepa-
ration, film preparation, powder calcination, doping and grinding. Sens. Actuators B 65, 166–168.
Dimitrov V. and Komatsu T. (2002) Classification of simple oxides: a polarizability approach. J. Solid State Chem.
163, 100–112.
Dmitriev S., Lilach Y., Button B., Moskovits M., and Kolmakov A. (2007) Nanoengineered chemiresistors: the inter-
play between electron transport and chemisorption properties of morphologically encoded SnO2 nanowires.
Nanotechnology 18, 055707.
Doll T. and Eisele I. (1998) Gas detection with work function sensors. In: Proceeding of SPIE Conference on Chemical
Microsensors and Applications, Boston, November 1998, vol. 3539, pp. 96–105.
Doll T., Lechner J., Eisele I., Schierbaum K.D., and Gopel W. (1996) Ozone detection in the ppb range with work
function sensors operating at room temperature. Sens. Actuators B 34, 506–510.
Dufour L.C. and Nowotny J. (eds.) (1988) Surface and Near Surface Chemistry of Oxide Materials. Elsevier, Amsterdam.
Eglitis R.I., Heifets E., Kotomin E.A., Majer J., and Borstel G. (2003) First-principles calculations of perovoskite
thin film. Mater. Sci. Semin. Proc. 5, 129–134.
Eisele I., Doll T., and Burgmair M. (2001) Low power gas detection with FET sensors. Sens. Actuators B 78, 19–25.
El-Azab A., Gan S., and Liang Y. (2002) Binding and diffusion of Pt nanoclusters on anatase TiO2(001)-(1x4)
surface. Surf. Sci. 506, 93–104.
Emiroglu S., Barsan N., Weimar U., and Hoffmann V. (2001) In-situ diffuse reflectance infrared spectroscopy study
of CO adsorption on SnO2. Thin Solid Films 391, 176–185.
Epling W.S., Peden C.H.F., Henderson M., and Diebolod U. (1998) Evidence for oxygen adatoms on TiO2 (110)
resulting from O2 dissociation at vacancy sites. Surf. Sci. 412/413, 333–343.
Esch H., Huyberechts G., Mertens R., Maes G., Manca J., DeCeuninck W., and De Schepper L. (2000) The stability
of Pt heater and temperature sensing elements for silicon integrated tin oxide gas sensors. Sens. Actuators B 65,
190–192.
Fang Q., Chetwynd D.G., and Gardner J.W. (2002) Conducting polymer films by UV-photo processing. Sens.
Fleischer M. and Meixner H. (1997) Fast gas sensors based on metal oxides, which are stable at high temperatures.
Fleischer M. and Meixner H. (1998) Selectivity in high-temperature operated semiconductor gas-sensors. Sens.
Freund H.J., Kuhlenbeck H., and Staemmler V. (1996) Oxide surfaces. Rep. Prog. Phys. 59, 283–347.
Freund H.J., Kuhlenbeck H., Libuda J., Rupprechter G., Baumer M., and Hamann H. (2001) Bridging the pressure
and materials gaps between catalysis and surface science: clean and modified oxide surfaces. Topics Catal. 15
(2–4), 201–209.
Freund J. and Umbach E. (1993) (eds.) Adsorption on Ordered Surfaces of Ionic Solids and Thin Films. Springer Series in
Surface Sciences, vol. 33. Springer-Verlag, Heidelberg.
Gadomski A., Rub J.M., Luczka J., and Ausloos M. (2005) On temperature-and space-dimension dependent matter
agglomerations in a mature growing stage. Chem. Phys. 310, 153–161.
Galdikas A., Kaciulis S., Mattogno G., Mironas A., Senuliene D., and Setkus A. (1998) Stability and oxidation of the
sandwich type gas sensors based on thin metal films. Sens. Actuators B 48, 376–382.
Galdikas A., Mironas A., Senuliene D., and Setkus A. (1996) CO gas induced resistance switching in SnO2:ultra-
thin-Pt sandwich structure. Sens. Actuators B 32, 87–92.
Gardner J.M. and Bartlett P.N. (1999) Electronic Noses: Principles and Applications. Oxford University Press, Oxford.
Gardner J.M. and Bartlett P.N. (eds.) (1992) Sensors and Sensory Systems for an Electronic Nose. Kluwer, Dordrecht, The
Netherlands.
Gas’kov A.M. and Rumyantseva M.N. (2001) Nature of gas sensitivity in nanocrystalline metal oxides. Russ. J. Appl.
Chem. 74(3), 440–444.
Geistlinger H. (1993) Electron theory of thin film gas sensors. Sens. Actuators B 17, 47–60.
Geistlinger H. (1994) Accumulation layer model for Ga2O3 thin-film gas sensors based on the Volkenstein theory
of catalysis. Sens. Actuators B 18–19, 125–131.
Gellinger P.J. and Bouwmeester H.J.M. (2000) Solid state aspects of oxidation catalysis. Catal. Today 58, 1–53.
Gellinger P.J. and Bouwmeester H.J.M. (eds.) (1997) The CRC Handbook of Solid State Electrochemistry. CRC Press,
Boca Raton, FL.
Gentry S.J. and Walsh P.T. (1982) The theory of poisoning of catalytic flammable gas-sensing elements. In: Mosely
P.T. and Tofied B.C. (eds.), Solid State Gas Sensors. Adam Hilger. Bristol and Philadelphia, pp. 32–50.
Gillet M., Aguir K., Bendahan M., and Mennini P. (2005) Grain size effect in sputtered tungsten trioxide thin films
on the sensitivity to ozone. Thin Solid Films 484, 358–363.
Golovanov V., Maki-Jaskari M.A, Rantala T.T., Korotcenkov G., Brinzari V., Cornet A., and Morante J. (2005a)
Experimental and theoretical studies of the indium oxide-based gas sensors deposited by spray pyrolysis. Sens.
Actuators 106, 563–571.
Golovanov V., Pekna T., Kiv A., Litovchenko V., Korotcenkov G., Brinzari V., Cornet A., and Morante J. (2005b)
The influence of structural factors on sensitivity of SnO2-based gas sensors to CO in humid atmosphere. Ukr.
Phys. J. 50, 374–380.
Gopel W. (1996) Ultimate limits in the miniaturization of chemical sensors. Sens. Actuators A 56, 83–102.
Gopel W. and Reinhard G. (1998) Metal oxide sensors: new devices through tailoring interfaces on the atomic scale.
In: Sensor Update, Part 3. Wiley-VCH, Weinheim, 47–121.
Gopel W., Hesse J., and Zemel J.N. (eds.) (1989–1995) Sensors, A Comprehensive Survey, Vols. 1–9. Wiley-VCH,
Weinheim.
Gordon M.J., Gaur S., Kelkar S., and Baldwin R.M. (1996) Low temperature incineration of mixed wastes using
bulk metal oxide catalysts. Catal. Today 28, 305–317.
Gurlo A., Ivanovskaya M., Barsan N., Schweizer-Berberich M., Weimar U., Gopel W., and Dieguez A. (1997) Grain
size control in nanocrystalline In2O3 semiconductor sensors. Sens. Actuators B 44, 327–333.
Hadjipanayis G.C. and Siegel R.W. (eds.) (1994) Nanophase Materials. Kluwer, Dordrecht, The Netherlands.
Hagen J.E. and Kvaal K. (1998) Electronic nose and artificial neural network. Meat Sci. 49(1), S273–S286.
Hammer B. and Norskov J.K. (1995) Why gold is the noblest of all the metals. Nature 376, 238–240.
Hamnett A. and Goodenough J.B. (1984) Binary transition-metal oxides. In: Madelung O. (ed.), Semiconductors:
Physics of Nontetrahedrally Bonded Binary Compounds III, Vol. 17, part g, Landolt-Börnstein New Series, Group III,
Springer-Verlag, Berlin.
Han P.G., Wong H., and Poon M.C. (2001) Sensitivity and stability of porous polycrystalline silicon gas sensor.
Colloids Surf. A 179, 171–175.
Hanys P., Janecek P., Matolin V., Korotcenkov G., and Nehasil V. (2006) XPS and TPD study of Rh/SnO2 system-
reversible process of substrate oxidation and reduction. Surf. Sci. 600, 4233–4238.
Harsanyi G. (1994) Polymer Films in Sensor Applications. Technomic, Lancaster.
Harsanyi G. (2000) Polymer films in sensor applications: a review of present uses and future possibilities. Sensor
Rev. 20(2), 98–105.
Haruta M. (1997) Size- and support dependency in the catalysis of gold. Catal. Today 36, 153.
Heeg J., Kramer C., Wolter M., Michaelis S., Plieth W., and Fisher W.J. (2001) Polythiophene-O3 surface reactions
studied by XPS. Appl. Surf. Sci. 180, 36–41.
Henrich V.E. and Cox P.A. (1994) The Surface Science of Metal Oxides. Cambridge University Press, Cambridge.
Heo Y.W., Norton D.P., Tien L.C., Kwon Y., Kang B.S., Ren F., Pearton S.J., and LaRoche J.R. (2004) ZnO nanowire
growth and devices. Mater. Sci. Eng. R 47, 1–47.
Holzinger M., Maier J., and Sitte W. (1996) Fast CO2-selective potentiometric sensor with open reference electrode.
Solid State Ionics 86–88, 1055–1062.
Horiuchi T., Hidaka H., Fukui T., Kubo Y., Horio M., Suzuki K., and Mori T. (1998) Effect of added basic metal
oxides on CO2 adsorption on alumina at elevated temperatures. Appl. Catal. A 167, 195–202.
Houser E.J., Mlsna T.E., Nguyen V.K., Chung R., Mowery E.L., and McGill R.A. (2001) Rational materials design
of sorbent coatings for explosives: applications with chemical sensors. Talanta 54, 469–485.
Huang X.J. and Choi Y.K. (2007) Chemical sensors based on nanostructured materials. Sens. Actuators B 122,
659–671.
Hyoda T., Nishida N., Shimizu Y., and Egashira M. (2002) Preparation and gas sensing properties of thermally
stable mesoporous SnO2. Sens. Actuators B 83, 209–215.
Hyodo T., Abe S., Shimuzu Y., and Egashira M. (2003) Gas sensing properties of ordered mesoporous SnO2 and
effects of coatings thereof. Sens. Actuators B 93, 590–600.
Ihokura K. and Watson J. (1994) The Stannic Oxide Gas Sensor, Principle and Applications. CRC Press, Boca Raton, FL.
Ishihara T. and Matsubara S. (1998) Capacitive type gas sensors. J. Electroceram. 2, 215–228.
Ishizaki K., Komarneni S., and Nanko M. (1998) Porous Materials: Process Technology and Applications. Kluwer,
Dordrecht, The Netherlands, p. 202.
Ivanovskaya M., Kotsikau D., Faglia G., and Nelli P. (2003) Influence of chemical composition and structural fac-
tors of Fe2O3/In2O3 sensors on their selectivity and sensitivity to ethanol. Sens. Actuators B 96, 498–503.
Jamnik J., Kamp B., Merkle R., and Maier J. (2002) Space charge influenced oxygen incorporation in oxides: in how
far does it contribute to the drift of Taguchi sensors? Solid State Ionics 150, 157–166.
Jianping L., Yue W., Xiaoguang., Qing M., Li W., and Jinghong H. (2000) H2S sensing properties of the SnO2-based
thin films. Sens. Actuators B 65, 111–113.
Jin C.J., Yamazaki T., Shirai Y., Yoshizawa T., Kikuta T., Nakatani N., and Takeda H. (2005) Dependence of NO2
gas sensitivity of WO3 sputtered films on film density. Thin Solid Films 474, 255–260.
Jin Z., Zhou H.J., Jin Z.L., Savinell R.F., and Lin C.C. (1998) Application of nano-crystalline porous tin oxide film
for CO sensing. Sens. Actuators B 52, 188–194.
Johnson C.L., Schwank J.W., and Wise K.D. (1994) Integrated ultra-thin-film gas sensors. Sens. Actuators B 20,
55–62.
Johnson D.A. (1982) Some Thermodynamic Aspects of Inorganic Chemistry. Cambridge University Press, Cambridge.
Jolivet J.P. (2000) Metal Oxide Chemistry and Synthesis: From Solution to Solid State. Wiley, Chichester, UK.
Jullien R. and Botet R. (1987) Aggregation and Fractal Aggregates. World Scientific, Singapore.
Kabbabi A., Gloaguen F., Andolfatto F., and Durand R. (1994) Particle size effect for oxygen reduction and metha-
nol oxidation on Pt/C inside a proton exchange membrane. J. Electroanal. Chem. 373, 251–254.
Kanazawa E., Sakai G., Shimanoe K., Kanmura Y., Teraoka Y., Miura N., and Yamazoe N. (2001) Metal oxide semi-
conductor N2O sensors for medical use. Sens. Actuators B 77, 72–77.
Kanda K. and Maekawa T. (2005) Development of a WO3 thick-film-based sensors for the detection of VOC. Sens.
Kaur M., Gupta SA.K., Betty C.A., Saxena V., Katti V.R., Gadkari S.C., and Yakhmi J.V. (2005) Detection of
reducing gases by SnO2 thin films: an impedance spectroscopy study. Sens. Actuators B 107, 360–365.
Kawabe T., Tabata K., Suzuki E., Ichikawa Y., and Nagasawa Y. (2001) Morphological effects of SnO2 thin film on
the selective oxidation of methane. Catal. Today 71, 21–29.
Kaye B.H. (1989) A Random Walk Through Fractal Dimensions. VCH, Weinheim.
Kindery W.D., Bowen H.K., and Uhlmann D.R. (1976) Introduction to Ceramics. 2nd ed., Wiley, New York.
Kocemba I., Szafran S., Rynkowski J., and Paryjczak T. (2001) The properties of strongly pressed tin oxide-based
gas sensors. Sens. Actuators B 79, 28–32.
Koch C.C. (1997) Synthesis of nanostructured materials by mechanical milling: problems and opportunities.
NanoStruct. Mater. 9, 13–22.
Kohl D. (1990) The role of noble metals in the chemistry of solid-state gas sensors. Sens. Actuators B 1, 158–165.
Kohl D. (2001) Function and application of gas sensors. J. Phys. D 34, R125–R149.
Kong X.H., Sun X.M., and Li Y.D. (2003) Synthesis of ZnO nanobelts by carbothermal reduction and their photo-
luminescence properties. Chem. Lett. 32, 546–547.
Korotcenkov G. (2005) Gas response control through structural and chemical modification of metal oxides: State
Korotcenkov G. (2007a) Metal oxides for solid state gas sensors. What determines our choice? Mater. Sci. Eng. B
139, 1–23.
Korotcenkov G. (2007b) Practical aspects in design of one-electrode semiconductor gas sensors: Status report.
conductometric-type gas sensors. Mater. Sci. Eng. Rev. 61, 1–39.
Korotcenkov G. and Stetter J.R. (2008) Comparative study of SnO2- and In2O3-based ozone sensors. ECS Trans.
6, 29–41.
Korotchenkov G., Brynzari V., and Dmitriev S. (1998) Kinetics characteristics of SnO2 thin film gas sensors for
environmental monitoring. In: Buettgenbach S. (ed.), Chemical Microsensors and Applications. Proc. SPIE 3539,
196–204.
Korotcenkov G., Brinzari V., DiBattista M., Schwank J., and Vasiliev A. (2001) Peculiarities of SnO2 thin film de-
position by spray pyrolysis for gas sensor application. Sens. Actuators B 77, 244–252.
Korotcenkov G., Cerneavschi A., Brinzari V., Cornet A., Morante J., Cabot A., and Arbiol J. (2002) Crystallographic
characterization of In2O3 films deposited by spray pyrolysis. Sens. Actuators B 84, 37–42.
Korotcenkov G., Brinzari V., Boris Y., Ivanov M., Schawnk J., and Morante J. (2003a) Surface Pd doping influence
on gas sensing characteristics of SnO2 thin films deposited by spray pyrolysis. Thin Solid Films 436, 119–126.
Korotcenkov G., Macsanov V., Tolstoy V., Brinzari V., Schwank J., and Faglia G. (2003b) Structural and gas re-
sponse characterization of nano-size SnO2 films deposited by SILD method. Sens. Actuators B 96, 602–609.
Korotcenkov G., Brinzari V., Cerneavschi A., Ivanov M., Golovanov V., Cornet A., Morante J., Cabot A., and
Arbiol J. (2004a) The influence of film structure on In2O3 gas response. Thin Solid Films 460, 315–323.
Korotcenkov G., Brinzari V., Cerneavschi A., Ivanov M., Cornet A. Morante J., Cabot A., and Arbiol J. (2004b)
In2O3 films deposited by spray pyrolysis: Gas response to reducing (CO, H2) gases. Sens. Actuators B 98(2–3),
236–243.
Korotcenkov G., Cerneavschi A., Brinzari V., Vasiliev A., Cornet A. Morante J., Cabot A., and Arbiol J. (2004c)
In2O3 films deposited by spray pyrolysis as a material for ozone gas sensors. Sens. Actuators B 99, 304–310.
Korotcenkov G., Boris I., Brinzari V., Luchkovsky Yu., Karlotsky G., Golovanov V., Cornet A., Rossinyol E.,
Rodriguez J., and Cirera A. (2004d) Gas sensing characteristics of one-electrode gas sensors on the base of
doped In2O3 ceramics. Sens. Actuators B 103, 13–22.
Korotcenkov G., Brinzari V., Golovanov V., and Blinov Y. (2004e) Kinetics of gas response to reducing gases of
SnO2 films deposited by spray pyrolysis. Sens. Actuators B 98, 41–45.
Korotcenkov G., Cornet A., Rossinyol E., Arbiol J., Brinzari V., and Blinov Y. (2005a) Faceting characterization
of SnO2 nanocrystals deposited by spray pyrolysis from SnCl4-5H2O water solution. Thin Solid Films 471,
310–319.
Korotcenkov G., Brinzari V., Ivanov M., Cerneavschi A., Rodriguez J., Cirera A., Cornet A., and Morante J. (2005b)
Structural stability of In2O3 films deposited by spray pyrolysis during thermal annealing. Thin Solid Films 479,
38–51.
Korotcenkov G., Golovanov V., Cornet A., Brinzari V., Morante J., and Ivanov M. (2005c) Distinguishing feature of
metal oxide films’ structural engineering for gas sensor application. J. Phys.: Confer. Ser. (IOP) 15, 256–261.
Korotcenkov G., Blinov I., and Stetter J.R. (2007a) Kinetics of indium oxide-based thin film gas sensor response: The
role of “redox” and adsorption/desorption processes in gas sensing effects. Thin Solid Films 515, 3987–3996.
Korotcenkov G., Brinzari V., Stetter J.R., Blinov I., and Blaja V. (2007b) The nature of processes controlling the
kinetics of indium oxide-based thin film gas sensor response. Sens. Actuators B 128, 51–63.
Korotcenkov G., Blinov I., Ivanov M., and Stetter J.R. (2007c) Ozone sensors on the base of SnO2 films deposited
by spray pyrolysis. Sens. Actuators B 120, 679–686.
Kosima I., Adachi H., and Yasumori I. (1983) Electronic structures of the LaBO3 (B = Co, Fe, Al) perovskite oxi-
des related to their catalysis. Surf. Sci. 130, 50.
Kotsikau D., Ivanovskaya M., and Orlik D. (2004) Gas-sensitive properties of thin and thick film sensosrs on
Fe2O3-SnO2 nanocomposites. Sens. Actuators B 101, 227–231.
Koudelka-Hep M. (ed.) (2001) Proceedings of the 8th International Meeting on Chemical Sensors, Part 1. 2–5 July
2000, Basel, Switzerland. Sens. Actuators B 76(1–3), 1–673.
Kreuer K.D. (1997) On the development of proton conducting materials for technological aaplications. Solid State
Ionics 97, 1–15.
Krilov O.V. and Kisilev V.F. (1981) Adsorption and Catalysis on the Transition Metals and Their Oxides. Chemistry Press,
Moscow.
Kulkarni D. and Wachs I.E. (2002) Isopropanol oxidation by pure metal oxide catalysts: number of active surface
sites and turnover frequencies. Appl. Catal. A 237, 121–137.
Kulwicki B.M. (1991) Humidity sensors. J. Am. Ceram. Soc. 74, 697–708.
Kumar D. and Sharma R.C. (1998) Advances in conductive polymers. Eur. Polym. J. 34(8), 1053–1060.
Kupriyanov L.Y. (ed.) (1996) Semiconductor Sensors in Physico-Chemical Studies. Elsevier, Amsterdam.
Labeau M., Schmatz U., Delabouglise G., Roman J., Vallet-Regi M., and Gaskov A. (1995) Capacitance effects and
gaseous adsorption on pure and doped polycrystalline tin oxide. Sens. Actuators B 26–27, 49.
Labidi A., Jacolin C., Bendahan M., Abdelghani A., Guerin J., Aguir K., and Maaref M. (2005) Impedance spectro-
scopy on WO3 gas sensor. Sens. Actuators B 106, 713–718.
Lai X., St. Clair T.P., and Goodman D.W. (1998) Scanning tunneling microscopy studies of metal clusters supported
on TiO2(110): morphology and electronic structure. Prog. Surf. Sci. 59, 25–52.
Lamoreaux R.H., Hildenbrand D.L., and Brewer L. (1987) High-temperature vaporization behavior of oxides: II.
Oxides of Be, Mg, Ca, Cr, Ba, B, Al, Ga, In, Ti, Si, Ge, Sn, Pb, Zn, Cd, and Hg. J. Phys. Chem. Perf. Data 16(3),
419–443.
Lampe U., Fleischer M., Reitmeier N., Meixner H., McMonagle J.B., and Marsh A. (1996) New materials for metal
oxide sensors. In: Sensors Update, Part 2. Wiley-VCH, Weinheim.
Lantto V., Rantalla T.S., and Rantalla T.T. (2000) Experimental and theoretical studies on the receptor and transdu-
cer function of SnO2 gas sensors. Electron. Technol. 33(1/2), 22–30.
Lauritsen J.V., Vang R.T., and Besenbacher F. (2006) From atom-resolved scanning tunneling microscopy (STM)
studies to the design of new catalysts. Catal. Today 111, 34–43.
Leblanc E., Perier-Camby L., Thomas G., Gibert R., Primet M., and Gelin P. (2000) NOx adsorption onto dehydro-
xylated or hydroxylated tin dioxide surface. Application to SnO2-based sensors. Sens. Actuators B 62, 67–72.
Lee G.G. and Kang S.-J.K. (2005) Formation of large pores and their effect on electrical properties of SnO2 gas
sensors. Sens. Actuators B 107, 392–396.
Lee J.H. and Park S.J. (1990) Temperature dependence of electrical conductivity in polycrystalline tin oxide. J. Am.
Ceram. Soc. 73(9), 2771–2774.
Li C., Zhang D., Han S., Liu X., Tang T., and Zhou C. (2003) Diameter-controlled growth of single-crystalline
In2O3 nanowires and their electronic properties. Adv. Mater. 15, 143–146.
Li G.J., Zhang X.H., and Kawi S. (1999) Relationships between sensitivity, catalytic activity and surface areas of
SnO2 gas sensors. Sen. Actuators B 60, 64–70.
Liang C., Meng G., Lei Y., Phillip F., and Zhang L. (2001) Catalytic growth of semiconducting In2O3 nanofibers.
Adv. Mater. 13, 1330–1333.
Litzelman S.J., Rothschild A., and Tuller H.L. (2005) The electrical properties and stability of SrTi0.65Fe0.35O3−δ thin
films for automotive oxygen sensor applications. Sens. Actuators B 108, 231–237.
Lu J.G., Chang P., and Fan Z. (2006) Quasi-one-dimensional metal oxide materials—synthesis, properties and ap-
plications. Mater. Sci. Eng. R. 52, 49–91.
Lucas E., Decker S., Khaleel A., Seitz A., Fultz S., Ponce A., Li W., Carnes C., and Klabunde K. (2001) Nanocrystalline
metal oxides as unique chemical reagent/sorbents. Chem. Eur. J. 7, 2505–2510.
Lundstrem I. (1996) Approaches and mechanisms to solid state based sensing. Sens. Actuators B 35–36, 11–19.
Madey T.E., Pelhos K., Wu Q., Barnes R., Ermanoski I., Chen W., Kolodziej J.J., and Rowe J.E. (2002) Nanoscale
surface chemistry. Proc Natl. Acad. Sci. USA 99, 6503–6508.
Madou M.J. and Morrison S.R. (1987) Chemical Sensing with Solid State Devices. Academic Press, Harcourt Brace
Jovanovich, Boston, New York.
Maier J., Holzinger M., and Sitte W. (1994) Fast potentiometric CO2 sensors with open reference electrodes. Solid
State Ionics 74, 5–9.
Maier J., Schiotz J., Liu P., Norskov J.K., and Stimming U. (2004) Nano-scale effects in electrochemistry. Chem. Phys.
Lett. 390, 440–444.
Majoo S., Gland J.L., Wise K.D., and Schwank J.W. (1996) A silicon micromachined conductometric gas sensor with
a maskless Pt sensing film deposited by selected-area CVD, Sens. Actuators B 35–36, 312–319.
Maki-Jaskar M. and Rantalla T.T. (2001) Band structure and optical parameters of the SnO2 (110) surface. Phys. Rev.
B 64, 075407-1-7.
Maki-Jaskar M. and Rantalla T.T. (2002) Theoretical study of oxygen-deficient SnO2 (110) surfaces. Phys. Rev. B 65,
245428-1-8.
Mašek K., Libra J., Skála T., Cabala M., Matolín V., Cháb V., and Prince K.C. (2006) SRPES investigation of tung-
sten oxide in different oxidation states. Surf. Sci. 600, 1624–1627.
Masel R.I. (1996) Introduction to surface reactions. In: Principles of Adsorption and Reaction on Solid State Surfaces.
Wiley, New York, pp. 438–481.
Matko I., Gaidi M., Chenevier B., Charai A., Saikaly W., and Labeau M. (2002) Pt doping of SnO2 thin films.
J. Electrochem. Soc. 149, H153–H158.
Matsuguchi M., Kuroiwa T., Miyagishi T., Suzuki S., Ogura T., and Sakai Y. (1998) Stability and reliability of capaci-
tive-type relative humidity sensors using crosslinked polyimide films. Sens. Actuators B 52, 53–57.
Matsushima S., Teraoka Y., Miura N., and Yamazoe N. (1988) Electronic interaction between metal additives and
tin dioxide in tin dioxide-based gas sensors. Jap. J. Appl. Phys. 27, 1798–1802.
McAleer J.F., Moseley P.T., Norris J.O., and Williams D.E. (1987) Tin dioxide gas sensors: I. Aspects of the surface
chemistry revealed by electrical conductance variations. J. Chem. Soc. Faraday Trans. I, 83, 1323–1346.
Meier J., Schiotzb J., Liub P., Norskov J.K., and Stimminga U. (2004) Nano-scale effects in electrochemistry. Chem.
Phys. Lett. 390, 440–444.
Meixner H. and Lampe U. (1996) Metal oxide sensors. Sens. Actuators B 33, 198–202.
Menesklou W., Schreiner H.J., Moos R., Hardtl K.H., and Ivers-Tiffee E. (2000) Sr(Ti,Fe)O3: material for a tempera-
ture independent resistive oxide sensors. In: Wun-Fogle M., Uchino K., Ito Y., and Gotthardt R. (eds.), Materials
for Smart Systems III, MRS Proceedings Vol. 604. Pittsburgh, PA, pp. 305-310.
Michel H.J., Leiste H., Schierbaum K.D., and Halbritter J. (1998) Adsorbates and their effects on gas sensing pro-
perties of sputtered SnO2 films. Appl. Surf. Sci. 126, 57–64.
Min B.K. and Choi S.D. (2004) SnO2 thin film gas sensor fabricated by ion beam deposition. Sens. Actuators B 98,
239–246.
Min B.K., Wallace W.T., and Goodman D.W. (2006) Support effects on the nucleation, growth and morphology of
gold nano-clusters. Surf. Sci. 600, L7–L11.
Minot C. (2001) Theoretical approaches of the reactivity at MgO(100) and TiO2(110) surfaces. In: Chaer
Nascimento M.A. (ed.), Progress in Theoretical Chemistry and Physics, Vol. 7,. Kluwer, Dordrecht, The Netherlands,
pp. 241–249.
Minot C., Fahmi A., and Ahdjoudj J. (1995) Periodic HF calculations of the adsorption of small molecules on
TiO2. In: Farrugia, L.J. (ed.), The Synergy Between Dynamics and Reactivity at Clusters and Surfaces. Kluwer, Drymen,
Scotland, pp. 257–270.
Monkman G. (2000) Monomolecular Langmuir-Blodgett films—tomorrow’s sensors? Sensor Rev. 20(2), 127–131.
Morrison S.R. (1982) Chemisorption on nonmetalic surfaces. In: Anderson J.R., and Boudart M. (eds.), Catalysid
Science and Technology. Springer-Verlag, Berlin.
Morrison S.R. (1987) Mechanism of semiconductor gas sensor operation. Sens. Actuators 11, 283–287.
Moseley P.T. (1997) Solid state gas sensors. Meas. Sci. Technol. 8, 223–237.
Moseley P.T., Norris J.O.W. and Williams D.E. (eds.) (1991) Techniques and Mechanisms in Gas Sensing. Adam Hilger,
Bristol, UK.
Mrowec S. (1978) On the defect structure in nonstoichiometric metal oxides. Ceram. Int. 4(2), 47–58.
Murch G.E. and Nowick A.S. (eds.) (1984) Diffusion in Crystalline Solids. Academic Press, New York.
Najafi N., Wise K.D., and Schwank J.W. (1994) A micromachined ultra-thin-film gas detector. IEEE Trans. Electron.
Devices 41, 1770–1777.
Nayral C., Viala E., Colliere V., Fau P., Senocq F., Maisonnat A., and Chaudret B. (2000) Synthesis and use of a novel
SnO2 nanomaterial for gas sensing. Appl. Surf. Sci. 164, 219–226.
Nguyen C. and Do D.D. (1999) Adsorption of supercritical gases in porous media: determination of micropore
size distribution. J. Phys. Chem. B 103, 6900–6908.
Ning J.L., Jiang D.M., Kim K.H., and Shim K.B. (2007) Influence of texture on electrical properties of ZnO cera-
mics prepared by extrusion and spark plasma sintering. Ceram. Int. 33, 107–114.
Nitta T. (1981) Ceramic humidity sensor. Ind. Eng. Chem. Prod. Res. Dev. 20, 669–674.
Nitta T., Terada Z., and Hayakawa S. (1980) Humidity-sensitive electrical conduction of MgCr2O4-TiO2 porous
ceramics. J. Am. Ceram.Soc. 63, 295–300.
Noguera C. (1995) Physics and Chemistry at Oxide Surfaces. Cambridge University Press, Cambridge.
Nowick A.S. (1991) Defects in ceramic oxides. MRS Bull. 16(11), 38–41.
Obando L.A. and Booksh K.S. (1999) Tunning dynamic range and sensing of white-light multimode, fiber-optic
surface plasmon resonance sensors. Anal. Chem. 71, 5116–5122.
Oelerich W., Klassen T., and Bormann R. (2001) Metal oxides as catalysts for improved hydrogen sorption in nano-
crystalline Mg-based materials. J. Alloys Compounds 315, 237–242.
Ogawa H., Nishikawa M., and Abe A. (1982) Hall measurement studies and an electrical conduction model of tin
oxide ultrafine particle films. J. Appl. Phys. 53, 4448–4455.
Pan C.A. and Ma T.P. (1980) Work function of In2O3 film as determined from internal photoemission. Appl. Phys.
Lett. 37, 714–716.
Park S.S. and Mackenzie J.D. (1996) Thickness and microstructure effect on alcohol sensing of tin oxide thin films.
Thin Solid Films 274, 154–159.
Pasierb P., Komornicki S., Kozinski S., Gajerski R., and Rekas M. (2004) Long-term stability of potentiometric CO2
sensors based on Nasicon as a solid electrolyte. Sens. Actuators B 101, 47–56
Pauling L. (1929) The principles determining the structure of complex ionic crystals. J. Am. Chem. Soc. 51(4),
1010–1026.
Pederson F.A., Greely J., and Norskov J.K. (2005) Understanding the effects of steps, strain, poisons, and alloying
metane activation on Ni surfaces. Catal. Lett. 105, 9–13.
Penza M., Antolini F., and Vittori-Antisari M. (2004) Carbon nanotubes as SAW chemical sensors materials. Sens.
Post M.L., Tunney J.J., Yang D., Du X., and Singleton D.L. (1999), Material chemistry of perovskite compounds as
chemical sensors. Sens. Actuators B 59, 190–194.
Rabek J.F. (1995) Polymer Photodegradation: Mechanism and Experimental Methods. Chapman & Hall, London.
Rao C.N.R., Kulkarni G.U., Thomas P.J., and Edwards P.P. (2002) Size-dependent chemistry: Properties of nano-
crystals. Chem. Eur. J. 8, 28–35.
Razumovskii S.D. and Zaikov G.Y. (1982) Effect of ozone on saturated polymers. Polymer Sci. U.S.S.R. 24(10),
2805–2325.
Rothschild A. and Komem Y. (2004) The effect of grain size on the sensitivity of nanocrystalline metal oxide gas
sensors. J. Appl. Phys. 95, 6374–6380.
Rothschild A., Litzelman S.J., Tuller H.L., Menesklou W., Schneider T., and Ivers-Tiffee E. (2005) Temperature-
independent resistive oxygen sensors based on SrTi1−xFexO3−δ solid solutions. Sens. Actuators B 108, 223–230.
Ruiz A.M., Cornet A., Shimanoe K., Morante J.R., and Yamazoe N. (2005) Effects of various metal additives on
the gas sensing performances of TiO2 nanocrystals obtained from hydrothermal treatments. Sens. Actuators B
108(1–2), 34–40.
Rumyantseva M.N., Gaskov A.M., Rosman N., Pagnier T., and Morante J.R. (2005) Raman surface vibration modes
in nanocrystalline SnO2: correlation with gas sensor performances. Chem. Mater. 17, 893–901.
Ryabtsev S.V., Shaposhnick A.V., Lukin A.N., and Domashevskaya E.P. (1999) Application of semiconductor gas
sensors for medical diagnostics. Sens. Actuators B 59, 26–29.
Sadaoka Y. (1992) Organic semiconductor gas sensors. In: Sberveglieri G. (ed.), Gas Sensors. Kluwer, Dordrecht,
The Netherlands, pp. 187–218.
Samsonov G.V. (1973) The Oxide Handbook. IFI/Plenum, New York.
Sandler S.R. and Karo W. (1974) Polymer Synthesis. Academic Press, New York.
Satterfield C.N. (1991) Hreterogeneous Catalysis in Industrial Practice. McGraw-Hill New York.
Sberveglieri G. (1992b) Classical and novel techniques for the preparation of SnO2 thin-film gas sensors. Sens.
Sberveglieri G. (1995) Recent developmenta insemiconducting thin film gas sensors. Sens. Actuators B 23, 103–109.
Sberveglieri G. (ed.) (1992a) Gas Sensors—Principles Operation and Development. Kluwer, Dordrecht, The Netherlands.
Sberveglieri G., Gropelli S., Nelli P., and Camanzi A. (1991) A new technique for the preparation of highly sensitive
hydrogen sensors based on SnO2(Bi2O3) thin films. Sens. Actuators B 5, 253–255.
Schierbaum K.D., Weimar U., and Gopel W. (1992) Comparison of ceramic, thick-film and thin-film chemical sen-
sors based upon SnO2. Sens. Actuators B 7, 709–716.
Schierbaum K.D., Weimar U., Gopel W., and Kowalkowski R. (1991) Conductance, work function and catalytic
activity of SnO2-based gas sensors. Sens. Actuators B 3, 205–214.
Schmelzer J., Ropke G., and Mahnke R. (1999) Aggregation Phenomena in Complex Systems. Wiley-VCH, Weinheim.
Schneider W.F., Hass K., Miletic M., and Gland J.I. (2002) Dramatic cooperative effects in adsorption of NOx on
MgO(001). Phys. Chem. B 106(30), 7405–7413.
Semancik S., Cavicchi R.E., Kreider K.G., Suehle J.S., and Chaparala P. (1996) Selected-area deposition of multiple
active films for conductometric microsensors arrays. Sens. Actuator B 34, 209–212.
Seo M.G., Kwang B.H., Chai Y.S., Song K.D., and Lee D.D. (2000) CO2 gas sensor using lithium ionic conductor
with inside heater. Sens.Actuators B 65, 346–348.
Shaw M.P. (1985) Handbook on Semiconductors. North Holland, Amsterdam, pp. 51–60.
Shek C.H., Lai J.K.L., and Lin G.M. (1999) Investigation of interface defects in nanocrystalline SnO2 by positron
annihilation. J. Phys. Chem. Solids 601, 189–193.
Shimizu Y. and Egashira M. (1999) Basic aspects and challenges of semiconductor gas sensors. MRS Bull. 24, 18–24.
Shimuzu Y., Maekawa T., Nakamura T., and Egashira M. (1998) Effects of gas diffusivity and reactivity on gas
sensing properties of thick film SnO2-based sensors. Sens. Actuators B 46, 163–168.
Skotheim T.A., Elsenbaumer R.L., and Reynolds J.R. (eds.) (1998) Handbook of Conducting Polymers. Marcel Dekker,
New York.
Skouras E.D., Burganos V.N., and Payatakes A.C. (1999) Simulation of gas diffusion and sorption in nanoceramic
semiconductors. J. Chem. Phys.110(18), 9244–9253.
Sol C. and Tilley J.D. (2001) Ultraviolet laser irradiation induced chemical reactions of some metal oxides. Mater.
Chem. 11, 815–820.
Somorjai G. (1981) Chemistry in Two Dimensions Surfaces. Cornell University Press. Ithaca, NY.
Sorensen O.T. (ed.) (1981) Non-stoichiometric Oxides. Academic Press, New York.
Spivey J.J. (1987) Complete catalytic oxidation of volatile organics. Ind. Eng. Chem. Res. 26, 2165–2180.
Steffes H., Imawan C., Solzbacher F., and Obermeier E. (1999) Reactively RF-sputtered In2O3 thin films for the
detection of NO2. In: Proceedings of European Conference Eurosensors XIII, The Hague, The Netherlands, 12–15
September 1999, pp. 871–874.
Suzuki T. and Yamazaki T. (1990) Effect of annealing on the gas sensitivity of thin oxide ultra-thin films. J. Mater.
Sci. Lett. 9, 750–751.
Szuber J. and Gopel W. (2000) Photoemission studies of the electronic properties of the space charge layer of SnO2
(110) surface. Electron. Technol. 33(1/2), 216–281.
Talazac L., Brunet J., Battut V., Blanc J.P., Pauly A., Germain J.P., Pellier S., and Soulier C. (2001) Air quality evalua-
tion by monolithic InP-based resistive sensors. Sens. Actuators B 76, 258–264.
Tamaki J., Shimanoe K., Yamada Y., Yamamoto Y. Miura N., and Yamazoe N. (1998) Dilute hydrogen sulfide
sensing properties of CuO-SnO2 thin film prepared by low-pressure evaporation method. Sens. Actuators B 49,
121–125.
Tan O.K., Cao W., Hu Y., and Zhu W. (2004) Nanostructured oxides by high-energy ball milling technique: applica-
tion as gas sensing materials. Solid State Ionics 172, 309–316.
Tan O.K., Zhu W., Yan Q., and Kong L.B. (2000) Size effect and gas sensing characteristics of nanocrystalline
xSnO(1-x) α-Fe2O3 ethanol sensors. Sens. Actuators B 65, 361–365.
Tanner R.E., Liang Y., and Altman E.I. (2002) Structure and chemical reactivity of adsorbed carboxylic acids on
anatase TiO2 (001). Surf. Sci. 506, 251–271.
Tesfamichael T., Motta N., Bostrom T., and Bell J.M. (2007) Development of porous metal oxide thin films by co-
evaporation. Appl. Surf. Sci. 253, 4853–4859.
Tess M.E. and Cox J.A. (1999) Chemical and biochemical sensors based on advances in materials chemistry.
J. Pharmaceut. Biomed. Anal. 19, 55–68.
Thompson M. and Stone D.C. (1997) Surface-Launched Acoustic Wave Sensors: Chemical Sensing and Thin-Film
Characterization. Wiley, New York.
Trasatti S. (1987) Oxide/aqueous solution interfaces. Interplay of surface chemistry and electrocatalysis. Mater.
Chem. Phys. 16, 157–174.
Trasatti S. (ed.) (1980 and 1981) Electrodes of Conductive Metallic Oxides. Part A and B. Elsevier, Amsterdam.
Traversa E., Gnappi G., Montenero A., and Gusmanoa G. (1996) Ceramic thin films by sol-gel processing as novel
materials for integrated humidity sensors. Sens. Actuators B 31, 59–70.
Tsiulyanu D., Marian S., Liess H.D., and Eisele I. (2004) Effect of annealing and temperature on the NO2 sensing
properties of tellurium based films. Sens. Actuators B 100, 380–386.
Tsud N., Johanek V., Stara I., Veltruska K., and Matolin V. (2001) XPS, ISS and TPD study of Pd-Sn interaction on
Pd-SnO2 systems. Thin Solid Films 391, 204–208.
Ustaze S., Guillemont L., Verucchi R., and Esaulov V.A. (1998) Electron capture on surfaces with electronegative
adsorbates and surface poisoning. Surf. Sci. 397, 361–473.
Valden M., Lai X., and Goodman D.W. (1998) Onset of catalytic activity of gold clusters on titania with the ap-
pearance of nonmetallic properties. Science 281, 1647.
Valentini L., Cantalini C., Armentano I., Kenny J.M., Lozzi L., and Santucci S. (2004) Highly sensitive and selective
sensors based on carbon nanotubes thin films for molecular detection. Diamond Rel. Mater. 13, 1301–1305
Valentini L., Cantalini C., Lozzi L., Armentano I., Kenny J.M., and Santucci S. (2003) Reversible oxidation effects
on carbon nanotubes thin films for gas sensing applications. Mater. Sci. Eng. C 23, 523–529.
Van de Krol R. and Tuller H.L. (2002) Electroceramics—the role of interfaces. Solid State Ionics 150, 167–179.
Varghese O.K., Gong D., Paulose M., Grimes C.A., and Dickey E.C. (2003) Crystallization and high-temperature
structural stability of titanium oxide nanotube arrays. J. Mater. Res. 18(1), 156–165.
Varghese O.K., Kichambre P.D., Gong D., Ong K.G., Dickey E.C., and Grimes C.A. (2001) Gas sensing
characteristics of multi-wall carbon nanotubes. Sens. Actuators B 81, 32–41.
Vasiliev R.B., Dorofeev S.G., Rumyantseva M.N., Ryabova L.I., and Gaskov A.M. (2006) Impedance spectroscopy
of the ultradisperse SnO2 ceramic with variable grain size. Phys. Techn. Semicond. 40, 108–111 (in Russian).
Veltruska K., Tsud N., Brinzari V., Korotcenkov G., and Matolin V. (2001) CO adsorption on Pd clusters deposited
on pyrolytically prepared SnO2 studied by XPS. J. Vacuum 61, 129–134.
Voorhoeve R.J., Remeika J.P., Freeland P.E., and Matthias B.T. (1972) Rare earth oxides of manganese and cobalt
rival platinum for the treatment of carbon monoxide in auto exhaust. Science 177(46), 353.
Wachs I.E. (1996) Raman and IR studies of surface metal oxide species on oxide supported metal oxide catalysts.
Catal. Today 27, 437–455.
Wahlstrom E., Lopez N., Schaub R., Thostrup P., Ronnau A., Africh C., Lagsgaard E., Norskov J.K., and Besenbacher
F. (2003) Bonding of gold nanoclusters to oxygen vacancies on rutile TiO2(110). Phys. Rev. Lett. 90, 026101.
Wahlstrom E., Vestergaard E.K., Schaub R., Ronnau A., Vestergaard M., Lagsgaard E., Stensgaard I., and
Besenbacher F. (2004) Electron transfer–induced dynamics of oxygen molecules on the TiO2 (110) surface.
Science 303, 511–513.
Wallace R.M. (2004) Challenges for the characterization and integration of high-K dielectrics. Appl. Surf. Sci. 231–
232, 543–551.
Walton D.J. (1990) Electrically conducting polymers. Mater. Design 11(3), 142–152.
Wang C.C., Akbar S.A., and Madou M.J. (1998) Ceramic based resistive sensors. J. Electrocer. 2(4), 273–282.
Wang J., Gan M., and Shi J. (2007b) Detection and characterization of penetrating pores in porous materials. Mater.
Character. 58, 8–12.
Wang X., Fei Y., Lu H., Jin K.J., Zhu X.D., Chen Z., and Yang G. (2007a) In-diffusion of oxygen vacancies near step
edges dominates the oxidation of perovskite films. J. Phys.: Cond. Matter 19, 026206.
Wang X., Yee S.S., and Carey W.P. (1995) Transition between neck-controlled and grain-boundary-controlled sensi-
tivity of metal-oxide gas sensors. Sens. Actuators B 25, 454–457.
Weast R.C., Astle M., and Beyer W. (eds.) (1988) CRC Handbook of Chemistry and Physics, 69th ed. CRC Press, Boca
Raton, FL.
Werle P., Slemr F., Maurer K., Kormann R., Mucke R., and Janker B. (2002) Near-and mid-infrared laser-optical
sensors for gas analysis. Optics Lasers Eng. 37, 101–1114.
Williams D. (1999) Semiconducting oxides as gas-sensitive resistors. Sens. Actuators B 57, 1–16.
Williams D.E. and Pratt K.F.E. (1997) Self-diagnostic gas-sensitive resistors in sour gas applications. Sens. Actuators
B 45, 147–153.
Williams D.E. and Pratt K.F.E. (1998) Classification of reactive sites on the surface of polycrystalline tin dioxide.
J. Chem. Soc. Faraday Trans. 94, 3493–3500.
Williams D.E. and Pratt K.F.E. (2000) Microstructure effects on the response of gas-sensitive resistors based on
semiconducting oxides. Sens. Actuators B 70, 214–221.
Williams G. and Coles G.S.V. (1998b) Gas sensing properties of nanocrystalline metal oxide powders produced by
a laser evaporation technique. J. Mater. Chem. 8, 1657–1664.
Wilson D.M., Dunman K., Roppel T., and Kalim R. (2000) Rank extraction in tin-oxide sensor arrays. Sens. Actuators
B 62, 199–210.
Woormann H., Kohl D., and Heiland G. (1979) Work function and band bending on clean cleaved zinc oxide sur-
faces. Surf. Sci. 80, 261–264.
Xu C., Tamaki J., Miura N., and Yamazoe N. (1990) Relationship between gas sensitivity and microstructure of
porous SnO2. J. Electrochem. Soc. Jpn. 58, 1143–1148.
Xu C., Tamaki J., Miura N., and Yamazoe N. (1991) Grain size effects on gas sensitivity of porous SnO2-based
elements. Sens. Actuators B 3, 147–155.
Yakovlev Y.P., Baranov A.N., Imenkov A.N., and Mikhailova M.P. (1991) Optoelectronic LED-photodiode pairs
for moisture and gas sensors in spectral range 1.8–4.8 μm. In: Wolfbeis S. (ed.), Chemical and Medical Sensors.
Proc. SPIE 1510, 170–177.
Yamazoe N. (1991) New approaches to improvement semiconducting gas sensors. Sens. Actuators B 5, 7–18.
Yamazoe N. and Miura N. (1992a) Some basic aspects of semiconductor gas sensors. In: Yamauchi S. (ed.) Chemical
Sensors Technology, vol. 4. Kodansha, Tokyo, pp. 20–41.
Yamazoe N. and Miura N. (1992b) New approaches in the devices of gas sensors. In: Sberveglieri G. (ed.) Gas
Sensors. Kluwer, Dordrecht, The Netherlands.
Yamazoe N., Fuchigami J., Kishikawa M., and Seiyama T. (1979) Interactions of tin oxide surface with O2, H2O
and H2. Surf. Sci. 86, 335–344.
Yamazoe N., Kurokawa Y., and Seiyama T. (1983) Effects of additives on semiconductor gas sensors.
Sens. Actuators 4, 283–289.
Yamazoe N., Matsushima S., Maekawa T., Tamaki J., and Miura N. (1991) Control of Pd dispersion in SnO2-based
sensors. Catal. Sci. Technol. 1, 201–205.
Yanagida, H. (1990) Intelligent ceramics. Ferroelectrics 102, 251–257.
Zambelli T., Trost J., Wintterlin J.G., and Ertl G. (1996) Diffusion and atomic hopping of N atoms on Ru(0001)
studied by scanning tunneling microscopy. Phys. Rev. Lett. 76, 795–798.
Zapola P., Jaffe J.B., and Anthony C. Hess A.C. (1999) Ab-initio study of hydrogen adsorption on the ZnO (101:0)
Zemel J. (1988) Theoretical description of gas film interaction on SnOx . Thin Solid Films 163, 89–195.
Zhang J. and Gao L. (2004) Synthesis of antimony-doped tin oxide (ATO) nanoparticles by the nitrate–citrate
combustion method. Mater. Res. Bull. 39, 2249–2255.
Zhang S., Yongqing D.S., and Du F.H. (2003) Recent advances of superhard nanocomposite coatings: a review. Surf.
Coat. Technol. 167, 113–119.
Zhang Y.C., Tagawa H., Asakura S., Mizusaki J., and Narita H. (1997) Solid-state CO2 sensor with Li2CO3–Li3PO4–
LiAlO2 electrolyte and LiCoO2–Co3O4 as a solid reference electrode. J. Electrochem. Soc. 144(12), 4345–4350.
Ziolek M., Kujawa J., Saur O., Aboulayt A., and Lavalley J.C. (1996) Influence of sulfur dioxide adsorption on the
surface properties of metal oxides. J. Mol. Catal. A: Chem. 112, 125–132.
CHAPTER 3
COMBINATORIAL CONCEPTS FOR

DEVELOPMENT OF SENSING MATERIALS
R. A. Potyrailo
1. INTRODUCTION
Rational design of sensing materials based on prior knowledge is a very attractive approach because
it may avoid time-consuming synthesis and testing of numerous materials candidates (Honeybourne
2000; Shtoyko et al. 2004; Njagi et al. 2007). However, to be quantitatively successful, rational design
(Newnham 1988; Akporiaye 1998; Ulmer II et al. 1998; Lavigne and Anslyn 2001; Suman et al. 2003;
Hatchett and Josowicz 2008) requires detailed knowledge regarding relation of intrinsic properties of
sensing materials to a set of their performance properties. This knowledge is typically obtained from
extensive experimental and simulation data. However, with the increase of structural and functional
complexity of materials, the ability to rationally define the precise requirements that result in a desired
set of performance properties becomes increasingly limited (Schultz 2003). Thus, in addition to rational
design, a variety of sensing materials, ranging from dyes and ionophores to biopolymers, organic and
hybrid polymers, and nanomaterials, have been discovered using detailed experimental observations
or simply by chance (McKusick et al. 1958; Pedersen 1967; Bühlmann et al. 1998; Svetlicic et al. 1998;
Walt et al. 1998; Steinle et al. 2000; Martin et al. 2001; Hu et al. 2004; Potyrailo and Sivavec 2004). Such
an approach in development of sensing materials reflects a more general situation in materials design
that is “still too dependent on serendipity” and with only limited capability for rational materials design
(Eberhart and Clougherty 2004).
Conventionally, detailed experimentation with sensing materials candidates for screening and opti-
mization consumes a tremendous amount of time and project cost. Thus, developing sensing materials
159
is a recognized challenge because it requires extensive experimentation to achieve not only the best
short-term performance but also long-term stability, manufacturability, and other practical require-
ments. Numerous practical challenges in rational sensing material design provide tremendous prospects
for the application of combinatorial methodologies for the development of sensing materials.
This chapter demonstrates the broad applicability of combinatorial technologies in discovery
and optimization of new sensing materials. We discuss general principles of combinatorial materials
screening, followed by a discussion of the opportunities facilitated by combinatorial technologies for
discovery and optimization of new sensing materials. We further critically analyze results of materials
development using discrete and gradient materials arrays and provide examples from a wide variety of
sensors based on various energy-transduction principles that involve radiant, mechanical, and electrical
types of energy.
2. GENERAL PRINCIPLES OF COMBINATORIAL

MATERIALS SCREENING
Combinatorial materials screening is a process that couples the capability for parallel production of
large arrays of diverse materials together with different high-throughput measurement techniques for
various intrinsic and performance properties, followed by navigation through the collected data to iden-
tify “lead” materials (Jandeleit et al. 1999; Takeuchi et al. 2002; Xiang and Takeuchi 2003; Potyrailo and
Amis 2003; Koinuma and Takeuchi 2004; Potyrailo et al. 2004a; Potyrailo and Takeuchi 2005b; Potyrailo
and Maier 2006). The terms combinatorial materials screening and high-throughput experimentation are typically
applied interchangeably to all types of automated parallel and rapid sequential evaluation processes for
materials and process parameters that include truly combinatorial permutations or selected subsets.
Individual aspects of accelerated materials development have been known for decades. These in-
clude combinatorial and factorial experimental designs (Birina and Boitsov 1974), parallel synthesis of
materials on a single substrate (Kennedy et al. 1965; Hoffmann 2001), screening of materials for per-
formance properties (Hoogenboom et al. 2003), and computer data processing (Anderson and Moser
1958, Eash and Gohlke 1962). However, in 1970, Hanak (1970) suggested an integrated materials-
development workflow. Its key aspects included (1) complete compositional mapping of a multicom-
ponent system in one experiment; (2) simple, rapid, nondestructive, all-inclusive chemical analysis; (3)
testing of properties by a scanning device; and (4) computer data processing. In 1995, Xiang, Schultz,
and co-workers (1995) initiated applications of combinatorial methodologies in materials science. Since
then, combinatorial tools have been employed to discover and optimize a wide variety of materials (see
Table 3.1).
A typical combinatorial materials development cycle is outlined in Figure 3.1 (Potyrailo and Mirsky
2008, 2009). Compared to the initial idea of Hanak (1970), the modern workflow has several new and
important aspects, such as design/planning of experiments, data mining, and scale-up. In combinatorial
screening of materials, concepts originally conceived as highly automated have been recently refined to
have more human input, with only an appropriate level of automation. For the throughput of 50–100 ma-
terials formulations per day, it is acceptable to perform certain aspects of the process manually (Potyrailo
et al. 2002, 2003a). To address numerous materials-specific properties, a variety of high-throughput
COMBINATORIAL CONCEPTS FOR DEVELOPMENT OF SENSING MATERIALS • 161
Table 3.1. Examples of materials developed using combinatorial

screening techniques
MATERIALS REFERENCE
Superconductor materials Xiang et al. 1995
Ferroelectric materials Chang et al. 1998
Magnetoresistive materials Briceño et al. 1995
Luminescent materials Danielson et al. 1998
Agricultural materials Wong and Robertson 1999
Structural materials Zhao 2001
Hydrogen storage materials Olk 2005
Organic light-emitting materials Zou et al. 2001
Ferromagnetic shape-memory alloys Takeuchi et al. 2003
Thermoelastic shape-memory alloys Cui et al. 2006
Heterogeneous catalysts Holzwarth et al. 1998
Homogeneous catalysts Cooper et al. 1998
Polymerization catalysts Lemmon et al. 2001
Electrochemical catalysts Reddington et al. 1998
Electrocatalysts for hydrogen evolution Greeley et al. 2006
Polymers Brocchini et al. 1997
Zeolites Lai et al. 2001
Metal alloys Ramirez and Saha 2004
Materials for methanol fuel cells Jiang et al. 2005
Materials for solid oxide fuel cells Lemmon et al. 2004
Materials for solar cells Hänsel et al. 2002
Automotive coatings Chisholm et al. 2002
Waterborne coatings Wicks and Bach 2002
Vapor-barrier coatings Grunlan et al. 2003
Marine coatings Stafslien et al. 2006
Fouling-release coatings Ekin and Webster 2007
characterization tools are required. Characterization tools are used for rapid and automated assessment
of single or multiple properties of the large number of samples fabricated together as a combinatorial
array or “library” (MacLean et al. 2000; Potyrailo and Amis 2003; Potyrailo and Takeuchi 2005a).
In addition to the parallel synthesis and high-throughput characterization instrumentation, which
differ significantly from conventional equipment, the data management approaches also differ from
conventional data evaluation (Potyrailo and Maier 2006). In a well-developed combinatorial workflow,
design and synthesis protocols for materials libraries are computer-assisted, materials synthesis and
library preparation are carried out with computer-controlled manipulators, and property screening and
materials characterization are also software-controlled. Further, materials synthesis data as well as prop-
erty and characterization data are collected into a materials database. This database contains informa-
tion on starting components, their descriptors, process conditions, materials testing algorithms, and per-
formance properties of libraries of sensing materials. Data in such a database are not just stored, they
Figure 3.1. Typical process for combinatorial materials development.
are also processed with the proper statistical analysis, visualization, modeling, and data-mining tools.
Combinatorial synthesis of materials provides a good possibility for formation of banks of combinato-
rial materials (Potyrailo and Mirsky 2009). Such banks can be used, for example, for further investigation
of materials of interest for new applications or as reference materials.
3. OPPORTUNITIES FOR SENSING MATERIALS

The development process for a sensor system with a new sensing material can be described using tech-
nology readiness levels (TRLs) as shown in Figure 3.2. The concept of TRLs is an accepted way to as-
sess technology maturity (2005). These TRLs provide a scale from TRL 1 (least mature) to TRL 9 (most
mature) that describes the maturity of a technology with respect to a particular use. Sensor development
includes several phases, including discovery with initial observations, feasibility experimentation, and
laboratory-scale detailed evaluation (TRLs 1–4), followed by validation of components and a whole-
system prototype in the field (TRLs 5–6), and then testing of the system prototype in the operational
environment (TRL 7) and tests and end-use operation of the actual system (TRLs 8–9).
At the initial concept stage, performance of the sensing material is matched with the appropriate
transducer for signal generation. The stage of laboratory-scale evaluation is very labor-intensive because
it involves detailed testing of sensor performance. Several key aspects of this evaluation include opti-
mization of the sensing material composition and morphology, its deposition method, detailed evalua-
tion of response accuracy, stability, precision, selectivity, shelf-life, long term stability of the response,
and key noise parameters (e.g., material instability because of temperature, potential poisons). Thus, as
illustrated in Figure 3.2, combinatorial methodologies for the development of sensing materials have
broad opportunities in TRLs 1–5.
4. DESIGNS OF COMBINATORIAL LIBRARIES OF

SENSING MATERIALS
The broad goals of combinatorial development of sensing materials are to discover and optimize per-
formance parameters and to optimize fabrication parameters of sensing materials. The key perfor-
mance and fabrication parameters of sensing materials are outlined in Figure 3.3. Factors affecting
performance of sensing material films are also summarized in Figure 3.3 and can be categorized as
those originating from the sample, the sample/film interface, the bulk of the film, and the film/sub-
strate interface. Depending on the real-world application, the qualities of the sensing materials are often
Figure 3.2. Opportunities for combinatorial development of sensing materials across technology
readiness levels.
Figure 3.3. Broad goals of combinatorial development of sensing materials and examples of factors affecting materials performance.
weighted differently. For example, response speed with millisecond time resolution is critical in gas
sensors for intensive care, while a much slower response speed is sufficient in home blood glucose bio-
sensors (Pickup and Alcock 1991; Newman and Turner 2005). Specific requirements for medical in vivo
sensors and bioprocess sensors include sample compatibility (Meyerhoff 1993; Clark and Furey 2006).
Resistance to gamma radiation during sterilization, drift-free performance, and cost are the most critical
specific requirements for sensors in disposable bioprocess components (Clark and Furey 2006).
Combinatorial experimentation is performed by arranging materials candidates as discrete and gra-
dient sensing materials arrays. A wide variety of array fabrication methods have been reported, as sum-
marized in Table 3.2. The specific type of library layout depends on the required density of space to be
explored, available library-fabrication capabilities, and capabilities of high-throughput characterization
techniques. Upon array fabrication, the array is exposed to an environment of interest and steady-state
or dynamic measurements are acquired to assess materials performance. Serial scanning analysis (e.g.,
by optical or impedance spectroscopy) is often performed to provide more detailed information about
materials properties than parallel analysis (e.g., imaging) does. When monitoring a dynamic process (e.g.,
response/recovery time, aging) of sensing materials arranged in an array with a scanning system, the
maximum number of elements in the sensor library that can be measured with the required temporal
resolution may be limited by the data-acquisition ability of the scanning system (Potyrailo and Hassib
2005). In addition to measurements of materials performance parameters, it is important to characterize
intrinsic materials properties (Göpel,1998).
To demonstrate the broad applicability of combinatorial technologies in discovery and optimiza-
tion of sensing materials, in the following sections we critically analyze results of materials develop-
ment using discrete and gradient materials arrays and provide examples from a wide variety of sensors
based on various energy-transduction principles that involve radiant, mechanical, and electrical types
of energy.
5. DISCOVERY AND OPTIMIZATION OF SENSING MATERIALS USING

DISCRETE ARRAYS
5.1. RADIANT ENERGY TRANSDUCTION SENSORS
Sensors based on radiant energy transduction can be categorized on the basis of the five parameters
that completely describe a light wave: amplitude, wavelength, phase, polarization state, and time-depen-
dent waveform. The majority of the development of sensing materials for these types of sensors relies
on colorimetric and fluorescent materials properties. At present, organic fluorophores dominate sens-
ing applications because of the diversity of their functionality and well-understood synthesis methods,
but new semiconducting nanocrystal labels have several advantages (photostability, relatively narrow
emission spectra, and broad excitation spectra (Alivisatos 2004; Medintz et al. 2005)) over organic fluo-
rophores. Thus, finding a solution to complement the existing organic fluorescent reagents with more
photostable yet chemically or biologically responsive nanocrystals is very attractive. It is known that a
variety of photoluminescent materials are sensitive to the local environment (Ko and Meyer 1999). In
particular, polished or etched bulk CdSe semiconductor crystals (Lisensky et al. 1988; Seker et al. 2000 )
Table 3.2. Examples of fabrication methods of discrete and gradient

materials arrays
TYPES OF ARRAYS OF
SENSING MATERIAL FABRICATION METHODS REFERENCES
Discrete arrays Ink jet printing Lemmo et al. 1997; Calvert 2001;
de Gans and Schubert 2004
Robotic liquid dispensing Apostolidis et al. 2004;
Hassib and Potyrailo 2004
Robotic slurry dispensing Scheidtmann et al. 2005
Microarraying Schena 2003
Automated dip-coating Potyrailo et al. 2004d
Electropolymerization Mirsky and Kulikov 2003;
Mirsky et al. 2004
Chemical vapor deposition Taylor and Semancik 2002
Pulsed-laser deposition Aronova et al. 2003
Spin coating Cawse et al. 2003; Amis 2004
Screen printing Potyrailo et al. 2007
Electrospinning Yoon et al. 2009
Gradient arrays In situ photopolymerization Dickinson et al. 1997
Microextrusion Potyrailo et al. 2003d, 2005b;
Potyrailo and Wroczynski 2005
Solvent casting Potyrailo et al. 2003c, 2004;
Potyrailo and Hassib 2005
Colloidal self-assembly Potyrailo et al. 2008b
Surface-grafted orthogonal Bhat et al. 2004
polymerization
Ink jet printing Turcu et al. 2005
Temperature-gradient Taylor and Semancik 2002
chemical vapor deposition
Thickness-gradient Sysoev et al. 2004
chemical vapor deposition
2-D thickness-gradient Klingvall et al. 2005
evaporation of two metals
Gradient surface coverage Baker et al. 1996
and gradient particle size
and nanocrystals (Nazzal et al. 2003; Vassiltsova et al. 2007) were shown to be sensitive to environmen-
tal changes. To better understand the environmental sensitivity of semiconductor nanocrystals upon
their incorporation into polymer films, mixtures of multisize CdSe nanocrystals were incorporated into
numerous rationally selected polymeric matrices (see Table 3.3) to produce thin films. These films were
further screened for their photoluminescence (PL) response to vapors of different polarity upon excita-
tion with a 407-nm laser (Potyrailo and Leach 2006; Leach and Potyrailo 2006).
It was discovered that CdSe nanocrystals of different sizes (2.8 and 5.6 nm in diameter) and pas-
sivated with tri-n-octylphosphine oxide had dramatically different PL response patterns upon exposure
to methanol and toluene after incorporation into polymeric matrices (see Figure 3.4A). As an exam-
ple, Figure 3.4B shows response patterns of gas-dependent PL of the two-size CdSe nanocrystals in
poly(methyl methacrylate) (PMMA, polymer 2 in Table 3.3) sensor film. The difference in the response
patterns of the nanocrystals was attributed to the combined effects of the dielectric medium surrounding
the nanocrystals, their size, and their surface oxidation state. The sensing films were tested for 16 h under
continuous laser excitation and exhibited high stability of PL intensity (Potyrailo and Leach 2006).
To evaluate polymer matrices quantitatively, K-nearest-neighbor (KNN) cluster analysis was em-
ployed as a data-mining tool. Cluster analysis is often used in assessing the diversity of materials accord-
ing to their compositional or performance properties and in developing structure–property relationship
models (Otto 1999; Potyrailo 2001). In KNN analysis, links are made between nearest neighbors of ad-
joining clusters. A measure that accounts for the different scales of variables and their correlations is the
Mahalanobis distance (Otto 1999). Results of cluster analysis of PL response patterns upon exposure to
methanol and toluene after incorporation into polymeric matrices are demonstrated in the dendrogram
in Figure 3.4C. The dendrogram was constructed by performing principal-component analysis on the
data from Figure 3.4A and then using Mahalanobis distance on three principal components (PCs). From
this dendrogram, it is clear that polymers 6 and 7 (see Table 3.3) were the most similar in their vapor re-
sponse the the CdSe nanocrystals studied, as demonstrated by a very small distance to K-nearest neigh-
bors between them. Polycaprolactone (polymer 4 in Table 3.3) was the most different from the rest of
the polymers, as indicated by the largest diversity distance to the K-nearest neighbor. Such data-mining
tools provide a means to evaluate polymer matrices quantitatively. Coupled with quantitative structure–
property relationship simulation tools that incorporate molecular descriptors, new knowledge generated
from high-throughput experiments may provide additional insights for rational design of gas sensors
Table 3.3. Polymer matrices for incorporation of different-size CdSe nanocrystals
POLYMER NO. POLYMER TYPE RATIONALE FOR SELECTION AS SENSOR MATRIX

1 Poly(trimethyl-silyl) propyne Polymer with largest known solubility of oxygen,
candidate for efficient oxidation of CdSe nanocrystals
2 Poly(methyl methacrylate) Polymer for solvatochromic dyes
3 Silicone block polyimide Polymer with very high partition coefficient for sorbing
organic vapors
4 Polycaprolactone Polymer for solvatochromic dyes
5 Polycarbonate Polymer with high Tg for sorbing of organic vapors
6 Polyisobutylene Polymer with low Tg for sorbing of organic vapors
7 Poly(dimethylaminoethyl) Polymer for surface passivation of semiconductor
methacrylate nanocrystals
8 Polyvinylpyrrolidone Polymer for sorption of polar vapors
9 Styrene-butadiene ABA Polymer for sorbing of nonpolar vapors
block copolymer
Source: Data from Leach and Potyrailo 2006.
Figure 3.4. Diversity of steady-state photoluminescence response of two-size (2.8- and 5.6-nm)
mixtures of CdSe nanocrystals to polar (methanol) and nonpolar (toluene) vapors. (A) Magnitude
of PL change in nine polymer matrices listed in Table 3.3. (B) Gas-dependent PL of two-size CdSe
nanocrystals sensor film (polymer #2) with emission of 2.8-nm nanocrystals at 511 nm and emission
of 5.6-nm nanocrystals at 617 nm. (C) Results of KNN cluster analysis of PL response patterns upon
exposure to methanol and toluene after incorporation into nine polymer matrices. Numbers 1 and 2 in
(B) are replicate exposures of sensor film to methanol (6% vol.) and toluene (1.5% vol.), respectively.
[(A) Reprinted with permission from Leach and Potyrailo 2006. Copyright 2006 Materials Research
Society. (B) Reprinted with permission from Potyrailo and Leach 2006. Copyright 2006 American
Institute of Physics.]
based on incorporated semiconductor nanocrystals. In the future, such work promises to complement
existing solvatochromic organic dye sensors with more photostable and reliable sensor materials.
Optimizing sensor materials that have been formulated is a cumbersome process because theoreti-
cal predictions are often limited by practical issues, such as poor solubility and compatibility of formula-
tion components (Mills et al. 1998; Bedlek-Anslow et al. 2000; Wang et al. 2003; Apostolidis et al. 2004).
These practical issues represent significant knowledge gaps that prevent a more efficient rational design
of sensor materials. Thus, combinatorial methodologies have been demonstrated for the development
of multicomponent sensor materials for gaseous (Potyrailo and Brennan 2004; Amis 2004; Apostolidis
et al. 2004) and ionic (Potyrailo 2004; Hassib and Potyrailo 2004, 2005a; Chojnacki et al. 2004) species.
Because polymer matrices (Hartmann and Trettnak 1996; Draxler et al. 1995; Mohr and Wolfbeis 1996;
Conway et al. 1997; Dickinson et al. 1997; Walt et al. 1998; Kolytcheva et al. 1999; Amao 2003) and
plasticizers (Preininger et al. 1996; Kolytcheva et al. 1999; Legin et al. 2002; Peper et al. 2003; Penco et
al. 2004) are known to affect the response of sensors for gases and liquids, automated screening was
applied to determine which polymers and plasticizers were best to use in constructing oxygen-sensing
materials based on Ru(4,7-diphenylphenanthroline) fluorophore. Structures of polymers 1–9 and plas-
ticizers 10–13 are presented in Figures 3.5 and 3.6, respectively. Following the initial study screening
polymer matrices 1–9 (Figure 3.7A), focused libraries were constructed with plasticizers 10–13 to tune
sensor sensitivity (Figure 3.7B). While in general the sensitivity of the sensor coatings increased with the
plasticizer concentration, because of an increase in the permeability of oxygen in the polymer matrix,
it was unexpectedly found that plasticizers 12 and 13 showed an initial decrease of sensitivity at low
concentrations. By combining manual and automated steps in the preparation of discrete sensor film
arrays, it was possible to reduce the time needed to screen sensor materials by at least 1000 (Apostolidis
et al. 2004).
Applying polymers with an intrinsic conductivity also permits development of chemical and bio-
logical sensors (Leclerc 1999; McQuade et al. 2000; Janata and Josowicz 2002; Dai et al. 2002; Bobacka
et al. 2003). A variety of conjugated organic monomers readily undergo polymerization and form linear
polymers. For example, acetylene, p-phenylenevinylene, p-phenylene, pyrrole, thiophene, furane, and
aniline form conducting polymers that are widely employed in sensors (Bidan 1992; Albert et al. 2000;
McQuade et al. 2000; Gomez-Romero 2001). However, as prepared, conducting polymers lack selectiv-
ity and often are unstable. Thus, such polymers are chemically modified to reduce these undesirable
effects. Modification methods include side-group substitution of heterocycles, doping of polymers,
charge compensation upon polymer oxidation by incorporation of functionalized counterions, forma-
tion of organic–inorganic hybrids, incorporation of various biomaterials (e.g., enzymes, antibodies,
nucleic acids, cells), and others (Bidan 1992; Gill and Ballesteros 1998; Gill 2001). Variations in po-
lymerization conditions (oxidation potential, oxidant, temperature, solvent, electrolyte concentration,
monomer concentration, etc.) can be also employed to produce diverse polymeric materials from the
same monomer, because polymerization conditions affect sensor-related polymer properties (morphol-
ogy, molecular weight, connectivity of monomers, conductivity, band gap, etc.) (McQuade et al. 2000;
Barbero et al. 2003).
Recently, a combinatorial approach for the colorimetric differentiation of organic solvents was
developed (Yoon et al. 2009). A polydiacetylene (PDA)–embedded electrospun fiber mat, prepared with
aminobutyric acid–derived diacetylene monomer (PCDA-ABA), displayed colorimetric stability when
Figure 3.5. Structures of polymers 1–9 employed for construction of oxygen-sensing materials
based on Ru(4,7-diphenylphenanthroline) fluorophore.
exposed to common organic solvents. In contrast, a fiber mat prepared with the aniline-derived diacety-
lene (PCDA-AN) exhibited a solvent-sensitive color transition. Arrays of PDA-embedded microfibers
were constructed by electrospinning poly(ethylene oxide) solutions containing various ratios of two
diacetylene monomers. Unique color patterns were developed when the conjugated polymer-embedded
electrospun fiber arrays were exposed to common organic solvents in a manner which enabled di-
rect colorimetric differentiation of the solvents. Results of these experiments are presented in Figure
3.8. Scanning electron microscopy (SEM) images of electrospun fiber mats encapsulated with DA
monomers prepared from pure PCDA-ABA, pure PCDA-AN, and a 1:1 molar mixture of PCDA-ABA
and PCDA-AN are presented in Figure 3.8A. No significant morphological differences were observed
among these electrospun fiber mats and polymer fibers with an average diameter of ~1 μm. Color pat-
terns of the combinatorial arrays of fiber mats derived from different combinations of DA monomers
(see Figure 3.8B) demonstrated the significance of the combinatorial approach for sensor development.
This methodology enables the generation of a compositionally diverse array of sensors starting with
only two DA monomers for visual differentiation of organic solvents.
In addition to sensing materials for radiant sensors, plasmonic nanostructures have also attracted
significant interest for chemical and biological sensing because of their potential for a dramatic im-
provement of sensor performance with a concurrent simplification of instrument design. Significant
improvements have been accomplished in developing nanoplasmonic structures for surface-enhanced
Raman spectroscopic (SERS) sensors and for localized surface plasmon resonance (LSPR) sensors.
Combinatorial approaches have been applied to optimize SERS substrates based on nanoparticles.
These approaches include 2-D variation of nanoparticle density to discover the existence of an opti-
mal surface coverage for the most effective SERS enhancement (Baker et al. 1996) and combinatorial
exploration of the roughness effects of metallic substrates (Kahl et al. 1998). It was proposed (Baker et
al. 1996) that the SERS enhancement of optimized periodic structures can be much larger than that of
simple island films, and the power of solution-based combinatorial approaches was demonstrated for
synthesis of surfaces exhibiting nanometer-scale variation in mixed-metal composition and architecture.
The SERS response with variable surface coverage of colloidal gold and the amount of silver coating
on Au nanoparticles (Ag staining) was studied in detail. A gradient in particle coverage was produced
along the direction of immersion of a glass slide by fixed-rate immersion of the (3-mercaptopropyl)
trimethoxysilane-coated glass slide into an aqueous solution of 12-nm-diameter colloidal Au particles.
The slide was further rotated by 90° and fixed-rate immersion was performed into an Ag ion–contain-
ing solution for Ag staining of Au. This second immersion step produced a gradient in particle size
over the new immersion direction. The resulting surface exhibited a continuous variation in nanometer-
Figure 3.6. Structures of plasticizers 10–13 employed for construction of oxygen-sensing materials
based on Ru(4,7-diphenylphenanthroline) fluorophore.
Figure 3.7. Results of combinatorial screening of steady-state responses of formulated optical gas
sensor materials. (A) Stern-Volmer plots of oxygen quenching of Ru(4,7-diphenylphenanthroline)–
based fluorophore in polymers 1–9 as changes in the fluorescence decay time. (B) Effect of type
and concentration of plasticizers 10–13 on the sensitivity of fluorescent oxygen-sensing materials
in polymer 1. (Reprinted with permission from Apostolidis et al. 2004. Copyright 2004 American
Chemical Society.)
scale morphology as defined by particle coverage and particle sizes. The SERS signal for adsorbed p-
nitrosodimethylaniline was measured over a 2 × 2 cm sample and exhibited continuous gradients in Au
coverage and Ag cladding thickness as shown on a spatial map of background-corrected SERS intensity
for the phenyl-nitroso stretch at 1168 cm−1. Detailed investigation revealed a region that was over 103-
fold more SERS-enhancing than the least active sites (see Figure 3.9). The nanometer-scale morphology
at positions of interest was determined by atomic force microscopy, and the results showed significant
changes in SERS enhancement factor with only small alterations in surface morphology.
To provide more fundamental insight into design of nanoplasmonic sensors, several computational
methods have been applied (Christensen and Fowers 1996, Schatz 2007, Montgomery et al. 2008, Zhao
et al. 2008). For simulations of LSPR sensors, computational data on refractive index sensitivity of
LSPR structures agrees qualitatively with experiments (Stewart et al. 2006; McMahon et al. 2007; Zheng
et al. 2008; Lee et al. 2008a; Chen et al. 2009; Potyrailo et al. 2009a). However, numerous practical
factors affect the refractive index sensitivity of LSPR structures, which are difficult to assess quantita-
tively. For example, the discrepancy between experimental and theoretical results has been attributed to
surface contamination during fabrication or during experiments (Zheng et al. 2008), to small but very
sensitive deviations between the experimental array structures and the theoretical idealized configura-
tions (McMahon et al. 2007), to the effects of directly transmitted background (Chen et al. 2009), and
to contribution from the edges of small-dimension nanohole arrays (Genet and Ebbesen 2007; Yang
et al. 2008). Overall, examples of important factors can be grouped as (1) possible edge effects of
relatively small-size arrays; (2) surface roughness effects; (3) surface chemistry (contamination) effects;
(4) imperfections in the fabrication of individual features; (4) array defects; and (5) general effects from
optical setup alignment (Potyrailo et al. 2009a). Thus, in addition to simulations, often different LSPR
structures are explored experimentally as combinatorial arrays.
The dependence of the visible-light transmission efficiency on the shape and size of nanoaper-
tures in Au films was explored in quantitative detail experimentally by Matteo et al. (2004). The goal of
these experiments was to find the shape and size of a subwavelength aperture with the largest visible-
light transmission efficiency (see Figure 3.10). The aperture shapes studied included C apertures, same-
area squares, same-area rectangles, and square apertures of the same near-field spot size (see Figure
3.10A). These studies were inspired by computational results and microwave experiments showing that
C-shaped apertures can overcome large light attenuation due to their cross-sectional shape. However,
utilizing these apertures in the visible regime presented design challenges. To make these subwavelength
apertures at optical frequencies, focused ion beam milling (FIB) was employed (see Figure 3.10B). The
apertures were formed in 200/5-nm (Au/Cr) film with the critical dimension d for all the apertures
ranging from 40 to 55 nm in 5-nm steps. The transmission properties of these individual apertures were
measured using white light for comparison of the performance of C apertures to square and rectan-
gular apertures. Figures 3.10C and 3.10D show transmitted colors and intensities of studied arrays for
horizontal and vertical polarizations, respectively. Both the C apertures and the rectangular apertures
were strongly resonant at the horizontal polarization and brighter than the square apertures. The C ap-
ertures changed from red to a blue-green in appearance as their size decreased, while the other apertures
Figure 3.8. Combinatorial approach for colorimetric differentiation of organic solvents based
on conjugated polymer-embedded electrospun fibers. (A) SEM images of electrospun fiber mats
embedded with (I) PCDA-ABA, (II) PCDA-AN, and (III) 1:1 molar ratio of PCDA-ABA and PCDA-AN
after UV irradiation. (B) Photographs of the polymerized PDA-embedded electrospun fiber mats after
exposure to organic solvents at 25°C for 30 s. (Color photographs can be seen in the original paper.)
(Reprinted with permission from Yoon et al. 2009. Copyright 2009 Wiley-VCH.)
Figure 3.9. Combinatorial discovery of the most active SERS region from a 2-D gradient of variable surface coverage of 12-nm-diameter
Au nanoparticles and variable thickness of Ag shell. (Left) Background-subtracted SERS intensity map for the 1168-cm−1 phenyl-nitroso
stretch of p-nitrosodimethylaniline. Each shaded box represents one SERS spectrum collected in a 1- × 1-mm2 area. (Right) AFM images
of nanoparticles (500 × 500 nm2) from regions A–D. (Color photographs can be seen in the original paper.) (Reprinted with permission
from Baker et al. 1996. Copyright 1996 American Chemical Society.)
Figure 3.10. Quantitative analysis of arrays of subwavelength LSPR apertures of different geometries.
(A) Design details and (B) SEM images of C aperture, same-area square, same-area rectangle,
and square aperture aimed to produce the same near-field spot size. (C) and (D) Photographs
taken through microscope eyepiece of the entire sample under horizontal and vertical polarization,
respectively. Each group of three rows was fabricated with the same shape but with decreasing d
size, from 55 nm to 40 nm in 5-nm steps. (Color photographs can be seen in the original paper.)
(Reprinted with permission from Matteo et al. 2004. Copyright 2004 American Institute of Physics.)
exhibited a much smaller size-dependent resonance shift, with the rectangular apertures strongly trans-
mitting red light and the same-area square apertures having a dim blue color. Strikingly, under vertical
polarization, the size-dependent color of the C apertures was lost, with all sizes having a dim blue color,
while the rectangular apertures shorted out and the squares remained unchanged. The small square
apertures, designed to produce the same spot size as the C aperture, did not even transmit enough light
to exceed the noise level of the detector. Throughput enhancements of the C-shaped apertures over
square apertures of the same area ranging from 13 times for the d = 40-nm apertures to 22 times for the
d = 55-nm apertures were observed. The reason for the improved performance of the larger apertures
was the rounding and tapering effects of smaller structures, which is inherent in FIB milling.
Round apertures (round nanoholes) have been demonstrated in numerous LSPR sensors (Anker et
al. 2008; Stewart et al. 2008; Potyrailo et al. 2009a) because of the relative simplicity of their fabrication.
To demonstrate multiplexed, high-spatial-resolution, and real-time analysis of biorecognition events,
small-dimension (≤20 μm2) ordered arrays of subwavelength holes were made with FIB milling (Yang
et al. 2008). Nanohole arrays were perforated on a super-smooth Au film (220 nm thickness, roughness
rms < 2.7 Å) deposited onto a fluoropolymer (fluorinated ethylene propylene copolymer, FEP) sub-
strate. The smooth Au surface provided a superb environment for fabricating nanometer features and
uniform immobilization of biomolecules. The fluorinated FEP polymer chosen as the substrate was
chemically inert, transparent in the visible region, and had a refractive index of 1.341 at λ = 590 nm,
close to that of biological solutions. With the matched refractive indices on both sides of the Au film,
the intensity of extraordinary transmission through the nanoholes was shown to be strongly enhanced
(Martín-Moreno et al. 2001; Krishnan et al. 2001; Yang et al. 2008). The refractive index matching be-
tween FEP and biological solutions contributed to ~20% improvement in sensing performance. For
quantitative assessment of sensing performance, a series of nanohole patterns with different nanohole
diameters, periodicities, and numbers was fabricated. Figure 3.11A shows a SEM micrograph of the top
view of a typical nanohole array. Figure 3.11B illustrates an intensity image of a pattern of thirty 9 × 9
nanohole arrays spaced by 11 μm. In the fabricated arrays, the hole diameter was changed from 130 to
180 nm and the hole periodicity/diameter ratio was varied from 1.67 to 3.00. By analyzing the transmis-
sion spectral features from series of nanohole arrays, it was determined that the hole diameters only
slightly affected the transmission maxima, and the periodicity was the dominant factor.
Figure 3.11. Fabricated series of 30-nanohole patterns with different nanohole diameters and hole
periodicity/diameter ratios. (A) SEM micrograph of the top view of a typical 9 × 9 nanohole array. (B)
Intensity image at 540 nm of a pattern of thirty 9 × 9 nanohole arrays spaced by 11 μm. In the fabri-
cated arrays, the hole diameter was changed from 130 to 180 nm and the hole periodicity/diameter
ratio was varied from 1.67 to 3.00. (Color photographs can be seen in the original paper.) (Reprinted
with permission from Yang et al. 2008. Copyright 2008 American Chemical Society.)
Application of nanoslits in LSPR sensors promises significant further improvement in sensor per-
formance (Lee et al. 2007; Potyrailo et al. 2008a; Lee et al. 2008b; Wu et al. 2009). Detection sensitivi-
ties of gap plasmons in gold nanoslit arrays were studied and compared with surface plasmons on the
outside surface (Lee et al. 2008b). The nanoslit arrays (each 150 μm × 150 μm) were fabricated in a
130-nm-thick Au film with various slit widths ranging from 20 to 100 nm and different periods between
the nanoslits ranging from 400 to 900 nm (see Figure 3.12). With transverse-magnetic (TM) incident
waves, the nanoslit arrays showed two distinguishable transmission peaks corresponding to the reso-
nances of gap plasmons and surface plasmons. The surface sensitivities for both modes were compared
by coating thin SiO2 film and different biomolecules on the nanoslit arrays. Experimental results dem-
onstrated that gap plasmons were more sensitive than conventional surface plasmons, with an increase
of detection sensitivity with a decrease of slit width. Experiments with BSA biomolecules (molecular
weight 66 kDa) demonstrated that the detection sensitivity of SPR mode was not significantly improved
when the slit width was changed, whereas the cavity mode was very sensitive when slit width was smaller
than 50 nm. Interestingly, a slight decrease in sensitivity of the cavity mode was observed for the small-
est slits tested, with widths of 20–30 nm.
5.2. MECHANICAL ENERGY TRANSDUCTION SENSORS
Sensors based on mechanical energy transduction can be categorized on the basis of transducer func-
tionality and include cantilevers and acoustic-wave devices. The mass loading and/or changes in the
viscoelastic properties of the sensing materials lead to the transducer response.
A 2-D multiplexed cantilever array platform was developed for an elegant combinatorial screening
of vapor responses of alkane thiols with different functional end groups (Lim et al. 2006; Raorane et al.
2008). The cantilever sensor array chip (size 2.5 × 2.5 cm) had ~720 cantilevers and was fabricated using
surface and bulk micromachining techniques. The optical readout has been developed for parallel analy-
sis of deflections from individual cantilevers. Figure 3.13A illustrates a general view of the 2-D cantilever
array system. To evaluate the performance of this 2-D sensor array for screening of sensing materials,
nonpolar and polar vapors such as toluene and water vapor were selected as analytes. The screening sys-
tem was tested with three candidate alkane thiol materials with different functional end groups as sens-
ing films: mercaptoundecanoic acid SH–(CH2)10–COOH (MUA), mercaptoundecanol SH–(CH2)11–OH
(MUO), and dodecanethiol SH–(CH2)11–CH3 (DOT). Each type of sensing films had different chemi-
cal and physical properties because the –COOH group is acidic in nature and can dissociate to give a
–COO− group, the –OH group does not dissociate easily but can form hydrogen bonds with polar
molecules, and the –CH3 group is inert to polar molecules, so the only interactions it can have result from
van der Waals and hydrophobic effects. Results of these experiments are presented in Figure 3.13B.
Since toluene is an organic solvent, it is likely to act via van der Waals interaction with the thiol
film on the gold side. Thus, the van der Waals intermolecular interaction is generally attractive: It brings
thiol molecules close to toluene molecules. This in turn brings the thiol molecules closer to each other,
inducing shrinkage in the Au layer, and this results in upward deflection as shown in Figure 3.13B. In
the case of DOT, because the –CH3 group is nonpolar, it has maximum contact area with toluene and
thus exhibits van der Waals interaction. This tendency is reduce as the end groups become more polar
Figure 3.12. Experiments with arrays of nanoslits in a gold film with variable slit width and periodicity. (A) Example of a fabricated nanoslit
array (period 600 nm, slit width 80 nm). (B) Transmission optical images of various nanoslit arrays in air and in water with variable array
periods (400–900 nm) and slit widths (20–100 nm). Area of each nanoslit array was 150 μm × 150 μm. The colors of the nanoslit arrays
(shown in shades of gray) were due to extraordinary transmissions of surface plasmon resonances. (Color photographs can be seen in
the original paper.) (Reprinted with permission from Lee et al. 2008b. Copyright 2008 Elsevier.)
Figure 3.13. Combinatorial vapor-response screening of alkane thiols with different functional end groups using a 2-D multiplexed canti-
lever array system. (A) General view of the fabricated cantilever array chip. (B) Steady-state deflection values of cantilevers upon expo-
sure to toluene vapor at four concentration levels (3, 6, 9, and 12% by mass). (Reprinted with permission from Lim et al. 2006. Copyright
2006 Elsevier.)
in nature. Hence, the –OH group of MUO undergoes more van der Waals interaction than the –COOH
group of MUA. As a result, the induced stress in the Au layer is maximum for DOT, medium for MUO,
and least for MUA. The water vapor experiment was performed with relative humidity (RH) levels rang-
ing from 8.8 to 61.8% RH. An upward deflection was recorded for all thiols, indicating that the Au film
was under compression. The response ranking of three thiols to water vapor was opposite compared to
the response to toluene. The largest response was of cantilevers coated with MUA, followed by those
coated with MUO, and the smallest response was of cantilevers coated with DOT. This multiplexed
cantilever sensor platform was further proposed as a search tool for sensing materials with improved
selectivity (Lim et al. 2006).
Polymeric materials are widely used for sensing because they provide the ability for room-temper-
ature sensor operation, rapid response and recovery times, and long-term stability over several years
(Hierlemann et al. 1995; Potyrailo and Sivavec 2004). In gas sensing with polymeric materials, polymer–
analyte interaction mechanisms include dispersion, dipole induction, dipole orientation, and hydrogen
bonding (Grate et al. 1997; Grate 2000). These mechanisms facilitate partial selectivity of response of
different polymers to diverse vapors. An additional molecular selectivity in response is added by ap-
plying molecular imprinting of target vapor molecules into polymers and formulating polymers with
molecular receptors. Several models have been developed to calculate polymer responses (Grate and
Abraham 1991; Abraham 1993; Maranas 1996; Wise et al. 2003; Belmares et al. 2004), but the most
widely employed model is based on linear solvation energy relationships (LSER) (Grate and Abraham
1991; Abraham 1993). This LSER method has been applied as a guide to select a combination of avail-
able polymers to construct an acoustic-wave sensor array based on thickness shear mode (TSM) resona-
tors for determination of organic solvent vapors in the headspace above groundwater (Potyrailo et al.
2004b). Field testing of the sensor system (Potyrailo et al. 1999) demonstrated that its detection limit
with available polymers was too high (several parts per million) to meet the requirements for detection
of groundwater contaminants. However, a new polymer for sensing has been found (silicone block
polyimide 14, Figure 3.14) that has a partition coefficient of >200,000 to part-per-billion concentra-
tions of trichloroethylene (TCE) and that provides at least 100 times more sensitive response for de-
tection of chlorinated organic solvent vapors than other known polymers (Sivavec and Potyrailo 2002;
Potyrailo and Sivavec 2004).
For development of materials for more selective part-per-billion detection of chlorinated solvent
vapors in the presence of interferencing substances, six families of polymeric materials were fabricated
based on polymer 14. Performance of these polymeric materials was evaluated with respect to the
differences in partition coefficients to the analytes perchloroethylene (PCE), trichloroethylene (TCE),
and cis-dichloroethylene (cis-DCE) and interfering carbon tetrachloride, toluene, and chloroform. For
quantitative screening of sensing materials candidates, a 24-channel TSM sensor system was built that
matched a 6 × 4 microtiter wellplate format (Figures 3.15A and 3.15B). The sensor array was positioned
in a gas flow-through cell and kept in an environmental chamber. Comprehensive materials screening
was performed at three levels (Potyrailo et al. 2003b, 2004c). In the primary (discovery) screening, ma-
terials were exposed to a single analyte concentration. In the secondary (focused) screening, the best
materials subset was exposed to analytes and interfering compounds. Finally, in the tertiary screening,
the remaining materials were tested under conditions mimicking long-term application. While all the
screenings were valuable, the tertiary screening provided the most intriguing data because aging of base
Figure 3.14. Structure of vapor-sensing polymer 14.
Figure 3.15. Approach for high-throughput evaluation of sensing materials for field applications. (A)
Setup schematic of a 24-channel TSM sensor array for gas-sorption evaluation of sorbing polymeric
films. (B) Photo of 24 sensor crystals (including two reference sealed crystals) in a gas flow cell. [(A)
Reprinted with permission from Potyrailo et al. 2004c. Copyright 2004 American Institute of Physics.]
polymers and copolymers is difficult or impossible to model (Ulmer II et al. 1998). From the tertiary
screening, the decrease in materials response to the nonpolar analyte vapors and the increase in re-
sponse to a polar interference vapor were quantified.
For detailed evaluation of diversity of the fabricated materials, principal-components analysis
(PCA) tools (Beebe et al. 1998) were applied as shown in Figure 3.16. PCA is a multivariate data analysis
tool that projects the data set onto a subspace of lower dimensionality with collinearity removed. PCA
achieves this objective by explaining the variance of the data matrix in terms of the weighted sums
of the original variables with no significant loss of information. These weighted sums of the original
variables are called principal components (PCs). The capability of discriminating among six vapors us-
ing eight types of polymers was evaluated using a scores plot (see Figure 3.16A). It demonstrated that
these six vapors are well separated in the PCA space when these eight types of polymers are used for
determinations. To understand what materials induce the most diversity in the response, a loadings plot
was constructed (see Figure 3.16B). The bigger the distance between the films of the different types, the
better was the differences between these films. The loadings plot also demonstrates the reproducibility
of the response of replicate films of the same materials. Such information served as an additional input
into materials selection for the tertiary screening. However, material selection on the basis of PCA alone
does not guarantee optimal discrimination of particular vapors in the test set, because PCA measures
variance, not discrimination (Grate 2000). Thus, cluster analysis tools—e.g., those shown in Figure
3.4C—can be also applied.
This 24-channel TSM sensor array system was further applied for high-throughput screening of
solvent resistance of a family of polycarbonate copolymers prepared from the reaction of bisphenol A
(BPA), hydroquinone (HQ), and resorcinol (RS) with the goal of using these copolymers as solvent-re-
sistant supports for deposition of solvent-containing sensing formulations (Potyrailo et al. 2005a). The
mass increase of the crystal was determined during periodic exposure of the TSM crystals to polymer/
Figure 3.16. Application of PCA tools for determination of differences in the response pattern of the
sensor materials toward analytes (PCE, TCE, and cis-DCE) and interferences (carbon tetrachloride,
toluene, and chloroform): (left) scores and (right) loadings plots of the first two principal components.
(Adapted with permission from Potyrailo et al. 2004c. Copyright 2004 American Institute of Physics.)
Figure 3.17. Application of the 24-channel TSM sensor array system for mapping of solvent resis-
tance of polycarbonate copolymers. (A) General view of the screening system with a 6 × 4 micro-
titer wellplate positioned below the sensor array. (B) Example of property/composition mapping of
solvent-resistance of polycarbonate copolymers in tetrahydrofuran. Numbers in the contour lines
are normalized sensor frequency shift values (hertz per milligram of polymer in a well of the micro-
titer wellplate). [(A) Reprinted with permission from Potyrailo et al. 2004d. Copyright 2004 American
Chemical Society. (B) Reprinted with permission from Potyrailo et al., 2006a. Copyright 2006
American Chemical Society.]
solvent combinations (Figure 3.17A) (Potyrailo et al. 2004d) and was found to be proportional to the
amount of polymer dissolved and deposited onto the sensor from a polymer solution. The high mass
sensitivity of the resonant TSM sensors (10 ng), use of only a minute volume of solvent (2 mL), and
parallel operation (matching a layout of available 24 microtiter wellplates) made this system a good fit
with available polymer combinatorial synthesis equipment. These parallel determinations of polymer–
solvent interactions also eliminated errors associated with serial determinations. The data were further
mined to construct detailed solvent-resistance maps of polycarbonate copolymers and to determine
quantitative structure–property relationships (see Figure 3.17B) (Potyrailo et al. 2006a). The application
of this sensor-based polymer screening system provided a lot of stimulating data, which would be dif-
ficult to obtain using a conventional one-sample-at-a-time approach.
5.3. ELECTRICAL ENERGY TRANSDUCTION SENSORS

Sensors based on electrical energy transduction are applicable for combinatorial screening of sensing
materials when these materials undergo electrically detectable changes, for example, changes in resis-
tance or conductance during polymerization reactions and exposure to species of interest, changes
in resistance due to swelling of polymers, interactions of metal-oxide semiconducting surfaces with
oxidizing or reducing species, etc. Typical devices for these applications include electrochemical and
electronic transducers (Zemel 1990; Suzuki 2000; Wang 2002; Hagleitner et al. 2002, 2003).
The simplicity of microfabrication of electrode arrays and their subsequent application as trans-
ducer surfaces makes sensors based on electrical energy transduction among the most employed tools in
combinatorial materials screening. The possibility of regulating polymerization on solid conductive sur-
faces by application of corresponding electrochemical potentials suggested a realization of this process
in the form of multiple polymerization regions on multiple electrodes of an electronic sensor system
(Mirsky and Kulikov 2003; Kulikov and Mirsky 2004). Arranging such polymerization electrodes in an
array eliminated the need for dispensing systems and allowed an electrically addressable immobilization.
This approach has been demonstrated on electropolymerization of aniline that was performed inde-
pendently on different electrodes of the array (Mirsky and Kulikov 2003; Kulikov and Mirsky 2004).
Thin-layer polymerization of defined mixtures of monomers was performed directly on the 96 inter-
digital addressed electrodes of an electrode array on an area of less than 20 mm × 20 mm (see Figure
3.18A). The electrodes had an interdigital configuration designed for four-point measurements and were
fabricated by lithography on an oxidized silicon wafer. Computer-controlled addition of analyte species
provided automated investigation of the influence of different substances on the synthesized polymers.
This system has been applied for combinatorial electrochemical copolymerization of mixtures of the
Figure 3.18. Application of a microfabricated electrode sensor array for multiple electropolymeriza-
tions and characterization of resulting conducting polymers as sensor materials. (A) Layout of the
interdigital addressed electrode array. Inset, detailed structure of the single electrode for four-point
measurements. (B) Current kinetics during combinatorial electropolymerization at constant potential
for different ratios of nonconductive additive to conductive monomer in the polymerization mixture.
(Adapted with permission from Kulikov and Mirsky 2004. Copyright 2004 Institute of Physics.)
nonconductive monomer aminobenzoic acid and the conductive monomer aniline at various ratios to
form diverse polymers. Six groups of polymers were formed at the same conditions on 12 electrodes
per group with ratios of monomers ranging from 100:1 to 4:1 to study the effects of the nonconductive
aniline derivative (see Figure 3.18B). Electrical characterization of the polymers demonstrated that incor-
poration of aminobenzoic acid as a nonconductive additive in the polymer structure disturbed the con-
ductive polymer chains and led to a strong decrease of polymer conductance, corresponding to predicted
behavior. Exposure to HCl gas showed that polymer conductance increased with HCl concentration.
Numerous copolymers have been screened using this electropolymerization system. Example
screening results for different binary copolymers are presented in Figure 3.19 (Potyrailo and Mirsky
2008). Introduction of nonconductive monomers into the polymer decreased the polymer conductance
and therefore decreased the difference between conductive and insulating polymer states. This caused
a decrease of absolute sensitivity (Figure 3.19A). Normalization to the polymer conductance without
Figure 3.19. Selected results of screening of sensing materials for their response to HCl gas: (A)
best absolute sensitivity; (B) best relative sensitivity; (C) best response rate; (D) best recovery ef-
ficiency, performed by heating. Sensor materials: ANI indicates polyaniline; 4ABA, 3ABSA, 3ABA, AA
indicate polymers synthesized from aniline and 4-aminobenzoic acid, 3-aminobenzenesulfonic acid,
3-aminobenzoic acid, and anthranilic acid, respectively. Gray and black bars are the results obtained
by 2- and 4-point techniques, respectively. (Reprinted with permission from Potyrailo and Mirsky
2008. Copyright 2008 American Chemical Society.)
analyte exposure compensated for this effect and demonstrated that the polymer synthesized from the
mixture of aminobenzoic acid and aniline possessed the highest relative sensitivity (Figure 3.19B). This
effect may be explained by the strong dependence of polymer conductance on the defect number in the
polymer chains. In comparison with pure polyaniline, this copolymer had better recovery efficiency but
slower response time (Figures 3.19C and 3.19D). The high-throughput screening system was capable of
reliably ranking sensing materials and required only ~20 min of manual interaction with the system and
~14 h of computer-controlled combinatorial screening, compared to ~2 weeks of laboratory work us-
ing traditional electrochemical polymer synthesis and materials evaluation (Mirsky and Kulikov 2003).
An interesting approach has been recently reported (Setasuwon et al. 2008) to enhance the di-
versity of responses of sensors with conventional polymer films (see Figure 3.20). Eight sensors with
unique characteristics were made by combinatorial stacking of three layers of two types of carbon
black–formulated composites such as carbon black with ethylene-co-vinyl acetate (composite E) and
carbon black with poly(vinyl alcohol) (composite A). Eight sensors were constructed on interdigitated
electrodes with 200-μm spacing on 10- × 15-mm glass substrates by spin-coating three consecutive
layers of composites A and E. As shown in Figure 3.20A, eight sensors resulted from all combinations
of three layers of two different composites, EEE, EEA, EAE, EAA, AAE, AEA, AEE, and AAA,
where the first letter indicates the first layer applied onto the electrodes. These sensors were subjected
to 15 solvent vapors with a broad range of dielectric constants from 2 to 80. The test set of solvents
consisted of water, dimethylformamide (DMF), methanol, ethanol, acetone, isopropanol, 1-butanol,
1,2-dichlorobenzene, 1-octanol, tetrahydrofuran (THF), ethyl acetate, chlorobenzene, chloroform, tolu-
ene, and hexane. Results of the response patterns of all sensors to 15 tested solvents are presented in
Figure 3.20B. The responses of the sensors were processed for calculation of resolution factor, the
pairwise parameter representing the ability to resolve the responses of one analyte vapor from the oth-
ers. If the detector responses are assumed to have a normal distribution, the resolution factor values
of 1.0, 2.0, and 3.0 indicate 76%, 92%, and 98% confidence, respectively, of correctly identifying one
analyte from the other of a specific pair. It was found that by stacking layers of two types of carbon
black–formulated composites, it was possible to improve the selectivity of vapor measurements using
resistive sensors. The resolution factor was affected by both the dielectric constant and the boiling point
of the solvents tested.
Semiconducting metal oxides are another type of sensing materials that benefit from combinato-
rial screening technologies. Semiconducting metal oxides are typically used as gas-sensing materials that
change their electrical resistance upon exposure to oxidizing or reducing gases. Over the years, significant
technological advances have been made that resulted in practical and commercially available sensors, and
now new materials are being developed that further improve the performance of these sensors. To en-
hance response selectivity and stability, an accepted approach is to formulate multicomponent materials
that contain additives in metal oxides. Introducing additives into base metal oxides can change a variety
of materials properties, including concentration of charge carriers, energy spectra of surface states, ener-
gy of adsorption and desorption, surface potential and intercrystallite barriers, phase composition, sizes
of crystallites, catalytic activity of the base oxide, stabilization of a particular valence state, formation of
active phases, stabilization of the catalyst against reduction, the electron exchange rate, etc. Dopants
can be added at the preparation stage (bulk dopants) that will affect the morphology, electronic proper-
ties of the base material, and catalytic activity. However, the fundamental effects of volume dopants
Figure 3.20. Method for enhancement of the diversity of responses of sensors with conventional polymer films. (A) Combinatorial stack-
ing of three layers of two types of carbon black–formulated composites, such as carbon black with ethylene-co-vinyl acetate (composite
E) and carbon black with poly(vinyl alcohol) (composite A), onto interdigitated electrodes. Eight resulting sensor configurations are EEE,
EEA, EAE, EAA, AAE, AEA, AEE, and AAA. (B) Results of the response patterns of eight sensors, EEE, EEA, EAE, EAA, AAE, AEA, AEE,
and AAA, to 15 tested solvents. (Reprinted with permission from Setasuwon et al. 2008. Copyright 2008 American Chemical Society.)
on base materials are not yet predictable (Siemons et al. 2007). Addition of dopants to the preformed
base material (surface dopants) can lead to different dispersion and segregation effects, depending on
the mutual solubility (Franke et al. 2006) and influence on the overall oxidation state of the metal oxide
surface (Korotcenkov 2005; Franke et al. 2006; Barsan et al. 2007; Siemons et al. 2007).
To improve the productivity of materials evaluation by using combinatorial screening, a 36-ele-
ment sensor array was employed to evaluate various surface-dispersed catalytic additives on SnO2 films
(Semancik 2002, 2003). Catalysts were deposited by evaporation to nominal thicknesses of 3 nm, and
then the microhotplates were heated to effect the formation of a discontinuous layer of catalyst par-
ticles on the SnO2 surfaces. The layout of the fabricated 36-element library is shown in Figure 3.21A.
The response characteristics of SnO2 with different surface-dispersed catalytic additives are presented
in Figure 3.21B. These radar plots show sensitivity results to benzene, hydrogen, methanol, and ethanol
for operation at three temperatures.
To expand the capabilities of screening systems, it is attractive to characterize not only the con-
ductance of the sensing materials with DC measurements but also their complex impedance spectra
(Barsoukov and Macdonald 2005). The use of complex impedance spectroscopy provides the capability
to test both ion- and electron-conducting materials and to study electrical properties of sensing materi-
als that are determined by the material microstructure, such as grain boundary conductance, interfacial
polarization, and polarization of the electrodes (Simon et al. 2002, 2005). A 64-multielectrode array has
been designed and built for high-throughput impedance spectroscopy (10–107 Hz) of sensing materials
(see Figure 3.22A) (Simon et al. 2002). In this system, an array of interdigital capacitors was screen-
printed onto a high-temperature-resistant Al2O3 substrate. To ensure high quality of the determinations,
parasitic effects caused by the leads and contacts were compensated by a software-aided calibration
(Simon et al. 2002). After the system validation with doped In2O3 and automation of the data evalua-
tion (Simon et al. 2005), the system was implemented for screening of a variety of additives and matri-
ces with the long-term goal of developing materials with improved selectivity and long-term stability.
Sensing films were applied using robotic liquid-phase deposition based on optimized sol-gel synthesis
procedures. Surface doping was achieved by the addition of appropriate salt solutions followed by
library calcination. Screening results at 350°C of thick films of In2O3 base oxide surface doped with
various metals are presented as bar diagrams in Figure 3.22B (Sanders and Simon 2007). It was found
that some doping elements lead to changes in both the conductivity in air as well as the gas-sensing
properties toward oxidizing (NO2, NO) and reducing (H2, CO, propene) gases. Correlations between
the sensing and the electrical properties in reference atmosphere indicated that the effect of the dop-
ing elements was due to an influence on the oxidation state of the metal oxide surface rather that to an
interaction with the test gas. This accelerated approach for generating reliable systematic data was then
coupled to data-mining statistical techniques that resulted in the development of (1) a model associating
the sensing properties and the oxidation state of the surface layer of the metal oxide based on oxygen
spillover from doping element particles to the metal oxide surface and (2) an analytical relation for the
temperature-dependent conductivity in air and nitrogen that described the oxidation state of the metal
oxide surface taking into account sorption of oxygen (Sanders and Simon 2007).
This high-throughput complex impedance screening system was further employed for reliable
screening of a wide variety of less explored material formulations. Polyol-mediated synthesis has been
known as an attractive method for preparation of nanoscale metal oxide nanoparticles (Feldmann 2001).
Figure 3.21. Combinatorial study of effects of surface dispersion of metals into CVD-deposited SnO2
films. (A) Layout of a 36-element library for study of the sensing characteristics of SnO2 films with
3 nm of surface-dispersed Pt, Au, Fe, Ni, or Pd (Con. = control). Each sample was made with six
replicates. (B) Radar plots of sensitivity results to benzene, hydrogen, methanol, and ethanol for
operation at 150, 250, and 350°C. (Color photographs can be seen in the original paper.) (Reprinted
with permission from Semancik 2002.)
Figure 3.22. Screening of sensor metal oxide materials using complex impedance spectroscopy and a multielectrode 64-sensor array.
(A) Layout of 64-sensor array. (B) Relative gas sensitivities at 350°C of the In2O3 base oxide materials library surface-doped with multiple
salt solutions, concentration 0.1 atom % if not denoted otherwise; ND = undoped. Sequence of test gases and their concentrations (with
air in between) was H2 (25 ppm), CO (50 ppm), NO (5 ppm), NO2 (5 ppm), propene (25 ppm). [(A) Reprinted with permission from Simon
et al. 2002. Copyright 2002 American Chemical Society. (B) Reprinted with permission from Sanders and Simon 2007. Copyright 2007
American Chemical Society.]
It requires only low annealing temperatures and provides the opportunity to tune the composition of
the materials by mixing the initial components on the molecular level (Siemons and Simon 2006, 2007).
To explore previously unknown combinations of p-type semiconducting nanocrystalline CoTiO3 with
different volume dopants as sensing materials, the polyol-mediated synthesis method was used to synthe-
size nanometer-sized CoTiO3 followed by volume doping with Gd, Ho, K, La, Li, Na, Pb, Sb, and Sm (all
at 2 at%). The SEM-estimated primary particle size of the volume-doped CoTiO3 materials was in the
range from 30 to 140 nm, with the smallest particle size for CoTiO3:La and the largest for CoTiO3:K.
The significant amount of data collected during experiments with numerous sensing materials
candidates facilitated successful efforts to develop data-mining techniques (Sieg et al. 2006, 2007) and a
database system (Frantzen et al. 2005). The data-mining tools, based on hierarchical clustering maps (see
Figure 3.23), have been applied to identify several promising materials candidates such as In99.5Co0.5Ox,
W99Co0.5Y0.5Ox, W98.3Ta0.2Y1Mg0.5Ox, W99.5Ta0.5Ox, and W99.5Rh0.5Ox with different gas-selectivity pat-
terns (Frenzer et al. 2006).
Figure 3.23. Hierarchical clustering map of 2112 responses of diverse sensing materials to H2,
CO, NO, and propene (Prop.) at four temperatures established from high-throughput constant-
current measurements and processed with Spotfire data-mining software (clustering algorithm was
“complete linkage” of the Euclidean distances). (Color photographs can be seen in the original
paper.) (Reprinted with permission from Frenzer et al. 2006. Copyright 2006 Molecular Diversity
Preservation International.)
6. OPTIMIZATION OF SENSING MATERIALS USING

GRADIENT ARRAYS
Sensor material optimization can be performed using gradient sensor materials. Spatial gradients in
sensing materials can be generated by varying the nature and concentration of starting components,
processing conditions, thickness, and some other parameters. Once a gradient sensor array is formed,
it is important to estimate the possibilities to adequately measure the variation of properties along the
gradient. These can be intrinsic (thickness, chemical composition, morphology, etc.) or performance
(response magnitude, selectivity, stability, immunity to poisoning, etc.) properties.
6.1. VARIABLE CONCENTRATION OF REAGENTS
Optimization of concentrations of formulation components can require significant effort because of

the nonlinear relationship between additive concentration and sensor response (Papkovsky et al. 1995;
Dickinson et al. 1997; Collaudin and Blum 1997; Mills 1998; Levitsky et al. 2001; Eaton 2002; Florescu
and Katerkamp 2004; Basu et al. 2005). Concentration-gradient sensor material libraries have been
employed for detailed optimization of sensor materials. One-, two-, and three-component composition
gradients were made by flow-coating individual liquid formulations onto a flat substrate and allowing
them to merge under diffusion control while still containing solvents (Potyrailo and Hassib 2005). This
method combines the fabrication of gradients of materials composition with recording the materi-
als’ response before and after analyte exposure and taking the ratio or difference of responses. These
gradient films were applied for optimization of sensor materials formulations for analysis of ionic and
gaseous species (Potyrailo and Hassib 2005). A very low reagent concentration in the film is expected to
produce only a small signal change. A small signal change is also expected when the reagent concentra-
tion is too high. Thus, the optimal reagent concentration depends on the analyte concentration and the
activity of the immobilized reagent.
Concentration optimization of a colorimetric reagent was performed in a polymer film for detec-
tion of trace concentrations of chlorine in water. A concentration gradient of a near-infrared cyanine
dye was formed in a poly(2-hydroxyethyl methacrylate) hydrogel sensing film. The optical absorption
profile A0(x) was obtained before analyte exposure to map the reagent concentration gradient in the
film. A subsequent scanning across the gradient after analyte exposure (1 ppm of chlorine) resulted
in determination of the optical response profile AE(x). The difference in responses, ΔA(x) = A0(x)
− AE(x), revealed the spatial location of the optimal concentration of the reagent that produced the
largest signal change (see Figure 3.24A). Sensing films with the optimized concentration of the cyanine
dye for chlorine determinations in industrial water were further screen-printed as a part of sensing ar-
rays (Potyrailo et al. 2007) onto conventional optical disks. The quantitative readout of changes in film
absorbance was performed in a conventional optical disk drive in a recently developed lab-on-a-disk
system (Potyrailo et al. 2006b, 2007).
This concentration-optimization method was also applied to optimize sensor material formulations
for analysis of gaseous species. Figure 3.24B shows optimization of concentration of Pt octaethylpor-
phyrin in a polystyrene film for detection of oxygen by fluorescence quenching. These data demonstrate
Figure 3.24. Optimization of formulated sensing materials using sensing films with gradient re-
agent concentration along the film length. (A) Concentration optimization of a colorimetric chlorine-
responsive reagent in a formulated polymeric poly(2-hydroxyethyl methacrylate) hydrogel sensing
film for detection of ions in water. Exposure, 1 ppm of chlorine. (B) Concentration optimization of
an oxygen-responsive Pt octaethylporphyrin fluorophore in polystyrene sensing film for detection of
oxygen in air.
the simplicity, yet tremendous value, of such determinations for rapid assessment of sensor film formu-
lations. One can see whether the optimal concentration has been reached or exceeded, depending on
the nonlinearity and decrease of the sensor response at the highest tested additive concentration. Unlike
traditional concentration optimization approaches (Papkovsky et al. 1995; Basu et al. 2005), this new
method provides a more dense evaluation mesh and opens opportunities for time-affordable optimiza-
tion of concentration of multiple formulation components with tertiary and higher gradients.
6.2. VARIABLE THICKNESS OF SENSING FILMS

The effect of thickness of sensing films on the stability of the response in water to ionic species has
been also evaluated using gradient-thickness sensing films (Potyrailo and Hassib 2005). Sensor reagent
stability in a polymer matrix upon water exposure is one of the key requirements. For deposition of
gradient sensor regions, several sensor coatings were flow-coated onto a 2.5-mm-thick polycarbon-
ate sheet. Typical coating dimensions were 1–1.5 cm wide and 10–15 cm long. To produce thickness
gradients, the coatings were positioned vertically until solvent evaporation in air at room temperature.
The coating thickness was further evaluated using optical absorbance or profilometry. An example of
a gradient-sensor coating array is shown in Figure 3.25A. The gradient thickness of sensing films was
determined from the absorbance of the film-incorporated bromothymol blue reagent. When these
arrays were further exposed to a pH 10 buffer (Figure 3.25B), an “activation” period was observed
before leaching of the reagent from the polymer matrix was detected from the absorbance decrease.
This activation period was roughly proportional to the film thickness. However, the leaching rate was
independent of the film thickness, as indicated by the similar slopes of the response curves at 3–9.5 h
exposure time.
Figure 3.25. Application of gradient-thickness sensor film arrays for evaluation of reagent leach-
ing kinetics. (A) Three gradient-thickness sensor film arrays with different loadings of an analyte-
sensitive indicator. (B) Reagent leaching kinetics at pH 10. (Reprinted with permission from Potyrailo
and Hassib 2005. Copyright 2005 American Institute of Physics.)
6.3. VARIABLE 2-D COMPOSITION

Sensors based on the change in the work function of the catalytic metal gate (Pd, Pt, Rh, Ir) due to
chemical reactions on the metal surface (Lundström et al. 1975, 1993) are attractive for detection of
various gases (hydrogen sulfide, ethylene, ethanol, various amines, and others). The chemical reaction
mechanisms in these sensors depend on the specific gas molecules. Optimization of materials for these
sensors involves several degrees of freedom (Eriksson et al. 2006). To simplify screening of the desired
material compositions and to reduce the common problem of batch-to-batch differences of hundreds
of individually made sensors for materials development, the scanning light pulse technique (SLPT) has
been developed (Lundström et al. 1991, 2007; Löfdahl et al. 2000). In the SLPT, a focused light beam is
scanned over a large-area, semitransparent, catalytic metal–insulator–semiconductor structure, and the
photocurrent generated in the semiconductor depletion region is measured to create a 2-D response
pattern of the sensing film (a.k.a. “chemical image”).
These chemical images were used to optimize properties such as chemical sensitivity, selectivity,
and stability (Klingvall et al. 2005). When combined with surface-characterization methods, this infor-
mation led to increased knowledge of gas response phenomena. It was suggested that a 2-D gradient
made from two types of metal films as a double-layer structure should provide new capabilities for
sensor materials optimization, capabilities that are not available with thickness gradients of single metal
films (Klingvall et al. 2003). To make a 2-D gradient, the first metal film was evaporated on the insulator
with the linear thickness variation in one dimension by moving a shutter with a constant speed in front
of the substrate during evaporation. On top of the first gradient-thickness film, a second metal film was
evaporated with a linear thickness variation perpendicular to that of the first film. As validation of the
2-D array deposition, the responses of devices with 1-D thickness gradients of Pd, Pt, and Ir films to
several gases have been studied with SLPT and demonstrate results similar to those of corresponding
discrete components (Klingvall et al. 2005).
The 2-D gradients have been used for studies and optimization of two-metal structures (Klingvall
et al. 2003, 2005) and for determination of the effects of insulator surface properties on the magnitude
of sensing response (Eriksson et al. 2005). Two-dimensional gradients of Pd/Rh film compositions
were also studied to identify materials compositions for the most stable performance (Klingvall et al.
2005). The Pd/Rh film compositions were tested for their response stability to 1000 ppm of hydrogen
upon aging for 24 h at 400°C while exposed to 250 ppm of hydrogen. This accelerated aging experi-
ment of the 2-D gradient film surface demonstrated the existence of two most stable local regions. One
region was a “valley” of stable response, shown as a dark color in Figure 3.26. The other region was a
thicker part of the two-component film, with a ~20-nm-thick Rh film and a ~23-nm-thick Pd film. This
new knowledge inspired new questions about the position stability of the “valley” and the possibility of
improving sensor stability by using an initial annealing process.
Figure 3.26. Results of accelerated aging of 2-D combinatorial library of Rh/Pd film. Chemical re-
sponse image to 1000 ppm of hydrogen as the differential response after and before the accelerated
aging; the most stable regions have the darkest color. (Reprinted with permission from Klingvall et al.
2005. Copyright 2005 The Institute of Electrical and Electronics Engineers, Inc.)
6.4. VARIABLE OPERATION TEMPERATURE AND DIFFUSION-

LAYER THICKNESS
At present, in conductometric sensors, semiconducting metal oxides are used as gas-sensing materials
that change their electrical resistance upon exposure to oxidizing or reducing gases. Over the years,
significant technological advances have been made that have resulted in practical and commercially
available sensors, and now new materials are being developed that further improve the performance
of these sensors. Realizing the opportunities that arise with the temperature dependence of the sensor
response, temperature gradient–based sensors that utilize a single metal-oxide thin film segmented by
electrodes have been developed (Goschnick et al. 1998, 2005; Arnold et al. 2002a, 2002b; Koronczi et
al. 2002; Schneider et al. 2004; Sysoev et al. 2004). In addition to the spatial temperature-gradient heater,
one design for the sensor chip also had a SiO2 or Al2O3 membrane with a gradient thickness from 2
to 50 nm (see Figure 3.27A). These ceramic membranes provided an additional response selectivity
(Goschnick 2001) through thickness-dependent gas transport.
To fabricate such a temperature- and membrane-gradient sensor, a gas-sensitive SnO2:Pt film (Pt
content of 0.8 at%) was deposited onto a thermally oxidized Si wafer by RF magnetron sputtering us-
ing a shadow mask. Next, Pt strip electrodes and two meander-shaped thermoresistors were sputtered
on the same side of the substrate as the SnO2 film, under a shadow mask for structuring the films. The
arrangement of the electrodes subdivided the monolithic SnO2 film into 38 sensor segments on an area
of 4 × 8 mm2. Finally, Pt heaters were deposited onto the back of the substrate to operate the chip with
a 50°C temperature gradient from 310 to 360°C (Sysoev et al. 2004). The application of a temperature
gradient increased the gas discrimination power of the sensor by 35%. The sensor with a SiO2 gradient-
thickness membrane was employed for detection of gaseous precursors of smoldering fires induced by
overheated cable insulation (see Figure 3.27B) (Arnold et al. 2002b).
7. EMERGING WIRELESS TECHNOLOGIES FOR COMBINATORIAL

SCREENING OF SENSING MATERIALS
At present, advances in wireless electronic technologies promise to add attractive new capabilities for
combinatorial screening of sensing materials. Wireless proximity chemical, biological, and physical sen-
sors provide several key advantages over wired sensors, such as (1) noncontact and noncontamination
measurements, (2) operation though packaging, and (3) rapid reading of multiple sensors with a single
reader. Several proximity-sensing approaches based on thickness shear-mode (TSM) and surface acous-
tic-wave (SAW) sensors connected to antennas as well as passive radio-frequency identification (RFID)
sensors have been developed (see Figure 3.28) with possible applications for combinatorial screening
of sensing materials.
To eliminate the direct wiring of individual TSM sensors and to permit materials evaluation in en-
vironments where wiring is not desirable or adds a prohibitively complex design, a wireless TSM sensor
array system (Potyrailo and Morris 2007b) was developed in which each sensor resonator was coupled
to a receiver antenna coil and an array of these coils was scanned with a transmitter coil (Figure 3.29A).
Using this wireless sensor system, sensing materials can be screened for their gas-sorption properties,
Figure 3.27. Double-gradient sensor microarray for selective gas detection. (A) Sensor schematic
illustrating a single metal-oxide thin film segmented by electrodes and arranged on a temperature-
gradient heater. The sensing film is further covered with a gradient-thickness ceramic membrane.
(B) Results of linear discrimination analysis of the signal patterns in practical tests to detect gaseous
precursors of smoldering fires induced by overheated cable insulation (ETFE, ethylene tetrafluorine
ethylene). [(A) Used with kind permission from Goschnick. (B) Reprinted with permission from Arnold
et al. 2002b. Copyright 2002 The Institute of Electrical and Electronics Engineers, Inc.]
analyte binding in liquids, and for changes in chemical and physical properties upon weathering and
aging tests. The applicability of the wireless sensor materials screening approach has been demon-
strated for the rapid evaluation of the effects of conditioning of polymeric sensing films at different
temperatures on the vapor-response patterns. In one set of high-throughput screening experiments,
Nafion film-aging effects on the selectivity pattern were studied. Evaluation of this and many other
polymeric sensing materials lacks the detailed studies on the change of the chemical selectivity patterns
as a function of temperature conditioning and aging. Conditioning of Nafion-coated resonators was

performed at 22, 90, and 125°C for 12 h. Temperature-conditioned sensing films were exposed to water
(H2O), ethanol (EtOH), and acetonitrile (ACN) vapors, all at concentrations (partial pressures) ranging
from 0 to 0.1 of the saturated vapor pressure P0, as shown in Figures 3.29B and 3.29C. It was found
that conditioning of sensing films at 125°C compared to room-temperature conditioning provided (1)
an improvement in linearity in response to EtOH and ACN vapors, (2) an increase in relative response
to ACN, and (3) a 10-fold increase of the contribution to principal component #2. The latter point
signifies an improvement in the discrimination ability between different vapors upon conditioning of
the sensing material at 125°C. This new knowledge will be critical in designing sensors for practical ap-
plications where the need exists to preserve sensor response selectivity over long exploitation time or
when there is temperature cycling for accelerated sensor-film recovery after vapor exposure.
Recently, ubiquitous and cost-effective passive RFID tags have been adapted for chemical sensing
(Potyrailo and Morris 2007a, Potyrailo et al. 2009b). By applying a sensing material onto the resonant
antenna of the RFID tag and measuring the complex impedance of the RFID resonant antenna, it
was possible to correlate impedance response to chemical properties of interest. A variety of sens-
ing materials was evaluated, including Nafion polymeric compositions formulated with plasticizers,
as shown in Figure 3.30. When a sensing film is deposited onto the resonant antenna (Figure 3.31A),
the analyte-induced changes in the dielectric and dimensional properties of this sensing film affect the
complex impedance of the antenna circuit through changes in film resistance and capacitance between
the antenna turns. Such changes provide selectivity in response of an individual RFID sensor and
provide the opportunity to replace a whole array of conventional sensors with a single RFID sensor
(Potyrailo and Morris 2007a). For this selective analyte quantitation using individual RFID sensors,
Figure 3.28. Examples of proximity wireless sensors applicable for combinatorial screening of sens-
ing materials: (A) TSM sensor resonating at 10 MHz; (B) SAW sensor resonating at 915 MHz; (C)
RFID sensor resonating at 13.56 MHz.
Figure 3.29. Concept for wireless high-throughput screening of materials properties using thickness
shear mode resonators. (A) Configuration of a wireless proximity resonant sensor array system for
high-throughput screening of sensing materials with a single transmitter coil that scans across an ar-
ray of receiver coils attached to resonant sensors. (B–D) Evaluation of selectivity of Nafion sensing
films to several vapors after conditioning at different temperatures: (B) 22°C and (C) 125°C. Vapors:
H2O (water), EtOH (ethanol), and ACN (acetonitrile). Concentrations of vapors are 0, 0.02, 0.04, 0.07,
and 0.10 P/P0. Arrows indicate the increase of concentrations of each vapor. (Reprinted with permis-
sion from Potyrailo and Morris 2007b. Copyright 2007 American Institute of Physics.)
complex impedance spectra of the resonant antenna are measured. Several parameters from the mea-
sured real and imaginary portions of the complex impedance are further calculated. These parameters
include Fp and Zp (the frequency and magnitude of the maximum of the real part of the complex
impedance, respectively) and F1 and F2 (the resonant and antiresonant frequencies of the imaginary
part of the complex impedance, respectively). By applying multivariate analysis of the full complex
impedance spectra or the calculated parameters, quantitation of analytes and rejection of interferences
is performed with individual RFID sensors.
Figure 3.30. Structures of plasticizers 15–19 employed for construction of vapor-sensing materials
based on Nafion polymer.
Because temperature effects are important at all stages of sensor fabrication, testing, and end use,
understanding of temperature effects can provide the ability to build robust temperature-corrected
transfer functions of sensor performance in order to preserve response sensitivity, response selectivity,
and response baseline stability. RFID sensors were applied for the combinatorial screening of sens-
ing materials to evaluate combined effects of plasticizers in polymeric formulated films and annealing
temperature (Potyrailo et al. 2009c). As a model system, a 6 × 8 array of polymer-coated RFID sensors
was constructed as shown in Figure 3.31B. A solid polymer electrolyte Nafion was formulated with five
types of phthalate plasticizers as shown in Figure 3.30, including dimethyl phthalate (15), butyl benzyl
phthalate (16), di-(2-ethylhexyl) phthalate (17), dicapryl phthalate (18), and diisotridecyl phthalate (19).
These sensing film formulations and control sensing films without a phthalate plasticizer were depos-
ited onto RFID sensors, exposed to eight temperatures ranging from 40 to 140°C using a gradient-
temperature heater, and evaluated for their response stability and gas-selectivity response patterns. The
interrogation of RFID sensors in the array was done with a single transmitter (pick-up antenna) coil
positioned on an X–Y translation stage and connected to a network analyzer.
To evaluate temperature effects on sensor response selectivity, the 6 × 8 array of temperature-
annealed sensing films was exposed to H2O and ACN vapors. Acetonitrile was selected as a simulant for
blood chemical warfare agents (CWAs) (Lee et al. 2005) and water vapor was selected as an interference.
In these experiments, the partial pressures of H2O and ACN vapors were 0.4 of the saturated vapor
pressure P0. Nafion sensing films were used previously for detection of humidity (Feng et al. 1997;
Tailoka et al. 2003) and organic vapors (Sun and Okada 2001). Conductance and dielectric properties
of Nafion have been shown to be vapor-dependent (Cappadonia et al. 1995; Wintersgill and Fontanella
1998). Figures 3.31C and 3.31D show representative ΔZp and ΔFp responses to H2O and ACN upon
annealing at two temperatures, 40 and 110°C. The patterns of ΔF1, ΔF2, ΔFp, and ΔZp responses of
sensing films to H2O and ACN vapors upon temperature annealing were further examined using prin-
cipal components analysis (PCA). From these screening experiments, it was found that different plasti-
cizers affect the response diversity to different extents; however, Nafion sensing films formulated with
Figure 3.31. Combinatorial screening of sensing film compositions using passive RFID sensors. (A) Strategy for adaptation of conven-
tional passive RFID tags for chemical sensing by deposition of a sensing film onto the resonant circuit of the RFID antenna. Inset, analyte-
induced changes in the film material affect film resistance (RF) and capacitance (CF) between the antenna turns. (B) Photo of an array
of 48 RFID sensors prepared for temperature-gradient evaluations of response of Nafion/phthalates compositions. (C and D) Examples
of RFID sensor response to water (H2O) and acetonitrile (ACN) vapors after annealing at different temperatures (40 and 110°C), (C) ΔZp
response and (D) ΔFp response. Nafion sensing film compositions: 1, control without plasticizer; 2, dimethyl phthalate; 3, butyl benzyl
phthalate; 4, di-(2-ethylhexyl) phthalate; 5, dicapryl phthalate; 6, diisotridecyl phthalate. (Color photographs can be seen in the original
paper.) (Reprinted with permission from Potyrailo et al. 2009c. Copyright 2009 American Chemical Society.)
the dimethyl phthalate plasticizer improve response diversity of the sensing films to H2O and ACN.
Overall, this study demonstrated that this RFID-based sensing approach permits rapid, cost-effec-
tive combinatorial screening of dielectric properties of sensing materials. As was pointed out earlier
(Potyrailo et al. 2003d), in general, an increase of the level of environmental stress may be problematic
if the correlation with traditional test methods is lost. To avoid this situation, it will be critical to plan
detailed accelerated-aging, high-throughput experiments with positive and negative controls.
8. SUMMARY AND OUTLOOK
Combinatorial technologies in materials science have been successfully developed by research groups
in academia and governmental laboratories that have overcome the entry barrier of dealing with new
and emerging aspects of materials research, such as automation and robotics, computer programming,
informatics, and materials data mining. The main driving forces for combinatorial materials science in
industry include broader and more detailed exploration of materials and process parameters and faster
time to market. Industrial research laboratories working on new catalysts and inorganic luminescent
materials were among the first adopters of combinatorial methodologies in industry. The classical ex-
ample, the effort of Mittasch, who spent 10 years (over 1900–1909) conducting 6500 screening experi-
ments with 2500 catalyst candidates to find a catalyst for industrial ammonia synthesis (Ertl, 1990) will
never happen again because of the availability and affordability of modern tools for high-throughput
synthesis and characterization.
In the area of sensing materials, reported examples of significant screening efforts are less dra-
matic yet also breathtaking. For example, a decade ago, Cammann, Shulga, and co-workers (Buhlmann
et al. 1998) reported an “extensive systematic study” of more than 500 compositions to optimize vapor-
sensing polymeric materials. Walt and co-workers (1998) reported screening over 100 polymer candi-
dates for “their ability to serve as sensing matrices” for solvatochromic reagents. Seitz and co-workers
(Conway et al. 1997) investigated the influence of multicomponent compositions on the properties of
pH-swellable polymers by designing 3 × 3 × 3 × 2 factorial experiments. Clearly, combinatorial tech-
nologies have been introduced at the right time to make the search for new materials more intellectually
rewarding. Naturally, numerous academic groups that were involved in the development of new sensing
materials turned to combinatorial methodologies to speed up knowledge discovery (Dickinson et al.
1997; Cho et al. 2002; Simon et al. 2002; Apostolidis et al. 2004; Frantzen et al. 2004; Mirsky et al. 2004;
Lundström et al. 2007).
Among numerous results achieved using combinatorial and high-throughput methods, the most
successful have been in the areas of molecular imprinting, polymeric compositions, catalytic metals for
field-effect devices, and metal oxides for conductometric sensors. In those materials, the desired selec-
tivity and sensitivity have been achieved by the exploration of multidimensional chemical composition
and process parameters at a previously unavailable level of detail and in a fraction of the time required
for conventional one-at-a-time experiments. These new tools have provided the opportunity for more
challenging, yet more rewarding, explorations that previously were too time-consuming to pursue.
Future advances in combinatorial development of sensing materials will be related to several key
remaining unmet needs that prevent researchers from having a complete combinatorial workflow. At
present, data management of the combinatorial workflow is perhaps the weakest link. However, over
the last several years, there have been a growing number of reports on data mining in sensing materials
(Villoslada and Takeuchi 2005; Frenzer et al. 2006; Mijangos et al. 2006; Potyrailo et al. 2006a). “Searching
for a needle in the haystack” was popular in the early days of combinatorial materials science (Jandeleit
et al. 1999; Jansen 2002). It has now been realized that screening the entire materials and process para-
meters space is still too costly and time prohibitive, even with the availability of existing tools. Instead,
designing high-throughput experiments to discover relevant descriptors will become more attractive.
A modern combinatorial scientist is acquiring skills as diverse as experimental planning, automated
synthesis, basics of high-throughput materials characterization, chemometrics, and data mining. These
new skills can be now obtained through the growing network of practitioners and through the new gen-
eration of scientists educated across the world in combinatorial methodologies. Combinatorial and high-
throughput experimentation have been able to bring together several previously disjoined disciplines
and to combine valuable complementary attributes from each of them into a new scientific approach.
9. ACKNOWLEDGMENTS
I gratefully acknowledge GE components for support of our combinatorial sensor research.
REFERENCES
Abraham M.H. (1993) Scales of solute hydrogen bonding: their construction and application to physicochemical
and biochemical processes. Chem. Soc. Rev. 22, 73–83.
Akporiaye D.E. (1998) Towards a rational synthesis of large-pore zeolite-type materials? Angew. Chem. Int. Ed. 37,
2456–2457.
Albert K.J., Lewis N.S., Schauer C.L., Sotzing G.A., Stitzel S.E., Vaid T.P., and Walt D.R. (2000) Cross-reactive
chemical sensor arrays. Chem. Rev. 100, 2595–2626.
Alivisatos, A.P. (2004) The use of nanocrystals in biological detection. Nature Biotechnol. 22, 47–52.
Amao Y. (2003) Probes and polymers for optical sensing of oxygen. Microchim. Acta 143, 1–12.
Amis E.J. (2004) Combinatorial materials science reaching beyond discovery. Nature Mater. 3, 83–85.
Anderson F.W. and Moser J.H. (1958) Automatic computer program for reduction of routine emission spectrogra-
phic data. Anal. Chem. 30, 879–881.
Anker J.N., Hall W.P., Lyandres O., Shah N.C., Zhao J., and Van Duyne R.P. (2008) Biosensing with plasmonic
nanosensors. Nature Mater. 7, 442–453.
Apostolidis A., Klimant I., Andrzejewski D., and Wolfbeis O.S. (2004) A combinatorial approach for development
of materials for optical sensing of gases. J. Comb. Chem. 6, 325–331.
Arnold C., Andlauer W., Häringer D., Körber R., and Goschnick J. (2002a) Gas analytical gradient microarrays com-
promising low price and high performance for intelligent consumer products. Proc. IEEE Sens. 1, 426–429.
Arnold C., Harms M., and Goschnick J. (2002b) Air quality monitoring and fire detection with the Karlsruhe elec-
tronic micronose KAMINA. IEEE Sens. J. 2, 179–188.
Aronova M.A., Chang K. S., Takeuchi I., Jabs H., Westerheim D., Gonzalez-Martin A., Kim J., and Lewis B.
(2003) Combinatorial libraries of semiconductor gas sensors as inorganic electronic noses. Appl. Phys. Lett. 83,
1255–1257.
Baker B.E., Kline N.J., Treado P.J., and Natan M.J. (1996) Solution-based assembly of metal surfaces by combina-
torial methods. J. Am. Chem. Soc. 118, 8721–8722.
Barbero C., Acevedo D.F., Salavagione H.J., and Miras M.C. (2003) Synthesis, properties and applications of func-
tionalized conductive polymers. Jornadas Sam/Conamet/Simposio Materia 2003, C-12.
Barsan N., Koziej D., and Weimar U. (2007) Metal oxide-based gas sensor research: how to? Sens. Actuators B 121,
18–35.
Barsoukov E. and Macdonald J.R. (Eds.) (2005) Impedance Spectroscopy: Theory, Experiment, and Applications. Wiley,
Hoboken, NJ.
Basu B.J., Thirumurugan A., Dinesh A.R., Anandan C., and Rajam K.S. (2005) Optical oxygen sensor coating based
on the fluorescence quenching of a new pyrene derivative. Sens. Actuators B 104, 15–22.
Bedlek-Anslow J.M., Hubner J.P., Carroll B.F., and Schanze K.S. (2000) Micro-heterogeneous oxygen response in
luminescence sensor films. Langmuir 16, 9137–9141.
Beebe K.R., Pell R.J., and Seasholtz M. B. (1998) Chemometrics: A Practical Guide. Wiley, New York.
Belmares M., Blanco M., Goddard W.A. III, Ross R.B., Caldwell G., Chou S.-H., Pham J., Olofson P.M., and
Thomas C. (2004) Hildebrand and Hansen solubility parameters from molecular dynamics with applications to
electronic nose polymer sensors. J. Comput. Chem. 25, 1814–1826.
Bhat R.R., Tomlinson M.R., and Genzer J. (2004) Assembly of nanoparticles using surface-grafted orthogonal
polymer gradients. Macromol. Rapid Commun. 25, 270–274.
Bidan G. (1992) Electroconducting conjugated polymers: new sensitive matrices to build up chemical or electroche-
mical sensors. A review. Sens. Actuators B 6, 45–56.
Birina G.A. and Boitsov K.A. (1974) Experimental use of combinational and factorial plans for optimizing the
compositions of electronic materials. Zavodskaya Laboratoriya (in Russian) 40, 855–857.
Bobacka J., Ivaska A., and Lewenstam A. (2003) Potentiometric ion sensors based on conducting polymers.
Electroanalysis 15, 366–374.
Briceño G., Chang H., Sun X., Schultz P.G., and Xiang X.-D. (1995) A class of cobalt oxide magnetoresistance
materials discovered with combinatorial synthesis. Science 270, 273–275.
Brocchini S., James K., Tangpasuthadol V., and Kohn J. (1997) A combinatorial approach for polymer design. J.
Am. Chem. Soc. 119, 4553–4554.
Buhlmann K., Schlatt B., Cammann K., and Shulga A. (1998) Plasticised polymeric electrolytes: new extremely versatile
receptor materials for gas sensors (VOC monitoring) and electronic noses (odour identification:discrimination).
Bühlmann P., Pretsch E., and Bakker E. (1998) Carrier-based ion-selective electrodes and bulk optodes. 2.
Ionophores for potentiometric and optical sensors. Chem. Rev. 98, 1593–1687.
Calvert P. (2001) Inkjet printing for materials and devices. Chem. Mater. 13, 3299–3305.
Cappadonia M., Erning J.W., Niaki S.M.S., and Stimming U. (1995) Conductance of Nafion 117 membranes as a
function of temperature and water content. Solid State Ionics 77, 65–69.
Cawse J.N., Olson D., Chisholm B.J., Brennan M., Sun T., Flanagan W., Akhave J., Mehrabi A., and Saunders D.
(2003) Combinatorial chemistry methods for coating development V: Generating a combinatorial array of
uniform coatings samples. Prog. Org. Coat., 47, 128–135.
Chang H., Gao C., Takeuchi I., Yoo Y., Wang J., Schultz P.G., Xiang X.-D., Sharma R.P., Downes M., and Venkatesan
T. (1998) Combinatorial synthesis and high throughput evaluation of ferroelectric/dielectric thin-film libraries
for microwave applications. Appl. Phys. Lett. 72, 2185–2187.
Chen H.M., Pang L., Kher A., and Fainman Y. (2009) Three-dimensional composite metallodielectric nanostructure
for enhanced surface plasmon resonance sensing. Appl. Phys. Lett. 94, 073117.
Chisholm B.J., Potyrailo R.A., Cawse J.N., Shaffer R.E., Brennan M.J., Moison C., Whisenhunt D.W., Flanagan W.P.,
Olson D.R., Akhave J.R., Saunders D.L., Mehrabi A., and Licon M. (2002) The development of combinato-
rial chemistry methods for coating development I. Overview of the experimental factory. Prog. Org. Coat. 45,
313–321.
Cho E.J., Tao Z., Tang Y., Tehan E.C., Bright F.V., Hicks W.L. Jr., Gardella J.A. Jr., and Hard R. (2002) Tools to
rapidly produce and screen biodegradable polymer and sol-gel-derived xerogel formulations. Appl. Spectrosc.
56, 1385–1389.
Chojnacki P., Werner T., and Wolfbeis O. S. (2004) Combinatorial approach towards materials for optical ion sen-
sors. Microchim. Acta 147, 87–92.
Christensen D. and Fowers D. (1996) Modeling SPR sensors with the finite-difference time-domain method. Biosens.
Bioelectron. 11, 677–684.
Clark K.J.R. and Furey J. (2006) Suitability of selected single-use process monitoring and control technology.
BioProcess Intern. 4(6), S16–S20.
Collaudin A.B. and Blum L.J. (1997) Enhanced luminescence response of a fibre-optic sensor for H2O2 by a high-
salt-concentration medium. Sens. Actuators B 38–39, 189–194.
Conway V.L., Hassen K.P., Zhang L., Seitz W.R., and Gross T.S. (1997) The influence of composition on the pro-
perties of pH-swellable polymers for chemical sensors. Sens. Actuators B 45, 1–9.
Cooper A.C., McAlexander L.H., Lee D.-H., Torres M.T., and Crabtree R.H. (1998) Reactive dyes as a method for
rapid screening of homogeneous catalysts. J. Am. Chem. Soc. 120, 9971–9972.
Cui J., Chu Y. S., Famodu O. O., Furuya Y., Hattrick-Simpers J., James R.D., Ludwig A., Thienhaus S., Wuttig M.,
Zhang Z., and Takeuchi I. (2006) Combinatorial search of thermoelastic shape-memory alloys with extremely
small hysteresis width. Nature Mater. 5, 286–290.
Dai L., Soundarrajan P., and Kim T. (2002) Sensors and sensor arrays based on conjugated polymers and carbon
nanotubes. Pure Appl. Chem. 74, 1753–1772.
Danielson E., Devenney M., Giaquinta D.M., Golden J.H., Haushalter R.C., McFarland E.W., Poojary D.M., Reaves
C.M., Weinberg W.H., and Wu X.D. (1998) A rare-earth phosphor containing one-dimensional chains identified
through combinatorial methods. Science 279, 837–839.
De Gans B.-J. and Schubert U.S. (2004) Inkjet printing of well-defined polymer dots and arrays. Langmuir 20,
7789–7793.
Department of Defense (2005) Technology Readiness Assessment (TRA) Deskbook, May 2005, Department of Defense,
Prepared by the Deputy Under Secretary of Defense for Science and Technology (DUSD(S&T), https://acc.
dau.mil/CommunityBrowser.aspx?id=18545.
Dickinson T.A., Walt D.R., White J., and Kauer J.S. (1997) Generating sensor diversity through combinatorial poly-
mer synthesis. Anal. Chem. 69, 3413–3418.
Draxler S., Lippitsch M.E., Klimant I., Kraus H., and Wolfbeis O.S. (1995) Effects of polymer matrices on the time-
resolved luminescence of a ruthenium complex quenched by oxygen. J. Phys. Chem. 99, 3162–3167.
Eash M.A. and Gohlke R.S. (1962) Mass spectrometric analysis. A small computer program for the analysis of mass
spectra. Anal. Chem. 34, 713–713.
Eaton K. (2002) A novel colorimetric oxygen sensor: dye redox chemistry in a thin polymer film. Sens. Actuators B
85, 42–51.
Eberhart M.E. and Clougherty D.P. (2004) Looking for design in materials design. Nature Mater. 3, 659–661.
Ekin A. and Webster D.C. (2007) Combinatorial and high-throughput screening of the effect of siloxane composi-
tion on the surface properties of crosslinked siloxane-polyurethane coatings. J. Comb. Chem. 9, 178–188.
Eriksson M., Klingvall R., and Lundström I. (2006) A combinatorial method for optimization of materials for
gas sensitive field-effect devices. In: Combinatorial and High-Throughput Discovery and Optimization of Catalysts and
Materials, Potyrailo R.A. and Maier W.F. (eds.). CRC Press, Boca Raton, FL, pp. 85–95.
Eriksson M., Salomonsson A., Lundström I., Briand D., and Åbom A.E. (2005) The influence of the insulator
surface properties on the hydrogen response of field-effect gas sensors. J. Appl. Phys. 98, 034903.
Ertl G. (1990) Elementary steps in heterogeneous catalysis. Angew. Chem. Int. Ed. 29, 1219–1227.
Feldmann C. (2001) Polyol mediated synthesis of oxide particle suspensions and their application. Scripta Mater.
44, 2193–2196.
Feng C.-D., Sun S.-L., Wang H., Segre C.U., and Stetter J.R. (1997) Humidity sensing properties of Nafion and sol-
gel derived SiO2/Nafion composite thin films. Sens. Actuators B 40, 217–222.
Florescu M. and Katerkamp A. (2004) Optimisation of a polymer membrane used in optical oxygen sensing. Sens.
Franke M.E., Koplin T.J., and Simon U. (2006) Metal and metal oxide nanoparticles in chemiresistors: does the
nanoscale matter? Small 2, 36–50.
Frantzen A., Sanders D., Scheidtmann J., Simon U., and Maier W.F. (2005) A flexible database for combinatorial and
high-throughput materials science. QSAR & Comb. Sci. 24, 22–28.
Frantzen A., Scheidtmann J., Frenzer G., Maier W.F., Jockel J., Brinz T., Sanders D., and Simon U. (2004) High-
throughput method for the impedance spectroscopic characterization of resistive gas sensors. Angew. Chem.
Int. Ed. 43, 752–754.
Frenzer G., Frantzen A., Sanders D., Simon U., and Maier W.F. (2006) Wet chemical synthesis and screening of thick
porous oxide films for resistive gas sensing applications. Sensors 6, 1568–1586.
Genet C. and Ebbesen T.W. (2007) Light in tiny holes. Nature 445, 39–46.
Gill I. (2001) Bio-doped nanocomposite polymers: sol-gel bioencapsulates. Chem. Mater. 13, 3404–3421.
Gill I. and Ballesteros A. (1998) Encapsulation of biologicals within silicate, siloxane, and hybrid sol-gel polymers:
an efficient and generic approach. J. Am. Chem. Soc. 120, 8587–8598.
Gomez-Romero P. (2001) Hybrid organic-inorganic materials—in search of synergic activity. Adv. Mater. 13, 163–174.
Göpel W. (1998) Chemical imaging: I. Concepts and visions for electronic and bioelectronic noses. Sens. Actuators
B 52, 125–142.
Goschnick J. (2001) An electronic nose for intelligent consumer products based on a gas analytical gradient micro-
array. Microelectron. Eng. 57–58, 693–704.
Goschnick J., Frietsch M., and Schneider T. (1998) Non-uniform SiO2 membranes produced by ion beam-assisted
chemical vapor deposition to tune WO3 gas sensor microarrays. Surf. Coat. Technol. 108–109, 292–296.
Goschnick J., Koronczi I., Frietsch M., and Kiselev I. (2005) Water pollution recognition with the electronic nose
KAMINA. Sens. Actuators B 106, 182–186.
Grate J.W. (2000) Acoustic wave microsensor arrays for vapor sensing. Chem. Rev. 100, 2627–2648.
Grate J.W. and Abraham M.H. (1991) Solubility interactions and the design of chemically selective sorbent coatings
for chemical sensors and arrays. Sens. Actuators B 3, 85–111.
Grate J.W., Abraham H., and McGill R.A. (1997) Sorbent polymer materials for chemical sensors and arrays. In:
Handbook of Biosensors and Electronic Noses. Medicine, Food, and the Environment, Kress-Rogers E. (ed.). CRC Press,
Boca Raton, FL, pp. 593–612.
Greeley J., Jaramillo T.F., Bonde J., Chorkendorff I., and Nørskov J.K. (2006) Computational high-throughput
screening of electrocatalytic materials for hydrogen evolution. Nature Mater. 5, 909–913.
Grunlan J.C., Mehrabi A.R., Chavira A.T., Nugent A.B., and Saunders D.L. (2003) Method for combinatorial scree-
ning of moisture vapor transmission rate. J. Comb. Chem. 5, 362–368.
Hagleitner C., Hierlemann A., Brand O., and Baltes H. (2002) CMOS single chip gas detection systems—Part I. In:
Sensors Update, Vol. 11, Baltes H., Göpel W., and Hesse J. (eds.). VCH, Weinheim, pp. 101–155.
Hagleitner C., Hierlemann A., Brand O., and Baltes H. (2003) CMOS single chip gas detection systems—Part II. In:
Sensors Update, Vol. 12, Baltes H., Göpel W., and Hesse J. (eds.). VCH, Weinheim, pp. 51–120.
Hanak J.J. (1970) The “Multiple-Sample Concept” in materials research: synthesis, compositional analysis and te-
sting of entire multicomponent systems. J. Mater. Sci. 5, 964–971.
Hänsel H., Zettl H., Krausch G., Schmitz C., Kisselev R., Thelakkat M., and Schmidt H.-W. (2002) Combinatorial
study of the long-term stability of organic thin-film solar cells. Appl. Phys. Lett. 81, 2106–2108.
Hartmann P. and Trettnak W. (1996) Effects of polymer matrices on calibration functions of luminescent oxygen
sensors based on porphyrin ketone complexes. Anal. Chem. 68, 2615–2620.
Hassib L. and Potyrailo R.A. (2004) Combinatorial development of polymer coating formulations for chemical
sensor applications. Polymer Preprints 45, 211–212.
Hatchett D.W. and Josowicz M. (2008) Composites of intrinsically conducting polymers as sensing nanomaterials.
Chem. Rev. 108, 746–769.
Hierlemann A., Weimar U., Kraus G., Schweizer-Berberich M., and Göpel W. (1995) Polymer-based sensor ar-
rays and multicomponent analysis for the detection of hazardous organic vapours in the environment. Sens.
Hoffmann R. (2001) Not a library. Angew. Chem. Int. Ed. 40, 3337–3340.
Holzwarth A., Schmidt H.-W., and Maier W. (1998) Detection of catalytic activity in combinatorial libraries of
heterogeneous catalysts by IR thermography. Angew. Chem. Int. Ed. 37, 2644–2647.
Honeybourne C.L. (2000) Organic vapor sensors for food quality assessment. J. Chem. Educ. 77, 338–344.
Hoogenboom R., Meier M.A.R., and Schubert U.S. (2003) Combinatorial methods, automated synthesis and high-
throughput screening in polymer research: past and present. Macromol. Rapid Commun. 24, 15–32.
Hu Y., Tan O.K., Pan J.S., and Yao X. (2004) A new form of nanosized SrTiO3 material for near-human-body
temperature oxygen sensing applications. J. Phys. Chem. B 108, 11214–11218.
Janata J. and Josowicz M. (2002) Conducting polymers in electronic chemical sensors. Nature Mater. 2, 19–24.
Jandeleit B., Schaefer D.J., Powers T.S., Turner H.W., and Weinberg W.H. (1999) Combinatorial materials science
and catalysis. Angew. Chem. Int. Ed. 38, 2494–2532.
Jansen M. (2002) A concept for synthesis planning in solid-state chemistry. Angew. Chem. Int. Ed. 41, 3746–3766.
Kahl M., Voges E., Kostrewa S., Viets C., and Hill W. (1998) Periodically structured metallic substrates for SERS.
Kennedy K., Stefansky T., Davy G., Zackay V.F., and Parker E.R. (1965) Rapid method for determining ternary-
alloy phase diagrams. J. Appl. Phys. 36, 3808–3810.
Klingvall R., Lundstrom I., Lofdahl M., and Eriksson M. (2003) 2D-evaluation of the gas response of a RhPd-MIS
device. Proc. IEEE Sensors 2, 1114–1115.
Klingvall R., Lundström I., Löfdahl M., and Eriksson M. (2005) A combinatorial approach for field-effect gas sen-
sor research and development. IEEE Sensors J. 5, 995–1003.
Ko M.C. and Meyer G.J. (1999) Photoluminescence of inorganic semiconductors for chemical sensor applications.
In: Optoelectronic Properties of Inorganic Compounds, Roundhill D.M. and Fackler J.P. Jr. (eds.). Plenum Press, New
York, pp. 269–315. Koinuma H. and Takeuchi I. (2004) Combinatorial solid state chemistry of inorganic ma-
terials. Nature Mater. 3, 429–438.
Kolytcheva N.V., Müller H., and Marstalerz J. (1999) Influence of the organic matrix on the properties of membra-
ne coated ion sensor field-effect transistors. Sens. Actuators B 58, 456–463.
Koronczi I., Ziegler K., Kruger U., and Goschnick J. (2002) Medical diagnosis with the gradient microarray of the
KAMINA. IEEE Sensors J. 2, 254–259.
Korotcenkov G. (2005) Gas response control through structural and chemical modification of metal oxide films:
state of the art and approaches. Sens. Actuators B 107, 209–232.
Krishnan A., Thio T., Kim T.J., Lezec H.J., Ebbesen T.W., Wolff P.A., Pendry J., Martín-Moreno L., and Garcia-Vidal
F.J. (2001) Evanescently coupled resonance in surface plasmon enhanced transmission. Opt. Commun. 200, 1–7.
Kulikov V. and Mirsky V.M. (2004) Equipment for combinatorial electrochemical polymerization and high-through-
put investigation of electrical properties of the synthesized polymers. Meas. Sci. Technol. 15, 49–54.
Lai R., Kang B.S., and Gavalas G.R. (2001) Parallel synthesis of ZSM-5 zeolite films from clear organic-free soluti-
ons. Angew. Chem., Int. Ed. 40, 408–411.
Lavigne J.J. and Anslyn E.V. (2001) Sensing a paradigm shift in the field of molecular recognition: from selective to
differential receptors. Angew. Chem. Int. Ed. 40, 3119–3130.
Leach A.M. and Potyrailo R.A. (2006) Gas sensor materials based on semiconductor nanocrystal/polymer compo-
site films. In: Combinatorial Methods and Informatics in Materials Science, MRS Symposium Proceedings, Vol. 894, Wang
Q., Potyrailo R.A., Fasolka M., Chikyow T., Schubert U.S., and Korkin A. (eds.). Materials Research Society,
Warrendale, PA, pp. 237–243.
Leclerc M. (1999) Optical and electrochemical transducers based on functionalized conjugated polymers. Adv.
Mater. 11, 1491–1498.
Lee W.S., Lee S.C., Lee S.J., Lee D.D., Huh J.S., Jun H.K., and Kim J.C. (2005) The sensing behavior of SnO2-based
thick-film gas sensors at a low concentration of chemical agent simulants. Sens. Actuators B 108, 148–153.
Lee K.-L., Lee C.-W., Wang W.-S., and Wei P.-K. (2007) Sensitive biosensor array using surface plasmon resonance
on metallic nanoslits. J. Biomed. Opt. 12, art. no. 044023.
Lee K.-L., Wang W.-S., and Wei P.-K. (2008a) Comparisons of surface plasmon sensitivities in periodic gold na-
nostructures. Plasmonics 3, 119–125.
Lee K.-L., Wang W.-S., and Wei P.-K. (2008b) Sensitive label-free biosensors by using gap plasmons in gold nanos-
lits. Biosens. Bioelectron. 24, 210–215.
Legin A., Makarychev-Mikhailov S., Goryacheva O., Kirsanov D., and Vlasov Y. (2002) Cross-sensitive chemical
sensors based on tetraphenylporphyrin and phthalocyanine. Anal. Chim. Acta 457, 297–303.
Lemmo A.V., Fisher J.T., Geysen H.M., and Rose D.J. (1997) Characterization of an inkjet chemical microdispenser
for combinatorial library synthesis. Anal. Chem. 69, 543–551.
Lemmon J.P., Wroczynski R.J., Whisenhunt D.W. Jr., and Flanagan W.P. (2001) High throughput strategies for mo-
nomer and polymer synthesis and characterization. Polymer Preprints 42, 630–631.
Levitsky I., Krivoshlykov S.G., and Grate J.W. (2001) Rational design of a nile red/polymer composite film for
fluorescence sensing of organophosphonate vapors using hydrogen bond acidic polymers. Anal. Chem. 73,
3441–3448.
Lim S.-H., Raorane D., Satyanarayana S., and Majumdar A. (2006) Nano-chemo-mechanical sensor array platform
for high-throughput chemical analysis. Sens. Actuators B 119, 466–474.
Lisensky G.C., Meyer G.J., and Ellis A.B. (1988) Selective detector for gas chromatography based on adduct-
modulated semiconductor photoluminescence. Anal. Chem., 60, 2531–2534.
Löfdahl M., Eriksson M., and Lundström I. (2000) Chemical images. Sens. Actuators B 70, 77–82.
Lundström I., Shivaraman S., Svensson C., and Lundkvist L. (1975) A hydrogen-sensitive MOS field-effect transi-
stor. Appl. Phys. Lett. 26, 55–57.
Lundström I., Erlandsson R., Frykman U., Hedborg E., Spetz A., Sundgren H., Welin S., and Winquist F. (1991)
Artificial ‘olfactory’ images from a chemical sensor using a light pulse technique. Nature 352, 47–50.
Lundström I., Sundgren H., Winquist F., Eriksson M., Krantz-Rülcker C., and Lloyd-Spetz A. (2007) Twenty-five
years of field effect gas sensor research in Linköping. Sens. Actuators B 121, 247–262.
Lundström I., Svensson C., Spetz A., Sundgren H., and Winquist F. (1993) From hydrogen sensors to olfactory
images—twenty yerars with catalytic field-effect devices. Sens. Actuators B 13–14, 16–23.
MacLean D., Baldwin J.J., Ivanov V.T., Kato Y., Shaw A., Schneider P., and Gordon E.M. (2000) Glossary of terms
used in combinatorial chemistry. J. Comb. Chem. 2, 562–578.
Maranas C.D. (1996) Optimal computer-aided molecular design: a polymer design case study. Ind. Eng. Chem. Res.
35, 3403–3414.
Martin P.D., Wilson T.D., Wilson I.D., and Jones G.R. (2001) An unexpected selectivity of a propranolol-derived
molecular imprint for tamoxifen. Analyst 126, 757–759.
Martín-Moreno L., García-Vidal F.J., Lezec H.J., Pellerin K.M., Thio T., Pendry J.B., and Ebbesen T.W. (2001) Theory
of extraordinary optical transmission through subwavelength hole arrays. Phys. Rev. Lett. 86, 1114–1117.
Matteo J.A., Fromm D.P., Yuen Y., Schuck P.J., Moerner W.E., and Hesselink L. (2004) Spectral analysis of strongly
enhanced visible light transmission through single C-shaped nanoapertures. Appl. Phys. Lett. 85, 648–650.
McKusick B.C., Heckert R.E., Cairns T.L., Coffman D.D., and Mower H.F. (1958) Cyanocarbon chemistry. VI.
Tricyanovinylamines. J. Am. Chem. Soc. 80, 2806–2815.
McMahon J.M., Henzie J., Odom T.W., Schatz G.C., and Gray S.K. (2007) Tailoring the sensing capabilities
of nanohole arrays in gold films with Rayleigh anomaly-surface plasmon polaritons. Opt. Express 15,
18119–18129.
McQuade D.T., Pullen A.E., and Swager T.M. (2000) Conjugated polymer-based chemical sensors. Chem. Rev. 100,
2537–2574.
Medintz I.L., Uyeda H.T., Goldman E.R., and Mattoussi H. (2005) Quantum dot bioconjugates for imaging, label-
ling and sensing. Nature Mater. 4, 435–446.
Meyerhoff M.E. (1993) In vivo blood-gas and electrolyte sensors: progress and challenges. Trends Anal. Chem. 12,
257–266.
Mijangos I., Navarro-Villoslada F., Guerreiro A., Piletska E., Chianella I., Karim K., Turner A., and Piletsky S.
(2006) Influence of initiator and different polymerisation conditions on performance of molecularly imprinted
polymers. Biosens. Bioelectron. 22, 381–387.
Mills A. (1998) Controlling the sensitivity of optical oxygen sensors. Sens. Actuators B 51, 60–68.
Mills A., Lepre A., and Wild L. (1998) Effect of plasticizer-polymer compatibility on the response characteristics
of optical thin CO2 and O2 sensing films. Anal. Chim. Acta 362, 193–202.
Mirsky V.M. and Kulikov V. (2003) Combinatorial electropolymerization: concept, equipment and applications. In:
High Throughput Analysis: A Tool for Combinatorial Materials Science, Potyrailo R.A. and Amis E.J. (eds.). Kluwer
Academic/Plenum, New York, pp.431–446.
Mirsky V.M., Kulikov V., Hao Q., and Wolfbeis O.S. (2004) Multiparameter high throughput characterization of
combinatorial chemical microarrays of chemosensitive polymers. Macromol. Rapid Commun. 25, 253–258.
Mohr G.J. and Wolfbeis O.S. (1996) Effects of the polymer matrix on an optical nitrate sensor based on a polarity-
sensitive dye. Sens. Actuators B 37, 103–109.
Montgomery J.M., Lee T.-W., and Gray S.K. (2008) Theory and modeling of light interactions with metallic na-
nostructures. J. Phys.: Condens. Matter. 20, art. no. 323201.
Nazzal A.Y., Qu L., Peng X., and Xiao M. (2003) Photoactivated CdSe nanocrystals as nanosensors for gases. Nano
Lett. 3, 819–822.
Newman J.D. and Turner A.P.F. (2005) Home blood glucose biosensors: a commercial perspective. Biosens. Bioelectron.
20, 2435–2453.
Newnham R.E. (1988) Structure-property relationships in sensors. Cryst. Rev., 1, 253–280.
Njagi J., Warner J., and Andreescu S. (2007) A bioanalytical chemistry experiment for undergraduate students: bio-
sensors based on metal nanoparticles. J. Chem. Educ. 84, 1180–1182.
Olk C.H. (2005) Combinatorial approach to material synthesis and screening of hydrogen storage alloys. Meas. Sci.
Technol. 16, 14–20.
Otto M. (1999) Chemometrics: Statistics and Computer Application in Analytical Chemistry. Wiley-VCH, Weinheim,
Germany.
Papkovsky D.B., Ponomarev G.V., Trettnak W., and O’Leary P. (1995) Phosphorescent complexes of porphyrin
ketones: optical properties and applications to oxygen sensing. Anal. Chem. 67, 4112–4117.
Pedersen C.J. (1967) Cyclic polyethers and their complexes with metal salts. J. Am. Chem. Soc. 89, 7017–7036.
Penco M., Sartore L., Bignotti F., Sciucca S.D., Ferrari V., Crescini P., and D’Antone S. (2004) Use of compatible
blends to fabricate carbon black composite vapor detectors. J. Appl. Polymer Sci. 91, 1816–1821.
Peper S., Ceresa A., Qin Y., and Bakker E. (2003) Plasticizer-free microspheres for ionophore-based sensing and ex-
traction based on a methyl methacrylate-decyl methacrylate copolymer matrix. Anal. Chim. Acta 500, 127–136.
Pickup J.C. and Alcock S. (1991) Clinicians’ requirements for chemical sensors for in vivo monitoring: a multinatio-
nal survey. Biosens. Bioelectron. 6, 639–646.
Potyrailo R.A. (2001) Combinatorial screening. In: Encyclopedia of Materials: Science and Technology, Vol. 2, Buschow
K.H.J., Cahn R.W., Flemings M.C., Ilschner B., Kramer E.J., and Mahajan S. (eds.). Elsevier, Amsterdam, pp.
1329–1343.
Potyrailo R.A. (2004c) Expanding combinatorial methods from automotive to sensor coatings. Polymeric Mater Sci.
Eng. Polymer Preprints 90, 797–798.
Potyrailo R.A. and Amis E.J. (eds.) (2003b) High Throughput Analysis: A Tool for Combinatorial Materials Science. Kluwer
Academic/Plenum, New York.
Potyrailo R.A. and Brennan M.J. (2004b) Method and apparatus for characterizing the barrier properties of mem-
bers of combinatorial libraries, U.S. Patent 6,684,683(B2).
Potyrailo R.A. and Hassib L. (2005) Analytical instrumentation infrastructure for combinatorial and high-through-
put development of formulated discrete and gradient polymeric sensor materials arrays. Rev. Sci. Instrum. 76,
062225.
Potyrailo R.A. and Leach A.M. (2006) Selective gas nanosensors with multisize CdSe nanocrystal/polymer compo-
site films and dynamic pattern recognition. Appl. Phys. Lett., 88, 134110.
Potyrailo R.A. and Maier W.F. (eds.) (2006) Combinatorial and High-Throughput Discovery and Optimization of Catalysts
and Materials. CRC Press, Boca Raton, FL.
Potyrailo R.A. and Mirsky V.M. (2008) Combinatorial and high-throughput development of sensing materials: the
first ten years. Chem. Rev. 108, 770–813.
Potyrailo R.A. and Mirsky V.M. (2009) Introduction to combinatorial methods for chemical and biological sensors.
In: Combinatorial Methods for Chemical and Biological Sensors, Potyrailo R.A. and Mirsky V.M. (eds.). Springer-Verlag,
New York, pp. 3–24.
Potyrailo R.A. and Morris W.G. (2007a) Multianalyte chemical identification and quantitation using a single radio
frequency identification sensor. Anal. Chem. 79, 45–51.
Potyrailo R.A. and Morris W.G. (2007b) Wireless resonant sensor array for high-throughput screening of materials.
Rev. Sci. Instrum. 78, 072214.
Potyrailo R.A. and Sivavec T.M. (2004) Boosting sensitivity of organic vapor detection with silicone block polyimi-
de polymers. Anal. Chem. 76, 7023–7027.
Potyrailo R.A. and Takeuchi I. (2005a) Role of high-throughput characterization tools in combinatorial materials
science. Meas. Sci. Technol. 16, 1–4.
Potyrailo R.A. and Takeuchi I. (eds.) (2005b) Special feature on combinatorial and high-throughput materials re-
search. Meas. Sci. Technol. 16, 1–316.
Potyrailo, R.A. and Wroczynski, R.J. (2005) Spectroscopic and imaging approaches for evaluation of properties of
one-dimensional arrays of formulated polymeric materials fabricated in a combinatorial microextruder system.
Rev. Sci. Instrum. 76, 062222.
Potyrailo R.A., Sivavec T.M., and Bracco A.A. (1999) Field evaluation of acoustic wave chemical sensors for moni-
toring of organic solvents in groundwater. Proc. SPIE, 3856, 140–147.
Potyrailo R.A., Chisholm B.J., Olson D.R., Brennan M.J., and Molaison C.A. (2002) Development of combinatorial
chemistry methods for coatings: high-throughput screening of abrasion resistance of coatings libraries. Anal.
Chem. 74, 5105–5111.
Potyrailo R.A., Chisholm B.J., Morris W.G., Cawse J.N., Flanagan W.P., Hassib L., Molaison C.A., Ezbiansky K.,
Medford G., and Reitz H. (2003a) Development of combinatorial chemistry methods for coatings: high-
throughput adhesion evaluation and scale-up of combinatorial leads. J. Comb. Chem. 5, 472–478.
Potyrailo R.A., Morris W.G., and Wroczynski R.J. (2003b) Acoustic-wave sensors for high-throughput screening
of materials. In: High Throughput Analysis: A Tool for Combinatorial Materials Science, Potyrailo R.A. and Amis E.J.
(eds.). Kluwer Academic/Plenum, New York, pp. 219–246.
Potyrailo R.A., Olson D.R., Brennan M.J., Akhave J.R., Licon M.A., Mehrabi A.R., Saunders D.L., and Chisholm B.J.
(2003c) Systems and methods for the deposition and curing of coating compositions. U.S. Patent 6,544,334 B1.
Potyrailo R.A., Wroczynski R.J., Pickett J.E., and Rubinsztajn M. (2003d) High-throughput fabrication, perfor-
mance testing, and characterization of one-dimensional libraries of polymeric compositions. Macromol. Rapid
Commun. 24, 123–130.
Potyrailo R.A., Karim A., Wang Q., and Chikyow T. (eds.) (2004a) Combinatorial and Artificial Intelligence Methods in
Materials Science II. Materials Research Society, Warrendale, PA.
Potyrailo R.A., May R.J., and Sivavec T.M. (2004b) Recognition and quantification of perchloroethylene, trichlo-
roethylene, vinyl chloride, and three isomers of dichloroethylene using acoustic-wave sensor array. Sensor Lett.
2, 31–36.
Potyrailo R.A., Morris W.G., and Wroczynski R.J. (2004c) Multifunctional sensor system for high-throughput prima-
ry, secondary, and tertiary screening of combinatorially developed materials. Rev. Sci. Instrum. 75, 2177–2186.
Potyrailo R.A., Morris W.G., Wroczynski R.J., and McCloskey P.J. (2004d) Resonant multisensor system for high-
throughput determinations of solvent-polymer interactions. J. Comb. Chem. 6, 869–873.
Potyrailo R.A., McCloskey P.J., Ramesh N., and Surman C.M. (2005a) Sensor devices containing co-polymer sub-
strates for analysis of chemical and biological species in water and air. U.S. Patent Application 2005133697.
Potyrailo R.A., Szumlas A.W., Danielson T.L., Johnson M., and Hieftje G.M. (2005b) A dual-parameter optical
sensor fabricated by gradient axial doping of an optical fibre. Meas. Sci. Technol. 16, 235–241.
Potyrailo R.A., McCloskey P.J., Wroczynski R.J., and Morris W.G. (2006a) High-throughput determination of quan-
titative structure-property relationships using resonant multisensor system: solvent-resistance of bisphenol A
polycarbonate copolymers. Anal. Chem., 78, 3090-3096.
Potyrailo R.A., Morris W.G., Leach A.M., Sivavec T.M., Wisnudel M.B., and Boyette S. (2006b) Analog signal acqui-
sition from computer optical drives for quantitative chemical sensing. Anal. Chem. 78, 5893–5899.
Potyrailo R.A., Morris W.G., Leach A.M., Sivavec T.M., Wisnudel M.B., Krishnan K., Surman C., Hassib L.,
Wroczynski R., Boyette S., Xiao C., Agree A., and Cecconie T. (2007a) A multiplexed, ubiquitous chem/bio
detection platform on DVD. Am. Lab. 32–35.
Potyrailo R.A., Morris W.G., Leach A.M., Hassib L., Krishnan K., Surman C., Wroczynski R., Boyette S., Xiao C.,
Shrikhande P., Agree A., and Cecconie T. (2007b) Theory and practice of ubiquitous quantitative chemical
analysis using conventional computer optical disk drives. Appl. Opt. 46, 7007–7017.
Potyrailo R.A., Barash E., Dovidenko K., and Lorraine P.W. (2008a) Methods and systems for detecting biological
and chemical materials on a submicron structured substrate. U.S. Patent Application 20080280374.
Potyrailo R.A., Ding Z., Butts M.D., Genovese S.E., and Deng T. (2008b) Selective chemical sensing using structu-
rally colored core-shell colloidal crystal films. IEEE Sensors J. 8, 815–822.
Potyrailo R.A., Dovidenko K., Le Tarte L.A., Surman C., and Pris A. (2009a) Theoretical and experimental deve-
lopment of label-free biosensors based on localized plasmon resonances on nanohole and nanopillar arrays.
Proc. SPIE 7322, 73220M1-11.
Potyrailo R.A., Morris W.G., Sivavec T., Tomlinson H.W., Klensmeden S., and Lindh K. (2009b) RFID sensors
based on ubiquitous passive 13.56-MHz RFID tags and complex impedance detection. Wireless Commun. Mobile
Comput. 2009, DOI: 10.1002/wcm.711.
Potyrailo R.A., Surman C., and Morris W.G. (2009c) Combinatorial screening of polymeric sensing materials using
RFID sensors: combined effects of plasticizers and temperature. J. Comb. Chem. 11, 598–603.
Preininger C., Mohr G., Klimant I., and Wolfbeis O.S. (1996) Ammonia fluorosensors based on reversible lactoniza-
tion of polymer-entrapped rhodamine dyes, and the effects of plasticizers. Anal. Chim. Acta 334, 113–123.
Ramirez A.G. and Saha R. (2004) Combinatorial studies for determining properties of thin-film gold–cobalt alloys.
Appl. Phys. Lett. 85, 5215–5217.
Raorane D., Lim S.-H., and Majumdar A. (2008) Nanomechanical assay to investigate the selectivity of binding
interactions between volatile benzene derivatives. Nano Lett. 8, 2229–2235.
Reddington E., Sapienza A., Gurau B., Viswanathan R., Sarangapani S., Smotkin E.S., and Mallouk T.E. (1998)
Combinatorial electrochemistry: a highly parallel, optical screening method for discovery of better electroca-
talysts. Science 280, 1735–1737.
Sanders D. and Simon U. (2007) High-throughput gas sensing screening of surface-doped In2O3. J. Comb. Chem. 9,
53–61.
Schatz G.C. (2007) Using theory and computation to model nanoscale properties. Proc. Natl. Acad. Sci. USA 104,
6885–6892.
Scheidtmann J., Frantzen A., Frenzer G., and Maier W.F. (2005) A combinatorial technique for the search of solid
state gas sensor materials. Meas. Sci. Technol. 16, 119–127.
Schena M. (2003) Microarray Analysis. Wiley, Hoboken, NJ.
Schneider. T., Betsarkis K., Trouillet V., and Goschnick J. (2004) Platinum-doped nanogranular tin dioxide layers
prepared by spin-coating from colloidal dispersions as basis for gradient gas sensor micro arrays. Proc. IEEE
Sensors 1, 196–197.
Schultz P.G. (2003) Commentary on combinatorial chemistry. Appl. Catal. A 254, 3–4.
Seker F., Meeker K., Kuech T.F., and Ellis A.B. (2000) Surface chemistry of prototypical bulk II-VI and III-V se-
miconductors and implications for chemical sensing. Chem. Rev., 100, 2505–2536.
Semancik S. (2002) Correlation of Chemisorption and Electronic Effects for Metal Oxide Interfaces: Transducing
Principles for Temperature Programmed Gas Microsensors. Final Technical Report Project Number: EMSP
65421, Grant Number: 07-98ER62709, U.S. Department of Energy Information Bridge.
Semancik S. (2003) Temperature-dependent materials research with micromachined array platforms. In: Combinatorial
Materials Synthesis, Xiang X.-D. and Takeuchi I. (eds.). Marcel Dekker, New York, pp. 263–295.
Setasuwon P., Menbangpung L., and Sahasithiwat S. (2008) Eight combinatorial stacks of three layers of carbon
black/PVA-carbon black/EVA composite as a vapor detector array. J. Comb. Chem. 10, 959–965.
Shtoyko T., Zudans I., Seliskar C.J., Heineman W.R., and Richardson J.N. (2004) An attenuated total reflectance
sensor for copper: an experiment for analytical or physical chemistry. J. Chem. Educ. 81, 1617–1619.
Sieg S., Stutz B., Schmidt T., Hamprecht F., and Maier W.F. (2006) A QCAR approach to materials modeling. J.
Molec. Mod. 12, 611–619.
Sieg S.C., Suh C., Schmidt T., Stukowski M., Rajan K., and Maier W.F. (2007) Principal component analysis of cata-
lytic functions in the composition space of heterogeneous catalysts. QSAR & Comb. Sci. 26, 528–535.
Siemons M., Koplin T.J., and Simon U. (2007) Advances in high throughput screening of gas sensing materials.
Appl. Surf. Sci. 254, 669–676.
Siemons M. and Simon U. (2006) Preparation and gas sensing properties of nanocrystalline La-doped CoTiO3. Sens.
Siemons M. and Simon U. (2007) Gas sensing properties of volume-doped CoTiO3 synthesized via polyol method.
Simon U., Sanders D., Jockel J., and Brinz T. (2005) Setup for high-throughput impedance screening of gas-sensing
materials. J. Comb. Chem. 7, 682–687.
Simon U., Sanders D., Jockel J., Heppel C., and Brinz T. (2002) Design strategies for multielectrode arrays applicable
for high-throughput impedance spectroscopy on novel gas sensor materials. J. Comb. Chem., 4, 511–515.
Sivavec T.M. and Potyrailo R.A. (2002) Polymer Coatings for Chemical Sensors, U.S. Patent 6,357,278 B1.
Stafslien S.J., Bahr J.A., Feser J.M., Weisz J.C., Chisholm B.J., Ready T.E., and Boudjouk P. (2006) Combinatorial
materials research applied to the development of new surface coatings I: a multiwell plate screening method for
the high-throughput assessment of bacterial biofilm retention on surfaces. J. Comb. Chem. 8, 156–162.
Steinle E.D., Amemiya S., Bühlmann P., and Meyerhoff M.E. (2000) Origin of non-Nernstian anion response slo-
pes of metalloporphyrin-based liquid/polymer membrane electrodes. Anal. Chem. 72, 5766–5773.
Stewart M.E., Anderton C.R., Thompson L.B., Maria J., Gray S.K., Rogers J.A., and Nuzzo R.G. (2008)
Nanostructured plasmonic sensors. Chem. Rev. 108, 494–521.
Stewart M.E., Mack N.H., Malyarchuk V., Soares J.A.N.T., Lee T.-W., Gray S.K., Nuzzo R.G., and Rogers J.A. (2006)
Quantitative multispectral biosensing and 1D imaging using quasi-3D plasmonic crystals. Proc. Natl. Acad. Sci.
USA 103, 17143–17148.
Suman M., Freddi M., Massera C., Ugozzoli F., and Dalcanale E. (2003) Rational design of cavitand receptors for
mass sensors. J. Am. Chem. Soc. 125, 12068–12069.
Sun L.-X. and Okada T. (2001) Studies on interactions between Nafion and organic vapours by quartz crystal mi-
crobalance. J. Memb. Sci. 183, 213–221.
Suzuki H. (2000) Advances in the microfabrication of electrochemical sensors and systems. Electroanalysis, 12,
703–715.
Svetlicic V., Schmidt A.J., and Miller L.L. (1998) Conductometric sensors based on the hypersensitive response of
plasticized polyaniline films to organic vapors. Chem. Mater. 10, 3305–3307.
Sysoev V.V., Kiselev I., Frietsch M., and Goschnick J. (2004) Temperature gradient effect on gas discrimination
power of a metal-oxide thin-film sensor microarray. Sensors 4, 37–46.
Tailoka F., Fray D.J., and Kumar R. V. (2003) Application of Nafion electrolytes for the detection of humidity in a
corrosive atmosphere. Solid State Ionics 161, 267–277.
Takeuchi I., Famodu O.O., Read J.C., Aronova M.A., Chang K.-S., Craciunescu C., Lofland S.E., Wuttig M.,
Wellstood F.C., Knauss L., and Orozco A. (2003) Identification of novel compositions of ferromagnetic sha-
pe-memory alloys using composition spreads. Nature Mater. 2, 180–184.
Takeuchi I., Newsam J.M., Wille L.T., Koinuma H., and Amis E.J. (eds.) (2002) Combinatorial and Artificial Intelligence
Methods in Materials Science. Materials Research Society, Warrendale, PA.
Taylor C.J. and Semancik,S. (2002) Use of microhotplate arrays as microdeposition substrates for materials explo-
ration. Chem. Mater. 14, 1671–1677.
Turcu F., Hartwich G., Schäfer D., and Schuhmann W. (2005) Ink-jet microdispensing for the formation of gradi-
ents of immobilised enzyme activity. Macromol. Rapid Commun. 26, 325–330.
Ulmer C.W. II, Smith D.A., Sumpter B.G., and Noid D.I. (1998) Computational neural networks and the rational
design of polymeric materials: the next generation polycarbonates. Comput. Theor. Polym. Sci. 8, 311–321.
Vassiltsova O.V., Zhao Z., Petrukhina M.A., and Carpenter M.A. (2007) Surface-functionalized CdSe quantum dots
for the detection of hydrocarbons. Sens. Actuators B 123, 522–529.
Villoslada F.N. and Takeuchi T. (2005) Multivariate analysis and experimental design in the screening of combina-
torial libraries of molecular imprinted polymers. Bull. Chem. Soc. Jpn. 78, 1354–1361.
Walt D. R., Dickinson T., White J., Kauer J., Johnson S., Engelhardt H., Sutter J., and Jurs P. (1998) Optical sensor
arrays for odor recognition. Biosens. Bioelectron., 13, 697–699.
Wang J. (2002) Electrochemical detection for microscale analytical systems: a review. Talanta 56, 223–231.
Wang J., Musameh M., and Lin Y. (2003) Solubilization of carbon nanotubes by Nafion toward the preparation of
amperometric biosensors. J. Am. Chem. Soc. 125, 2408–2409.
Wicks D.A. and Bach H. (2002) The coming revolution for coatings science: high throughput screening for formu-
lations. In: Proceedings of the 29th International Waterborne, High-Solids, and Powder Coatings Symposium, 29, p. 1–24.
Wintersgill M.C. and Fontanella J.J. (1998) Complex impedance measurements on Nafion. Electrochim. Acta 43,
1533–1538.
Wise B.M., Gallagher N.B., and Grate J.W. (2003) Analysis of combined mass- and volume-transducing sensor
arrays. J. Chemometrics 17, 463–469.
Wong D.W. and Robertson G.H. (1999) Combinatorial chemistry and its applications in agriculture and food. Adv.
Exp. Med. Biol. 464, 91–105.
Wu X., Zhang J., Chen J., Zhao C., and Gong Q. (2009) Refractive index sensor based on surface-plasmon interfe-
rence. Opt. Lett. 34, 392–394.
Xiang X.-D., Sun X., Briceño G., Lou Y., Wang K.-A., Chang H., Wallace-Freedman W.G., Chen S.-W., and Schultz
P.G. (1995) A combinatorial approach to materials discovery. Science 268, 1738–1740.
Xiang X.-D. and Takeuchi I. (eds.) (2003) Combinatorial Materials Synthesis. Marcel Dekker, New York.
Yang J.-C., Ji J., Hogle J.M., and Larson D.N. (2008) Metallic nanohole arrays on fluoropolymer substrates as small
label-free real-time bioprobes. Nano Lett. 8, 2718–2724.
Yoon J., Jung Y.-S., and Kim J.-M. (2009) A combinatorial approach for colorimetric differentiation of organic
solvents based on conjugated polymer-embedded electrospun fibers. Adv. Funct. Mater. 19, 209–214.
Zemel J.N. (1990) Microfabricated nonoptical chemical sensors. Rev. Sci. Instrum. 61, 1579–1606.
Zhao J., Pinchuk A.O., McMahon J.M., Li S., Ausman L.K., Atkinson A.L., and Schatz G.C. (2008) Methods for
describing the electromagnetic properties of silver and gold nanoparticles. Acc. Chem. Res. 41, 1710–1720.
Zhao J.-C. (2001) A combinatorial approach for structural materials. Adv. Eng. Mater. 3, 143–147.
Zheng Y.B., Juluri B.K., Mao X., Walker T.R., and Huang T.J. (2008) Systematic investigation of localized surface
plasmon resonance of long-range ordered Au nanodisk arrays. J. Appl. Phys., 103, 014308.
Zou L., Savvate'ev V., Booher J., Kim C.-H., and Shinar J. (2001) Combinatorial fabrication and studies of intense
efficient ultraviolet–violet organic light-emitting device arrays. Appl. Phys. Lett. 79, 2282–2284.
CHAPTER 4
SYNTHESIS AND DEPOSITION OF

SENSOR MATERIALS
G. Korotcenkov
B. K. Cho
1. DEPOSITION TECHNOLOGY: INTRODUCTION AND OVERVIEW

The production of high-quality materials suitable for chemical sensors is one of the most important
tasks of modern materials science. As shown in previous chapters, materials used in chemical sensors
need to fulfill a range of requirements related to the crystallographic structure, chemical composi-
tion, electrophysical properties, catalytic activity, and so on. These materials also show a great deal of
variation. Materials for chemical sensors can come in a variety of forms, including films, ceramics, or
powders. Their structure may be amorphous, glassy, nanocrystalline, polycrystalline, monocrystalline, or
epitaxial. They may be either dense or porous. These materials may be elementary substances, complex
compounds, or composites. Metals, dielectrics, and semiconductors can also be used as materials for
chemical sensors. They may be either organic or inorganic in nature.
This vast amount of variation indicates that it is impossible to produce such a wide range of
materials using just one method. The possible differences in the physical-chemical properties of the
materials are too great; so too are the resulting differences in the conditions required for the synthesis
and deposition of these materials. Therefore, the aim of this survey is to provide a brief overview of
the basic technological methods that may be used in sensor technology. The purpose is not to analyze
these methods in detail, as there already exists a vast array of quality reviews devoted to the subject
(Randhaw 1991; Hitchman and Jensen 1993; Hecht et al. 1994; Bunshah 1994; Brinker et al. 1996;
215
Glocker and Shah 1995; Nenov and Yordanov 1996; Artur 2002; Willmott 2004; Tay et al. 2006; Vahlas
et al. 2006; Vayssiere 2007; Viswanathan et al. 2006; Christen and Eres 2008; Jaworek and Sobszyk 2008;
Michev 2008; Tiemann 2008). This chapter will instead devote considerable attention to discussing the
advantages and disadvantages of the different methods used in the fabrication of sensor materials.
The particulars of various technologies for synthesizing specific materials, such as one-dimensional
structures (nanotubes, nanowires), metal nanoparticles, mesoporous materials, and nanocomposites, are
discussed in other chapters.
2. VACUUM EVAPORATION AND VACUUM DEPOSITION
2.1. PRINCIPLES OF FILM DEPOSITION BY THE VACUUM

EVAPORATION METHOD
The standard vacuum evaporation method includes both thermal evaporation and electron-beam evap-
oration. This method has seen widespread use in the deposition of metal films for many years. In this
case, the metal can be evaporated by suspending it in a coil or coiled filament made of a material with
a high melting temperature that is heated to the required temperature. Typical evaporation sources and
their operational principles are shown schematically in Figure 4.1.
The surface temperature required for evaporation of a material depends on its vapor pressure,
which is a strong function of temperature. For different materials, this pressure may vary within the
range of 1 to 9 orders of magnitude at a fixed temperature (Westwood 1989). As is well known (Powell
et al. 1967), the evaporation rate (Vevap ) is first determined by the pressure of the saturated vapor over
the vaporizable substance:
P
Vevap = 2.635 × 10 24 (4.1)
MT
where P is the pressure of the saturated vapor, M is the molecular weight, and T is the temperature.
The vapor pressure required to obtain a useful evaporation rate is usually about 1 Pa (Westwood 1989).
The temperatures and the rates of evaporation of several metals used in chemical sensor fabrication
are given in Table 4.1.
The thermal vacuum evaporation method is not widely used for deposition of composite com-
pounds, because dealing with the retention of the initial stoichiometry of the material in the deposited
film poses a considerable problem. Note that introducing oxygen to the evaporation chamber to obtain
metal oxides is not possible with vacuum evaporation. As a result, metal oxides formed by thermal
evaporation are nonstoichiometric and therefore an additional step of thermal annealing in an oxygen
atmosphere is required to acchieve the correct stoichiometry.
2.2. DISADVANTAGES OF THE VACUUM EVAPORATION METHOD

Several disadvantages of this method are well known, including:
SYNTHESIS AND DEPOSITION OF SENSOR MATERIALS • 217
Figure 4.1. Variants of vacuum evaporation: (a) thermal evaporation; (b) electron-gun evaporation;
(c) evaporation from Knudsen cell; (d) flash evaporation; (e) deposition of compounds using three
Knudsen cells.
• Possible contamination from the evaporator

• The need to periodically load the evaporator
• Inconstancy of evaporation rates during the deposition process
• A considerable difference in composition between the evaporated and deposited materials
• Difficulties in the evaporation of materials that have high melting temperature and low saturated
vapor pressure.
To resolve the problem of deposition for both multicomponent compounds and alloys using the
thermal evaporation method, it has commonly been suggested to deposit these films using either a flash
sputtering method (see Figure 4.1d) or by evaporation of the material’s components from different in-
dependent sources (Figure 4.1e). During the flash evaporation method, the stoichiometry of the depos-
ited film is improved at the expense of fast and complete evaporation of a small amount of deposited
substance, delivered discretely at the evaporator. In the area above the evaporator, all components of the
vaporizable material are presented in a vapor phase in the same ratio as the initial materials before evapo-
ration. Therefore, a film with the desired composition will automatically be obtained on the substrate
surface. It necessary to note, however, that the complications of controlling these processes reduces the
reproducibility of the results. The indicated disadvantages considerably restrict the areas of potential ap-
plication for this method. However, the opportunity to attain a high rate of deposition, which is not ac-
cessible with many modern methods, maintains this method’s appeal for a certain practical applications.
2.3. FILM DEPOSITION BY THERMAL EVAPORATION

The main parameters of the film deposition process using the vacuum evaporation method are the level
of the vacuum, the rate of deposition, the temperature of the substrate, and the structure of the sub-
Table 4.1. Temperatures and rates of evaporation for metals
EVAPORATION
RATE OF
TEMPERATURE
EVAPORATION RECOMMENDED MATERIALS
METAL °С К (10−4 g cm−2 s−1 ) FOR FILAMENT OR BOAT
Rb (L) 176 449 2.55

Zn 342 615 1.90 W, Ta, Mo, Nb, (Al2O3 )
Sr 530 803 1.93 W, Ta, Mo, Nb, (Al2O3 )
In (L) 947 1220 1.79 W, Fe, (Al2O3)
Mn 947 1220 1.24 W, Ta, Mo, Nb, (Al2O3 )
Ag (L) 1032 1305 1.67 Ta, Mo, Nb, Fe
Ga (L) 1957 1330 1.34
Al (L) 1207 1480 0.788 W, Ta, Mo, Nb, (BN )
Sn (L) 1227 1500 1.64 (Al2O3 )
Cu (L) 1272 1545 1.18 Nb, Mo, Ta, W, (Al2O3 )
Au (L) 1332 1605 2.05 W, Mo
Cr 1392 1665 1.03 W, Ta, (Al2O3 )
Fe 1467 1740 1.05 W, (Al2O3, BeO)
Ni (L) 1497 1770 1.06 W, (Al2O3, BeO)
Co (L) 1516 1789 1.06 Nb, (Al2O3, ZrO2, BeO)
Pd (L) 1547 1820 1.41 W, (Al2O3 )
Ti 1703 1980 0.90 W, Ta
La (L) 1740 2013 1.53
V (L) 1845 2118 0.95 W, Mo
Pt (L) 2077 2350 1.68 W, (Al2O3)
Rh (L) 2027 2300 1.23 (Al2O3)
Zr 2397 2670 1.08 W, (Al2O3 )
Ta 2996 3269 0.686
W 3324 3607 1.32
(L), in liquid state.
strate. All of these factors need to be strictly controlled, and their proper choice is of great importance
during film deposition. For deposition from a vapor phase on a nonheated substrate, the deposited
materials usually form amorphous structures (Edelman et al. 2000; Kimura et al. 2001). The maximum
temperature at which the amorphous phase remains stable can vary within the range of 100–400°С,
depending on the melting temperature of the deposited material and the substrate type. Amorphous
thin films, stable at room temperature, usually have a disordered and dense packing structure. The
increase in the substrate temperature during deposition leads to the ordering of the film structure. As
the temperature of the substrate increases, atoms become more mobile and growing grains (crystallites)
enlarge. It is important to note that the increase in the substrate’s temperature during deposition intensi-
fies the enlargement of grains in comparison to increasing the temperature in the following annealing
of the deposited film.
As a rule, noticeable textures are revealed in films at temperatures where they are in a crystal state.
The degree and character of these textures may change sharply as the deposition temperature increases
(Palatnik et al. 1972). It has been established that, for textured film forming, the deposition should take
place at a reasonably slow rate of growth. Therefore, at high rates of growth, the film’s texture generally
appears weaker. An explanation for this effect is given by Palatnik et al. (1972).
It is well known that an increase in the concentration of impurities, oversaturation, overcooling,
and a decrease of film–substrate interface energy results in an increase in the rate of heterogeneous
nucleation. Therefore, low-vacuum conditions with a high rate of deposition and a reduced substrate
temperature will generally form finely dispersed (or amorphous) films (Palatnik et al. 1972).
A great deal of research has shown that, during thermal evaporation, the structure of thin films in
the starting stages of growth may be stable in form, close to monoatomic layers, or in the form of three-
dimensional blocks (Powell et al. 1967). As a general rule, the following correlation can be seen: At HL<HM
agglomerates are formed, whereas at HL<HM a homogeneous layer is formed. Here, HL is the heat of
vaporization of the metal atoms from the substrate, and HM is the heat of vaporization of the metal atoms
from metal. The majority of materials form agglomerates. If the size of the critical nucleating centers and
the free energy for their formation are large, the film consists of a small amount of large agglomerates.
If those magnitudes are small, the film consists of a large amount of small agglomerations. Such film
becomes continuous at a relatively small value of medium thickness. The large agglomerate structure is
preserved until a comparatively large film thickness is obtained. Hence, the exact conditions of the film
growth determine its microstructure. For example, the inclination for 3-D blocks forming is reduced as
the rate of film deposition increases. High levels of dispersion in the film and a large disorientation of the
growing crystallites on the film’s plane can be explained by a large amount of nucleating centers forming
on the substrate during the first stage of growth, from which a finely dispersed structure develops.
3. SPUTTERING TECHNOLOGY
3.1. PRINCIPLES OF DEPOSITION BY SPUTTERING
Sputtering technology has developed rapidly over the last decade to the point where it has become
established as the process of choice for the deposition of a wide range of industrially important coat-
ings. Sputtering deposition is widely used to grow alloys and component films in which one or more
of the constituent elements are volatile (Brown 1998; Tay et al. 2006; Aita 2008). Low-defect-density
films of high-melting-point materials can be grown on unheated substrates because phase formation is
governed primarily by kinetics rather than by thermodynamics. Physical sputtering of the target occurs
when positive ions from the plasma are accelerated across the dark space and strike the target surface
(Moisan and Pelletier 1992) (see Figure 4.2).
To realize a significant transfer of momentum, the incident particle must have a mass Mi compa-
rable to that of the target atoms (Mt). The sputtering rate may be calculated using Eq. (4.2), formulated
by Behrisch (1981):
4 Mi Mt E
S =C ⋅ ⋅ Sn ( Ei ) ⋅ i (4.2)
( Mi + M t ) U
where Ei is the kinetic energy of the incident particles, U is the surface binding energy, and Sn(Ei ) is
a function of both Ei and the particle masses. The incident particles are usually ions because they are
easily accelerated to energy Ei .
Note that the influence of the accelerated energy on the deposition rate, as shown in Figure 4.3,
has a threshold. Region IV is usually used for film deposition because the sputtering coefficient is
weakly dependent on the ion kinetic energy in this region. As a rule, the sputtering rate is independent
of the target temperature unless the target becomes liquid. Any material ejected by a momentum trans-
fer process remains mostly uncharged and moves between the electrodes, where it becomes thermal-
ized and condenses on any surface.
Power +
Reaction supply
chamber
Cathode
Plasma
Substrate
Gas Vacuum
pump
Figure 4.2. Typical construction of devices for DC and radio-frequency sputtering.
Ethr

Figure 4.3. Dependence of deposition rate on kinetic energy of incident particles.
A metal oxide film is grown by sputtering a metal target in a discharge containing oxygen, usually in
conjunction with a noble gas (Tay et al. 2006). The sputtering process may be accompanied by dissocia-
tion of the sputtered oxides. Therefore, the sputtered flux may consist of both metal atoms and metal
oxide molecules. The metal, metal oxide, and oxygen species that arrive at the substrate are adsorbed
and ultimately incorporated into a stable nucleus to form a continuous film. These processes are the
major factors determining film chemistry, short-range atom order, crystallography, and the microstruc-
ture of deposited films.
Since the vapor pressure of metals is very low, except at elevated temperatures, the sticking coef-
ficients for the different species in the sputtered flux tend to 1.0, and the film composition will be the
same as the composition of the flux. There is another situation, however, in relation to a compound’s
sputtering. All compounds do not have a low vapor pressure, so their sticking coefficients may not be
equal to 1.0. As a result, metal oxide films, for example, are oxygen-deficient (Bardos et al. 1983). This
nonstochiometry, of course, is not as strong as when it is observed during metal oxide deposition us-
ing thermal evaporation. To compensate for the oxygen deficiency in the film, oxygen is added to the
sputtering system.
Ceramic thin films can be deposited on various substrate materials. During sputtering, the substrate
temperature usually does not exceed 70°C. Annealing may be carried out after the sputtering deposition
to repair damage from sputtering and to study the effects of the annealing temperature on the structural
and electrical properties.
3.2. SPUTTERING TECHNIQUES

The main variations of sputtering techniques are
• DC cathodic sputtering
• Radio-frequency (RF) sputtering
• Magnetron sputtering
3.2.1. DC and RF Sputtering

A schematic diagram of the apparatus for DC and RF sputtering is shown in Figure 4.3. In this basic
sputtering process, a target (or cathode) plate is bombarded by energetic ions generated from a glow
discharge plasma situated in front of the target. Secondary electrons are also emitted from the target
surface as a result of the ion bombardment, and these electrons play an important role in maintaining
the plasma. Many materials have been successfully deposited using the DC and RF sputtering technique
(Gong et al. 2002; Bittencourt et al. 2002), but the process is limited by its low deposition rates (Nadel
and Greene 2001, 2003), low ionization efficiencies in the plasma, and high substrate heating effects.
The development of magnetron sputtering and, more recently, unbalanced magnetron sputtering have
helped to overcome these limitations.
3.2.2. Magnetron Sputtering

Schematic diagrams of various magnetron sputtering systems are shown in Figure 4.4. Magnetrons
make use of the fact that a magnetic field configured parallel to the target surface can constrain second-
ary electron motion to the vicinity of the target (Kelly and Arnell 2000). The magnets are arranged in
such a way that one pole is positioned at the central axis of the target and the other pole is formed by a
ring of magnets around the target’s outer edge. Trapping the electrons in this way substantially increases
the probability of an ionizing electron–atom collision occurring. The increased ionization efficiency of
Figure 4.4. Schematic representation of devices for magnetron sputtering: (a) conventional magnetron
(ion current density < 1 mA cm−2); (b) unbalanced magnetron (ion current density ~ 2–10 mA cm−2).
Figure 4.5. Dual unbalanced closed-field magnetron configuration: (a) vertically opposed configura-
tion; (b) co-planar configuration.
a magnetron results in a dense plasma in the target region. This in turn leads to increased ion bombard-
ment of the target, giving higher sputtering rates and, therefore, higher deposition rates at the substrate.
In addition, the increased ionization efficiency achieved from the magnetron mode allows the discharge
to be maintained at a lower operating pressure (typically, 10−3 mbar, compared to 10−2 mbar) and a lower
operating voltage (typically, −500 V, compared to −2 to −3 kV) than is possible in the basic sputtering
mode. Other variants of magnetron sputtering systems can be seen in Figure 4.5.
In the mirrored case (see Figure 4.5), the field lines are directed toward the chamber walls. Secondary
electrons following these lines are lost, resulting in a low plasma density in the substrate region. Conversely,
in the closed-field configuration, the field lines are linked between the magnetrons. Losses to the chamber
walls are low and the substrate lies in a high-density plasma region. Operation in the closed-field mode
results in an ion-to-atom ratio incident at the substrate some two to three times greater than that obtained
under the same conditions using the mirrored, or single unbalanced magnetron configurations.
3.2.3. Pulsed Magnetron Sputtering

The problems associated with the reactive magnetron sputtering of highly insulating materials have
been widely reported (Bunshah 1994; Musil 1996; Kelly and Arnell 2000). As the deposition process
proceeds, areas on the target away from the main racetrack become covered with an insulating layer, as
do the target earth shields. The poisoned layers charge up until breakdown occurs in the form of an arc.
Arc events at the target can result in the ejection of droplets of material from the target surface, which
can cause defects in the growing film.
The recently developed pulsed magnetron sputtering process (PMS) overcomes many of the prob-
lems that occur from the reactive sputtering mode (Schiller et al. 1993; Posadowski, 1999). Studies have
shown that, when depositing insulating films, pulsing the magnetron discharge in the medium frequency
range (10–200 kHz) can significantly reduce the formation of arcs and, consequently, reduce the num-
ber of defects in the resulting film. Therefore, the PMS process now enables a high deposition rate of
defect-free ceramic films. As such, this process has attracted considerable commercial interest and has
led to the development of a new generation of magnetron power supplies and pulse units.
3.3. ADVANTAGES AND DISADVANTAGES OF

SPUTTERING TECHNOLOGY
Because the sputtering process is a low-temperature process, it can be used for the sputtering of almost
all materials used in chemical sensors. Additionally, the sputtering method presents several advantages
(Bardos et al. 1983; Kelly and Arnell 2000; Safi 2000; Tay et al. 2006):
• The opportunity to synthesize compounds that cannot be obtained by thermal evaporation of

materials in a high vacuum. This can be achieved by introducing various reactive gases into the
gas-discharge plasma.
• The high adhesion of a film, conditioned by the fact that the energy of sputtered particles is a
degree higher than the energy of thermally evaporated particles.
• The retention of the stoichiometry of multicomponent materials during sputtering. It is well
known that many materials have a sputtering rate that differs by no more than two to three times,
while thermal evaporation rates may change by several degrees.
• A high coefficient of use of the sputtered material.
• Becoming homogeneous through thickness coverings.
• The opportunity to create apparatuss and production lines for continuous operation, because
targets can have a nearly unlimited supply of sputtering material.
Among the primary disadvantages of this method, it is necessary to include such properties as:
(1) difficulties in plasma stabilization, particularly at low pressure and in large areas, (2) the rate’s de-
pendence on electrical power, and (3) a large particle energy resulting in surface damage due to surface
bombardment. The problem of surface damage is particularly important for the deposition of polymer
materials and films, which cannot be thermally processed for recovery to the initial state.
To resolve the problem of ion bombardment of the surface during the process of film deposition
by sputtering, three independent approaches are currently being used:
• Thermalization of sputtered particles

• Considerable reduction of process pressure
• Use an off-axis configuration for the substrate holder
The principles of these approaches can be understood simply from their names. It only needs to be
added that the thermalization principle indicates that, during deposition, conditions are being created in
which the sputtered particles lose their kinetic energy due to collisions with molecules of residual gas
and reach the substrate with the help of diffusion.
3.4. PROPERTIES OF FILMS DEPOSITED BY SPUTTERING
Ion bombardment during film growth provides sufficient activation energy to modify the film proper-
ties, and the plasma processes can be controlled to provide the required film properties (Randhaw 1991;
Kizling and Jaras 1996). The specific conditions for film deposition depend on the applications (Musil
1996; Aita 2008). For example, gas sensor applications require high porosity, a small grain size, and as
much adhesion of the sensing layer as possible. On the other hand, wear applications require a high
intrinsic strength, high hardness, and high adhesion.
The effects of ion bombardment on film growth have been studied in great detail over the last de-
cade. In general, ion bombardment results in films with improved adhesion and density. Interaction with
high-energy particles results in strong atomic mixing at the interface and the generation of high concen-
trations of radiation-induced defects. Substrate cleaning, however, also plays an important role in adhe-
sion. The increase in the substrate temperature resulting from ion bombardment further promotes a high
level of adhesion. The high-energy ions sputter the material from the surface biased with a high negative
voltage. This helps to remove any residual thin oxide layers and any foreign surface contaminations.
Ion bombardment also facilitates the low substrate temperatures needed during deposition by
activating the reaction species, particularly in reacted films (Bunshah 1982, 1994). The preferred orien-
tation of the films also depends on the method of deposition (Kelly and Arnell 2000). This has been
attributed to the differences in ionization and the activated reactions that occur in DC and RF plasmas.
Bombarding the growing films with energetic ions and neutral particles can also result in modification
of the film morphology, crystal orientation, grain size, etc. Films grown using plasma-assisted processes
are generally fine-grained and have a dense structure, resulting in very high hardness. Bombardment by
energetic atoms and ions may therefore be advantageous in obtaining a particular film property—for
example, morphology or crystal orientation. However, associated effects such as stress and defects must
also be considered (Safi 2000; Musil and Vlcek 2001).
The thickness of deposited films can range from monolayers up to micrometers, or even tens or
hundreds of micrometers in some cases. However, a thickness in the broad range of 0.1–10 μm is typi-
cal for this technique. The rate of deposition does not normally exceed tens of micrometers per hour.
When summarizing the results of binary oxide deposition on the substrate at room temperature
using metal target sputtering, the main conclusions for the most studied ZnO, In2O3, and SnO2 oxides
may be formulated as follows (Zeman and Takabayashi 2002; Bittencourt et al. 2002; Gong et al. 2002;
Lemire et al. 2002; Tay et al. 2006; etc.):
• As a rule, the increase of O2 partial pressure is accompanied by a decrease in the deposition rate.
This behavior is explained by the partial oxidation of the metal target. As is well known, metal
oxides have a smaller coefficient of sputtering than metals.
• Films deposited in an atmosphere of pure argon are metals. The growth of metal oxide will take
place in an oxygen-containing atmosphere which already has an O2 partial pressure higher than
6 × 10−4 torr (see Figure 4.6).
• The structure of deposited films depends on both the O2 partial pressure and the nature of the de-
posited oxides. ZnO films with a minimal O2 partial pressure are polycrystalline. At the same time,
In2O3 films with a minimal O2 pressure (5 × 10−4 torr) have an amorphous structure, and they
begin to show a polycrystalline structure only at a higher O2 partial pressure. SnO2 films deposited
at room temperature have an amorphous structure independent of the oxygen partial pressure.
• The grain size of deposited films depends on both the deposition parameters and the nature of
the oxide. The grain size of In2O3, ZnO, and SnO2 films deposited under the same conditions
decreases in the order In2O3 > ZnO > SnO2. The increase of O2 partial pressure is accompanied
by a decrease in grain size of up to two or three times.
• The rise of O2 partial pressure during sputtering promotes a decrease of film conductivity. This
takes place in exact accordance with an improvement in film stoichiometry.
Thus, metal oxide films formed by reactive sputtering at room temperature may have either an
amorphous or a polycrystalline structure (Gong et al. 2002). For the majority of metal oxides, the as-
deposited amorphous state is stable up to 300°C, and the first crystallization can be observed in metal
oxide films only after annealing at temperatures higher than 300°C.
Increase of crystallite size and a change in the film’s morphology also take place when the deposi-
tion temperature increases. For example, Gong et al. (Gong et al. 2002; Lemire et al. 2002) showed that
an increase in the substrate temperature from 100 to 400°C during ZnO deposition was accompanied
Figure 4.6. Influence of total and oxygen partial pressures in deposition chamber on TiO2 crystal
structure. (Reprinted with permission from Zeman and Takabayashi 2002. Copyright 2002 Elsevier.)
by a growth in grain size from 13 to 43 nm. During deposition on a heated substrate (Тsub >200°С),
the structure of the most sputtered films consists of columns. The stress in evaporated metal films
is usually tensile and is more apparent for refractory metals than for the lower-melting-point metals
(Westwood 1989). Compressive stress is usually observed in sputtered compound films.
4. THE RGTO TECHNIQUE
4.1. PARTICULARS OF THE RGTO METHOD
It is well known that both high porosity and a large surface area are important requirements for materi-
als used in gas sensor designs. The reothaxial growth and thermal oxidation (RGTO) technique was designed
by Sbervegliery et al. (1991, 1993) to fabricate the exact highly porous metal oxide films necessary for
solid-state conductometric gas sensors. This method consists of two main stages. During the first stage,
reothaxial growth, the sputter (or thermal evaporation) deposition of a metal that does not sublimate,
on top of an insulating substrate, is maintained at a temperature above the melting point of the metal
(232°C for tin). Variations of this method consist of deposition of metals in an oxidant atmosphere,
maintaining the substrate at a low temperature, or heating the substrate during deposition (Yoo et al.
1995). In the second stage, thermal oxidation, the sample is annealed under an oxygen atmosphere,
synthetic air, at a temperature high enough to form the metal oxide.
Rapid and complete oxidation of metal oxides takes place at temperatures higher than 700°C. The
technology for sensor fabrication, however, hinders the use of temperatures higher than 500–700°C
because of problems with material diffusion inside the sensor structure. Therefore, for the oxidation of
tin, temperatures above 500°C are the most suitable to allow the oxygen to diffuse into the metallic tin
layer within an acceptable time frame. In this case, the duration of the RGTO oxidation cycle ranges
from 10 to 30 h.
The complete oxidation cycle in RGTO consists of two steps. As applied to SnO2, the first step is
carried out at 250°C for 4 h, and the second step is usually done at 600°C for times up to 30 h. To reach
each stable temperature, 4 h is usually needed to avoid fast evaporation of the metallic tin, especially dur-
ing the first step. The first stabilization, at 250°C, allows a first oxidation of the film to be obtained. This
oxidation method is roughly the same as that used for the oxidation of other metals (Gupta et al. 2002;
Radecka et al. 2001). It has been shown that, for metal oxidation, a stronger oxidizer, such as ozone, can
also be used (Gupta et al. 2002). However, the application of ozone should be carefully considered if
oxidation of the substrate needs to be excluded, as is the case for silicon-based microelectronics. It has
been established that the thickness of interfacial SiO2 increases when ozone is used as an oxidant.
Scanning electron microscopy (SEM) images of Sn and SnO2 films in the various steps of the
RGTO process are shown in Figures 4.7 and 4.8. As can be seen, after reothaxial growth, the metal does
not form a continuous layer, but rather microspheres due to the liquid-phase tension. An increase in
the sphere’s diameter is also observed, together with the increase in deposition time. After thermal oxi-
dation of the metal, the surface is formed by uniformly distributed agglomerates. These agglomerates
have a spongelike structure with a large surface area, required for application in semiconductor gas sen-
sors. Research conducted by Dieguez et al. (1997, 1999) showed that the SnO2 crystallites usually begin
Figure 4.7. SEM images of as-deposited Sn films with different thicknesses (Tsub = 400°C): (a) d =
50 nm; (b) d = 100 nm; (c) d = 200 nm; (d) d = 400 nm. (Reprinted with permission from Dieguez et
al. 1999. Copyright 1999 ElectroChemical Society.)
Figure 4.8. SEM images of Sn films shown in Figure 4.7 after oxidation at 600°C during 30 h.
(Reprinted with permission from Dieguez et al. 1999. Copyright 1999 ElectroChemical Society.)
to grow from the surface of the microspheres in random orientations. The stresses produced by crystal-
lite lattice expansion give rise to the formation of spongy agglomerates (Dieguez et al. 1997, 1999).
Thus, RGTO is a simple procedure based on deposition of the metal directly from its metallic tar-
get and subsequent thermal oxidation. In principle, the RGTO technique can be applied to the growth
of oxides of all metals that do not evaporate above the melting point and that tend to form stable
oxides by means of a thermal oxidation cycle. As a result, this method has been successfully applied in
the growth of Bi2O3, SnxFe1−xOy, SnO2, ZnO, Ga2O3, In2O3, Al2O3 etc. (Bonzi et al. 1994; Faglia et al.
1994; Hellmich et al. 1995; Radecka et al. 2001; Gupta et al. 2002).
4.2. ADVANTAGES AND DISADVANTAGES OF RGTO

Clearly, the RGTO technique allows for practical use of any metal to form a metal oxide film. Further,
the coverings may have a very-high porosity, which provides them with the high surface activity neces-
sary for chemical sensor design. The two-stage technology used in forming metal oxide films is a defi-
nite advantage of this method. In this case, the technology used to form the required topology of the
metal oxide film with high chemical stability is simplified considerably. It is well known that metals are
not as chemically stable as metal oxides.
It is necessary to point out, however, that the RGTO technology has very low reproducibility.
The parameters of the films formed are highly dependent on the mode of deposition, particularly its
thickness. Moreover, the technological cycle has an unacceptably long duration. All of these factors
combined lead to the conclusion that this method has very limited applicability in the real technological
cycle of chemical sensors design.
5. LASER ABLATION OR PULSED LASER DEPOSITION
5.1. PRINCIPLES OF PULSED LASER DEPOSITION

There is a general consensus that for miniaturization of multicomponent chemical sensors, thin films
are desired. However, achieving a multicomponent composition within thin films requires special depo-
sition methods (Schubert et al. 2001; Willmott 2004; Christen and Fres 2008). The pulsed laser deposi-
tion (PLD) technique was first introduced in the late 1980s and is a standard deposition technique that
provides the possibility for a stoichiometric transfer of even multicomponent target compositions to a
given substrate (Payne and Pravman 1990; Ozegowski et al. 1999). Therefore, utilizing the PLD process
for the preparation of thin film layers for chemical sensor applications is a major challenge (Greer et al.
1997; Myslik et al. 1999; Que et al. 2003; Chrisey and Hubler 1994).
Conceptually and experimentally, the PLD process is extremely simple, since the equipment con-
sists of a target and a substrate holder in a vacuum chamber (see Figure 4.9). In addition, a high-power
focused laser beam is used as an external energy source to vaporize materials and to deposit the thin
films. Some of the parameters for the lasers used in the PLD process are shown in Table 4.2.
The PLD method utilizes a short-pulse high-energy laser beam to vaporize material from a target,
and then deposits this material onto a substrate placed in a vacuum. During a pulse length of 30 ns,
Figure 4.9. Schematic off-axis configuration of the apparatus for laser ablation.
several stages of interaction between the laser beam and the target material occur (Schubert et al. 2001).
Optical irradiation causes the target surface to heat. At the high power levels (107–108 W cm−2) typically
used in PLD, any material will react with the heated surface during the interaction with the laser beam.
The surface heating is followed by a surface melting in the next stage. For efficient material removal
and subsequent evaporation of multicomponent targets, a short wavelength (typically λ = 248 nm) and
short pulse widths (approximately 30 ns) are used. High temperatures generated at the target surface
cause the emission of many species from the target. If the energy density of the laser radiation and the
composition and pressure of the surrounding atmosphere are suitable, a plasma plume will form above
the exposed target area, extending to a distance of several tens of millimeters. The formation of plasma
is a result of photoionization of the evaporated material and the multiphonon processes. Once gener-
ated at the surface, the plasma expands away perpendicularly from the target and then interacts with the
laser beam, absorbing its optical radiation. These ablation products are deposited on a closely localized
substrate and form a thin layer of deposited substance.
Table 4.2. Lasers used for PLD
PULSE DURATION MAXIMAL MAX. REP.

LASER WAVELENGTH ON TARGET ENERGY FREQUENCY
Excimer 193; 248; 308 nm 6–50 ns 400 mJ 20–50 Hz

Q-Switched Nd:YAG 266; 355; 532; 1064 nm 5–20 ns 300–600 mJ 20 Hz
Free-running Nd:YAG 1064 nm 0.1–0.6 ms 9000 mJ 1 Hz
TEA-CO2 10.6 μm 80 ns (3 μs) 900 mJ 20 Hz
5.2. ADVANTAGES AND DISADVANTAGES OF PLD
PLD technology can be used with great success for the deposition of thin films of almost any avail-
able inorganic material (Saender 1993; Chrisey and Hubler 1994; Hopfe et al. 1996; Greer et al. 1997;
Schoninget et al. 1999, 2001; Dolbec et al. 2002; Willmott 2004; Christen and Eres 2008), including
polymers (Myslik et al. 1999; Que et al. 2003). Employing PLD technology in the deposition of active
layers for sensor devices provides the following advantages:
• PLD technology requires no solvents (typical for thick-layer technology).

• PLD permits precise control of deposition of multilayer structures in situ in one technological
step (impossible in RF sputtering).
• With PLD, any metal, ceramic, alloy, or intermetallic compound, as well as fully reacted metals,
can be deposited on virtually any substrate, including plastic, paper, metals, and ceramics.
Occasionally, other processes cannot be used for these purposes.
• Films can be deposited at either low (including room temperature) or high temperature.
• The method is very convenient for the quick preparation and study of new sensing materials.
• PLD is compatible with oxygen and other reactive gases.
• PLD technology creates the possibility for controlled deposition of ultrathin coverings, which is
very attractive for surface modification of the materials for chemical sensors.
Nevertheless, until now, this technique has been scarcely employed for industrial applications in the
area of chemical sensor fabrication. This is largely due to the following disadvantages of PLD:
• Low productivity of the method.

• Relatively high investment costs.
• The composition and thickness depend on many deposition conditions. The processing parameters,
such as wavelength, energy and shape of the laser pulse, focusing geometry, process atmosphere,
and substrate temperature, all greatly influence the parameters of the ablative particle fluxes and,
hence, the properties of the deposited films and the efficiency of the deposition process.
• Difficulties in the deposition of thick layers.
• Difficulties in attaining the necessary stoichiometry of materials containing volatile components.
• Difficulties in scaling up to large wafers.
• As a result of repeated interaction of the laser beam with the target, structural changes occur
on the surface with craters forming. Therefore, the composition and properties of the deposited
material will depend on the duration of the deposition process.
5.3. TECHNICAL APPROACHES TO IMPROVING PLD RESULTS
To overcome the above-mentioned disadvantages of the PLD method, several technical approaches
have been suggested to improve PLD technology. These approaches include: (1) rotating the substrate,
or the laser beam scanning, over the surface of the target during the deposition process; (2) forming
the composite target from several materials; (3) creating a multiple-beam laser system; and (4) creating a
double-chamber system consisting of a deposition chamber and a film-growth chamber.
Conditions to control film stoichiometry can be created through the use of device that combines
a laser, a rotating target drum with various separate metal oxide targets, and a computer controller to
select (or skip) each target and adjust the laser energy needed to ablate, forming either buffer or contact
layers. Repeated ablations of different target materials may result in the deposition of a multicompo-
nent thin film with the required composition.
A schematic diagram of this two-chamber device is shown in Figure 4.10. In this scheme, a laser
beam vaporizes atoms from the target on the rotating drum. The adjacent film-growth chamber is filled
with volatile material, such as oxygen or another gas, which is kept at a uniform temperature and pres-
sure to facilitate film growth. A nonvolatile material from the laser ablation chamber is deposited on the
substrate as it passes by the opening in the film-growth chamber. The layers of atoms deposited on the
substrate then react with the gas in the film-growth chamber and spin toward the opening again. This
procedure may be repeated many times, promoting the formation of films with high stoichiometry.
5.4. SOME PARTICULARS OF FILM DEPOSITION BY THE PLD METHOD

Pulsed laser deposition is a vapor deposition technique; however, it has been established that, during pulse
lased deposition, there are far more controlling parameters than in other PVD methods. During PLD,
there are additional deposition parameters, such as plasma density, deposition rate per pulse, plasma com-
position, degree of ionization, and kinetic energy of the ions. Ozegowski et al. (1999) have shown that
when the wavelength and laser power density increase, the kinetic energy of the ions and the number of
multicharged ions also increases, while the number of ions in the plasma flux decreases. It was observed

Figure 4.10. Setup for impulse laser deposition using dual-beam scheme and double-chamber
system.
that the microhardness of metal oxide films, such as Al2O3, as well as the elastic modulus of alumina
films, increased when they were fabricated using a long-wavelength laser providing higher ion energy. The
dependence of the morphology of manganin films on the laser wavelength is much weaker. This behavior
is due to differences in the absorption of the various laser wavelengths for both materials. It has been
observed that the laser parameters actually have a negligible influence on the film composition.
For the example of SnO2 deposition, it was shown that an oxygen background pressure is a pre-
requisite for the growth of pure metal oxide phases. During deposition under vacuum, metal oxides are
nonstoichiometric. Moreover, it was established that by using a relatively high laser power (~5 J cm−2 ),
polycrystalline SnO2 can be grown at deposition temperatures as low as room temperature.
6. ION-BEAM–ASSISTED DEPOSITION (IBAD)

6.1. INTRODUCTION
In recent years it has become clear that low-energy ion beams play an important role in influencing
the processing of semiconductor thin films, thereby providing an additional mechanism to control
the parameters of semiconductors (Kizling and Jaras 1996). It is evident that, since ions of sufficient
energy striking a film physically displace atoms, the film microstructure is directly affected. Figure 4.11
summarizes the typical ion energies involved in various processes.
Low-energy and high-energy ion beams are used to sputter, clean and etch surfaces, enhance dop-
ant incorporation, modify strain, and change the growth modes of films, changing their morphology
and increasing densification and compactness (Cuomo et al. 1989; Armour 1994). Moreover, low-ener-
gy ion beams can be used to reduce thin-film deposition temperature, suppressing dopant diffusion and
preventing segregation (Pouch and Alerovitz 1992).
6.2. PRINCIPLES OF THE IBAD METHOD

IBAD is a physical vapor deposition method using electron-gun evaporation, thermal evaporation, or
sputtering, and is assisted by an ion beam bombarding the growing film (Hubler 1989). In this method,
accelerated ions and evaporated atoms arrive simultaneously at the surface of the substrate. Typical IBAD
equipment consists of a low-energy ion source, a deposition unit, and a target holder within a vacuum
10-2 1 101 102 103 104

Figure 4.11. Ion energies of various plasma-assisted deposition processes.
chamber. For example, the equipment shown schematically in Figure 4.12 consists of an electron-beam
evaporation source, a Kaufman source for the oxygen ion beam, and a rotary drum for the substrates.
The deposition rate of metal oxide may be monitored with a quartz-crystal thickness monitor.
Operationally, IBAD takes many forms, such as dual-ion-beam deposition, accelerated PVD, and
accelerated molecular beam epitaxy (MBE) (Hubler 1989). Other deposition methods derive benefits
from the addition of energetic particles, such as biased sputtering, ion plating, and accelerated vacuum
arc deposition. The latter methods work well for volume production but are not as well characterized in
terms of the energy and flux of the particles.
In general, Ar+ or Ne+ ions are used. However, oxygen ions can also be used. Metal oxide films pro-
duced with oxygen IBAD usually have improved adhesion as a consequence of the promotion of the
electronic processes and reactions on the bombarded surface. Cuomo et al. (1989) assumed that the en-
ergetic ion bombardment can generate significant changes in crystal size and orientation, defect density,
electrical and optical properties, chemical stoichiometry, and surface morphology. For example, Song
(1999) has shown that the increase of oxygen ion energy during SnO2 deposition at room temperature
actually leads to a change in crystal structure from amorphous to a texture of rutile SnO2, to increases
in film crystallinity and grain size. In addition, oxygen ions lead to superior metal oxide stoichiometry.
It is clear that IBAD will make a rapid entrance into the marketplace. The many advantages of ion-
beam-assisted deposition have been enumerated in several excellent reviews (Cuomo et al. 1989; Hubler
1989). The main advantages are increased film density, greater adhesion, and lower substrate tempera-
Figure 4.12. Schematic diagram of typical configuration of IBAD apparatus, which utilizes a Kaufman
ion gun.
tures. Other advantages are the ability to produce layered structures, such as multilayer films, and the pro-
duction of metastable films, but these are often overlooked. This is due in part to the present limitations
of IBAD, which include small sample areas that are uniformly bombarded with ions, a generally poor
understanding of the process, inadequate process monitoring and control, and line-of-sight processing.
7. CHEMICAL VAPOR DEPOSITION
7.1. INTRODUCTION
Chemical vapor deposition (CVD) is one of the variants of the deposition method from vapor phase.
The previously discussed physical vapor deposition (PVD) techniques relate to this method; however, in
spite of their common nature, several considerable differences exist between CVD and PVD methods.
First, in PVD methods, the gas phase usually has the same composition as the deposited material, while
in the CVD method there is a considerable difference between the composition of the gas phase and
that of the deposited film. In the CVD process, films are formed as a result of chemical reactions taking
place near or on the surface of the substrate. A second difference is that the PVD deposition process
takes place in vacuum, while in CVD, a chemically active mixture of gases serves as a deposition agent
and needs to be activated to begin the deposition process of a stable solid product. For activation of the
deposition process, heat, light, or plasma can be used (Glocker and Shan 1977; Sherman 1987; Pierson
1992; Hitchman and Jensen 1993).
CVD is a relatively mature technique that was first used in 1893. Around this period, the CVD pro-
cess was developed as an economically viable industrial process for the preparation of many advanced
products, including bulk materials, coatings and films from metals, semiconductors, metal oxides, and
composites (Blocher 1981; Spitsyn et al. 1988; Hocking et al. 1989; Pochet et al. 1997; Wahl et al.
1998; Nalwa 2000). The details of the CVD process are discussed in numerous comprehensive reviews
(Schuergraf 1988; Levy 1989; Morosanu 1990; Glocker and Shan 1995; Vincenzini 1995; Nalwa 2000;
Choy 2003), many of which were consulted during the preparation of this section.
7.2. PRINCIPLES OF THE CVD PROCESS
A schematic diagram of CVD coating is shown in Figure 4.13. CVD can be performed in either a
“closed” or “open” system. In a “closed” system, both the reactants and the products are recycled. This
process is normally used where reversible chemical reactions can occur with a temperature difference, for
example, in the purification of metals, or where there is a difference in the chemical activity of an isother-
mal system (e.g., pack cementation for aluminizing, chromizing, etc.). The “closed reactor” CVD process
is of less importance nowadays, since only a small fraction of CVD processes are performed using this
system. Most CVD processes utilize an “open” system, in which the reaction chemicals are removed from
the reactor after deposition and recovery of the reactants is carried out only if the expense justifies it.
The CVD system usually includes a chemical vapor precursor supply system, a CVD reactor, and an
effluent gas-handling system. There are various reactor configurations, such as horizontal, vertical, barrel,

Figure 4.13. Schematic diagram of a CVD apparatus for preparation of complex compounds.
and multiple wafers. It is necessary to keep in mind that CVD reactants and products are often corrosive,
toxic, poisonous, hygroscopic, inflammable, readily oxidizing, and have high vapor pressures. Therefore,
the postdeposition section of the reactor system must be efficient to render these chemicals harmless be-
fore disposal. Special precautions need to be taken, and toxic gas monitors are required to be installed.
In general, the CVD process is conducted using the following key steps (Nalwa 2000):
1. Generation of active gaseous reactant species

2. Transport of the gaseous species into the reaction chamber
3. Gas-phase reactions that form intermediate species:
a. At a high temperature above the decomposition temperatures of intermediate species in-
side the reactor, homogeneous gas-phase reaction forms powders and volatile by-products
in the gas phase. Powder is collected on the substrate surface and may act as crystallization
centers. The deposited film in this case may have poor adhesion.
b. At temperatures below the dissociation temperature of the intermediate phase, diffusion/
convection of the intermediate species across the boundary layer (a thin layer close to the
substrate surface) occurs. These intermediate species subsequently undergo steps 4–7.
4. Absorption of gaseous reactants onto the heated substrate and heterogeneous reaction at the gas–
solid interface (i.e., the heated substrate), which produces the deposit and by-product species
5. Diffusion of the deposits along the heated substrate surface, forming the crystallization center
and growth of the film
6. Removal of the gaseous by-products from the boundary layer by diffusion or convection
7. Transport of unreacted gaseous precursors and by-products away from the deposition chamber
7.3. CHEMICAL PRECURSORS AND REACTION CHEMISTRY
Thermal decomposition (pyrolysis), reduction, oxidation, hydrolysis, nitridation, disproportionation,

and synthesis are the main reactions involved in CVD processes. More detailed descriptions of the
various chemical reactions can be found elsewhere (Glocker and Shan 1995; Blocher 1981; Pochet et al.
1997; Putkonen et al. 2003; etc.).
According to Choy (2003), the selection criteria of a suitable chemical precursor for coating ap-
plications are that the precursor:
1. Is stable at room temperature

2. Has a low vaporization temperature and a high saturation vapor pressure
3. Can generate vapor that is stable at low temperature (i.e., before decomposing or reacting at a
higher temperature)
4. Has a suitable deposition rate, with a low deposition rates for thin-film applications (e.g., in the
semiconductor industry) and a high deposition rate for thick coating applications
5. Undergoes decomposition/chemical reaction at a temperature below the melting temperature or
phase-transformation temperature of the substrate, depending on the engineering application
6. Has low toxicity and explosivity, and is inflammable for safely handling the chemicals and dispos-
ing of unreacted precursor
7. Is cost-effective for thin-film or coating deposition
8. Is readily available commercially, at high purity
Common precursors used in the CVD process are metals and metal hydrides, halides and halo-
hydrides, and metal-organic compounds (see Table 4.3). Generally, metal halides and halohydrides are
more stable than the corresponding hydrides. The metal-organic precursors offer the advantage of
lower reaction and deposition temperatures than halides and hydrides, and they are less toxic and pyro-
phoric. Therefore, they are currently gaining wider application, especially in the deposition of semicon-
ductors and metal oxides. Precursors can be in the gas, liquid, or solid state, although gaseous precursors
are the simplest to use. When using liquid or solid sources, it is necessary to use special equipment for
the saturation of carrier gas by the vapors of those sources.
7.4. PARTICULARS OF CVD TECHNOLOGY
The study of CVD has revealed that the temperature at which the film is deposited is critical, because
it controls both the thermodynamics and the kinetics of the coating process. The kinetics of a CVD
Table 4.3. Typical precursors for CVD deposition of some metals and metal oxides
DEPOSITED MATERIAL REACTANT DEPOSITION TEMPERATURE (°C)

SnO2 SnCl2 + O2 400–450
SnCl4 + H2O 400–600
Sn(acac) + air 400–600
In2O3 In(acac) + air 300–650
InCl3 + H2O 350–550
TiO2 TiCl4 + O2 500
Isopropyltitanate 260–370
CeO2 CeIV-β-diketonate 300
Al2O3 Aluminum alcoholates 550
Al(acac) 400–550
Aluminum isopropoxide 500
SiO2 Tetraethylorthosilicate 500
SiCl4 + H2O 200
Fe2O3 Fe(acac) 400
Fe(CO)5 400
BeO Beryllium carboxylate 400
ZrO2 ZrCl4 + H2O 180–600
ZrI4 + H2O + H2O2 250–500
Zr(thd)4 + O3 275–500
Zr[OC(CH3)3]4 + H2O 150–400
Pt Pt(acac)2 350
(CH3)3CH3CpPt 350
PdO Pd(acac) 300–500
PdCl2 300–500
Ni Ni(CO)4 + H2S 100
Mo Mo(CO)6 350–600
W W(CO)6 350–600
WF6 + H2 400
involves processes such as chemical reactions in the gas phase and on the substrate surface, chemisorp-
tion, and desorption. The overall reaction rate is limited by the slowest reaction step. The deposition
temperature must be achieved and maintained in order for the reaction to occur on the substrate and
not in the gas phase, and with an appropriate microstructure (e.g., grain size and shape). Small changes
in temperature (e.g., ±25°C) can potentially change the reaction and/or its kinetics, resulting in an in-
ferior coating.
The ability of the reactant gases to reach the substrate surface is important in determining the unifor-
mity of the coating. Transportation of the reactant gases to the substrate surface is controlled by the par-
tial pressure, the total reactor pressure, the reactor geometry, and the substrate architecture. Optimizing
the gas flow to supply the reactant to the substrate is crucial for achieving satisfactory deposition.
CVD processes can be carried out from atmospheric pressure to high vacuum. To reduce the
dependence of the growth rate and the film composition on the hydrodynamics in the CVD reactor,
many CVD processes are carried out at total gas pressures well below 1 atm (Levy 1989). Under these
circumstances, chemical reactions become more important in determining the characteristics of the
deposited films. The deposition rate can thus be described by Arrhenius’s law:
⎛ E ⎞
Vdep = A exp ⎜ − a ⎟ (4.3)
⎝ RT ⎠
where A is a constant that depends on the parameters, including reactor pressure, gas flow rate, gas
characteristics, reactant gas concentration, and reactor geometry; Ea is the apparent activation energy; R
is the gas constant; and T is the deposition temperature.
While transport through the gas phase is rapid and relative to the rate of reaction, the concentra-
tion gradient perpendicular to the surface is small and the process is believed to be kinetically controlled.
However, as the substrate temperature rises, the interdiffusion of reactants and reaction products in the
gas phase becomes limited. In these conditions, an appreciable concentration gradient exists in the gas
phase normal to the surface. Since the concentrations of reactants and reaction products at the surface
tend toward equilibrium, the deposition process becomes much less dependent on temperature and is
said to be diffusion-limited.
As previously stated, CVD and PVD methods are similar in nature. Therefore, many of the regu-
larities established for PDV are being applied in film deposition with CVD. In CVD processes, the film
structures also strongly depend on deposition parameters. They can be very porous or gas-impene-
trable, amorphous or polycrystalline. Furthermore, all of the previously mentioned structures can be
obtained on any thermally stable substrate. For example, it is only necessary to use monocrystalline
substrates, having a certain crystallographic correspondence with the deposited layer, to obtain epitaxial
layers (Arthur 2002). In particular, single-crystal sapphire Al2O3(0001) wafers are good substrates for
SnO2 heteroepitaxial deposition (Poirier et al. 1993; Lee et al. 2001), and yttrium-stabilized zirconia
(YSZ)(100) wafers are optimal substrates for In2O3 and ITO heteroepitaxial deposition (Kwok et al.
1998). Basic regularities in the influence of deposition mode on films parameters, established as a result
of extensive research, are presented in Figure 4.14. They demonstrate once again that the properties of
deposited films are very sensitive to the deposition parameters.
It has been shown that the depletion of reactants during deposition can result in a nonuniform
coating thickness, and the composition variation across the coating parallel to the direction of reac-
tants can be overcome by pulsing the gaseous reactant. It has also been found that, to improve coating
uniformity, it is necessary to translocate/rotate the substrate, to improve precursor mixing by stirring
the reactants and/or periodically reversing the gas flow direction, tilting the substrate (e.g., ~45°) to en-
hance the projection of downstream substrates into the boundary layer, and/or creating a temperature
gradient across the substrate.
7.5. ADVANTAGES AND DISADVANTAGES OF CVD

Choy (2003) and Pochet et al. (1997) give the following distinctive advantages of the CVD process:

Figure 4.14. Basic regularities of the influence of deposition parameters on the structural parameters of deposited films.
• CVD can be environmentally friendly, with easily neutralized waste products.

• CVD has the ability to coat complex shapes internally and externally because it is a non–line-of-
sight process with strong throwing power. Therefore, it can be used to uniformly coat complex-
shaped components and deposit films with sufficient conformal coverage.
• CVD can produce uniform films with strong reproducibility and adhesion at reasonably high
deposition rates.
• CVD can produce multilayered coatings and coatings for a variety of metals, alloys, and com-
pounds that are not readily obtainable by other means.
• CVD provides the ability to control crystal structure, surface morphology, and the orientation of
the products by controlling the process parameters.
• Coatings are dense, and their purity can be controlled.
There are also process disadvantages that need to be considered when choosing this process. Some
of these are as follows (Pochet et al. 1997; Choy 2003):
• Up-front capital costs can be high, with complex handling, safety, and automatic systems.
• High temperature requirements may limit substrate choices.
• Some substrates can be attacked by the coating gases.
• Poor adhesion or lack of metallurgical bonding is possible.
• Masking portions that are not to be coated can be difficult.
7.6. VARIANTS OF CVD METHODS
7.6.1. Thermally Activated Chemical Vapor Deposition

Thermally activated CVD is a conventional CVD process in which the chemical reactions are initiated
by thermal energy in a hot-wall or cold-wall reactor using inorganic chemical precursors. The thermal
energy can be provided in the form of RF heating, infrared radiation, or resistive heating. The heating
or cooling is usually performed at a normal rate.
CVD of oxide films is typically carried out at temperatures ranging from 325 to 900°C. Conventional
CVD reactor systems are relatively simple to operate, economical, and effective, but they require a great
deal of cleanup maintenance to minimize any possible contamination from particulates. This method
is acceptable for many applications that do not require highly conformal step coverage or the lowest
possible particle density.
7.6.2. Plasma-Enhanced Chemical Vapor Deposition (PECVD)

Plasma-enhanced chemical vapor deposition (PECVD) is also known as glow-discharge chemical vapor
deposition (Oehrlein 1997). Supplying electrical power at a sufficiently high voltage to a gas at reduced
pressures (<1.3 kPa) results in the breaking down of the gas and generates a glow-discharge plasma
consisting of electrons, ions, and electronically excited species (Moisan and Pelletier 1992). The vapor
reactants are ionized and dissociated by the electron impact. This generates chemically active ions and
radicals that undergo a heterogeneous chemical reaction at or near the heated substrate surface and
deposit the thin film.
The main advantage of PECVD over other CVD methods is that the deposition can occur at
relatively low temperatures and on large areas. The PECVD process can deposit any metal, ceramic, or
alloy on virtually any substrate, including plastics, integrated circuits, metals, and ceramics. It also offers
flexibility in the microstructure of the film, and the deposition rate can be controlled separately. The ion
bombardment can contribute to obtaining the required film density.
The main disadvantage of PECVD is the appearance of surface damage, resulting from the ion bom-
barding during the film-deposition process. The other drawback of PECVD is that it requires the use of a
vacuum system to generate the plasma and a more sophisticated reactor to contain it. Therefore, PECVD
is often more expensive than thermally activated CVD. However, PECVD may find applications where the
value of the technology outweighs the cost of fabrication, and where low deposition temperatures are re-
quired for temperature-sensitive substrates, a feature that cannot be met by conventional CVD mehods.
7.6.3. Metal-Organic Chemical Vapor Deposition
Metal-organic chemical vapor deposition (MOCVD) is a variant of CVD that is classified according to
its use of metal-organics as precursors (Glocker and Shan 1995). Compounds containing metal atoms
bond to organic radicals and become “metal-organics.” Metal-organic compounds that are suitable as
precursors for most main-group and transition-metal elements are now available from a number of
commercial sources (Morton International 1993). Some examples of commercially available or labora-
tory-prepared precursor compounds used in the MOCVD of oxides are listed in Table 4.4. Information
from various references (Sladek and Gilbert 1972; Bradley et al. 1978; Schierber et al. 1991; Xie and Raj
1993; Zhang et al. 1994; Peng and Desu 1992; Si et al. 1994; Kim et al. 1995; Broorse and Burlitch 1994;
Gardiner et al. 1994; Chour et al. 1994; Xu 1997; Choy 2003) was used in preparing this table.
The metallic elements are usually in their correct oxidation state for diketonates, alkoxides, and
carboxylates; thus, they do not require the presence of oxygen during deposition. It is necessary to
remember that, when alkylmetals are used in MOCVD, an oxidizing environment must be used. This is
sometimes dangerous because of the ignitability and toxicity of most alkylmetal compounds.
Because of the low volatility inherent to common precursors, most oxide depositions are achieved at
reduced pressures, ranging from a few torr to a few tens of torr. Occasionally, atmospheric vapor deposi-
tion becomes feasible when the precursor is sufficiently volatile. Such is the case for ZnO and SnO2 depo-
sition using Zn(acac)2 and Sn(acac) as a precursors. The catalytic effect of H2O in alkoxide pyrolysis sig-
nificantly enhances the deposition rate and the deposition quality of oxides. In previous research, Al2O3,
Nb2O5, Sb2O3, TiO2, and ZrO2 films have all been deposited through vapor-phase hydrolysis of alkoxides
(Sladek and Gilbert 1972; Bradley et al. 1978). The composition of the deposited films is highly depen-
dent on the individual precursor partial pressure immediately above the substrate surface. Therefore, one
important aspect of multicomponent oxide deposition is control of the vapor-phase composition.
MOCVD can be used to deposit a wide range of materials, including semiconducting materials,
metallic films, dielectric films, and conductive metal oxides in the form of amorphous, epitaxial, and
Table 4.4. Metal-organic compounds for metal oxide deposition
ALKYLATE,
METAL DIKETONATE ALKOXIDE CARBOXYLATE OXIDE EXAMPLES
Main group
IA (Li) thd Oeth, n-Obut LiTaO3, LiNbO3
IIA (Mg, Sr, Ba) thd Oeth Bi2Sr2Ba2Cu3Ox, MgAl2O4
IIIA (Al, In,Ga) acac Oeth Benzonate In2O3, Al2O3, MgAl2O4, Ga2O3
IVA (Si, Sn, Pb) acac, thd Oeth TEPb, eha SiO2, SnO2, PbTiO3, etc.
VA (Bi) Phenyl Bi2Sr2Ba2Cu3Ox
Transition metals
IB(Cu) thd Bi3Sr2Ba2Cu3Ox
IIB (Zn) acac Ethyl ZnO
IIIB (Y, La, Eu) thd, tfa Eu2O3, Yba2Cu3O6, etc.
IVB (Ti, Zr) thd, tfa, acac n-Obut TiO2, ZrO2, etc.
VB (Nb, Ta) n-Obut, Oeth LiTaO3, LiNbO3, Ta2O5, etc.
VIB (Cr) acac Cr2O3
VIIB (Mn) acac Mn2O3, Lax(Sr,Ca)y MnO3
VIIIB (Co) acac LaSrCoO3
thd, 2,2,6,6-tetramethyl-3,5-heptanedionate (β-diketonates); acac, acetoacetonate; Oeth, alkoxide (e.g.,
ethoxide); n-OBut, butoxide; eha, ethylhexanoate.
polycrystalline films. In general, metal-organic precursors have lower decomposition or pyrolysis tem-
peratures than halides, hydrides, or halohydrides, thus enabling the MOCVD process to perform at a
lower deposition temperature than conventional CVD, which generally uses halides or hydrides.
Besides oxygen, it has been established that ozone can also be used as an effective oxygen source
for metal oxide deposition. Typically, ozone is the preferred oxygen source for β-diketonate precursors,
which need a highly reactive oxidizer to produce films with low impurity content. For example, while a
Ga(acac)3/H2O combination produced Ga2O3 films containing over 30 at% of carbon, films deposited
by the analogous ozone system had a carbon content of only 1 at% (Nieminen et al. 1996). On the other
hand, either water or hydrogen peroxide can be used as a source of oxygen if the metal precursor—a
halide or organometallic compound, for example—has sufficiently high reactivity.
At the same time, it is important to note that metal-organic precursors tend to be more expensive
compared to halides, hydrides, and halohydrides, and they are not widely available commercially for
some coating systems. Furthermore, most metal-organics are volatile liquids and thus require accurate
pressure control.
7.6.4. Microchemical Vapor Deposition of Metal Oxides
The microchemical vapor deposition method was developed for local deposition of metal oxides that
have high chemical stability during the process of elaboration of solid–gas sensors (Semancik et al.
1996; Majoo et al. 1996). Research has shown that а powered microheater can be used to localize the
deposition of a sensing film onto its associated micro-hotplate using chemical vapor deposition. When
CVD gases are set to flow over an array, only heated micro-hotplates initiate the reactions to deposit a
film. Simply turning off the micro-heater can stop deposition.
This self-lithography is an easy method for depositing different sensing materials on different
micro-hotplates. It is advantageous because the process is simple, requires minimal contamination of
the surface, avoids the reduced yield associated with multiple photolithography steps, and allows for
high-temperature processing of the sensing film to optimize the properties of the materials. The fact
that electronic switching of heater currents and straightforward gas handling replace multiple mask-
ing, photolithography, and lift-off procedures makes the approach both convenient and cost-effective,
especially when many active films are being incorporated into an array device. Surveys of processing–
property relationships can be performed in a highly effective manner to determine optimal fabrication
methods for multiply active material arrays.
An electrical contact to the growing film allows conductance measurements to be taken during
deposition. An electrical contact during growth serves as an in situ characterization of the growth. This
method has been used successfully to form various SnO2– and Pt-Ti–based sensors (Semancik et al.
1996; Majoo et al. 1996) and has been found to be very promising for the fabrication of various sensor
arrays aimed at “electronic nose” applications.
7.6.5. Photo-Assisted Chemical Vapor Deposition
The problem of local deposition can also be resolved using photo-assisted chemical vapor deposition
(PACVD). PACVD is a process that relies on the absorption of light to raise the substrate temperature
and cause thermal decomposition of the precursor in the gas phase and/or substrate surface (Hitchman
and Jensen 1993; Glocker and Shan 1977; Tamura et al. 1999) and can be performed at atmospheric
or reduced pressure (e.g., 0.01–1 atm) (Tamura et al. 1999). Unlike the thermally heated reactor in the
conventional CVD process, there is a transmitting window for the excitation wavelength in the deposi-
tion chamber, which allows for optical access. A variety of light sources can be used for PCVD, includ-
ing arc lamps, CO2 lasers, Nd-YAG lasers, excimer lasers, argon-ion lasers, and excimer UV lamps (from
354 to 126 nm).
Decomposition and/or chemical reactions can be initiated via a photothermal (pyrolytic) or pho-
tolytic mechanism. The photothermal mechanism is often used for selected-area deposition, which can
be achieved either by laser scanning or projection imaging using a pulsed laser source. The use of pulsed
laser sources can provide localized surface heating without heating the bulk substrate, this avoiding any
thermal damage to the substrate while achieving the sufficiently high transient temperatures required
for photothermal decomposition. Another aspect of the photolytic mechanism is that the wavelengths
(<250 nm) in the UV spectrum can cause the non–thermal-decomposition formation of radicals from
chemical precursors. This allows the deposition to occur at a lower temperature than with photother-
mal decomposition. It is necessary to note that direct photolysis is possible only if the gaseous reactant
molecules absorb resonant energy from the electromagnetic radiation of a specific wavelength, causing
decomposition and a chemical reaction.
Several possibilities for application of UV sources have already been identified, including material
deposition, polymer etching, and surface modification. Experimental results have shown that excimer
V-UV and UV sources are ideal for initiating photopolymerization processes because they offer rapid
polymerization at ambient temperature. For example, this method was used for the deposition of poly-
pyrrole (PPY) conducting polymer thin films with thicknesses ranging from 100 to 2000 nm (Fang et
al. 2002).
The following key advantages of PACVD provided the motivation for the development of this
method:
• The possibility of local deposition

• A low deposition temperature that minimizes the dopant and interlayer diffusion, defect forma-
tion, and thermal stress that are created with high-temperature processing
• Low excitation energies (typically <5 eV), which leave the deposited films free from the ion-
bombardment damage that may be encountered with plasma-assisted CVD
• Better potential for reducing unwanted reactions compared to plasma-enhanced or thermally
activated CVD, due to the narrow energy distribution of photons emitted from lasers or a spec-
trally filtered lamp
8. DEPOSITION FROM AEROSOL PHASE
8.1. INTRODUCTION
The design of effective technological methods for the deposition of complex metal oxide phases has
been the focus of numerous articles dealing with the study of solid-state chemical sensors. In many
cases, however, traditional approaches have not been effective. Thin-film deposition methods, such
as CVD, MOCVD, molecular beam epitaxy, magnetron sputtering, or vacuum evaporation, are very
cumbersome and inefficient for the deposition of complex materials. In most methods, for example, it
is necessary to have three or more vaporizers and gas channels, complicated vacuum and gas-handling
systems, or chemically stable precursors with high volatility. Furthermore, throughout the deposition
process, it is necessary to keep the gas flow rates and evaporation rates from every vaporizer at fixed
levels with high accuracy. In the case of magnetron sputtering, an expensive, custom-made target with
predetermined elemental composition is needed, and films with only this fixed composition can be de-
posited using this target. The complexity and high cost of these traditional film deposition techniques
severely limits the ability to experiment systematically with the modification of film composition over
wide ranges and to quickly identify film compositions with optimum sensing characteristics. This limits
the ability to study systematically the nature of chemical sensing of metal-oxide thin films as well.
For deposition of such complex metal oxides, using the deposition from precursor materials in the
form of aerosols has been suggested as an alternative to using vapor-phase deposition or sol-gel methods
(Lefebvre 1989). The application of aerosol particles to the film deposition process simplifies the tradi-
tional multisource process, from requiring several precursors to needing only a single-source deposition
process. Because of the constant concentration of precursors in the solution during the entire deposition
process and the small dimensions of the aerosol droplets, these aerosol particles may be considered as a
single component precursor. Sol-gel methods offer the possibility to deposit complex metal-oxide films
from one source (Brinker and Scherer 1990; Ivanovskaya and Bogdanov 2001). Such methods, however,
include multistage, lengthy processes and are scarcely involved in modern sensor technologies.
Previous research has shown that thin films prepared using the spray pyrolysis (SP) method have
useful properties for a wide variety of potential applications, including chemical sensors (Tarasevich et
al. 1981; Leo et al. 1999; Oktik 1988; Tomar and Garcia 1981; Lalauze et al. 1991; Labeau et al. 1994;
Chen et al. 1998; Leo et al. 1999; Korotcenkov et al. 1999, 2001, 2004b–2004d). It has also been estab-
lished that chemical SP offers a highly competitive alternative for the production of various thin-film
coatings, including thin films of noble metals, metal oxides, spinel oxides, and complex compounds
(Kadam and Patil 2001; Castaneda et al. 2002; Mahmand 2002; Todorovska et al. 2002; Perednis and
Ganckler 2005). For example, this technique was successfully employed in the deposition of MgAl2O4
and LiTaO3 from the nonvolatile and thermally unstable alkoxide complexes MgAl2(OC2H5)8 (Zhang et
al. 1994) and LiTa(OC2H5)6 (Xie and Raj 1993). The general applicability of such liquid-delivery evapo-
ration systems has also been demonstrated for platinum (Goswami et al. 2003) and other oxide systems,
such as BaTiO3, YBa2Cu3O7-d, YSZ, and LaSrCoO3, using mixtures of the respective precursors in
the solution (Gardiner et al. 1994). Some review articles pertaining to SP processing and the range of
thin films deposited for various applications have also appeared in the literature (Blandenet et al. 1981;
Lefebvre, 1989; Lavernia and Wu 1996; Patil 1999).
8.2. PRINCIPLES AND MECHANISM OF THE DEPOSITION PROCESS
Deposition from an aerosol phase is basically a chemical deposition process in which fine droplets of
the desired material solution are sprayed onto a heated substrate. Figure 4.15 shows a typical system
used for spray pyrolysis.
The main components of this system are a spray nozzle (or ultrasonic atomizer), a solution tank,
a pyrolysis furnace–substrate heater, a temperature controller, an air compressor or gas propellant, and
an exhaust system for the gases. The pyrolysis furnace may be cylindrical (carousel type) or linear (run-
through-conveyor type) in shape. The exhaust system is a depressurizing system that must satisfy two
aims. First, for safety reasons, it must remove the effluent gases, because the chemicals used may be
toxic. Second, it must ensure that the gas mixture circulates near the substrates to be coated, particu-
larly with a run-through-conveyor-type setup. To measure the flow of the precursor solution and air,
liquid and gas flow meters are used. Vertical and sloping spray deposition arrangements with either a
stationary or a linearly moving spray nozzle are also frequently used in this technique. To achieve uni-
form deposition, moving arrangements have been used for the nozzle, the substrates, or both. Other
approaches for improving the spray pyrolysis technology can be found in Patil (1999).
The spray pattern and droplet size, as a function of the process parameters, can be determined
and monitored using a “high-spec” imager and a droplet-size analyzer (e.g., a Malvern Mastersizer),
respectively.
Figure 4.16 illustrates the various decomposition modes of a solution in relation to temperature.
In essence, the process consists of two stages, droplet drying and precursor decomposition. If these
Figure 4.15. Basic scheme of apparatus for metal-oxide film deposition by pneumatic spray pyroly-
sis: 1, air compressor; 2, spray atomozer; 3, substrate holder; 4, reaction chamber; 5, furnace; 6,
metering tank; 7, substrate. (Reprinted with permission from Korotcenkov et al. 2001b. Copyright
2001 Elsevier.)
stages take place before the droplet reaches the substrate, spray pyrolysis can be understood as a chemi-
cal vapor deposition with gaseous species reacting out and reacting at the surface. If they take place
after the droplet reaches the substrate, then the SP is understood as a hydrothermal growth taking place

Figure 4.16. Decomposition as a function of deposition temperature and droplet size. (Adapted with
permission from Blandenet et al. 1981. Copyright 1981 Elsevier.)
on the surface covered by a thin liquid layer and separated from the substrate by a thin layer of vapor
phase. The outcome will depend on the specific heat of the solution, the amount of liquid reaching
the film per unit time, and the substrate temperature. It is clear that aerosol droplets behave differently
depending on their size (Sears and Gee 1988). Therefore, it is easy to understand that, if the deposition
conditions are set so that the major fraction of the aerosol is a gas when it contacts the substrate, the
larger droplets will not have vaporized completely, whereas the smaller droplets will have already been
decomposed by pyrolysis. This means that different reactions, such as vapor–solid, liquid–solid, and
solid–solid, can take place simultaneously during deposition from an aerosol phase.
8.3. ATOMIZATION TECHNIQUES
The critical operations for deposition from an aerosol phase are the preparation of uniform and fine
droplets, and the controlled thermal decomposition of these droplets in terms of environment, loca-
tion, and time. Therefore, the method of atomization used during film deposition from an aerosol
phase is a main factor in controlling the conditions of film growth.
8.3.1. Spray Nozzle

The spray nozzle consists of a pipe having a diameter large enough to prevent the droplets from co-
alescing. It directs the aerosol into the pyrolysis zone, and its temperature is regulated to prevent prema-
ture vaporization of the atomized solution. Varying the atomization gas flow rate allows the droplet-size
distribution of the spray to be modified. According to the general relationship developed for twin-fluid
atomizers (Lavernia and Wu 1996), an increase of the atomization gas flow rate ( Qg ) for a given liquid
flow rate ( Ql ) lowers the average droplet size (Ddroplet ) of the spray, as shown in Eq. (4.4):
3/2
⎛Q ⎞
Ddroplet ~ ⎜ l ⎟ (4.4)
⎜ Qg ⎟
⎝ ⎠
However, such nozzle atomizers are not sufficient for obtaining reproducible micrometer or submi-
crometer-size droplets, or for controlling their size distribution.
8.3.2. Ultrasonic Nebulized Atomization

To obtain uniformly distributed micrometer- and submicrometer-size droplets, an ultrasonic nebulizer
has been used (Chen et al. 1998). The histograms in Figure 4.17 illustrate the droplet-size distribution
in the two types of spraying, namely, ultrasonic and pneumatic (Blandenet et al. 1981). As shown, the
ultrasonic nebulizer produces a more narrow size distribution in comparison with pneumatic spraying.
The precursor solutions are usually vaporized with an ultrasonic nebulizer (the ultrasonic power
of the mist generator is about 100–300 W) operated at a frequency of 1.0–2.56 MHz. Blandenet et
Figure 4.17. Droplet spectrum for (1) pneumatic spraying and (2) ultrasonic spraying (volume
distribution), frequencies of 0.1 and 3 MHz. (Adapted with permission from Blandenet et al. 1981.
al. (1981) indicated that this power is sufficient for spraying approximately 1 liter h−1 of liquid at a
frequency between 0.8 and 1.0 MHz. By focusing an intense ultrasonic beam close to the surface of a
liquid, an ultrasonic “geyser” is formed. The joint effects of cavitation within the liquid and vibrations
on its surface lead to the formation of the aerosol. Such droplets have very small sizes with a narrow
size distribution and no inertia in their movement, so they can be transported by the carrier gases with-
out heating. The properties of this aerosol will depend on the nature of the liquid, the frequency, and
the intensity of the ultrasonic beam. Lang (Blandenet et al. 1981) further established the relationship
between the size of droplets and the parameters of the ultrasonic nebulizer and the sprayed solution:
1/3
⎛ 2 πγ ⎞
Ddroplet ≈ k⎜ 2 ⎟ (4.5)
⎝ ρf ⎠
where k is a constant, γ is the surface tension, and ρ is the density. In other words, the diameter of the
droplets is a function of the ultrasonic frequency, f. For example, the average diameter of droplets for
water is 4.5 μm at 800 kHz, and for butanol it is 3.6 μm at the same excitation frequency. The aerosol
phase that is generated is transported by the carrier gas or by air through a pipe to the heated substrate.
The basic scheme of the apparatus used for ultrasonic spray deposition is shown in Figure 4.18. The
advantage of this method is that the gas flow rate is independent of the aerosol flow rate, which is not
the case with pneumatic spraying.
8.3.3. Corona Spray Pyrolysis

The most significant limitation of the conventional pneumatic spray pyrolysis technique is its low depo-
sition efficiency, defined as the ratio of atoms effectively deposited to those supplied. Lampkin (1979)
showed that the presence of an electric field across the aerosol stream improves the deposition process.
Gas flow controller Aerosol Pyrolysis zone
Aerosol
Substrate
Purified air
Solution Substrate heater
H.F. Generator Piezoelectric ceramic
Figure 4.18. Basic scheme of apparatus for ultrasonic spray deposition. (Adapted with permission
from Labeau et al. 1994. Copyright 1994 Elsevier.)
This conclusion is consistent with results reported by Siefert (1984), who demonstrated that the depo-
sition efficiency can be increased by up to 80% with the aid of an electric field superimposed on the
stream. In this process, the aerosol droplets are charged by the corona current and accelerated along the
lines of the electric field force to the substrate, which is acting as a counter electrode. Thus, transporting
the droplets to the substrate can be more readily controlled using corona spray pyrolysis, because the
charged droplets must follow the electric field lines.
8.3.4. Electrostatic Spray Pyrolysis

The electrostatic spray pyrolysis (ESD) technique was developed several years ago to prepare various
metal oxide films for energy conversion, storage systems, and chemical gas sensors (Chen et al. 1998).
In this technique, a positive high voltage of up to 12.5 kV is applied to a nozzle (a hollow needle or
capillary), through which the precursor solution is forced to flow and from which a positively charged
spray is generated. When a liquid is forced to flow through a small metal nozzle that is subjected to an
electric field, the different electrodynamic mechanisms will cause the liquid to leave the outlet of the
nozzle in different forms or modes. Some of these modes include dripping mode, cone-jet mode, and
spindle mode. In which mode a spray operates depends primarily on the electric potential applied to
the nozzle and the flow rate, the electrical conductivity, and the surface tension of the liquid. A detailed
overview of this technique for the synthesis of oxide films and powders has been presented in several
comprehensive works (Gafian-Calvo et al. 1996; Choy 2000, 2003; Jaworek and Sobczyk 2008).
For film deposition, the cone-jet mode (see Figure 4.19) is the most suitable, because it is a continu-
ous mode that allows for the formation of homogeneous fine spray droplets. This cone is extended by
a jet and breaks up into spray droplets to generate an aerosol of the precursor liquid. Under mainly the

Figure 4.19. Cone-jet spraying of methanol containing a small amount of hydrochloric acid.
(Reprinted with permission from Pantano et al. 1994. Copyright 1994 Elsevier.)
electrostatic force (other forces, such as electric wind, play a minor role), the spray droplets tend to move
to the hot substrate, where pyrolysis takes place at or near the surface of the substrate. Subsequently, the
aerosol is electrohydrodynamically directed to a heated substrate to form a thin film.
A schematic diagram of the ESD system as modified by Kim et al. (2000) is shown in Figure 4.20.
The higher the electric field strength, the smaller the droplets formed at the tip of the nozzle, and the
Figure 4.20. Schematic diagram of the system for ESP. (Adapted with permission from Kim et al.
number of jets also increases with an increase in the applied voltage. The droplet diameter and the cur-
rent through the liquid cone can be estimated using Eqs. (4.6) and (4.7), as given by Fernandez de la
Mora and Loscertales (Lavernia and Wu 1996):
1/3
⎛ Q εr εo ⎞
Ddroplet = b1 ( ε r ) ⎜ ⎟ (4.6)
⎝ Gl ⎠
1/2
⎛ γQGl ⎞
I = b2 ( ε r ) ⎜ ⎟ (4.7)
⎝ εr ⎠
where, Ddroplet is the droplet diameter, Q is the liquid flow rate, Gl is the liquid conductivity, γ is the
surface tension, εr is the relative permittivity of the liquid, I is the current, and b1 and b2 are functions
of the liquid permittivity. These relationships are often called the scaling laws for electrohydrodynamic
atomization in the cone-jet mode and are valid only for liquid cones with a flat radial profile of the axial
liquid velocity in the jet. Equations (4.6) and (4.7) are applicable only when the liquid has high conduc-
tivity and viscosity.
The main process parameters for ESD of films are the deposition temperature, the field strength,
the stand-off distance between the heated substrate and the precursor atomizer, the precursor flow rate,
and the size of the spray droplets. Depending mainly on the substrate temperature and the composition
of the precursor liquid, various film morphologies can be obtained. In addition, the ESD method has
high deposition efficiency (>90%) because the precursor is directed to the substrate under the electric
field, minimizing the loss of the precursor to the surrounding. For synthesizing of nanostructured
films, the preferred deposition temperature is below 550°C, the field strength is within the range of
4–25 kV, and the precursor flow rate varies from 10 to 30 ml h−1.
8.4. ADVANTAGES AND DISADVANTAGES OF DEPOSITION FROM AN

AEROSOL PHASE
Using relatively simple fluid flow and reaction engineering models, it is possible to identify conditions
in which oxide film formation takes place only on the substrate surface and without complications from
thermal diffusion processes, polymerization, and decomposition of precursors in the vapor phase. The
deposition rate and the thickness of the films can be easily controlled over a wide range by changing
the spray parameters, thus eliminating the major drawbacks of chemical methods, such as sol-gel, which
produce films of limited thickness. Operating at moderate temperatures (100–500°C), SP can produce
films on less robust materials. Unlike high-power methods, such as radio-frequency magnetron sputter-
ing, SP causes no local overheating, which can be detrimental to the materials to be deposited. There are
virtually no restrictions on the substrate material, its dimensions, or its surface profile.
For aerosol deposition, it is possible to use much simpler, less expensive, and less toxic precursors,
which do not possess the high pressure of vapor under saturation. Precursors with poor volatility and
those inclined to polymerization may also be used. This creates the possibility to apply some salts and
metal-organic compounds that cannot be used with standard methods such as CVD and MOCVD. In
particular, this concerns the coordination compounds of some metal catalysts such as Pd and Pt, and
the whole row of rare-earth metals (Y, La, Gd, etc.).
The required equipment for aerosol deposition is both simple and safe, and the process is straight-
forward, rapid, reliable, and inexpensive. The deposition of films can be carried out at atmospheric
pressure, with a wide range of processing parameters, including the concentration of precursors, tem-
perature, and deposition rates, all of which may be controlled independently. Unlike closed vapor depo-
sition methods, SP does not require high-quality targets and/or substrates, nor does it require a vacuum
at any stage. This is a strong advantage if the technique is to be scaled up for industrial application.
This method allows the composition of deposited complex films to be controlled with high re-
producibility. In the case of mixed oxide layers, the composition is very close to the composition of
the original solution. It offers an extremely simple way to dope films with virtually any element in any
proportion, merely by adding it in some form to the spray solution. Changing the composition of the
spray solution during the spray process enables it to be used in making layered films and films having
composition gradients throughout the thickness.
There are some disadvantages of the spray pyrolysis technology, beginning with its fairly low re-
producibility, and the homogeneity of the film thickness distribution over the substrate’s area. These
problems, however, are easily resolved by substrate rotating or the aerosol stream scanning. Other tech-
nological problems are associated with the plotting of sensitive material on small areas. Low effective-
ness in the use of the precursor during the film deposition process is typical for spray pyrolysis.
8.5. TECHNOLOGY OF THE PYROLYSIS PROCESS
The source compound for a pyrolysis process must satisfy a number of essential conditions. It should be
stable at room temperature, it should not oxidize in air or in the presence of water vapor, and it should
have a decomposition temperature below 500°C but preferably higher than its boiling or sublimation
temperature. Generally, the products employed are organometallic compounds (acetyl acetonates, alco-
holates, etc.) and mineral salts (chlorides, fluorides, etc.). Typical precursors used during spray pyrolysis
of metal oxides are shown in Table 4.5.
The type of solvent is chosen in relation to the source compound in order to obtain a concentra-
tion of more than approximately 0.1–1.0 M, because below this concentration the rate of deposition
is too slow. The viscosity of the solution must also permit a high aerosol flow rate. Furthermore, it is
advisable to select a solvent that is not highly flammable, to avoid sudden combustion in the pyrolysis
furnace. The most commonly used solvents are acetylacetone, methanol, and distilled water.
The carrier gas is chosen for its suitability in relation to the chemical reaction taking place during
film deposition. For example, air may be used in the deposition of an oxide so as to supply the oxygen
needed for the reaction. In other cases, a neutral gas, such as nitrogen, argon, or helium, may be used.
The rate of film deposition depends on a number of experimental parameters, including the tem-
perature of the substrate (see Figure 4.21), the flow rates of the aerosol and the carrier gas, the speed
at which the substrate runs through, moves, or rotates in the reaction zone, the distance between the
nozzle and the substrate, the concentration of the compound in the solution, and the dopant content
Table 4.5. Common precursors and deposition parameters of sensing materials
PYROLYSIS
TEMPERATURE
PRECURSOR SOLVENT ADDITIVES (°C) SENSING MATERIAL
Acetilacetonates of
Sn, In, Co, Pd, Cu, Fe, Acetilacetone; 300–600 In2O3, SnO2, ZnO, Al2O3,
Al, Ti, Cr, V, etc. methanol; PdO; Fe2O3; Cr2O3; TiO2;
ethanol V2O3, CuO
Clorides
(SnCl4; SnCl2; InCl3; Distilled water; HCl 300–550 In2O3, SnO2, ZnO, Al2O3,
AlCl3; WCl6; ZnCl2; ethanol; alcohol; PdO; Fe2O3; Cr2O3; TiO2;
FeCl3; MgCl2; CoCl2; methanol MgO; Co2O3; CeO2; Mn2O3;
TiCl4; CuCl2, etc.) CuO; Ta2O5
Nitrates
(Co(NO3)2; Bi(NO3)2; Distilled water HNO3 300–650 Bi2O3; Co3O4; Ag2O; MgO;
Ni(NO3)2; La(NO3)2; NiCo2O4; LaSrMnO3; Mn2O3;
Sr(NO3)2; MnNO3; etc.) ZnO; In2O3; etc.
(Patil 1999; Shanthi et al. 1999; Smith 2000; Korotcenkov et al. 2001; Thangaraju 2002; Prince et al.
2002; Perednis and Gauckler 2005). Obviously, the substrate temperature is the most important factor
and must be controlled with precision. A temperature variation of a few degrees will result in consider-
able differences in the thickness of the coating.

90 90
SnO2
3
60 60
1
(SEM)
2
30
4
30
(XRD)
a) b)
0 0
300400500600 300400500 600

Figure 4.21. Relationship between Tpyr and SnO2 film parameters (VS = 2 ml): 1, 2, thickness; 3, 4,
size of crystallites; 1, 3, 4, just-prepared SnCl4–water solution; 2, aged solution (tag = 24 h); 5, ultra-
sonic spray, XRD. (Adapted with permission from Labeau et al. 1994 and Korotcenkov et al. 2005.
Copyright 1994, 2005 Elsevier.)
Figure 4.22. Influence of both volume and concen-

tration of InCl3 in the sprayed solution on the thick-
ness of the In2O3 films (Tpyr = 475°C, PS = 2.0 atm, L =
20 cm): 1, 1.0 M InCl3–water solution; 2, 02 M solu-
tion. (Reprinted with permission from Korotcenkov et
al. 2004b. Copyright 2004 Elsevier.)
The range of applications is restricted because of the limited thickness (usually less than 1–10 μm)
of the coatings obtained. Therefore, this process does not provide a solution to the problems of wear
and corrosion, which require greater thicknesses. As a rule, film thickness depends linearly on both the
value of the sputtered solution (see Figure 4.22) (Korotcenkov et al. 2001, 2004b–2004d) and the spray-
ing time (Moris and McElnea 1996). This means that the mentioned parameters can be used to control
of the thickness of deposited film. It is necessary to remember that the film thickness also depends on
the time of solution aging.
Shanthi et al. (1999) showed that the thermal behavior of the droplets depends on their mass, and
different deposition processes will occur depending on the size of the droplets. In an ideal deposition
condition, the droplet approaches the substrate just as the solvent is completely removed. It has been
shown that, depending on its velocity and flow direction, a droplet will either flatten, skip along the sur-
face, or hover motionless. Optimal deposition conditions must be determined based on the interaction
of the various parameters mentioned above.
During spray pyrolysis deposition, it necessary to remember that because cooling of the substrate
takes place, the period and volume of the sprayed solution must be fixed at 1–3 s and 0.5–2.0 ml, re-
spectively. After recovering the substrate temperature, the same operation may be carried out again.
This procedure can be repeated anywhere from 2 to 100 times for one process, depending on the
required film thickness.
8.6. REGULARITIES OF METAL OXIDE GROWTH DURING SPRAY

PYROLYSIS DEPOSITION
As a rule films, deposited at low temperatures (Тpyr < 350°С) are finely dispersed, with a grain size less
of than 1–2 nm (Lee and Lin 1996; Mahmoud 2002; Korotcenkov et al. 2000, 2002), while films grown
at temperatures higher than 350°C show the crystalline phase with corresponding diffraction peaks on
x-ray diffraction patterns. The intensity of the peaks increases with the growth temperature, indicating
enhancement of the crystallites. The thickness of the transition layer between the deposited film and
the substrate may reach 10–40 nm (Korotcenkov et al. 2000). Low-temperature films (Тpyr < 350°C)
usually consist of disoriented crystallites, whereas high-temperature films (Тpyr > 400°С) are generally
textured (Wang et al. 2000; Korotcenkov et al. 2000, 2004). The texturing is determined by the teme-
prature, the precursor used (Thangaraju 2002; Dien et al. 1999), and the type of substrate (Prince et al.
2002; Smith 2000). Using monocrystalline metal-oxide substrates promotes more oriented growth, and
films can show epitaxial properties when the substrate is optimally selected.
Grain size and film morphology depend on the temperature of pyrolysis and the thickness of the
film. Increases in these two parameters cause considerable increase in the grain size (Labeau et al. 1994;
Lalauze et al. 1991; Brinzari et al. 2002; Korotcenkov et al. 2002, 2004b–2004d) (see Figure 4.23). During
high-temperature pyrolysis, the thickness has a much stronger influence on the grain size. An increase in
precursor concentration in a sprayed solution also leads to growth in the grain size, while a decrease in
drop size results in the grain size decreasing. This means that films deposited by ultrasonic spray pyrolysis,
as a rule, have a smaller grain size than films deposited by the pneumatic spay method (Lalauze et al. 1991;
Labeau et al. 1994) (see Figure 4.21). In some experiments, it was demonstrated that the surface rough-
ness also decreases when the solution used has a lower concentration of precursor (Wang et al. 2000).
During grain-size growth, a transition from spherulites to crystallites with definite faceting takes
place. Alterations to the grain size are accompanied by a change in faceting of the grains, which usu-
ally become apparent through changes in the morphology and film texture (Korotcenkov et al. 2002,
2004b–2004d). For example, it was previously determined that, depending on Tpyr, SnO2 crystallites can
have a needled, prismatic, or bipyramidal shape. The influences of both the film thickness and pyrolysis
temperatures on SEM images of SnO2 films deposited by spray pyrolysis are shown in Figure 4.24.

Figure 4.23. Influence of film thickness on grain size of SnO2 films deposited by spray pyrolysis:
(left) grain size determined by x-ray diffraction (XRD); (right) grain size determined by scanning elec-
tron microscopy (SEM). 1, Tpyr = 350–375°C; 2, Tpyr = 450–475°C; 3, Tpyr = 510–535°C. (Reprinted
with permission from Korotcenkov et al. 2004b. Copyright 2004 Elsevier.)

Figure 4.24. Influence of deposition parameters on morphology of SnO2 films deposited by spray
pyrolysis. (Reprinted with permission from Korotcenkov et al. 2004c. Copyright 2004 Elsevier.)
Low-temperature films are inclined to agglomeration (Korotcenkov et al. 2002, 2004b–2004d),

and the size of the agglomerates may vary anywhere from 10–20 nm to 200–500 nm. A typical view of
agglomerated structures is shown in Figure 4.25. It is important to note that the agglomeration process
has unique features that also depend on Tpyr. For example, for SnO2 films it was determined that, if in
the low-temperatures range (<360°C) agglomerates are formed from randomly oriented crystallites in a
spherelike envelope, then at higher temperatures, close to the second boundary (~400–450°C), a mutual
orientation of crystallites forming agglomerates takes place. As the transmission electron microscopy
(TEM) images shown in Figure 4.26 indicate, the SnO2 films deposited at Tpyr < 400°C have a granular
a b
Figure 4.25. Typical SEM images of agglomerated metal oxide films. (Reprinted with permission
from Brinzari et al. 2002. Copyright 2002 Elsevier.)
(a)
SnO2
Si substrate SiO2
(b)
SnO2 layer
(c)
SnO2 Figure 4.26. TEM cross-section micro-
graphs of SnO2 films deposited by spray
pyrolysis: (a) Tpyr = 330–350°C, d ≈ 70–80
nm; (b) Tpyr = 475°C, d ≈ 300 nm; (c) Tpyr
= 510°C, d ≈ 75–100 nm. (Reprinted with
Si substrate permission from Korotcenkov et al. 2004c.
structure. High-temperature films usually have a columnar structure whose density increases with an
increase in pyrolysis temperature (Korotcenkov et al. 2002, 2004). The SnO2 columnar structure of the
crystals grown at Tpyr > 400°C can be observed in the cross-section TEM micrographs in Figure 4.25.
The film’s gas-permeability is determined mainly by the thickness and the deposition temperature.
It is important to note that the regularities analyzed in this section are characteristic of all methods
of metal oxide deposition. Therefore, they can be applied for optimization of the structural parameters
of metal oxides during deposition by other methods as well.
9. DEPOSITION FROM AQUEOUS SOLUTIONS
9.1. INTRODUCTION
Techniques for the synthesis of ceramic thin films from aqueous solutions at low temperatures are
emerging as possible alternatives to vapor-phase and chemical-precursor techniques. Lower tempera-
tures allow films to be deposited on substrates that may not be chemically or mechanically stable at high
temperatures. Unlike vapor-phase processes, techniques using liquids as the deposition medium do not
rely on line-of-sight deposition, which allows nonplanar substrates to be coated. The equipment for liq-
uid-based techniques is simple and much less costly than, for example, vacuum systems and glove boxes.
Finally, with aqueous solutions and their readily available reagents, there is reduced reliance on expensive
or sensitive organometallic precursors, and the potential for a reduced environmental impact, in com-
parison with many chemical routes. Single-crystalline films can be obtained via either aqueous or organic
routes, as recently reviewed by Lange (Sol and Tilley 2001; Vayssieres 2007), but all require a thermal
decomposition step at temperatures usually between 150 and 500°C. On the other hand, the chemical
versatility of sol-gel and polymer-pyrolysis routes has enabled them to be utilized in the synthesis of
virtually any ceramic material of technological interest (Epifani et al. 2006; Korotcenkov et al. 2006).
The main techniques for synthesis of polycrystalline ceramic films from aqueous solutions at low
temperatures are as follows (Tolstoy 1977; Niesen and Guire 2001, 2002):
• Chemical bath deposition (CBD

• Selective ion-layer adsorption and reaction (SILAR) or successive ionic-layer deposition (SILD)
• Liquid-phase deposition (LPD)
• Electrodeless deposition (ED)
• Electrochemical deposition under oxidizing conditions (EDOC)
Some of the materials which can be produced using methods of deposition from aqueous solu-
tions are presented in Table 4.6.
9.2. CHEMICAL BATH DEPOSITION (CBD)

The term chemical bath deposition is used here to describe techniques that produce a solid film from a
single immersion by controlling the kinetics of formation of the solid. Many sulfides, selenides, and
Table 4.6. Methods of metal oxide film deposition from aqueous solutions
METHOD OF DEPOSITION METAL OXIDES DEPOSITED USING THIS METHOD

Chemical bath deposition CoO, Co3O4, NiO, AgO, Ag2O, ZnO, CdO, In2O3,
SnO2, Cd2SnO4, V2O5, ZrO2, etc.
Liquid-phase deposition SiO2, TiO2, ZrO2, V2O5, FeOOH/Fe2O3, Cu2O, ZnO,
NiFe2O4, LnMO3 (Ln = La, Nd; M = Cr, Mn, Fe, Co),
SnO2, Cr2O3, In2O3, Ta2O5, Nb2O5, etc.
Electroless deposition (ED) with catalyst (Ag+, Sn+2,or Pd+2) - MnO2, WO3, La1−xMnO3, ZnO,
In2O3, Tl2O3, PbO2, Li4Ti5O12, Cu2O, TiO2
Electrolytic deposition ZnO, Mn3O4, CuO, Fe2O3, SnO2, In2O3, CdO, RhO2,
ZrO2, NiO, Co3O4, CoFe2O4, In2O3/SnO2, TiO2, SmO,
RhO, Re2O3, TiO2, RuO2, Cr2O3, Al2O3, CeO2, Y2O3
Electrophoretic deposition TiO2, Al2O3, ZnO, In2O3, CeO2, Al2O3-Pd, ZrO2, NiO,
SnO2, MgO, LaMnO3, LaGaO3, CaSiO3, (LaSr), MnO3
Successive ionic-layer deposition or selective MnO2, FeOOH/Fe2O3, CuO, ZnO, ZnO:(Ni;Cu;Cr),
ion-layer adsorption and reaction (SILD or Tl2O3, NiO, SnO2, LaNbOx, CeO2, SnWO, PbO2,
SILAR) SnWPO, Ce4MnOx, PdO, PbO2, TiO2
oxides have been prepared using this technique (Lokhande 1991; Nair et al. 1998; Savadogo 1998; Mane
and Lokhande 2000).
Niesen and Guire (2001, 2002) give the following explanation of the CBD principle. The deposi-
tion medium for CBD consists of one or more salts of metal Mn+, a source for the chalcogenide X
(X = O, S, Se), and usually a complexing agent (e.g., ammonia, NH3) in an aqueous solution, while the
water itself provides oxygen, initially in the form of OH− ions. The processes that occur in the CBD
solution generally consist of the following steps: (1) equilibrium between the complexing agent and
water; (2) hydrolysis of the chalcogenide source; (3) formation/dissociation of the ionic metal–ligand
complexes [M(L)i ]n-ik, where Lk− denotes one or more ligands; and (4) formation of the solid.
As applied to the CBD of oxide films from a metal cation Mn+ complexed by i ligands Lk−, these
steps can be formulated as follows:
1. Complexant-H2O equilibrium:
2Lk − + 2H2O ↔ 2LH − k +1 + 2OH − (4.8)
2. Dissociation of water:
nH2O ↔ nOH − + nH + (4.9)
3. Displacement of ligands:
nOH − + M(L)(i n − ik )+ → M(OH)n ( s ) + iLk − (4.10)
4. Deprotonation to form oxide:
n
M(OH)n ( s ) → MOn/2 ( s ) + H2O (4.11)
2
yielding the net reaction:
n
M(L)(i n − ik )+ + H2O → MOn/2 ( s ) + nH + + iLk − (4.12)
2
The complexant (urea, NH3, KCN, etc.), when used, is usually chosen for the affinity of its ligands
toward the metal. This may make step 3 the rate-determining step, thus slowing down the rate of solid
formation, and therefore providing a degree of control over it.
The hydrolytic process depicted by reactions (4.11) and (4.12) is sometimes called forced hydrolysis.
It can be accelerated simply by heating the solution, which induces deprotonation of the hydrated
metal species [reaction (4.11)]. Hydrolysis can occur even in acidic solutions when the metal cation
is easily hydrolyzed. In contrast, the CBD of non-oxides has been carried out almost exclusively in
basic solutions.
In all CBD processes, whether oxide [reaction (4.12)] or non-oxide, a solvated metal complex
reacts with a chalcogenide source to form the desired solid product. The main difference is that, for
non-oxide films, the supply of chalcogenide anions, and thus the rate of hydrolysis, can be more readily
controlled through adjustments in the pH, temperature, and concentration of the chalcogenide source.
In the CBD of oxides, the ‘‘chalcogenide source’’ is water, so tighter control must be exerted over just
two parameters (pH and T ) to achieve a similar degree of control over the rate of hydrolysis and, by
implication, control over the film’s microstructure and properties.
One of the often-stated advantages of chemical bath deposition is that it can uniformly coat
substrates of many different natures and essentially any shape. The latter is a consequence of the fact
that most CBD processes are not diffusion-controlled. In common with other deposition techniques,
rough surfaces (rough on the microscale) tend to give more adherent films than corresponding smooth
surfaces (Hodes 2003). This is probably due to the greater actual surface area of contact per geometric
surface area and to the possibility of anchoring the initial deposit in pores of the substrate.
Glass is one of the most commonly-used substrates in CBD, and even though it is a relatively inert
material, its surface can be very reactive toward species in solution. The surface of glass, along with oth-
er oxides such as SnO2-coated glass, is hydroxylated in an aqueous solution and the surface hydroxide
groups can form fairly strong hydrogen bonds. A great deal of research has shown that glass substrates
can be sensitized on the surface, usually with a solution of SnCl2, which hydrolyzes to produce nuclei of
tin hydroxide or oxide. While such sensitization is not used or required in most CBD cases, there have
been many reports of it resulting in more adherent and/or homogeneous layers and faster deposition.
In general, metals make good substrates, either because chalcogenides tend to adsorb strongly
on many metals, or because the non-noble metals are covered with an oxide layer hydroxylated in the
deposition solution. A large variety of CBD films have been deposited on different polymer surfaces.
While deposition is sometimes satisfactory on a clean polymer, various activation treatments, such as
treatment with permanganate, have been used to improve the adhesion and homogeneity. Films have
also been deposited on monolayers. Hydrophilic and specific end groups on the monolayer molecules
can bind species from the solution through chemical or electrostatic interactions.
CBD requires supersaturated solutions, meaning that the product of the concentrations of metal
and chalcogenide must exceed the solubility product of the desired solid. In a closed system, growth
cannot continue once the reactants have been depleted below this point. Thus, the maximum obtain-
able thickness is limited by the reactant supply. Increasing the degree of supersaturation, for example,
by raising the pH in the case of oxides, increases the rate of film growth. Under these rapid conditions
for solid formation, however, some of the solid forms into a powder or relatively large particles that are
only weakly adsorbed into the film. Therefore, the film’s thickness is less than would be expected based
on maximum utilization of the reactants. Solutions that are less heavily supersaturated have a slower
growth rate but ultimately provide thicker films.
9.3. SELECTIVE ION-LAYER ADSORPTION AND

REACTION (SILAR) OR SUCCESSIVE IONIC-LAYER
DEPOSITION (SILD)
Selective ion-layer adsorption and reaction (SILAR) and successive ionic-layer deposition (SILD) are
the same method with different names. The distinguishing characteristic of the SILAR or SILD method
is the use of alternating aqueous solutions (a metal salt solution, followed by a hydrolyzing or sulfidiz-
ing solution). The SILD (SILAR) method consists primarily of repeated successive treatments of the
substrate surface with a solution of salts, such as acetates, chlorides, and nitrates of various metals
(Tolstoy et al. 1986, 1997, 2002, 2003; Tolstoi 1993; Korotcenkov et al. 2006). A significant limitation
in the SILD (SILAR) deposition of oxide thin films seems to be that, upon repeated treatment of the
growing film, which is necessary for synthesis of the next layer, the film is at least partially redissolved
in the metal salt solution. Therefore, Tolstoi (1993, 1997, 1999) suggested the use of a metal salt in a
lower oxidation state and to oxidize the metal ions during film formation.
Tolstoy et al. (Tolstoy et al. 1986, 1997, 2002, 2003; Tolstoi 1993; Korotcenkov et al. 2006) pro-
posed the following six different methods for synthesizing hydrated metal–oxygen compounds that are
difficult to dissolve:
1. Reaction of adsorbed metal cations with H2O2 (OH−), used to synthesize CuO1+x, ZnO1+x, and
La(OH)x(OOH)y.
2. Oxidation of the adsorbed metal layer using H2O2 solution, resulting in the subsequent formation
of a layer of metal peroxide [e.g., from the Ce(OH)x(OOH)4−x layers], metal oxide (SnO2·nH2O,
Tl2O3·nH2O, PbO2·nH2O), or metal hydroxide (FeOOH).
3. Reaction of adsorbed metal cations (Laaq3+) with peroxianion (NbO83−), resulting in formation
of the hard-to-dissolve compound LaxNbOy·nH2O.
4. Oxidation of adsorbed metal cations (Ceaq3+) by peroxianion (NbO83−), resulting in an inter-
action and formation of CexNbOy·nH2O.
5. Reduction of adsorbed anions by cations—for example, reduction of MnO4 by Mn2+ leads to

formation of an MnO2·nH2O layer, and reduction of MoO42− by a SnCl2 solution results in the
formation of SnxMoOy·nH2O.
6. Reduction of adsorbed anions of heteropoly oxometalates—for example, H7PW12O42 reduced by
a SnCl2 solution transforms into a hybride isopoly (Sn-O)-heteropoly (H-P-W-O-) compound.
This method is based on the adsorption of cations, and then anions, from solutions, whose interac-
tion on the substrate surface produces poorly soluble compounds, such as hydroxides, peroxides, and
oxide hydrates. The samples are then washed with distilled water to remove excess salt, treated with
an H2O2 (OH−) solution, and washed again. One such round of treatment comprises one deposition
cycle, and the duration of each treatment depends on the time needed for each stage of sorption and
removal of excess reagents. Depending on the materials, this time can vary from 1 to 10 min. The main
advantage of this chemical route stems from the fact that the cation metal adsorption during the sec-
ond step of the deposition cycle stops automatically once monolayer coverage is attained. Thus, a type
of self-assembly is observed. In theory, this allows ion-by-ion growth of the compound film through
sequential addition of individual atomic layers.
Ellipsometric measurements have shown that, after each cycle, a layer of peroxicomplexes or metal
hydroxide with a thickness of 0.4–1.5 nm can be deposited on the surface of the substrate. The layer
thickness depends on the type of metal deposited. The dependence of the film’s total thickness on the
number of deposition cycles follows a linear law. Thus, this method allows for the film thickness to be
controlled with high precision via the number of ionic deposition cycles. With relatively weak heating
(100–400°C), these compounds release water and transform into the corresponding oxides.
To realize the SILD method, it is convenient to use a special automatic plant, as shown in Figure
4.27 (Korotcenkov et al. 2006). This figure outlines the scheme of one plant designed for the synthesis

Figure 4.27. Block schematic of automated plant for synthesis of metal oxide and noble metal layers
by SILD method at the surface of monolithic substrates: 1, sample; 2, chemical cups with reagents;
3, electromechanical drive; 4, personal computer. (Reprinted with permission from Korotcenkov et al.
2006. Copyright 2006 Institute of Physics Publishing.)
of metal oxide and noble metal layers on the surface of monolithic substrates. It contains chemical
cups (2) with different reagents, an electromechanical drive (3) with a cassette for the samples (1), and a
control unit (4) on the base of a personal computer.
This synthesis can be carried out under “mild” conditions, at room temperature, and in a simple
apparatus. Moreover, simple, inexpensive precursors and solutions with a pH level near neutral can be
used for film deposition. The SILD method is a clean and innovative process with low emissions, which
is readily available and has low maintenance costs. Diluted solutions, including cation and anion re-
agents and stripping water after SILD synthesis, can be easily utilized through simple mixing and fi ltra-
tion. Because this method does not require electrical current, the particles used for surface modification
can be suspended in the solution. As a result, this procedure allows for the deposition of nanolayers, or
nanoclusters of noble metals and metal oxides, on the developed surface—for instance, on the surface
of porous materials and disperse grains (powders).
The SILD method does not have a high deposition rate in comparison with other chemical methods
of deposition, such as chemical or electrochemical deposition (Niesen and Guire 2001, 2002). However,
this is not required for surface modification, particularly by nanoclusters and nanolayers (Korotcenkov
et al. 2003, 2004a). Precise monitoring of both the size and the composition of the deposited clusters
is more important for surface modification. Thus, this method has great advantages over many other
methods used for surface modification of metal oxides. Therefore, the SILD-based technique should
have the potential for new nanoscale products and innovative materials with nano-modified surface
structures. Some of these potential applications include catalytic reactions, new surface coatings, ad-
vanced drug design, and smart materials.
9.4. LIQUID-PHASE DEPOSITION (LPD)

The term liquid-phase deposition refers to the formation of oxide thin films from an aqueous solution of
a metal–fluoro complex [MFn]m−n which is slowly hydrolyzed by adding water, boric acid (H3BO3), or
aluminum metal. While the addition of water directly forces the precipitation of the oxide, boric acid
and aluminum act as fluoride scavengers, destabilizing the fluorocomplex and forcing oxide precipita-
tion. Compared to chemical bath deposition of oxide films, this method allows for superior control of
the hydrolysis reaction and of the solution’s supersaturation. Additionally, film formation is performed
with strongly acidic solutions, in contrast to the usually basic or slightly acidic CBD baths. The differ-
ences mentioned above justify the use of a new term, despite the similarities to the CBD technique.
Simplified, the LPD technique is based on the overall equilibrium reaction
m
H(n − m) MFn + H2O ↔ MOm/2 + nHF (4.13)
2
where m is the charge of the metal cation. According to this reaction, the equilibrium will shift toward
the oxide if the concentration of water is increased or if the hydrogen fluoride concentration is de-
creased through the addition of fluorine scavengers, such as boric acid or aluminium metal.
In this simplified description, homogeneous nucleation and colloid formation occur when the
oxide’s solubility limit is reached, and thin films are formed through the attraction and attachment of
these colloidal particles. This method has been used previously for coating glass with titania, tin oxide,
zirconia, or a variety of 3d transition-metal oxides (V, Cr, Mn, Fe, Co, Ni, Cu, Zn, In, either individually
or combined) (Nagayama et al. 1988; Niesen and Guire 2001).
9.5. ELECTROLESS DEPOSITION (ED)
In its original form, electroless plating has been used to produce metal films on substrates that are
themselves not only metallic but also catalytic for the deposition process (e.g., Fe, Pd, Ni, Co, Au, Ag,
or Rh). Here, deposition is achieved by introducing a reducing agent in the bath. The process is auto-
catalytic and proceeds on the newly formed catalytically active surface. Details of the deposition of
these metals are discussed by Rao and Trivedi (2005). The parameters used for the deposition of the
main noble metals (Pt and Pd) are shown in Table 4.7 (data from Eriksson et al. 2004; Rao and Trivedi
2005; etc.).
The parameters presented in Table 4.7 have been confirmed experimentally. For example, the
composition of the solution used for Re deposition was chosen for several reasons (Petrovich et al.,
2002). First, the NH4+ ions prevent the presence in the solution of alkali-metal ions, which would be
an impediment for silicon technology. Second, the SO4− ions provide higher film uniformity during se-
lective deposition. Additionally, the F− ions improve the process compatibility with silicon technology,
particularly be ensuring the absence of an oxide layer on a silicon surface. Finally, owing to its buffer
properties, a hydrofluoric acid helps to maintain a constant pH level.
Essentially, the solutions for electroless deposition contain hypophosphite, borohydride, alkylam-
ine boranes, or hydrazine as reducing agent and the source of the metal to be deposited. Metals such
as copper that are not catalytic, but that are also not poisonous to catalysis, can be “activated” for sub-
sequent electroless plating by immersing them in an aqueous solution of PdCl2 and HCl. If adequate
electrical conductivity is exhibited by the coating itself, then the substrate can be an electrical insulator.
Such substrates can be made catalytic by nucleation in a colloidal palladium suspension, or by a two-
step process of sequential immersion in an aqueous SnCl2/HCl “sensitizer” and a PdCl2/HCl activa-
tor. Once deposition is initiated, the metal film must itself catalyze any further growth (“autocatalytic
electroless deposition”).
ED resembles CBD in most other respects, particularly in the use of specific oxidizing or hydrolyz-
ing agents to drive the formation of the desired solid phase. Particle sizes of 20–200 nm appear to be
typical for this technique, and the films are mostly nonoriented.
Over the last decade, it has been shown that the electroless deposition technique can also be used
for the deposition of metal oxide films (Niesen and Guire 2001). Metal oxides such as MnO2, TiO2,
In2O3, Tl2O3, PbO2, ZnO, and LaMnO3 have all been deposited using this method, with the thickness
of the films ranging between 20 and 10,000 nm and the growth rate being between 10 and 50,000 nm
h−1. In reference to this process, Niesen and Guire (2001) identify the ED of ceramics as a process that,
demonstrably or apparently, (1) produces films without the use of a counter electrode or connections
to an external electrical power source; (2) involves a change in the oxidation state of the metal cation,
dissolved in an aqueous solution, to an insoluble state, which is the neutral metallic state in electroless
plating; (3) requires participation of the deposition surface (the substrate and, later, the film) in the
Table 4.7. Data on electroless plating systems for some noble metals
CONC. OF METAL MAIN REDUCING
SALT OR COMPLEX ELECTROLYTES IN AGENT
METAL SOURCE (g l−1) (g l−1) MEDIUM (g l−1 ) T ( °C )
PdCl2 2.7 (i) NH4OH, 390 ml B Formaldehyde, 1.5 ml 64
(ii) Na2EDTA, 20
2.0 (i) NH4OH, 150 ml B Sodium hypophosphite 60
1.7 (i) Ethylenediamine, 4.8 A 60
Pd(NH3)4(NO3)2 2.5 NH4OH, 200 ml B Sodium phosphite 64
Na2EDTA, 20 Hydrazine, 0.8 mol
Pt(NH3)2(NO2)2 2.0 Acetic acid, pH 3 A 50–80
NH4ReO4 25 NH4F (30 g l−1); HF (pH ~1) A Hydrazine hydrate, 3 20
RhCl3 4H2O (0.01 M) Ethylenediamine (0.8 M ) (NH4)2SO4 (10 g l−1) 90
Dimethyl-glyoxime (0.025 M),
sodium hydroxide (1.5 M)
HReO4 Na2SO4 + H2SO4; pH 2 A
H[Ir(N2H5Cl)Cl4] pH ~2 A 60–90
A, acidic; B, basic.
Source: Adapted with permission from Eriksson et al. 2004. Copyright 2004 Elsevier.
“redox” process, usually as a path for electron transfer from the site of an oxidation to the site of a
reduction; and (4) requires catalysis to initiate and sustain the process.
9.6. ELECTROCHEMICAL DEPOSITION (ECD)
In sensor technology, the standard method of electrochemical deposition described in detail by Raub
and Muller (1967) is used primarily for the deposition of noble metals, such as Pt and Pd, which act
as contacts and catalysts in chemical sensors. When two electrodes are immersed in a bath containing
the metallic ions to be deposited, cations are attracted to the cathode (negative pole). Because of the
presence of electrons, these cations can be reduced to metallic atoms if the potential is greater than the
discharge potential of the ion, and can coat the surface of the cathode:
Mn+ + ne− → M (4.14)
If the anode is not of the same metal, the reverse reaction takes place, leading to the creation of metallic
ions. According to Faraday’s law, the deposition rate is proportional to the current density.
Because these metals are costly, electrodeposition of a noble metal is undertaken only under special
circumstances. Typical parameters used for the deposition of these metals can be found in Table 4.8,
and the details of this process are discussed by Rao and Trivedi (2004).
Electrochemical deposition has several important advantages over alternative techniques of vapor-
phase physical or chemical deposition. These advantages are related mostly to the higher rate of deposi-
tion, which allows the growth of thicker layers, easier control of the thickness and composition of the
layer, and lower cost of the experimental setup (Morante et al. 2002). Another advantage is the ability
to selectively deposit metal layers on the window regions of a mask layout, allowing the postprocess-
ing of CMOS-processed wafers. The method does, however, have three important disadvantages that
limit possibilities for application: (1) the substrate for deposition must be conductive; (2) there must
be electrical contact with the sample; and (b) there must be contact between the active surface and the
solution.
Electrochemical deposition under oxidizing conditions (EDOC) is a new method for the electro-
chemical preparation of metal oxide nanoparticles (Dierstein et al. 2001). The EDOC method is similar
to the electrochemical synthesis procedure, which also uses an organic electrolyte, but strictly excludes
oxygen, moisture, and the electrochemical formation of oxide layers. The EDOC procedure involves
three steps:
1. Anodic oxidation of a metal sacrificial electrode,
Metalbulk → Men+ + ne (4.15)
2. Cathodic reduction of metal ions and coating with the stabilizer,
Me n+ + ne + stabilizer → metalnano/stabilizer (4.16)

Table 4.8. Data on electroplating systems for noble metals
CONC. AS METAL MAIN ELECTROLYTE IN CURRENT DENSITY

Pt, Pd SOURCE (g l−1 ) (g l−1 ) MEDIUM (A dm−2 ) T (°C)
H2PtCl6 5–25 HCl, 180–300 ml A 2.5–3.5 45–90
(NH4)2PtCl6 6 Sodium citrate, 100 N 0.5–1.0 80–90
PtCl4 3 Na2HPO4, 100 B 0.3–1.0 70–90
K2Pt(OH)6 10 K2SO4, 40 B 0.3–1.0 70–90
Pt(NH3)2(NO2) 5–10 Ammonia, 50 B 0.3–2.0 90–95
H2Pt(NO2)2(SO4) 5.7 H2SO4, to pH 2 A 2.5 30–70
H3Pt(SO3)2OH H2SO4 A
PdCl2 1–40 Potassium phosphate, 50–100 B 0.1–50 40–70
1–3 Sodium chloride, 10–60 A 0.5–5.0 25–45
H2PdCl4 5 Sodium chloride, 40 A 0.4–1.0 50
Pd(NH3)4Cl2 20 Ammonium sulfate, 20 N–B 0.5–1.5 20–40
5–25 Ammonium chloride, 90 N–B >2.5 30–50
Pd(NH3)2(NO3)2 4 Sodium nitrate, 10–11 N–B 0.1–4.0 40–55
Pd(NH3)4(NO3)2 10 Ammonium nitrate, 90–100 B 0.5–2.0 60–80
Pd(NO3)2 2–15 Sulfuric acid, 98 A 0.5–8.0 20–35
Rh as sulfate concentrate 2 Sulfuric acid, 20 ml A 4 40-45
K2[Ru2O(H2PO4)] 6–10 Phosphoric acid, 2 mol A —
(NH4)2 IrCl6 6–8 Sulfuric acid, 0.8 mol A 0.1 18–25
IrCl3 5–15 Sulfamic acid, 25–50 A 0.1–0.2 50–80
K2OsCl6 10 Potassium bisulfate, 60 A 1–4 70
A, acidic; B, basic; N, neutral.
Source: Adapted with permission from Rao and Trivedi 2005. Copyright 2005 Elsevier.
3. Oxidation by air or pure oxygen,
Metalnano/stabilizer + O2 → Me-oxidenano/stabilizer (4.17)
EDOC is based on the reduction of metal ions originating from a sacrificial anode in a nonaque-
ous medium. The metal clusters that form are immediately coated with stabilizers, which prevent the
particles from agglomeration and realize the nanocrystalline state. Thus, numerous amorphous and
crystalline metal oxides, mixed oxides, and oxide mixtures in the nanometer size range and with a nar-
row particle-size distribution can be prepared.
Deposited small particles are disposed to agglomeration, and research has shown that increasing
the particle concentration leads to higher degrees of agglomeration. It has also been found that the
organic coating on the particles formed during step 2 stabilizes the nanoparticles, even at elevated tem-
peratures. For example, experiments have shown that ZnO nanoparticles remain stable even up to 723
K, whereas the uncoated material (washed with ether/ethanol) exhibits significant crystallite growth as
early as 473 K.
9.7. FERRITE PLATING
Ferrite plating, first described in the early 1980s by M. Abe, has been used extensively as a method
of magnetic oxide deposition (mostly spinel ferrites, MxFe3−xO4, where M = Mn, Fe, Co, Ni, or Zn).
Heated solutions of Fe(II) chloride, with additional divalent chlorides when necessary, are oxidized by
adding NO2 ions or oxygen gas, while the pH level is maintained close to 7 by the addition of a base.
Because the oxidation from Fe(II) to Fe(III) occurs in the bulk solution rather than at the substrate,
it is classified here as a CBD technique rather than ED. A variety of substrates can be used, including
glass, sapphire, alumina, and quartz. The hydroxyl groups present on these surfaces are believed to be
essential for film deposition. This technique is notable for the numerous variations developed from it,
including liquid flow deposition; drip deposition; electrically, ultrasonically, hydrothermally, and photo-
assisted deposition; and deposition of bilayers. Nonferrite films have also been deposited (Niesen and
Guire 2001). Detailed discussions are available in several reviews (Abe 2000).
9.8. LIQUID FLOW DEPOSITION (LFD)
In liquid flow deposition, the reactant solution flows past the substrate at a controlled rate. The reactants
are continuously replenished to avoid depletion, as this may cause growth to slow down or cease under
nonflowing conditions. The rate of flow of the deposition medium past the substrate becomes an ad-
ditional and easily controlled variable for adjusting the deposition process, along with the temperature,
pH level, and solution composition. In several cases, consistently high growth rates have been achieved
with no evidence of slowing over longer periods of times, implying that there is no limitation on the
maximum achievable film thickness. Alternatively, this process can be designed to vary the composition,
growth rate, and microstructure, even within a single film, as reported by Deki et al. (2001).
9.9. ELECTROPHORETIC DEPOSITION (EPD)
Electrophoretic deposition (EPD) is one of the colloidal processes used in ceramic production (Heavens
1990). In EPD, charged powder particles, either dispersed or suspended in a liquid medium, are attract-
ed and deposited onto a conductive substrate of the opposite charge as affected of a DC electric field
(Zhitomirsky 2002; Besra and Liu 2007). This means that the principal driving force for electrophoretic
deposition (EPD) is the charge on the particle and the electrophoretic mobility of the particles in the
solvent under the influence of an applied electric field. The charge on the particles can be controlled
by a variety of charging agents, such as acids, bases, and specifically adsorbed ions or polyelectrolytes,
which can be added to the suspension (Zarbov et al. 2002). These additives act through different mecha-
nisms. The main criteria for selecting a charging agent are the preferred polarity and deposition rate of
the particles.
The main differences between EPD and the previously discussed ECD process are presented in
Table 4.9. The basic distinction is that EPD is based on the suspension of particles in a solvent, whereas
ECD is based on a solution of salts (i.e., ionic species) (Heavens 1990; Zhitomirsky 2002). The funda-
mental aspects of EPD, including the factors influencing the deposition process, kinetic aspects, types
of EPD, driving forces, preparation of the electrophoretic suspension, stability, and the mechanisms
involved in EPD, are discussed in detail by Besra and Liu (2007).
Numerous researchers (Heavens 1990; Sarkar and Nicholson 1996; Sato et al. 2001; Zhitomirsky
2002; Besra and Liu 2007) believe that EPD provides the advantages of short formation time, simple
apparatus, little restriction on the shape of the substrate, and no requirement for binder burnout,
because the green coating contains few or no organics. Compared to other advanced shaping tech-
niques, the EPD process is very versatile, since it can be easily modified for a specific application. For
example, deposition can be performed on flat, cylindrical, or any other shape of substrate with only
minor changes to the electrode design and positioning. One such example is given by Dougami and
Takada (2003), who adapted an electrophoretic deposition method to the design process for SnO2-
based gas sensors.
Table 4.9. Main differences between electrophoretic and electrochemical

deposition techniques
DEPOSITION PROCESS
PROPERTY ELECTROCHEMICAL ELECTROPHORETIC
Medium for deposition Solution Emulsion
Main parameter, controlling the rate of deposition Current Voltage
Moving species Ions Solid particles
Charge transfer on deposition Ion reduction None
Presence of surface reactions Yes None
Required conductance of liquid medium High Low
Preferred liquid Water Organic
Source: Adapted with permission from Besra and Liu 2007. Copyright 2007 Elsevier.
The fabrication of surface coatings performed by Dougami and Takadab (2003) using the EPD
method consisted of several steps. The SnO2-based beads were first inserted into a socket on the elec-
trode. A specific amount of coating material was then uniformly diffused in an ethanol solution using
ultrasonic vibration. This suspension was poured into a container, where the Pt mesh electrode was ar-
ranged uniformly at the bottom and was churned once again by ultrasonic vibration. After churning, the
SnO2-based bead electrode was immersed in the suspension. A specific voltage was applied between the
Pt mesh electrode and another electrode from the SnO2-based beads. After a specific interval, the power
supply was turned off and the SnO2-based bead electrode was removed from the container. Lastly, the
coated beads were dried at room temperatures and sintered at 600°C for 15 min in air using Joule heat-
ing of the bead-type sensor. Using the electrophoretic method, coating layers such as TiO2, Al2O3, ZnO,
In2O3, CeO2, and Al2O3–Pd were formed at the surface of the SnO2-based bead.
As shown by Zhitomirsky (2002), because of the high rate of deposition, EPD is an important tool
for the preparation of thick ceramic films, while the ELD method is more applicable for the formation
of nanostructured thin ceramic films. Electrophoretic deposition also offers important advantages for
the deposition of complex compounds and ceramic laminates. The degree of stoichiometry in the elec-
trophoretic deposit is controlled by the degree of stoichiometry in the powder used. This means that
the technological details of the process for producing the powders used in EPD play a very important
role in the formation of films with required parameters.
It is necessary to note, however, that despite the clear advantages the EPD process, there are also
several disadvantages. For example, EPD requires a judicious choice of solvent medium to develop an
appreciable magnitude of surface charge on the suspended powder surface. This is necessary to ensure
the stability of the suspension as well as to facilitate high electrophoretic mobility. Another disadvantage
is that EPD cannot use water as the liquid medium, because application of a voltage to water causes the
formation of hydrogen and oxygen gases at the electrodes, which might adversely affect the quality of
the deposits formed. The difficulties of film deposition on nonconducting substrates and the deposi-
tion of ultrathin films are other important disadvantages of EPD. In addition, the mechanisms of EPD
are still completely not clear.
9.10. PHOTOCHEMICAL DEPOSITION (PCD). APPLYING EXTERNAL

FORCES OR FIELDS
An additional degree of control can be exerted over these synthesis routes if one of the reaction steps is
designed to be triggered photochemically with an external light source (Ichimura et al. 1999; Matsumoto
2000; Sol and Tilley 2001). By controlling the solution conditions and carefully aiming the illumination,
the formation of the solid can be confined to regions on or near the substrate. Photochemical deposi-
tion also offers the possibility of optically patterning (or, by rastering a laser, ‘‘writing’’) the film, with
or without the use of a photoresist.
Ultrasonic agitation can be used to suppress the agglomeration of the colloidal particles that form,
allowing the creation of compact, smooth thin films from solutions that would otherwise not be con-
sidered optimal. Magnetic and electric fields have been applied during CBD of CdS films, leading to
changes in their grain size, crystallinity, and orientation.
9.11. SUMMARY
Just as in the other methods that we have discussed, there are many variations in the methods of deposi-
tion from aqueous solutions. More specific details of the parameters of the compounds obtained from
these methods are shown in Table 4.10 (Niesen and De Guire 2002).
10. THE SOL-GEL PROCESS
10.1. INTRODUCTION
One process that has been intensively studied in the field of chemical sensors is the sol-gel reaction. It
has been established that, through this process, homogeneous inorganic oxide materials with desirable
properties of hardness, optical transparency, chemical durability, tailored porosity, and thermal resis-
tance can be produced at room temperatures. This is directly contrary to the much higher temperatures
required in the production of conventional inorganic materials (Hench and Ulrich 1984; Brinker et al.
1984, 1991; Brinker and Scherer 1990; Warren et al. 1990; Pope et al. 1995; Komarneni et al. 1998).
The specific uses of these sol-gel–produced glasses and ceramics are derived from the various material
shapes generated in the gel state, specifically, monoliths, films, fibers, and monosized powders. Materials
prepared using sol-gel technology have found application in the design of various chemical sensors,
including pH sensors, sensors for ionic species, gas sensors, glucose sensors, oxygen sensors, solid
matrixes for proteins and enzymes, and selective electrodes (Lin and Brown 1997; Ivanovskaya 2000;
Ivanovskaya and Bogdanov 2001).
Table 4.10. Summary of techniques for deposition of oxide thin films from aqueous
media at low temperatures
TYPICAL MAXIMUM
DEPOSITION GROWTH RATE FILM THICKNESS
TECHNIQUE T (°C) DEPOSITION pH (nm h−1) (nm)
Chemical bath deposition 20–100 1–14 2–5,000 50–900
(CBD)
Successive ionic-layer 20–100 4–9 0.2–5 nm 10–1,000
deposition (SILD) per cycle
Liquid-phase deposition (LPD) 15–80 Acidic (1–3) 2–400 100–1,000
Electroless deposition (ED) 20–60 3–10 10–5,000 500–10,000
Electrolytic deposition (ELD) 20–80 3–10 1–103 ~10,000
Ferrite plating (FP) 20–100 5–11 30–20,000 350–5,000
Liquid flow deposition (LFD) 50– 90 2–8 50–2,400 800–1,000
Electrophoretic deposition 20–40 2–11 102–106 >100,000
(EPD)
Source: Adapted with permission from Niesen and De Guire 2002. Copyright 2002 Elsevier.
Figure 4.28. Typical stages of sol-gel route.
10.2. PRINCIPLES OF THE SOL-GEL PROCESS

A sol is a dispersion of solid particles (~0.01–1 μm) in a liquid in which only Brownian motion suspends
the particles. In a gel, liquid and solid are dispersed within each other, presenting a solid network con-
taining liquid components.
The starting materials used in the preparation of the sol are usually inorganic metal salts or metal-
organic compounds such as metal alkoxides. In a typical sol-gel process, the precursor is subjected to a
series of hydrolysis and polymerization reactions to form a colloidal suspension called a sol (see Figure
4.28). Further heat treatment leads to the transformation of gel to xerogels, powders, or ceramic materi-
als in various forms. Thin films can be produced on a piece of substrate by spin-coating or dip-coating.
When the sol is cast into a mold, a wet gel forms. With further drying and heat treatment, the gel is con-
verted into dense ceramic or glass particles. If the liquid in a wet gel is removed under a supercritical con-
dition, a highly porous and extremely low-density material called an aerogel is obtained. As the viscosity
of the sol is adjusted into a proper viscosity range, ceramic fibers can be drawn from the sol. Ultrafine
and uniform ceramic powders are formed by precipitation, spray pyrolysis, or emulsion techniques.
Details of the sol-gel process are discussed more extensively in several excellent review articles (Livage et
al. 1988; Brinker and Scherer 1990; Minh and Takahashi 1995; Brinker et al. 1996; Bagwell and Messing
1996; Livage 1997; Narendar and Messing 1997; Troczynski and Yang 2001; Olding et al. 2001).
10.3. SOL-GEL CHEMISTRY AND TECHNOLOGY

The details of the chemistry involved in the sol-gel process have been reviewed by Brinker and Scherer
(1990). At the functional-group level, three reactions are generally used to describe the sol-gel process:
(I) hydrolysis, (II) water condensation, and (III) alcohol condensation:
I. M-O-R + H2O → M-OH + R-OH (hydrolysis) (4.18)
II. M-OH + HO-M → M-O-M + H2O (water condensation) (4.19)
III. M-O-R + HO-M → M-O-M + R-OH (alcohol condensation) (4.20)

(M = Si, Zr, Ti, Sn, In, Na, Ba, Cu, Ge, W, Y, etc.)
Synthesis using sol-gel processing techniques is generally based on the hydrolytic polycondensation
of metal alkoxides, which corresponds to the nucleophilic attack of H2O on the metal, producing metal
hydroxide groups. In the following steps, the metal hydroxide groups react with either the alkoxide
(heterocondensation) or the hydroxide groups (homocondensation), giving rise to the M–O–M bridges
(Zeigler and Fearon 1990; Brinker and Scherer 1990).
During sol-gel formation, the viscosity of the solution gradually increases, so the gel may exhibit
spontaneous shrinkage, called syneresis, or aging. Depending on the sol-gel processing conditions, such
as the Me/H2O ratio, the type and concentration of the catalyst, alkoxide precursors, and temperature,
gel formation can take place on a time scale ranging from a number of seconds to a number of months.
The network is initially supple, allowing for further condensations, and bond formation induces con-
struction of the network and expulsion of liquid from the pores. Gels which have been dried by evapo-
ration under normal conditions are called xerogels, and they are interesting materials because of their
porosity and high surface area, making them useful as catalyst substrates, sensing materials, and filters.
They are also useful in the preparation of dense ceramics by sintering (densification driven by interfacial
energy) (Brinker et al. 1984).
The characteristics and properties of a sol-gel inorganic network are related to a number of factors
that affect the rate of hydrolysis and condensation reactions: the pH level, the temperature and time of
the reaction, the reagent concentrations, the nature and concentration of the catalyst, the H2O/M mo-
lar ratio (R), the aging temperature, and the drying time. Among these factors, the pH level, the nature
and concentration of the catalyst, the H2O/M molar ratio (R), and the temperature have been identified
as being the most important (Brinker and Scherer 1990). However, it has been shown that, regardless
of the pH level, hydrolysis occurs through nucleophilic attack of the oxygen contained in water on the
M atom, as proven by the reaction of isotopically labeled water with tetraethyl orthosilicate (TEOS),
which produces only unlabeled alcohol in both acid- and base-catalyzed systems (Bradley et al. 1978;
Livage et al. 1988). Although hydrolysis can occur without the addition of an external catalyst, it is more
rapid and complete when a catalyst is employed. Mineral acids (HCl) and ammonia are used most often;
however, other catalysts may be used as well, including acetic acid, KOH, amines, KF, and HF. Under
basic conditions, the hydrolysis reaction is first-order in the base concentration. With weaker bases,
such as ammonium hydroxide and pyridine, measurable speeds of reaction are produced only if large
concentrations are present. Compared to acidic conditions, base hydrolysis kinetics is more strongly
affected by the nature of the solvent. From Eq. (4.19), an increased value of R (H2O/M molar ratio) is
expected to promote the hydrolysis reaction.
As with hydrolysis, condensation is possible without a catalyst; however, it has been shown that
these reactions are acid- and base-specific. In addition, it has been established that under more ba-
sic conditions, gel times increase. It has also been shown that the typical sequence of the condensa-
tion products is monomer, dimer, linear trimer, cyclic trimer, cyclic tetramer, and higher-order rings.
According to Iler (Brinker and Scherer 1990), sol-gel polymerization occurs in three stages. First, the
monomers are polymerized to form particles, followed by the growth of particles, and finally the par-
ticles are linked into chains. Networks then extend throughout the liquid medium, thickening into a
gel. The rate of the ring-opening polymerization and monomer addition reactions is dependent on the
environmental pH level.
It is necessary to note that, in addition to metal-organic precursors, inorganic precursors [salts
such as In(NO3)3 ] may also be used in the sol-gel process. Taurino et al. (2003) believe that the choice
between these two classes should be based on a number of factors. First, handling of inorganic precur-
sors does not generally require any special equipment such as a glove-box. Their chemistry is also more
extensively known and more easily manipulated than that of metal-organic precursors, and the spin-
coating of their solutions does not require any special conditions such as low moisture. Additionally, the
solutions have long-term stability against the increase of viscosity. On the other hand, films prepared
from solutions based on metal-organic precursors are much more uniform than those using inorganic
precursors, and they can be easily deposited onto various substrates, including silicon. Metal-organic
precursors are also free from possible contaminants such as chlorine or sulfur.
10.4. ADVANTAGES AND DISADVANTAGES OF

SOL-GEL TECHNIQUES
Based on the information presented in the above-mentioned reviews, sol-gel technology provides the
following advantages:
• Sol-gel technology has attracted increasing attention for chemical sensor applications because
of its low cost, easier fabrication of films or coatings of complex oxides, and easy control
of composition and microstructure of the deposited coating compared to conventional thin-
film technology.
• Sol-gel methods are generally recognized as efficient routes in fabricating ceramic thin films,
with excellent adhesion between the metallic substrate and the top coating. At the same time,
sol-gel is suitable for preparing thick porous ceramics necessary for various sensor applications,
because the reaction proceeds with the coexisting solvent phase in which evaporation leaves
numerous cavities.
• This method can easily shape materials into complex geometries in a gel state.
• The sol-gel process can produce high-purity products because the organometallic precursor of
the desired ceramic oxides can be mixed, dissolved in a specified solvent, and hydrolyzed into a
sol, and subsequently a gel, and the composition is highly controllable.
• Sol-gel technology has low-temperature sintering capability, usually 200–600°C.
• The fundamental property of the sol-gel process is that it allows for the generation of ceramic
material at a temperature close to room temperature. Therefore, such a procedure opens the
possibility of incorporating soft dopants, such as fluorescent dye molecules and organic cro-
mophores, in these glasses.
• This method provides the possibility of modifying the chemical composition of the films with
multiple different dopants. Reagents can be readily incorporated in a stable host matrix by simply
adding them to the sol prior to its gelation. Alternatively, reagents can be added by copolymer-
izing. In some cases, the matrix actually stabilizes the entrapped reagent from photodegradation
or caustic solution environments. This last feature is particularly interesting in the field of gas
sensors, where the addition of suitable catalysts to the sensing layer may improve the response
of the sensor in terms of both response time and the selectivity of the sensor.
Despite these advantages, the sol-gel technique has never reached its full industrial potential due to
several limitations, including weak bonding, low wear resistance, high permeability, and difficulty in con-
trolling porosity. In particular, because the crack-free property is an essential requirement, the maximum
coating thickness is limited to 0.5 μm (Olding et al. 2001). The consequences of the cracking effect in
In2O3 ceramics are shown in the SEM images presented in Figure 4.29.
Atkinson and Guppy (1991) observed that the crack spacing increased with film thickness and at-
tributed this behavior to a mechanism in which partial delamination accompanies crack propagation.
According to several studies (Schmidt et al. 1988; Takahashi et al. 1990; Garino 1990; Pope et al. 1995),
there are several strategies to avoid the cracking effects, including increasing the fracture toughness of
the film, reducing the modulus of the film, reducing the volume fraction of the solvent at the solidifica-
tion point, and reducing the film thickness.
Plasticizers are often added to organic polymer films to reduce the stiffness of the film and thus
avoid cracking. For sol-gel systems, analogous results are obtained by organic modification of the alkox-
ide precursors, chelation by multidentate ligands such as β-diketinates, or a reduction in the extent of
the hydrolysis of the alkoxide precursors. It should be noted that for particulate films, the maximum
film thickness that can be obtained without cracks decreases linearly with a reduction in particle size.
The organics trapped by the thick coating often cause failure during the thermal process. The present
sol-gel technique is very substrate-dependent, and the thermal expansion mismatch prevents its wide
application. However, this method is one of most effective methods for preparing sensing materials for
chemical sensors. For these applications, the presence of the cracking effect is not a factor which limits
the sensing properties.
b)
(a) (b) (c)

Figure 4.29. Cracking effects in In2O3 ceramics prepared by sol-gel method: (a) In2O3:Bi; (b)
In2O3:Mn; (c) In2O3:Cu. (Reprinted with permission from Korotcenkov et al. 2004e. Copyright 2004
Elsevier.)
10.5. CALCINATION OF SOL-GEL–OBTAINED OXIDES

Research has shown that stabilized colloidal metal hydroxides are synthesized during the sol-gel pro-
cess. This means that the deposited materials are very far from thermodynamic equilibrium. All of
these products are very unstable at temperatures over 150°C and do not have a crystalline structure.
Furthermore, depending on the formation procedure, the material also presents traces of other chemi-
cal species, such as chlorides or traces of organic precursors. Therefore, thermal treatment (calcination)
is required to transfer these hydroxides in their stoichiometric oxide form. The modes of thermal treat-
ment commonly used in the calcination of some oxide phases are presented in Table 4.11 (Tahar et al.,
1997; Chung et al. 1998; Sangaletti et al. 1999; Li et al. 1999; Nayral et al. 2000; Ivanovskaya 2000; Llobet
et al. 2000; Yue and Gao 2000; Leiti et al. 2001; Zakrzewska 2001; Kaya et al. 2002).
For the majority of oxides, thermal annealing changes the properties of the oxides in a consistent
manner. These changes generally are as follows:
• The concentration of chlorine and traces of other precursors decreases as the annealing tem-
perature increases. This concentration falls below the detection limit of x-ray photoelectron
spectroscopy (XPS), which is ~0.1%, at temperatures above 600–700°C.
• As a rule, the crystalline phase determined by x-ray diffraction (XRD) appears at temperatures
above 400°C. When the annealing temperature is increased, the crystallinity of the samples also
increases. This process is accompanied by grain-size growth, in which strong increases have been
shown experimentally at temperatures above 450°C (Zhang and Liu 2000; Cirera et al. 2000;
Leite et al. 2001, Cabot et al. 2001). In treatments at about 1000°C, the grains size may be as large
Table 4.11. Some metal oxides prepared by the sol-gel method and their parameters
PRECURSOR, TEMPERATURE TEMPERATURE AND TIME

METAL OXIDE OF HYDROLYSIS OF CALCINATION GRAIN SIZE
SnO2 [Sn(NMe2)2]2, 135°C 600°C, 6 h 10–20 nm
SnO2 SnCl4, 150°C 700°C, 1 h ~10 nm
In2O3 InCl3, 110°C 600°C, 1 h 20–27 nm
In2O3 In(NO3)3, 60°C 500–800°C, 1 h 18–50 nm
In2O3 In(OAc)3, 3 days 400–700°C, 30 min 23–37 nm
WO3 W(OC2H5)6, RT, 1 h 400–700°C
WO3 (NH4)10W12O41 5H2O 300–1000°C, 3–5 h
TiO2 Ti(OC2H5)4, RT, 1 h 350–500°C, 4 h 10–15 nm
TiO2 TiCl4 750°C, 2 h Whiskers
Ta2O5 TaCl5, 24 h 600–700°C, 4 h ~50 nm
Fe2O3 Fe(OC2H4OCH3)3, 90°C 400°C, 2 h ~5–30 nm
ZrO2 Zr(acac), 220°C 600°C, 4 h ~20–60 nm
TiO2-WO3 Ti(OC3H7)4 + H2WO4 700–900°C
TiO2-SnO2 Ti(OC3H7)4 + SnCl2 600–700°C 10–50 nm
BaSnO3 Ba(OH)2 + K2SnO3 3H2O 1000°C 200 nm
as 100–150 nm (see Figure 4.30). Taking into account the straight correlation between the sur-
face activity of a material and the grain size (Leite et al. 2001), this metal oxide behavior requires
a careful choice in the method by which materials are stabilized using sol-gel technology.
• At low calcination temperatures, nanoparticles are characterized by a quasi-spherical size, strong
disorder, and water content, which progressively disappear with increasing calcination tempera-
ture. When the transition temperature is above 450°C, changes occur very quickly, and practically
no water is found, meaning that the nanoparticles are of good crystalline quality and are faceted
(Dieguez et al., 1999b, 2000).
• As a rule, the majority of the particles are in a highly agglomerated state (Kaya et al. 2002; Nayral
et al. 2000; Dieguez et al. 2000).
• Nanoparticles prepared using low calcination temperature are characterized by a strong distor-
tion in the crystalline structure. Mechanical strains observed in metal oxide powders are taken
off during the annealing process at Т > 400–600°С (Dieguez et al. 1999b; Cirera et al. 2000).
• The introduction of different additives for sol stabilization can change the particulars of oxide-
phase formation (Ivanovskaya 2000; Cabot et al. 2001) because the additives are incorporated
into the sol micelle, and further, into the crystalline-phase structures. Additives introduced to
change the electrophysical and catalytic properties of oxides (Sangaletti et al. 1999; Cabot et al.
2001) show the same influence.
The fundamentals of the processes taking place in nanoceramics during sintering are given by
Lu (2008).
SnO2
Figure 4.30. Influence of calcination temperature on size of SnO2 powders prepared by sol-gel
method.
SiX3
X3Si X3Si SiX3

A B X3Si
C SiX3
X=OR, Cl,
Figure 4.31. Types of Si(X)3 groups that can be used for preparing organic–inorganic hybrid solids.
10.6. ORGANIC–INORGANIC HYBRID MATERIALS (OIHM)
Currently, there is growing interest in organic–inorganic hybrid solids. Hybrid materials can be de-
scribed as nanocomposites (Cerveau and Corriu 1998; Yano et al. 1998; Collinson 2002; Viswanathan
et al. 2006; Maya et al. 2007; Schubert et al. 2007) in which the mean size of the organic and inorganic
phases can be of the order of a few nanometers. This type of material unites the advantages of both
organic polymers and ceramics, meaning that organic–inorganic hybrid solids have unique characteris-
tics and can optimize a series of important chemical sensor parameters (Livage 1997). Several reviews
consider the design and applications of these novel organic–inorganic hybrid materials, particularly
from a materials point of view (Rappoport and Apeloig 1998; Collinson 2002).
Previous research has shown how that this unique phase of solids can be prepared from a molecu-
lar precursor using hydrolytic polycondensation, which is also used in the sol-gel preparation of oxides
from alkoxide precursors (Cerveau and Corriu 1998; Yano et al. 1998). Thermally unstable organic
polymers can be incorporated with inorganic materials because the sol-gel process proceeds under mild
conditions at low temperature and atmospheric pressure.
The simplest way to prepare organic–inorganic hybrid solids is the elaboration of an oxide matrix
starting with a molecular precursor containing stable metal–carbon bonds. For these purposes, Si(X)3
groups can be used as shown in Figure 4.31. These bonds must be stable enough to survive hydrolytic
sol-gel conditions, oxidation in the air, and the drying temperature. The number of metals that allow con-
servation of the covalent metal–carbon bonds is limited. For instance, the elements Si, P, Ge, As, and Sn
are all possibilities, whereas transition metals and Group III elements cannot be used. Silicon, however,
is the most convenient element, because its chemistry is well known, which allows methods to introduce
the Si(X)3 groups into most organic molecules; the identification of its compounds is well established;
and it has the ability to form a wide range of oxide matrices, permitting easy access to materials.
It has been established that MOIH solids are metastable solids of which their formation is kineti-
cally controlled. All parameters with the ability to modify the kinetics of the hydrolysis polycondensa-
tion process are of importance for both the physical and chemical properties of the solids. The organic
moieties are also a primary factor in both the short-range organization and the texture of the solid.
Even weak interactions are able to control the formation of both the solid and the substructure be-
tween the organic units. These materials are also a convenient way to introduce new physical properties
to the oxide matrix. Further details regarding the forming of these phases and their properties are given
by Livage (1997), Cerveau and Corriu (1998), and Yano et al. (1998).
10.7. SUMMARY
By controlling the parameters of hydrolysis, polymerization, and condensation, it is possible to vary over
wide ranges the structure and properties of the sol-gel–derived inorganic network or organic–inorganic
network. At the same time, such strong dependence of the produced materials on the parameters of the
sol-gel process shows the necessity to follow the synthesis and annealing conditions strictly to obtain
reproducible powders, ceramics, and hybrid materials with the required properties. These conditions
provide optimal sensor parameters and are determined empirically for every composition.
11. POWDER TECHNOLOGY
11.1. INTRODUCTION
Nanoscaled powders are the main materials used to prepare ceramics for a variety of applications in
chemical sensing. As shown previously, the various synthesis techniques used for producing nanoscaled
powders can generally classed into four major types: (1) assembly of clusters/nanoparticles produced
by chemical routes or gas-phase synthesis; (2) crystallization of amorphous structures under appropri-
ate conditions; (3) electrolytic deposition; and (4) introduction of a critical concentration of defects into
crystalline powders by mechanical milling.
The challenge in synthesizing nanostructure powders using wet chemical routes is to control and
engineer the physical properties (size, shape, composition, etc.) of the starting powders, since these
affect the properties of the final products. Many advanced processing techniques, such as sol-gel syn-
thesis, spray pyrolysis, and precipitation, have been used to produce fine powders (Hench and Ulrich
1984; Lavernia and Wu 1996; Niesen and De Guire 2001; Vahlas et al. 2006). From an industrial point
of view, however, cost-effective and less complicated synthesis techniques are required for large-scale
applications. Therefore, the gas-phase synthesis methods have been developed quite extensively over
the last few years, including large-scale industrial applications, with considerable progress being made in
the commercialization of new products. In all of these techniques, the primary products are nanosized
powders, which are subsequently transformed by various consolidation techniques to materials in either
bulk form or coatings, with or without porosity.
Some of the most widely used techniques are summarized schematically in Figure 4.32. The pow-
der synthesis processes shown are (1) gas condensation (GPC), (2) chemical vapor condensation (CVC),
(3) microwave plasma (MPP), and (4) low-pressure combustion flame synthesis (CFS). These processes
can all be interchanged and used in similar chamber designs.
11.2. GAS PROCESSING CONDENSATION (GPC)
In gas processing condensation, a metallic or inorganic material, such as a suboxide, is vaporized using
thermal evaporation in an atmosphere of inert gas (at 1–50 mbar) (Hahn 1997). Clusters form in the
vicinity of the source by homogeneous nucleation in the gas phase and grow through coalescence and
O2, N2
Figure 4.32. Techniques for gas-phase synthesis of nanocrystalline powders. (Adapted with permis-
sion from Hahn 1997. Copyright 1997 Elsevier.)
incorporation of atoms from the gas phase. The cluster or particle size depends critically on the resi-
dence time of the particles in the growth regime and can be influenced by the gas pressure, the type of
inert gas, and the evaporation rate/vapor pressure of the evaporating material. The average size of the
nanoparticles can be increased by increasing the gas pressure, vapor pressure, and mass of the inert gas.
Log-normal size distributions have been found experimentally and explained theoretically by the growth
mechanisms of the particles. Even in more complex processes, such as low-pressure combustion flame
synthesis, in which a number of chemical reactions are involved, the size distributions are log-normal.
This method, however, can only be used in a system designed for gas flow (GFC gas flow conden-
sation). In such a system a dynamic vacuum is generated by continuous pumping and the inlet of gas
takes place via a mass flow controller. One major advantage over convectional gas flow is the improved
control of the particle sizes. The particle-size distribution in gas flow systems shift toward smaller aver-
age values, with an appreciable reduction in the standard deviation of the distribution.
Nanocrystalline oxide powders are formed through controlled postoxidation of the primary nano-
particles of a pure metal (e.g., Ti to TiO2) or a suboxide (e.g., ZrO to ZrO2). In some cases, direct
evaporation of oxides, namely, SnO2, Al2O3, CeO2, has resulted in average particle sizes as small as
3 nm with essentially no agglomeration. The velocity of the postoxidation process, influenced by the
heat of the reaction, determines the degree of agglomeration of the primary powder particles and is
essential for the evolution of the microstructure during compaction and sintering. Since the reaction is
performed in a gas flow, there is no agglomeration during the reaction of the isolated particles, a clear
advantage of this method.
11.3. CHEMICAL VAPOR CONDENSATION (CVC)
The CVC process, shown schematically in Figure 4.32, was originally designed to adjust the parameter
field during synthesis in order to suppress film formation and enhance homogeneous nucleation of the
particles in a gas flow. In a certain range of residence time, both particle and film formation can be ob-
tained. The residence time of the precursor molecules can be adjusted by changing the gas flow rate, the
pressure difference between the precursor delivery system and the main chamber, and the temperature
of the hot-wall reactor. These changes result in mass production of nanosized particles of metals and
ceramics instead of thin films, as in CVD processing (Tschope et al. 1997). The mechanisms involved
in the process of particle formation for two sources, thermal evaporation and electrospraying, are pre-
sented schematically in Figure 4.33 (Nakaso et al. 2003).
In the simplest form of the CVC process, a metal-organic precursor is introduced into the hot
zone of the reactor using a mass flow controller. In spite of the increased number of factors control-
Figure 4.33. Mechanisms in the process of particle formation. (Reprinted from Vahlas et al. 2006.
ling the CVC process in comparison to CFC, a wider range of ceramics, including nitrides and carbides,
can be synthesized.
In addition to the formation of single-phase nanoparticles, the CVC reactor allows the synthesis
of mixtures of (1) nanoparticles of two phases or doped nanoparticles by supplying two precursors at
the front end of the reactor, and (2) coated nanoparticles, for example, n-ZrO2 powders coated with
n-Al2O3 or vice versa, by supplying a second precursor in a second stage of the reactor. In the latter
case, nanoparticles formed by homogeneous nucleation are coated by heterogeneous nucleation in a
second stage of the reactor.
Because the CVC processing is continuous, production capabilities are much greater than with
GPC processing. The microstructure of nanoparticles and the properties of materials obtained by CVC
have been identical to the GPC-prepared powders for all oxides studied.
11.4. MICROWAVE PLASMA PROCESSING (MPP)
Microwave plasma processing is similar to the previously discussed CVC method, but employs plasma
instead of high temperature for the decomposition of the metal-organic precursors. In comparison to
thermal activation, the major advantage of plasma-assisted pyrolysis is the low-temperature reaction,
which reduces the tendency for agglomeration of the primary particles.
11.5. COMBUSTION FLAME SYNTHESIS (CFS)
The CFS process involves the combustion of liquid or gaseous precursors injected/delivered into dif-
fused or premixed flames, where the liquid precursor either decomposes or vaporizes and then under-
goes a chemical reaction and/or combustion in the flame. The flame source and the combustion process
provide the required thermal environment for vaporization, decomposition, and chemical reaction.
The fuel for the CFS process can be hydrogen or a hydrocarbon (C2H2), mixed with O2. The use
of a hydrocarbon often leads to the formation of soot, whereas the combustion of hydrogen is a faster
process and produces no condensed species. The flame temperature is usually high, typically 1727–
2727°C, which often causes a homogeneous gas-phase reaction to occur, leading to the deposition of
powders. Therefore, CFS is widely used in industrial applications to produce large quantities of powders
from metal chloride precursors in hydrocarbon flames (Hampikian and Carter 1997; Carter 1999).
The main process parameters that can be optimized to control the crystal structure and particle
size are the flame temperature and its distribution, the choice of precursor and its residence time in the
flame, and the ratio of precursor to fuel (Bunshah 1994; Vincenzini 1995; Carter 1999). Additives may
also be introduced into the flame to alter the size, phase, and shape of the powders.
A typical CFS system is shown in Figure 4.34. It consists of a water-cooled vacuum chamber con-
tinuously pumped with a roughing pump. A constant dynamic pressure is maintained with a closed-loop
pressure controller. Inside the chamber, a burner and a substrate are rigidly fixed to a mount to allow the
distance between the burner and substrate to be varied. The fuel flows upstream along with vapors of
the precursor carried by a carrier gas, such as helium or another inert gas. As the configuration in Figure

Figure 4.34. Schematic of burner–substrate assembly for combustion flame chemical vapor con-
densation. (Reprinted with permission from Nakaso et al. 2003. Copyright 2003 Elsevier.)
4.34 indicates, the flow is highly one-dimensional, and the temperature and chemical species concentra-
tions vary only in the axial direction (i.e., perpendicular to the burner face). The precursor pyrolyzes in
the preheat zone of the flame, and particles condense as the flame gases cool upon approaching the
substrate. The high degree of uniformity in these flames ensures that the particle-size distribution in
the deposit is narrow, since all particles experience a similar time/temperature history. The chamber
may be sited on an optical table which houses a tunable dye laser system, to probe the flame for the gas
temperature and radical species concentrations.
The CFS method presents the following advantages:
• The high flame temperatures allow the use of volatile and less volatile chemical precursors to
form a chemical vapour.
• The formation of the reaction product is a single-step process with no postprocessing treat-
ments, such as calcination.
• The mixing of reactants on a molecular scale takes place quickly, thus significantly reducing
the processing time and enabling better control of the stoichiometry of the multicompo-
nent powders.
• The vaporization, decomposition, and chemical reactions occur rapidly, leading to a high deposi-
tion rate.
• CFS technology is a relatively low-cost compared to conventional methods and can be per-
formed in an open atmosphere without the need for a sophisticated reactor or vacuum system.
The main drawback to the CFS method is the large temperature fluctuation of the flame source dur-
ing deposition, due to the large temperature gradient in the flame. Efforts have been made to minimize
the instability of the flame temperature by developing specially designed burners, such as the counter-
flow flame burner and the reduced-pressure flat flame burner, to produce a very flat and uniform flame
in the horizontal plane (Skandan et al. 1999; Choy 2003). Such burners allow for better control over
the microstructure, the particle size, and its distribution, leading to the fabrication of nanocrystalline
powders that are difficult to produce using the conventional CFS method (Glumac et al. 1999).
At the same time, however, the high flame temperature is another problem of conventional CFS.
The high gas pressure during the reaction produces highly agglomerated powders, which is disadvan-
tageous for subsequent processing. Therefore, the low-pressure combustion flame synthesis method
(LPCFS) has been developed over the past several years. The basic concept of LPCFS is to extend the
pressure range to the pressures used in gas-phase synthesis and thus to reduce or avoid agglomeration.
Just such an approach to CFS was shown in Figure 4.34.
11.6. NANOPOWDER COLLECTION
A range of collection devices has been used to separate nanoparticles from the gas (Choy 2003; Vahlas
et al. 2006). Traditionally, a rotating cylindrical device cooled with liquid nitrogen has been employed for
particle collection. The nanoparticles are subsequently removed from the surface of the cylinder with a
scraper, usually in the form of a metallic plate.
However, the simplest method of collecting nanopowders is by to use a mechanical filter with
a small pore size. The disadvantage of this method is that parts of the filter material are often
incorporated into the nanocrystalline powder, resulting in impurities and flaws in the bulk nano-
crystalline material.
Nanoparticles ranging from 2 to 50 nm in size may be extracted from the gas flow by thermopho-
retic forces from an applied permanent temperature gradient, and then deposited loosely on the surface
of the collection device as a powder of low density with no agglomeration.
In addition, a variety of devices developed for aerosol science and pollution control can be used
to collect nanoparticles effectively. Among these techniques, corona discharge and electrostatic devices
(Karthikeyan et al. 1997) are the most appropriate for highly effective large-scale separation and con-
tinuous processing. In all of these techniques, the nanopowders can be taken out of the system and
further processed by mixing, pressing, and sintering to obtain bulk nanocrystalline materials with vari-
able microstructures.
11.7. MECHANICAL MILLING OF POWDERS
For some applications, especially solid-state gas sensor design, it is necessary to use nanoscaled pow-
ders with sizes smaller than 5–10 nm (Nayral et al. 2000; Leite et al. 2001). In many cases, however, the
deposited powders are agglomerated and are large in size. Mechanical milling is an effective method of
resolving this problem (Hadjipanayis and Siegel 1994; Koch 1997).
Mechanical attrition, or ball milling, which induces heavy cyclic deformation in powders, is a tech-
nique that produces nanostructures not by cluster assembly, but by the structural decomposition of
coarser-grained structures as the result of severe plastic deformation. This has become a popular meth-
od for producing nanocrystalline materials because of its simplicity, the relatively inexpensive equip-
ment (on the laboratory scale) needed, and the applicability to essentially all classes of materials. The
most commonly cited advantage is the possibility for easily scaling up material to tonnage quantities for
various applications.
On the other hand, the most commonly noted problems are contamination from milling media
and/or atmosphere and the need in many applications to consolidate the powder product without
coarsening the nanocrystalline microstructure. In fact, the contamination problem is often presented as
a reason to dismiss this method, at least for some materials. If steel balls and containers are used, iron
contamination becomes a problem. This is most serious for the highly energetic mills and depends on
the mechanical behavior of the powder being milled as well as its chemical affinity for the milling media.
Lower-energy mills result in much less, often negligible, iron contamination. Other milling media, such
as tungsten carbide or ceramics, may be used, but contamination is also possible from such media (Koch
1997). Surfactants (process control agents) may also be used to minimize contamination.
At longer milling times, the grain size decreases steadily. The minimum grain size obtainable by
milling has been attributed to a balance between the defect/dislocation structure introduced by the
plastic deformation of milling and its recovery through thermal processes. For metal powders, the mini-
mum grain size induced by milling scales inversely with the melting temperature of the group of metals
studied (see Figure 4.35), as shown in the data presented by Hadjipanayis and Siegel (1994).
Fecht (Hadjipanayis and Siegel 1994) summarizes the phenomenology of the development of a
nanocrystalline microstructure by mechanical milling into the following three stages:
Stage 1: Deformation is localized in shear bands containing a high dislocation density.
Figure 4.35. Minimum grain size for elements during milling process versus melting temperatures.
(Adapted with permission from Koch 1997. Copyright 1997 Elsevier.)
Stage 2: Dislocation annihilation/recombination/rearrangement forms a cell/subgrain structure

with nanoscale dimensions. Further milling extends this structure throughout the sample.
Stage 3: The orientation of the grains becomes random.
It has been established that the mill energy is not critical to the final microstructure, but the kinetics
of the process are, of course, dependent on the energy. The amount of time needed to attain the same
microstructure can be several orders of magnitude longer in low-energy mills than in high-energy mills.
It has also been assumed that the total strain introduced by milling, rather than the milling energy or ball–-
powder–ball collision frequency, is responsible for determining the nanocrystalline grain size (Koch 1997).
Thus, as a result of mechanical milling, additional mechanical tensions appear in the material. The
appearance of these mechanical tensions and the change of microstructure can be an important ad-
ditional factor that influences gas-sensing characteristics. Mechanical tensions should therefore be con-
sidered during the development of procedures for sensor fabrication. These tensions are only relieved
through thermal treatment at T ≈ 600–800°C (Cirera et al. 2000). Thus, an additional thermal treatment
after mechanical milling may be required.
11.8. SUMMARY
Figure 4.36 summarizes the capabilities of the various gas-phase synthesis methods that have been
described. It is evident that all materials can be prepared by means of gas-phase synthesis in a nano-
Figure 4.36. Typical nanocrystalline materials synthesized using various powder technology methods.
crystalline microstructure. For each case, it is necessary to determine which technique is most ap-
propriate in terms of cleanliness of the powder surfaces, degree of agglomeration, particle size and
distribution, phases, and quantities. In addition to single-phase materials, some of the techniques are
also capable of synthesizing metal/metal, metal/ceramic, and ceramic/ceramic composites, as well as
coated nanoparticles, potentially leading to interesting applications in the near future.
12. POLYMER TECHNOLOGY
12.1. INTRODUCTION
There are two main options for incorporating polymers into chemical sensors (Gardner and Bartlett
1995; Kumar and Sharma 1998; Harsanyi 1995, 2000):
1. Preprocessed polymer films are synthesized and shaped by extrusion, stretched into sheet forms,
and covered by metal film separately from the sensor structures. They can then be attached (typi-
cally by gluing) to inorganic sensor surfaces.
2. Polymerization occurs directly on the sensor surfaces. The synthesis and shaping process occur
on the sensor surface.
Of course, the second method is more progressive and is therefore the more commonly used technique
for the fabrication of thin polymer layers used in the majority of chemical sensors.
12.2. METHODS OF POLYMER SYNTHESIS
There is no single method for synthesizing polymers targeted for chemical sensors. Most polymers, ex-
cept ionomeric polymers, can be synthesized using standard methods of polymerization, both conven-
tional and specific routes, including the polycondensation process and metal-catalyzed polymerization
techniques. Thus, polymers may be synthesized by any of the following techniques:
• Chemical or radical polymerization

• Electrochemical polymerization
• Photochemical polymerization
• Metathesis polymerization
• Concentrated emulsion polymerization
• Inclusion polymerization
• Solid-state polymerization
• Plasma polymerization
• Pyrolysis
• Soluble-precursor polymer preparation
These methods are described in detail in various comprehensive reviews (Malkin and Siling 1991;
Skolhein 1986; Kumar and Sharma 1998; Gurunathan et al. 1999; Malinauskas 2001; Reisinger and
Hillmyer 2002). A summary of the reported literature highlighting the polymerization techniques of
some widely used conductive polymers is presented in Table 4.12.
Among all of the above-listed techniques, chemical polymerization is the most used for preparing
large amounts of conductive polymers, since it is performed without electrodes (Malinauskas 2001).
Chemical polymerization (oxidative coupling) is followed by the oxidation of monomers to a cation
radical and their coupling to form dications. The repetition of this process generates a polymer, and
all classes of conjugated polymers can be synthesized using this technique. Chemical polymerization
is conducted with relatively strong chemical oxidants, such as ammonium peroxydisulfate (APS), fer-
ric ions, permanganate or bichromate anions, or hydrogen peroxide. The reaction is controlled by the
concentration and oxidizing power of the oxidant. Chemical polymerization occurs in the bulk of the
solution, and the resulting polymers precipitate as insoluble solids. The polymer formed by chemical
synthesis is generally a black powder.
Another chemical method used in the preparation of polymers is chemical vapor deposition. This
method uses reagents in the gaseous phase, with the polymer being formed on a substrate present in
Table 4.12. Synthesis techniques for some conductive polymers used in chemical
sensor fabrication
POLYMER METHOD
Polyacetylene Chemical polymerization
Polythiophene Chemical polymerization
Electrochemical polymerization
Polyaniline Chemical polymerization
Polyisoprene Inclusion polymerization
Polybutadiere Inclusion polymerization
Polysiloxane Pyrolysis
Poly(2,3-dimethyl-butadiene) Inclusion polymerization
Polypurrole Chemical polymerization
Photochemical polymerization
Poly(p-phenylene-terephthalamide) Electrochemical polymerization
Polypyrrole-polymide composites Electrochemical polymerization
PVC Chemical polymerization
Polystyrene Concentrated emulsion polymerization
Tetraphenylporphin Vacuum polymerization
Poly(p-phenylene) Chemical polymerization
Poly(α-naphthylamine) Electrochemical polymerization
3-Octylthiophene-3-methylthiophene Electrochemical polymerization
Poly(1,4-phenylene) Electrochemical polymerization
the reagent vapor. An alternative is to have only the monomer in gaseous form, with the oxidant being
present in liquid form on the surface of the substrate to be coated.
Electrochemical polymerization is usually carried out on the working electrode in a single-
compartment cell with a three-electrode configuration, including a working electrode (generally Pt,
but may vary according to the final requirements), a reference electrode (saturated calomel electrode),
and a secondary electrode (Pt, Ni, or C). Generally, organic solvents are used in electrochemical po-
lymerization, but aqueous solutions have been employed as well. Electrochemical polymerization
can be carried out potentiometrically using a suitable power supply (potentiogalvanostat). Generally,
potentiostatic conditions are recommended to obtain thin films, while galvanostatic conditions are
recommended to obtain thick films. The advantage of this method is that precise flow control and
rate of film deposition can be maintained by varying the potential/current conditions of the working
electrode in the system.
The voltage potential is the most important parameter in controlling the polymerization process.
For electrochemical oxidation, a certain electropolymerization potential (EP) must be applied to the
solution for the monomer to be oxidized. Table 4.12 gives the peak oxidation potentials for some of
the aromatic compounds that can produce conducting polymers using the electrochemical technique
(Miller 1982). Table 4.13 shows that the electrochemically polymerizable monomers reported to date
have relatively lower anodic oxidation potential peaks, that is, oxidation potentials of less than 2.1 V. No
polymerization occurs below the electropolymerization potential. Above the EP, the rate of polymeriza-
tion increases with the potential, which may be affected by a number of parameters, including monomer
concentration, electrolyte concentration, and the nature of the electrode. For example, the EP for pyr-
role is generally between 0.6 and 0.8 V. During pyrrole polymerization, lower current densities lead to
the formation of a more crystalline polymer with fewer cross-linkages, while higher current densities
produce rougher films.
Table 4.13. Electrochemical data for some heterocyclic and aromatic monomers
used for electrochemical polymerization
MONOMER OXIDATION POTENTIAL (V)

Pyrrole 0.6–0.8
Bipyrrole 0.55
Terpyrrole 0.26
Thiophene 2.07
Biothiophene 1.31
Terthiophene 1.05
Azulene 0.91
Pyrene 1.30
Carbazole 1.82
Fluorene 1.62
Fluoranthene 1.83
Aniline 0.71
H2, gas
Figure 4.37. Configuration of device for polymer film deposition by reactive sputtering.
Using this technique, a variety of conductive polymers has been generated, such as polypyrrole,
polythiophene, polyaniline, polyphenylene oxide pyrrole, and polyaniline/polymeric acid composite.
The degree of polymer doping during electrochemical polymerization depends on the dopant concen-
tration, the amount of charge passed, and the voltage applied.
Photochemical polymerization takes place in the presence of UV irradiation. This technique utiliz-
es photons to initiate a polymerization reaction in the presence of photosensitizers. Pyrrole has recently
been photopolymerized using a ruthenium(II) complex as the photosensitizer. Under photoirradiation,
Ru(II) is oxidized to Ru(III), and the polymerization is initiated by a one-electron-transfer oxidation
process. Polypyrrole (PPY) films can be obtained through the photosensitized polymerization of pyr-
role using a copper complex as the photosensitizer. Photopolymerization of benzo(C)thiophene has
been carried out in accetonitrile using CCl4 and tetrabutylammoniumbromide. Photosensitive polymers
have a unique advantage in processing, as polymerization and shaping can occur simultaneously with
UV illumination. Photosensitive polymers can be applied and patterned with the same technology used
by photoresists.
Plasma polymerization is a technique that is used in the preparation of ultrathin uniform layers
(5–10 nm) that adhere strongly to an appropriate substrate. An electric glow discharge is used to create
low-temperature “cold” plasma. A schematic diagram of the apparatus used for plasma polymerization
is provided in Figure 4.37. The device used for plasma polymerization is commonly constructed with
an anode, a mesh cathode, and a monomer supply ring. The stage is cooled by water because the prob-
ability that the monomer radicals will stick to the wafer decreases at high temperature, and the plasma-
polymerization rate decreases along with the temperature. The monomer liquid is cooled at 0°C to
maintain the vapor pressure. Argon gas is supplied to the chamber through the holes in the anode and
is ionized between the anode and the mesh cathode. The advantage of this technique is that it eliminates
a number of steps needed in the conventional coating process (Favia and De Agostino 1998).
Metathesis polymerization is unique, differing from all other polymerizations in that all the double
bonds in the monomer remain in the polymer. It is a natural outgrowth of Ziegler-Natta polymerization
in that the catalysts used are similar, and often identical. This usually involves a transition-metal com-
pound plus an organometallic alkylating agent. Metathesis polymerization is further divided into three
classes: ring-opening metathesis of cycloolefins (ROMP); metathesis of alkynes, acyclic or cyclic; and
metathesis of diolefins. By far the greatest amount of work has been done on ROMP.
Pyrolysis is probably one of the oldest approaches utilized to synthesize conductive polymers by
heating the polymer to form extended aromatic structures, thus eliminating heteroatoms. The product
of polymer hydrolysis can be a film, a powder, or a fiber, depending on the form and nature of the
standing polymer and the pyrolysis condition.
Nevertheless, conductive polymers have also been synthesized using other techniques, such as
chain polymerization, step-growth polymerization, chemical vapor deposition, solid-state polymeriza-
tion, soluble-precursor polymer preparation, and concentrated emulsion polymerization, to name just a
few. Most of these techniques, however, are time-consuming and involve the use of costly chemicals.
12.3. FABRICATION OF POLYMER FILMS
Thin polymer films can be deposited using a variety of techniques that vary in complexity and ap-
plicability (Gardner and Bartlett 1995; Matsumoto et al. 1998; Pique et al. 2003; Chrisey and Hubler
1994). The choice of deposition technique depends on the physicochemical properties of the material,
the film quality requirements, and the substrate being coated. The simplest methods involve the ap-
plication of a liquid-solution polymer in a volatile solvent, including aerosol, dipping, spin coating, and
Langmuir-Blodgett (LB) dip coating (Harsanyi 1995).
The Langmuir-Blodgett technique is based on the transfer of an insoluble polymer monolayer on
a substrate with a hydrophilic surface as it is raised from the liquid covered by this polymer monolayer
(Gaines 1966; Osada and DeRossi 2000). It is also possible to dip a substrate with a hydrophobic surface
in water covered by a polymer monolayer and then slowly draw it back out. The possibility of deposit-
ing ordered films with known and controlled thickness (in the range ±2.5 nm) is the main advantage
of the Langmuir-Blodgett technique. In principle, the LB technique can be used to prepare mono- and
multimolecular layers and architectures with high perfection, different layer symmetries, and molecular
orientations. This method, however, has very low technological effectiveness, making its application
in real chemical sensor fabrication processes unlikely. The number of polymers which can be used for
preparing polymer films with the LB method is also very limited.
A part of polymers formed by chemical polymerization can deposit spontaneously on the surface
of various materials immersed in the polymerization solution. The distribution of the resulting poly-
mers between the precipitated and deposited forms depends on many variables and varies within a
broad range. To coat materials with a polymer layer, it is desirable to shift this distribution toward the
surface-deposited form, whereas bulk polymerization should be diminished as much as possible. This
can usually be achieved by choosing appropriate reaction conditions, such as the concentration of the
solution components, the concentration ratio of oxidant to monomer, the reaction temperature, and an
appropriate treatment of the surface of the material to be coated by conducting polymers. Although
a bulk polymerization cannot be completely suppressed, a reasonably high yield of surface-deposited
polymers can be achieved by adjusting the reaction conditions (Malinauskas 2001).
Screen printing is a widely used technology in processing polymer composite materials available
in paste form. The dipping methods will be discussed in detail in Volume 2 of this series. The dipping
method and spin-coating technique depend on the solubility of the conductive polymers, previously
synthesized by the chemical polymerization technique. In this case, the surface to be coated is enriched
either with a monomer or an oxidizing agent, and is then treated with a solution of either oxidizer or
monomer, respectively. A major advantage of this process is that the polymerization occurs almost
exclusively at the surface; no bulk polymerization takes place in the solution. For some polymer materi-
als, the surface can be enriched with a monomer by its sorption from the solution. Enrichment of the
surface by an oxidizer can be achieved either by using an ion-exchange mechanism or by deposition of
an insoluble layer of oxidizer. The disadvantage of this process is that it is limited by materials that can
be covered or enriched with a layer of either monomer or oxidizer in a separate stage, preceding the
surface polymerization.
Several techniques are applicable to bulk polymer materials, such as vacuum deposition technolo-
gies. Vacuum deposition processes may be used to obtain thin polymer films that have high density,
thermal stability, and insolubility in organic solvents, acids, and alkalis. These polymer deposition tech-
niques involve in situ polymerization on a substrate surface affected by various factors. The layers can
be deposited on any substrates that cannot be damaged by the vacuum processes. The following are all
examples of vacuum deposition processes (Skolheim 1986; Harsanyi 2000):
• Vacuum pyrolysis consisting of sublimation, a pyrolysis, and a deposition-polymerization process

• Vacuum polymerization stimulated by electron bombardment
• Vacuum polymerization initiated by UV irradiation
• Vacuum evaporation using either a resistance-heated solid polymer source or, more effectively,
an electron beam
• Radio-frequency sputtering of a polymer target in a plasma composed of polymer fragments
with argon added to the plasma
• Plasma or glow discharge polymerization of monomer gases or vapors
Most of the currently available techniques, however, are not generic enough to be capable of
simultaneously depositing polymer thin films without affecting their chemical integrity and physico-
chemical properties while also producing thin, uniform, and solvent-free coatings in a discrete or con-
tinuous fashion. Furthermore, most of these techniques are not appropriate for the fabrication of
multilayers, since they rely on the application of a solvent solution containing the material of interest,
which may dissolve any previously deposited layers. Therefore, intensive research has been carried out
over the last few years attempting to perfect the methods for the deposition of polymer films.
The most successful approach so far has been a modification of the pulsed laser deposition (PLD)
method, described earlier. Previous work with UV PLD showed the ability of this technique to deposit
thin films of various types of polymer materials. However, it also has been established that the PLD of
polymers in a standard variant is limited to a small class of materials.
The recently developed matrix-assisted pulsed laser evaporation (MAPLE) method considerably
extends the possibilities of PLD for the deposition of polymeric materials. MAPLE differs from PLD
in the way the target is prepared and in the laser energy regime under which the laser–material interac-
tions at the target take place (Pique et al. 2003). The MAPLE deposition process has been used suc-
cessfully to deposit various types of polymer and organic materials, including chemoselective polymers.
MAPLE offers significant advantages for the fabrication of chemical sensors, since it allows for the
deposition of solvent-free chemoselective polymers on a variety of substrate surfaces. For example,
using MAPLE, highly uniform films of siloxane fluoroalcohol (SXFA) have been deposited on the
surface of surface acoustic wave (SAW) resonators. The performance of these MAPLE-coated sensors
was comparable to that of spray-coated SAW devices.
It is necessary to note that in the CVD method with laser activation, the method of cooling the sub-
strate during polymer deposition is also effective. This cooling method creates conditions that hamper
the polymer’s degradation and allow for a considerable increase in the deposition rate. For example, by
irradiating a cooled substrate with an excimer laser in an organic gas environment, polymethylmethacry-
late (PMMA) films have been selectively deposited with a high deposition rate (Takashima et al. 1994).
It needs to be highlighted that the appearance of effective dry methods for polymer deposition is
an important achievement, because the standard wet processes do not promote device integration into
modern semiconductor processing (Harsanyi 2000).
13. DEPOSITION ON FIBERS
13.1. SPECIFICS OF FILM DEPOSITION ON FIBERS
In principle, it is possible to use any of the previously mentioned methods for deposition of sensitive
materials on fibers. It is only necessary to take into account the following factors:
• Deposition takes place on a small area and, as a result, the process has very low effectiveness of
use of the deposited material.
• All sides of the fiber need to be covered uniformly. An SEM image of a fiber with a covering
deposited using the PVD method without rotation is shown in Figure 4.38.
Figure 4.38. Schematic diagrams and SEM image for fiber coating by thermal evaporation.
CVD
reactor
Fibre Figure 4.39. Apparatus for CVD on fiber.
• The fiber may be too long and have poor thermal contact with the substrate’s holder.
• The fiber may be polymeric, which may impose a series of limitations on the thermal parameters
of deposition.
• The fiber has lowered mechanical durability, while the covering needs to have high adhesion
because the fiber may bend.
13.2. COATING DESIGN AND TOOLING
Coating methods for fibers do not require specific changes in the deposition process. Therefore, to
deposit sensing material on the fiber, the same modes may be used as for deposition on substrates
with large particle sizes. The details of the deposition on the fiber become apparent only when specific
constructions of the substrate’s holders and heaters are used, which fix the fiber’s position, rotation,
uniform heating, and displacement (Kashima et al. 1991; Kaneko and Nittono 1997).
At present there no fixed rules for elaborating these units. Every designer resolves this problem in-
dividually, taking into account the deposition method being used and available equipment. For example,
the fiber may be pulled through a reactor with hot walls in the process of CVD. Alternatively, the fiber
may be heated to the reaction temperature by an electrical current applied through a mercury electrode.
In one conventional technique, the fiber was heated in a waveguide-type microwave applicator. A hot-
wall CVD fiber-coating system capable of coating a fiber in a continuous manner is depicted in Figure
4.39. This system may operate at atmospheric pressure and has separate furnaces for desizing the fiber
prior to deposition and for thermal treatment of the coating after deposition. A supply spool and mo-
torized take-up spool complete the system.
Of course, all of these approaches have specific advantages and disadvantages. For example, the
disadvantages of the direct heating technique are that it is limited to electrically conductive fibers, each
reaction chamber can accommodate only one fiber at a time, and the mercury used for electrical contact
with the fiber has a highly toxic vapor. The microwave technique is not limited to electrically conductive
fibers and does not involve mercury. However, it is limited to one fiber per applicator, and the wave-
guide applicator is intrinsically energy-inefficient because its proper operation depends on the absorp-
tion of a substantial portion of the incident microwave power in a dummy load. It should be kept in
mind, however, that the effectiveness of the microwave-cavity applicator has improved considerably.
The fibers can be electrically conductive or nonconductive, so there is no need for mercury, and mi-
crowave energy is being utilized more efficiently than in the older waveguide/dummy-load microwave-
applicators. An effective way of improving the uniformity of a fiber’s covering using the PDV method
is through simultaneous use of several deposition sources located around the fiber (Kashima et al.
1991), as shown in Figure 4.40.
Another area that is critical to the successful production of fiber coatings is the tool used to hold
the fibers in the coating chamber. Because most fibers must be coiled and packed into the tool, it is im-
portant that the fiber remain unstressed. This requirement may necessitate tooling in a standard coating
chamber that reduces the distance from the source to the substrate. Another approach is to have tooling
in the form of a drum. Even in such cases, however, the number of fibers that can be coated is limited
by the amount of epoxy in the fiber bundles. Even cured epoxy will outgas in a vacuum chamber, with
most of the gas being water vapor absorbed from ambient air. Such outgassing of the epoxy affects the
adhesion of the coating to the fiber, the packing density of the coating, and the refractive index of the
deposited films.
During deposition of sensing material on polymer fibers using magnetron sputtering, it is nec-
essary to take into account that, depending on the evaporation conditions, PVD-coated fibers can
present quite different surface properties (Dietzel et al. 2000). Deposition of reactive metals induces
both textural and chemical changes on the fiber surface. Nitrogen and carbon react with the polymer
surface and form new structures as a result of an etching process. A higher concentration of reactive
gas leads to a more drastic recombination of the polymer surface and results in a fibroid fiber struc-
ture. Dietzel et al. (2000) established that different noble and reactive gases produce different layer
adhesion. Thus, nitrogen and acetylene plasmas lead to better layer adhesion than argon plasma does.
Higher N2 concentrations produce superior layer adhesion, and pure Zr–based layers adhere better
than Ti-based layers.
Temperature control of the fiber during coating is another parameter that must be monitored.
Research has shown that the refractive indices of a “cold” coated film are usually less than that of a
hot film, even with the use of ion-assisted deposition. In many cases, these refractive indices are also
Fiber
Electrode
Electrode
Fiber
Figure 4.40. Apparatus with multielectrode configuration. (Adapted from Kashima et al. 1991.
Figure 4.41. SEM images of carbon fiber after (a) 2 h, (b) 4 h, and (c) 6 h exposure to a TiO2 LPD
solution. (Reprinted with permission from Kern and Gadow 2004. Copyright 2004 Elsevier.)
nonuniform. If the fibers have epoxy or plastic jackets, as is usually the case, they cannot be exposed
to high temperatures. The temperature controller of the coating system is usually programmed to a
temperature between 35 and 90°C; however, the heat from the electron-beam gun or ion gun will raise
the temperature of the fiber during coating. If the temperature of the fiber exceeds a safe limit, the ion
or electron gun must be shut off to allow the fiber to cool down.
To resolve this problem and establish a successful fiber-coating system, the coating chamber is set
up with either a broad-band, high-current, low-voltage ion gun or a suitable plasma system to densify
the coating. This is known as an ion-assist process and can produce dense films on cold substrates, such
as fibers.
In addition to densifying the coating, an ion gun may also be used to clean the surface prior to coat-
ing, thus improving adhesion of the coating to the fiber tips. For this to be possible, the ion gun must
be capable of handling oxygen gases, because an oxide coating is the most-robust that can be applied
to a fiber. However, in spite of the intensive development of dry deposition methods for various films
on a fiber, the dip-coating method continues to be the most widely used (Kaneko and Nittono 1997).
So far, compared to wet deposition methods, this method is a more commercially attractive alterna-
tive to costly vapor deposition technologies (Kern and Gadow 2002; Herbig and Libmann 2004). The
process offers several unique properties, including the ability to use a wide range of compositions and
the ability to apply coatings to complex substrates, both of which make this process advantageous for
coating fibers with sensing materials. For example, Herbig and Libmann (2004) found that by using a
dip-coating procedure, the sol-gel technique could be used to apply a coating to individual fibers. SEM
measurements confirmed that an effective coating had been applied to the fibers (see Figure 4.41). Kern
and Gadow (2004) further showed that the continuous liquid-phase coating method is technically and
economically feasible for the production of coated carbon fibers used in the manufacture of various
composite materials.
Some fibers can be coated with polymer layers by immersing them into an electrolyte, where elec-
trochemical polymerization takes place, or by vapor-phase treatment of oxidant-containing carriers
with the monomers (Malinauskas 2001). In this latter case, fibrous materials are charged by sorption
with FeCl3, provided from aqueous or ether solutions, and then treated with a pyrrole vapor, either in
vacuum or from its solution in toluene.
14. OUTLOOK
This chapter has demonstrated the vast array of methods which can be used effectively for the form-
ing and deposition of various materials used in chemical sensors. These techniques differ mainly in
deposition rates, substrate temperature during deposition, precursor materials, the necessary equip-
ment, expenditure, and in the quality of the resulting films. Table 4.14, for example, compares two main
deposition methods for films used in the design of chemical sensors.
It is clear that every method has its own advantages and disadvantages, making it necessary to
choose the appropriate method for each specific engineering application, taking into account the prop-
erties of the deposited material and the possible consequences for the sensor’s parameters during the
method’s application. The ideal method in any application should meet the following criteria:
• Compatibility with the process of manufacture of the chemical sensor

• No impairment of, or effect on, the properties of the bulk materials used in the device
• Ability to deposit the required type of material with the required thickness and structure
• Improvement in the quality of the designed sensor
• Ability to coat the engineering components uniformly with respect to both size and shape
• Cost-effectiveness in terms of the cost of the substrate, depositing material, and coating
technique
• Ecologically clean and safe for attending personnel
It is obvious that there is currently no perfect method that meets all these requirements. Will et
al. (2000) considered and analyzed currrent methods and their possibilities for forming thin-film solid
electrolytes. The results of that comparison are presented in Table 4.15.
Basing on this analysis, it can be concluded that only simple stoichiometric compounds can be de-
posited using standard CVD and PVD methods, as each component has to be evaporated at different
temperatures due to their different vapor pressures. Moreover, the CVD process also applies very toxic
precursors. The constituents have to be deposited from independently controlled sources, adding com-
plexity to the system. At the same time, PVD is a line-of-sight process, meaning that it has difficulty in
coating complex-shaped components.
Other methods also have limitations. Magnetron sputtering of multicomponent materials requires
expensive sintered solid targets and produces surface damage. Ceramic powder methods and spray
pyrolysis methods, on the other hand, have the potential to be good candidates for complicated stoi-
chiometric compositions or for mixtures of materials. However, sol-gel coated films tend to crack and
have a thickness limitation for each layer, meaning that the process needs to be repeated to obtain the
required thickness. Spray pyrolysis has limitations in repeatability. The investment cost for CVD and
PVD apparatus is high compared to the droplet and powder techniques, whereas the setup for spray
pyrolysis is inexpensive and simple. Standard CVD methods, however, have the advantage of being
able to coat large areas uniformly, and they feature easily controlled deposition rates and film thick-
nesses. On the other hand, all liquid-precursor methods, such as sol-gel and slurry coating, are time-,
labor-, and energy-intensive, because coating and drying/sintering have to be repeated in order to avoid
crack formation.
Table 4.14. Comparison of chemical vapor deposition and physical vapor deposition coating techniques
CHEMICAL VAPOR DEPOSITION (CVD) PHYSICAL VAPOR DEPOSITION (PVD)

Sophisticated reactor and/or vacuum system. Sophisticated reactor and vacuum system.
Simpler deposition rigs with no vacuum system has been adopted in
variants of CVD, AACVD, ESAVD, FACVD, and CCVD.
Expensive techniques for LPCVD, PECVD, PACVD, MOCVD, EVD, Expensive techniques.
ALE, UHVCVD; relatively low-cost techniques for AACVD
and FACVD.
Non–line-of-sight process. Line-of-sight process.
Therefore, complex-shaped components can be coated, and coatings can Therefore, it is difficult to coat complex-shaped
be deposited with good conformal coverage. components and to produce conformal coverage.
Tends to use volatile/toxic chemical precursors. Tends to use expensive sintered solid targets/sources.
Less volatile/more environmentally friendly precursors have been adopted Potential difficulties in large-area deposition and in varying
in variants of CVD such as AACVD, ESAVD, and CCVD. the composition or stoichiometry of the deposits.
Multisource precursors tend to produce nonstoichiometric films. Both single and multiple targets do not guarantee the
deposition of stoichiometric films because different
Single-source precursors (AACVD, PICVD) have overcome such elements will evaporate or sputter at different rates, except
problems. with the laser ablation method.
High deposition temperatures in conventional CVD. Low to medium deposition temperatures.
Low to medium deposition temperatures can be achieved using variants
of CVD such as PECVD, PACVD, MOCVD, AACVD, ESAVD.
CCVD, combustion chemical vapor deposition; MOCVD, metal-organic–assisted CVD; PECVD, plasma-enhanced CVD; FACVD,
flame-assisted CVD; AACVD, aerosol-assisted CVD; ESAVD, electrostatic-atomization CVD; LPCVD, low-pressure CVD; APCVD,
atmospheric-pressure CVD; PACVD, photo-assisted CVD; TACVD, thermal-activated CVD; EVD, electrochemical vapor deposition;
RTCVD, rapid thermal CVD; UHVCVD, ultrahigh-vacuum CVD; ALE, atomic-layer epitaxy; PICVD, pulsed-injection CVD.
Source: Adapted with permission from Choy 2003. Copyright 2003 Elsevier.
Table 4.15. Comparison of methods for producing thin and dense solid electrolytes
FILM CHARACTERISTICS
DEPOSITION RATE PROCESS FEATURES
TECHNIQUE STRUCTURE OR THICKNESS COST CHARACTERISTICS AND LIMITATIONS
Vapor phase
Thermal spray 100–500 μm h−1 High deposition rates, various compositions, possible thick and porous
technologies coatings, high temperatures necessary
EVD Columnar structures 3–50 μm h−1 Expensive High reaction temperatures necessary, corrosive gases equipment and
processing costs
CVD Columnar structures 1–10 μm h−1 Expensive Various precursor materials possible, high reaction equipment
temperatures necessary, corrosive gases
PVD (RF and Amorphous to 0.25–2.5 μm h−1 Expensive Tailor-made films, dense and crack-free, low films equipment deposition
magnetron) polycrystalline temperatures, multipurpose technique, relatively low deposition rate
sputtering)
Laser ablation Amorphous to Expensive Intermediate deposition temperatures, difficult equipment upscaling, time-
polycrystalline sharing of laser, relatively low deposition rate.
Spray pyrolysis Amorphous to 5–60 μm h−1 Economical Robust technology, upsealing possible, easy control of polycrystalline
polycrystalline parameters, corrosive salts, post-thermal treatment usually necessary
Liquid phase
Sol-gel, liquid Polycrystalline 0.5–1 μm for each Economical Various precursors possible, very thin films, low temperature sintering,
precursor route coating coating and drying/heating have to be repeated 5–10 times, crack
processes formation during drying, many process parameters
Solid phase
Tape casting Polycrystalline 25–200 μm Robust technology, upscaling possible, crack formation
Slip casting and Polycrystalline 25–200 μm Economical Robust technology, crack formation, slow
slurry coating
Tape calendering Polycrystalline 5–200 μm Upscaling possible, co-calendering possible
EPD Polycrystalline 1–200 μm Short formation time, little restriction to shape of substrate, suitable
for mass production, high deposition rates, inhomogeneous thickness
Transfer printing Polycrystalline 5–100 μm Economical Robust technology, rough substrate surfaces possible, adhesion
on smooth substrates difficult
Screen printing Polycrystalline 10–100 μm Economical Robust technology, upscaling possible, crack formation
Source: Adapted with permission from Will et al. 2000. Copyright 2000 Elsevier.
Therefore, in practice it is always necessary to search for the best compromise and choose the
method which meets as many requirements as possible. At the same time, however, these requirements
may change considerably during the transition from the device’s elaboration phase to industrial produc-
tion. Properties that may be preferred during elaboration, such as multifunctionality, an opportunity to
vary the parameters of the technological process and the deposited material, and the speed of reorganiz-
ing the technological process, differ greatly from the properties necessary during industrial fabrication,
such as compatibility with basic technological processes, reproducibility, productiveness, and cost. Such
differences in requirements are inevitable, and it is necessary to take them into account during research
targeted toward the elaboration of devices designed for application in the chemical sensor market.
15. ACKNOWLEDGMENTS
This work was supported by the World Class University (WCU) Program at the Gwangju Institute of
Science and Technology (GIST) through a grant (Project No. R31-20008-000-10026-0) provided by the
Ministry of Education, Science and Technology (MEST) of Korea, for the project titled “Development
of Maritime Environmental Sensor Using Nano and Photonic technology,” funded by the Ministry of
Land, Transport and Maritime Affairs, Korea, and by the Korean Science and Engineering Foundation
(KOSEF) NCRC grant funded by the Korean government (MEST) (No. R15-2008-006-01002-0). G.
Korotcenkov and B. K. Cho are also thankful to the Korean BK21 Program for the support of their
scientific research.
REFERENCES
Abe M. (2000) Ferrite plating: a chemical method preparing oxide magnetic films at 24-100°C, and its application.
Electrochim. Acta 45, 3337–3343.
Aita C.R. (2008) Reactive sputter deposition of metal oxide nanolaminates. J. Phys.: Condens. Matter 20, 264006.
Armour D.G. (1994) Ion beam deposition. Nuclear Instrum. Meth. Phys. Res. B 89, 325–331.
Arthur J.A. (2002) Molecular beam epitaxy. Surf. Sci. 500, 189–217.
Atkinson A. and Guppy R.M. (1991) Mechanical stability of sol-gel films. J. Mater. Sci. 26, 3869–3875.
Bagwell R.B. and Messing G.L. (1996) Critical factors in the production of sol–gel derived porous alumina. Key
Eng. Mater. 115, 45–64.
Bardos L., Dragila R., Loncar G., and Musil J. (1983) Creation of Thin Oxide Films in a Microwave Oxygen Plasma.
Academia, Praha.
Behrisch I.R. (ed.) (1981) Sputtering by Particle Bombardment. Springer-Verlag, Berlin.
Besra L. and Liu M. (2007) A review on fundamentals and applications of electrophoretic deposition (EPD). Prog.
Mater. Sci. 52, 1-61.
Bittencourt C.B., Landers R., Llabet E., Molas G., Correig X., Silva M.A.P., Sueras J.E., and Calderer J. (2002)
Effects of oxygen partial pressure and annealing temperature on the formation of sputtered tungsten oxide
films. J. Electrochem. Soc. 149(3), H81–H86.
Blandenet G., Court M., and Lagarde Y. (1981) Thin layers deposited by the pyrosol process. Thin Solid Films 77,
81–90.
Blocher J.M. (1981) Coating of glass by chemical vapor deposition. Thin Solid Films 77, 51–63.
Bonzi P., Depero L.E., Parmigiani F., Perego C., Sberveglieri G., and Quattroni G. (1994) Formation and structure
of tin-iron oxide thin film CO sensors. J. Mater. Res. 9, 1250–1256.
Bradley D.C., Mehrotra R.C., and Gaur D.P. (1978) Metal Alkoxides. Academic Press, London.
Brinker C.J, Clark D.E., and Ullrich D.R. (eds.) (1984) Better Ceramics Through Chemistry. Elsevier-North Holland,
New York.
Brinker C.J. and Scherer G.W. (1990) Sol-Gel Science: The Physics and Chemistry of Sol-Gel Processing. Academic Press,
San Diego, CA.
Brinker C.R., Hurd A.J., Frye G.C., Schunk P.R., and Ashley C.S. (1991) Sol-gel thin film formation. J. Ceram. Soc.
Japan 99, 862–877.
Brinker C.J., Hurd A.J., Schunk P.R., Ashely C.S, Cairncross R.A., Samuel J., Chen K.S., Scotto C., and Schwartz
R.A. (1996) Sol-gel derived ceramic films—Fundamentals and applications. In: Metallurgical and Ceramic Protective
Coatings, K. Stern (ed.). Chapman & Hall, London, 112–151.
Brinzari V., Korotcenkov G., Schwank J., Lantto V., Saukko S., and Golovanov V. (2002) Morphological rank of
51–58.
Broorse R.S. and Burlitch J.M. (1994) MOCVD route to stable, oxygen-rich, chromium oxide films and their con-
version to epitaxial Cr2O3. Chem. Mater. 6, 1509–1515.
Brown I.G. (1998) Cathodic arc deposition of films. Annu. Rev. Mater. Sci. 28, 243–269.
Bunshah R.F. (ed.) (1982) Deposition Technologies for Films and Coatings. Noyes, Park Ridge, NJ.
Bunshah R.F. (ed.) (1994) Handbook of Deposition Technologies for Films and Coatings, 2nd ed. Noyes, Park Ridge, NJ.
Cabot A., Dieguez A., Romano-Rodriguez A., Morante J., and Barsan N. (2001) Influence of the catalytic intro-
duction procedure on the nano-SnO2 gas sensor performance. Where and how stay the catalytic atoms? Sens.
Carter W.B. (1999) Flame assisted deposition of oxide coatings. In: Intermetallic and Ceramic Coatings, N.B. Dahotre
and T.S. Sudarshan (eds.). Marcel Dekker, New York, 233–266.
Castaneda L., Alonso J.C., Ortiz A., Andrade E., Saniger J.M., and Banuelos J.G. (2002) Spray pyrolysis deposition
and characterization of titanium oxide thin films. Mater. Chem. Phys. 77, 938–944.
Cerveau G. and Corriu R.J.P. (1998) Some recent developments of polysilsesquioxanes chemistry for material sci-
ence. Coord. Chem. Rev. 178–180, 1051–1071.
Cirera A., Cornet A., Morante J.R., Olaizola S.M., Castano E., and Garcia J. (2000) Comparative structural study
between sputtered and liquid pyrolysis nanocrystalline SnO2. Mater. Sci. Eng. B 69–70, 406–410.
Chen C.H., Yuan F.L., and Schoonman J. (1998) Spray pyrolysis routes to electroceramic powders and thin films.
Eur. J. Solid State Inorg. Chem. 35, 189–196.
Chour K.W., Wang G.D., and Xu R. (1994) Vapor deposition of lithium tantalate with volatile double alkoxide
precursors. Mater. Res. Soc. Symp. Proc., 335, 65–71.
Choy K.L. (2000) Vapour processing of nanostructured materials. In: Nelwa H.S. (ed.), Handbook of Nanostructured
Materials and Nanotechnology, vol. 1, pp. 533–577. Academic Press, San Diego, CA.
Choy K.L. (2003) Chemical vapour deposition of coatings. Prog. Mater. Sci. 48(2), 57–170.
Christen H.M. and Eres G. (2008) Recent advances in pulsed-laser deposition of complex oxides. J. Phys.: Condens.
Matter 20, 264005.
Chrisey D. and Hubler G. (eds.) (1994) Pulsed Laser Deposition of Thin Films. Wiley, New York.
Chung W.K., Sakai G., Shimanoe K., Miura N., Lee D.D., and Yamazoe N. (1998) Preparation of indium oxide thin
film by spin-coating method and its gas-sensing properties. Sens. Actuators B 46, 139–145.
Collinson M.M. (2002) Recent trends in analytical applications of organically modified silicate materials. Trends
Anal. Chem. 21, 31–39.
Cuomo J.J., Rossnagel S.M., and Kaufman H.R. (1989) Handbook of Ion Beam Processing Technology. Noyes, Park Ridge, NJ.
Deki S., Iizuka S., Akamatsu K., Mizuhata M., and Kajinami A. (2001) Novel fabrication method for Si1-xTxO2 thin
films with graded composition profiles by liquid phase deposition. J. Mater. Chem. 11, 984–986.
Dieguez A., Romani-Rodriguez A., Morante J., Nelli P., Depero L.E., Sangaletti L., and Sberveglieri G. (1997)
Electron microscopy analysis of the RGTO technique for high sensitivity gas sensor development. Inst. Phys.
Conf. Ser. 157, 593–596.
Dieguez A., Romano-Rodriguez A., Morante J.R., Nelli P., Sangaletti L., and Sberveglieri G. (1999a) Analysis of
the thermal oxidation of tin droplets and its implications on gas sensor stability. J. Electrochem. Soc. 146(9),
3527–3535.
Dieguez A., Romano-Rodriguez A., Morante J.R., Kappler J., Barsan N., and Gopel W. (1999b) Nanoparticle en-
gineering for gas sensor optimisation: improved sol-gel fabricated nanocrystalline SnO2 thick film gas sensor
for NO2 detection by calcination, catalytic metal introduction and grinding treatments. Sens. Actuators B 60,
125–137.
Dieguez A., Romano-Rodriguez A., Alay J.L., Morante J.R., Barsan N., Kappler J., Weimar U., and Gopel W. (2000)
Parameter optimization in SnO2 gas sensors for NO2 detection with low cross-sensitivity to CO: sol-gel prepa-
ration, film preparation. Powder calcination, doping and grinding. Sens. Actuators B 65, 166–168.
Dien E., Laurent J.M., and Smith A. (1999) Comparison of optical and electrical characteristics of SnO2-based thin
films deposited by pyrosol from different tin precursors. J. Eur. Ceram. Soc. 19, 787–789.
Dierstein A., Natter H., Meyer F., Stephan H.O., Kropf Ch., and Hempelmann R. (2001) Electrochemical deposi-
tion under oxidizing conditions (EDOC): a new synthesis for nanocrystalline metal oxides. Scripta Mater. 44,
2209–2212.
Dietzel Y., Przyborowski W., Nocke G., Offermann P., Hollstein F., and Meinhardt J. (2000) Investigation of PVD
arc coatings on polyamide fabrics. Surf. Coat. Technol. 135, 75–81.
Doblec R., El.Khakari M.A., Serventi A.M., Trudeau M., and Saint-Jacques R.G. (2002) Microstructure and physical
properties of nanostructured tin oxide thin films grown by means of pulsed laser deposition. Thin Sold Films
419, 230–236.
Dougami N. and Takada T. (2003) Modification of metal oxide semiconductor gas sensor by electrophoretic depo-
sition. Sens. Actuators B 93, 316–320.
Edelman F., Rothshild A., Komem Y., Mikhelashvili V., Chack A., and Cosandey F. (2000) E-gun sputtered and
reactive ion sputtered TiO2 thin films for gas sensors. Electron. Technol. 33(1/2), 89–107.
Epifani M., Díaz R., Arbiol J., Comini E., Sergent N., Pagnier T., Siciliano P., Faglia G., and Morante J.R. (2006)
Nanocrystalline metal oxides from the injection of metal oxide sols in coordinating solutions: synthesis,
characterization, thermal stabilization, device processing, and gas-sensing properties. Adv. Funct. Mater. 16,
1488–1498.
Eriksson S., Nylen U., Rojas S., and Boutonnet M. (2004) Preparation of catalysts from microemulsions and their
applications in heterogeneous catalysis. Appl. Catal. A: General 265, 207–219.
Faglia G., Nelli P., and Sberveglieri G. (1994) Frequency effect of highly sensitive NO2 sensors based on RGTO
SnO2(Al) thin films. Sens. Actuators B 18, 497–499.
Fang Q., Chetwynd D.G., and Gardner J.W. (2002) Conducting polymer films by UV-photo processing. Sens.
Favia P. and De Agostino R. (1998) Plasma treatments and plasma deposition of polymers for biomedical applica-
tions. Surf. Coat. Technol. 98, 1102–1106.
Gaines G.L. Jr. (1966) Insoluble Monolayers at Liquid-Gas Interfaces. Wiley-Interscience, New York.
Gafian-Calvo A.M., Davila J., and Barrero A. (1997) Current and droplet size in the electrospraying of liquids.
Scaling laws. J. Aerosol Sci. 28, 249–275.
Gardiner R.A., Van Buskirk P.C., and Kirlin P.S. (1994) Liquid delivery of low vapor pressure MOCVD precursors.
Mater. Res. Soc. Symp. Proc. 335, 221–226.
Gardner J.W. and Bartlett P.N. (1995) Application of conducting polymer technology in microsystems. Sens.
Garino T.J. (1990) The cracking of sol-gel thin films during drying. In: Better Ceramics Through Chemistry IV. MRS
vol. 180, 497–502.
Glocker D.A. and Shah I. (eds.) (1995) Handbook of Thin Film Process Technology. Institute of Physics Publishing,
Bristol, UK.
Glumac N.G., Skandan G., Chen Y.J., and Kear B.H. (1999) Particle size control during flat flame synthesis of na-
nophase oxide powders. NanoStruct. Mater. 12, 253–258.
Gong H., Wang Y., Yan Z., and Yang Y. (2002) The effect of deposition conditions on structure properties of radio
frequency reactive sputtered polycrystalline ZnO films. Mater. Sci. Semicond. Proc. 5, 31–34.
Goswami J., Wang C.G., Cao W., and Dey S.K. (2003) MOCVD of platinum films from (CH3)3CH3CpPt and
Pt(acac)2: nanostructure, conformality, and electrical resistivity. Chem. Vap. Deposition 9(4), 213–220.
Greer J.A., Tabat M.D., and Lu C. (1997) Future trends for large-area pulsed laser deposition. Nuclear Instrum. Meth.
Phys. Res. B 121, 357–362.
Gupta R.K., Shridhar N., and Katiyar M. (2002) Structure of ZnO films prepared by oxidation of metallic zinc.
Mater. Sci. Semicond. Proc. 5, 11–15.
Gurunathan K., Vadivel Murugan A., Marimuthu R., Mulik U.P., and Amalnerkar D.P. (1999) Electrochemically
synthesised conducting polymeric materials for applications towards technology in electronics, optoelectronics
and energy storage devices. Mater. Chem. Phys. 61, 173–191.
Hahn H. (1997) Gas phase synthesis of nanocrystalline materials. NanoStruct. Mater. 9, 3–12.
Hampikian J.M. and Carter W.B. (1999) The combustion chemical vapor deposition of high temperature materials.
Mater. Sci. Eng. A 267, 7–18.
Hadjipanayis G.C. and Siegel R.W. (eds.) (1994) Nanophase Materials. Kluwer, Dordrecht, The Netherlands.
Harsanyi G. (1995) Polymeric sensing films: new horizons in sensorics? Sens. Actuators A 46–47, 85–88.
Harsanyi G. (2000) Sensors in Biomedical Applications: Fundamentals, Technology and Applications. Technomic, Basel.
Heavens N. (1990) Chapter 7: Electrophoretic deposition as a processing route for ceramics. In: Binner G.P. (ed.),
Advanced Ceramic Processing and Technology, Vol. 1, 255–283. Noyes, Park Ridge, NJ.
Hecht G., Richter F., and Hahn J. (eds.) (1994) Thin Films. DGM Informationgessellschaft, Oberursel, Germany.
Hellmich W., Braunmuhl Ch.B., Muller G., and Sberveglieri G. (1995) The kinetics of formation of gas-sensitive
RGTO-SnO2 films. Thin Solid Films 263, 231–237.
Hench L.L. and Ulrich D.R. (eds.) (1984) Ultrastructure Processing of Ceramics, Glasses, and Composites. Wiley, New
York.
Herbig B. and Loebmann P. (2004) TiO2 photocatalysts deposited on fiber substrates by liquid phase deposition. J.
Photochem. Photobiol. A: Chem. 163, 359–365.
Hitchman M.L. and Jensen K.F. (1993) CVD: Principles and Applications. Academic Press, San Diego, CA.
Hocking M.G., Vasantasree V., and Sidky P.S. (1989) Metallic and Ceramic Coatings: Production, High Temperature Properties
and Applications. Longman, Essex, UK, and Wiley, New York.
Hodes G. (2003) Substrate effects in chemical solution deposition. In: Proceedings of 203rd Electrochemical Society
Meeting, April 27–May 2, 2003, Paris, France, Abstract 570.
Hopfe V., Jiickel R., and Schisnfeld K. (1996) Laser based coating and modification of carbon fibres: application of
industrial lasers to manufacturing of composite materials. Appl. Surf. Sci. 106, 60–66.
Hubler G.K. (1989) Fundamentals of ion-beam-assisted deposition: technique and film properties. Mater. Sci. Eng.
A ll5, 181–192.
Ichimura M., Goto F., Ono Y., and Arai E. (1999a) Deposition of CdS and ZnS from aqueous solutions by a new
photochemical technique. J. Cryst. Growth 198/199, 308–312.
Ichimura M., Goto F., and Arai E. (1999b) Photochemical deposition of CdS from aqueous solutions. J. Electrochem.
Soc. 146, 1028–1034.
Ivanovskaya M. (2000) Ceramic and film metaloxide sensors obtained by sol-gel method: structural features and
gas-sensitive properties. Electron. Technol. 33(1/2), 108–112.
Ivanovskaya M. and Bogdanov P. (2001) The role of catalytic additives in gas-sensitivity of SnO2-Mo based thin
film sensors. Sens. Actuators B 77, 268–274.
Jaworek A. and Sobczyk A.T. (2008) Electrospraying route to nanotechnology: an overview. J. Electrostatics 66,
197–219.
Kadam L.D. and Patil P.S. (2001) Thickness-dependent properties of sprayed cobalt oxide thin films. Mater. Chem.
Phys. 68, 225–232.
Kaneko T. and Nittono O. (1997) Improved design of inverted magnetrons used for deposition of thin films on
wires. Surf. Coat. Technol. 90, 268–274.
Kamarneni S., Sakka S., Phule P.P., and Laine R.M. (1998) Sol-Gel Synthesis and Processing. Ceramic Transactions vol.
94, The American Ceramic Society.
Karthikeyan J., Berndt C.C., Tikkannen J., Wang J.Y., King A.H., and Herman H. (1997) Preparation of nanophase
materials by thermal spray processing of liquid precursors. NanoStruct. Mater. 9, 137–140.
Kashima T., Matsuda Y., and Fujiyama H. (1991) Development of the quadrupole plasma chemical vapour deposi-
tion method for low temperature, high speed coating on an optical fibre. Mater. Sci. Eng. A 139, 79–84.
Kaya C., He J.Y., Gu X., and Butler E.G. (2002) Nanostructured ceramic powders by hydrothermal synthesis and
their applications. Microporous Mesoporous Mater. 54, 37–49.
Kelly P.J. and Arnell R.D. (2000) Magnetron sputtering: a review of recent developments and applications. Vacuum
56, 159–172.
Kern F. and Gadow R. (2002) Liquid phase coating process for protective ceramic layers on carbon fibers. Surf.
Coat. Technol. 151–152, 418–423.
Kern F. and Gadow R. (2004) Deposition of ceramic layers on carbon fibers by continuous liquid phase coating.
Surf. Coat. Technol. 180–181, 533–537.
Kim J.S., Marzouk H.A., and Reucroft P.J. (1995) Deposition and structural characterization of ZrO2 films by che-
mical vapor deposition. Thin Solid Films 254, 33–38.
Kim S.G., Kim J.Y., and Kim H.J. (2000) Deposition of MgO thin films by modified electrostatic spray pyrolysis
method. Thin Solid Films 376, 110–114.
Kimura Y., Kobayashi T., Hanamoto K., Sasaki M., Kimura S., Nakada T., Nakayama Y., and Kaito C. (2001) Effect
of SR irradiation on crystallization of amorphous tin oxide film. Nuclear Instrum. Meth. Phys. Res. A 467–468,
1221–1224.
Kizling M.B. and Jaras S.G. (1996) A review of the use of plasma techniques in catalyst preparation and catalytic
reactions. Appl. Catal. A: General 147, 1–21.
Koch C.C. (1997) Synthesis of nanostructured materials by mechanical milling: problems and opportunities.
Nanostuct. Mater. 9, 13–22.
Komarneni S., Sakka S., Phule P.P., and Laine R.M. (1998) Sol-Gel Synthesis and Processisng. Ceramic Transactions, Vol.
95, The American Ceramic Society.
Korotcenkov G. (2005) Gas response control through structural and chemical modification of metal oxides: state
Korotchenkov G., Brynzari V., and Dmitriev S. (1999a) SnO2 films for thin film gas sensor design. J. Mater. Sci. Eng.
B 63, 195–204.
Korotchenkov G., Brynzari V., and Dmitriev S. (1999b) Processes development for low cost and low power consu-
ming SnO2 thin film gas sensors (TFGS). Sens. Actuators B 54, 202–209.
Korotcenkov G., DiBattista M., Schwank J., and Brinzari V. (2000) Structural characterization of SnO2 gas sensing
films deposited by spray pyrolysis. J. Mater. Sci. Eng. B 77, 33–39.
Korotcenkov G., Brinzari V., Schwank J., and Cerneavschi A. (2001a) Possibilities of aerosol technology for depo-
sition of SnO2-based films with improved gas sensing characteristics. J. Mater. Sci. Eng. C 19, 73–77.
Korotcenkov G., Brinzari V., DiBattista M., Schwank J., and Vasiliev A. (2001b) Peculiarities of SnO2 thin film
deposition by spray pyrolysis for gas sensor application. Sens. Actuators B 77, 244–252.
Korotcenkov G., Cerneavschi A., Brinzari V., Cornet A., Morante J., Cabot A., and Arbiol J. (2002) Crystallographic
characterization of In2O3 films deposited by spray pyrolysis. Sens. Actuators B 84, 37–42.
Korotcenkov G., Macsanov V., Tolstoy V., Brinzari V., Schwank J., and Faglia G. (2003a) Structural and gas response
characterization of nano-size SnO2 films deposited by SILD method. Sens. Actuators B 96, 602–609.
Korotcenkov G., Macsanov V., Boris Y., Brinzari V., Tolstoy V., Schwank J., and Morante J. (2003b) Using of SILD
technology for surface modification of SnO2 films for gas sensor applications. In: Surface Engineering, Synthesis,
Characterization, and Applications, A. Kumar, W.J. Merg, and Y.T. Cheng (eds.), MRS Proc. 750, Y5.25.1–Y.5.25.6.
Korotcenkov G., Macsanov V., Brinzari V., Tolstoy V., Schwank J., Cornet A., and Morante J. (2004a) Influence of
Cu, Fe, Co, and Mn oxide nanoclusters on sensing behavior of SnO2 films. Thin Solid Films 467, 209–214.
Korotcenkov G., Cerneavschi A., Brinzari V., Vasiliev A., Cornet A., Morante J., Cabot A., and Arbiol J. (2004b)
Korotcenkov G., Cornet A., Rossinyol E., Arbiol J., Brinzari V., and Blinov Y. (2004c) Faceting characterization
310–319.
Korotcenkov G., Brinzari V., Cerneavschi A., Ivanov M., Golovanov V., Cornet A., Morante J., Cabot A., and
Arbiol J. (2004d) The influence of film structure on In2O3 gas response. Thin Solid Films 460, 315–323.
Rodriguez J., and Cirera A. (2004e) Gas sensing characteristics of one-electrode sensors on the base of doped
In2O3 ceramics. Sens. Actuators B 103, 13–22.
Korotcenkov G., Brinzari V., Golovanov V., and Blinov Y. (2004f) Kinetics of gas response to reducing gases of
Korotcenkov G., Tolstoy V., and Schwank J. (2006) Successive ionic layer deposition (SILD) as a new sensor tech-
nology: synthesis and modification of metal oxides. Meas. Sci. Technol. 17, 1861–1869.
Korotcenkov G., Brinzari V., Stetter J.R., Blinov I., and Blaja V. (2007a) The nature of processes controlling the
kinetics of indium oxide-based thin film gas sensor response. Sens Actuators B 128, 51–63.
Korotcenkov G., Ivanov M., Blinov I., and Stetter J.R. (2007b) Kinetics of In2O3-based thin film gas sensor re-
sponse: The role of “redox” and adsorption/desorption processes in gas sensing effects. Thin Solid Films 515,
3987–3996.
Korotcenkov G., Blinov I., Brinzari V., and Stetter J.R. (2007c) Effect of air humidity on gas response of SnO2 thin
film ozone sensors. Sens. Actuators B 122, 519–526.
Korotcenkov G. and Cho B.K. (2009) Thin film SnO2-based gas sensors: Film thickness influence. Sens. Actuators
B 142, 321–330.
Kumar D. and Sharma R.C. (1998) Advances in conductive polymers. Eur. Polym. J. 34(8), 1053–1060
Kwok H.S., Sun X.W., and Kim D.H. (1998) Pulsed laser deposited crystalline ultrathin indium tin oxide films and
their conduction mechanisms. Thin Solid Films 355, 299–302.
Labeau M., Gas’kov A.M., Gautheron B., and Senateu J.P. (1994) Synthesis of Pd-doped SnO2 films on silicon and
interaction with ethanol and CO. Thin Solid Films 248, 5–11.
Lalauze R., Breuil P., and Pijolat C. (1991) Thin films for gas sensors. Sens. Actuators B 3, 175–181.
Lampkin C.M. (1979) Aerodynamics of nozzles used in spray pyrolysis. Prog. Crystal Growth Charact. 1, 405–416.
Lavernia E. and Wu Y. (1996) Spray Atomization and Deposition. Wiley, Chichester, UK.
Lee C.H. and Lin L.Y. (1996) Characteristics of spray pyrolytic ZnO thin films. Appl. Surf. Sci. 92, 163–166.
Lee D.S., Rue G.H., Huh J.S., Choi S.D., and Lee D.D. (2001) Sensing characteristics of epitaxially-grown tin oxide
gas sensors on sapphire substrate. Sens. Actuators B 77, 90–94.
Lefebvre, A. H. (1989) Atomization Spray. Taylor & Francis, London.
Lemire C., Lollman D.B.B., Mohammad A.A., Gillet E., and Aguir K. (2002) Reactive R.F. magnetron sputtering
deposition of WO3 thin films. Sens. Actuators B 84, 43–48.
Leite E.R., Cerri J.A., Longo E., Valera J.A., and Paskocima C.A. (2001) Sintering of ultrafine undoped SnO2 pow-
ders. J. Eur. Ceram. Soc. 21, 669–.675
Leo G., Rella R., Siciliano P., Capone S., Alonso J.C., Pankov V., and Ortiz A. (1999) Sprayed SnO2 thin films for
NO2 sensors. Sens. Actuators B 58, 370–374.
Levy R.A. (ed.) (1989) Microelectronic Materials and Processes. Kluwer, Dordrecht, The Netherlands.
Li G.L. Wang G.H., and Hong J.M. (1999) Morphologies of rutile form TiO2 twins crystals. J. Mater. Sci. Lett. 18,
1243–1246.
Lin J. and Brown C.W. (1997) Sol-gel glass as a matrix for chemical and biochemical sensing. Trends Anal. Chem. 16,
200–211.
Livage J., Henry M., and Sanchez C. (1988) Sol-gel chemistry of transition metal oxides. Prog. Solid State Chem. 18,
259–342.
Livage J. (1997) Sol-gel processes. Solid State Mater. Sci. 2, 132–136.
Llobet E., Molas G., Molinas P., Calderer J., Vilanova X., Brezmes J., Sueiras J.E., and Correig X. (2000) Fabrication
of highly selective tungstem oxide ammonia sensors. J. Electrochem. Soc. 147(3), 776–779.
Lokhande C.D. (1991) Chemical deposition of metal chalcogenide thin films. Mater. Chem. Phys. 27, 1–43.
Lu K. (2008) Sintering of nanoceramics. Int. Mater. Rev. 53, 21–38.
Mahmoud S.A. (2002) Characterization of thorium dioxide thin films prepared by the spray pyrolysis technique.
Solid State Sci. 4, 221–228.
Majoo S., Gland J.L., Wise K.D., and Schwank J.W. (1996) A silicon micromachined conductometric gas sensor with
a maskless Pt sensing film deposited by selected-area CVD. Sens. Actuators 35–36, 312–319.
Malinauskas A. (2001) Chemical deposition of conducting polymers. Polymer 42, 3957–3972.
Malkin A.Y. and Siling M.I. (1991) Scientific principles of present-day and future technologies of synthesis and
processing polycondensation polymers. Rev. Polymer Sci. 33, 2135–2160.
Mane R.S. and Lokhande C.D. (2000) Chemical deposition method for metal chalcogenide thin films. Mater. Chem.
Phys. 65, 1–31.
Matsumoto Y., Yoshida K., and Ishida M. (1998) A novel deposition technique for fluorocarbon films and its appli-
cations for bulk- and surface-micromachined devices. Sens. Actuators A 66, 308–314.
Matsumoto Y. (2000) Electrochemical and photoelectrochemical processing for oxide films. MRS Bull. 25, 47–50.
Milchev A. (2008) Electrocrystallization: nucleation and growth of nano-clusters on solid surfaces. Russ. J.
Electrochem. 44, 619–645.
Miller J.S. (ed.) (1982) Catalysis and Electrocatalysis. Am. Chem. Soc. Symp. Ser. Vol. 192. American Chemical Society,
Washington, DC.
Minh N.Q. and Takahashi T. (1995) Science and Technology of Ceramic Fuel Cells. Elsevier, Amsterdam.
Moisan M. and Pelletier J. (eds.) (1992) Microwave Excited Plasmas. Elsevier, Amsterdam.
Morante J.R., Barbiers D., and Michel B. (eds.) (2002) Proceedings of E-MRS Spring Meeting 2001, June 5–8, 2001,
Strasbourg, France. Sens. Actuators A 99, 1–223.
Morris G.C. and McElnea A.E. (1996) Fluorine doped tin oxide films from spray pyrolysis of stannous fluoride
solutions. Appl. Surf. Sci. 92, 167–170.
Morosanu C.E. (1990) Thin Films by Chemical Vapor Deposition. Elsevier, Amsterdam.
Morton International. (1993) CVD Metalorganics for Vapour Phase Epitaxy, Product Guide and Literature Review. Advanced
Materials, Danvers, MA.
Moya J.S., Lopez-Esteban S., and Pecharroman C. (2007) The challenge of ceramic/metal microcomposites and
nanocomposites. Prog. Mater. Sci. 52, 1017–1090.
Musil J. and Vlcek J. (2001) Magnetron sputtering of hard nanocomposite coatings and their properties. Surf. Coat.
Technol. 142, 557–566.
Musil J. (1996) Deposition of thin films using microwave plasmas: present status and trends. Vacuum 47(2),
145–155.
Myslik V., Vrnata M., Vyslouzil F., Jelinek M., and Novotna M. (1999) Pulsed-laser deposition of metal acetyl ace-
tonates for sensor devices. J. Appl. Polym. Sci. 74, 1614–1622.
Nadel S.J. and Greene P.G. (2001) Strategies for high rate reactive sputtering. Thin Solid Films 392, 174–183.
Nadel S.J., Greene P., Rietzel J., and Strumpfel J. (2003) Equipment, materials and processes: a review of high rate
sputtering technology for glass coating. Thin Solid Films 442, 11–14.
Nagayama H., Honda H., and Kawahara H. (1988) A new process for silica coating. J. Electrochem. Soc. 135,
2013–2016.
Nair P.K., Nair M.T.S., Garcia V.M., Arenas O.L., Pena Y., Castillo A., Ayala I.T., Gomezdaza O., Sanchez A.,
Campos J., Hu H., Suarez R., and Rincon M.E. (1998). Sol. Energy Mater. Sol. Cells 52, 313–344.
Nakaso K., Han B., Ahn K.H., Choi M., and Okuyama K. (2003) Synthesis of non-agglomerated nanoparticles by
an electrospray assisted chemical vapor deposition (ES-CVD) method. J. Aerosol Sci. 34, 869–881.
Nalwa H.S. (ed.) (2000) Handbook of Nanostructured Materials and Nanotechnology, Vol. 1: Synthesis and Processes. Academic
Press, San Diego, CA.
Narendar Y. and Messing G.L. (1997) Mechanisms of phase separation in gel-based synthesis of multicomponent
metal oxides. Catal. Today 35, 247–268
Nayral C., Viala E., Fau P., Senocq F., Jumas J.C. Maisonnat A., and Chaudret B. (2000) Synthesis of tin and tin
oxide nanoparticles of low size dispersity for application in gas sensing. Chem. Eur. J. 6, 4082–4090.
Nenov T.G. and Yordanov S.P. (1996) Ceramic Sensors: Technology and Applications, Technomic, Basel.
Nieminen M., Niinisto L., and Lanhala E. (1996) Growth of gallium oxide thin films from gallium acetylacetonate
by atomic layer epitaxy. J. Mater. Chem. 6, 27.
Niesen T.P. and De Guire M.R. (2001) Review: deposition of ceramic thin films at low temperatures from aqueous
solutions. J. Electroceram. 6, 169–207.
Niesen T.P. and De Guire M.R. (2002) Review: deposition of ceramic thin films at low temperatures from aqueous
solutions. Solid State Ionics 151, 61–68.
Oehrlein G.S. (1997) Surface processes in low pressure plasmas. Surf. Sci. 386, 222–230.
Oktik S. (1988) Low cost, non vacuum techniques for the preparation of thin/thick films for photovoltaic applica-
tions. Prog. Cystal Growth Charact. 17, 171–240.
Olding T., Sayer M., and Barrow D. (2001) Ceramic sol-gel composite coatings for electrical insulation. Thin Solid
Films 398–399, 581–586.
Osada Y. and DeRossi D.E. (eds.) (2000) Polymer Sensors and Actuators. Springer-Verlag, Berlin.
Ozegowski M., Meteva K., Metev S., and Sepold G. (1999) Pulsed laser deposition of multicomponent metal and
oxide films. Appl. Surf. Sci. 138–139, 68–74.
Palatnik L.S., Fuks M.I., and Kosevich V.M. (1972) Formation Mechanism and Substructure of Condensed Films. Nauka,
Moscow (in Russian).
Pantano C., Ganan-Calvo A.M., and Barrero A. (1994) Electrohydrostatic solution for electrospraying in cone-jet
mode. J. Aerosol Sci. 25, 1065–1077.
Patil P.S. (1999) Versatility of chemical spray pyrolysis technique. Mater. Chem. Phys. 59, 185–198.
Payne D.C. and Pravman J.C. (eds.) (1990) Laser Ablation for Materials Synthesis. Proceedings of the Material Research
Society Symposium, Vol. 191. MRS, Pittsburgh, PA.
Peng C.H. and Desu S.B. (1992) Low temperature metalorganic chemical vapor deposition of perovskite Pb(ZrxTi1-x )
O3 thin films. Appl. Phys. Lett. 61(1), 16–18.
Perednis D. and Gauckler L.J. 2005) Thin film deposition using spray pyrolysis. J. Electroceram. 14, 103–111.
Petrovich V., Haurylau M., and Volchek S. (2002) Rhenium deposition on a silicon surface at the room temperature
for application in microsystems. Sens. Actuators A 99, 45–48.
Pierson H.O. (1992) Handbook of Chemical Vapor Deposition. Noyes, Park Ridge, NJ.
Pique A., Auyeung R.C.Y., Stepnowsk J.L., Weir D.W., Arnold C.B., McGill R.A., and Chrisey D.B. (2003) Laser
processing of polymer thin films for chemical sensor applications. Surf. Coat. Technol. 163–164, 293–299.
Pochet L.F., Howard P., and Safaie S. (1997) CVD coatings: from cutting tools to aerospace applications and its
future potential. Surf. Coating Technol. 94–95, 70–75.
Poirier G.E., Cavicchi R.E., and Semancik S. (1993) Ultrathin heteroepitaxial SnO2 films for use in gas sensors.
J. Vacuum. Sci. Technol. A 11(4), 1392–1395.
Pope E.J.A., Sakka S., and Klein L.C. (eds.) (1995) Sol-Gel Science and Technology. Ceramic Transactions, Vol. 55, The
American Ceramic Society.
Posadowski W.M. (1999) Pulsed magnetron sputtering of reactive compounds. Thin Solid Films 343–344, 85–89.
Pouch J.J. and Alerovitz S.A. (eds.) (1992) Plasma Processing Technology. Material Science Forum Trans. Tech. Pub.,
Aedermannorf, Switzerland.
Powell C.F., Oxley J.H., and Blocher J.M. Jr. (eds.) (1967) Vapor Deposition. Wiley, New York.
Prince J.J., Ramamurthy S., Subramanian B., and Sanjeeviraja C. (2002) Spray pyrolysis growth and materials proper-
ties of In2O3 films. J. Crystal Growth 240, 142–151.
Putkonen M., Niinisto J., Kukli K., Sajavaara T., Karppinen M., Yamauchi H., and Niinisto L. (2003) ZrO2 thin
films grown on silicon substrates by atomic layer deposition with Cp2Zr(CH3)2 and water precursors. Chem.
Vapor Deposition 9, 207–211.
Que A.P., Auyeung R.C.Y., Stepnowsk J.L., Weir D.W., Arnold C.B., McGill R.A., and Chrisey, D.B. (2003) Laser
processing of polymer thin films for chemical sensor applications. Surf. Coat. Technol. 163–164, 293–299.
Radecka M., Przewoznik J., and Zakrzewska K. (2001) Microstructure and gas-sensing properties of (Sn,Ti)O2 thin
films deposited by RGTO technique. Thin Solid Films 391, 247–254.
Randhaw H. (1991) Review of plasma-assisted deposition processes. Thin Solid Films 196, 329–349.
Rao C.R.K. and Trivedi D.C. (2005) Chemical and electrochemical depositions of platinum group metals and their
applications. Coord. Chem. Rev. 249, 613–631.
Rappoport Z. and Apelog Y. (eds.) (1998) The Chemistry of Organic Silicon Compounds, Vol. 2. Wiley, New York.
Raub E. and Muller K. (1967) Fundamentals of Electrodeposition. Elsevier, Amsterdam.
Reisinger J.J. and Hillmyer M.A. (2002) Synthesis of fluorinated polymers by chemical modification. Prog. Polym.
Sci. 27, 971–1005.
Saender K.L. (1993) Pulsed laser deposition: Part 1. A review of process characteristics and capabilities. Proc. Adv.
Mater. 3, 1–24.
Safi I. (2000) Recent aspects concerning DC reactive magnetron sputtering of thin films: a review. Surf. Coat. Technol.
127, 203–219.
Sangaletti L., Depero L.E., Allieri B., Pioselli F., Angelucci R., Poggi A., Tagliani T., and Nicoletti S. (1999)
Microstructural development in pure and V-doped SnO2 nanopowders. J. Eur. Ceram. Soc. 19, 2073–2077.
Sarkar P. and Nicholson P.S. (1996) Electrophoretic deposition (EPD): mechanisms, kinetics and application to
ceramics. J. Am. Ceram. Soc. 79, 1987–2002.
Sato N., Kawachi M., Noto K., Yoshimoto N., and Yoshizawa M. (2001) Effect of particle size reduction on crack
formation in electrophoretically deposited YBCO films. Physica C 357–360, 1019–1022.
Savadogo O. (1998) Chemically and electrochemically deposited thin films for solar energy materials. Sol. Energy
Mater. Sol. Cells 52, 361–388.
Sberveglieri G., Gropelli S., Nelli P., and Camanzi A. (1991) A new technique for the preparation of highly sensitive
hydrogen sensors based on SnO2(Bi2O3) thin films. Sens. Actuators B 5, 253–255.
Sberveglieri S., Groplli S., Nelli P., Perego C., Valdre G., and Camanzi A. (1993) Detection of sub-ppm H2S concen-
trations by means of SnO2(Pt) thin films grown by RGTO technique. Sens. Actuators B 15–16, 86–89.
Schieber M., Han S.C., Ariel Y., Chokron S., Tsach T., Maharizi M., Deutscher C., Racah D., Raizman A., and
Rotter S. (1991) Comparison of thin films of YBa2Cu3O7-d deposited by physical (laser ablation) and chemical
(OMCVD) methods for device applications. J. Cryst. Growth 115, 31–42.
Schiller S., Goedicke K., Reschke J., Kirchhoff V., Schncider S., and Milde F. (1993) Pulsed magnetron sputter
technology. Surf. Coat. Technol. 61, 331–337.
Schuegraf K.K. (ed.) (1988) Handbook of Thin Film Deposition Processes and Techniques. Noyes, Park Ridge, NJ.
Schmidt H., Rinn G., Nass R., and Sporn D. (1988) Film formation by inorganic-organic sol-gel synthesis. In: Better
Ceramics Through Chemistry III. MRS Vol. 121, 743–752.
Schoning M.J., Mourzina Y.G., Schubert J., Zander W., Legin A., Vlasov Y.G., and Luth H. (2001) Pulsed laser de-
position: an innovative technique for preparing inorganic thin films. Electroanalysis 13(8–9), 275–280.
Schoning M.J., Schubert J., Zander W., Veggian M.M., Legin A.V., Vlasov Y.G., Kordos P., and Luth H. (1999)
Pulsed laser deposition as a novel preparation technique for chemical microsensors. In: Proceedings of SPIE
Conference on Chemical Microsensors and Applications II, Boston, September 1999, vol. 3857, pp. 124–133.
Schoonman J., Dekker J.P., Briers J.W., and Kiwiet N.J. (1991) Electrochemical vapor deposition of stabilized zirco-
nia and interconnection materials for solid oxide fuel cells. Solid State Ionics 46, 299–308.
Schubert. J., Schoning M., Mourzina Y.G., Legin A.V., Vlasov Y.G., Zander W., and Luth H. (2001) Multicomponent thin
films for electrochemical sensor applications prepared by pulsed laser deposition. Sens. Actuators B 76, 327–334.
Schubert U., Gao Y., and Kogler F.R. (2007) Tuning the properties of nanostructured inorganic-organic hybrid
polymers obtained from metal oxide clusters as building blocks. Prog. Solid State Chem. 35, 161–170.
Semancik S., Cavicchi R.E., Kreider K.G., Suehle J.S., and Chaparala P. (1996) Selected-area deposition of multiple
active films for conductometric microsensor arrays. Sens. Actuators B 34, 209–212.
Shanthi S., Subramanian C., and Ramasamy P. (1999) Preparation and properties of sprayed undoped and fluorine
doped tin oxide films. Mater. Sci. Eng. B 57, 127–134.
Sherman A. (1987) Chemical Vapor Deposition for Microelectronics. Noyes, Park Ridge, NJ.
Sears W.M. and Gee M.A. (1988) Mechanics of film formation during spray pyrolysis of tin oxide. Thin Solid Films
165, 265–277.
Siefert W. (1984) Properties of thin In2O3 and SnO2 films prepared by corona spray pyrolysis, and a discussion of
the spray pyrolysis process. Thin Solid Films 120, 275.
Si J., Desu S., and Tsai C.Y. (1994) Metalorganic chemical vapor deposition of ZrO2 films using Zr(thd)4 as precur-
sor. J. Mater. Res. 9(7), 1721–1727.
Sladek K.J. and Gilbert W.W. (1972) Low temperature metal oxide deposition by alkoxide hydrolysis. In: Glaski
F.A. (ed.), Proceedings of 3rd International Conference on Chemical Vapour Deposition, Salt Lake City, Utah. American
Nuclear Society, Hinsdale, IL, pp. 215–231.
Skandan G., Chen Y.J., Glumac N., and Kear B.H. (1999) Synthesis of oxide nanoparticles in low pressure flames.
NanoStruct. Mater. 11, 149–158.
Skolheim T.A. (ed.) (1986) Handbook of Conducting Polymers. Marcel Dekker, New York.
Smith A. (2000) Pyrosol deposition of ZnO and SnO2 based thin films: the interplay between solution chemistry,
growth rate and film morphology. Thin Solid Films 376, 47–55.
Sol C. and Tilley R.J.D. (2001) Ultraviolet laser irradiation induced chemical reactions of some metal oxides, J.
Mater. Chem. 11, 815–820.
Song S.K. (1999) Characteristics of SnO2 films deposited by reactive-assisted deposition. Phys. Rev. B 60,
11137–11148.
Song J.-W., Kim J., Yoon Y.-H., Choi B.-S., Kim J.-H., and Han C.-S. (2008) Inkjet printing of single-walled carbon
nanotubes and electrical characterization of the line pattern. Nanotechnology 19, 095702.
Spitsyn B.V., Bouilov L.L., and Derjaguin B.V. (1988) Diamond and diamond-like films: deposition from the vapour
phase, structure and properties. Prog. Crystal Growth Charact. 17, 79–170.
Tahar R.B.H., Ban T., Ohya Y., and Takahashi Y. (1997) Optical, structural and electrical properties of indium oxide
thin films prepared by the sol-gel method. J. Appl. Phys. 82, 865–870.
Takahashi Y., Matsuoka Y., Yamaguchi K., Matsuki M., and Kobayashi K. (1990) Dip coating of PT, PZ and PZT
films using in alkoxide-diethanolamine method. J. Mater. Sci. 25, 3960–3964.
Takashima K., Minami K., Esashib M., and Nishizawa J. (1994) Laser projection CVD using the low temperature
condensation method. Appl. Surf. Sci. 79/81, 366–374.
Tamura S., Ishida T., Magara H., Mihara T., Tabata O., and Tatsuta T. (1999) Transparent conductive tin oxide films
by photochemical vapour deposition. Thin Solid Films 343–344, 142–144.
Tarasevich M.R. and Efremov B.N. (1981) Properties of spinel-type oxide electrodes. In: Trasatti S. (ed.), Electrodes
of Conductive Metallic Oxides, vol. 1, Elsevier, Amsterdam, pp. 221–259.
Taurino A.M., Epifani M., Taccoli T., Iannotta S., and Siciliano P. (2003) Innovative aspects in thin film technologies
for nanostructured materials in gas sensor devices. Thin Solid Films 436, 52–63.
Tay B.K., Zhao Z.W., and Chua D.H.C. (2006) Review of metal oxide films deposited by filtered cathodic vacuum
arc technique. Mater. Sci. Eng. R 52, 1–48.
Tiemann M. (2008) Repeated templating. Chem. Mater. 20, 961–971.
Thangaraju B. (2002) Structural and electrical studies on highly conducting spray deposited fluorine and antimony
doped SnO2 thin films from SnCl2 precursor. Thin Solid Films 402, 71–78.
Todorovska R.V., Groudeva-Zotova S., and Todorovsky D.S. (2002) Spray pyrolysis deposition of a-Fe2O3 thin
films using iron (III) citric complexes. Mater. Lett. 56, 770–774.
Tolstoy V.P., Bogdanova L.P., and Mityukova G.Y. (1986) USSR Patent No. 1386600. Prior. 06.11.1986.
Tolstoi V.P. (1993) Synthesis of superthin SnO2·nH2O layers on the surface of Si by layer by layer ionic deposition.
Russ. J. Inorg. Chem. 38, 1063–1065.
Tolstoy V.P. (1997a) The peroxide route of the successive ionic layer deposition procedure for synthesizing nano-
layers of metal oxides, hidroxides and peroxides. Thin Solid Films 307, 10–13.
Tolstoy V.P. and Ehrlich A.G. (1997b) The synthesis of CeO2+n nH2O nanolayers on silicon and fused-quartz sur-
faces by the successive ionic layer deposition technique. Thin Solid Films 307, 60–64.
Tolstoy V.P., Murin I.V., and Reller A. (1997c) The synthesis of Mn(IV) oxide nanolayers by the successive ionic
layer deposition method. Appl. Surf. Sci. 112, 255–257.
Tolstoy V.P. (1999) The synthesis of FeOOH nanolayers on the surface of SiO2 using layer by layer method. Russ.
J. Appl. Chem. 72(8), 1259–1261.
Tolstoy V.P. (2002) Synthesis by SILD of Ce4MnOX·nH2O nanolayers. Russ. J. Appl. Chem. 75(4), 673–678.
Tolstoy V.P., Gulina L.B., Korotcenkov G., and Brinzari V. (2003) Synthesis of nanolayers of hybrid-type hydroxo-
(SnxOyHz ) and heteropoly-(HxPWyOz ) compounds on silica surfaces by successive ionic layer deposition me-
thod. Appl. Surf. Sci. 221, 197–202.
Tomar M.S. and Garcia F.J. (1981) Spray pyrolysis in solar cells and gas sensors. Prog. Crystal Growth Charact. 4,
221–248.
Troczynski T. and Yang Q. (2001) Process for Making Chemically Bonded Sol-Gel Ceramics. U.S. Patent no.
6,284,682, May 2001.
Tschope A., Schaadt D., Birringer R., and Yying J. (1997) Catalytic properties of nanostructured metal oxides.
Synthesized by inert gas condensation. NanoStruct. Mater. 9, 423–432.
Vahlas C., Caussat B., Serp P., and Angelopoulos G.N. (2006) Principles and applications of CVD powder techno-
logy. Mater. Sci. Eng. R 53, 1–72.
Vayssieres L. (2007) An aqueous solution approach to advanced metal oxide arrays on substrates. Appl. Phys. A 89,
1–8.
Vincenzini P. (ed.) (1995) Advances in Inorganic Films and Coatings. Techna SRL., Italy.
Viswanathan V., Laha T., Balani K., Agarwal A., and Seal S. (2006) Challenges and advances in nanocomposite
processing techniques. Mater. Sci. Eng. R 54, 121–285.
Vossen J.V. and Kern W. (eds.) (1978) Thin Film Processes. Academic Press, New York.
Wahl G., Pulver M., Decker W., and Klippe L. (1998) CVD process for coating and surface modifications. Surf. Coat.
Technol. 100–101, 132–141.
Wang S., Wang W., Wang W., Jiao Z., Liu J., and Qian Y. (2000) Characterization and gas-sensing properties of na-
nocrystalline iron III oxide films prepared by ultrasonic spray pyrolysis on silicon. Sens. Actuators B 69, 22–27.
Warren W.L., Lenahan P.M., Brinker C.J., Shaffen G.R., Ashley C.S., and Reed S.T. (1990) Sol-gel thin film electronic
properties. In: Better Ceramics Through Chemistry IV. MRS Proceedings Vol. 180, pp. 413–419.
Westwood W.D. (1989) Physical vapor deposition. In: Microelectronic Materials and Processes. Proc. NATO Adv. Study
Inst. Kluwer Academic Publishers, Dorderecht, The Netherlands.
Will J., Mitterdorfer A., Kleinlogel C., Perednis D., and Gauckler L.J. (2000) Fabrication of thin electrolytes for
second-generation solid oxide fuel cells. Solid State Ionics 131, 79–96.
Willmott P.R. (2004) Deposition of complex multielemental thin films. Prog. Surf. Sci. 76, 163–217.
Xie H. and Raj R. (1993) Epitaxial LiTaO3 thin film by pulsed metalorganic chemical vapor deposition from a single
precursor. Appl. Phys. Lett. 63, 3146–3148.
Xu R. (1997) The challenge of precursor compounds in the MOCVD of oxides. JOM 49(10).
Yano S., Iwata K., and Kurita K. (1998) Physical properties and structure of organic-inorganic hybrid materials
produced by sol-gel process. Mater. Sci. Eng. C 6, 75–90.
Yoo K.S., Cho N.W., Song H.S., and Jung H. (1995) Surface morphology of SnO2-x thin films oxidised from Sn
films. Sens. Actuators B 24–25, 474–477.
Yue Y. and Gao Z. (2000) Synthesis of mesoporous TiO2 with crystalline frame work. Chem. Commun. 1755–1756.
Zakrzewska K. (2001) Mixed oxides as gas sensors. Thin Solid Films 391, 229–238.
Zarbov M., Schuster I., and Gal-Or L. (2002) Methodology for selection of charging agents for electrophoretic de-
position of ceramic particles. In: Proceedings of the International Symposium on Electrophoretic Deposition: Fundamentals
and Applications. The Electrochemical Society, vol. 21, pp. 39–46.
Zeigler J.M. and Fearon F.W.G. (eds.) (1990) Silicon Based Polymer Science: A Comprehensive Resource. ACS Advances in
Chemistry Series No. 224. American Chemical Society, Washington, DC.
Zeman P. and Takabayashi S. (2002) Effect of total and oxygen partial pressure on structure of photocatalytic TiO2
films sputtered on unheated substrate. Surf. Coating Technol. 153, 93–99.
Zhang J., Stauf G.T., Gardiner R., Van Buskirk P., and Steinbeck J. (1994) Single molecular precursor metalorganic
chemical vapor deposition of MgAl2O4 thin films. J. Mater. Res. 9(6), 1333–1336.
Zhang G. and Liu M. (2000) Effect of particle size and dopant on properties of SnO2-based gas sensors. Sens.
Zhitomirsky I. (2002) Cathodic electrophoretic deposition of ceramic and organoceramic materials—fundamental
aspects. Adv. Colloid Interface Sci. 97, 279–317.
CHAPTER 5
MODIFICATION OF SENSING MATERIALS

METAL OXIDE MATERIALS ENGINEERING
G. Korotcenkov
1. INTRODUCTION
As we have seen in other chapters, modern applications of solid-state gas sensors still face problems of
high cross-sensitivity to gases (i.e., low selectivity), high sensitivity to the humidity of the ambient at-
mosphere, long-term drift related to oxygen bulk diffusion, and transformation of the crystal structure
of polycrystalline materials (Morrison 1986; Gopel and Scherbaum 1995; Meixner and Lampe 1996).
Poor selectivity toward the monitored gas or cross-sensitivity with other gases may make a sensor’s
output unreliable. Also, both aging and poor stability may cause a sensor’s calibration curve to drift with
time (Wang et al. 1998). In comparison to the selectivity issue, which has been discussed extensively,
the problem of stability has rarely been addressed in the literature. However, this does not mean that
stability has less impact on the usefulness of a gas sensor. It is rather an engineering issue, which is often
addressed at the prototyping and manufacturing stages, and is difficult to resolve mostly because of the
lack of precise knowledge of the underlying physical and chemical processes.
Current efforts to resolve these issues tend to integrate results obtained from both empirical and
basic science approaches. They focus on various stages of sensor manufacturing, including the devel-
opment of new material systems, improvement of sensor fabrication processes, elaboration of smart
sensors (arrays) with signal conditioning, and filtration (Koudelka 2001; Sberveglieri 2001; Wang et
al. 1998). To solve the various problems, suggestions have been forwarded to change from ceramic
(Figaro type) to thin-film gas sensors; to modify known metal oxides; to use modern information
processing technologies; to elaborate new modes of measurement; and so on (Gopel and Scherbaum
313
1995; Sberveglieri 1995; Barsan et al. 1999; Williams 1999). In our classification, by thin films we mean
films with thickness usually less than 1 μm, which have been deposited using methods of microelectron-
ics such as chemical vapor deposition (CVD), spraying, sputtering, and so on.
This chapter focuses on discussion of the basic effects that are responsible for the gas-sensing
properties of metal oxides. Attention is directed mainly toward establishing some general regularities, to
provide a correlation between the structural, surface, and physical-chemical properties of metal oxides
and the gas-sensing characteristics of solid-state gas sensors. Possible methods suitable for modifying
the structural and physical-chemical properties of metal oxides are described in a general way, without
detailed descriptions, because we feel that excessive detail may hamper understanding of the nature of
the processes taking place. More information about these methods may be found in other chapters.
2. CONTROL OF SENSOR RESPONSE THROUGH STRUCTURAL

ENGINEERING OF METAL OXIDES
2.1. STRUCTURAL ENGINEERING—WHAT DOES IT MEAN?
At present, structural engineering of metal oxides is the most effective means available for optimizing
solid-state gas sensors. Considerable improvement of operating parameters such as sensor response,
selectivity, stability, and rate of response can be achieved by optimizing both the bulk and surface struc-
ture of the metal oxides used (Li et al. 1999; Korotcenkov 2007, 2008).
However, it is necessary to recognize that the technical potential of structural engineering to im-
prove nanoscale material parameters has so far not been well studied. For the most part, during gas
sensor design the main efforts of designers have focused on searching for optimal methods of doping
metal oxides using nanoscale metal catalyst additives and on decreasing crystallite (grain) size to the
nanometer scale. A dramatic increase in sensitivity when the sizes of both crystallites and intercrystallite
necks are reduced to the magnitude of the Debye length has now been demonstrated experimentally
many times (Suzuki and Yamazaki 1990; Yamazoe and Miura 1992; Brinzari et al. 2001; Korotcenkov
et al. 2004).
From our point of view, however, this is a simplified approach to structural engineering. In this
chapter we will show that gas sensitivity is more complicated and includes as important factors not only
the size, but also the crystallite shape, microscopic structure, and crystallographic orientation of the
crystallite planes forming the gas-sensing matrix. This conclusion is based on the following findings
(Masel 1996; Lambert and Pacchioni 1996; Henrich and Cox 1996; Lucas et al. 2001; Brinzari et al. 2001,
2002; Golovanov et al. 2002; Korotcenkov et al. 2005a):
1. The growth of grains, especially in the range from nanometers to micrometers, while the tran-
sition from spherulites to nanocrystallites, nanocrystals, and crystals takes place, may be ac-
companied by a change in the size and external shape of the crystallites (Brinzari et al. 2002;
Korotcenkov et al. 2005a). It is known that even crystals with the same crystallographic sym-
metry can have different external forms depending on their size. At the same time, the results of
structural studies show that even spherulities have real microplanes and facets.
MODIFICATION OF SENSING MATERIALS • 315
2. The external planes of nanocrystals participate in gas–solid interaction, and therefore it is these
planes that determine the gas-sensing properties of nanostructural materials. In addition, nano-
crystal shape may determine such parameters as which crystallographic planes participate in
formation of intergrain contacts, the areas of intergrain contact, the gas permeability of the
gas-sensing matrix, and others.
3. Every crystallographic crystal form has its own combination of crystallographic planes, fram-
ing the nanocrystal, and every crystallographic plane has its own overall combination of surface
electron parameters. These include surface state density, energy levels induced by adsorbed spe-
cies, adsorption/desorption energies of interacting gas molecules, concentration of adsorption
surface states, position of the surface Fermi level, activation energy of native point defects, and
so on. This means that chemisorption characteristics change noticeably in going from a crystal
surface with one orientation to another, indicating a striking variation of chemical bonding of
adsorbed particles with the surface that depends on atomic structure.
Brinzari et al. (2000a, 2001) developed a qualitative phenomenological model of thin-film
gas sensors (GSs) which was based on the chemisorption theory proposed by Volkenstein (1963).
This model allowed understanding of the main regularities of GS behavior. Theoretical simula-
tions conducted in the frame of this model (Brinzari et al. 1999, 2000a, 2001, 2002a) have shown
that the above-mentioned parameters determine the degree of gas sensitivity, the operating tem-
perature at which sensor response reaches maximum, and the rate of response and recovery
processes. It was established that a change of adsorption/desorption parameters might change
the gas sensitivity, and shift greatly the temperature of maximum sensor response.
4. Adsorption/desorption processes that determine gas sensitivity have activation energies, and the
parameters that affect these processes are orientation- and grain size–dependent A decrease in
crystal size to the nanometer range notably strengthens the influence of crystallite shape on ad-
sorption properties. Both the shape and the size of nanocrystals have a profound impact on the
amount of species that are adsorbed, and on the type of bonding to the surface that takes place.
For example, surface species with different types of bonding may have preferred adsorption on
the edge/corner sites or on the flat planes.
All of this means that, depending on external form of the nanocrystallites, the nanostructure gas-sensing
matrix will have its own individual total combination of structural, electronic, and adsorption/desorption
process parameters.
Detailed studies (Brinzari et al. 2002a; Korotcenkov et al. 2005a) conducted on SnO2 films de-
posited by spray pyrolysis have confirmed the appropriateness of the approach we have discussed for
modifying the gas-sensing properties of metal oxides. It was found that, depending on both deposition
parameters and film thickness, SnO2 nanocrystallites may take various forms, being faceted in different
crystallographic planes. As an example, the shape of SnO2 nanocrystals, deposited by spray pyrolysis
at Tpyr = 425°C is shown in Figure 5.1. For this example, crystallographic models of SnO2 crystallites
were simulated on the basis of the results of x-ray diffraction and high-resolution transmission electron
microscopy (HRTEM) measurements (Korotcenkov et al. 2005a).
One can see in Figure 5.1 (Korotcenkov et al. 2005a) that, in addition to the most stable plane
(110), the external facets of SnO2 nanocrystals can be formed by crystallographic planes with other
Figure 5.1. Results of computer image simulation of SnO2 nanocrystallites deposited by spray pyroly-
sis at Tpyr = 425°C: (a) d ≈ 40–80 nm; (b) d ≈ 300 nm. (Reprinted with permission from Korotcenkov
et al. 2005a. Copyright 2005 Elsevier.)
--
structural and electronic configuration planes, such as (111), (200), (101), (011), (112), and (210). At that
the combination of these planes depends on the grain size. In other words, the results in Figure 5.1
show that even a slight increase in SnO2 grain size can produce a marked change in the film structure,
which is responsible for both the catalytic and gas-sensing properties of metal oxides (Golovanov et
al. 2002).
Examples of some crystallographic planes observed in SnO2 films deposited by spray pyrolysis are
shown in Figure 5.2. From an electronic point of view, the (110) surface is likely to be the most stable
one, because it has the lowest density of dangling bonds. On this surface we find both fivefold- and
(a) (c)
(b) (d)
Figure 5.2. Models of unrelaxed rutile surfaces: (a) (110); (b) (100); (c) (001); (d) (101). Lighter
balls represent tin atoms and darker balls represent oxygen atoms. (Reprinted with permission from
Korotcenkov et al. 2005d. Copyright 2005 IOP.)
sixfold-coordinated tin. The (110) surface is not atomically flat. Symmetric rows of “bridging” O ions
lie an equal distance above and below the surface plane. The bridging O ions can be easily removed
and replaced, depending on surface treatment. In the case of a stoichiometric surface, the fivefold-
coordinated surface tin atoms are of particular interest in providing active sites for adsorbates.
The (100) plane shown in Figure 5.2b is a rather rough surface, with O rows lying above the surface
plane. All of the cations on this surface are fivefold-coordinated with O ligands and therefore may serve
as surface sites for adsorbates. Formation of bridging oxygen vacancies results in fourfold-coordinated
tin ions. Deep reduction of the surface may lead to the appearance of threefold-coordinated cations.
The (001) surface shown in Figure 5.2c, in spite of the high symmetry of the lattice in that direc-
tion, is the least stable of the low-index faces. The nearest-neighbor ligand coordination number of
the surface cations has been reduced from six to four. Although this surface is nonpolar, the low ligand
coordination of the surface cations favors reconstruction to increase that coordination.
The ideal cleavage of the (101) atomic face, shown in Figure 5.2d is also a nonpolar, microscopi-
cally rough surface, with nonsymmetrically coordinated O rows lying above the surface plane and form-
ing a rectangular pattern. The surface concentration of such twofold-coordinated “bridging” oxygens
is higher, and they located closer to the atomic plane than at the (110) face. All of the cations on this
surface are fivefold-coordinated with O ligands and therefore may serve as surface sites for adsorbates.
The involvement of weakly bound bridging oxygens in the surface reaction results in fourfold-coordi-
nated tin ions.
As follows from reported results (Sunagawa 1990; Hartman 1993; Kawamura et al. 2001), the (110)
and (101) planes of the SnO2 crystal are F (flat) faces, while the (111) plane is K (kinked). According to
the periodic bond chains (PBC) theory (Hartman 1993), surface atoms of F faces are strongly bound to
each other in directions parallel to the face, so these faces have a slight tendency to react with arriving
atoms. In reality, this statement is not entirely correct, because besides “in-plane” oxygen on the (110)
SnO2 surface, there are also “bridging” oxygen ions, located at equal distances above and below the sur-
face plane. However, the statement that the (110) surface plane is the most stable is correct. Indeed, the
faces parallel to only one PBC (stepped S faces) or to none (kinked K faces) are highly reactive, because
there are many unsaturated bonds cutting their surfaces. It was also established that the (110) and (101)
surfaces of SnO2 grow according to a layer-by-layer growth mechanism, while the (111) surface grows
by the addition of chains, or the piling up of edges of growth layers on the (110) surface (Sunagawa
1990; Hartman 1993; Kawamura et al. 2001). This means that the (111) plane has a much rougher sur-
face than either the (110) or (101) plane.
The catalytic activity of atomic faces is to a large extent determined by the surface concentration
of nonsaturated cations and weakly bound bridging oxygen. The (110) face as shown in Figure 5.2a is
the most stable and therefore the most extended surface in SnO2 films. However, the contribution from
the other faces may be rather important under certain conditions. Analysis carried out by Brinzari et
al. (2002a) allows one to propose the following ranking for catalytic activity (CA) of ideal SnO2 atomic
planes: CA(110) < CA(001), CA(100) < CA(101).
On the other hand, simple estimates show that dissociative chemisorption on the SnO2 surface is
also orientation-dependent. Various crystallographic planes have different distances between Sn atoms,
which follow the following ranking: d(110) ≈ d(100) < d(101) < d(001) (Brinzari et al. 2002). Tin atoms are cen-
ters of oxygen chemisorption, and therefore the change in distance influences the rate of dissociative
oxygen chemisorption, which in many cases is a main process of gas-sensing phenomena (Moseley and
Tofield 1987; Madou and Morrison 1989; Gopel et al. 1991; Barsan et al. 1999; Brinzari et al. 2000a).
Thus we have confirmation that an observed change in SnO2 crystal faceting may lead to considerable
modification of the gas-sensing characteristics of SnO2 films. Of course, the same reasoning can be
carried out for other sensing materials as well. In other words, optimization of metal-oxide crystal fa-
ceting may be an effective method for improving the gas-sensing characteristics of metal oxide–based
gas sensors.
It has also been determined that, like grain size and shape, both film thickness and agglomeration
play important roles in sensor response (Korotcenkov et al. 2003b, 2005a). An example of an SnO2 ag-
glomerated film is shown in Figure 5.3.
The gas-sensing matrix of an agglomerated polycrystalline film can be represented schematically as
an equivalent circuit (see Figure 5.4). Here, by “gas-sensing matrix” we understand a three-dimensional
matrix of crystallites (grains) forming the gas sensing layer. In this case, R(a–a) represents the resistance
of interagglomerate contacts, Rc is the resistance of intercrystallite (intergrain) contacts, Rb is the bulk
resistance of crystallites (grains), and Ragl is the resistance of the agglomerate.
From this scheme, one can see that in our matrix we have four gas-sensitive elements that deter-
mine the gas-sensing characteristics of this structure: (1) intergrain (intercrystallite) contacts; (2) inter-
agglomerate contacts; (3) agglomerates; and (4) grains. Here the agglomerate resistance is an integral
resistance, representing the three-dimensional grain network. Thus it is seen that by influencing the
porosity of both agglomerates and the gas-sensing matrix, we can change the role of either gas-sensing
element in affecting sensor response, and in that way can affect the gas-sensing parameters. For exam-
ple, decreasing grain size increases the role of grains in gas-sensing effects. Low gas permeability of
agglomerates promotes an increase in the influence of interagglomerate contacts on sensor response.
Figure 5.3. Typical SEM image of agglomerated SnO2 film; Tpyr = 370°C; d ≈ 70 nm. (Adapted with
permission from Korotcenkov et al. 2005a. Copyright 2005 Elsevier.)
Figure 5.4. Principal electrical circuit of agglomerated polycrystalline film. (Adapted with permission
from Korotcenkov 2008. Copyright 2008 Elsevier.)
Simulation results (Williams and Pratt 2000) have shown that the agglomeration of small crystallites
into large masses within the fabricated devices is a key phenomenon leading to great variation of ap-
parent response characteristics.
Based on the analysis of SnO2 sensors fabricated using selective ionic-layer deposition (SILD),
Korotcenkov et al. (2003b) concluded that small crystallite (grain) size is an essential, but not sufficient,
condition for achieving both maximum gas sensitivity and fast response. First of all, the gas-sensing
matrix should be highly porous, and the size of the agglomerates must be minimum. Moreover, a para-
doxical situation appears for strongly agglomerated structures. In such structures, small grain (crystal-
lite) size is not an advantage. Agglomerates from smaller grains are more densely packed, i.e., they have
less gas penetrability. Therefore, such structures may actually have worse gas-sensing characteristics
than films with larger crystallites.
In most cases, metal oxide films deposited using various technologies are agglomerated (Barsan et
al. 1999). In this context it becomes clear that during comparative analysis of gas-sensing characteristics
of metal oxide–based sensors it is necessary to consider primarily the size of agglomerates and their
porosity, and only after that the grain size and film thickness.
So, in applying structural engineering, we should control the following geometric parameters of the
metal-oxide gas-sensing matrix (Brinzari et al. 2000a, 2001):
• Grain size
• Agglomeration
• Area of intergrain and interagglomerate contact
• Thickness
• Porosity
• Dominant orientation and faceting of crystallites that form the gas-sensing surface
However, it is necessary to note that these parameters are only some of the metal-oxide matrix
parameters which can influence the gas-sensing properties (see Figure 5.5). In addition to geometric
Structural engineering
Geometry control Chemical composition control
Thickness Grain size Phase Elemental

composition composition
Intergrain
Grain faceting Bulk
contacts Surface
stoichiometry architecture
Grain network Porosity

Kind of Type of
Agglomeration additives conductivity
Figure 5.5. Structural parameters for optimization of metal oxides. (Adapted with permission from
Korotcenkov 2008. Copyright 2008 Elsevier.)
parameters, we should also take into account the physical-chemical parameters of the gas-sensing ma-
trix, including:
• Chemical and phase composition

• Type of conductivity of additives
• Bulk and surface oxygen vacancy concentrations (stoichiometry)
• Size and density of both metal catalyst particles and single atoms on the surface of the
metal oxide
• Surface architecture
• Type and concentration of noncontrolled impurities
As shown in previous chapters, these parameters control the electrophysical, adsorption, and cata-
lytic properties of metal oxides. Therefore, as well as geometric parameters, they exert a great influence
on metal-oxide gas-sensing characteristics.
With regard to noncontrolled impurities that may influence gas sensitivity, it is necessary to high-
light carbon and its compounds; various forms of water; and interfering chemisorbed species between
water and carbon-containing species (Michel et al. 1998; Brinzari et al. 2000b). For example, results
presented by Brinzari et al. (2000b) have shown that, on one hand, the concentration of carbon may
be high enough to block many centers of adsorption at the SnO2 surface, and, on the other hand, the
carbon concentration may be restored quickly even after treatments targeted to removing carbon from
the SnO2 surface (see Figure 5.6). As seen in Figure 5.5, the initial carbon concentration on the SnO2
surface is already being restored after 10 min of exposure in air at Тan = 300°С.
2.2. STRUCTURAL ENGINEERING OF METAL OXIDES—TECHNICAL

APPROACHES
A large variety of methods may be used to control the parameters of metal oxide sensors. These meth-
ods are summarized in Figure 5.7.
2.2.1. Control of Deposition Parameters

Control of metal oxide properties during their synthesis or deposition may be achieved by controlling
parameters such as pyrolysis temperature, substrate temperature, precursor concentration, gas flow,
oxygen concentration in the deposition chamber, rate of deposition, and so on.
When selecting technical parameters for film deposition, it is important to remember that the
deposition temperature is the principal parameter, in that it may affect film parameters such as grain
size and faceting, agglomeration, both the surface and bulk stoichiometry of crystallites, texture, and so
on (Brinzari et al. 2002a; Korotcenkov et al. 1999, 2001, 2004). For example, in terms of dependence
on Tpyr, SnO2 films may have granular structure with a grain size ~5–7 nm (Figure 5.8a) or columnar
structure with a grain size of more 60-100 nm (Figure 5.8b).
SnO2
Figure 5.6. Influence of various thermal treatments on intensities of main peaks of SnO2 x-ray pho-
toelectron spectra (φ = 90°): A, treatments in vacuum; B, treatments in air, Tan = 300°C, tan = 10 min;
A, dry air; B, humid air.
Figure 5.7. Methods of structural engineering to optimize parameters of solid-state gas sensors.
Through pyrolysis temperature control from 325 to 550°C, it was found to be possible to de-
posit SnO2 films textured in the (101), (110), or (200) direction (Brinzari et al. 2002a; Korotcenkov et
al. 2005a). As a result, it was possible to control the temperature of the sensor response maximum,
Smax(Toper), in the range from 280 to 450°С (Korotcenkov et al. 2001a) (see Figure 5.9).
The same effect was observed earlier in a study of gas-sensing properties of SnO2 films prepared
by chemical vapor deposition (Lalauze et al. 1991). In this research the temperature of maximum sensor
response was shifted in the range 200–360°C, depending on the deposition temperature. This large shift
of the maximum sensor response indicates that modification of film structure through control of the
technology really can affect the surface parameters that determine gas-sensing characteristics.
This means that determination of crystallographic planes with optimal combinations of
adsorption/desorption and catalytic parameters, and the development of methods for metal oxide de-

(a)
SnO2
Si substrate SiO2
(b)
SnO2
Si substrate
Figure 5.8. TEM cross-section micrographs of SnO2 films deposited by spray pyrolysis: (a) Tpyr = 330–
350°C, d ≈ 70–80 nm; (b) Tpyr = 510°C, d ≈ 75–100 nm. (Adapted with permission from Korotcenkov
et al. 2005a. Copyright 2005 Elsevier.)
position with specified faces of grains, can be considered one of the main contemporary goals of thin-
film technology for gas sensor applications. Accomplishing this task will improve selectivity and allow
high sensor response to be obtained in the range of low operating temperatures without surface modi-
fication of the metal oxides by noble metals. This last factor, which will be discussed later, is important
to improve the stability of solid-state gas sensors.
Tpyr= 350oC Tpyr=530 oC
Figure 5.9. Influence of SnO2 deposition parameters on normalized temperature dependence of

gas response to CO: (a) influence of pyrolysis temperature; (b) and (c) influence of film thickness.
(Adapted with permission from Korotcenkov et al. 2001a. Copyright 2001 Elsevier.)
410oC
Figure 5.10. Influence of pyrolysis temperature on grain size and number of particles in agglo-
merates during SnO2 deposition. (Reprinted with permission from Korotcenkov 2005c. Copyright
2005 Elsevier.)
In choosing a temperature for film deposition, it is necessary to take into account the dual nature of
the influence of pyrolysis temperature on the film structure (Figure 5.10). To obtain films with minimum
grain size, it is necessary to use low deposition temperatures (Brinzari et al. 2002a; Korotcenkov et al.
2001b). However, at low Тpyr, the probability of crystallite agglomeration increases sharply, which in turn
adversely affects the rate of gas response; Korotcenkov et al. (1998) established that an increase of both
agglomerate size and film thickness after some critical size always leads to an increase of response time.
Therefore, sensors based on thin films with minimum agglomeration will always have a maximum
rate of response. As Тpyr increases, film agglomeration decreases sufficiently. However, considerable
growth of both grain size and film density also takes place, which is undesirable for attaining maximum
gas sensitivity and minimum response time. Therefore, we have to seek a compromise between grain
size and the possibility of agglomerates forming in growing films. Many researchers have shown that
the easiest way to solve this problem is to use ultrathin films, with thicknesses of less than 40–60 nm.
In such films, neither the agglomerate size nor the film thickness reaches the critical value at which their
gas penetration ability starts to influence the kinetics of sensor response.
When using sputtering technologies, the partial pressure of oxygen in the deposition chamber is
of great importance as well. Lalauze et al. (1991) established that at low oxygen pressure the films, as a
rule, are very compact, and increasing pressure is accompanied by decreasing metal oxide density. For
example, during SnO2 films deposition by magnetron sputtering, Lalauze et al. (1991) determined that
the film density decreased by a factor of ∼100 when oxygen pressure increased from 2 to 150 Pa. It is
known that film density (porosity) is related directly to the area of active surface. Therefore, decreas-
ing the density of deposited films is undoubtedly an important factor in improving the gas response
characteristics of sputtered films.
2.2.2. Selection of Substrate

A great number of substrates are suitable for deposition of gas-sensing metal oxide films. They include
various dielectric polycrystalline Al2O3-based substrates, Si, MgO, ZrO, Al2O3 single crystals with spe-
cific crystallographic orientation, glass ceramics, and so on.
By studying the structure of SnO2 and In2O3 films deposited by spray pyrolysis on Si, SiO2, and
Al2O3 substrates, Korotcenkov et al. (2000) established that these films are mechanically strained. Even
in cases of high texture between the substrate and the film, there is a transitional layer with a thickness
of ~10–20 nm and consisting of randomly oriented crystallites. A simplified structure of such films is
shown in Figure 5.11. It was found that mechanical strains in SnO2 films deposited by spray pyrolysis
on Si substrates are not reduced after postdeposition thermal treatments (Korotcenkov et al. 2000,
2005), and it was concluded that the strains were conditioned by the difference in thermal expansion
coefficients between the substrate and the deposited film.
It appears that, by selecting substrates with corresponding lattice parameters and thermal expan-
sion coefficients, it is possible to minimize the mechanical strains and the thickness of transition layers
which are the sources of temporal instability of the film (Korotcenkov et al. 2005b). Moreover, under
proper conditions one can provide even epitaxial growth of metal oxide films. Descriptions of some
technical aspects of such film deposition have been published (Chambers 2000; Kamei et al. 2001; Lee
et al. 2001; Tarre et al. 2002). From our point of view, such an approach to deposition of gas-sensing
layers is very promising, because a transition from polycrystalline layers to epitaxial ones would resolve
one problem of film structure instability. Some work in this direction has been carried out, and there
may soon be some breakthrough in elaboration of epitaxial metal-oxide gas sensors. Already, first ex-
periments with epitaxial SnO2 films have shown that gas-sensing characteristics of such films are stable
under ambient conditions and, in particular, in the presence of water (Gopel and Scherbaum 1995; Lee
et al. 2001).
Another interesting trend in research which may provide good results in structure control of de-
posited metal oxide films is the search for ways to manage the process of metal-oxide grain growth
during film deposition. Grain growth conditions depend strongly on the initial state of the substrate
surface (particularly, the concentration of the centers of growth) (Palatnik et al. 1972). Therefore, modi-
Figure 5.11. Simplified view of metal oxide film with transient layer. (Reprinted with permission from
Korotcenkov 2005c. Copyright 2005 Elsevier.)
fying the substrate surface using certain nanoparticles (Panchapakesan et al. 2001) may allow control of
grain growth, especially during the initial stage of deposition. Panchapakesan et al. (2001) showed that
depositing nanoparticle seeds of those metals with the highest melting temperatures (Co, Fe, Ce, and Ni
oxides) on the surface of a dielectric substrate leds to a considerable decrease in the grain size of SnO2
films and improved their thermal stability. As a result of preliminary substrate surface modification by
Co and Ni, the grain size of SnO2 during chemical vapor deposition (T = 500°C) on a micro-hotplate
gas sensor platform was decreased by more than three times, from 75 to 20–25 nm.
2.2.3. Postdeposition Treatments

Treatments used after deposition include postdeposition annealing (Suzuki and Yamazoe 1990; Serrini
et al. 1997), various radiation treatments (Rosenfeld et al. 1993), and other processes.
Korotcenkov et al. (2005b) showed that during annealing, the smallest grains are being joined first.
Therefore, thermal treatments can be used to stabilize the properties of already-deposited films. For
example, research on the structural stability of In2O3 films has shown that the change of In2O3 film
structure during thermal treatment in an oxygen-containing atmosphere goes through four standard
stages of structural transformation of polycrystalline materials: (1) structural stability of the films; (2)
coalescence of grains to form agglomerates; (3) local structural reconstruction; and (4) global (com-
prehensive) structural reconstruction (Korotcenkov et al. 2005b). These stages are shown in Figure
5.12 (Korotcenkov et al. 2005b). Results obtained allow asserting that annealing at temperatures of
450–500°С is sufficient for structural stabilization of deposited films.
However, it is necessary to consider that thermal treatment, in addition to influencing both the
size (Dieguez et al. 1997) and shape of crystallites (see Figure 5.13), is accompanied by a change in the
concentration of surface point defects, which are adsorption centers of chemisorbed oxygen and other
surface species participating in the reactions of gas detection.
The influence of postdeposition treatment on metal-oxide film parameters was also analyzed by El-
Azab et al. (2002). Studying SnO2 films deposited by magnetron sputtering at room temperature (RT),
Figure 5.12. Stages of structure transformation of In2O3 films during thermal treatment. (Reprinted
with permission from Korotcenkov et al. 2005b. Copyright 2005 Elsevier.)
a b
c . d
Figure 5.13. AFM images of In2O3 films (d ≈ 400 nm) after thermal treatment at: (a) 600°C; (b)
900°C; (c) 1000°C; (d) 1100°C. (Adapted with permission from Korotcenkov et al. 2005b. Copyright
2005 Elsevier.)
El-Azab et al. (2002) found that thermal treatment at Tan = 375°C in air did not greatly change SnO2
grain size, but it improved appreciably both the crystallinity and the adsorption of oxygen species on
the film surface. Serrini et al. (1997) established that these processes act the same way, i.e., they promote
an improvement of gas sensitivity of SnO2 films with thickness 200–400 nm to both NO2 and CO. The
sensor response of SnO2-based devices after such treatments increased more than 5–20 times.
However, the mentioned Tan probably is not universal, and depends on either the deposition tech-
nique used or the initial film structure. For example, according to Suzuki and Yamazoe (1990), long-
term thermal annealing of ultrathin SnO2 films (d ≈ 15–18 nm) at Tan = 500°C has led to even more
impressive results. Sensor response to H2 (0.5%) after annealing for 100 h increased by more than 103
times. It is believed that one of a consequences of the indicated thermal treatment was pre-aging of
as-deposited SnO2 films, i.e., an improvement of their stoichiometry, and, as a result, a sharp decrease
of bulk concentration of free charge carriers. After-effects of such a drop in the concentration of free
charge carriers for sensor response were considered earlier by Brinzari et al. (2000, 2001).
2.2.4. One-Dimensional Nanostructures

At present, the use of individual one-dimensional nanostructures is considered one of the most prom-
ising directions of structural engineering research for new-generation chemical sensors. There are many
types of one-dimensional structures, including nanowires, nanobelts, nanotubes, nanoribbons, nano-

whiskers, nanorings, nanorods, etc. (Dai et al. 2002; Hu et al. 2002; Li et al. 2003; Kam et al. 2004; Zhang
et al. 2004). Increased attention to such structures is understandable because the use of one-dimensional
structures may resolve the most important problem of chemical sensors: increasing the temporal stabil-
ity of their parameters. One-dimensional structures are monocrystals, and so the problem of instability
of grain size, typical of polycrystalline materials, does not occur.
In addition, one-dimensional structures are crystallographically perfect and have clear faceting with
a fixed set of planes. However, as much research has shown, the faceting of one-dimensional struc-
tures can change depending on the fabrication parameters. For example, Kong et al. (2003) reported
that, using solid–vapor deposition, they were able to synthesize In2O3 nanobelts with (100) and (120)
growth directions. Both types of nanobelts had top and bottom surfaces of (001), while the (100)-type
nanobelts had side surfaces of (010) and a rectangular cross- section. The (120)-type nanobelts had a
parallelogram cross section. In addition, it was reported (Liang et al. 2001; Li et al. 2003) that In2O3
nanobelts may have growth directions of (111) and (110) with side and top surfaces being (100) as well.
As was demonstrated earlier, this fact offers additional opportunities for control of sensor parameters.
Typical growth directions and facets for some metal oxide nanobelts are shown in Table 5.1.
In analyzing the opportunities of one-dimensional structures of various types for their practical
use in chemical sensors, it is necessary to admit that nanobelts (nanoribbons) might be the most de-
manding. Nanobelts are thin and plain belt-type structures with rectangular cross section (see Figure
5.14). At present, nanobelts are available for practically for all basic oxides used in chemical sensors,
There is information about synthesis of nanobelts based on SnO2, In2O3, ZnO, Ga2O3, TiO2, etc. (Li
Table 5.1. Crystallographic parameters of some metal oxide nanobelts

(nanoribbons)
METAL OXIDE GROWTH DIRECTION FACETS REFERENCE

ZnO [0001] (2-1-10) and (01-10) (Pan et al. 2001)
[01-10] (0001) and (2-1-10) (Pan et al. 2001)
SnO2 [101] (010) and (0-10) (Dai et al. 2001)
[101] (10-1) and (-101) (Dai et al. 2001)
[110] (Hu et al. 2002)
[203] (Hu et al. 2002)
In2O3 [101] (100) (Pan et al. 2001)
[100] (001) and (010) (Kong et al. 2003)
[120] (001) and (2-10) (Kong et al. 2003)
[110] (100) (Liang et al. 2001)
[111] (Li et al. 2003)
CdO [100] (001) and (010) (Pan et al. 2001)
Ga2O3 [010] (10-1) and (101) (Lee et al. 2002)
Figure 5.14. SEM image of SnO2 nanoribbons. (Reprinted with permission from Kong and Li 2005.
et al. 2000; Pan et al. 2001; Dai et al. 2002; Lee et al. 2002; Konga and Wang 2003; Varghese et al. 2003;
Jung et al. 2003; Kong et al. 2003; Guha et al. 2004). Typical nanobelts have widths of 30–300 nm, thick-
nesses of 10–15 nm, and lengths from several micrometers to hundreds, or even some thousands, of
micrometers (Dai et al. 2001; Pan et al. 2001). Synthesis of nanobelts can be accomplished using various
methods (Liu et al. 2001; Wang et al. 2002; Li et al. 2002; Xu et al. 2002), which creates good conditions
for widening the research into such nano-size materials.
Nanobelts do not have the mechanical strength of nanotubes. However, their structural homoge-
neity and crystallographic perfection are very good attributes. Because there are zero defects in their
structure, there is no problem, as there is with nanotubes, that defects may destroy quantum-level prop-
erties. It is necessary to emphasize, however, that important advantages of nanobelts for mass manu-
facturing are their suitable geometry (see Figure 5.15), high structural homogeneity, and long length.
Nanobelts are flexible and therefore can be curved over 180° without being damaged. This fact gives
additional advantages to those materials for devices design.
However, it is necessary to note that considerable success in manufacturing of gas sensors based
on individual one-dimensional structures has not been attained as yet. A significant part of research
in this direction has been dedicated to elaborating methods of synthesis and control of nanostructure
parameters. At the same time, one must admit that this field of research is being developed successfully
enough (Comini et al. 2002; Li et al. 2003; Kolmakov et al. 2003; Maiti et al. 2003; Kong and Li 2004).
Research reported in these articles has shown that chemical sensors based on one-dimensional metal
oxide structures can show high sensitivity and great stability at room temperatures. So far, these sensors
have too long response times and especially recovery times. However, one can expect that resolving the
problem of operating temperature increase will promote the decrease of response and recovery times
to acceptable values. It is important that the observed sensitivity of one-dimensional metal oxide sen-
Nanowires Nanobelts
Nanotubes Nanoribbons
Figure 5.15. Position of (a) nanowires or nanotubes and (b) nanobelts on the contact pad. (Reprinted
with permission from Korotcenkov 2008. Copyright 2008 Elsevier.)
sors is significantly higher than the reported sensitivity of carbon nanotubes (Kong et al. 2000). It is
assumed that the difference is related to the nature of the metal oxide surface, which can react readily
with ambient species, compared to the inert sidewall of carbon nanotubes (Kong et al. 2000).
What is more, taking into account the specificity of chemical sensors, one can expect that wide
applications of one-dimensional structures in sensor technology will come sooner than in standard
microelectronics. As was shown in a previous chapter, chemical sensors can have very simple construc-
tion, which does not require a lot of connections as in integrated circuits. In addition, requirements for
ohmic contacts in chemical sensors are not so strict. Chemical sensors permit considerably more range
in the parameters of sensing materials as well.
It is also necessary to emphasize that semiconducting oxide nanobelts with well-defined geometry
and perfect crystallinity could become a perfect model materials family for systematic experimental
study and theoretical understanding of the fundamentals of gas sensing effects in metal oxides.
2.2.5. Other Approaches to Structural Engineering of Metal-Oxide

Gas Sensors
Many other technical approaches, besides those discussed so far, may be used to optimize gas-sensing
characteristics. Many of them were presented in previous chapters of this book. For example, attempts
have been made to improve parameters by creating a special geometry and particular contact positions
(Williams and Pratt 1996; Cabot et al. 2002b), creating two- or multilayered structures from gas-sensing
materials (Hyodo et al. 2001; Kim et al. 2002), and so on. However, many such methods are just techni-
cal in character and so will not be discussed in this chapter. Some approaches will be discussed in other
volumes of this series. For example, elsewhere a chapter will be devoted to discussion of the synthesis
and application of mesoporous materials. Detailed descriptions of mesoporous structures and hollow
oxide nanostructures can also be found in reviews by Tiemann (2007) and Lee (2009). Methods for
chemically modifying metal oxides to optimize gas-sensing characteristics are discussed in detail in the
next section.
3. SENSOR RESPONSE CONTROL THROUGH MODIFICATION OF

METAL OXIDE COMPOSITION
3.1. PHASE MODIFICATION OF METAL OXIDES

Research has shown that simple metal oxides cannot meet all the requirements of a perfect gas-sensing
matrix. Therefore, to overcome the imperfections of current solid-state sensors, in many works it has
been suggested that designers turn away from simple binary gas-sensitive materials, such as TiO2, SnO2,
ZnO, In2O3, Fe2O3, or WO3, and use more complex, multicomponent materials instead (Yamazoe et
al. 1983; Yamazoe and Miura 1992; Meixner and Lampe 1996; Yamaura et al. 1996, 2000; Kanazawa et
al. 2001; Shimizu et al. 2001; Wang et al. 2002; Korotcenkov et al. 2004e). In addition to the base oxide
matrix, these multicomponent materials will contain various additives, introduced into the metal oxides
during their synthesis or deposition, to improve the gas-sensing parameters.
This approach is based on the assumption that including elements with different physical-chemical
properties in the gas-sensing matrix provides additional factors with which to influence the important
parameters of the metal oxides (see Figure 5.16). H. Gleiter (1992, 2000) was one of the first research-
ers to indicate that the development of well-defined materials, such as metal oxides, can be attained by
advancement of microstructures through the incorporation of additional phases (different oxides) in
nanocrystalline systems. For these purposes may be used both catalytic active additives (noble metals
and transition-metal oxides) and inert impurities. As shown in Figure 5.16, these additives can modify
the catalytic activity of the base oxide, stabilize a particular valence state, favor formation of active
Figure 5.16. Parameters that may be changed as a result of metal oxide doping during preparation
(deposition). (Reprinted with permission from Korotcenkov 2005c. Copyright 2005 Elsevier.)
phases, stabilize the catalyst against reduction, or increase the electron exchange rate. Introduction of
additives into base metal oxides can change such parameters as the concentration of charge carriers,
the chemical and physical properties of the metal oxide matrix, the electronic and physical-chemical
properties of the surface (energetic spectra of surface states, energy of adsorption and desorption,
sticking coefficients, etc.), surface potential and intercrystallite barriers, phase composition, sizes of
crystallites, and so on (Yamazoe et al. 1983; Morrison 1986; Gopel and Scherbaum 1995; Ratko et al.
1998; Kawamura et al. 2001; Skala et al. 2003; Korotcenkov et al. 2004e).
3.2. METHODS OF PHASE MODIFICATION

To date, a great many additives have been tested for phase modifications of metal oxides (see Table 5.2).
For example, Kanazawa et al. (2001) analyzed the influence of more than 25 additives.
Phase modification of metal oxide materials can be done using three basic approaches:
1. During the process of synthesis or deposition of initial material

2. Using layer-by-layer deposition of chosen materials
3. Mixing already-synthesized materials in certain proportions
It is necessary to note here that the second and the third approaches suppose a subsequent high-
temperature treatment, which will reduce the effectiveness of the methods and limit their application.
It is obvious that the third approach cannot be used with thin-film technology; this approach is limited
mainly to ceramics, for which high-temperature annealing is one of the usual steps in the manufacture
of chemical sensors. The second and third approaches also suppose that at least one of the components
participating in solid-phase reactions has sufficient mobility at the annealing temperatures used.
Regarding the first approach, for its realization one could use all methods of synthesis and deposi-
tion, as described earlier in this volume. However, the most effective for this approach are sol-gel syn-
thesis processes and aerosol-phase deposition methods. The advantages of those methods for forming
multiple-component materials and precursors have been discussed in Chapter 4.
Table 5.2. Elements often used for phase modification of metal oxides
Noble metals Pd, Pt, Au, Rh, Ru, Ag

Alkali metals K, Na, Rb
Alkaline-earth metals Mg, Ca, Sr, Ba
Transition metals Fe, Co, Ni, Mn, Cu. Mo, Cr, V
Rare-earth metals La, Ce, Pr, Nb, Gd, Er
Other metals Sn, In, Ga, Zn, Ti, Al, Bi, W
Metalloids B, Si, Ge, Sb
Nonmetals P
Halogens F
Table 5.3. Influence of additives and annealing temperature on SnO2 grain size
AVERAGE GRAIN SIZE (nm)

MATERIAL 550°C 900°C 1100°C
SnO2 (undoped) 12.7 44.7 158.7
SnO2:Ce 11.7 16.9 70.3
SnO2:La 6.2 8.7 50.4
SnO2:Y 5.2 11.1 47.3
Source: Data from Carreno et al. 2002.
3.3. INFLUENCE OF ADDITIVES ON STRUCTURAL PROPERTIES OF

MULTICOMPONENT METAL OXIDES
Incorporation of a second phase, even in small quantity, can change the conditions of base oxide
growth. For example, doping of SnO2 by In, Sn, and Nb (0.1–4 mol%) caused a decrease of crystallite
size from 220 nm for pure SnO2 to about 30 nm for Nb (0.1 mol%)–doped samples (temperature of
calcination was 900°C) (Szezuka et al. 2001). Doping of SnO2 by Ce, Y, and La (Carreno et al. 2002)
and by Ca and K (Choi and Lee 2001) led to the same effect, i.e., a reduction in grain size (see Tables
5.3 and 5.4). This means that Ca, Ce, La, and Y are very effective inhibitors of SnO2 crystallite growth
above 350°C.
The same effect was observed by Ivanovskaya et al. (2001a, 2001b) for In2O3 and by Ferroni et al.
(2001) for TiO2. Their results are presented in Tables 5.4 and 5.5. The In2O3 and TiO2 were prepared
by sol-gel technology, and the In2O3 and TiO2 grain sizes were determined by x-ray diffraction (XRD)
after calcination at 600°C.
Kawamura et al. (2001) established that the presence of Cr3+ accelerates the SnO2 growth rate in
the (111) direction and suppresses it in the (110) direction. Estimates made by Kawamura et al. (2001)
Table 5.4. Influence of additives on grain size of In2O3 and SnO2 metal oxides
prepared using sol-gel method
AVERAGE GRAIN SIZE

MATERIAL (nm) REFERENCE
In2O3 (undoped) 20–25 Ivanovskaya et al. 2001
In2O3-Mo 15–20
In2O3-Ni 10–15
SnO2 (undoped) 16 Choi and Lee 2001
SnO2:Mg 58.5
SnO2:Ca 9.7
SnO2:K 14.2
Table 5.5. Effects of additives and annealing temperature on TiO2 grain growth
AVERAGE GRAIN SIZE (nm)

MATERIAL 650°C 850°C 1050°C
TiO2 (undoped) 16 110 214
TiO2:Nb 24 35 109
TiO2:Ta 18 45 63
TiO2:V 183 397 2012
Source: Data from Ferroni et al. 2001.
indicated that when trivalent cations are adsorbed on the site of Sn4+, the oxygen in the neighborhood
of the Sn4+ site is destabilized, which results in desorption of oxygen from the matrix to maintain elec-
trically neutrality. Desorption of oxygen on the S (stepped) face was identical to transformation of an
S face to a rough K (kinked) face.
It is known that the sintering rate of ceramic materials is usually controlled by the grain boundary
transport kinetics. This means that incorporation of a second phase, through the properties of grain
boundary phases and precipitates, often have a controlling influence on the mechanism and kinetics
of sintering and the final properties of the ceramic materials (Dufour and Nowotny 1988). The intro-
duction of volatile impurities causes a process which improves the porosity of the gas-sensing matrix.
Alkali and alkaline-earth metals are known to be electron exchange promoters, presumably because
of the higher partial charge on the neighboring oxygen, whereas transition-metal promoters favor the
formation of active sites (Morrison 1986). The latter is connected with a possibility for cations in
transition-metal oxides to be in several valence states.
So, phase modification opens up exciting additional possibilities for varying the structure, electro-
physical, and gas-sensing properties of metal oxide materials. Experimental data on the influence of in-
dividual additives on the gas-sensing properties of various metal oxides may be found in various reviews
dedicated to both chemical and solid-state gas sensors (Yamazoe et al. 1983; Moseley and Tofield 1987;
McAleer et al. 1988; Madou and Morrison 1989; Gopel et al. 1991; Yamazoe and Miura 1992; Gopel
and Scherbaum 1995; Meixner and Lampe 1996).
Usually, the second components of the nanocomposite are selected by considering parameters
such as ability to form oxides of different oxidation states, the absence of chemical reactions resulting
in the formation of ternary metal oxide compounds up to 500°C, catalytic activity, volatility; value and
type of conductivity, and their limited mutual solubility with the base metal oxide. This last is especially
important for structural engineering, because this parameter determines the phase state of additives to
the base metal oxide, i.e., the possibility of clusters forming on the surface of the base oxide at certain
concentrations of doping impurity. According to one estimate, the limited solubility of many oxides in
SnO2 and In2O3 does not exceed 1–2 wt% (Nowotny 1988). For some oxides, such as Ga2O3 in In2O3
(Ratko et al. 1998, and In2O3 in SnO2 (Szezuka et al. 2001), one can consider that limited solubility is at
the level of ~10%.
The surface segregation process that takes place during bulk doping of metal oxides leads to en-
richment of interfaces (free surfaces, grain boundaries, dislocations, etc.) with species coming from the
bulk phase (Nowotny 1988). In contrast to metals, in which segregation leads to changes in composi-
tion over a distance of about 1 nm, the composition gradient produced by segregation in metal oxides
involves many lattice layers. Research on such a two-phase systems, in which the concentration of the
second oxide phase is much less than the base oxide concentration, has shown that the second phase,
as a rule, is finely dispersed and is being on the surface of the base oxide’s grains (Szezuka et al. 2001;
Pagnier et al. 2001; Carreno et al. 2002). Possible versions of segregation layers of foreign cations on
the surface of SnO2 grains are shown in Figure 5.17 (Varela et al. 1999; Carreno et al. 2002). It is neces-
sary to note that the rate of surface segregation of doping elemets is controlled by lattice diffusion, and
therefore segregation equilibrium may be established at elevated temperatures.
It is important that, according to experimental data (Nowotny 1988), the near-surface solubility of
doping elements is not, as a rule, the same as the bulk solubility. For example, the surface solubility limit
of Cr in CoO is higher than the bulk solubility limit by a factor of 10 or more. A difference between the
bulk and the near-surface solubility has also been reported for Ce in NiO. Cerium can only be dissolved
within the grain boundary region, while its bulk solubility is negligibly low. This effect has a strong im-
pact on interpretation of phase diagrams for very-fine-grain ceramics. It means that the phase diagrams
for fine-grain ceramics may be entirely different from the bulk-related data available in the literature, if
those data were determined for single crystals or large-grain ceramics. It can be generally concluded that
bulk formulas of chemical compounds are not valid for the interface region, which exhibits different
composition, crystalline ordering, and resulting properties.
It has been established that the degree of segregation depends strongly on the composition of
the bulk phase (Barret and Dufour 1985; Dufour and Nowotny 1988). Experiments with the Cr–CoO
system have shown that the coefficient of Cr solubility in the near-surface layer of Cr-doped CoO
increases strongly in comparison with the bulk concentration. This effect has a severe practical result:
Even impurities that are present in the bulk of grains at a negligible level may segregate at the surface
Y-SnO2
Ce-SnO2 In-SnO2
La-SnO2
CeO2 MexSnOy
SnO2
Figure 5.17. Formation of segregation layer on the surface of SnO2 grains as dependent on the
nature of the doping elements. (Reprinted with permission from Korotcenkov 2005c. Copyright
2005 Elsevier.)
of grains and reach a high concentration in the surface layer of a crystal. High concentrations of Al and
Ca in the surface layer of Y-stabilized zirconia may serve as an example of this effect.
Research has also shown that the degree of nonstoichiometry and resulting concentrations of
defects both in the bulk and within the near-surface (grain boundary) region is controlled by the com-
position of the surrounding gas. The near-surface segregation profile of intrinsic defects has a strong
effect on the segregation of extrinsic defects. This phenomenon was observed for CoO–Cr, NiO–Cr,
and other systems. The strong effect of oxygen activity in the gas phase on the segregation profile
of doping elements in metal oxides has been demonstrated (Barret and Dufour 1985; Dufour and
Nowotny 1988).
The formation of bi-dimensional structures, i.e., structures with a new surface phase different
from that of bulk material, is another important aspect of surface segregation, which must be taken into
account in the design of multicomponent metal oxides for chemical sensors.
It has been established that the thermal stability of nanocomposites may be significantly higher
than that of the homogeneous metal oxides (see Figure 5.18). The appearance of additional intergrain
boundaries between adjacent grains of Me1O and Me2O inhibits cation interdiffusion and crystallite
growth in spite of the surface energy stored at the interface. For example, Yang et al. (2000) showed that
NiO, MoO3, CuO, and Cr2O3 oxides dispersed at the surface of SnO2 can retard both the decrease in
the specific surface area of the samples and the increase in crystallite size of SnO2 during calcination at
Tan = 500°C. However, when choosing additives for film structure stabilization, it is necessary to know
that the optimizing influence of additives does not become so evident with an increase in annealing
temperature. Panchapakesan et al. (2001) established that additives with low melting temperature, such
as CuO, upon high-temperature treatment (Tan > 800°C) could, on the contrary, stimulate the growth
of the SnO2 grains. The same regularity was observed for other metal oxides. For example, Ferroni et al.
(2001) established that additives of V stimulated the growth of TiO2 grains as well (see Table 5.5).
SnO2:Co SnO2
SnO2:La
SnO2:Ba
Figure 5.18. Influence of SnO2 doping on grain size during thermal treatment. (Adapted with permis-
sion from Yamazoe 1991. Copyright 1991 Elsevier.)
3.4. GAS-SENSING PROPERTIES OF MULTICOMPONENT

METAL OXIDES
Results of attempts at structural engineering of metal oxides intended for gas sensor applications
through bulk doping are often unpredictable. Depending on both the concentrations of additives and
crystallization modes, during bulk doping one can observe effects such as a simultaneous structure
modification accompanied by a change in grain size (see Tables 5.3, 5.4, and 5.5), the appearance of
second phase in the base oxide, i.e., heterostructures, and a change in the electrophysical and surface
properties of the metal oxides (Yamazoe et al. 1983; Cabot et al. 2001, 2002; Ivanovskaya et al. 2001).
If this complex influence on metal oxide properties is not taken into account, one can easily reach a
false conclusion. Therefore, in trying to determine the influence of a mechanism of doping on the gas-
sensing characteristics of metal oxides, one needs to take into account the changes in all the parameters
of the controlled films. Only in this case will it be possible to understand that the changes that are so
often observed in gas-sensing characteristics are a consequence of either a change of grain size (t ) or
of the concentration of free charge carriers (Nd ) (Fliegel et al. 1994): That is, observed modifications
are not conditioned on the appearance of additional active catalytic centers. For example, according to
Szezuka et al. (2001), SnO2 doping by Nb and Sb in the range 0.01–1.0 mol% during sol-gel preparation
with subsequent annealing at T = 900°C led to a drop of film resistance of 102 and 104 times, respec-
tively, while doping by In was accompanied by a rise of film resistance by a factor of approximately
100. The same result was observed for SnO2:Mo (Ivanovskaya et al. 2001). Ivanovskaya and co-workers
(2001) established that SnO2 doping by Mo increases the resistance of SnO2-based ceramics more than
103–105 times. The considerable growth of SnO2 resistance was observed after doping by Mn (especially
at concentration > 0.1 mol%) as well (Fliegel et al. 1994).
Under certain conditions, the influence of the indicated factors (t and Nd ) may be much stronger
than the influence of catalyst particles. Results of theoretical simulations, presented in Figure 5.19 for
SnO2-based thin-film gas sensors, show how strong this effect may be. It can be seen that an Nd decrease
below 1017 cm−3 promotes considerable growth of sensor response. At such concentrations of free

SnO2
104
103
102
1
Figure 5.19. Results of theoretical simulation of
10 film thickness influence on the high limit of SnO2
gas response for different concentrations of free
100 charge carries. Nd : 1, 1 × 1017 cm−3; 2, 3 × 1017
cm−3; 3, 1 × 1018 cm−3; 4, 3 × 1018 cm−3; 5, 1 × 1019
cm−3. (Reprinted with permission from Brinzari et al.
1999a. Copyright 1999 SPIE.)
charge carriers, even films with grain size exceeding 60 nm become gas-sensing ones. Earlier, Yamazoe
and co-workers (1991) obtained similar results for ceramic metal oxide gas sensors (see Figure 5.20).
During the fabrication of ceramic and thick-film gas sensors, the problem of additives influencing
grain size and bulk concentration of free charge carriers can be resolved by blending in additional com-
ponents in a final stage of either ceramic or paste preparation. For example, this process may include
mechanical mixing of components, such as SnO2 and Al2O3 powders (Ihokura and Watson 1994). In
this case there is small probability that the properties of SnO2 will be determined by incorporation of
doping impurities (Al3+) into the tin dioxide lattice. In such a system, Al2O3 added to tin dioxide prob-
ably plays the role of diluent, which only separates SnO2 crystallites, preventing their sintering. As a
result, both stability and sensitivity of sensors may be increased.
In thin-film technology it is impossible to avoid the mutual influence of additives and base oxide
during deposition. This, of course, presents problems in selecting an optimal composition for the gas-
sensing matrix. However, with careful selection of additives, the effects mentioned above can increase
the porosity of the nanostructured gas-sensing matrix, stabilize the crystal structure, modify bulk and
surface properties, decrease the influence of the ambient atmosphere, improve sensitivity and selectiv-
ity, and contribute to long-term maintenance of activity.
It is necessary to note that the mechanism of gas sensitivity for the nanocomposites Me1O–Me2O
is much more complicated than that for homogeneous oxides and has special features. According to
Gaskov and Rumyantseva (2001), the solid–gas interaction in such systems involves selective chemical
reactions at the grain boundaries with participation of the chemisorbed molecules and lattice oxygen.
When Me2 is a transition element (d element), these reactions can change the Me2 oxidation state, co-
ordination number, and cation distribution between the surface and lattice positions. The reversible
variations of Me2 oxidation state and position in the presence of specific gas molecules can result in a
sharp change in the electronic state of the grain boundaries and the transport properties of the nano-
crystalline materials. Thus, the grain boundary state, which is responsible for the electrical properties

Nd(SnO2:Sb)>Nd(SnO2)>Nd(SnO2:Al
800 ppm H
SnO2 SnO2:Al
SnO2:Sb
Figure 5.20. Influence of grain size of doped and undoped SnO2 on gas response to H2. Toper =
300°C. (Adapted with permission from Yamazoe 1991. Copyright 1991 Elsevier.)

S( H 2 ) S(CH 4 )

S(CH 4 ) S( H 2 )
S( H 2 ) S(CH 4 )

S(CO) S(CO)
Figure 5.21. Influence of In2O3 doping on gas-response selectivity of one-electrode gas sensors
relative to H2, CO, and CH4 detection. (Reprinted with permission from Korotcenkov et al. 2004e.
of nanocomposites, is self-adjusting, depending on the reduction–oxidation (redox) properties of the

atmosphere as well as the crystal/chemical properties of Me2 cations, such as their electronic configura-
tion and ionic radius.
Some results from study of gas-sensing properties of both doped In2O3 ceramics prepared by the
sol-gel method and doped SnO2 films deposited by spray pyrolysis are presented in Figures 5.21 and
5.22. The In2O3 sensors were fabricated as one-electrode variants (Korotcenkov et al. 2004e), while the
SnO2 sensors were fabricated as standard thin-film gas sensors (Korotcenkov et al. 1999, 2001). One
can see that the doping of In2O3 ceramics and SnO2 films can significantly improve both the sensitiv-
ity and selectivity of gas sensors. From these results, some important conclusions have been drawn, as
discussed in the following paragraphs.
1. The effect of doping elements on the gas-sensing properties of metal-oxide gas sensors does not always correspond
to the catalytic activity of these additives. This is especially clear for the transition metals (Cr, Mn, Fe, Ni, Co,
Cu, Zn, and Ga), which are widely used in catalysis (see Figure 5.23). For example, an improvement of
conductivity response of In2O3-based sensors containing a second oxide at 2–10 wt% was observed for
elements with minimal catalytic activity.
In principle, this effect is not unexpected, because gas sensitivity and catalytic activity are results of
entirely different mechanisms, even though both are surface reactions. If the result of catalysis is a new
chemical substance which is generated as a result of reactions taking place at the surface of the metal

SnO2:Me 105 SnO2:Me
104
103
102
1 ppm O3
1000 ppm H2
101
(a) (b)
Figure 5.22. Influence of SnO2 doping by Co, Ni, Cu, and Fe during film deposition by spray pyrolysis
on gas response to (a) H2 and (b) ozone. Tpyr = 410–420°C; d ≈ 45–55 nm; 1, SnO2:Fe; 2, SnO2:Co;
3, SnO2:Cu. (Reprinted with permission from Korotcenkov 2005c. Copyright 2005 Elsevier.)
oxide, the sensor response is a change of resistance of the oxide matrix (Morrison 1986). Moreover,
surface catalysis can occur with a conduction band of the oxide matrix, without electron exchange,
while electron exchange is obligatory for the appearance of gas sensitivity. That is why high catalytic
activity is not always accompanied by high sensor response. Probably, this situation is realized for the
gas sensitivity of the In2O3 ceramics mentioned above, which had been doped with transition metals.
Figure 5.23. Comparison of gas response of one-electrode sensor based on doped In2O3 ceramics
and catalytic activity of transition metals used as additives for In2O3 modification. (Reprinted with
permission from Korotcenkov et al. 2004e. Copyright 2004 Elsevier.)
2. The optimal concentration of catalytically active additives to attain maximum conductivity response of SnO2-
and In2O3-based gas sensors lies in the range <1–3 wt%. As an example, some results of film doping by Co,
Cu, and Fe on the response to both ozone and CO of SnO2 sensors are presented in Figure 5.19. SnO2
films were doped during the process of spray pyrolysis deposition.These results, and others (Xingqin
et al. 1993), show that superfluous concentration of these additives, as a rule, sharply reduces the ef-
fectiveness of the modification. There is a hypothesis that, for maximum effect, catalytically active
additives should have atomic or finely dispersed distribution in the gas-sensing matrix. This conclusion
was drawn by Yamaura et al. (2000), who showed that an increase of concentration of Со in an In2O3
matrix from 0.5 wt% to 5.0 wt% was accompanied by a drop of gas response of more that 20 times
(see Figure 5.24).
For sensors in which both “acceptor” and “transducer” functions are realized by added oxides, the
concentration of the second phase can attain 5 wt% or more. This is realized, for example, in H2S sen-
sors based on a SnO2-CuO system whose function is based on a change of chemical composition of the
second oxide (CuO → CuS) during interaction with H2S (Pagnier et al. 2001). The mechanism of this
gas sensitivity is shown in Figure 5.25. In these sensors, CuO is the main gas-sensing element, because
its interaction with the surrounding atmosphere determines the behavior of the gas sensor. The SnO2
crystallites act as a stable conductive matrix, guaranteeing the necessary porosity of the sensing struc-
ture and the possibility to form fine, dispersed, and isolated CuO grains on their surface. However, it
is necessary to remember that CuS is unstable; the crystal structure of CuS is changeable. For instance,
at T ≈ 103°C, CuS transforms to Cu2S, which is an ionic conductor with higher resistance, especially
In2O3:Co
In2O3
In2O3 In2O3:Co
Figure 5.24. Sensitivity to 1000 ppm CO in wet air for In2O3-based sensors loaded with various ad-
ditives (Toper = 300°C). (Adapted with permission from Yamaura et al. 1996, 2000. Copyright 1996,
2000 Elsevier.)

Air Air + H2S
SnO2 CuO SnO2 SnO2 CuS SnO2
CuO
p-n junction
R(p-n) Rg(CuO) Rg(SnO2) Rg(SnO2) Rg(CuS)
R(total) ~ R(p-n) + Rg(CuO) R(total) ~ Rg(CuS)
Rg(CuO)>>Rg(CuS)
Figure 5.25. Mechanism of SnO2-CuO–based sensor operation.
at temperatures > 220°C (Zhou et al. 2003). Therefore, the sensitivity of CuO-SnO2 sensors becomes
lower with increasing operating temperature.
According to some research, the worsening of gas-sensing characteristics of sensors with high
concentrations of a second oxide can be related to the fact that at certain concentrations, exceeding the
limited mutual solubility, the second phase—with less gas sensitivity—starts determining the electro-
conductivity of the gas-sensing matrix. This worsening can also be related to the considerable disorder
of the base oxide. In such oxides, the density of surface states may increase sharply, which may lead,
according to theoretical estimations, to stabilization of the surface Fermi-level position and, therefore,
to a decrease in sensor response (Brinzari et al. 2000a, 2001).
The appearance of additional surface states in the SnO2 band gap during the doping process
was observed experimentally both on x-ray photoelectron spectra and on optical absorption spectra.
The same results were presented earlier by Cabot et al. (2001). By analyzing transmission spectra of
SnO2 films modified by Cu, Co, Ni, and Fe, it was established that dispersion of edge absorption was
already taking place for additives at the level of 1–4 wt%. Korotcenkov (2005c) believes that this effect
is connected with the appearance of tails of density-state distribution in the SnO2 band gap (Figure
5.26). It was established that this effect became especially noticeable with SnO2 doping by Cu and Fe.
Apparently, surface stoichiometry worsening is the reason for the relative intensity drop of the surface
mode in SnO2 Raman spectra, observed by Pagnier et al. (2001), at Cu doping of more than 1 wt%.
3. Doping may be accompanied by the formation of new compounds with fundamentally different physical-chemical
properties. In experiments discussed by Korotcenkov et al. (2004e), such an effect was seen most clearly
for In2O3 doping by phosphorus. Raman spectroscopy of In2O3:P2O5 ceramics showed the formation
in this system of such compounds as InPO4. As a result, doping of In2O3 even by such a hygroscopic
SnO2:Cu
Undoped SnO2
Figure 5.26. Transmission spectra of SnO2 films doped by Cu. Tpyr = 410°C; d ≈ 300 nm; 1, undoped
film; 2, 1 at%; 3, 4 at%; 4, 16 at%. (Reprinted with permission from Korotcenkov 2005c. Copyright
2005 Elsevier.)
additive as P2O5, did not lead to a strengthening of the influence of air humidity on the characteristics
of In2O3:P2O5-based gas sensors. It necessary to note that such processes also take place in multicom-
ponent SnO2-based materials. Formation of new phases has been observed for SnO2 doped with Ba
(BaSnO3), Y (Sn2Y2O7), La (Sn2La2O7), Ca (CaSnO3), etc. These effects are discussed by Azad and Hon
(1998) and Carreno et al. (2002).
Changes that may take place in a two-phase metal oxide matrix while a concentration of the second
phase is being increased are demonstrated in Figure 5.27. Here, for simplification, Rf.ph and Rs.ph are the
grains’ resistance to the first and second oxide phases, including resistance of both bulk and intergrain
contacts. One can see that at low concentration of additive, the second phase only modifies the surface
Rf.ph Rs.ph
Me1Ox
Me2Ox
Figure 5.27. Change of film structure and electrical scheme of film conductivity during phase modifica-
tion of metal oxides. (Reprinted with permission from Korotcenkov 2005c. Copyright 2005 Elsevier.)
properties of the base oxide. At somewhat higher concentration it can contribute to limiting the electro-
conductivity of the metal oxide matrix. And at the final stage, at a certain combination of electrocon-
ductivity and gas sensitivity for two metal oxide phases in the gas-sensing matrix, the second oxide phase
can produce either full blockage of interaction of the base oxide with the surrounding atmosphere, or
shunting of the matrix of the base oxide through a more conductive second metal oxide phase. All this
will certainly lead to a significant change of both the electrophysical and gas-sensing properties of the
metal oxide matrix. In this case, these properties will not be determined by the base oxide.
After analyzing results connected with the study of gas-sensing properties of SnO2 modified by
Ni, Fe, Co, and Cu, it was concluded that at excess concentration of the second oxide phase, the wors-
ening of gas-sensing characteristics may also be conditioned by formation of additional p-n heterojunc-
tions in the gas-sensing matrix. This conclusion was made on the basis of the following considerations
(Morrison 1986; Williams 1999):
a. Metal oxides formed by transition metals from the end of Period IV in the lower oxidation state,
i.e., with a deficiency of oxygen, exhibit usually p-type conductivity.
b. The adsorption of oxygen with formation of acceptor-like species on p-type semiconductors is
difficult due to the lack of conduction electrons.
c. Charge transfer in transition metal oxides, where the number of d electrons exceeds 3, is not de-
scribed by the zone model. In this case we have to take into account the Geitler-London model
of localized electrons.
d. The sign of conductivity response to oxidizing and reducing gases for p-type oxides is different
than that with n-type oxides (SnO2, In2O3), i.e., such changes may compensate each other.
The above results are summarized in Table 5.6, and leave us skeptical about wide use of mixed p-,
n-type oxide matrixes for standard solid-state chemisorption-type sensor design.
4. SENSOR RESPONSE CONTROL THROUGH SURFACE

MODIFICATION OF METAL OXIDES
The gas-sensing effect is a surface phenomenon, in that changes in the concentration of conduction
electrons in metal oxides result from surface chemical reactions, such as chemisorption, reduction/
reoxidation (“redox”), and/or catalysis, which take place during interaction with surrounding gas. Both
experimental and theoretical studies have demonstrated that various surface sites participate in combus-
tion, electrical response, and interactions with water vapor and adsorbed species. Different gases inter-
act differently with different surface sites as well. For example, as an explanation of SnO2 gas response
to methane, Williams and Pratt (1998) proposed that four different types of reactive surface sites may
participate in gas-sensing effects on polycrystalline SnO2: the electrically neutral oxygen, which cata-
lyzes methane; two electrically charged oxygen states, which respectively mediate methane response and
dissociative water chemisorption; and a site where molecular chemisorption of water occurs. Williams
and Pratt (1998) believed that the neutral site could be attributed to a coordinately unsaturated lattice
oxygen. The two different electrically charged sites could be associated with oxygen adsorbed on or at
Table 5.6. Influence of bulk additives on gas-sensing characteristics of SnO2 and

In2O3 sensors
ADDITIVE EFFECT NATURE

Noble metals (<5 wt%) Increased response to reducing gases Catalytic effect
(Pd, Pt, Rh, Ag, Au) Decreased operating temperature Change in A/D parameters
Decreased response time (decreased O2 dissociation
temperature)
Al2O3, SiO2 Increased sensitivity Decreased grain size
Improved thermal stability Decreased area of intergrain
contacts
Increased porosity
Ag (Ag2O), Cu (Cu2O) Increased sensitivity to H2S, SO2 Two-phase system
Phase transformations during gas
detection
Ga(Ga2O3), Zn( ZnO) Increased gas response Decreased grain size
Increased porosity
P, B Improved selectivity Creation of new phase
Ca, K, Rb, Mg Increased sensitivity Decreased grain size
Improved thermal stability
La, Ba, Y, Ce Improved thermal stability Stabilization of grain size
(forming new phases)
Decreased grain size
Transition-metal oxides Increased sensor response Catalytic effect
(<0.5 wt%) Improved selectivity Change of electron concentration
(Co, Mn, Sr, Ni, Fe) Change in A/D parameters
Change of grain size
Fe (Fe2O3) Increased sensitivity to alcohol Change in oxidation state
two different types of surface vacancy associated with reduced (Sn2+) surface cations. All these surface
sites are related directly to the surface stoichiometry of metal oxides, and thus it is clear that surface
modification, i.e., surface engineering, is a powerful tool for control of sensor response. An explanation
of the influence of this surface stoichiometry on gas-sensing characteristics of metal oxides was also
given earlier by Brinzari et al. (1999).
4.1. METHODS OF SURFACE MODIFICATION

Any method of influencing the surface state can be used to modify the metal oxide surface, including
(Wada and Egashira 1998; Kiv et al. 2001; Kim et al. 2002; Zhang et al. 2003):
• Oxygen plasma treatment

• UV irradiation
• Treatment with various chemical reagents
• Low-energy ion irradiation
and so on. As a result of such treatments, the surface stoichiometry, work function, binding energy
of atoms, concentration of carbon contaminants, chemical activity of surface, and so on, can all be
changed. However, in spite of the availability of numerous methods by which to influence the surface
state, surface modification of metal oxides using noble and transition metals remains the most-often-
used method.
Introducing some foreign atoms at the surface is a classical way to change the surface scenario
and obtain a response with unusual kinetics, enhanced stability toward aging, and higher sensitivity. To
achieve that goal one can use any deposition method, as described earlier in this volume, from thermal
vacuum evaporation to chemical vapor deposition and electroless deposition.
The choice of method will be determined by the objects being subjected to surface modification,
the materials used, and the requirements for surface additives. For example, Diaz et al. (2002) estab-
lished that for surface modification of powders by noble metals, the most appropriate technique is to
use both impregnation and an electroless method able to provide uniform covering of the powder from
every side. Precursors which can be used for these purposes are presented in Table 5.7.
It is necessary to underline that, independent of the method used for surface modification (impreg-
nation, electroless, vacuum evaporation, spraying, etc.), after deposition of noble metals it is necessary
to carry out a subsequent annealing in the temperature range 300–600°С (Zhang and Colbow 1997).
This annealing promotes the formation of metallic clusters, improves the homogeneity of their distri-
bution across the layer thickness, and stabilizes the properties of the gas-sensing matrix. As has been
Table 5.7. Precursors for surface modification by selected metals and their oxides
ADDITIVE PRECURSORS
Au HAuCl4 4H2O
Ag (Ag2O) AgNO3; Ag(acac)
Pd (PdO) PdCl2; Pd(acac)2; Pd(NH3)4; Pd(NO3)2
Pt PtCl2; H2PtCl6 6H2O; (CH3)3(CH3C5H4)Pt;
(NH4)2PtCl4
Cu (CuO) Cu(NO3)2; Cu(acac)2
Ni (NiO) NiCl2; Ni(NO3)2
Rh (Rh2O3) RhCl3 3H2O
Ru (RuO) RuCl3
Co (Co2O3) Co(acac)2
Fe (Fe2O3) FeCl3; Fe(NO3)3
Si (SiO2) Diethoxydimethylsilane (DMES)
Al2O3 Al(acac)
noted in (McAleer et al. 1988; Cabot et al. 2000), during this final stage of preparation of the gas-sens-
ing matrix, the temperature is raised only to a value sufficient to decompose the salt used to impregnate
the grains, and hence the additive should be very finely divided. This is important for achieving high
sensor response. High-temperature treatment sufficient to cause agglomeration of catalyst particles
resulted in the loss of optimizing effect.
The nature of the noble metals, their electronic state (either oxidized or not), and their distribution
both over the oxide surface and in the film are determining factors in promoting gas-sensor sensitivity
and selectivity. However, it is necessary to acknowledge that when using physical methods to deposit
noble metals, it is very difficult to attain a homogeneous distribution through the film thickness. For
that is required either too high temperature treatment, or too long annealing. In using metal oxides for
surface modification, this task becomes even more complicated.
For example, a profile of Pd distribution through SnO2 film thickness (d ≈ 40 nm) after annealing
at 500°C (tan = 15 min) is shown in Figure 5.28. The profiles in this figure were obtained using Auger
spectroscopy with Ar sputtering (Korotcenkov et al. 2003c). Studied SnO2:Pd films were formed using
separate deposition of SnO2 and Pd layers by spray pyrolysis. One can see that the maximum of Pd
concentration is observed at the surface of the SnO2 film. This means that surface modification is ef-
fective only for ultrathin and thin films. As film thickness increases, the role of added surface catalysts
in gas-sensing effects is sharply reduced. In practice, the problem of improving noble metal distribution
in the film thickness is often resolved by using multiple, layer-by-layer depositions of metal oxide and
noble metal. Such an approach was used successfully (Holody et al. 2001; Kim et al. 2001) to form a
Pt/SnO2 gas-sensing matrix with a thickness of ~400 nm.
d ~ 40 nm
SnO2 Si
SnO2
Figure 5.28. Typical depth profile distribution of Pd and C in SnO2 film (d ≈ 40 nm) deposited on Si
substrate. Pd was deposited by spray pyrolysis. (Adapted with permission from Korotcenkov et al.
Regarding ion implantation for optimization of gas-sensing characteristics, it is an interesting mod-

ification method, but it does not find wide use—possibly because it requires investment in expensive
equipment. Designers may also be concerned about generating a high concentration of radiation de-
fects, which with long-term use may become unstable.
4.2. INFLUENCE OF SURFACE MODIFICATION ON GAS-SENSING

PROPERTIES OF METAL OXIDES
Nanoscaled particles of noble metals (Pd, Rh, Pt, Au, Ag), and oxides of other elements (Co, Fe, Zn,
B, Se, Cu, etc.) that are deposited on the surface of metal oxides can act as surface sites for adsorbates,
as promoters for surface catalysis, or as elements promoting improvement in the thermal stability of
the film nanostructure. These additives can also act as inhibitors or activators of the surface reactions.
They can provide adsorption sites and surface electronic states which can mediate the electronic trans-
fer processes (Morrison 1986; Willams and Pratt 1998; Barsan et al. 1999; Williams 1999). As a result,
parameters of gas sensors such as gas sensitivity, rate of response, and selectivity have been dramatically
improved through surface modifications (McAleer et al. 1988; Cabot et al. 2000; Brinzari et al. 2002c;
Korotcenkov et al. 2003d, 2003c, 2004a) (see Figures 5.29 and 5.30).
A shift of sensor response maximum in the range of low operating temperatures during surface
modification by noble metals takes place as well (Figure 5.31). For example, in experiments with SnO2
films modified with Pd, this shift reached 100–200°С (Korotcenkov et al. 2001a, 2003c; Brinzari et al.
2002c). Table 5.8 summarizes the effects of surface modification on sensor response of metal oxide–
based devices.

103
SnO2
102
101
Figure 5.29. Influence of SnO2 surface modification by Pd on 1. magnitude, and 2, time constants of
gas response to 2000 ppm H2. RH = 50%; Toper = 215°C. (Reprinted with permission from Korotcenkov
et al. 2003c. Copyright 2003 Elsevier.)

SnO2 104
103
102
101
Figure 5.30. Influence of SnO2 surface modification with Ag and Pd on gas response to CO and
ozone. Ag and Pd particles were deposited by SILD. (Adapted with permission from Korotcenkov et
al. 2003d. Copyright 2003 MRS.)
It is important that, depending on the metal deposited, the loading method, and the interacting
gas, surface clusters may be in metallic or oxidized form, which have different catalytic activities and
therefore different effects on sensing charactersitics (Kohl 1990). At present, two types of sensitiza-
tion mechanisms—electronic and chemical—are the most accepted ones. Chemical sensitization takes
place via a spilover effect, which is now rather familiar in catalytic chemistry. The promoter in this case
activates a test gas to facilitate its catalytic oxidation on the semiconductor surface. In other words, the
promoter does not affect the resistance of the element directly, leaving the gas-sensing mechanism
essentially the same as in the case without it. The promoter increases the gas sensitivity, as it inceases
the rate of the chemical processes, leading to a decrease in the concentration of the negatively charged
adsorbed oxygen. On the other hand, electronic sensitization is due to direct electronic interaction
between the metal additive and the semiconductor surface. In this case there is no mass transfer be-
Figure 5.31. Temperature dependence of gas

response of SnO2 films with surface modification
by Pd. Pd was deposited by spray pyrolysis. 1,
initial SnO2; 2, 0.6% ML; 3, 1.1% ML. (Reprinted
from Korotcenkov et al. 2003c. Copyright 2003
Elsevier.)
Table 5.8. Influence of surface modification of metal oxides (SnO2) on sensitivity

and selectivity of gas response
IMPROVED SENSOR RESPONSE

AND SELECTIVITY TO THE
ADDITIVE FOLLOWING GASES AND VAPORS REFERENCE
Ru Hydrocarbons, hydrogen Chaudhary et al. 1998
Ag H2, H2S Tong et al. 2000
Fe (Fe2O3) Alcohol Ivanovskaya 2000
La (La2O3) Ethanol Yamazoe 1991
La/Pd Ethanol Yamazoe 1991
Pd CO, H2, CH4 Cabot et al. 2002
Pt CO, H2, CH4 Cabot et al. 2002
Cu (CuO) H2S Yamazoe 1991
Ru/Pd H2 Chaudhary et al. 1998
Co CO Yamaura et al. 2000
La/Au CO
SiO2/Pt NO Benard et al. 2005
Co/Au CO Yamaura et al. 2000
Co; Fe O3 Takada et al. 1993
tween the foreign particles and the semiconductor. Instead, there is an electronic interaction between
them, through the space charge created in the semiconductor by the presence of the surface clusters.
Additives at the surface of the metal oxides act as “receptors,” while the semiconductor oxide acts as a
“transducer” of the changes taking place at the surface under gas adsorption. This type of sentization
has been proposed for SnO2 elements modified with Ag, Pd, and Cu, which form stable metal oxides
when they are exposed to air (Yamazoe et al. 1983). The chemical mechanism of sentization usually is
applied to explain the behavior of metal oxides (SnO2) modified with Pt (Kohl 1990).
Although the above mechanisms are the most accepted ones to explain most research results
(Matsushima et al. 1988; McAleer et al. 1988; Brinzari et al. 2002c), some authors propose other mecha-
nisms to explain their results. For example, it has been claimed that improved response to gas with
noble-metal loading can arise mainly through the interaction of the gas molecules with electron states
introduced near the valence band edge by the additives (Cabot et al. 2000), i.e., via control of the popu-
lation of such states and the consequent change of the band bending. Actually, all the above mecha-
nisms are present at the same time. However, several factors determine their activity, and one mecha-
nism may be dominant in some cases. For example, it was found that electronic interactions actually
appear between Ag, Pd, and Cu in oxide form and SnO2, and disapear when these oxides are converted
to metals. In addition, it was shown that for a given additive, its distribution through the support and
loading also have important effects on the device response, in addition to that arising from the nature
of the catalytic material itself.
When choosing additives it is also necessary to remember that high catalytic activity of additives
is an essential, but not sufficient, requirement. To achieve high sensor response, the method of surface
modification should create optimal conditions for both electron and ion (spillover) exchange between
the surface nanoclusters and the metal oxide support (Figure 5.32). Only in this case will catalytic reac-
tion on the surface of the additives be accompanied by a change in the film electroconductivity, which
controlled in metal oxide films during gas detection. This last is necessary for effective influence of
surface nanoclusters on the gas response of solid-state nanostructure gas sensors. Research (Tsud et
al. 2001; Skala et al. 2003; Nehasil et al. 2003) has shown that in the system of either Pd(Rh)-SnO2 or
Pd(Rh)-In2O3, the formation of alloys of the noble metal-Sn(In) may promote an improvement of
both electron and ion exchange between the nanoclusters of the noble metal and the metal oxide sup-
port. For example, a decrease of CO desorption energy from the SnO2:Rh surface after Rh-Sn alloy for-
mation has been observed routinely (Tsud et al. 2001; Skala et al. 2003; Nehasil et al. 2003). According
to Tsud et al. (2001), specific properties of such alloys (Pd-Sn) were provided by the strong interaction
between Pd and Sn through the hybridization of Pd d and Sn s states, leading to the formation of a Pd-
Sn alloy with a noble metal–like electronic structure.
This effect is apparently general for bimetallic catalysts. For example, Rodriguez (1996), analyzing
the interaction of Co with Pd-based bimetallic catalysts, established that bimetallic surfaces have lower
CO desorption temperature than Pd (see Table 5.9).
It is important to emphasize that the use of alloys of palladium for surface modification is current-
ly the most promising direction for surface modification of sensitive materials. For example, Chaudhary
et al. (1998) have shown that alloys of Ru/Pd really are much more effective surface catalysts than the
same metals alone for attaining the required selectivity of gas sensors. It was established that, by using
alloys of Ru/Pd, it is possible to achieve a considerably greater increase of sensitivity and selectivity for
hydrogen detection with sensors based on tin oxide than with sensors modified using Ru or Pd metal
(1–2 wt%) separately (Chaudhary et al. 1999). The optimum seems to be at a ratio Ru/Pd ≈ 1.28.
One may judge the effectiveness of such modification by studying the results given in Table 5.10.
The tested ceramic sensors were fabricated using sol-gel technology. Ru and Pd were deposited on the
surface of SnO2 powders from millimolar RuCl3 and PdCl2 aqueous solutions using techniques de-
O2g O2g
O2g
OsOs
Os Os
Os
O2s O o2s OS O s Os

Os O s
Figure 5.32. Influence of Pd on adsorbed species on the SnO2 surface. (Reprinted with permission
from Korotcenkov et al. 2003c. Copyright 2003 Elsevier.)
Table 5.9. Effects of bimetallic bonding on properties of Pd
BIMETALLIC CO DESORPTION SHIFT IN Pd(3d5/2)

CATALYST TEMPERATURE (K) BINDING ENERGY (eV)
Pd(100) 480 0.00
Pd/Re 370 0.65
Pd/Ru 335 0.30
Pd/Mo 320 0.90
Pd/W 300 0.85
Pd/Cu 280 0.60
Pd/Zn 250 0.70
Pd/Ta 245 0.90
Source: Data from Benard et al. 2005.
scribed by Chaudhary et al. (1997). These techniques included dipping for 5 h, drying, and then heating
at 350°C for 2 h. The selectivity was calculated as the ratio S(H2)/S(gas).
A considerable increase of promotion effect was also observed using a Co/Au alloy. Studying the
influence of additives on gas-sensing characteristics of In2O3-based sensors, Yamaura et al. (2000) es-
tablished that a binary catalyst consisting of 0.5 wt% Co and 0.04 wt% Au gave dramatically improved
sensitivity and selectivity to CO detection (see Table 5.11).
It is necessary to note that bimetallic catalysts that contain Group 10 metals such as Pt, Pd, or Ni
are ideal catalysts for many applications: isomerization of hydrocarbons, olefin hydrogenation, CO oxi-
dation, alcohol synthesis, etc. With an overlayer of Group 10 metals, there is very good correlation be-
tween the changes in the electronic and chemical properties induced by bimetallic bonding (Rodriguez
1996). Comparing the behavior of Pt, Pd, and Ni, it was found that among the Group 10 metals, Pd
showed the strongest effects, while Ni exhibited the weakest (Ni < Pt < Pd). For these admetals, the
Table 5.10. Influence of surface modification with Ru/Pd (Ru/Pd ≈ 1.28) alloy on
sensitivity and selectivity of H2 detection by SnO2-based gas sensors
PURE SnO2 SnO2 MODIFIED WITH Ru/Pd

GASES AND VAPORS, 1000 ppm
(Toper = 250°C) SENSITIVITY SELECTIVITY (H2) SENSITIVITY SELECTIVITY (H2)
Hydrogen 7.6 1.0 1350 1.0
LPG 3.0 0.39 375 0.28
Ammonia 0.12 0.02 0.14 0.0001
Nitrogen dioxide 0.40 0.05 0.24 0.0002
Alcohol 6.5 0.86 20.50 0.015
Carbon monoxide 4.0 0.53 0.08 0.00006
Source: Data from Chaudhary et al. 1997.
Table 5.11. Gas response of In2O3-based sensors modified with Co and Co/Au
SENSOR RESPONSE
(Toper = 250°C; 1000 ppm)
MATERIAL CO H2 S(CO)/S(H2)
Undoped In2O3 7 4 1.5–2.0
In2O3:Co (0.5 wt%) 50–60 6–10 5–10
In2O3:Co/Au 160 4–6 25–40
Source: Data from Yamaura et al. 2000.
magnitude of the electronic and chemical changes increased when the fraction of empty states in the
valence band of the metal substrate rose.
However, at present, the phenomena responsible for the catalytic behavior of these bimetallic
systems are not fully understood. Rodriguez (1996) gave the following explanation of the properties of
bimetallic catalysts. A Group 10 adatom in contact with the surface of an s,p or early transition metal
exhibits large perturbations in its electronic and chemical properties. In this type of system, there is an
important redistribution of charge that shifts d electrons from around the Group 10 metal to the inter-
face region between the admetal and the substrate, producing an accumulation of electrons around the
bimetallic bonds. This redistribution of charge affects the stability of the core levels and the valence d
band of the Group 10 metal. The larger the movement of d electrons from the Group 10 metal toward
the admetal–substrate interface, the stronger is the bimetallic bond, and the lower is the ability of the
Group 10 metal to bond CO through π back-donation.
In choosing deposition modes, it is also necessary to take into account the influence of dimen-
sional effects, such as the size of surface clusters and the average distance between them. Even during
early study of semiconductor gas sensors, it was noted (Kawamura et al. 2001) that the kind of catalyst
distribution, both throughout the device and in terms of catalyst particle size, has an important influ-
ence on sensor response in addition to that arising from the nature of the catalyst material itself. For
example, it was established by McAleer et al. (1988) that, to achieve optimal effect, surface clusters
size should not exceed 1–5 nm. According to some estimates (Brinzari et al. 2002; Korotcenkov et al.
2003), an optimal distance between clusters approximately equals the oxygen surface diffusion length at
the operating temperature (Figure 5.29). Research on the Pd-SnO2 system showed that, as a rule, these
conditions are reached at an average surface concentration of noble metals of ~1% of the monolayer
(Brinzari et al. 2002; Korotcenkov et al. 2003). Apparently, at that concentration was attained an optimal
covering of spillover zones around noble metal clusters (Figure 5.33). Simple calculations showed that
at such a degree of coverage, the average distance between clusters with size from 1 to 5 nm may change
from 6 to 12 nm correspondingly. In reality, this distance may be even less. Attempts to evaluate the
size of Pd particles on the surface of SnO2 films deposited by spray pyrolysis were made by Veltruska
et al. (2001), but they were successful only for films with a coverage level over 18%. According to esti-
mates made by Veltruska et al. (2001), when the Pd grew three-dimensionally, the average size of a Pd
cluster was ~3.5 nm for ~18%Pd coverage, and ~4.0 nm for ~30%Pd coverage. Tsud et al. (2001) later
confirmed this estimate. The results of this research showed that with a small amount of deposited Pd,
SnO2
Figure 5.33. Spillover zone overlapping on Pd-

modified surface. (Reprinted with permission from
Korotcenkov et al. 2003c. Copyright 2003 Elsevier.)
equivalent to 1 monolayer (ML), Pd forms an atomically dispersed layer on the SnO2 surface, and only
higher quantities, several MLs of Pd, led to the formation of three-dimensional Pd particles, which had
mixed metallic and Pd-Sn bimetallic character.
The initial structure of modified grains should also be taken into account when evaluating the
influence of surface modification on gas-sensing properties. Recent work has provided direct experi-
mental confirmation that noble-metal clusters accumulate at step edges of metal oxides (El-Azab et al.
2002). Due to this behavior of noble metals, an increase of terrace area should be accompanied by an
increase in the size of clusters, which can be accumulated at step edges of the terrace. This means that
for the same degree of surface coverage by noble metals, the number of clusters will be smaller, while
the distance between the clusters will be greater on the surface that has larger terrace area. Illustration
of this process is presented in Figure 5.34. It follows from this that the appearance of extended, atomi-
cally flat surface planes facilitates the process of cluster formation. At the same time, the presence of
atomically stepped–like surfaces in the case of small nanograins provides with high probability the
conditions for atomic dispersion of palladium. Results obtained during SnO2 surface modification by
Pd have confirmed this conclusion. Optimal influence of Pd on gas-sensing characteristics of SnO2
films containing small grains (spherulites) was observed at considerably lower Pd concentration than for
larger crystallites with clearly visible faceting.
(b)
(a)
(c)
Figure 5.34. Influence of surface morphology on the size and distribution of Pd clusters; (a), (b), and
(c) refer to structures with different sizes of flat surfaces. (Reprinted with permission from Korotcenkov
Detailed analysis of the noble-metal impregnation of as-obtained SnO2 sol-gel carried out by Cabot
et al. (2000) confirmed many earlier conclusions. The existence of different chemical states of the noble
metals localized at the grain surface, which give rise to different sensing sites, was established. These
sites responded in a different way to different gas molecules such as CO or CH4, as well as to humidity,
showing different sensitivities at the working temperatures. Moreover, it was also found that additives
can introduce a strong distortion at the grain surface (see Figure 5.35). This lattice distortion presents a
significant difference for the palladium, whose distortion seems to be extended inside the grain.
All this means that the introduction of additives must be controlled with high accuracy in order
to fix the type and density of sites where the gas-molecule detection takes place. According to Cabot
et al. (2000), it is necessary to define not only the best noble metal to be used as catalytic additive, but
also its introduction technique and technical steps to be applied, such as grinding, calcinating, and firing.
Earlier, Yamazoe (1991) reached the same conclusion. Analyzing the influence of different methods of
Pd loading on SnO2 gas response to H2, Yamozoe (1991) established that the fixation method gives the
best results in comparison with impregnation and colloidal methods. It was found that this effect was
related to the size of PdO particles formed on the surface of SnO2 grains. The finer the PdO particles,
the greater was the promoting effect at a fixed loading.
4.3. SURFACE ADDITIVES AS ACTIVE AND PASSIVE FILTERS

One should also remember that at excessive thickness of catalytically active additives, they can change
their functions, turning into either shunting layers, as shown in Figure 5.36, or active membranes (fil-
SnO2
Figure 5.35. XPS valence-band spectra of tin dioxide nanopowders after sputtering and surface
modification with Au, Pd, and Pt (2%). Tcal = 450°C. A valence band of an unmodified material
treated at 800°C is shown as a reference (Adapted with permission from Cabot et al. 2000.
Me1Ox
Me2Ox
Figure 5.36. Change of film structure and electrical scheme of film conductivity during surface modi-
fication of metal oxides: (a), (b), (c), and (d) correspond to structures with different degrees of surface
covering of base oxide by grains of the second oxide. (Reprinted with permission from Korotcenkov
ters), obstructing the penetration of detecting gas on the surface of the gas-sensing matrix. At certain
conditions this quality can be used to improve the selectivity of the gas sensor (Wang et al. 1998; Barsan
et al. 1999; Frietsch et al. 2000). This concept of filtering the sensor upstream is now used in several
commercial CO and CH4 sensors to avoid the interfering effects of HC and alcohol.
So far, the most often used filters have been passive membranes, having different diffusion pa-
rameters according to the adsorption affinity of the gas molecules on the sieve material and the pore–
molecule size relationship. For example, a simple filter that resists permeation by unwanted molecules
can be fabricated with several protective coatings such as SiO2, Al2O3, zeolites, ZrO2, and Ga2O3, or
with various polymeric membranes. These porous materials should have pore diameter ranging from 0.2
to 1 nm, which is in very good agreement with the general sizes of gaseous molecules: H2 = 0.218 nm,
CO = 0.380 nm, C6H6 = 0.6 nm. However, this type of adsorbent filter may become saturated with a
high concentration of interfering gas (Cabot et al. 2003). The use of such membranes also reduces the
sensor’s general sensitivity. Only during H2 detection, due to the great difference in H2 and O2 diffusion
coefficients in the membrane, was an increase of sensor response observed (Shimuzu et al. 1998).
Selective catalytic reaction mechanisms in the filtering membrane are expected to overcome
these limitation, by means of catalytic conversion of the interfering species into innocuous molecules
(Bernard et al. 2005). High resistance, stability, catalytic activity, and porosity are the best conditions for
realization of such properties of surface layers. For these purposes, either noble metals or metal oxides
may be used. However, the use of metal oxides is preferable. In such systems, one can avoid mutual dif-
fusion influencing the stability of parameters of the base sensitive matrix. In addition, above 30–50 nm,
short-circuiting of the semiconductor film by the metallic film takes place. Therefore, in many cases,
metallic filters are not thick enough for effective filtration.
In order to solve this problem, a protective layer has been developed. The goal for this layer was to
have a very porous layer without any effect on the gases to be detected, but with sufficient efficiency to
stop water drop, metallic particles, and soot. In addition, catalytically active layers, such as Pt, Pd, or Rh
layers on top of a passive membrane (Al2O3 or SiO2) utilize catalytic activity in the conversion of certain
gases as a pretreatment. Reactive gases are burned outside (in the catalytic layer), while the less reac-
tive gases reach and interact with the SnO2 layer (Barsan et al. 1999). The presence of active catalysts
on the surface of the passive membrane can stimulate the dissociation of some gases and affect their
permeability through the membrane as well. Such technical choices promote even higher selectivity of
gas response. However, it is necessary to remember that catalysts derived from noble metals can be
poisoned by many organic and inorganic chemicals that contain sulfur (H2S, SO2, thiols) or phosphorus
(Symons et al. 1992).
It’s necessary to note that requirements for such catalytically active filters, consisting of a passive
membrane and a catalyst, depend on the nature of the detecting gas. For example, Bernard et al. (2005)
established that the most adequate support for NO oxidation should weakly adsorb NO and not store
NO2 in the form of nitrates. Bernard et al. (2005) showed that silica is the ideal candidate for this pur-
pose, and platinum particles deposited on its surface should present a large size to prevent the forma-
tion of platinum oxides. Moreover, the number of Pt sites should be sufficient to favor the reaction
between NO molecules coming from the support and oxygen adsorbed on the Pt. Pt-supported silica
containing 2 wt% of low-dispersed Pt was proposed as an additional layer for a NOx sensor.
The role of various coatings on the surface of metal oxides in gas-sensing effects is summarized
in Table 5.12.
5. IMPROVED OPERATING CHARACTERISTICS OF GAS SENSORS

THROUGH MATERIALS ENGINEERING OF METAL OXIDES: WHAT
DETERMINES THE CHOICE?
As has been clearly established, the choice of method and technical parameters for modifying the prop-
erties of metal oxide films should be based on requirements for the sensor. For example, when develop-
ing gas-sensor technology it is necessary to take into account factors such as the intended application of
Table 5.12. Influence of surface modification on gas-sensing characteristics of

SnO2 and In2O3 sensors
SURFACE ADDITIVE EFFECT ROLE OF ADDITIVE

Noble metals (d < 10 nm) Increased gas response to reducing Catalyst
(Pd, Pt, Au, Rh) gases
Decreased operating temperature
Decreased response time
Noble metals (d > 10 nm) Improved selectivity Catalytically active membrane
Al2O3, SiO2 Improved selectivity Passive membrane
Increased sensitivity to H2
(Al2O3, SiO2) + noble metals Improved selectivity Catalytically active membrane
Transition-metal oxides Improved selectivity Catalytically active membrane
the device; the nature of the gas to be detected; the required rate, sensitivity, and selectivity of the gas
response; environmental conditions of use; and so on.
5.1. DEVICE APPLICATION
At present, designed solid-state gas sensors are intended for use in alarm systems or in meteorological
devices. In the case of semiconductor gas sensors used in meteorological devices, requirements for
stability of the gas-sensing matrix parameters are increasing sharply.
The main reasons for inadequate long-term stability of gas-sensor characteristics is a change in
the properties of the metal oxide, meaning a change in crystallite size, irreversible reaction with the gas
phase, or reaction with the substrate (Meixner and Lampe 1996). In this case, if the requirements for
sensor sensitivity allow, it is desirable not to use additional catalysts, especially noble metals. As research
has shown, this is a sourse of additional instability. As shown by Gopel et al. (1991), the addition of
catalytic metals results in loss of sensitivity over long periods of time due to fouling or cooking of
catalytic metals when exposed to certain gases and vapors. According to Madou and Morrison (1989),
sulfur (as H2S, for example) is a potential poison that can block the catalytic activity of the Pd surface.
Other poisons include many reactive species, for example, chlorine gas. In addition, it was established
that the activity of a catalyst depends on its oxidation state (Kohl, 1990, 2001). After a prolonged time
at the operating temperature, semiconductor gas sensors for combustibles in the absence of target gas
may become “dormant” as a result of the formation of palladium oxide. Only exposure to a combus-
tible gas reduces the palladium and reduces its spillover function. This is a considerable shortcoming of
modified sensors, since uncertainty in display may occur.
As shown by McAleer et al. (1988), metal oxide in modified semiconductor gas sensors has two
functions. The first is its action as a transducer that is sensitive to processes taking place at the surface
of metal clusters; the second is its serving to stabilize the very small catalyst particles necessary for opti-
mal gas-sensing effect. However, at high temperatures, metal oxide support stops executing the second
function, and noble metal clusters obtain increased mobility on the surface of metal oxides. Usually the
clusters move from terraces along preferential directions such as atomic rows to line defects such as
steps or domain boundaries. This process is accompanied by cluster redistribution, which implies size
and distribution density changes due to cluster coalescence (El-Azab et al. 2002). As a rule, the interac-
tion with active gas, especially reducing gas, facilitates this surface reorganization of nanoparticles. So,
with long use, growth of noble-metal clusters is possible. According to results presented by Skala et al.
(1999) and Veltruska et al. (2001), at certain conditions, such as strong surface reduction, it is even pos-
sible for noble-metal clusters encapsulate on the support. This is an additional reason for instability of
the gas-sensing characteristics of modified sensors, associated with noble metals. This problem can be
partly resolved by introducing additional components to stabilize the position of the noble-metal clus-
ters. For example, Holody et al. (2001) showed that Nb introduced into a Pt/SnO2 matrix suppresses
the growth of Pt clusters during annealing at Tan < 900°C. However, even in this case, it was not pos-
sible to completely exclude the growth of Pt clusters.
The presence of a finely dispersed fraction with grain size smaller then 1–5 nm also leads to some
structure instability of the metal oxide matrix in the range of operating temperatures (Т < 600°С), even
for films with an average grain size of more than 50 nm. As one can see in Figure 5.37, small grains first
intergrow with bigger grains. This means that thermal stability of the film structure is controlled by the
presence of a finely dispersed phase that is inclined to coalescence at low temperature, and not by grains
with average size. The bigger the finely dispersed phase is, the stronger are the changes in film structure.
Therefore, production of both metal oxide powders and deposition of films with small dispersion of
grain sizes is one of the most effective methods of structural engineering to improve temporal stability
of solid-state gas sensors. Utilizing thin films also contributes to improved stability of the gas-sensing
matrix. As research has shown (see Figure 5.38), grain size during annealing changes considerably less
in thin films than in thick ones (Korotcenkov et al. 2005).
5.2. THE NATURE OF THE GAS TO BE DETECTED

In designing the technical parameters for preparing a gas-sensing matrix, one should take into account
first of all the reactivity of the gas to be detected, and its tendency to dissociate. For example, ozone
dissociates easily on the surface of metal oxides, and therefore the use of a thin film is preferable for
a solid-state ozone sensor (see Figure 5.39, curve 1). Other gases with high catalytic reactivity may also
combust catalytically at the sensor surface and therefore never reach the interior of the metal oxide
matrix (McAleer et al. 1988).
In addition to reactivity, diffusivity (i.e., the difference in diffusivity through a porous sensor be-
tween the detected gas and oxygen) was found to be another important factor determining the sensor’s
sensitivity and selectivity (Shimuzu et al. 1998). The limited diffusion of O2 in the interior of a thick
film in comparison with that of the detected gas—for example, H2—leads to lower oxygen partial pres-
sure inside the gas-sensing matrix and then to decreased coverage of chemisorbed oxygen on the metal
oxide surface. This should result in increased sensitivity to H2. Experiments conducted by Shimuzu

In2O3
Figure 5.37. Influence of annealing temperature on area of In2O3 film surface occupied by grains of
different sizes (Tpyr = 400°C; d ≈ 50 nm): 1, t = 10–15 nm; 2, t = 20–25 nm; 3, t = 30–35 nm; 4, t =
40–45 nm. (Reprinted with permission from Korotcenkov et al. 2005b. Copyright 2005 Elsevier.)

In2O3
Figure 5.38. Influence of annealing temperature

on average grain size, estimated from AFM mea-
surements: 1, d ≈ 200 nm, Tpyr = 520°C; 2, d ≈
50 nm, Tpyr = 520°C; 3, d ≈ 50 nm, Tpyr = 400°C.
(Reprinted with permission from Korotcenkov
et al. 2005b. Copyright 2005 Elsevier.)
et al. (1998) with thick SnO2 films (d > 100 μm) covered by a SiO2 layer confirmed this assumption.
Thus, control of porosity is one the method of influencing gas-sensor selectivity to H2 and other gases
when H2 is one of the products of dissociation. At corresponding porosity of the gas-sensing matrix,
gas molecules with diameter larger than H2 and O2 will not be able to penetrate into the interior of the
metal oxide film, and therefore the gas sensor will not be sensitive to their appearance in the surround-
ing atmosphere. Apparently, this method can be used only in elaboration of thick and ceramic gas sen-
sors, where the thickness of the gas-sensing matrix exceeds 10–100 μm.

In2O3
Figure 5.39. Dependence of In2O3 gas response

to ozone (1, Toper = 270°C) and H2 (2, Toper = 370°C)
on the thickness of In2O3 films deposited by spray
pyrolysis from InCl3 solutions (Tpyr = 475°C).
(Adapted with permission from Korotcenkov et al.
2004d. Copyright 2004 Elsevier.)
The character of the adsorbed gas molecule bonding with the metal oxide is also an important
factor (Kohl 2001). For example, Lucas et al. (2001) observed that monodentate adsorption on MgO
and CaO is preferred on the edge/corner sites, which is specific to small nanocrystallites (spherulites),
whereas bidentate adsorption is favored by flat planes, which are more prevalent on the larger nano-
crystals. This feature of surface species adsorption indicates that for improved gas-response selectivity,
control of surface roughness and grain shape may be as effective as metal oxide doping using catalytic
active additives.
Depending on the nature of the detecting gas, the requirements for deposition modes of the oxide
used may also change. For example, Korotcenkov et al. (2003a, 2004d) established that, for design of
ozone sensors, it is better to use In2O3 films deposited at high pyrolysis temperatures. However, for
detection of reducing gases, the use of In2O3 films deposited at low pyrolysis temperatures is better.
The nature of the detecting gas can also influence requirements for the gas-sensing surface. For
example, according to Kohl (2001), methane can be activated on a smooth SnO2 surface, where it forms
a CH3 radical. Activation means excitation of a bond and, as a possible consequence, ionization, disso-
ciation, or formation of radicals. H2 molecules are not activated on the smooth SnO2 surfaces of single
crystals. The same effect was observed for WO3 and V2O5. However, on the rough surface of sintered
SnO2, the specimen’s activation and reaction to H2O occurs at 470 K.
5.3. DETECTION MECHANISM
Gas detection reactions may have different mechanisms, depending on the gas-sensing material used and
the gas to be detected. For example, in SnO2-based sensors under standard conditions, the “chemisorp-
tion” mechanism of gas detection is being realized (Brinzari et al. 1999a, 2000a); whereas in In2O3-
based sensors, the “redox” mechanism prevails (Ivanovskaya and Bogdanov 2001; Korotcenkov et al.
2003a). Each mechanism has its own particulars of gas molecule interaction with the metal oxide sur-
face, and therefore its own specific correlation between gas response and film structure. For example,
minimizing grain size is indispensable for achieving of both high gas sensitivity and minimum response
time with SnO2-based gas sensors (Figures 5.40 and 5.41). However, for In2O3-based gas sensors de-
signed to detect reducing gases (Figure 5.39, curve 2), this is not obligatory (Korotcenkov et al. 2003a,
2004c, 2004d). This means that in the last case there is no necessity to achieve minimum grain size, as is
required for ozone detection (Figure 5.42). Therefore temporal and thermal stabilization of grain size
is not a problem.
5.4. ENVIRONMENTAL CONDITIONS DURING USE
The inconstancy of operating conditions, such as changes in air humidity and oxygen concentration
in the surrounding atmosphere, drives the need to decrease the dependence of film parameters on the
state of the surrounding atmosphere, and in many cases determines the selection of sensing materials.
Results of research presented by Korotcenkov et al. (2004c) have shown that the mentioned parameters
are also structure-dependent, and the stability of the film properties may be improved by optimizing

103 SnO2
102
101
Figure 5.40. Influence of SnO2 film thickness on

100
gas response to CO and H2. 1, Tpyr = 430°C; 2, Tpyr =
340°C. (Reprinted with permission from Korotcenkov
et al. 2001b. Copyright 2001 Elsevier.)
the structure. For instance, it was established that an increase of both In2O3 grain size and pyrolysis
temperature during In2O3 deposition (see Figure 5.43) reduces the influence of water vapor on In2O3
film conductivity and, therefore, on gas response.
As shown in Figure 5.43, films deposited at high pyrolysis temperatures show very small differ-
ences in conductivity response to CO in both wet and dry air in comparison with films deposited at low
pyrolysis temperatures. The same influence of pyrolysis temperature on sensitivity to water vapor was
observed for SnO2 sensors (Brinzari et al. 2002b). Results of Fourier transmission infrared (FTIR) mea-
surements also confirmed the higher water sensitivity of low-pyrolysis-temperature SnO2 films. FTIR
spectra demonstrated larger amounts of water species on the surface of these films compared to films
deposited at high Tpyr (Brinzari et al. 2002).

SnO2
Figure 5.41. Influence of SnO2 film thickness on

response time during CO (1, Toper = 340°C) and
CH4 (2, Toper = 430°C) detection. (Adapted with
permission from Korotcenkov et al. 1998. Copyright
1998 SPIE.)
~t2
Figure 5.42. Influence of grain size on response time (1) and gas response (2) of In2O3 films dur-
ing ozone detection. Tpyr = 475°C; Toper = 270°C. (Reprinted with permission from Korotcenkov et al.
2004d. Copyright 2004 Elsevier.)
In this context, a result described by Gopel and Scherbaum (1995) is important: They showed that
epitaxial SnO2 films are less sensitive to changes in air humidity than polycrystalline films constructed
from nanocrystallites.
Emiroglu et al. (2001) confirmed that water adsorption on the surface of metal oxides is structure-
sensitive. They also observed a considerable decrease of adsorbed water concentration on nanoscaled
In2O3
Figure 5.43. Influence of air humidity on In2O3 gas response to 5000 ppm H2. Toper = 370°C; 1, wet
air (35–40% RH); 2, dry air (1–2% RH). (Adapted with permission from Korotcenkov et al. 2003a.
SnO2 powders (grain size ∼7.5 nm) in comparison with commercial Merck SnO2 powders, which are
at least two times larger in size. It is important that this distinction involved not only concentration,
but also where (on what sites) the hydroxyl groups bonded with the SnO2 surface. Because OH groups
participate in reactions of gas detection, and often water adsorption/desorption processes control the
kinetics of sensor response (Michel et al. 1998; Barsan et al. 1999; Korotcenkov et al. 2004b), then, on
the basis of the given results, one can draw an important conclusion: If minimum sensitivity of the
solid-state gas sensor to air humidity is required, the size of the crystallites forming the gas-sensing
matrix can be as large as necessary; that is, there is no need to try for excessive grain size decrease.
The choice of technique for producing crystallites with the needed crystallite faceting can also pro-
vide a means of reducing the influence of humidity on the gas sensor. For example, Golovanov et al.
(2002), using results of a Mulliken population analysis, have shown that chemisorption of OH groups at
the (110) face of SnO2 is accompanied by localization of negative charge there to a greater extent than
in the case of OH groups chemisorbed at the (011) surface. This means that adsorption/desorption
processes and surface reactions with water participation will be different on different SnO2 crystallo-
graphic faces.
It is necessary to note here that the insertion of catalytically active additives into In2O3 and SnO2
metal oxide matrixes usually does not improve the stability of their parameters during a change in air
humidity. For example, from a study of gas-sensing characteristics of In2O3-based sensors modified
with Cu, Zn, B, Ga, P, and Al, it was established that parameters of gas sensors based on undoped In2O3
ceramics and In2O3, doped with Al, were most stable. These results are shown in Figure 5.44. Doping
with other elements basically resulted in strengthening the influence of air humidity on the resistance of
the gas sensors. The same effect was observed in other research as well (Harbeck et al. 2003).
In2O3:Me
Figure 5.44. Sensitivity of In2O3 one-electrode sensos to air humidity in cycles (dry air ↔ wet air),
1, In2O3; 2, Ga2O3 (2%); 3, P2O5 (10%); 4, Al2O3 (4%); 5, ZnO (2%); 6, Ga2O3 (4%); 7, B2O3 (1%); 8,
Cu2O (2%). (Reprinted with permission from Korotcenkov et al. 2004e. Copyright 2004 Elsevier.)

In2O3
103
102
101
Figure 5.45. Experimental dependence of In2O3
gas sensitivity and response time on operating
temperature. Tpyr = 550°C; 0.2 M InCl3 solution;
100 d ≈ 100 nm; ~1 ppm ozone; ~40% RH. (Reprinted
with permission from Korotcenkov 2005c.
5.5. REQUIRED RATE OF SENSOR RESPONSE

The rate of response of solid-state gas sensors is controlled by such processes as water and oxygen
adsorption/desorption, oxygen intercrystallite diffusion, gas diffusion inside the porous structure, and
bulk oxygen diffusion (Barsan et al. 1999; Brinzari et al. 1999a; Korotcenkov et al. 2003b, 2004c). Water
and oxygen adsorption/desorption processes underlie the gas-sensing effect, and opportunities to influ-
ence them are limited. Considerable improvement of the rate of response can be attained only at the
expense of increasing the operating temperatures (see Figure 5.45).
Furthermore, this method has two essential limitations: First, excessive increase of operating tem-
perature may lead to a considerable drop of gas sensitivity; and second, by increasing the working tem-
perature, one may create conditions at which sensor response will be determined by changes in the bulk
properties of the material (see Figure 5.46). As a rule, diffusion processes have large time constants,
and therefore an increase in operating temperature in this case may be accompanied by an increase in
the response time.
With an increase of both film thickness and amount of agglomeration, the influence of diffusion
processes is increased, while sensor response can be reduced sharply (Korotcenkov et al. 1999, 2003,
2004). Therefore, improving porosity and decreasing film thickness—i.e., switching from thick films to
thin ones—is a basic method for improving the kinetics of solid-state gas sensor response. The weak
effect of annealing on grain size in thin films (see Figure 5.38) represents another significant advantage
of using thin films in the design of devices that require grain-size stabilization during long-term use.
Introducing noble metals such as Pd, Pt, or Au (Figures 5.29 and 5.47), which promote dissocia-
tive oxygen adsorption at low temperatures, is also an effective means of shortening both response and
recovery times. In particular, Pd additive in a La2O3-SnO2 matrix (Yamazoe and Miura 1992) and Au in
a CoO3-In2O3 matrix (Yamaura et al. 2000) played important roles in improving the gas response rate
of SnO2- and In2O3-based sensors.
In thick films with developed porosity of the sensitive layer, a deceleration of sensor response rate
is also possible because of accumulation inside the metal oxide matrix of products of catalytic reac-
Figure 5.46. Surface and bulk processes taking place in metal oxides during gas detection, and
their consequences for the properties of polycrystalline materials. (Reprinted with permission from
Korotcenkov 2005c. Copyright 2005 Elsevier.)
tions. As a result, in the presence of fairly reactive catalytically active products of surface reactions, it is
possible for self-heating, temperature destruction, change of phase composition of a layer, and irrevers-
ible sensor degradation to occur, particularly in H2S sensors, in which extremely active components are
formed during oxidation, such as SO2 and SO3.
5.6. REQUIRED SENSITIVITY

Sensitivity and stability are interconnected parameters. As a rule, the smaller the crystallite size, the
higher the sensitivity is. At the same time, however, the stability of the gas-sensor parameters decreases.
It was shown that for SnO2 grains with size about 1–4 nm, grain growth begins at temperatures of
~200–400°С (Korotcenkov et al. 2005b). This is also true for materials that have been modified with
catalytically active additives. As was shown earlier, catalytically active additives, especially noble metals,
can essentially increase sensitivity. However, they may also be one of the reasons for observed time in-
stability of gas-sensor parameters. Therefore, a compromise between high sensitivity and high temporal
and thermal stability is required. It is foreseen that in the future the design of methods for grain-size
stabilization during long-term use of nano-scale devices will gain priority over the design of methods
that produce nano-scale materials with minimal grain size.
As discussed earlier, some studies (Panchapakesan et al. 2001) have considered the possibility of in-
troducing microadditives of various impurities to limit the mobility of adatoms and stabilize grain size.
This is an interesting area of research. It is well known in metallurgy that impurities or second-phase
particles are very effective in altering and inhibiting grain growth. In accordance with results of research
carried out in this field, additives such as K, Ca, Al, or Si salts might prove to be such grain-growth
inhibitors (McAleer et al. 1988; Ihokura and Watson 1994).
Both experimental research and theoretical simulations have established that in the presence of
a second phase, this effect takes place when the average domain size is comparable to the average
interparticle distance. Therefore, controlled addition of impurities to the metal oxide is one way to
achieve the particular grain size, morphology, and structural stability necessary for practical applications.
However, using this approach, it should be borne in mind that even small quantities of gas-sensing ma-
terial additives used for grain-size stabilization can affect the catalytic and adsorption-surface properties.
As a result, though grain size may be stabilized, there may be a large change in both the electrophysical
and the gas-sensing properties of the metal oxides. However, this research is still in the early stages, so
detailed studies of applied metal oxides are needed to develop real technology for the stabilization of
nano-scale materials.
5.7. SENSOR RESPONSE SELECTIVITY

Solid-state gas sensors by nature are not selective. Therefore, it is not realistic to attempt to create a
highly selective sensor, that is, one that is sensitive to only one gas. By modifying the film properties it
In2O3:Pd
Figure 5.47. Influence of In2O3 surface modification with Pd on rate of gas response to reducing
gases. (Reprinted with permission from Korotcenkov 2005c. Copyright 2005 Elsevier.)
is possible only to intensify or inhibit the material’s sensitivity to a particular gas. In previous chapters
many techniques for achieving this goal have been described, so there is no need to discuss these tech-
niques here. One can only remind the reader that, besides changing the operating temperature, there
are two main ways to improve for gas response selectivity. One can use surface membranes (filters), so
that only the gas of interest can reach the sensor; and one can add to the gas-sensing matrix suitable
elements or oxides which selectively catalyzes the particular redox reaction of choice (Morrison 1986;
Madou and Morrison 1989; Kohl 1990; Frietsch et al. 2000). Using these methods, it is possible to create
highly selective sensors for detection of H2, CO, CH4, H2S, and many other gases.
Unfortunately, in spite of a great amount of work in the area of selective catalysis, the catalytic
approach is still highly empirical and usually proprietary. Consequences of metal oxide modification
are unpredictable, and therefore a choice of optimal additives is reached usually by considering many
elements (Yamazoe and Miura 1992; Yamaura et al. 2000). A partial listing of additives examined for
this purpose has been presented by Kohl (1990) and Sberveglieri (1995). It is difficult to make recom-
mendations, but for selective sensor elaboration the best results were obtained while using additives of
transition metals. With Pd and Pt catalysts the gas response selectivity is relatively poor, because they
catalyze the oxidation of most hydrocarbons, CO, and H2 (Madou and Morrison 1989). However, it
is necessary to remember that noble metals in alloying catalysts often yield other properties with well-
expressed selectivity.
A more effective way to attain selective gas detection is to create an electronic “nose” on based on
both an array of sensors set at different operating temperatures and special methods of measurement
(Gardner et al. 1992; Capone et al. 2001; Nakata et al. 2002). However, these approaches overstep the
limits of this survey, and therefore they will not be considered here. Electronic “nose” will be discussed
in detail in other volumes of this series.
5.8. COMPATIBILITY WITH PERIPHERAL MEASURING DEVICES

Thin-film sensors possess maximum rate of response, but they may have very high layer or film resis-
tance, which can complicate their use with measuring equipment. However, we have to admit that the
level of modern electronics development is so high that this limitation may be ignored. A specifically
designed contact geometry can also considerably simplify this problem.
6. SUMMARY
Materials engineering provides great possibilities to improve selectivity, increase sensitivity, decrease
response and recovery times, shift the maximum of sensitivity into the range of lower operation tem-
peratures, and design nanostructured gas sensors with new consumer properties. However, because of
lack of understanding of the nature and mechanism of gas sensitivity, this method remains empirical,
especially for selection of catalytic additives. From analysis of contemporary works devoted to solid-
state gas sensors, our understanding of the nature and mechanism of many processes responsible
for the effectiveness of gas-detection reactions remains at the level of 1986 (Morrison 1986), when J.
Morrison provided his views on the problem of sensitivity and selectivity of semiconductor gas sen-
sors. Therefore, though the addition of a second phase seems to be a feasible approach for achieving
high sensitivity and selectivity, the process of choosing the additive becomes a challenge due to the lack
of basic understanding. It is thus clear that more effort devoted to basic studies for better understand-
ing of the sensing mechanism is an important condition for achieving progress in designing solid-state
gas sensors acceptable for practical use. However, it is necessary to note that the science and technology
of metal oxide gas sensors have advanced considerably in the last decade. The theoretical understand-
ing of the surface processes involved in reactions of gas detection is developing apace as well (Oviedo
and Gillan 2000; Maki-Jaskari and Rantala 2002, 2003), and therefore a detailed understanding of the
elementary processes underlying sensor behavior now seems achievable.
This review has also shown that there is no universal answer to simultaneous optimization of
all sensor parameters. As a rule, an improvement in one parameter is accompanied by worsening of
another. Therefore, one should always seek a compromise among high sensitivity, high stability, and
high selectivity of designed devices (Figure 5.48). This will allow the development of sensors for
specific or focused applications, rather than multipurpose devices targeted at all markets. The last
aspect is important because it can help reduce research and development expenses, elevate the ac-
ceptance level of solid-state sensors, and promote further development of specific sensors for use
in various environments.
It follows that using more than one method of structural engineering and more than one addi-
tive to the metal oxide may yield better results, though at the expense of other effects of combination.
However, it is necessary to admit that success in this direction is unpredictable. It depends greatly
on the developer’s ability to select correctly the necessary combination of additives and methods of
structural engineering for optimal results. We hope that this survey will provide an opportunity to solve
Figure 5.48. Possible future directions for optimization of gas sensor parameters through structural
and chemical modification. (Reprinted with permission from Korotcenkov 2005c. Copyright 2005
Elsevier.)
this problem with greater success. New device concepts and advances in sensor technologies, derived
from improved theoretical understanding of both surface processes and mechanisms of the influence
of metal oxide structure on response give hope that future solid-state gas sensors will be stable, simple,
inexpensive, and reliable devices with high sensitivity, selectivity, and rapid response.
7. ACKNOWLEDGMENTS
G. Korotcenkov thanks the Korean BK21 Program for support of his research.
REFERENCES
Azad A.M. and Hon N.C. (1998) Characterization of BaSnO3-based ceramics. Part 1. Synthesis, processing and
microstructural development. J. Alloys Compounds 270, 95–106.
Barret P. and Dufour L.C. (eds.) (1985) Reactivity of Solids. Elsevier, Amsterdam.
Barsan N., Schweizer-Berberich M., and Gopel W. (1999) Fundamental and practical aspects in the design of nano-
scaled SnO2 gas sensors: a status report. Fresenius J. Anal. Chem. 365, 287–304.
Benard S., Retailleau L., Gaillard F., Vernoux P., and Giroir-Fendler A. (2005) Supported platinum catalysts for
nitrogen oxide sensors. Appl. Catal. B: Environ. 55, 11–21.
Brinzari V., Korotcenkov G., and Schwank J. (1999a) Optimization of thin film gas sensors for environmental
monitoring through theoretical modeling. In: Buettgenbach S. (ed.), Proceedings of SPIE Conference on Chemical
Microsensors and Applications II, vol. 3857, pp.186–197.
Brinzari V., Korotchenkov G., and Dmitriev S. (1999b) Simulation of thin film gas sensor kinetics. Sens. Actuators
B 61(1–3), 143–153.
Brinzari V., Korotchenkov G., and Dmitriev S. (2000a) Theoretical study of semiconductor thin film gas sensitivity:
Attempt to consistent approach. J. Electron. Technol. 33, 225–235.
Brinzari V., Korotcenkov G., Veltruska K., Matolin V., Tsud N., and Schwank J. (2000b) XPS study of gas sensitive
SnO2 thin films. In: Proceedings of Semiconductor International Conference CAS ’2000, 10–14 October 2000, Sinaia,
Romania, vol. 1, pp.127–130.
Brinzari V., Korotcenkov G., and Golovanov V. (2001) Factors influencing the gas sensing characteristics of tin
dioxide films deposited by spray pyrolysis: understanding and possibilities for control. Thin Solid Films 391,
167–175.
Brinzari V., Korotcenkov G., Schwank J., Lantto V., Saukko S., and Golovanov V. (2002a) Morphological rank of
51–58.
Brinzari V., Korotcenkov G., Blinov Y., Cerneavshi A., Tolstoy V., and Ivanov M. (2002b) The role of water ad-
sorption in sensing mechanism of SnO2 gas sensors. In: Proceedings of International Conference on Microelectronics
and Computer Science, October 2002, Chisinau, Moldova, vol. 1, pp. 336–339.
Brinzari V., Korotcenkov G., Schwank J., and Boris Y. (2002c) Chemisorptionoal approach to kinetic analysis of
SnO2:Pd-based thin film gas sensors (TFGS). J. Optoelect. Adv. Mater. (Romania) 4(1), 147–150.
Cabot A., Arbiol J., Morante J., Weimar U., Barsan N., and Gopel W. (2000) Analysis of the noble metal catalytic
additives introduced by impregnation of as obtained SnO2 sol-gel nanocrystals for gas sensors. Sens. Actuators
B 70, 87–100.
Cabot A., Dieguez A., Romano-Rodriguez A., Morante J.R., and Barsan N. (2001) Influence of the catalytic intro-
duction procedure on the nano-SnO2 gas sensor performances. Where and how stay the catalytic atoms? Sens.
Cabot A., Vila A., and J.R.Morante (2002a) Analysis of the catalytic activity and electrical characteristics of different
modified SnO2 layers for gas sensors. Sens. Actuators B 84, 2–20.
Cabot A., Morante J.R., Pratt K.F.E., Chabanis G., and Williams D. (2002b) Selectivity enhancement by electrode
geometry and catalytic additives in In2O3 gas sensors. In: Proceedings of 16th European Conference on Solid State
Transducers, Eurosensors XVI, September 15–18, 2002, Prague, Czech Republic, WP55, pp. 1089–1092.
Cabot A., Arbiol J., Cornet A., Morante J., Che F., and Liu M. (2003) Mesoporous catalytic filters for semiconductor
gas sensors. Thin Solid Films 436, 64–69.
Capone S., Siciliano P., Barsan N., Weimar U., and Vasanelli L. (2001) Analysis of CO and CH4 gas mixtures by
using a micromachined sensor array. Sens. Actuators B 78, 40–48.
Carreno N.L.V., Maciel A.P., Leite E.R., Lisboa-Filho P.N., Longo E., Valentino A., Probst L.E.D., Paiva-Santos
C.O., and Schreiner W.H. (2002) The influence of cations segregation on the metahol decomposition on na-
nostructured SnO2. Sens. Actuators B 86, 185–192.
Chambers S.A. (2000) Epitaxial growth and properties of thin film oxides. Surf. Sci. Rep. 39, 105–180.
Chaudhary V.A., Mulla I.S., and Vijayamohanan K. (1997) Synergetic sensitivity effects in surfcae modified tin oxide
hydrogen sensors using ruthenium and palladium oxides. J. Mater. Sci. Lett. 16, 1819–1821.
Chaudhary V.A., Mulla I.S., Sainkar S.R., Belhekar A.A., and Vijayamohanan K. (1998) Surface rutenated tin oxide
as a novel hydrocarbon sensor. Sens. Actuators A 65, 197–202.
Chaudhary V.A., Mulla I.S., and Vijayamohanan K. (1998) Comparative studies of doped and surface modified tin
oxide towards hydrogen sensing: Synergistic effects of Pd and Ru. Sens. Actuators B 50, 45–51.
Chaudhary V.A., Mulla I.S., and Vijayamohanan K. (1999) Selective hydrogen sensing properties of surface func-
tionalized tin oxide. Sens. Actuators B 55, 154–160.
Choi S.D. and Lee D.D. (2001) CH4 sensing characteristics of K-, Ca-, Mg-impregnated SnO2 sensors. Sens. Actuators
B 77, 335–338.
Comini E., Faglia G., Sberveglieri G. Pan Z., and Wang Z. (2002) Stable and highly sensitive gas sensors based on
semiconducting oxide nanobelts. Appl. Phys. Lett. 81, 1869–1871.
Dai Z.R., Pan Z.W., and Wang Z.L. (2001) Ultra-long single crystalline nanoribbons of tin oxide. Sol. Stat. Commun.
119, 351–354.
Dai Z.R., Gole J.L., Stout J.D., and Wang Z.L. (2002) Tin oxide nanowires, nanoribbons, and nanotubes. Phys. Chem.
B 106, 1274–1279.
Diaz R., Arbiol J., Sanz F., Cornet A., and Morante J.R. (2002) Electroless addition of platinum to SnO2 nanopow-
ders. Chem. Mater. 14, 3277–3283.
Dieguez A., Alay J.L., Kappler J., Romano-Rodriguez A., Vila A., Barsan N., Weimar U., Gopel W., and Morante
J.R. On the stoichiometry, surface composition and structural changes of SnO2 nanoparticles with calcinations
treatments for gas sensor applications (1997) In: Gonsalves K.E., Baraton M.I., and Singh R. (eds.), Surface-
Controlled Nanoscaled Materials for High-Added-Value Applications. MRS Symposium Proceedings, vol. 501, pp.
53–58.
Dufour L.C. and Nowotny J. (eds.) (1988) Surface and Near-Surface Chemistry of Oxide Materials. Elsevier,
Amsterdam.
El-Azab A., Gan S., and Liang Y. (2002) Binding and diffusion of Pt nanoclusters on anatase TiO2(001)-(1×4)
Emiroglu S., Barsan N., Weimar U., and Hoffmann V. (2001) In-situ diffuse reflectance infrared spectroscopy study
of CO adsorption on SnO2. Thin Solid Films 391, 176–185.
Ferroni M., Carotta M.C., Guidi V., Martinelli G., Ronconi F., Sacerdoti M., and Traversa E. (2001) Preparation and
characterization of nanosized titania sensing film. Sens. Actuators B 77, 163–166.
Fliegel W., Behr G., Werner J., and Krabbes G. (1994) Preparation, development of microstructure , electrical and
ga-sensitivr properties of pure and doped SnO2 powders. Sens. Actuators B 18–19, 474–477.
Frietsch M., Zudock F., Goschnick J., and Bruns M. (2000) CuO catalytic membrane as selectively trimmer for
metal oxides gas sensors. Sens. Actuators B 65, 379–381.
Gardner J.W. and Bartlett P.N. (1992). Odour sensors for and electronic nose. In: Gardner J.W. and Bartlett P.N.
(eds.), Sensors and Sensory Systems for Electronic Noses, NATO ASI Series E: Applied Sciences, vol. 212, pp. 31–52.
Kluwer, Dordrecht, The Netherlands.
Gaskov A.M. and Rumyantseva M.N. (2001) Nature of gas sensitivity in nanocrystalline metal oxides. Russ. J. Appl.
Chem. 74(3), 440–444.
Gleiter H. (1992) Materials with ultrafine microstructures: Retrospective and perspective. Nanostruct. Mater. 1, 1–19.
Gleiter H. (2000) Nanostructured materials: Basic concepts and microstructure. Acta Mater. 48, 1–29.
Golovanov V., Korotcenkov G., Brinzari V., Cornet A., Morante J., Arbiol J., and Rossyniol E. (2002) CO-water
interaction with SnO2 gas sensors: role of orientation effects. In: CD Proceedings of 16th International Conference on
Transducers, EUROSENSORS-XVI, September 2002, Prague, Czech Republic, pp. 926–929.
Gopel W., Hese J., and Zemel J.N. (eds.) (1991) Sensors: A. Comprehensive Survey, Chemical and Biochemical Sensors, Part
1, Vol. 2. VCH, Weinheim, Germany.
Gopel W. and Scherbaum K.D. (1995) SnO2 sensors: current status and future prospects. Sens. Actuators B 26–27, 1–2.
Guha P., Chakrabarti S., and Chaudhuri S. (2004) Synthesis of β-Ga2O3 nanowire from elemental Ga metal and its
photoluminescence study. Physica E 23, 81–85.
Ihokura K. and Watson J. (1994) The Stannic Oxide Gas Sensor, Principle and Applications. CRC Press, Boca Raton, FL.
Ivanovskaya M. and Bogdanov P. (2001a) The role of catalytic additives in gas-sensitivity of SnO2-Mo based thin
film sensors. Sens. Actuators B 77, 268–274.
Ivanovskaya M., Bogdanov P., Faglia G. Nelli P., Sberveglieri G., and Taroni A. (2001b) On the role of catalytic
additives in gas-sensitivity of SnO2-Mo based thin film sensors. Sens Actuators B 77, 268–274.
Ivanovskaya M., Gurlo A., and Bogdanov P. (2001c) Mechanism of O3 and NO2 detection and selectivity of In2O3
sensors. Sens. Actuators B 77, 264–267.
Jung S.W., Park W.I., Yi G.C., and Kim M. (2003) Fabrication and controlled magnetic properties of Ni/ZnO
nanorod heterostructures. Adv. Mater. 15(15), 1358–1361.
Harbeck S., Szadvanyi A., Barsan N., Weimar U., and Hoffmann V. (2003) Effects of the Pd dopant on the surface
reactions of SnO2 thick film sensors. In: Proceeding of 2003 Meeting of Electrochemical Society, April–May 2003,
Paris, France, Abstract 2878.
Hartman P. (ed.) (1993) Crystal Growth—An Introduction. North-Holland, Amsterdam.
Henrich V.E. and Cox P.A. (1996) The Surface Science of Metal Oxides, Cambridge University Press, Cambridge, UK.
Holody P.R.J., Soltis R.E., and Hangas J. (2001) Limiting particle growth in platinum/tin oxide nanocomposites.
Scripta Mater. 44, 1821–1824.
Hu J.Q, Ma X.L., Shang N.G., Xie Z.Y., Wong N.B., Lee C.S., and Lee S.T. (2002) Large-scale rapid oxidation syn-
thesis of SnO2 nanoribbons. J. Phys. Chem. B 106, 3823–3826.
Hyodo T., Mori T., Kawahara A., Katsuki H., Shimizu Y., and Egashira M. (2001) Gas sensing properties of semi-
conductor heterolayer sensors fabricated by slide-off transfer printing. Sens. Actuators B 77, 41–47.
Kam K.C., Deepak F.L., Cheetham A.K., and Rao C.N.R. (2004) In2O3 nanowires, nanobouquets and nanotrees.
Chem. Phys. Lett. 397, 329–334.
Kamei M., Enomoto H., and Yasui I. (2001) Origin of the crystalline orientation dependence of the electrical
properties in tin-doped indium oxide films. Thin Solid Films 392, 265–268.
Kanazawa E., Sakai G., Shimanoe K., Kanmura Y., Teraoka Y., Miura N., and Yamazoe N. (2001) Metal oxide semi-
conductor N2O sensor for medical use. Sens. Actuators B 77, 72–77.
Kawamura F., Yasui I., Kamei M., and Sunagawa I. (2001) Habit modifications of SnO2 crystals in SnO2-Cu2O flux
system in the presence of trivalent impurity cations. J. Am. Ceram. Soc. 84, 1134.
Kawamura F., Takahashi T., Yasui I., and Sunagawa I. (2001) Impurity effect on [111] and [110] directions of gro-
wing SnO2 single crystals in SnO2-Cu2O flux system. J. Cryst. Growth 233, 259–268.
Kim D.H., Lee S.H., and Kim K.H. (2001) Comparison of CO-gas sensing characteristics between mono- and
multi-layer Pt/SnO2 thin films. Sens. Actuators B 77, 427–431.
Kim H., Lee J., and Park C. (2002) Surface characterization of O2-plasma-treated indium-tin-oxide (ITO) anodes
for organic light-emitting-device applications. J. Korean Phys. Soc. 41(3), 395–399
Kiv A.E., Maximova T.I., and Soloviev V.N. (2001) MD simulation of the ion-stimulated processes in Si surface
layers. In: Baraton M.I. and Uvarova I.V. (eds.), Functional Gradient Materials and Surface Layers Preparated by Fine
Particles Technology. Kluwer, Dordrecht, The Netherlands, pp. 297–303.
Kohl D. (1990) The role of noble metals in the chemistry of solid-state gas sensors. Sens. Actuators B 1, 158–165.
Kohl D. (2001) Function and applications of gas sensors. J. Phys. D: Appl. Phys. 34, R125–R149.
Kolmakov A., Zhang Y., Cheng G., and Moskovits M. (2003) Detection of CO and O2 using tin oxide nanowire
sensors. Adv. Mater. 15, 997–1000.
Konga Y.Y. and Wang Z.L. (2003) Structures of indium oxide nanobelts. Sol. State Commun. 128, 1–4.
Kong J., Franklin N.R., Zhou C., Chapline M.G., Peng S., Cho K., and Dai H. (2000) Nanotube molecular wires as
chemical sensors. Science 287, 622–625.
Kong X.H., Sun X.M., and Li Y.D. (2003) Synthesis of ZnO nanobelts by carbothermal reduction and their pho-
toluminescence properties. Chem. Lett. 32, 546.
Kong X. and Li Y. (2005) High sensitivity of CuO modified SnO2 nanoribbons to H2S at room temperatures. Sens.
Korotchenkov G., Brynzari V., and Dmitriev S. (1998) Kinetics characteristics of SnO2 thin film gas sensors for
environmentalmonitoring. In: Buettgenbach S. (ed.), Chemical Microsensors and Applications, Proceedings of SPIE,
vol. 3539, pp. 196–204.
Korotchenkov G., Brynzari V., and Dmitriev S. (1999a) SnO2 films for thin film gas sensor design. J. Mater. Sci. Eng.
B 63, 195–204.
Korotchenkov G., Brynzari V., and Dmitriev S. (1999b) Processes development for low cost and low power consu-
ming SnO2 thin film gas sensors (TFGS). Sens. Actuators B 54, 202–209.
Korotcenkov G., DiBattista M., Schwank J., and Brinzari V. (2000) Structural characterization of SnO2 gas sensing
films deposited by spray pyrolysis. J. Mater. Sci. Eng. B 77, 33–39.
Korotcenkov G., Brinzari V., Schwank J., and Cerneavschi A. (2001a) Possibilities of aerosol technology for depo-
sition of SnO2-based films with improved gas sensing characteristics. J. Mater. Sci. Eng. C 19, 73–77.
Korotcenkov G., Brinzari V., DiBattista M., Schwank J., and Vasiliev A. (2001b) Peculiarities of SnO2 thin film
deposition by spray pyrolysis for gas sensor application. Sens. Actuators B 77, 244–252.
Korotcenkov G., Brinzari V., Cerneavschi A., Ivanov M., Cornet A., Morante J., Cabot A., and Arbiol J. (2003a) In2O3
films deposited by spray pyrolysis: gas response to reducing (CO, H2) gases. Sens. Actuators B 98, 236–243.
Korotcenkov G., Macsanov V., Tolstoy V., Brinzari V., Schwank J., and Faglia G. (2003b) Structural and gas re-
sponse characterization of nano-size SnO2 films deposited by SILD method. Sens. Actuators B 96, 602–609.
Korotcenkov G., Brinzari V., Boris Y., Ivanov M., Schwank J., and Morante J. (2003c) Surface Pd doping influence
on gas sensing characteristics of SnO2 thin films deposited by spray pyrolysis. Thin Solid Films 436, 119–126.
Korotcenkov G., Macsanov V., Boris Y., Brinzari V., Tolstoy V., Schwank J., and Morante J. (2003d) Using of SILD
technology for surface modification of SnO2 films for gas sensor applications. In: Kumar A., Merg W.J., and
Cheng Y.T. (eds.), Surface Engineering, Synthesis, Characterization, and Applications. MRS Proceedings, vol. 750,
Y5.25.1–Y.5.25.6.
Korotcenkov G., Macsanov V., Brinzari V., Tolstoy V., Schwank J., Cornet A., and Morante J. (2004a) Influence of
Cu, Fe, Co, and Mn oxide nanoclusters on sensing behavior of SnO2 films. Thin Solid Films 467, 209–214.
Korotcenkov G., Brinzari V., Golovanov V., and Blinov Y. (2004b) Kinetics of gas response to reducing gases of
Korotcenkov G., Brinzari V., Cerneavschi A., Ivanov M., Golovanov V., Cornet, Morante J., Cabot A., and Arbiol J.
(2004c) The influence of film structure on In2O3 gas response. Thin Solid Films 460, 315–323.
Korotcenkov G., Cerneavschi A., Brinzari V., Vasiliev A., Cornet A., Morante J., Cabot A., and Arbiol J. (2004d)
Rodriguez J., and Cirera A. (2004e) Gas sensing characteristics of one-electrode gas sensors on the base of
doped In2O3 ceramics. Sens. Actuators B 103, 13–22.
Korotcenkov G., Cornet A., Rossinyol E., Arbiol J., Brinzari V., and Blinov Y. (2005a) Faceting characterization
310–319.
Korotcenkov G., Brinzari V., Ivanov M., Cerneavschi A., Rodriguez J., Cirera A., Cornet A., and Morante J. (2005b)
Structural stability of In2O3 films deposited by spray pyrolysis during thermal annealing. Thin Solid Films 479,
38–51.
Korotcenkov G. (2005c) Gas response control through structural and chemical modification of metal oxides: State
Korotcenkov G., Golovanov V., Cornet V., Brinzari V., Morante J., and Ivanov M. (2005d) Distinguishing feature
of metal oxide films’ structural engineering for gas sensor application. J. Phys.: Confer. Series (IOP) 15, 256–261.
conductometric-type gas sensors. Mater. Sci. Eng. R. 61, 1–39.
Koudelka-Hep M. (ed.) (2001) Proceedings of the 8th International Meeting on Chemical Sensors, Part 1, 2—July
2000, Basel, Switzerland. Sens. Actuators B 76, 1–673.
Lalauze R., Breuil P., and Pijolat C. (1991) Thin films for gas sensors. Sens. Actuators B 3, 175–182.
Lambert R.M. and Pacchioni G. (eds.) (1996) Chemisorption and Reactivity on Supported Clusters and Thin Films. NATO
ASI Series, vol. 331. Kluwer, London.
Lee D.S., Rue G.H., Huh J.S., Choi S.D., and Lee D.D. (2001) Sensing characteristics of epitaxialy-grown tin oxide
gas sensor on saphire substrates. Sens. Actuators B 77, 90–94.
Lee J.S., Park K, Nahm S., Kim S.W., and Kima S. (2002) Ga2O3 nanomaterials synthesized from ball-milled GaN
powders. Crystal Growth 244, 287–295.
Lee J.H. (2009) Gas sensors using hierarchical and hollow oxide nanostructures: Overview. Sens. Actuators B 140,
319–336.
Li C., Zhang D., Han S., Liu X., Tang T., and Zhou C. (2003) Diameter-controlled growth of single-crystalline
In2O3 nanowires and their electronic properties. Adv. Mater. 15, 143–146.
Li G.J., Zhang X.H., and Kawi S. (1999) Relationships between sensitivity, catalytic activity and surface areas of
SnO2 gas sensors. Sens. Actuators B 60, 64–70.
Li J.Y., Qiao Z.Y., Chen X.L., Chen L., Cao Y.G., He M., Li H., Cao Z.M., and Zhang Z. (2000) Synthesis of
β-Ga2O3 nanorods. J. Alloys Compounds 306, 300–302.
Li Y., Bando Y., and Golberg D. (2003) Single-crystalline In2O3 nanotubes filled with In. Adv. Mater. 15, 581–585.
Li Z.J., Li H.J., Chen X.L., Li L., Xu Y.P., and Li K.Z. (2002) β-Ga2O3 nanowires on unpatterned and patterned
MgO single crystal substrates. J. Alloys Compounds 345, 275–279.
Liang C., Meng G., Lei Y., Phillip F., and Zhang L. (2001) Catalytic growth of semiconducting In2O3 nanofibers.
Adv. Mater. 13, 1330–1333.
Liu Y., Zheng C., Wang W., Zhan Y., and Wang G. (2001) Production of SnO2 nanorods by redox reaction. Crystal
Growth 233, 8–12.
Lucas E., Decker S., Khaleel A., Seitz A., Fultz S., Ponce A., Li W., Carnes C., and Klabunde K. (2001) Nanocrystalline
metal oxides as uniquie chemical reagent/sorbents. Chem. Eur. J. 7(12), 2505–2510.
Madou M.J. and Morrison S.R. (1989) Chemical Sensing with Solid State Devices. Academic Press, San Diego, CA, and
London.
Maiti A., Rodriguez J.A., Law M., Kung P., Mckinney J.R., and Yang P.D. (2003) SnO2 nanoribbons as NO2 sensors:
insights from first principles calculations. Nano Lett. 3, 1025–1029.
Maki-Jaskari M.A. and Rantala T.T. (2003) Density function study of Pd adsorbates at SnO2(110) surfaces. Surf.
Sci. 537, 168–178.
Maki-Jaskari M.A. and Rantala T.T (2002) Theoretical study of oxygen-deficient SnO2(110) surfaces. Phys. Rev. B
65, 245428-1-6.
Mallory G.O. and Hajdu J.B. (1990) Electroless Plating: Fundamentals and Applications. AESF Society, Orlando, FL.
Masel R.I. (1996) Principles of Adsorption and Reaction on Solid Surfaces. Wiley–Interscience, New York.
Matsushima S., Teraoka Y., Miura N., and Yamazoe N. (1988) Electronic interaction between metal additives and
tin dioxide-based gas sensors. Jpn. J. Appl. Phys. 27, 1798.
McAleer J.F., Moseley P.T., Norris J.O.W., Williams D.E., and Tofield B.C. (1988) Tin dioxide gas sensors Part
2—The role of surface additives. J. Chem. Soc., Faraday Trans. 1 84(2), 441–457.
Meixner H. and Lampe U. (1996) Metal oxide sensors. Sens. Actuators B 37, 198–202.
Michel H.J., Leiste H., Schierbaum K.D., and Halbritter J. (1998) Adsorbates and their effects on gas sensing pro-
perties of sputteres SnO2 films. Appl. Surf. Sci. 126, 57–64.
Morrison S.R. (1986) Selectivity in semiconductor sensors. In: Proceedings of the 2nd International Meeting on Chemical
Sensors, Bourdeaux, France, pp. 39–48.
Moseley P.T. and Tofield B.C. (eds.) (1987) Solid State Gas Sensors. Adam Hilgen, Bristol and Philadelphia.
Nakata S., Neya K., and Hashimoto T. (2002) A semiconductor gas sensors based on nonlinearity: Utilization of
the effect of competition on the sensor response to gases mixtures. Electroanalysis 14, 881–887.
Nehasil V., Janecek P., Korotchenkov G., and Matolin V. (2003) Investigation of behavior of Rh deposited onto
polycrystalline SnO2 by means of TPD, AES and EELS. Surf. Sci. 532–535, 415–419.
Nowotny J. (1988) Surface segregation of defects in oxide ceramic materials. Sol. State Ionics 28–30, 1235–1243.
Oviedo J. and Gillan M.J. (2000) Energetics and structure of stoichiometric SnO2 surfaces studied by first principles
calculations. Surf. Sci. 463, 93–-101.
Pagnier T., Boulova M., Galerie A., Gaskov A., and Lucazeau G. (2001) Reactivity of SnO2-CuO nanocrystalline
materials with H2S: a coupled electrical and Raman spectroscopic study. Sens. Actuators B 71, 134–139.
Palatnik L.S., Fuks M.I., and Kosevich V.M. (1972) Mechanism of Formation and Substructure of Condensed Films. Nauka,
Moscow (in Russian).
Pan Z.W., Dai Z.R., and Wang Z.L. (2001) Nanobelts of semiconducting oxides. Science 291, 1947–1949.
Panchapakesan B., De Voe D.L., Widmaier M.R., Cavicchi R., and Semancik S. (2001) Nanoparticle engineering and
control of tin oxide microstructures for chemical microsensor applications. Nanotechnology 12, 336–349.
Rao C.R.K. and Trivedi D.C. (2005) Chemical and electrochemical depositions of platinum group metals and their
applications. Coord. Chem. Rev. 249, 613–631.
Ratko A., Babushkin O., Baran A., and Baran S. (1998) Sorption and gas sensitive properties of In2O3 based cera-
mics doped with Ga2O3. J. Eur. Ceram. Soc. 18, 2227–2232.
Raub E. and Muller K. (1967) Fundamentals of Electrodeposition. Elsevier, Amsterdam.
Rodriguez J.A. (1996) Electronic and chemical properties of Pt, Pd and Ni in bimetallic surfaces. Surf. Sci. 345,
347–362.
Rosenfeld D., Sanjines R., Schreiner W.H., and Levy F. (1993) Gas sensitive and selective SnO2 thin polycrystalline
films doped by ion implantation. Sens. Actuators B 15–16, 406–409.
Sberveglieri G. (ed.) (2001). Proceedings of the Eighth International Meeting on Chemical Sensors, Basel,
Switzerland, 2–5 July 2000. Part 2. Sens. Actuators B 77, 1–591.
Serrini P., Briois V., Horrillo M.C., Traverse A., and Manes L. (1997) Chemical composition and crystalline structure
of SnO2 thin films used as gas sensors. Thin Solid Films 304, 113–122.
Shek C.H., Lai J.K.L., and Lin G.M. (1999) Investigation of interface defects in nanocrystalline SnO2 by positron
annihilation. J. Phys. Chem. Sol. 601, 189–193.
Shimuzu Y., Maekawa T., Nakamura T., and Egashira M. (1998) Effects of gas diffusivity and reactivity on gas
sensing properties of thick film SnO2-based sensors. Sens. Actuators B 46, 163–168.
Shimizu Y., Matsunaga N., Hyodo T., and Egashira M. (2001) Improvement of SO2 sensing properties of WO3 by
noble metal loading. Sens. Actuators B 77, 35–40.
Skala T., Veltruska K., Moroseac M., Matolinova I., Korotcenkov G., and Matolin V. (2003) Study of Pd-In interac-
tion during Pd deposition on pyrolytically prepared In2O3. Appl. Surf. Sci. 205, 196–205.
Sunagawa I. (ed.) (1990) Morphology of Crystals. Terra Science, Tokyo.
Suzuki T. and Yamazaki T. (1990) Effect of annealing on the gas sensitivity of thin oxide ultra-thin films. J. Mater.
Sci. Lett. 9, 750–751.
Symons E.A. (1992) Catalytic gas sensors. In: Sberveglieri G. (ed.), Gas Sensors. Kluwer, Dordrecht, The Netherlands,
pp. 169–185.
Szezuko D., Werner J., Oswald S., Behr G., and Wetzing K. (2001) XPS investigations of surface segregation of
doping elements in SnO2. Appl. Surf. Sci. 179, 301–306.
Takada T., Suzuki K., and Nakane M. (1993) Highly sensitive ozone sensor. Sens. Actuators B 13–14, 40–407.
Tarre A., Rosental A., Aidla A., Aarik J., Sundqvist J., and Harsta A. (2002) New route to SnO2 heteroepitaxy.
Vacuum 67, 571–575.
Tiemann M. (2007) Porous metal oxides as gas sensors. Chem. Eur. J. 13, 8376–8388.
Tong M.S., Dai G.R., Wu Y.D., and Gao D.S. (2000) High sensitivity and switching-like respnse behavior of SnO2-
Ag-SnO2 element to H2S at room temperature. J. Mater. Sci.: Mater. Electron. 11, 661–665.
Tsud N., Johanek V., Stara I., Veltruska K., and Matolin V. (2001) XPS, ISS and TPD study of Pd-Sn interaction on
Pd-SnO2 systems. Thin Solid Films 391, 204–208.
Varela J.A., Cerri J.A., Leite E.R., Longo E., Shamsuzzoha M., and Bradt R.C. (1999) Microstructural evolution
during sintering of CoO doped SnO2 ceramics. Ceram. Int. 25, 253–256.
Varghese O.K., Gong D., Paulose M., Grimes C.A., and Dickey E.C. (2003) Crystallization and high-temperature
structural stability of titanium oxide nanotube arrays. J. Mater. Res. 18, 156–165.
Veltruska K., Tsud N., Brinzari V., Korotcenkov G., and Matolin V. (2001) CO adsorption on Pd clusters deposited
on pyrolytically prepared SnO2 studied by XPS. J. Vacuum 61, 129–134.
Volkenstein Th. (1963) The Electron Theory of Catalysis on Semiconductors. Macmillan, New York.
Wada K. and Egashira M. (1998) Improvement of gas-sensing properties of SnO2 by surface chemical modification
with diethoxydimethylsilane. Sens. Actuators B 53, 147–154.
Wang C.C., Akbar A.A., and Madou M.J. (1998) Ceramic based resistive sensors. J. Electroceram. 2(4), 273–282.
Wang W., Xu C., Wang X., Liu Y., Zhan Y., Zheng C., Song F., and Wang G. (2002) Preparation of SnO2 nanorods
by annealing SnO2 powder in NaCl flux. J. Mater. Chem. 12, 1922–1925.
Williams D.E. and Pratt K.F.E. (1996) Resolving combustible gas mixtures using gas sensitive resistors with arrays
of electrodes. J. Chem. Soc., Faraday. Trans., 92, 4496–4504.
Williams D.E. and Pratt K.F.E. (1998) Classification of reactive sites on the surface of polycrystalline tin dioxide.
J. Chem. Soc., Faraday Trans. 94, 3493–3500.
Williams D.E. (1999) Semiconducting oxides as gas-sensitive resistors. Sens. Actuators B 57, 1–16.
Williams D.E. and Pratt K.F.E. (2000) Microstructure effects on the response of gas-sensitive resistors based on
semiconducting oxides. Sens. Actuators B 70, 214–221.
Xingqin L., Chunhua C., Wendong X., Yusheng S., and Guangyao M. (1993) Study on SnO2-Fe2O3 gas sensing
system by a.c. impedance technique. Sens. Actuators B 17, 1–5.
Xu C., Xu G., Liu Y., Zhao X., and Guanghou W.G. (2002) Preparation and characterization of SnO2 nanorods by
thermal decomposition of SnC2O4 precursor. Scripta Mater. 46, 789–794.
Yamaura H., Jinkawa T., Tamaki J. Moriya K., Miura N., and Yamazoe N. (1996) Indium oxide-based gas sensor for
selective detection of CO. Sens. Actuators B 35–36, 325–332.
Yamaura H., Moriya K., Miura N., and Yamazoe N. (2000) Mechanism of sensitivity promotion in CO sensors
using indium oxide and cobalt oxide. Sens. Actuators B 65, 39–41.
Yamazoe N. (1991) New approaches for improving semiconductor gas sensors. Sens. Actuators B 5, 7–19.
Yamazoe N., Kurokawa Y., and Seiyama T., (1983) Effects of additives on semiconductor gas sensors. Sens. Actuators
4, 283–289.
Yamazoe N. and Miura N. (1992a) Some basic aspects of semiconductor gas sensors. In: Yamauchi S. (ed.), Chemical
Sensors Technology, vol. 4. Kodansha, Tokyo, pp.20–41.
Yamazoe N. and Miura N. (1992b) New approaches in the devices of gas sensors. In: Sberveglieri G. (ed.), Gas
Sensors. Kluwer, Dordrecht, The Netherlands.
Yang G., Haibo Z., and Biying Z. (2000) Monolayer dispersion of oxide additives on SnO2 and their promoting
effect on thermal stability of SnO2 ultrafine particles. J. Mater. Sci. 35, 917–923.
Yoshinobu T., Ecken H., Poghossian A., Luth H., Iwasaki H., and Schoning M.J. (2001) Alternative sensor materials
for light-addressable potentiometric sensors. Sens. Actuators B 76, 388–392.
Zhang J. and Colbow K. (1997) Surface silver clusters as oxidation catalysts on semiconductor gas sensors. Sens.
Zhang D., Li C., Han S., Liu X., Tang T., Jin W., and Zhou C. (2003) Ultraviolet photodetection properties of indi-
um oxide nanowires. Appl. Phys. A 76, 1–4.
Zhang Y., Ago H., Liu J., Yumura M., Uchida K., Ohshima S., Iijima S., Zhu J., and Zhang X. (2004) The synthesis
of In, In2O3 nanowires and In2O3 nanoparticles with shape-controlled. J. Crystal Growth 264, 363–368.
Zhou X.H., Cao Q.X., Huang H., Yang P., and Hu Y. (2003) Study on sensing mechanism of CuO-SnO2 gas sen-
sors. Mater. Sci. Eng. B 99, 44–47.
INDEX
2-D composition, 194 agglomeration, 23, 121, 131, 140, 142, 144, 257,
absorption-based optical sensors, 39 269, 271, 282, 283, 285, 288, 318, 319, 321,
absorption-based sensors, 39 324, 347, 365
absorption fiber optic chemical sensors, 39 Al2O3, 66–68, 71–73, 76, 79, 85, 89, 90, 93, 109,
acetonitrile (ACN), 198–202 111, 112, 143, 188, 196, 218, 229, 233, 238,
ACN. See acetonitrile 239, 242, 243, 254, 260, 271, 282, 283, 325,
acoustic wave propagation, 27, 29 338, 345, 346, 356, 357, 364
acoustic wave sensor, 24, 27, 29, 30 alcohol condensation, 273, 274
active and passive filters, 355 alloy, 83, 131, 161, 217, 220, 231, 241, 242, 351,
active surface area, 131, 134, 139, 140, 142 352
ADFETs. See adsorption-based FETs amperometric sensor, 2–4
adsorption, 9, 17, 20, 24, 32, 70, 71, 75–77, analyte, 2, 3, 8, 9, 17, 21–26, 33, 34, 36–39, 42,
79–85, 88, 90, 94, 96, 98, 105, 112, 119, 120, 44, 64, 74, 84, 111, 127, 177, 180, 182, 184,
121, 124, 129, 132, 135, 136, 186, 259, 260, 186, 192, 194, 197–199, 201
262, 263, 315, 320, 322, 326, 327, 332, 344, aqueous solutions, 259, 260, 262, 272, 290, 351
348, 350, 356, 361, 363, 364, 365, 367 atomic force microscopy (AFM), 32, 174, 327,
adsorption ability, 64, 129 360
adsorption-based FETs (ADFETs), 12 atomization techniques, 248
adsorption/desorption parameters, 79, 80–82, Au nanoparticles, 127, 171, 174
315
adsorption processes, 64, 76, 81 band gap, 9, 69–71, 74, 106–109, 169, 342
adsorption properties, 70, 124, 315 baseline sensor drift, 24
aerosol phase, 245, 246, 248, 249, 252 biological sensors, 32, 51, 169
AFM. See atomic force microscopy bridging oxygen, 103, 317
Ag, 90, 171, 172, 174, 218, 260, 265, 332, 345, bulk stoichiometry, 321
346, 348–350 calcination, 135, 188, 277, 278, 284, 333, 336
agglomerate, 131, 140–142, 144, 219, 227, 229, calorimetric sensors, 48
257, 318, 319, 324, 326 cantilever, 9, 32, 33, 83, 177, 179, 180
agglomerated, 140–142, 257, 258, 278, 285, 318, capacitance sensors, 8, 9
319 carbon-containing species, 320
379
380 • INDEX
carbon nanotubes, 21, 330 chemiluminescence, 33, 34, 35

catalyst, 19, 20, 23, 45, 46, 47, 82, 83, 105, 124, chemisorption, 21, 22, 69, 76, 77, 79, 80, 88,
125, 127, 128, 131, 161, 186, 188, 202, 253, 96–98, 101–103, 121, 124, 238, 315, 317,
260, 267, 274, 276, 291, 314, 320, 332, 337, 318, 344, 361, 364
347, 351, 352, 353, 357, 358, 368 cladding-based chemical fiber-optic sensors, 37
catalytic activity, 64, 70, 82, 84, 87, 88, 101, 105, Clark oxygen sensor, 3
121, 127, 186, 215, 317, 331, 334, 339, 340, CO, 3, 9, 40, 70, 79–84, 87, 96, 98, 103, 111, 127,
349, 350, 356, 358 128, 133, 135, 136, 139, 142, 188, 190, 191,
catalytic activity of sensing materials, 84 238, 323, 327, 339, 341, 349–353, 355, 356,
catalytic bead, 19 362, 368
catalytic sensors, 19 coating design and tooling, 295
cationic site, 75 colorimetric sensors, 39
CBD. See chemical bath deposition combinatorial experimentation, 165
CdS, 78, 271 combinatorial library, 163, 195
CdSe, 165–168 combinatorial materials screening, 160, 184
CFS. See combustion flame synthesis combinatorial screening, 160, 172, 177, 183, 186,
CH4, 40, 96, 121, 339, 350, 355, 356, 362, 368 188, 196, 198, 200, 202
charging agent, 270 combinatorial technologies, 160, 165, 202
chemFET, 1, 12, 15 combustion flame synthesis (CFS), 280, 281,
chemFET-based sensors, 15 283–285
chemical activity, 92, 93, 235, 346 complexant, 260, 261
of metal oxides, 93 composite, 101, 105, 186, 187, 215, 216, 232,
chemical bath deposition (CBD), 259–262, 264, 235, 288, 289, 291, 292, 297
265, 269, 271, 272 conducting polymer, 110, 111, 169, 184, 245, 290,
chemical polymerization, 289, 292, 293 292
chemical precursor, 237, 241, 244, 284, 299 conductivity, 1, 4, 5, 20, 21, 23, 24, 27, 28, 46, 47,
chemical sensitization, 349 48, 64, 70, 74, 79, 80, 82, 92, 95, 96, 97, 98,
chemical sensor, 1, 17, 19, 25, 31, 32, 33, 34, 36, 99, 100, 101, 102, 103, 104, 105, 106, 107,
37, 38, 39, 45, 52, 63, 64, 65, 69, 70, 78, 80, 108, 109, 110, 111, 114, 115, 128, 131, 169,
84, 88, 90, 92, 94, 95, 96, 97, 103, 106, 107, 188, 226, 250, 252, 265, 320, 334, 339, 341,
109, 110, 112, 113, 120, 144, 215, 216, 224, 343, 344, 356, 362
229, 231, 245, 246, 267, 272, 275, 276, 279, conductivity type, 101, 102
288, 289, 292, 294, 298, 301, 327, 328, 329, conductometric sensor, 4, 20–23, 107, 109, 196, 202
330, 332, 336 conjugated organic monomers, 169
chemical stability, 92, 229, 243 corona spray pyrolysis, 250
chemical vapor condensation (CVC), 280, CoTiO3, 191
282–284 counter electrode, 3, 4, 250, 265
chemical vapor deposition (CVD), 189, 235, 236, Cr2-xTixO3, 103
237, 238, 239, 241, 242, 243, 244, 245, 247, cracking effect, 276
253, 282, 289, 292, 294, 295, 298, 299, 300, crystal faceting, 318
314, 322, 326, 346 crystallite, 88, 95, 98, 113, 115, 116, 118, 119,
chemical warfare agents (CWAs), 200 120, 121, 128, 132, 135, 136, 139, 141, 142,
INDEX • 381
186, 219, 226, 227, 229, 254, 256, 257, 269, ED. See electroless deposition
314, 315, 318, 319, 321, 324, 325, 326, 332, EDOC. See electrochemical deposition under
333, 336, 338, 341, 354, 358, 364, 366 oxidizing conditions
crystal shape, 120 electrical energy transduction, 183, 184
crystal structure, 65, 67, 103, 104, 120, 226, 234, electrical energy transduction sensors, 183
241, 283, 313, 338, 341 electrochemical cell, 2–5, 98
CVC. See chemical vapor condensation electrochemical deposition, 259, 264, 267, 270
CVD. See chemical vapor deposition electrochemical deposition under oxidizing
CWAs. See chemical warfare agents conditions (EDOC), 259, 267, 269
electrochemical polymerization, 288–291, 297
DC and RF sputtering, 222 electrochemical sensors, 1–3, 38, 64, 94, 105, 110
Debye length (LD), 114 electroconductivity, 69, 70, 74, 95, 107, 109, 110,
dendrogram, 167 141, 342, 344, 351
deposition from aerosol phase, 245 electroless deposition (ED), 259, 260, 265, 269,
deposition parameters, 226, 232, 239, 240, 254, 272
257, 315, 323 electrolyte, 2–5, 16, 95, 98, 169, 200, 266–268,
deposition process, 16, 124, 217, 218, 223, 231, 290, 297, 298, 300
235, 239, 242, 245, 246, 249, 253, 265, 269, electron-beam evaporation, 216, 234
270, 293, 295 electron configuration, 68, 69, 74, 87
deposition rate, 220–225, 232, 234, 237, 239, 241, electron exchange, 186, 332, 334, 340
242, 252, 253, 264, 267, 270, 284, 294, 298, electronic “nose,” 368
300 electronic properties, 71–73, 77, 78, 96, 186
deposition technology, 121, 215, 293, 297 of metal oxide surfaces, 77
desorption, 77, 79–82, 98, 119–121, 186, 238, electronic sensitization, 349
315, 322, 332, 334, 351, 352, 364, 365 electronic structure, 68, 70, 74, 106, 129, 351
detection mechanism, 38, 116, 361 of metal oxides, 68, 70, 106
device application, 358 electrophoretic deposition (EPD), 270–272, 300
dielectric constant, 8, 9, 43, 106, 112, 114, 186 electrophysical properties, 68, 97, 113, 215
diffusion, 18, 20, 26, 67, 96–98, 100, 103, 120, of sensing materials, 97
124, 131, 132, 138–142, 192, 196, 225, 227, electrostatic spray deposition (ESD), 250–252
233, 236, 237, 239, 245, 252, 261, 313, 335, electrostatic spray pyrolysis. See electrostatic spray
353, 356, 359, 365 deposition
diffusion-layer thickness, 196 environmental conditions, 358, 361
dimension effect, 114, 115 EPD. See electrophoretic deposition
discrete arrays, 165, 166 ESD. See electrostatic spray deposition
dissociation, 18, 44, 75, 89, 105, 124, 138, 221, ethanol (EtOH), 198, 199
236, 260, 345, 357, 360, 361 EtOH. See ethanol
dissociative chemisorption, 124, 317 evaporation rate, 216, 281
doping, 13, 23, 82, 105, 116, 126, 127, 130, 138, 169,
188, 191, 291, 314, 331, 333–342, 361, 364 Fe2O3, 66, 68–70, 72, 73, 78, 85, 87–91, 93, 105,
doping process, 342 108, 111, 131, 238, 254, 260, 277, 331, 345,
droplet size, 246–248 346, 350
382 • INDEX
Fermi level, 9, 77, 78, 109, 130, 315 188, 314–316, 319, 322, 334, 337, 339, 344,
Ferrite plating, 269, 272 354, 367
FET. See field-effect transistor of metal oxides, 314, 315, 316, 348
F (flat) faces, 317 of multicomponent metal oxides, 337
fiber, 34–39, 41, 109, 110, 112, 169–171, 173, gas sensor, 2, 9, 11, 13, 14, 18, 21–24, 27, 31,
272, 273, 292, 294–297 45, 63, 64, 69, 70, 74, 77–79, 81, 82, 84, 87,
fiber Bragg gratings, 39 90, 92, 94–101, 103, 106, 107, 109–113,
fiber coating, 294, 296 116, 119–121, 123, 124, 127, 130, 132, 134,
fiber optic chemical sensor, 36–39 135, 138, 139, 165, 167, 172, 225, 227, 243,
fiber optic sensor, 36, 38 250, 270, 272, 276, 285, 313, 314, 315, 318,
field-effect transistor (FET), 11–15, 51, 94 322, 323, 325, 326, 329, 334, 337–339, 341,
field-effect transistor sensors, 12 343, 348, 351–353, 356, 358–361, 364, 365,
filament thermistor, 48 367–370
film agglomeration, 324 GCS. See grating coupler sensor
film deposition, 128, 216, 218–220, 224, 225, gelation, 276
232, 233, 239, 245, 247, 248, 250, 253, 260, GPC. See gas processing condensation
264, 269, 271, 290, 291, 294, 321, 324, 325, gradient arrays, 166, 192
340 gradient sensor material, 192
by thermal evaporation, 218 grain, 21–23, 88, 97, 98, 100, 107, 113–121,
film microstructure, 233 127–135, 141–144, 188, 219, 225–227, 234,
film texture, 23, 127–129, 256 238, 255, 256, 264, 271, 277, 278, 286, 287,
film thickness influence, 128, 337 314–316, 318, 319, 321, 323–328, 333–338,
finely dispersed fraction, 118, 358 341, 343, 345, 347, 354–356, 358–367
flame ionization detector, 49 grain boundary, 22, 98, 100, 115, 120, 121, 128,
flame ionization sensor, 49, 50 188, 334–336, 338
flammable gas, 19, 140 grain faceting, 121
flexural plate wave (FPW), 25, 31 grain-growth inhibitors, 367
fluorescence-based optical chemical sensor, 38 grain size, 23, 113–121, 129, 131–135, 142–144,
fluorescence fiber optic chemical sensors, 38 225–227, 234, 238, 255, 256, 271, 277, 278,
fluorescent chemical sensors, 34 286, 287, 315, 316, 318, 319, 321, 324, 326–
FPW. See flexural plate wave 328, 333, 334, 336–338, 345, 358–365, 367
grain-size growth, 256, 277
Ga2O3, 19, 66, 71–73, 85, 87, 89, 93, 98, 100, 102, “grains” model, 113, 115
104, 109, 229, 243, 328, 334, 345, 356, 364 grating coupler sensor, (GCS), 35
GaAs, 19, 78, 94, 95, 108, 111, 112 growth directions, 328
GaN, 19, 95, 108
gap plasmons, 177 H2, 5, 70, 76, 81, 83, 87, 88, 90, 111, 117, 119,
gas condensation, 280 124, 133, 134, 137–139, 142, 188, 190,
gas processing condensation (GPC), 280, 283 191, 238, 327, 338–340, 348, 350, 352, 353,
gas-sensing effect, 63, 79, 80, 115, 121, 123, 131, 355–357, 359, 360–363, 368
140, 142, 318, 344, 347, 357, 358, 365 H2S, 3, 92, 105, 238, 341, 345, 350, 357, 358, 366,
gas-sensing properties, 114, 115, 121, 129, 137, 368
INDEX • 383
heat of formation of the most stable oxide, 88 Lamb wave sensors, 30

humidity, 9, 24, 37, 39, 41, 48, 49, 64, 94, 96, 97, 99, Langmuir-Blodgett (LB) films, 51, 52
101–103, 106, 109, 111, 112, 119, 121, 131, Langmuir-Blodgett (LB) sensors, 1, 50
139, 180, 200, 313, 343, 355, 361, 363, 364 Langmuir-Blodgett (LB) technique, 292
hybrid materials, 279, 280 laser ablation, 21, 229, 230, 232, 299, 300
hydrolysis, 97, 117, 135, 237, 242, 260, 261, 264, lattice oxygen, 88, 103, 105, 106, 338, 344
273, 274, 276, 277, 279, 280, 292 LB. See Langmuir-Blodgett
hydroxyl groups, 77, 119, 269, 364 Lewis acidity, 70
Lewis basicity, 71, 76
IBAD. See ion-beam–assisted deposition LFD. See liquid flow deposition
IDTs. See interdigital transducers liquid flow deposition (LFD), 269, 272
In2O3, 66, 72, 73, 74, 85, 89, 93, 96, 98, 101, 102, liquid-phase deposition (LPD), 188, 259, 260,
109, 115–117, 119, 128, 129, 134, 136, 137, 264, 272, 297
139, 188, 190, 225, 226, 229, 238, 239, 243, long-term stability, 25, 64, 91, 94, 95, 97, 99,
254, 255, 260, 265, 271, 276, 277, 325–328, 142, 160, 180, 188, 275, 358
331, 333, 334, 339–345, 351–353, 357, Love-mode acoustic wave sensors, 30
359–365, 367 lower explosive limit (% LEL), 20
indium tin oxide (ITO), 373 LPD. See liquid-phase deposition
initial reduction temperature, 88, 90
InP, 19, 94, 95, 108, 111, 112 magnetron sputtering, 196, 222–224, 245, 252,
interagglomerate contact, 141, 142, 144, 318, 319 296, 298, 324, 326
interdigital transducers (IDTs), 27, 29–31 MAPLE. See matrix-assisted pulsed laser
interdigitated electrode, 8, 23, 186, 187 evaporation
intergrain contact, 115, 315, 343 mass-sensitive chemical sensors, 31
ion-beam–assisted deposition (IBAD), 233–235 material selection, 182
ionic drift, 95 materials engineering, 23, 313, 368
ionosorption, 74 matrix-assisted pulsed laser evaporation
ion-selective electrode (ISE), 5–7 (MAPLE), 293, 294
ion-sensitive field-effect transistor (ISFET), 12, mechanical energy transduction, 177
15, 16 mechanical energy transduction sensors, 177
IR spectroscopy, 39–41, 76, 111 mechanical milling, 124, 142–144, 280, 285–287
ISE. See ion-selective electrode melting temperature, 67, 88, 91, 94, 216, 217,
ISFET. See ion-sensitive field-effect transistor 219, 237, 286, 326, 336
ITO. See indium tin oxide membrane, 2, 3, 5–7, 9, 15, 16, 30, 31, 36, 47,
88, 101, 110, 196, 197, 355–357, 368
Kelvin probe, 10, 11 MEMS. See microelectromechanical systems
kinetics of sensor response, 324, 364 mesoporous materials, 216, 330
K (kinked) face, 334 metal-ion coordination numbers, 65
K-nearest-neighbor (KNN) cluster analysis, 167, metal-organic chemical vapor deposition
168 (MOCVD), 242, 243, 245, 253, 299
KNN. See K-nearest-neighbor cluster analysis metal-organic precursor, 237, 243, 275, 282,
Knudsen cell, 217 283
384 • INDEX
metal oxide, 9, 14, 21, 23, 24, 46, 63–72, 74–79, NO2, 3, 11, 21, 40, 43, 77, 106, 116, 128, 129,
81–90, 93, 95–98, 100–112, 114–116, 118, 135, 142, 188, 190, 266, 268, 269, 327, 357
119, 121, 124–127, 129–131, 134, 135, noble metal, 2, 20, 76, 82, 83, 125–127, 131, 246,
138–140, 142, 144, 186, 188, 190, 196, 202, 261, 263–268, 323, 331, 332, 345–348, 351,
216, 221, 225–227, 229, 232–235, 237, 238, 353–358, 365, 366, 368
242, 243, 245, 246, 250, 253, 255, 258–260, noble-metal surface clustering, 125
262–265, 267, 269, 277, 278, 313–316, non–transition-metal oxides, 68–70
318–325, 328–339, 343–348, 350, 351, 354, nucleophilic attack, 274
356–361, 363–370
composition, 331 OIHM. See organic–inorganic hybrid material
growth, 255 one-dimensional structures, 113, 121, 123, 216,
surface, 65, 75, 77, 79, 84, 127 328–330
metal-oxide semiconductor field-effect transistor operating characteristics, 113, 120
(MOSFET), 12–16 operating temperature, 19, 25, 80, 81, 84, 91, 95,
metathesis polymerization, 288, 291 97–100, 106, 107, 118, 132, 196,315, 323,
microchemical vapor deposition, 243 329, 342, 345, 348, 353, 357, 358, 365, 368
microelectromechanical systems (MEMS), 32 optical sensor, 33–35, 38, 39, 44, 64
microsensors, 9, 16 optode, 37, 38
microwave plasma processing (MPP), 280, 283 organic fluorophore, 165
MOCVD. See metal-organic chemical vapor organic–inorganic hybrid material (OIHM), 279
deposition oxidation, 3, 19, 20, 21, 46, 47, 68, 69, 74, 75, 81,
MOSFET. See metal-oxide semiconductor field- 84, 87, 90, 94, 96, 102, 103, 105, 106, 109,
effect transistor 121, 124, 126–128, 131, 167, 169, 188, 225,
MPP. See microwave plasma processing 227–229, 237, 242, 262, 265, 267, 269, 279,
multicomponent materials, 186, 224, 298, 331 289–291, 334, 338, 339, 344, 345, 349, 352,
multicomponent metal oxides, 333, 336, 337 357, 358, 366, 368
multielectrode array, 188 oxidation potential, 169, 290
oxidizing gases, 94, 101, 102, 106, 116, 144
Nafion, 94, 197–201 oxygen chemisorption, 21, 102, 317, 318
nanobelt, 123, 328–330 oxygen diffusion, 97, 98, 365
nanocomposite, 70, 144, 216, 279, 334, 336, 338, oxygen states, 344
339 oxygen vacancies, 75, 97, 100, 124, 129, 131, 317,
nanocrystal shape, 315 320
nanopowder collection, 285
nanorod, 328 PACVD. See photo-assisted chemical vapor
nanoslit, 177, 178 deposition
nanotube, 21, 144, 216, 328–330 passive membrane, 356, 357
nanowire, 21, 123, 144, 216, 328, 330 PCA. See principal-components analysis
“necks” model, 113, 115 PCD. See photochemical deposition
Nernst equation, 5, 7 Pd, 13, 14, 18, 66, 82, 90, 125–127, 130, 189, 194,
NIR spectroscopy, 39–41 195, 218, 238, 253, 254, 260, 265–268, 271,
NO, 3, 45, 76, 103, 124, 188, 190, 191, 350, 357 332, 345–358, 365, 367, 368
INDEX • 385
PECVD. See plasma-enhanced chemical vapor postdeposition annealing, 326

deposition post–transition-metal oxides, 68, 69, 74
pellistor, 19, 20, 48, 140 potentiometric sensor, 3, 5, 6, 95
peripheral measuring device, 103, 368 powder, 40, 116–118, 132, 133, 135, 136, 139, 140,
phase modification, 332, 334, 343 143, 215, 236, 250, 262, 264, 270–273, 278,
of metal oxides, 331, 332, 343 280–289, 292, 298, 338, 346, 351, 359, 364
photoacoustic sensors, 45 powder technology, 280, 287
photo-assisted chemical vapor deposition PPY. See polypyrrole
(PACVD), 244, 245, 299 pressure of the saturated vapor, 216
photochemical deposition (PCD), 271 pre–transition-metal oxides, 68
photochemical polymerization, 288, 289, 291 principal-components analysis (PCA), 182, 200
photolytic mechanism, 244 Pt, 13, 18–20, 23, 47, 82, 90, 91, 96, 109, 117,
photothermal mechanism, 244 125–127, 133, 138, 139, 189, 192–194, 196,
pH sensor, 2, 6, 272 218, 238, 244, 253, 265–268, 271, 290, 332,
physical sputtering, 220 345–348, 350, 352, 355–358, 365, 368
physical vapor deposition (PVD), 232–235, 239, pulsed laser deposition (PLD), 229–232, 293
294, 296, 298–300 pulsed magnetron sputtering, 223, 224
piezoelectric material, 24, 28 PVD. See physical vapor deposition
plasma-enhanced chemical vapor deposition pyroelectric effect, 46
(PECVD), 241, 242, 299 pyroelectric sensor, 47
plasma polymerization, 288, 291 pyrolysis, 21, 116, 119, 121, 129, 137, 140, 141,
plasmonic nanostructure, 171 237, 242, 243, 246–251, 253–259, 273, 280,
plasticizer, 169, 171, 172, 198, 200–202, 276 283, 288, 289, 292, 293, 298, 300, 315, 316,
PLD. See pulsed laser deposition 321–325, 339–341, 347, 349, 353, 360–362
point defect, 65–68, 97, 98, 107, 315, 326 pyrolysis temperature, 119, 243, 256, 259,
polycrystalline film, 129, 141, 243, 318, 319, 363 321–324, 361, 362
polymer, 8, 9, 21, 24, 30, 34, 36, 37, 47, 91, 92,
94, 97, 108, 110–112, 161, 166–173, 176, QCM. See quartz crystal microbalance
180–183, 185–187, 192, 193, 200, 202, 224, QCM crystal, 25, 26
245, 259, 261, 262, 276, 288–294, 296, 297 quartz crystal microbalance (QCM), 25, 26
polymer film, 9, 94, 166, 186, 187, 276, 288, 292,
293 radiant energy transduction, 165
polymerization, 94, 107, 161, 166, 169, 184, 245, radiant energy transduction sensor, 165
252, 273, 275, 280, 288–293, 297 radio-frequency identification (RFID) sensor,
polymer sensor, 94, 97, 110 196, 198–202
polymer synthesis, 186, 288 rate of sensor response, 365
polymer technology, 288 reaction chemistry, 237
polypyrrole (PPY), 94, 108, 112, 245, 289, 291 reactivity of elements toward oxygen, 90
porosity, 23, 121, 128, 131–135, 137–139, 142, reagent, 37, 38, 41, 82, 92–94, 165, 192–194, 202,
225, 227, 229, 272, 274, 276, 280, 318, 319, 259, 263, 264, 274, 276, 289, 290, 346
324, 334, 338, 341, 345, 356, 360, 365 recovery time, 1, 80, 81, 99, 117, 128, 129, 139,
porosity and active surface area, 131, 134 140, 142, 165, 180, 329, 365, 368
386 • INDEX
“redox” mechanism, 101, 361 181, 183, 185, 186, 188, 191–193, 196–200,
redox reaction, 10, 74, 368 202, 203, 231, 244, 254, 274, 276, 295–297,
reduction, 3, 28, 69, 74, 88–90, 95, 98, 102, 103, 318, 330, 361, 367
120, 121, 125, 127, 129, 135, 186, 224, 237, sensing properties, 21, 70, 114, 115, 121, 129,
263, 267, 269, 270, 276, 281, 317, 332, 333, 137, 188, 276, 314–316, 319, 322, 334, 339,
339, 344, 358 344, 354, 367
reference electrode, 2, 6, 7, 15, 16, 37, 95, 290 sensitivity, 1, 12–14, 16, 18, 21–25, 27, 28, 30,
refractive index, 34–37, 39, 43, 44, 71, 111, 134, 32, 34, 37–40, 44, 46, 63, 64, 78, 79, 81, 84,
172, 176, 296 86–88, 91, 92, 94–96, 98, 99–102, 104–106,
refractometric fiber optic chemical sensors, 39 109–112, 115–117, 119, 124, 125, 127, 129,
refractometric optical sensors, 39 131–134, 137–139, 142, 144, 166, 169,
reothaxial growth and thermal oxidation 172, 177, 183, 185, 186, 188–190, 194, 200,
(RGTO), 126, 131, 227, 229 202, 313–315, 319, 320, 324, 327, 329, 330,
response time, 44, 63, 98–100, 116, 128, 135, 137, 338–342, 344–352, 355–359, 361, 362,
139, 140, 144, 186, 276, 324, 329, 345, 357, 364–366, 368–370
361–363, 365 to air humidity, 102, 119, 131
Rayleigh sensor. See surface acoustic wave (SAW) sensor array, 99, 177, 180–184, 188, 190, 192,
or Rayleigh sensor 196, 199, 244
RFID. See radio-frequency identification sensor sensor parameters, 95, 96, 119, 129, 140, 279,
RF sputtering, 222, 231 280, 328, 366, 369
RGTO. See reothaxial growth and thermal sensor response, 17, 18, 24, 28, 80–84, 86, 88,
oxidation 114–117, 121, 123, 126, 128, 129, 131, 134–
137, 140–144, 192, 193, 196, 198, 200, 201,
Sauerbrey equation, 25, 27 314, 315, 318, 322–324, 327, 337, 340, 342,
SAW. See surface acoustic wave sensor 345, 347, 348, 350, 353, 356, 364, 365, 367
scanning light pulse technique (SLPT), 194, 195 control of, 331, 344
Schottky barrier, 17, 18, 78 selectivity of, 198, 200, 367
Schottky diode, 1, 16–18, 19 sensor stability, 95, 195
Schottky diode–based sensor, 16 serial scanning analysis, 165
screen printing, 21, 166, 292, 300 SERS. See surface-enhanced Raman
Seebeck effect, 45 spectroscopic sensor
selective ion-layer adsorption and reaction SH-APM. See shear-horizontal acoustic plate
(SILAR), 259, 260, 262 mode (SH-APM) sensor
self-lithography, 244 shear-horizontal acoustic plate mode (SH-APM)
semiconducting metal oxides, 96, 100, 186, 196 sensor, 29
sensing film, 11, 141, 167, 177, 188, 192–194, shear-horizontal surface acoustic wave (SH-SAW)
197–202, 244 sensor, 29, 30
sensing layer, 21, 22, 24, 31, 132, 135, 225, 276, SH-SAW. See shear-horizontal surface acoustic
318, 325 wave (SH-SAW) sensor
sensing material, 38, 46, 63, 65, 77, 79, 84, 87, 88, Si, 19, 46, 72, 78, 90, 94, 95, 108, 111, 112, 196,
92, 94, 97, 101, 102, 107–112, 115, 130–134, 242, 243, 274, 279, 325, 332, 346, 347, 367
138, 140, 144, 159–166, 169–172, 177, 180, SiC, 14, 15, 19, 89, 95, 108, 112
INDEX • 387
SILAR. See selective ion-layer adsorption and structural engineering, 119, 314, 319, 321, 322,
reaction 327, 334, 337, 359, 369
SILD. See successive ionic-layer deposition structural properties, 21, 100, 134, 333
sintering, 21, 66, 115, 128, 134–136, 274, 275, successive ionic-layer deposition (SILD), 142,
278, 282, 285, 298, 300, 334, 338 144, 259, 260, 262–264, 272, 319, 349
SLPT. See scanning light pulse technique surface acoustic wave (SAW) or Rayleigh sensor,
SnO2, 11, 23, 66, 68, 69, 72–74, 77–82, 85–87, 89, 25, 27–30, 83, 84, 110, 196, 198, 294
91–93, 95, 96, 98, 101–105, 109, 111, 112, surface additive, 355, 357
115–119, 121–124, 126, 127, 129–144, 188, surface catalysis, 22, 96, 340, 348
189, 196, 225–227, 229, 233, 234, 238, 239, surface charge, 15, 46, 79, 80, 129, 271
242–244, 254, 256–262, 270, 271, 277, 278, surface clustering, 125
282, 315–329, 331, 333–345, 347–355, 357, surface clusters, 125, 349, 350, 353
358, 360–366 surface diffusion, 124, 131, 353
crystallographic planes, 121 surface-enhanced Raman spectroscopic (SERS)
SnO2-CuO, 92, 341, 342 sensor, 171, 172, 174
sol-gel method, 116, 140, 245, 246, 275–278, 333, surface geometry, 23, 123–125
339 surface modification, 23, 103, 125, 231, 245, 264,
sol-gel polymerization, 275 323, 326, 345–352, 354–357, 367
solid electrolytes, 95, 298, 300 of metal oxides, 264, 344, 346, 350, 356
solid-state gas sensor, 63, 64, 78, 81, 87, 90, surface morphology, 124, 172, 234, 241, 354
95–98, 106, 113, 119, 121, 124, 130, 285, surface plasmon, 1, 35, 43, 44, 111, 171, 178
313, 314, 322, 323, 334, 358, 359, 364, 365, surface plasmon resonance (SPR), 1, 43–45, 111,
367–370 171, 177
space-charge region, 115 surface plasmon resonance sensor, 43
specific-ion electrodes, 6 surface plasmon resonance technique, 43
SPR. See surface plasmon resonance surface poisoning, 82, 103
spray nozzle, 246, 248 surface processes, 70, 369, 370
spray pyrolysis, 298, 300 of sensing materials, 77
spray pyrolysis deposition, 255, 341 surface properties, 70, 77, 79, 94, 95, 118, 130,
sputtering, 21, 196, 217, 219–226, 231, 233, 234, 139, 195, 296, 337, 338, 343, 367
245, 252, 291, 293, 296, 298, 300, 314, 324, surface reaction, 23, 83, 88, 98, 103, 121, 124,
326, 347, 355 270, 317, 339, 348, 364, 366
rate, 220, 223, 224 surface segregation, 334–336
techniques/technology, 219, 221, 224, 324 surface solubility, 335
SrFeO3, 100 surface state, 63, 74, 77–80, 121, 130, 186, 315,
stability, 21, 23–25, 44, 64, 69, 76, 78, 88–92, 332, 342, 345, 346
94–97, 99–102, 109, 111, 113, 118, 119, 121, surface stoichiometry, 130, 131, 342, 345, 346
123, 129, 131, 134, 140, 142, 160, 162, 167, disordering, 129
169, 180, 186, 188, 192–195, 200, 229, 243, swelling, 8, 24, 184
270, 271, 275, 293, 313, 314, 323, 326, 328,
329, 336, 338, 345, 346, 348, 353, 356, 358, technology readiness level (TRL), 162, 163
359, 361, 364, 366, 367, 369 temporal drift, 95, 98
388 • INDEX
thermal conductivity, 1, 47, 48 TSM. See thickness shear-mode sensor

thermal conductivity sensor, 47, 48
thermal evaporation, 216–219, 221, 224, 227, ultrasonic agitation, 271
233, 280, 282, 294 ultrasonic atomizer, 246
thermalization, 224, 225 UV absorption, 42
thermally activated CVD, 241, 242, 245 UV irradiation, 94, 173, 291, 293, 346
thermal program reduction (TPR), 88–90 UV spectroscopy, 42
thermal stability, 78, 89, 91, 94, 102, 131, 134,
293, 326, 336, 345, 348, 359, 366 vacuum deposition, 216, 293
thermodynamic stability, 64, 88–90 vacuum evaporation, 216–218, 245, 293, 346
thermoelectric effect, 45 vacuum polymerization, 289, 293
thermoelectric sensor, 45, 46 van der Waals interaction (force), 24, 177, 180
thick film, 118, 119, 138, 139, 188, 290, 359, vapor pressure, 198, 200, 216, 217, 221, 236, 237,
365 281, 291, 298
thick-film sensor, 21, 134, 140
thickness shear mode (TSM) sensor, 25, 27, water adsorption, 119, 124, 363, 364
180–183, 196, 198 water condensation, 139, 273, 274
thin film, 21, 23, 28, 47, 50, 109, 116, 118, 119, WE. See working electrode
123, 141, 166, 196, 197, 219, 221, 229, Wheatstone bridge, 5, 19
231–233, 242, 245, 246, 251, 259, 262, 264, wireless technologies, 196
271–273, 275, 282, 290, 293, 300, 314, 324, WO3, 66, 69, 70, 72, 84, 85, 89, 93, 98, 100–102,
347, 359, 365 108, 111, 112, 115, 124, 260, 277, 331, 361
thin-film sensor, 21, 129, 144, 368 work function, 1, 9, 10, 11, 13, 15, 17, 18, 77, 78,
threshold voltage, 12–16 106, 194, 346
TiO2, 66, 68, 69, 72–76, 78, 85, 87, 89, 90, 93, 95, work-function sensor, 9, 84, 88, 109
98, 102, 104, 108, 111, 112, 125–128, 131, working electrode (WE), 2–4
226, 238, 242, 243, 254, 260, 265, 271, 277,
281, 297, 328, 331, 333, 334, 336 ZnO, 30, 66, 68–70, 72–74, 78, 85, 87, 89, 90, 93,
TPR. See thermal program reduction 102, 109, 111, 112, 127, 128, 130, 131, 225,
transducer, 25, 27, 29, 45, 162, 177, 184, 341, 226, 229, 242, 243, 254, 260, 265, 269, 271,
350, 358 328, 331, 345, 364
transition-metal oxide, 66, 68–70, 74, 87, 106, ZnS, 78
127, 265, 331, 334, 345, 357 ZrO2, 66, 72, 73, 76, 78, 85, 89, 90, 93, 98, 108,
transmittance, 40–42 111, 112, 126, 127, 218, 238, 242, 243, 260,
TRL. See technology readiness level 277, 281, 283, 356
Chemical Sensors
Korotcenkov
Edited by Ghenadii Korotcenkov, Ph.D., Dr. Sci.
Momentum Press is proud to bring you Chemical Sensors Volume 1, the newest addition to The Sensors
Technology Series, edited by Joe Watson. In this long-awaited first volume, Fundamentals of Sensing
Material: General Approaches, you will find a comprehensive introduction to all key materials used for
chemical sensors, beginning with the essential foundations like desired properties, as well as methods of
synthesis, methods of deposition, and methods for modification of sensing materials properties. Inside,
you will find background and guidance on:
CHEMICAL SENSORS
• Basic principles of operation for all major classes of sensors, including electrochemical sensors,
Fundamentals of Sensing Material: General Approaches

capacitance sensors, field-effect transistor sensors, Schottky Diode-based sensors, and Langmuir-
Blodgett Film sensors
• How to attain desired properties for sensors like surface properties, stability properties, electrophysical
properties, and structural properties
• Synthesis methods for sensing materials, including vacuum evaporation and vacuum deposition,
sputtering, chemical vapor deposition , the Sol-Gel process, powder technologies and polymer
technologies
• How to improve the operating characteristics of sensors through metal oxides engineering
Chemical sensors are integral to the automation of a myriad industrial processes, as well as everyday
monitoring of such activities as public safety, testing and monitoring, medical therapeutics, and many
more. This massive reference work will cover all major categories of both the materials used for chemi-
cal sensors and their applications. This will become THE reference work on sensors used for chemical
detection and analysis.
The Chemical Sensors references books will span six volumes and cover in-depth details on both mate-
rials used for chemical sensors and their applications, with the first three volumes exploring the materials
used for chemical sensors — their properties, their behavior, their composition, and even their manufac-
turing and fabrication. Volumes 4 through 6 will explore the great variety of applications for chemical
sensors — from manufacturing and industry to biomedical uses.
About the Editor
Devices in 1976 and his Habilitate Degree (Dr. Sci.) in Physics and Mathematics of Semiconductors and
Dielectrics in 1990. He was for many years the leader in the Gas Sensor Group at the Technical Univer-
sity of Moldova. He is currently a research professor at Gwangju Institute of Science and Technology, in
Gwangju, Republic of Korea. Dr. Korotcenkov is the author of five previous books and has authored over
180 peer-reviewed papers. His research has received numerous awards and honors, including the Award
volume i
of the Supreme Council of Science and Advanced Technology of the Republic of Moldova.
A volume in the Sensor Technology Series

Edited by Joe Watson
Published by Momentum Press®
ISBN: 978-1-60650-103-0
90000
www.momentumpress.net 9 781606 501030

MPB0000026 PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

MPB0000026 PDF

Uploaded by

Copyright:

Available Formats

Chemical Sensors

Fundamentals of Sensing Material: General Approaches

A volume in the Sensor Technology Series

www.momentumpress.net 9 781606 501030

GWANGJU INSTITUTE OF SCIENCE AND TECHNOLOGY

MOMENTUM PRESS, LLC, NEW YORK

First published in 2010 by

ISBN-13: 978-1-60650-103-0 (hard back, case bound)

Cover design by Jonathan Pennell

First Edition: September 2010

Printed in Tawian, ROC

PREFACE TO CHEMICAL SENSORS: FUNDAMENTALS OF SENSING MATERIALS xi

2 DESIRED PROPERTIES FOR SENSING MATERIALS 63

5.4 Electroconductivity 107

3 COMBINATORIAL CONCEPTS FOR DEVELOPMENT OF SENSING MATERIALS 159

4 SYNTHESIS AND DEPOSITION OF SENSOR MATERIALS 215

1 Deposition Technology: Introduction and Overview 215

8.5 Technology of the Pyrolysis Process 253

9 Deposition from Aqueous Solutions 259

10 The Sol-Gel Process 272

11 Powder Technology 280

12 Polymer Technology 288

13 Deposition on Fibers 294

Beongki Cho (Chapter 4)

Ghenadii Korotcenkov (Chapters 1, 2, 4, and 5)

Massood Zandi Atashbar (Chapter 1)

Radislav A. Potyrailo (Chapter 3)

Sridevi Krishnamurthy (Chapter 1)

BASIC PRINCIPLES OF CHEMICAL

gas Gas inlet gas

Working electrode Counter electrode Counter electrode Reference electrode

2.1. AMPEROMETRIC SENSORS

O2 + 2H2O + 2e − → H2O2 + 2OH −

Table 1.1. Reactions at working and counter electrode in amperometric sensors,

WORKING ELECTRODE COUNTER ELECTRODE ELECTRODE

NASICON (Na3Zr2Si2PO12, a sodium super-ionic conductor) as an electrolyte material with a constant

2.2. CONDUCTOMETRIC SENSORS

Figure 1.2. CO2 amperometric transient response.

2.3. POTENTIOMETRIC SENSORS

P(H2)(I) > P(H2)(II)

2.3.1. Ion-Selective Electrodes

Figure 1.6. Illustration of energy levels in an energy-band diagram.

−φMet − q θMet = −φMat − q θMat (1.9)

M = φMet − φMat = −q (θMet − θMat )

Figure 1.7. Setup of a Kelvin probe. A metallic grid oscil-

5. FIELD-EFFECT TRANSISTOR SENSORS

Figure 1.9. Schematic of a four-terminal MOSFET.

Figure 1.11. “Classical” schematic drawing illustrat-

Figure 1.12. Schematic diagram of an ISFET.

7. SCHOTTKY DIODE–BASED SENSORS

Figure 1.16. Structure of a typical conductometric sensor.

10. ACOUSTIC WAVE SENSORS

10.1. THICKNESS SHEAR MODE SENSORS

10.2. SURFACE ACOUSTIC WAVE SENSORS

Figure 1.20. Schematic view of a SAW sensor.

where cm is a mass sensitivity factor.

10.2.1. Shear-Horizontal Surface Acoustic Wave Sensors

10.2.2. Shear-Horizontal Acoustic Plate Mode Sensors

Input IDT Output IDT

Figure 1.21. Schematic view of a SH-APM device structure.

10.2.3. Love-Mode Acoustic Wave Sensors

10.2.4. Lamb Wave Sensors

Interdigital Transducers Guiding Layer Substrate